1 \input texinfo @c -*-texinfo-*-
2 @comment %**start of header
3 @setfilename ../../info/eintr
4 @c setfilename emacs-lisp-intro.info
5 @c sethtmlfilename emacs-lisp-intro.html
6 @settitle Programming in Emacs Lisp
12 @c <<<< For hard copy printing, this file is now
13 @c set for smallbook, which works for all sizes
14 @c of paper, and with PostScript figures >>>>
20 @set print-postscript-figures
22 @c clear print-postscript-figures
25 @comment %**end of header
27 @c per rms and peterb, use 10pt fonts for the main text, mostly to
28 @c save on paper cost.
29 @c Do this inside @tex for now, so current makeinfo does not complain.
35 \global\hbadness=6666 % don't worry about not-too-underfull boxes
38 @set edition-number 3.10
39 @set update-date 28 October 2009
42 ## Summary of shell commands to create various output formats:
44 pushd /usr/local/src/emacs/lispintro/
48 makeinfo --paragraph-indent=0 --verbose emacs-lisp-intro.texi
50 ## ;; (progn (when (bufferp (get-buffer "*info*")) (kill-buffer "*info*")) (info "/usr/local/src/emacs/info/eintr"))
53 texi2dvi emacs-lisp-intro.texi
55 ## xdvi -margins 24pt -topmargin 4pt -offsets 24pt -geometry 760x1140 -s 5 -useTeXpages -mousemode 1 emacs-lisp-intro.dvi &
58 makeinfo --html --no-split --verbose emacs-lisp-intro.texi
60 ## galeon emacs-lisp-intro.html
63 makeinfo --fill-column=70 --no-split --paragraph-indent=0 \
64 --verbose --no-headers --output=emacs-lisp-intro.txt emacs-lisp-intro.texi
70 mtusb # mount -v -t ext3 /dev/sda /mnt
71 cp -v /u/intro/emacs-lisp-intro.texi /mnt/backup/intro/emacs-lisp-intro.texi
72 umtusb # umount -v /mnt
75 ## Other shell commands
77 pushd /usr/local/src/emacs/lispintro/
81 texi2dvi --pdf emacs-lisp-intro.texi
82 # xpdf emacs-lisp-intro.pdf &
84 ## DocBook -- note file extension
85 makeinfo --docbook --no-split --paragraph-indent=0 \
86 --verbose --output=emacs-lisp-intro.docbook emacs-lisp-intro.texi
88 ## XML with a Texinfo DTD -- note file extension
89 makeinfo --xml --no-split --paragraph-indent=0 \
90 --verbose --output=emacs-lisp-intro.texinfoxml emacs-lisp-intro.texi
92 ## PostScript (needs DVI)
93 # gv emacs-lisp-intro.ps &
94 # Create DVI if we lack it
95 # texi2dvi emacs-lisp-intro.texi
96 dvips emacs-lisp-intro.dvi -o emacs-lisp-intro.ps
99 # Use OpenOffice to view RTF
100 # Create HTML if we lack it
101 # makeinfo --no-split --html emacs-lisp-intro.texi
102 /usr/local/src/html2rtf.pl emacs-lisp-intro.html
105 /usr/bin/rtf2latex emacs-lisp-intro.rtf
111 @c ================ Included Figures ================
113 @c Set print-postscript-figures if you print PostScript figures.
114 @c If you clear this, the ten figures will be printed as ASCII diagrams.
115 @c (This is not relevant to Info, since Info only handles ASCII.)
116 @c Your site may require editing changes to print PostScript; in this
117 @c case, search for `print-postscript-figures' and make appropriate changes.
119 @c ================ How to Create an Info file ================
121 @c If you have `makeinfo' installed, run the following command
123 @c makeinfo emacs-lisp-intro.texi
125 @c or, if you want a single, large Info file, and no paragraph indents:
126 @c makeinfo --no-split --paragraph-indent=0 --verbose emacs-lisp-intro.texi
128 @c After creating the Info file, edit your Info `dir' file, if the
129 @c `dircategory' section below does not enable your system to
130 @c install the manual automatically.
131 @c (The `dir' file is often in the `/usr/local/share/info/' directory.)
133 @c ================ How to Create an HTML file ================
135 @c To convert to HTML format
136 @c makeinfo --html --no-split --verbose emacs-lisp-intro.texi
138 @c ================ How to Print a Book in Various Sizes ================
140 @c This book can be printed in any of three different sizes.
141 @c In the above header, set @-commands appropriately.
151 @c European A4 size paper:
156 @c ================ How to Typeset and Print ================
158 @c If you do not include PostScript figures, run either of the
159 @c following command sequences, or similar commands suited to your
162 @c texi2dvi emacs-lisp-intro.texi
163 @c lpr -d emacs-lisp-intro.dvi
167 @c tex emacs-lisp-intro.texi
168 @c texindex emacs-lisp-intro.??
169 @c tex emacs-lisp-intro.texi
170 @c lpr -d emacs-lisp-intro.dvi
172 @c If you include the PostScript figures, and you have old software,
173 @c you may need to convert the .dvi file to a .ps file before
174 @c printing. Run either of the following command sequences, or one
177 @c dvips -f < emacs-lisp-intro.dvi > emacs-lisp-intro.ps
181 @c postscript -p < emacs-lisp-intro.dvi > emacs-lisp-intro.ps
184 @c (Note: if you edit the book so as to change the length of the
185 @c table of contents, you may have to change the value of `pageno' below.)
187 @c ================ End of Formatting Sections ================
189 @c For next or subsequent edition:
190 @c create function using with-output-to-temp-buffer
191 @c create a major mode, with keymaps
192 @c run an asynchronous process, like grep or diff
194 @c For 8.5 by 11 inch format: do not use such a small amount of
195 @c whitespace between paragraphs as smallbook format
198 \global\parskip 6pt plus 1pt
202 @c For all sized formats: print within-book cross
203 @c reference with ``...'' rather than [...]
205 @c This works with the texinfo.tex file, version 2003-05-04.08,
206 @c in the Texinfo version 4.6 of the 2003 Jun 13 distribution.
209 \if \xrefprintnodename
210 \global\def\xrefprintnodename#1{\unskip, ``#1''}
212 \global\def\xrefprintnodename#1{ ``#1''}
214 % \global\def\xrefprintnodename#1{, ``#1''}
217 @c ----------------------------------------------------
219 @dircategory GNU Emacs Lisp
221 * Emacs Lisp Intro: (eintr).
222 A simple introduction to Emacs Lisp programming.
226 This is an @cite{Introduction to Programming in Emacs Lisp}, for
227 people who are not programmers.
229 Edition @value{edition-number}, @value{update-date}
231 Copyright @copyright{} 1990--1995, 1997, 2001--2013 Free Software
238 GNU Press, @hfill @uref{http://www.fsf.org/campaigns/gnu-press/}@*
239 a division of the @hfill email: @email{sales@@fsf.org}@*
240 Free Software Foundation, Inc. @hfill Tel: +1 (617) 542-5942@*
241 51 Franklin Street, Fifth Floor @hfill Fax: +1 (617) 542-2652@*
242 Boston, MA 02110-1301 USA
249 GNU Press, http://www.fsf.org/campaigns/gnu-press/
250 a division of the email: sales@@fsf.org
251 Free Software Foundation, Inc. Tel: +1 (617) 542-5942
252 51 Franklin Street, Fifth Floor Fax: +1 (617) 542-2652
253 Boston, MA 02110-1301 USA
258 @c Printed copies are available from @uref{http://shop.fsf.org/} for $35 each.@*
261 Permission is granted to copy, distribute and/or modify this document
262 under the terms of the GNU Free Documentation License, Version 1.3 or
263 any later version published by the Free Software Foundation; there
264 being no Invariant Section, with the Front-Cover Texts being ``A GNU
265 Manual'', and with the Back-Cover Texts as in (a) below. A copy of
266 the license is included in the section entitled ``GNU Free
267 Documentation License''.
269 (a) The FSF's Back-Cover Text is: ``You have the freedom to
270 copy and modify this GNU manual. Buying copies from the FSF
271 supports it in developing GNU and promoting software freedom.''
274 @c half title; two lines here, so do not use `shorttitlepage'
277 \hbox{}\vskip 1.5in \chaprm \centerline{An Introduction to}%
279 {\begingroup\hbox{}\vskip 0.25in \chaprm%
280 \centerline{Programming in Emacs Lisp}%
281 \endgroup\page\hbox{}\page}
286 @center @titlefont{An Introduction to}
288 @center @titlefont{Programming in Emacs Lisp}
290 @center Revised Third Edition
292 @center by Robert J. Chassell
295 @vskip 0pt plus 1filll
301 @evenheading @thispage @| @| @thischapter
302 @oddheading @thissection @| @| @thispage
306 @c Keep T.O.C. short by tightening up for largebook
309 \global\parskip 2pt plus 1pt
310 \global\advance\baselineskip by -1pt
320 @top An Introduction to Programming in Emacs Lisp
324 <p>The homepage for GNU Emacs is at
325 <a href="/software/emacs/">http://www.gnu.org/software/emacs/</a>.<br>
326 To view this manual in other formats, click
327 <a href="/software/emacs/manual/eintr.html">here</a>.
333 This master menu first lists each chapter and index; then it lists
334 every node in every chapter.
337 @c >>>> Set pageno appropriately <<<<
339 @c The first page of the Preface is a roman numeral; it is the first
340 @c right handed page after the Table of Contents; hence the following
341 @c setting must be for an odd negative number.
344 @c global@pageno = -11
347 @set COUNT-WORDS count-words-example
348 @c Length of variable name chosen so that things still line up when expanded.
351 * Preface:: What to look for.
352 * List Processing:: What is Lisp?
353 * Practicing Evaluation:: Running several programs.
354 * Writing Defuns:: How to write function definitions.
355 * Buffer Walk Through:: Exploring a few buffer-related functions.
356 * More Complex:: A few, even more complex functions.
357 * Narrowing & Widening:: Restricting your and Emacs attention to
359 * car cdr & cons:: Fundamental functions in Lisp.
360 * Cutting & Storing Text:: Removing text and saving it.
361 * List Implementation:: How lists are implemented in the computer.
362 * Yanking:: Pasting stored text.
363 * Loops & Recursion:: How to repeat a process.
364 * Regexp Search:: Regular expression searches.
365 * Counting Words:: A review of repetition and regexps.
366 * Words in a defun:: Counting words in a @code{defun}.
367 * Readying a Graph:: A prototype graph printing function.
368 * Emacs Initialization:: How to write a @file{.emacs} file.
369 * Debugging:: How to run the Emacs Lisp debuggers.
370 * Conclusion:: Now you have the basics.
371 * the-the:: An appendix: how to find reduplicated words.
372 * Kill Ring:: An appendix: how the kill ring works.
373 * Full Graph:: How to create a graph with labeled axes.
374 * Free Software and Free Manuals::
375 * GNU Free Documentation License::
380 --- The Detailed Node Listing ---
384 * Why:: Why learn Emacs Lisp?
385 * On Reading this Text:: Read, gain familiarity, pick up habits....
386 * Who You Are:: For whom this is written.
388 * Note for Novices:: You can read this as a novice.
393 * Lisp Lists:: What are lists?
394 * Run a Program:: Any list in Lisp is a program ready to run.
395 * Making Errors:: Generating an error message.
396 * Names & Definitions:: Names of symbols and function definitions.
397 * Lisp Interpreter:: What the Lisp interpreter does.
398 * Evaluation:: Running a program.
399 * Variables:: Returning a value from a variable.
400 * Arguments:: Passing information to a function.
401 * set & setq:: Setting the value of a variable.
402 * Summary:: The major points.
403 * Error Message Exercises::
407 * Numbers Lists:: List have numbers, other lists, in them.
408 * Lisp Atoms:: Elemental entities.
409 * Whitespace in Lists:: Formatting lists to be readable.
410 * Typing Lists:: How GNU Emacs helps you type lists.
414 * Complications:: Variables, Special forms, Lists within.
415 * Byte Compiling:: Specially processing code for speed.
419 * How the Interpreter Acts:: Returns and Side Effects...
420 * Evaluating Inner Lists:: Lists within lists...
424 * fill-column Example::
425 * Void Function:: The error message for a symbol
427 * Void Variable:: The error message for a symbol without a value.
431 * Data types:: Types of data passed to a function.
432 * Args as Variable or List:: An argument can be the value
433 of a variable or list.
434 * Variable Number of Arguments:: Some functions may take a
435 variable number of arguments.
436 * Wrong Type of Argument:: Passing an argument of the wrong type
438 * message:: A useful function for sending messages.
440 Setting the Value of a Variable
442 * Using set:: Setting values.
443 * Using setq:: Setting a quoted value.
444 * Counting:: Using @code{setq} to count.
446 Practicing Evaluation
448 * How to Evaluate:: Typing editing commands or @kbd{C-x C-e}
450 * Buffer Names:: Buffers and files are different.
451 * Getting Buffers:: Getting a buffer itself, not merely its name.
452 * Switching Buffers:: How to change to another buffer.
453 * Buffer Size & Locations:: Where point is located and the size of
455 * Evaluation Exercise::
457 How To Write Function Definitions
459 * Primitive Functions::
460 * defun:: The @code{defun} special form.
461 * Install:: Install a function definition.
462 * Interactive:: Making a function interactive.
463 * Interactive Options:: Different options for @code{interactive}.
464 * Permanent Installation:: Installing code permanently.
465 * let:: Creating and initializing local variables.
467 * else:: If--then--else expressions.
468 * Truth & Falsehood:: What Lisp considers false and true.
469 * save-excursion:: Keeping track of point, mark, and buffer.
473 Install a Function Definition
475 * Effect of installation::
476 * Change a defun:: How to change a function definition.
478 Make a Function Interactive
480 * Interactive multiply-by-seven:: An overview.
481 * multiply-by-seven in detail:: The interactive version.
485 * Prevent confusion::
486 * Parts of let Expression::
487 * Sample let Expression::
488 * Uninitialized let Variables::
490 The @code{if} Special Form
492 * if in more detail::
493 * type-of-animal in detail:: An example of an @code{if} expression.
495 Truth and Falsehood in Emacs Lisp
497 * nil explained:: @code{nil} has two meanings.
499 @code{save-excursion}
501 * Point and mark:: A review of various locations.
502 * Template for save-excursion::
504 A Few Buffer--Related Functions
506 * Finding More:: How to find more information.
507 * simplified-beginning-of-buffer:: Shows @code{goto-char},
508 @code{point-min}, and @code{push-mark}.
509 * mark-whole-buffer:: Almost the same as @code{beginning-of-buffer}.
510 * append-to-buffer:: Uses @code{save-excursion} and
511 @code{insert-buffer-substring}.
512 * Buffer Related Review:: Review.
515 The Definition of @code{mark-whole-buffer}
517 * mark-whole-buffer overview::
518 * Body of mark-whole-buffer:: Only three lines of code.
520 The Definition of @code{append-to-buffer}
522 * append-to-buffer overview::
523 * append interactive:: A two part interactive expression.
524 * append-to-buffer body:: Incorporates a @code{let} expression.
525 * append save-excursion:: How the @code{save-excursion} works.
527 A Few More Complex Functions
529 * copy-to-buffer:: With @code{set-buffer}, @code{get-buffer-create}.
530 * insert-buffer:: Read-only, and with @code{or}.
531 * beginning-of-buffer:: Shows @code{goto-char},
532 @code{point-min}, and @code{push-mark}.
533 * Second Buffer Related Review::
534 * optional Exercise::
536 The Definition of @code{insert-buffer}
538 * insert-buffer code::
539 * insert-buffer interactive:: When you can read, but not write.
540 * insert-buffer body:: The body has an @code{or} and a @code{let}.
541 * if & or:: Using an @code{if} instead of an @code{or}.
542 * Insert or:: How the @code{or} expression works.
543 * Insert let:: Two @code{save-excursion} expressions.
544 * New insert-buffer::
546 The Interactive Expression in @code{insert-buffer}
548 * Read-only buffer:: When a buffer cannot be modified.
549 * b for interactive:: An existing buffer or else its name.
551 Complete Definition of @code{beginning-of-buffer}
553 * Optional Arguments::
554 * beginning-of-buffer opt arg:: Example with optional argument.
555 * beginning-of-buffer complete::
557 @code{beginning-of-buffer} with an Argument
559 * Disentangle beginning-of-buffer::
560 * Large buffer case::
561 * Small buffer case::
563 Narrowing and Widening
565 * Narrowing advantages:: The advantages of narrowing
566 * save-restriction:: The @code{save-restriction} special form.
567 * what-line:: The number of the line that point is on.
570 @code{car}, @code{cdr}, @code{cons}: Fundamental Functions
572 * Strange Names:: An historical aside: why the strange names?
573 * car & cdr:: Functions for extracting part of a list.
574 * cons:: Constructing a list.
575 * nthcdr:: Calling @code{cdr} repeatedly.
577 * setcar:: Changing the first element of a list.
578 * setcdr:: Changing the rest of a list.
584 * length:: How to find the length of a list.
586 Cutting and Storing Text
588 * Storing Text:: Text is stored in a list.
589 * zap-to-char:: Cutting out text up to a character.
590 * kill-region:: Cutting text out of a region.
591 * copy-region-as-kill:: A definition for copying text.
592 * Digression into C:: Minor note on C programming language macros.
593 * defvar:: How to give a variable an initial value.
594 * cons & search-fwd Review::
599 * Complete zap-to-char:: The complete implementation.
600 * zap-to-char interactive:: A three part interactive expression.
601 * zap-to-char body:: A short overview.
602 * search-forward:: How to search for a string.
603 * progn:: The @code{progn} special form.
604 * Summing up zap-to-char:: Using @code{point} and @code{search-forward}.
608 * Complete kill-region:: The function definition.
609 * condition-case:: Dealing with a problem.
612 @code{copy-region-as-kill}
614 * Complete copy-region-as-kill:: The complete function definition.
615 * copy-region-as-kill body:: The body of @code{copy-region-as-kill}.
617 The Body of @code{copy-region-as-kill}
619 * last-command & this-command::
620 * kill-append function::
621 * kill-new function::
623 Initializing a Variable with @code{defvar}
625 * See variable current value::
626 * defvar and asterisk::
628 How Lists are Implemented
631 * Symbols as Chest:: Exploring a powerful metaphor.
636 * Kill Ring Overview::
637 * kill-ring-yank-pointer:: The kill ring is a list.
638 * yank nthcdr Exercises:: The @code{kill-ring-yank-pointer} variable.
642 * while:: Causing a stretch of code to repeat.
644 * Recursion:: Causing a function to call itself.
649 * Looping with while:: Repeat so long as test returns true.
650 * Loop Example:: A @code{while} loop that uses a list.
651 * print-elements-of-list:: Uses @code{while}, @code{car}, @code{cdr}.
652 * Incrementing Loop:: A loop with an incrementing counter.
653 * Incrementing Loop Details::
654 * Decrementing Loop:: A loop with a decrementing counter.
656 Details of an Incrementing Loop
658 * Incrementing Example:: Counting pebbles in a triangle.
659 * Inc Example parts:: The parts of the function definition.
660 * Inc Example altogether:: Putting the function definition together.
662 Loop with a Decrementing Counter
664 * Decrementing Example:: More pebbles on the beach.
665 * Dec Example parts:: The parts of the function definition.
666 * Dec Example altogether:: Putting the function definition together.
668 Save your time: @code{dolist} and @code{dotimes}
675 * Building Robots:: Same model, different serial number ...
676 * Recursive Definition Parts:: Walk until you stop ...
677 * Recursion with list:: Using a list as the test whether to recurse.
678 * Recursive triangle function::
679 * Recursion with cond::
680 * Recursive Patterns:: Often used templates.
681 * No Deferment:: Don't store up work ...
682 * No deferment solution::
684 Recursion in Place of a Counter
686 * Recursive Example arg of 1 or 2::
687 * Recursive Example arg of 3 or 4::
695 Regular Expression Searches
697 * sentence-end:: The regular expression for @code{sentence-end}.
698 * re-search-forward:: Very similar to @code{search-forward}.
699 * forward-sentence:: A straightforward example of regexp search.
700 * forward-paragraph:: A somewhat complex example.
701 * etags:: How to create your own @file{TAGS} table.
703 * re-search Exercises::
705 @code{forward-sentence}
707 * Complete forward-sentence::
708 * fwd-sentence while loops:: Two @code{while} loops.
709 * fwd-sentence re-search:: A regular expression search.
711 @code{forward-paragraph}: a Goldmine of Functions
713 * forward-paragraph in brief:: Key parts of the function definition.
714 * fwd-para let:: The @code{let*} expression.
715 * fwd-para while:: The forward motion @code{while} loop.
717 Counting: Repetition and Regexps
720 * @value{COUNT-WORDS}:: Use a regexp, but find a problem.
721 * recursive-count-words:: Start with case of no words in region.
722 * Counting Exercise::
724 The @code{@value{COUNT-WORDS}} Function
726 * Design @value{COUNT-WORDS}:: The definition using a @code{while} loop.
727 * Whitespace Bug:: The Whitespace Bug in @code{@value{COUNT-WORDS}}.
729 Counting Words in a @code{defun}
731 * Divide and Conquer::
732 * Words and Symbols:: What to count?
733 * Syntax:: What constitutes a word or symbol?
734 * count-words-in-defun:: Very like @code{@value{COUNT-WORDS}}.
735 * Several defuns:: Counting several defuns in a file.
736 * Find a File:: Do you want to look at a file?
737 * lengths-list-file:: A list of the lengths of many definitions.
738 * Several files:: Counting in definitions in different files.
739 * Several files recursively:: Recursively counting in different files.
740 * Prepare the data:: Prepare the data for display in a graph.
742 Count Words in @code{defuns} in Different Files
744 * lengths-list-many-files:: Return a list of the lengths of defuns.
745 * append:: Attach one list to another.
747 Prepare the Data for Display in a Graph
749 * Data for Display in Detail::
750 * Sorting:: Sorting lists.
751 * Files List:: Making a list of files.
752 * Counting function definitions::
756 * Columns of a graph::
757 * graph-body-print:: How to print the body of a graph.
758 * recursive-graph-body-print::
760 * Line Graph Exercise::
762 Your @file{.emacs} File
764 * Default Configuration::
765 * Site-wide Init:: You can write site-wide init files.
766 * defcustom:: Emacs will write code for you.
767 * Beginning a .emacs File:: How to write a @code{.emacs file}.
768 * Text and Auto-fill:: Automatically wrap lines.
769 * Mail Aliases:: Use abbreviations for email addresses.
770 * Indent Tabs Mode:: Don't use tabs with @TeX{}
771 * Keybindings:: Create some personal keybindings.
772 * Keymaps:: More about key binding.
773 * Loading Files:: Load (i.e., evaluate) files automatically.
774 * Autoload:: Make functions available.
775 * Simple Extension:: Define a function; bind it to a key.
776 * X11 Colors:: Colors in X.
778 * Mode Line:: How to customize your mode line.
782 * debug:: How to use the built-in debugger.
783 * debug-on-entry:: Start debugging when you call a function.
784 * debug-on-quit:: Start debugging when you quit with @kbd{C-g}.
785 * edebug:: How to use Edebug, a source level debugger.
786 * Debugging Exercises::
788 Handling the Kill Ring
790 * What the Kill Ring Does::
792 * yank:: Paste a copy of a clipped element.
793 * yank-pop:: Insert element pointed to.
796 The @code{current-kill} Function
798 * Code for current-kill::
799 * Understanding current-kill::
801 @code{current-kill} in Outline
803 * Body of current-kill::
804 * Digression concerning error:: How to mislead humans, but not computers.
805 * Determining the Element::
807 A Graph with Labeled Axes
810 * print-graph Varlist:: @code{let} expression in @code{print-graph}.
811 * print-Y-axis:: Print a label for the vertical axis.
812 * print-X-axis:: Print a horizontal label.
813 * Print Whole Graph:: The function to print a complete graph.
815 The @code{print-Y-axis} Function
817 * print-Y-axis in Detail::
818 * Height of label:: What height for the Y axis?
819 * Compute a Remainder:: How to compute the remainder of a division.
820 * Y Axis Element:: Construct a line for the Y axis.
821 * Y-axis-column:: Generate a list of Y axis labels.
822 * print-Y-axis Penultimate:: A not quite final version.
824 The @code{print-X-axis} Function
826 * Similarities differences:: Much like @code{print-Y-axis}, but not exactly.
827 * X Axis Tic Marks:: Create tic marks for the horizontal axis.
829 Printing the Whole Graph
831 * The final version:: A few changes.
832 * Test print-graph:: Run a short test.
833 * Graphing words in defuns:: Executing the final code.
834 * lambda:: How to write an anonymous function.
835 * mapcar:: Apply a function to elements of a list.
836 * Another Bug:: Yet another bug @dots{} most insidious.
837 * Final printed graph:: The graph itself!
845 Most of the GNU Emacs integrated environment is written in the programming
846 language called Emacs Lisp. The code written in this programming
847 language is the software---the sets of instructions---that tell the
848 computer what to do when you give it commands. Emacs is designed so
849 that you can write new code in Emacs Lisp and easily install it as an
850 extension to the editor.
852 (GNU Emacs is sometimes called an ``extensible editor'', but it does
853 much more than provide editing capabilities. It is better to refer to
854 Emacs as an ``extensible computing environment''. However, that
855 phrase is quite a mouthful. It is easier to refer to Emacs simply as
856 an editor. Moreover, everything you do in Emacs---find the Mayan date
857 and phases of the moon, simplify polynomials, debug code, manage
858 files, read letters, write books---all these activities are kinds of
859 editing in the most general sense of the word.)
862 * Why:: Why learn Emacs Lisp?
863 * On Reading this Text:: Read, gain familiarity, pick up habits....
864 * Who You Are:: For whom this is written.
866 * Note for Novices:: You can read this as a novice.
872 @unnumberedsec Why Study Emacs Lisp?
875 Although Emacs Lisp is usually thought of in association only with Emacs,
876 it is a full computer programming language. You can use Emacs Lisp as
877 you would any other programming language.
879 Perhaps you want to understand programming; perhaps you want to extend
880 Emacs; or perhaps you want to become a programmer. This introduction to
881 Emacs Lisp is designed to get you started: to guide you in learning the
882 fundamentals of programming, and more importantly, to show you how you
883 can teach yourself to go further.
885 @node On Reading this Text
886 @unnumberedsec On Reading this Text
888 All through this document, you will see little sample programs you can
889 run inside of Emacs. If you read this document in Info inside of GNU
890 Emacs, you can run the programs as they appear. (This is easy to do and
891 is explained when the examples are presented.) Alternatively, you can
892 read this introduction as a printed book while sitting beside a computer
893 running Emacs. (This is what I like to do; I like printed books.) If
894 you don't have a running Emacs beside you, you can still read this book,
895 but in this case, it is best to treat it as a novel or as a travel guide
896 to a country not yet visited: interesting, but not the same as being
899 Much of this introduction is dedicated to walkthroughs or guided tours
900 of code used in GNU Emacs. These tours are designed for two purposes:
901 first, to give you familiarity with real, working code (code you use
902 every day); and, second, to give you familiarity with the way Emacs
903 works. It is interesting to see how a working environment is
906 hope that you will pick up the habit of browsing through source code.
907 You can learn from it and mine it for ideas. Having GNU Emacs is like
908 having a dragon's cave of treasures.
910 In addition to learning about Emacs as an editor and Emacs Lisp as a
911 programming language, the examples and guided tours will give you an
912 opportunity to get acquainted with Emacs as a Lisp programming
913 environment. GNU Emacs supports programming and provides tools that
914 you will want to become comfortable using, such as @kbd{M-.} (the key
915 which invokes the @code{find-tag} command). You will also learn about
916 buffers and other objects that are part of the environment.
917 Learning about these features of Emacs is like learning new routes
918 around your home town.
921 In addition, I have written several programs as extended examples.
922 Although these are examples, the programs are real. I use them.
923 Other people use them. You may use them. Beyond the fragments of
924 programs used for illustrations, there is very little in here that is
925 `just for teaching purposes'; what you see is used. This is a great
926 advantage of Emacs Lisp: it is easy to learn to use it for work.
929 Finally, I hope to convey some of the skills for using Emacs to
930 learn aspects of programming that you don't know. You can often use
931 Emacs to help you understand what puzzles you or to find out how to do
932 something new. This self-reliance is not only a pleasure, but an
936 @unnumberedsec For Whom This is Written
938 This text is written as an elementary introduction for people who are
939 not programmers. If you are a programmer, you may not be satisfied with
940 this primer. The reason is that you may have become expert at reading
941 reference manuals and be put off by the way this text is organized.
943 An expert programmer who reviewed this text said to me:
946 @i{I prefer to learn from reference manuals. I ``dive into'' each
947 paragraph, and ``come up for air'' between paragraphs.}
949 @i{When I get to the end of a paragraph, I assume that that subject is
950 done, finished, that I know everything I need (with the
951 possible exception of the case when the next paragraph starts talking
952 about it in more detail). I expect that a well written reference manual
953 will not have a lot of redundancy, and that it will have excellent
954 pointers to the (one) place where the information I want is.}
957 This introduction is not written for this person!
959 Firstly, I try to say everything at least three times: first, to
960 introduce it; second, to show it in context; and third, to show it in a
961 different context, or to review it.
963 Secondly, I hardly ever put all the information about a subject in one
964 place, much less in one paragraph. To my way of thinking, that imposes
965 too heavy a burden on the reader. Instead I try to explain only what
966 you need to know at the time. (Sometimes I include a little extra
967 information so you won't be surprised later when the additional
968 information is formally introduced.)
970 When you read this text, you are not expected to learn everything the
971 first time. Frequently, you need only make, as it were, a `nodding
972 acquaintance' with some of the items mentioned. My hope is that I have
973 structured the text and given you enough hints that you will be alert to
974 what is important, and concentrate on it.
976 You will need to ``dive into'' some paragraphs; there is no other way
977 to read them. But I have tried to keep down the number of such
978 paragraphs. This book is intended as an approachable hill, rather than
979 as a daunting mountain.
981 This introduction to @cite{Programming in Emacs Lisp} has a companion
984 @cite{The GNU Emacs Lisp Reference Manual}.
987 @ref{Top, , The GNU Emacs Lisp Reference Manual, elisp, The GNU
988 Emacs Lisp Reference Manual}.
990 The reference manual has more detail than this introduction. In the
991 reference manual, all the information about one topic is concentrated
992 in one place. You should turn to it if you are like the programmer
993 quoted above. And, of course, after you have read this
994 @cite{Introduction}, you will find the @cite{Reference Manual} useful
995 when you are writing your own programs.
998 @unnumberedsec Lisp History
1001 Lisp was first developed in the late 1950s at the Massachusetts
1002 Institute of Technology for research in artificial intelligence. The
1003 great power of the Lisp language makes it superior for other purposes as
1004 well, such as writing editor commands and integrated environments.
1008 GNU Emacs Lisp is largely inspired by Maclisp, which was written at MIT
1009 in the 1960s. It is somewhat inspired by Common Lisp, which became a
1010 standard in the 1980s. However, Emacs Lisp is much simpler than Common
1011 Lisp. (The standard Emacs distribution contains an optional extensions
1012 file, @file{cl.el}, that adds many Common Lisp features to Emacs Lisp.)
1014 @node Note for Novices
1015 @unnumberedsec A Note for Novices
1017 If you don't know GNU Emacs, you can still read this document
1018 profitably. However, I recommend you learn Emacs, if only to learn to
1019 move around your computer screen. You can teach yourself how to use
1020 Emacs with the on-line tutorial. To use it, type @kbd{C-h t}. (This
1021 means you press and release the @key{CTRL} key and the @kbd{h} at the
1022 same time, and then press and release @kbd{t}.)
1024 Also, I often refer to one of Emacs's standard commands by listing the
1025 keys which you press to invoke the command and then giving the name of
1026 the command in parentheses, like this: @kbd{M-C-\}
1027 (@code{indent-region}). What this means is that the
1028 @code{indent-region} command is customarily invoked by typing
1029 @kbd{M-C-\}. (You can, if you wish, change the keys that are typed to
1030 invoke the command; this is called @dfn{rebinding}. @xref{Keymaps, ,
1031 Keymaps}.) The abbreviation @kbd{M-C-\} means that you type your
1032 @key{META} key, @key{CTRL} key and @key{\} key all at the same time.
1033 (On many modern keyboards the @key{META} key is labeled
1035 Sometimes a combination like this is called a keychord, since it is
1036 similar to the way you play a chord on a piano. If your keyboard does
1037 not have a @key{META} key, the @key{ESC} key prefix is used in place
1038 of it. In this case, @kbd{M-C-\} means that you press and release your
1039 @key{ESC} key and then type the @key{CTRL} key and the @key{\} key at
1040 the same time. But usually @kbd{M-C-\} means press the @key{CTRL} key
1041 along with the key that is labeled @key{ALT} and, at the same time,
1042 press the @key{\} key.
1044 In addition to typing a lone keychord, you can prefix what you type
1045 with @kbd{C-u}, which is called the `universal argument'. The
1046 @kbd{C-u} keychord passes an argument to the subsequent command.
1047 Thus, to indent a region of plain text by 6 spaces, mark the region,
1048 and then type @w{@kbd{C-u 6 M-C-\}}. (If you do not specify a number,
1049 Emacs either passes the number 4 to the command or otherwise runs the
1050 command differently than it would otherwise.) @xref{Arguments, ,
1051 Numeric Arguments, emacs, The GNU Emacs Manual}.
1053 If you are reading this in Info using GNU Emacs, you can read through
1054 this whole document just by pressing the space bar, @key{SPC}.
1055 (To learn about Info, type @kbd{C-h i} and then select Info.)
1057 A note on terminology: when I use the word Lisp alone, I often am
1058 referring to the various dialects of Lisp in general, but when I speak
1059 of Emacs Lisp, I am referring to GNU Emacs Lisp in particular.
1062 @unnumberedsec Thank You
1064 My thanks to all who helped me with this book. My especial thanks to
1065 @r{Jim Blandy}, @r{Noah Friedman}, @w{Jim Kingdon}, @r{Roland
1066 McGrath}, @w{Frank Ritter}, @w{Randy Smith}, @w{Richard M.
1067 Stallman}, and @w{Melissa Weisshaus}. My thanks also go to both
1068 @w{Philip Johnson} and @w{David Stampe} for their patient
1069 encouragement. My mistakes are my own.
1074 @email{bob@@gnu.org}
1081 @c ================ Beginning of main text ================
1083 @c Start main text on right-hand (verso) page
1086 \par\vfill\supereject
1089 \par\vfill\supereject
1091 \par\vfill\supereject
1093 \par\vfill\supereject
1097 @c Note: this resetting of the page number back to 1 causes TeX to gripe
1098 @c about already having seen page numbers 1-4 before (in the preface):
1099 @c pdfTeX warning (ext4): destination with the same identifier (name{1})
1100 @c has been already used, duplicate ignored
1101 @c I guess that is harmless (what happens if a later part of the text
1102 @c makes a link to something in the first 4 pages though?).
1103 @c E.g., note that the Emacs manual has a preface, but does not bother
1104 @c resetting the page numbers back to 1 after that.
1107 @evenheading @thispage @| @| @thischapter
1108 @oddheading @thissection @| @| @thispage
1112 @node List Processing
1113 @chapter List Processing
1115 To the untutored eye, Lisp is a strange programming language. In Lisp
1116 code there are parentheses everywhere. Some people even claim that
1117 the name stands for `Lots of Isolated Silly Parentheses'. But the
1118 claim is unwarranted. Lisp stands for LISt Processing, and the
1119 programming language handles @emph{lists} (and lists of lists) by
1120 putting them between parentheses. The parentheses mark the boundaries
1121 of the list. Sometimes a list is preceded by a single apostrophe or
1122 quotation mark, @samp{'}@footnote{The single apostrophe or quotation
1123 mark is an abbreviation for the function @code{quote}; you need not
1124 think about functions now; functions are defined in @ref{Making
1125 Errors, , Generate an Error Message}.} Lists are the basis of Lisp.
1128 * Lisp Lists:: What are lists?
1129 * Run a Program:: Any list in Lisp is a program ready to run.
1130 * Making Errors:: Generating an error message.
1131 * Names & Definitions:: Names of symbols and function definitions.
1132 * Lisp Interpreter:: What the Lisp interpreter does.
1133 * Evaluation:: Running a program.
1134 * Variables:: Returning a value from a variable.
1135 * Arguments:: Passing information to a function.
1136 * set & setq:: Setting the value of a variable.
1137 * Summary:: The major points.
1138 * Error Message Exercises::
1145 In Lisp, a list looks like this: @code{'(rose violet daisy buttercup)}.
1146 This list is preceded by a single apostrophe. It could just as well be
1147 written as follows, which looks more like the kind of list you are likely
1148 to be familiar with:
1160 The elements of this list are the names of the four different flowers,
1161 separated from each other by whitespace and surrounded by parentheses,
1162 like flowers in a field with a stone wall around them.
1163 @cindex Flowers in a field
1166 * Numbers Lists:: List have numbers, other lists, in them.
1167 * Lisp Atoms:: Elemental entities.
1168 * Whitespace in Lists:: Formatting lists to be readable.
1169 * Typing Lists:: How GNU Emacs helps you type lists.
1174 @unnumberedsubsec Numbers, Lists inside of Lists
1177 Lists can also have numbers in them, as in this list: @code{(+ 2 2)}.
1178 This list has a plus-sign, @samp{+}, followed by two @samp{2}s, each
1179 separated by whitespace.
1181 In Lisp, both data and programs are represented the same way; that is,
1182 they are both lists of words, numbers, or other lists, separated by
1183 whitespace and surrounded by parentheses. (Since a program looks like
1184 data, one program may easily serve as data for another; this is a very
1185 powerful feature of Lisp.) (Incidentally, these two parenthetical
1186 remarks are @emph{not} Lisp lists, because they contain @samp{;} and
1187 @samp{.} as punctuation marks.)
1190 Here is another list, this time with a list inside of it:
1193 '(this list has (a list inside of it))
1196 The components of this list are the words @samp{this}, @samp{list},
1197 @samp{has}, and the list @samp{(a list inside of it)}. The interior
1198 list is made up of the words @samp{a}, @samp{list}, @samp{inside},
1199 @samp{of}, @samp{it}.
1202 @subsection Lisp Atoms
1205 In Lisp, what we have been calling words are called @dfn{atoms}. This
1206 term comes from the historical meaning of the word atom, which means
1207 `indivisible'. As far as Lisp is concerned, the words we have been
1208 using in the lists cannot be divided into any smaller parts and still
1209 mean the same thing as part of a program; likewise with numbers and
1210 single character symbols like @samp{+}. On the other hand, unlike an
1211 ancient atom, a list can be split into parts. (@xref{car cdr & cons,
1212 , @code{car} @code{cdr} & @code{cons} Fundamental Functions}.)
1214 In a list, atoms are separated from each other by whitespace. They can be
1215 right next to a parenthesis.
1217 @cindex @samp{empty list} defined
1218 Technically speaking, a list in Lisp consists of parentheses surrounding
1219 atoms separated by whitespace or surrounding other lists or surrounding
1220 both atoms and other lists. A list can have just one atom in it or
1221 have nothing in it at all. A list with nothing in it looks like this:
1222 @code{()}, and is called the @dfn{empty list}. Unlike anything else, an
1223 empty list is considered both an atom and a list at the same time.
1225 @cindex Symbolic expressions, introduced
1226 @cindex @samp{expression} defined
1227 @cindex @samp{form} defined
1228 The printed representation of both atoms and lists are called
1229 @dfn{symbolic expressions} or, more concisely, @dfn{s-expressions}.
1230 The word @dfn{expression} by itself can refer to either the printed
1231 representation, or to the atom or list as it is held internally in the
1232 computer. Often, people use the term @dfn{expression}
1233 indiscriminately. (Also, in many texts, the word @dfn{form} is used
1234 as a synonym for expression.)
1236 Incidentally, the atoms that make up our universe were named such when
1237 they were thought to be indivisible; but it has been found that physical
1238 atoms are not indivisible. Parts can split off an atom or it can
1239 fission into two parts of roughly equal size. Physical atoms were named
1240 prematurely, before their truer nature was found. In Lisp, certain
1241 kinds of atom, such as an array, can be separated into parts; but the
1242 mechanism for doing this is different from the mechanism for splitting a
1243 list. As far as list operations are concerned, the atoms of a list are
1246 As in English, the meanings of the component letters of a Lisp atom
1247 are different from the meaning the letters make as a word. For
1248 example, the word for the South American sloth, the @samp{ai}, is
1249 completely different from the two words, @samp{a}, and @samp{i}.
1251 There are many kinds of atom in nature but only a few in Lisp: for
1252 example, @dfn{numbers}, such as 37, 511, or 1729, and @dfn{symbols}, such
1253 as @samp{+}, @samp{foo}, or @samp{forward-line}. The words we have
1254 listed in the examples above are all symbols. In everyday Lisp
1255 conversation, the word ``atom'' is not often used, because programmers
1256 usually try to be more specific about what kind of atom they are dealing
1257 with. Lisp programming is mostly about symbols (and sometimes numbers)
1258 within lists. (Incidentally, the preceding three word parenthetical
1259 remark is a proper list in Lisp, since it consists of atoms, which in
1260 this case are symbols, separated by whitespace and enclosed by
1261 parentheses, without any non-Lisp punctuation.)
1264 Text between double quotation marks---even sentences or
1265 paragraphs---is also an atom. Here is an example:
1266 @cindex Text between double quotation marks
1269 '(this list includes "text between quotation marks.")
1272 @cindex @samp{string} defined
1274 In Lisp, all of the quoted text including the punctuation mark and the
1275 blank spaces is a single atom. This kind of atom is called a
1276 @dfn{string} (for `string of characters') and is the sort of thing that
1277 is used for messages that a computer can print for a human to read.
1278 Strings are a different kind of atom than numbers or symbols and are
1281 @node Whitespace in Lists
1282 @subsection Whitespace in Lists
1283 @cindex Whitespace in lists
1286 The amount of whitespace in a list does not matter. From the point of view
1287 of the Lisp language,
1298 is exactly the same as this:
1301 '(this list looks like this)
1304 Both examples show what to Lisp is the same list, the list made up of
1305 the symbols @samp{this}, @samp{list}, @samp{looks}, @samp{like}, and
1306 @samp{this} in that order.
1308 Extra whitespace and newlines are designed to make a list more readable
1309 by humans. When Lisp reads the expression, it gets rid of all the extra
1310 whitespace (but it needs to have at least one space between atoms in
1311 order to tell them apart.)
1313 Odd as it seems, the examples we have seen cover almost all of what Lisp
1314 lists look like! Every other list in Lisp looks more or less like one
1315 of these examples, except that the list may be longer and more complex.
1316 In brief, a list is between parentheses, a string is between quotation
1317 marks, a symbol looks like a word, and a number looks like a number.
1318 (For certain situations, square brackets, dots and a few other special
1319 characters may be used; however, we will go quite far without them.)
1322 @subsection GNU Emacs Helps You Type Lists
1323 @cindex Help typing lists
1324 @cindex Formatting help
1326 When you type a Lisp expression in GNU Emacs using either Lisp
1327 Interaction mode or Emacs Lisp mode, you have available to you several
1328 commands to format the Lisp expression so it is easy to read. For
1329 example, pressing the @key{TAB} key automatically indents the line the
1330 cursor is on by the right amount. A command to properly indent the
1331 code in a region is customarily bound to @kbd{M-C-\}. Indentation is
1332 designed so that you can see which elements of a list belong to which
1333 list---elements of a sub-list are indented more than the elements of
1336 In addition, when you type a closing parenthesis, Emacs momentarily
1337 jumps the cursor back to the matching opening parenthesis, so you can
1338 see which one it is. This is very useful, since every list you type
1339 in Lisp must have its closing parenthesis match its opening
1340 parenthesis. (@xref{Major Modes, , Major Modes, emacs, The GNU Emacs
1341 Manual}, for more information about Emacs's modes.)
1344 @section Run a Program
1345 @cindex Run a program
1346 @cindex Program, running one
1348 @cindex @samp{evaluate} defined
1349 A list in Lisp---any list---is a program ready to run. If you run it
1350 (for which the Lisp jargon is @dfn{evaluate}), the computer will do one
1351 of three things: do nothing except return to you the list itself; send
1352 you an error message; or, treat the first symbol in the list as a
1353 command to do something. (Usually, of course, it is the last of these
1354 three things that you really want!)
1356 @c use code for the single apostrophe, not samp.
1357 The single apostrophe, @code{'}, that I put in front of some of the
1358 example lists in preceding sections is called a @dfn{quote}; when it
1359 precedes a list, it tells Lisp to do nothing with the list, other than
1360 take it as it is written. But if there is no quote preceding a list,
1361 the first item of the list is special: it is a command for the computer
1362 to obey. (In Lisp, these commands are called @emph{functions}.) The list
1363 @code{(+ 2 2)} shown above did not have a quote in front of it, so Lisp
1364 understands that the @code{+} is an instruction to do something with the
1365 rest of the list: add the numbers that follow.
1368 If you are reading this inside of GNU Emacs in Info, here is how you can
1369 evaluate such a list: place your cursor immediately after the right
1370 hand parenthesis of the following list and then type @kbd{C-x C-e}:
1376 @c use code for the number four, not samp.
1378 You will see the number @code{4} appear in the echo area. (In the
1379 jargon, what you have just done is ``evaluate the list.'' The echo area
1380 is the line at the bottom of the screen that displays or ``echoes''
1381 text.) Now try the same thing with a quoted list: place the cursor
1382 right after the following list and type @kbd{C-x C-e}:
1385 '(this is a quoted list)
1389 You will see @code{(this is a quoted list)} appear in the echo area.
1391 @cindex Lisp interpreter, explained
1392 @cindex Interpreter, Lisp, explained
1393 In both cases, what you are doing is giving a command to the program
1394 inside of GNU Emacs called the @dfn{Lisp interpreter}---giving the
1395 interpreter a command to evaluate the expression. The name of the Lisp
1396 interpreter comes from the word for the task done by a human who comes
1397 up with the meaning of an expression---who ``interprets'' it.
1399 You can also evaluate an atom that is not part of a list---one that is
1400 not surrounded by parentheses; again, the Lisp interpreter translates
1401 from the humanly readable expression to the language of the computer.
1402 But before discussing this (@pxref{Variables}), we will discuss what the
1403 Lisp interpreter does when you make an error.
1406 @section Generate an Error Message
1407 @cindex Generate an error message
1408 @cindex Error message generation
1410 Partly so you won't worry if you do it accidentally, we will now give
1411 a command to the Lisp interpreter that generates an error message.
1412 This is a harmless activity; and indeed, we will often try to generate
1413 error messages intentionally. Once you understand the jargon, error
1414 messages can be informative. Instead of being called ``error''
1415 messages, they should be called ``help'' messages. They are like
1416 signposts to a traveler in a strange country; deciphering them can be
1417 hard, but once understood, they can point the way.
1419 The error message is generated by a built-in GNU Emacs debugger. We
1420 will `enter the debugger'. You get out of the debugger by typing @code{q}.
1422 What we will do is evaluate a list that is not quoted and does not
1423 have a meaningful command as its first element. Here is a list almost
1424 exactly the same as the one we just used, but without the single-quote
1425 in front of it. Position the cursor right after it and type @kbd{C-x
1429 (this is an unquoted list)
1434 What you see depends on which version of Emacs you are running. GNU
1435 Emacs version 22 provides more information than version 20 and before.
1436 First, the more recent result of generating an error; then the
1437 earlier, version 20 result.
1441 In GNU Emacs version 22, a @file{*Backtrace*} window will open up and
1442 you will see the following in it:
1445 A @file{*Backtrace*} window will open up and you should see the
1450 ---------- Buffer: *Backtrace* ----------
1451 Debugger entered--Lisp error: (void-function this)
1452 (this is an unquoted list)
1453 eval((this is an unquoted list))
1454 eval-last-sexp-1(nil)
1456 call-interactively(eval-last-sexp)
1457 ---------- Buffer: *Backtrace* ----------
1463 Your cursor will be in this window (you may have to wait a few seconds
1464 before it becomes visible). To quit the debugger and make the
1465 debugger window go away, type:
1472 Please type @kbd{q} right now, so you become confident that you can
1473 get out of the debugger. Then, type @kbd{C-x C-e} again to re-enter
1476 @cindex @samp{function} defined
1477 Based on what we already know, we can almost read this error message.
1479 You read the @file{*Backtrace*} buffer from the bottom up; it tells
1480 you what Emacs did. When you typed @kbd{C-x C-e}, you made an
1481 interactive call to the command @code{eval-last-sexp}. @code{eval} is
1482 an abbreviation for `evaluate' and @code{sexp} is an abbreviation for
1483 `symbolic expression'. The command means `evaluate last symbolic
1484 expression', which is the expression just before your cursor.
1486 Each line above tells you what the Lisp interpreter evaluated next.
1487 The most recent action is at the top. The buffer is called the
1488 @file{*Backtrace*} buffer because it enables you to track Emacs
1492 At the top of the @file{*Backtrace*} buffer, you see the line:
1495 Debugger entered--Lisp error: (void-function this)
1499 The Lisp interpreter tried to evaluate the first atom of the list, the
1500 word @samp{this}. It is this action that generated the error message
1501 @samp{void-function this}.
1503 The message contains the words @samp{void-function} and @samp{this}.
1505 @cindex @samp{function} defined
1506 The word @samp{function} was mentioned once before. It is a very
1507 important word. For our purposes, we can define it by saying that a
1508 @dfn{function} is a set of instructions to the computer that tell the
1509 computer to do something.
1511 Now we can begin to understand the error message: @samp{void-function
1512 this}. The function (that is, the word @samp{this}) does not have a
1513 definition of any set of instructions for the computer to carry out.
1515 The slightly odd word, @samp{void-function}, is designed to cover the
1516 way Emacs Lisp is implemented, which is that when a symbol does not
1517 have a function definition attached to it, the place that should
1518 contain the instructions is `void'.
1520 On the other hand, since we were able to add 2 plus 2 successfully, by
1521 evaluating @code{(+ 2 2)}, we can infer that the symbol @code{+} must
1522 have a set of instructions for the computer to obey and those
1523 instructions must be to add the numbers that follow the @code{+}.
1525 It is possible to prevent Emacs entering the debugger in cases like
1526 this. We do not explain how to do that here, but we will mention what
1527 the result looks like, because you may encounter a similar situation
1528 if there is a bug in some Emacs code that you are using. In such
1529 cases, you will see only one line of error message; it will appear in
1530 the echo area and look like this:
1533 Symbol's function definition is void:@: this
1538 (Also, your terminal may beep at you---some do, some don't; and others
1539 blink. This is just a device to get your attention.)
1541 The message goes away as soon as you type a key, even just to
1544 We know the meaning of the word @samp{Symbol}. It refers to the first
1545 atom of the list, the word @samp{this}. The word @samp{function}
1546 refers to the instructions that tell the computer what to do.
1547 (Technically, the symbol tells the computer where to find the
1548 instructions, but this is a complication we can ignore for the
1551 The error message can be understood: @samp{Symbol's function
1552 definition is void:@: this}. The symbol (that is, the word
1553 @samp{this}) lacks instructions for the computer to carry out.
1555 @node Names & Definitions
1556 @section Symbol Names and Function Definitions
1557 @cindex Symbol names
1559 We can articulate another characteristic of Lisp based on what we have
1560 discussed so far---an important characteristic: a symbol, like
1561 @code{+}, is not itself the set of instructions for the computer to
1562 carry out. Instead, the symbol is used, perhaps temporarily, as a way
1563 of locating the definition or set of instructions. What we see is the
1564 name through which the instructions can be found. Names of people
1565 work the same way. I can be referred to as @samp{Bob}; however, I am
1566 not the letters @samp{B}, @samp{o}, @samp{b} but am, or was, the
1567 consciousness consistently associated with a particular life-form.
1568 The name is not me, but it can be used to refer to me.
1570 In Lisp, one set of instructions can be attached to several names.
1571 For example, the computer instructions for adding numbers can be
1572 linked to the symbol @code{plus} as well as to the symbol @code{+}
1573 (and are in some dialects of Lisp). Among humans, I can be referred
1574 to as @samp{Robert} as well as @samp{Bob} and by other words as well.
1576 On the other hand, a symbol can have only one function definition
1577 attached to it at a time. Otherwise, the computer would be confused as
1578 to which definition to use. If this were the case among people, only
1579 one person in the world could be named @samp{Bob}. However, the function
1580 definition to which the name refers can be changed readily.
1581 (@xref{Install, , Install a Function Definition}.)
1583 Since Emacs Lisp is large, it is customary to name symbols in a way
1584 that identifies the part of Emacs to which the function belongs.
1585 Thus, all the names for functions that deal with Texinfo start with
1586 @samp{texinfo-} and those for functions that deal with reading mail
1587 start with @samp{rmail-}.
1589 @node Lisp Interpreter
1590 @section The Lisp Interpreter
1591 @cindex Lisp interpreter, what it does
1592 @cindex Interpreter, what it does
1594 Based on what we have seen, we can now start to figure out what the
1595 Lisp interpreter does when we command it to evaluate a list.
1596 First, it looks to see whether there is a quote before the list; if
1597 there is, the interpreter just gives us the list. On the other
1598 hand, if there is no quote, the interpreter looks at the first element
1599 in the list and sees whether it has a function definition. If it does,
1600 the interpreter carries out the instructions in the function definition.
1601 Otherwise, the interpreter prints an error message.
1603 This is how Lisp works. Simple. There are added complications which we
1604 will get to in a minute, but these are the fundamentals. Of course, to
1605 write Lisp programs, you need to know how to write function definitions
1606 and attach them to names, and how to do this without confusing either
1607 yourself or the computer.
1610 * Complications:: Variables, Special forms, Lists within.
1611 * Byte Compiling:: Specially processing code for speed.
1616 @unnumberedsubsec Complications
1619 Now, for the first complication. In addition to lists, the Lisp
1620 interpreter can evaluate a symbol that is not quoted and does not have
1621 parentheses around it. The Lisp interpreter will attempt to determine
1622 the symbol's value as a @dfn{variable}. This situation is described
1623 in the section on variables. (@xref{Variables}.)
1625 @cindex Special form
1626 The second complication occurs because some functions are unusual and do
1627 not work in the usual manner. Those that don't are called @dfn{special
1628 forms}. They are used for special jobs, like defining a function, and
1629 there are not many of them. In the next few chapters, you will be
1630 introduced to several of the more important special forms.
1632 The third and final complication is this: if the function that the
1633 Lisp interpreter is looking at is not a special form, and if it is part
1634 of a list, the Lisp interpreter looks to see whether the list has a list
1635 inside of it. If there is an inner list, the Lisp interpreter first
1636 figures out what it should do with the inside list, and then it works on
1637 the outside list. If there is yet another list embedded inside the
1638 inner list, it works on that one first, and so on. It always works on
1639 the innermost list first. The interpreter works on the innermost list
1640 first, to evaluate the result of that list. The result may be
1641 used by the enclosing expression.
1643 Otherwise, the interpreter works left to right, from one expression to
1646 @node Byte Compiling
1647 @subsection Byte Compiling
1648 @cindex Byte compiling
1650 One other aspect of interpreting: the Lisp interpreter is able to
1651 interpret two kinds of entity: humanly readable code, on which we will
1652 focus exclusively, and specially processed code, called @dfn{byte
1653 compiled} code, which is not humanly readable. Byte compiled code
1654 runs faster than humanly readable code.
1656 You can transform humanly readable code into byte compiled code by
1657 running one of the compile commands such as @code{byte-compile-file}.
1658 Byte compiled code is usually stored in a file that ends with a
1659 @file{.elc} extension rather than a @file{.el} extension. You will
1660 see both kinds of file in the @file{emacs/lisp} directory; the files
1661 to read are those with @file{.el} extensions.
1663 As a practical matter, for most things you might do to customize or
1664 extend Emacs, you do not need to byte compile; and I will not discuss
1665 the topic here. @xref{Byte Compilation, , Byte Compilation, elisp,
1666 The GNU Emacs Lisp Reference Manual}, for a full description of byte
1673 When the Lisp interpreter works on an expression, the term for the
1674 activity is called @dfn{evaluation}. We say that the interpreter
1675 `evaluates the expression'. I've used this term several times before.
1676 The word comes from its use in everyday language, `to ascertain the
1677 value or amount of; to appraise', according to @cite{Webster's New
1678 Collegiate Dictionary}.
1681 * How the Interpreter Acts:: Returns and Side Effects...
1682 * Evaluating Inner Lists:: Lists within lists...
1686 @node How the Interpreter Acts
1687 @unnumberedsubsec How the Lisp Interpreter Acts
1690 @cindex @samp{returned value} explained
1691 After evaluating an expression, the Lisp interpreter will most likely
1692 @dfn{return} the value that the computer produces by carrying out the
1693 instructions it found in the function definition, or perhaps it will
1694 give up on that function and produce an error message. (The interpreter
1695 may also find itself tossed, so to speak, to a different function or it
1696 may attempt to repeat continually what it is doing for ever and ever in
1697 what is called an `infinite loop'. These actions are less common; and
1698 we can ignore them.) Most frequently, the interpreter returns a value.
1700 @cindex @samp{side effect} defined
1701 At the same time the interpreter returns a value, it may do something
1702 else as well, such as move a cursor or copy a file; this other kind of
1703 action is called a @dfn{side effect}. Actions that we humans think are
1704 important, such as printing results, are often ``side effects'' to the
1705 Lisp interpreter. The jargon can sound peculiar, but it turns out that
1706 it is fairly easy to learn to use side effects.
1708 In summary, evaluating a symbolic expression most commonly causes the
1709 Lisp interpreter to return a value and perhaps carry out a side effect;
1710 or else produce an error.
1712 @node Evaluating Inner Lists
1713 @subsection Evaluating Inner Lists
1714 @cindex Inner list evaluation
1715 @cindex Evaluating inner lists
1717 If evaluation applies to a list that is inside another list, the outer
1718 list may use the value returned by the first evaluation as information
1719 when the outer list is evaluated. This explains why inner expressions
1720 are evaluated first: the values they return are used by the outer
1724 We can investigate this process by evaluating another addition example.
1725 Place your cursor after the following expression and type @kbd{C-x C-e}:
1732 The number 8 will appear in the echo area.
1734 What happens is that the Lisp interpreter first evaluates the inner
1735 expression, @code{(+ 3 3)}, for which the value 6 is returned; then it
1736 evaluates the outer expression as if it were written @code{(+ 2 6)}, which
1737 returns the value 8. Since there are no more enclosing expressions to
1738 evaluate, the interpreter prints that value in the echo area.
1740 Now it is easy to understand the name of the command invoked by the
1741 keystrokes @kbd{C-x C-e}: the name is @code{eval-last-sexp}. The
1742 letters @code{sexp} are an abbreviation for `symbolic expression', and
1743 @code{eval} is an abbreviation for `evaluate'. The command means
1744 `evaluate last symbolic expression'.
1746 As an experiment, you can try evaluating the expression by putting the
1747 cursor at the beginning of the next line immediately following the
1748 expression, or inside the expression.
1751 Here is another copy of the expression:
1758 If you place the cursor at the beginning of the blank line that
1759 immediately follows the expression and type @kbd{C-x C-e}, you will
1760 still get the value 8 printed in the echo area. Now try putting the
1761 cursor inside the expression. If you put it right after the next to
1762 last parenthesis (so it appears to sit on top of the last parenthesis),
1763 you will get a 6 printed in the echo area! This is because the command
1764 evaluates the expression @code{(+ 3 3)}.
1766 Now put the cursor immediately after a number. Type @kbd{C-x C-e} and
1767 you will get the number itself. In Lisp, if you evaluate a number, you
1768 get the number itself---this is how numbers differ from symbols. If you
1769 evaluate a list starting with a symbol like @code{+}, you will get a
1770 value returned that is the result of the computer carrying out the
1771 instructions in the function definition attached to that name. If a
1772 symbol by itself is evaluated, something different happens, as we will
1773 see in the next section.
1779 In Emacs Lisp, a symbol can have a value attached to it just as it can
1780 have a function definition attached to it. The two are different.
1781 The function definition is a set of instructions that a computer will
1782 obey. A value, on the other hand, is something, such as number or a
1783 name, that can vary (which is why such a symbol is called a variable).
1784 The value of a symbol can be any expression in Lisp, such as a symbol,
1785 number, list, or string. A symbol that has a value is often called a
1788 A symbol can have both a function definition and a value attached to
1789 it at the same time. Or it can have just one or the other.
1790 The two are separate. This is somewhat similar
1791 to the way the name Cambridge can refer to the city in Massachusetts
1792 and have some information attached to the name as well, such as
1793 ``great programming center''.
1796 (Incidentally, in Emacs Lisp, a symbol can have two
1797 other things attached to it, too: a property list and a documentation
1798 string; these are discussed later.)
1801 Another way to think about this is to imagine a symbol as being a chest
1802 of drawers. The function definition is put in one drawer, the value in
1803 another, and so on. What is put in the drawer holding the value can be
1804 changed without affecting the contents of the drawer holding the
1805 function definition, and vice-verse.
1808 * fill-column Example::
1809 * Void Function:: The error message for a symbol
1811 * Void Variable:: The error message for a symbol without a value.
1815 @node fill-column Example
1816 @unnumberedsubsec @code{fill-column}, an Example Variable
1819 @findex fill-column, @r{an example variable}
1820 @cindex Example variable, @code{fill-column}
1821 @cindex Variable, example of, @code{fill-column}
1822 The variable @code{fill-column} illustrates a symbol with a value
1823 attached to it: in every GNU Emacs buffer, this symbol is set to some
1824 value, usually 72 or 70, but sometimes to some other value. To find the
1825 value of this symbol, evaluate it by itself. If you are reading this in
1826 Info inside of GNU Emacs, you can do this by putting the cursor after
1827 the symbol and typing @kbd{C-x C-e}:
1834 After I typed @kbd{C-x C-e}, Emacs printed the number 72 in my echo
1835 area. This is the value for which @code{fill-column} is set for me as I
1836 write this. It may be different for you in your Info buffer. Notice
1837 that the value returned as a variable is printed in exactly the same way
1838 as the value returned by a function carrying out its instructions. From
1839 the point of view of the Lisp interpreter, a value returned is a value
1840 returned. What kind of expression it came from ceases to matter once
1843 A symbol can have any value attached to it or, to use the jargon, we can
1844 @dfn{bind} the variable to a value: to a number, such as 72; to a
1845 string, @code{"such as this"}; to a list, such as @code{(spruce pine
1846 oak)}; we can even bind a variable to a function definition.
1848 A symbol can be bound to a value in several ways. @xref{set & setq, ,
1849 Setting the Value of a Variable}, for information about one way to do
1853 @subsection Error Message for a Symbol Without a Function
1854 @cindex Symbol without function error
1855 @cindex Error for symbol without function
1857 When we evaluated @code{fill-column} to find its value as a variable,
1858 we did not place parentheses around the word. This is because we did
1859 not intend to use it as a function name.
1861 If @code{fill-column} were the first or only element of a list, the
1862 Lisp interpreter would attempt to find the function definition
1863 attached to it. But @code{fill-column} has no function definition.
1864 Try evaluating this:
1872 You will create a @file{*Backtrace*} buffer that says:
1876 ---------- Buffer: *Backtrace* ----------
1877 Debugger entered--Lisp error: (void-function fill-column)
1880 eval-last-sexp-1(nil)
1882 call-interactively(eval-last-sexp)
1883 ---------- Buffer: *Backtrace* ----------
1888 (Remember, to quit the debugger and make the debugger window go away,
1889 type @kbd{q} in the @file{*Backtrace*} buffer.)
1893 In GNU Emacs 20 and before, you will produce an error message that says:
1896 Symbol's function definition is void:@: fill-column
1900 (The message will go away as soon as you move the cursor or type
1905 @subsection Error Message for a Symbol Without a Value
1906 @cindex Symbol without value error
1907 @cindex Error for symbol without value
1909 If you attempt to evaluate a symbol that does not have a value bound to
1910 it, you will receive an error message. You can see this by
1911 experimenting with our 2 plus 2 addition. In the following expression,
1912 put your cursor right after the @code{+}, before the first number 2,
1921 In GNU Emacs 22, you will create a @file{*Backtrace*} buffer that
1926 ---------- Buffer: *Backtrace* ----------
1927 Debugger entered--Lisp error: (void-variable +)
1929 eval-last-sexp-1(nil)
1931 call-interactively(eval-last-sexp)
1932 ---------- Buffer: *Backtrace* ----------
1937 (Again, you can quit the debugger by
1938 typing @kbd{q} in the @file{*Backtrace*} buffer.)
1940 This backtrace is different from the very first error message we saw,
1941 which said, @samp{Debugger entered--Lisp error: (void-function this)}.
1942 In this case, the function does not have a value as a variable; while
1943 in the other error message, the function (the word `this') did not
1946 In this experiment with the @code{+}, what we did was cause the Lisp
1947 interpreter to evaluate the @code{+} and look for the value of the
1948 variable instead of the function definition. We did this by placing the
1949 cursor right after the symbol rather than after the parenthesis of the
1950 enclosing list as we did before. As a consequence, the Lisp interpreter
1951 evaluated the preceding s-expression, which in this case was
1954 Since @code{+} does not have a value bound to it, just the function
1955 definition, the error message reported that the symbol's value as a
1960 In GNU Emacs version 20 and before, your error message will say:
1963 Symbol's value as variable is void:@: +
1967 The meaning is the same as in GNU Emacs 22.
1973 @cindex Passing information to functions
1975 To see how information is passed to functions, let's look again at
1976 our old standby, the addition of two plus two. In Lisp, this is written
1983 If you evaluate this expression, the number 4 will appear in your echo
1984 area. What the Lisp interpreter does is add the numbers that follow
1987 @cindex @samp{argument} defined
1988 The numbers added by @code{+} are called the @dfn{arguments} of the
1989 function @code{+}. These numbers are the information that is given to
1990 or @dfn{passed} to the function.
1992 The word `argument' comes from the way it is used in mathematics and
1993 does not refer to a disputation between two people; instead it refers to
1994 the information presented to the function, in this case, to the
1995 @code{+}. In Lisp, the arguments to a function are the atoms or lists
1996 that follow the function. The values returned by the evaluation of
1997 these atoms or lists are passed to the function. Different functions
1998 require different numbers of arguments; some functions require none at
1999 all.@footnote{It is curious to track the path by which the word `argument'
2000 came to have two different meanings, one in mathematics and the other in
2001 everyday English. According to the @cite{Oxford English Dictionary},
2002 the word derives from the Latin for @samp{to make clear, prove}; thus it
2003 came to mean, by one thread of derivation, `the evidence offered as
2004 proof', which is to say, `the information offered', which led to its
2005 meaning in Lisp. But in the other thread of derivation, it came to mean
2006 `to assert in a manner against which others may make counter
2007 assertions', which led to the meaning of the word as a disputation.
2008 (Note here that the English word has two different definitions attached
2009 to it at the same time. By contrast, in Emacs Lisp, a symbol cannot
2010 have two different function definitions at the same time.)}
2013 * Data types:: Types of data passed to a function.
2014 * Args as Variable or List:: An argument can be the value
2015 of a variable or list.
2016 * Variable Number of Arguments:: Some functions may take a
2017 variable number of arguments.
2018 * Wrong Type of Argument:: Passing an argument of the wrong type
2020 * message:: A useful function for sending messages.
2024 @subsection Arguments' Data Types
2026 @cindex Types of data
2027 @cindex Arguments' data types
2029 The type of data that should be passed to a function depends on what
2030 kind of information it uses. The arguments to a function such as
2031 @code{+} must have values that are numbers, since @code{+} adds numbers.
2032 Other functions use different kinds of data for their arguments.
2036 For example, the @code{concat} function links together or unites two or
2037 more strings of text to produce a string. The arguments are strings.
2038 Concatenating the two character strings @code{abc}, @code{def} produces
2039 the single string @code{abcdef}. This can be seen by evaluating the
2043 (concat "abc" "def")
2047 The value produced by evaluating this expression is @code{"abcdef"}.
2049 A function such as @code{substring} uses both a string and numbers as
2050 arguments. The function returns a part of the string, a substring of
2051 the first argument. This function takes three arguments. Its first
2052 argument is the string of characters, the second and third arguments are
2053 numbers that indicate the beginning and end of the substring. The
2054 numbers are a count of the number of characters (including spaces and
2055 punctuation) from the beginning of the string.
2058 For example, if you evaluate the following:
2061 (substring "The quick brown fox jumped." 16 19)
2065 you will see @code{"fox"} appear in the echo area. The arguments are the
2066 string and the two numbers.
2068 Note that the string passed to @code{substring} is a single atom even
2069 though it is made up of several words separated by spaces. Lisp counts
2070 everything between the two quotation marks as part of the string,
2071 including the spaces. You can think of the @code{substring} function as
2072 a kind of `atom smasher' since it takes an otherwise indivisible atom
2073 and extracts a part. However, @code{substring} is only able to extract
2074 a substring from an argument that is a string, not from another type of
2075 atom such as a number or symbol.
2077 @node Args as Variable or List
2078 @subsection An Argument as the Value of a Variable or List
2080 An argument can be a symbol that returns a value when it is evaluated.
2081 For example, when the symbol @code{fill-column} by itself is evaluated,
2082 it returns a number. This number can be used in an addition.
2085 Position the cursor after the following expression and type @kbd{C-x
2093 The value will be a number two more than what you get by evaluating
2094 @code{fill-column} alone. For me, this is 74, because my value of
2095 @code{fill-column} is 72.
2097 As we have just seen, an argument can be a symbol that returns a value
2098 when evaluated. In addition, an argument can be a list that returns a
2099 value when it is evaluated. For example, in the following expression,
2100 the arguments to the function @code{concat} are the strings
2101 @w{@code{"The "}} and @w{@code{" red foxes."}} and the list
2102 @code{(number-to-string (+ 2 fill-column))}.
2104 @c For GNU Emacs 22, need number-to-string
2106 (concat "The " (number-to-string (+ 2 fill-column)) " red foxes.")
2110 If you evaluate this expression---and if, as with my Emacs,
2111 @code{fill-column} evaluates to 72---@code{"The 74 red foxes."} will
2112 appear in the echo area. (Note that you must put spaces after the
2113 word @samp{The} and before the word @samp{red} so they will appear in
2114 the final string. The function @code{number-to-string} converts the
2115 integer that the addition function returns to a string.
2116 @code{number-to-string} is also known as @code{int-to-string}.)
2118 @node Variable Number of Arguments
2119 @subsection Variable Number of Arguments
2120 @cindex Variable number of arguments
2121 @cindex Arguments, variable number of
2123 Some functions, such as @code{concat}, @code{+} or @code{*}, take any
2124 number of arguments. (The @code{*} is the symbol for multiplication.)
2125 This can be seen by evaluating each of the following expressions in
2126 the usual way. What you will see in the echo area is printed in this
2127 text after @samp{@result{}}, which you may read as `evaluates to'.
2130 In the first set, the functions have no arguments:
2141 In this set, the functions have one argument each:
2152 In this set, the functions have three arguments each:
2156 (+ 3 4 5) @result{} 12
2158 (* 3 4 5) @result{} 60
2162 @node Wrong Type of Argument
2163 @subsection Using the Wrong Type Object as an Argument
2164 @cindex Wrong type of argument
2165 @cindex Argument, wrong type of
2167 When a function is passed an argument of the wrong type, the Lisp
2168 interpreter produces an error message. For example, the @code{+}
2169 function expects the values of its arguments to be numbers. As an
2170 experiment we can pass it the quoted symbol @code{hello} instead of a
2171 number. Position the cursor after the following expression and type
2179 When you do this you will generate an error message. What has happened
2180 is that @code{+} has tried to add the 2 to the value returned by
2181 @code{'hello}, but the value returned by @code{'hello} is the symbol
2182 @code{hello}, not a number. Only numbers can be added. So @code{+}
2183 could not carry out its addition.
2186 You will create and enter a @file{*Backtrace*} buffer that says:
2191 ---------- Buffer: *Backtrace* ----------
2192 Debugger entered--Lisp error:
2193 (wrong-type-argument number-or-marker-p hello)
2195 eval((+ 2 (quote hello)))
2196 eval-last-sexp-1(nil)
2198 call-interactively(eval-last-sexp)
2199 ---------- Buffer: *Backtrace* ----------
2204 As usual, the error message tries to be helpful and makes sense after you
2205 learn how to read it.@footnote{@code{(quote hello)} is an expansion of
2206 the abbreviation @code{'hello}.}
2208 The first part of the error message is straightforward; it says
2209 @samp{wrong type argument}. Next comes the mysterious jargon word
2210 @w{@samp{number-or-marker-p}}. This word is trying to tell you what
2211 kind of argument the @code{+} expected.
2213 The symbol @code{number-or-marker-p} says that the Lisp interpreter is
2214 trying to determine whether the information presented it (the value of
2215 the argument) is a number or a marker (a special object representing a
2216 buffer position). What it does is test to see whether the @code{+} is
2217 being given numbers to add. It also tests to see whether the
2218 argument is something called a marker, which is a specific feature of
2219 Emacs Lisp. (In Emacs, locations in a buffer are recorded as markers.
2220 When the mark is set with the @kbd{C-@@} or @kbd{C-@key{SPC}} command,
2221 its position is kept as a marker. The mark can be considered a
2222 number---the number of characters the location is from the beginning
2223 of the buffer.) In Emacs Lisp, @code{+} can be used to add the
2224 numeric value of marker positions as numbers.
2226 The @samp{p} of @code{number-or-marker-p} is the embodiment of a
2227 practice started in the early days of Lisp programming. The @samp{p}
2228 stands for `predicate'. In the jargon used by the early Lisp
2229 researchers, a predicate refers to a function to determine whether some
2230 property is true or false. So the @samp{p} tells us that
2231 @code{number-or-marker-p} is the name of a function that determines
2232 whether it is true or false that the argument supplied is a number or
2233 a marker. Other Lisp symbols that end in @samp{p} include @code{zerop},
2234 a function that tests whether its argument has the value of zero, and
2235 @code{listp}, a function that tests whether its argument is a list.
2237 Finally, the last part of the error message is the symbol @code{hello}.
2238 This is the value of the argument that was passed to @code{+}. If the
2239 addition had been passed the correct type of object, the value passed
2240 would have been a number, such as 37, rather than a symbol like
2241 @code{hello}. But then you would not have got the error message.
2245 In GNU Emacs version 20 and before, the echo area displays an error
2249 Wrong type argument:@: number-or-marker-p, hello
2252 This says, in different words, the same as the top line of the
2253 @file{*Backtrace*} buffer.
2257 @subsection The @code{message} Function
2260 Like @code{+}, the @code{message} function takes a variable number of
2261 arguments. It is used to send messages to the user and is so useful
2262 that we will describe it here.
2265 A message is printed in the echo area. For example, you can print a
2266 message in your echo area by evaluating the following list:
2269 (message "This message appears in the echo area!")
2272 The whole string between double quotation marks is a single argument
2273 and is printed @i{in toto}. (Note that in this example, the message
2274 itself will appear in the echo area within double quotes; that is
2275 because you see the value returned by the @code{message} function. In
2276 most uses of @code{message} in programs that you write, the text will
2277 be printed in the echo area as a side-effect, without the quotes.
2278 @xref{multiply-by-seven in detail, , @code{multiply-by-seven} in
2279 detail}, for an example of this.)
2281 However, if there is a @samp{%s} in the quoted string of characters, the
2282 @code{message} function does not print the @samp{%s} as such, but looks
2283 to the argument that follows the string. It evaluates the second
2284 argument and prints the value at the location in the string where the
2288 You can see this by positioning the cursor after the following
2289 expression and typing @kbd{C-x C-e}:
2292 (message "The name of this buffer is: %s." (buffer-name))
2296 In Info, @code{"The name of this buffer is: *info*."} will appear in the
2297 echo area. The function @code{buffer-name} returns the name of the
2298 buffer as a string, which the @code{message} function inserts in place
2301 To print a value as an integer, use @samp{%d} in the same way as
2302 @samp{%s}. For example, to print a message in the echo area that
2303 states the value of the @code{fill-column}, evaluate the following:
2306 (message "The value of fill-column is %d." fill-column)
2310 On my system, when I evaluate this list, @code{"The value of
2311 fill-column is 72."} appears in my echo area@footnote{Actually, you
2312 can use @code{%s} to print a number. It is non-specific. @code{%d}
2313 prints only the part of a number left of a decimal point, and not
2314 anything that is not a number.}.
2316 If there is more than one @samp{%s} in the quoted string, the value of
2317 the first argument following the quoted string is printed at the
2318 location of the first @samp{%s} and the value of the second argument is
2319 printed at the location of the second @samp{%s}, and so on.
2322 For example, if you evaluate the following,
2326 (message "There are %d %s in the office!"
2327 (- fill-column 14) "pink elephants")
2332 a rather whimsical message will appear in your echo area. On my system
2333 it says, @code{"There are 58 pink elephants in the office!"}.
2335 The expression @code{(- fill-column 14)} is evaluated and the resulting
2336 number is inserted in place of the @samp{%d}; and the string in double
2337 quotes, @code{"pink elephants"}, is treated as a single argument and
2338 inserted in place of the @samp{%s}. (That is to say, a string between
2339 double quotes evaluates to itself, like a number.)
2341 Finally, here is a somewhat complex example that not only illustrates
2342 the computation of a number, but also shows how you can use an
2343 expression within an expression to generate the text that is substituted
2348 (message "He saw %d %s"
2352 "The quick brown foxes jumped." 16 21)
2357 In this example, @code{message} has three arguments: the string,
2358 @code{"He saw %d %s"}, the expression, @code{(- fill-column 32)}, and
2359 the expression beginning with the function @code{concat}. The value
2360 resulting from the evaluation of @code{(- fill-column 32)} is inserted
2361 in place of the @samp{%d}; and the value returned by the expression
2362 beginning with @code{concat} is inserted in place of the @samp{%s}.
2364 When your fill column is 70 and you evaluate the expression, the
2365 message @code{"He saw 38 red foxes leaping."} appears in your echo
2369 @section Setting the Value of a Variable
2370 @cindex Variable, setting value
2371 @cindex Setting value of variable
2373 @cindex @samp{bind} defined
2374 There are several ways by which a variable can be given a value. One of
2375 the ways is to use either the function @code{set} or the function
2376 @code{setq}. Another way is to use @code{let} (@pxref{let}). (The
2377 jargon for this process is to @dfn{bind} a variable to a value.)
2379 The following sections not only describe how @code{set} and @code{setq}
2380 work but also illustrate how arguments are passed.
2383 * Using set:: Setting values.
2384 * Using setq:: Setting a quoted value.
2385 * Counting:: Using @code{setq} to count.
2389 @subsection Using @code{set}
2392 To set the value of the symbol @code{flowers} to the list @code{'(rose
2393 violet daisy buttercup)}, evaluate the following expression by
2394 positioning the cursor after the expression and typing @kbd{C-x C-e}.
2397 (set 'flowers '(rose violet daisy buttercup))
2401 The list @code{(rose violet daisy buttercup)} will appear in the echo
2402 area. This is what is @emph{returned} by the @code{set} function. As a
2403 side effect, the symbol @code{flowers} is bound to the list; that is,
2404 the symbol @code{flowers}, which can be viewed as a variable, is given
2405 the list as its value. (This process, by the way, illustrates how a
2406 side effect to the Lisp interpreter, setting the value, can be the
2407 primary effect that we humans are interested in. This is because every
2408 Lisp function must return a value if it does not get an error, but it
2409 will only have a side effect if it is designed to have one.)
2411 After evaluating the @code{set} expression, you can evaluate the symbol
2412 @code{flowers} and it will return the value you just set. Here is the
2413 symbol. Place your cursor after it and type @kbd{C-x C-e}.
2420 When you evaluate @code{flowers}, the list
2421 @code{(rose violet daisy buttercup)} appears in the echo area.
2423 Incidentally, if you evaluate @code{'flowers}, the variable with a quote
2424 in front of it, what you will see in the echo area is the symbol itself,
2425 @code{flowers}. Here is the quoted symbol, so you can try this:
2431 Note also, that when you use @code{set}, you need to quote both
2432 arguments to @code{set}, unless you want them evaluated. Since we do
2433 not want either argument evaluated, neither the variable
2434 @code{flowers} nor the list @code{(rose violet daisy buttercup)}, both
2435 are quoted. (When you use @code{set} without quoting its first
2436 argument, the first argument is evaluated before anything else is
2437 done. If you did this and @code{flowers} did not have a value
2438 already, you would get an error message that the @samp{Symbol's value
2439 as variable is void}; on the other hand, if @code{flowers} did return
2440 a value after it was evaluated, the @code{set} would attempt to set
2441 the value that was returned. There are situations where this is the
2442 right thing for the function to do; but such situations are rare.)
2445 @subsection Using @code{setq}
2448 As a practical matter, you almost always quote the first argument to
2449 @code{set}. The combination of @code{set} and a quoted first argument
2450 is so common that it has its own name: the special form @code{setq}.
2451 This special form is just like @code{set} except that the first argument
2452 is quoted automatically, so you don't need to type the quote mark
2453 yourself. Also, as an added convenience, @code{setq} permits you to set
2454 several different variables to different values, all in one expression.
2456 To set the value of the variable @code{carnivores} to the list
2457 @code{'(lion tiger leopard)} using @code{setq}, the following expression
2461 (setq carnivores '(lion tiger leopard))
2465 This is exactly the same as using @code{set} except the first argument
2466 is automatically quoted by @code{setq}. (The @samp{q} in @code{setq}
2467 means @code{quote}.)
2470 With @code{set}, the expression would look like this:
2473 (set 'carnivores '(lion tiger leopard))
2476 Also, @code{setq} can be used to assign different values to
2477 different variables. The first argument is bound to the value
2478 of the second argument, the third argument is bound to the value of the
2479 fourth argument, and so on. For example, you could use the following to
2480 assign a list of trees to the symbol @code{trees} and a list of herbivores
2481 to the symbol @code{herbivores}:
2485 (setq trees '(pine fir oak maple)
2486 herbivores '(gazelle antelope zebra))
2491 (The expression could just as well have been on one line, but it might
2492 not have fit on a page; and humans find it easier to read nicely
2495 Although I have been using the term `assign', there is another way of
2496 thinking about the workings of @code{set} and @code{setq}; and that is to
2497 say that @code{set} and @code{setq} make the symbol @emph{point} to the
2498 list. This latter way of thinking is very common and in forthcoming
2499 chapters we shall come upon at least one symbol that has `pointer' as
2500 part of its name. The name is chosen because the symbol has a value,
2501 specifically a list, attached to it; or, expressed another way,
2502 the symbol is set to ``point'' to the list.
2505 @subsection Counting
2508 Here is an example that shows how to use @code{setq} in a counter. You
2509 might use this to count how many times a part of your program repeats
2510 itself. First set a variable to zero; then add one to the number each
2511 time the program repeats itself. To do this, you need a variable that
2512 serves as a counter, and two expressions: an initial @code{setq}
2513 expression that sets the counter variable to zero; and a second
2514 @code{setq} expression that increments the counter each time it is
2519 (setq counter 0) ; @r{Let's call this the initializer.}
2521 (setq counter (+ counter 1)) ; @r{This is the incrementer.}
2523 counter ; @r{This is the counter.}
2528 (The text following the @samp{;} are comments. @xref{Change a
2529 defun, , Change a Function Definition}.)
2531 If you evaluate the first of these expressions, the initializer,
2532 @code{(setq counter 0)}, and then evaluate the third expression,
2533 @code{counter}, the number @code{0} will appear in the echo area. If
2534 you then evaluate the second expression, the incrementer, @code{(setq
2535 counter (+ counter 1))}, the counter will get the value 1. So if you
2536 again evaluate @code{counter}, the number @code{1} will appear in the
2537 echo area. Each time you evaluate the second expression, the value of
2538 the counter will be incremented.
2540 When you evaluate the incrementer, @code{(setq counter (+ counter 1))},
2541 the Lisp interpreter first evaluates the innermost list; this is the
2542 addition. In order to evaluate this list, it must evaluate the variable
2543 @code{counter} and the number @code{1}. When it evaluates the variable
2544 @code{counter}, it receives its current value. It passes this value and
2545 the number @code{1} to the @code{+} which adds them together. The sum
2546 is then returned as the value of the inner list and passed to the
2547 @code{setq} which sets the variable @code{counter} to this new value.
2548 Thus, the value of the variable, @code{counter}, is changed.
2553 Learning Lisp is like climbing a hill in which the first part is the
2554 steepest. You have now climbed the most difficult part; what remains
2555 becomes easier as you progress onwards.
2563 Lisp programs are made up of expressions, which are lists or single atoms.
2566 Lists are made up of zero or more atoms or inner lists, separated by whitespace and
2567 surrounded by parentheses. A list can be empty.
2570 Atoms are multi-character symbols, like @code{forward-paragraph}, single
2571 character symbols like @code{+}, strings of characters between double
2572 quotation marks, or numbers.
2575 A number evaluates to itself.
2578 A string between double quotes also evaluates to itself.
2581 When you evaluate a symbol by itself, its value is returned.
2584 When you evaluate a list, the Lisp interpreter looks at the first symbol
2585 in the list and then at the function definition bound to that symbol.
2586 Then the instructions in the function definition are carried out.
2589 A single quotation mark,
2596 , tells the Lisp interpreter that it should
2597 return the following expression as written, and not evaluate it as it
2598 would if the quote were not there.
2601 Arguments are the information passed to a function. The arguments to a
2602 function are computed by evaluating the rest of the elements of the list
2603 of which the function is the first element.
2606 A function always returns a value when it is evaluated (unless it gets
2607 an error); in addition, it may also carry out some action called a
2608 ``side effect''. In many cases, a function's primary purpose is to
2609 create a side effect.
2612 @node Error Message Exercises
2615 A few simple exercises:
2619 Generate an error message by evaluating an appropriate symbol that is
2620 not within parentheses.
2623 Generate an error message by evaluating an appropriate symbol that is
2624 between parentheses.
2627 Create a counter that increments by two rather than one.
2630 Write an expression that prints a message in the echo area when
2634 @node Practicing Evaluation
2635 @chapter Practicing Evaluation
2636 @cindex Practicing evaluation
2637 @cindex Evaluation practice
2639 Before learning how to write a function definition in Emacs Lisp, it is
2640 useful to spend a little time evaluating various expressions that have
2641 already been written. These expressions will be lists with the
2642 functions as their first (and often only) element. Since some of the
2643 functions associated with buffers are both simple and interesting, we
2644 will start with those. In this section, we will evaluate a few of
2645 these. In another section, we will study the code of several other
2646 buffer-related functions, to see how they were written.
2649 * How to Evaluate:: Typing editing commands or @kbd{C-x C-e}
2651 * Buffer Names:: Buffers and files are different.
2652 * Getting Buffers:: Getting a buffer itself, not merely its name.
2653 * Switching Buffers:: How to change to another buffer.
2654 * Buffer Size & Locations:: Where point is located and the size of
2656 * Evaluation Exercise::
2660 @node How to Evaluate
2661 @unnumberedsec How to Evaluate
2664 @i{Whenever you give an editing command} to Emacs Lisp, such as the
2665 command to move the cursor or to scroll the screen, @i{you are evaluating
2666 an expression,} the first element of which is a function. @i{This is
2669 @cindex @samp{interactive function} defined
2670 @cindex @samp{command} defined
2671 When you type keys, you cause the Lisp interpreter to evaluate an
2672 expression and that is how you get your results. Even typing plain text
2673 involves evaluating an Emacs Lisp function, in this case, one that uses
2674 @code{self-insert-command}, which simply inserts the character you
2675 typed. The functions you evaluate by typing keystrokes are called
2676 @dfn{interactive} functions, or @dfn{commands}; how you make a function
2677 interactive will be illustrated in the chapter on how to write function
2678 definitions. @xref{Interactive, , Making a Function Interactive}.
2680 In addition to typing keyboard commands, we have seen a second way to
2681 evaluate an expression: by positioning the cursor after a list and
2682 typing @kbd{C-x C-e}. This is what we will do in the rest of this
2683 section. There are other ways to evaluate an expression as well; these
2684 will be described as we come to them.
2686 Besides being used for practicing evaluation, the functions shown in the
2687 next few sections are important in their own right. A study of these
2688 functions makes clear the distinction between buffers and files, how to
2689 switch to a buffer, and how to determine a location within it.
2692 @section Buffer Names
2694 @findex buffer-file-name
2696 The two functions, @code{buffer-name} and @code{buffer-file-name}, show
2697 the difference between a file and a buffer. When you evaluate the
2698 following expression, @code{(buffer-name)}, the name of the buffer
2699 appears in the echo area. When you evaluate @code{(buffer-file-name)},
2700 the name of the file to which the buffer refers appears in the echo
2701 area. Usually, the name returned by @code{(buffer-name)} is the same as
2702 the name of the file to which it refers, and the name returned by
2703 @code{(buffer-file-name)} is the full path-name of the file.
2705 A file and a buffer are two different entities. A file is information
2706 recorded permanently in the computer (unless you delete it). A buffer,
2707 on the other hand, is information inside of Emacs that will vanish at
2708 the end of the editing session (or when you kill the buffer). Usually,
2709 a buffer contains information that you have copied from a file; we say
2710 the buffer is @dfn{visiting} that file. This copy is what you work on
2711 and modify. Changes to the buffer do not change the file, until you
2712 save the buffer. When you save the buffer, the buffer is copied to the file
2713 and is thus saved permanently.
2716 If you are reading this in Info inside of GNU Emacs, you can evaluate
2717 each of the following expressions by positioning the cursor after it and
2718 typing @kbd{C-x C-e}.
2729 When I do this in Info, the value returned by evaluating
2730 @code{(buffer-name)} is @file{"*info*"}, and the value returned by
2731 evaluating @code{(buffer-file-name)} is @file{nil}.
2733 On the other hand, while I am writing this document, the value
2734 returned by evaluating @code{(buffer-name)} is
2735 @file{"introduction.texinfo"}, and the value returned by evaluating
2736 @code{(buffer-file-name)} is
2737 @file{"/gnu/work/intro/introduction.texinfo"}.
2739 @cindex @code{nil}, history of word
2740 The former is the name of the buffer and the latter is the name of the
2741 file. In Info, the buffer name is @file{"*info*"}. Info does not
2742 point to any file, so the result of evaluating
2743 @code{(buffer-file-name)} is @file{nil}. The symbol @code{nil} is
2744 from the Latin word for `nothing'; in this case, it means that the
2745 buffer is not associated with any file. (In Lisp, @code{nil} is also
2746 used to mean `false' and is a synonym for the empty list, @code{()}.)
2748 When I am writing, the name of my buffer is
2749 @file{"introduction.texinfo"}. The name of the file to which it
2750 points is @file{"/gnu/work/intro/introduction.texinfo"}.
2752 (In the expressions, the parentheses tell the Lisp interpreter to
2753 treat @w{@code{buffer-name}} and @w{@code{buffer-file-name}} as
2754 functions; without the parentheses, the interpreter would attempt to
2755 evaluate the symbols as variables. @xref{Variables}.)
2757 In spite of the distinction between files and buffers, you will often
2758 find that people refer to a file when they mean a buffer and vice-verse.
2759 Indeed, most people say, ``I am editing a file,'' rather than saying,
2760 ``I am editing a buffer which I will soon save to a file.'' It is
2761 almost always clear from context what people mean. When dealing with
2762 computer programs, however, it is important to keep the distinction in mind,
2763 since the computer is not as smart as a person.
2765 @cindex Buffer, history of word
2766 The word `buffer', by the way, comes from the meaning of the word as a
2767 cushion that deadens the force of a collision. In early computers, a
2768 buffer cushioned the interaction between files and the computer's
2769 central processing unit. The drums or tapes that held a file and the
2770 central processing unit were pieces of equipment that were very
2771 different from each other, working at their own speeds, in spurts. The
2772 buffer made it possible for them to work together effectively.
2773 Eventually, the buffer grew from being an intermediary, a temporary
2774 holding place, to being the place where work is done. This
2775 transformation is rather like that of a small seaport that grew into a
2776 great city: once it was merely the place where cargo was warehoused
2777 temporarily before being loaded onto ships; then it became a business
2778 and cultural center in its own right.
2780 Not all buffers are associated with files. For example, a
2781 @file{*scratch*} buffer does not visit any file. Similarly, a
2782 @file{*Help*} buffer is not associated with any file.
2784 In the old days, when you lacked a @file{~/.emacs} file and started an
2785 Emacs session by typing the command @code{emacs} alone, without naming
2786 any files, Emacs started with the @file{*scratch*} buffer visible.
2787 Nowadays, you will see a splash screen. You can follow one of the
2788 commands suggested on the splash screen, visit a file, or press the
2789 spacebar to reach the @file{*scratch*} buffer.
2791 If you switch to the @file{*scratch*} buffer, type
2792 @code{(buffer-name)}, position the cursor after it, and then type
2793 @kbd{C-x C-e} to evaluate the expression. The name @code{"*scratch*"}
2794 will be returned and will appear in the echo area. @code{"*scratch*"}
2795 is the name of the buffer. When you type @code{(buffer-file-name)} in
2796 the @file{*scratch*} buffer and evaluate that, @code{nil} will appear
2797 in the echo area, just as it does when you evaluate
2798 @code{(buffer-file-name)} in Info.
2800 Incidentally, if you are in the @file{*scratch*} buffer and want the
2801 value returned by an expression to appear in the @file{*scratch*}
2802 buffer itself rather than in the echo area, type @kbd{C-u C-x C-e}
2803 instead of @kbd{C-x C-e}. This causes the value returned to appear
2804 after the expression. The buffer will look like this:
2807 (buffer-name)"*scratch*"
2811 You cannot do this in Info since Info is read-only and it will not allow
2812 you to change the contents of the buffer. But you can do this in any
2813 buffer you can edit; and when you write code or documentation (such as
2814 this book), this feature is very useful.
2816 @node Getting Buffers
2817 @section Getting Buffers
2818 @findex current-buffer
2819 @findex other-buffer
2820 @cindex Getting a buffer
2822 The @code{buffer-name} function returns the @emph{name} of the buffer;
2823 to get the buffer @emph{itself}, a different function is needed: the
2824 @code{current-buffer} function. If you use this function in code, what
2825 you get is the buffer itself.
2827 A name and the object or entity to which the name refers are different
2828 from each other. You are not your name. You are a person to whom
2829 others refer by name. If you ask to speak to George and someone hands you
2830 a card with the letters @samp{G}, @samp{e}, @samp{o}, @samp{r},
2831 @samp{g}, and @samp{e} written on it, you might be amused, but you would
2832 not be satisfied. You do not want to speak to the name, but to the
2833 person to whom the name refers. A buffer is similar: the name of the
2834 scratch buffer is @file{*scratch*}, but the name is not the buffer. To
2835 get a buffer itself, you need to use a function such as
2836 @code{current-buffer}.
2838 However, there is a slight complication: if you evaluate
2839 @code{current-buffer} in an expression on its own, as we will do here,
2840 what you see is a printed representation of the name of the buffer
2841 without the contents of the buffer. Emacs works this way for two
2842 reasons: the buffer may be thousands of lines long---too long to be
2843 conveniently displayed; and, another buffer may have the same contents
2844 but a different name, and it is important to distinguish between them.
2847 Here is an expression containing the function:
2854 If you evaluate this expression in Info in Emacs in the usual way,
2855 @file{#<buffer *info*>} will appear in the echo area. The special
2856 format indicates that the buffer itself is being returned, rather than
2859 Incidentally, while you can type a number or symbol into a program, you
2860 cannot do that with the printed representation of a buffer: the only way
2861 to get a buffer itself is with a function such as @code{current-buffer}.
2863 A related function is @code{other-buffer}. This returns the most
2864 recently selected buffer other than the one you are in currently, not
2865 a printed representation of its name. If you have recently switched
2866 back and forth from the @file{*scratch*} buffer, @code{other-buffer}
2867 will return that buffer.
2870 You can see this by evaluating the expression:
2877 You should see @file{#<buffer *scratch*>} appear in the echo area, or
2878 the name of whatever other buffer you switched back from most
2879 recently@footnote{Actually, by default, if the buffer from which you
2880 just switched is visible to you in another window, @code{other-buffer}
2881 will choose the most recent buffer that you cannot see; this is a
2882 subtlety that I often forget.}.
2884 @node Switching Buffers
2885 @section Switching Buffers
2886 @findex switch-to-buffer
2888 @cindex Switching to a buffer
2890 The @code{other-buffer} function actually provides a buffer when it is
2891 used as an argument to a function that requires one. We can see this
2892 by using @code{other-buffer} and @code{switch-to-buffer} to switch to a
2895 But first, a brief introduction to the @code{switch-to-buffer}
2896 function. When you switched back and forth from Info to the
2897 @file{*scratch*} buffer to evaluate @code{(buffer-name)}, you most
2898 likely typed @kbd{C-x b} and then typed @file{*scratch*}@footnote{Or
2899 rather, to save typing, you probably only typed @kbd{RET} if the
2900 default buffer was @file{*scratch*}, or if it was different, then you
2901 typed just part of the name, such as @code{*sc}, pressed your
2902 @kbd{TAB} key to cause it to expand to the full name, and then typed
2903 @kbd{RET}.} when prompted in the minibuffer for the name of
2904 the buffer to which you wanted to switch. The keystrokes, @kbd{C-x
2905 b}, cause the Lisp interpreter to evaluate the interactive function
2906 @code{switch-to-buffer}. As we said before, this is how Emacs works:
2907 different keystrokes call or run different functions. For example,
2908 @kbd{C-f} calls @code{forward-char}, @kbd{M-e} calls
2909 @code{forward-sentence}, and so on.
2911 By writing @code{switch-to-buffer} in an expression, and giving it a
2912 buffer to switch to, we can switch buffers just the way @kbd{C-x b}
2916 (switch-to-buffer (other-buffer))
2920 The symbol @code{switch-to-buffer} is the first element of the list,
2921 so the Lisp interpreter will treat it as a function and carry out the
2922 instructions that are attached to it. But before doing that, the
2923 interpreter will note that @code{other-buffer} is inside parentheses
2924 and work on that symbol first. @code{other-buffer} is the first (and
2925 in this case, the only) element of this list, so the Lisp interpreter
2926 calls or runs the function. It returns another buffer. Next, the
2927 interpreter runs @code{switch-to-buffer}, passing to it, as an
2928 argument, the other buffer, which is what Emacs will switch to. If
2929 you are reading this in Info, try this now. Evaluate the expression.
2930 (To get back, type @kbd{C-x b @key{RET}}.)@footnote{Remember, this
2931 expression will move you to your most recent other buffer that you
2932 cannot see. If you really want to go to your most recently selected
2933 buffer, even if you can still see it, you need to evaluate the
2934 following more complex expression:
2937 (switch-to-buffer (other-buffer (current-buffer) t))
2941 In this case, the first argument to @code{other-buffer} tells it which
2942 buffer to skip---the current one---and the second argument tells
2943 @code{other-buffer} it is OK to switch to a visible buffer.
2944 In regular use, @code{switch-to-buffer} takes you to an invisible
2945 window since you would most likely use @kbd{C-x o} (@code{other-window})
2946 to go to another visible buffer.}
2948 In the programming examples in later sections of this document, you will
2949 see the function @code{set-buffer} more often than
2950 @code{switch-to-buffer}. This is because of a difference between
2951 computer programs and humans: humans have eyes and expect to see the
2952 buffer on which they are working on their computer terminals. This is
2953 so obvious, it almost goes without saying. However, programs do not
2954 have eyes. When a computer program works on a buffer, that buffer does
2955 not need to be visible on the screen.
2957 @code{switch-to-buffer} is designed for humans and does two different
2958 things: it switches the buffer to which Emacs's attention is directed; and
2959 it switches the buffer displayed in the window to the new buffer.
2960 @code{set-buffer}, on the other hand, does only one thing: it switches
2961 the attention of the computer program to a different buffer. The buffer
2962 on the screen remains unchanged (of course, normally nothing happens
2963 there until the command finishes running).
2965 @cindex @samp{call} defined
2966 Also, we have just introduced another jargon term, the word @dfn{call}.
2967 When you evaluate a list in which the first symbol is a function, you
2968 are calling that function. The use of the term comes from the notion of
2969 the function as an entity that can do something for you if you `call'
2970 it---just as a plumber is an entity who can fix a leak if you call him
2973 @node Buffer Size & Locations
2974 @section Buffer Size and the Location of Point
2975 @cindex Size of buffer
2977 @cindex Point location
2978 @cindex Location of point
2980 Finally, let's look at several rather simple functions,
2981 @code{buffer-size}, @code{point}, @code{point-min}, and
2982 @code{point-max}. These give information about the size of a buffer and
2983 the location of point within it.
2985 The function @code{buffer-size} tells you the size of the current
2986 buffer; that is, the function returns a count of the number of
2987 characters in the buffer.
2994 You can evaluate this in the usual way, by positioning the
2995 cursor after the expression and typing @kbd{C-x C-e}.
2997 @cindex @samp{point} defined
2998 In Emacs, the current position of the cursor is called @dfn{point}.
2999 The expression @code{(point)} returns a number that tells you where the
3000 cursor is located as a count of the number of characters from the
3001 beginning of the buffer up to point.
3004 You can see the character count for point in this buffer by evaluating
3005 the following expression in the usual way:
3012 As I write this, the value of @code{point} is 65724. The @code{point}
3013 function is frequently used in some of the examples later in this
3017 The value of point depends, of course, on its location within the
3018 buffer. If you evaluate point in this spot, the number will be larger:
3025 For me, the value of point in this location is 66043, which means that
3026 there are 319 characters (including spaces) between the two
3027 expressions. (Doubtless, you will see different numbers, since I will
3028 have edited this since I first evaluated point.)
3030 @cindex @samp{narrowing} defined
3031 The function @code{point-min} is somewhat similar to @code{point}, but
3032 it returns the value of the minimum permissible value of point in the
3033 current buffer. This is the number 1 unless @dfn{narrowing} is in
3034 effect. (Narrowing is a mechanism whereby you can restrict yourself,
3035 or a program, to operations on just a part of a buffer.
3036 @xref{Narrowing & Widening, , Narrowing and Widening}.) Likewise, the
3037 function @code{point-max} returns the value of the maximum permissible
3038 value of point in the current buffer.
3040 @node Evaluation Exercise
3043 Find a file with which you are working and move towards its middle.
3044 Find its buffer name, file name, length, and your position in the file.
3046 @node Writing Defuns
3047 @chapter How To Write Function Definitions
3048 @cindex Definition writing
3049 @cindex Function definition writing
3050 @cindex Writing a function definition
3052 When the Lisp interpreter evaluates a list, it looks to see whether the
3053 first symbol on the list has a function definition attached to it; or,
3054 put another way, whether the symbol points to a function definition. If
3055 it does, the computer carries out the instructions in the definition. A
3056 symbol that has a function definition is called, simply, a function
3057 (although, properly speaking, the definition is the function and the
3058 symbol refers to it.)
3061 * Primitive Functions::
3062 * defun:: The @code{defun} special form.
3063 * Install:: Install a function definition.
3064 * Interactive:: Making a function interactive.
3065 * Interactive Options:: Different options for @code{interactive}.
3066 * Permanent Installation:: Installing code permanently.
3067 * let:: Creating and initializing local variables.
3069 * else:: If--then--else expressions.
3070 * Truth & Falsehood:: What Lisp considers false and true.
3071 * save-excursion:: Keeping track of point, mark, and buffer.
3077 @node Primitive Functions
3078 @unnumberedsec An Aside about Primitive Functions
3080 @cindex Primitive functions
3081 @cindex Functions, primitive
3083 @cindex C language primitives
3084 @cindex Primitives written in C
3085 All functions are defined in terms of other functions, except for a few
3086 @dfn{primitive} functions that are written in the C programming
3087 language. When you write functions' definitions, you will write them in
3088 Emacs Lisp and use other functions as your building blocks. Some of the
3089 functions you will use will themselves be written in Emacs Lisp (perhaps
3090 by you) and some will be primitives written in C@. The primitive
3091 functions are used exactly like those written in Emacs Lisp and behave
3092 like them. They are written in C so we can easily run GNU Emacs on any
3093 computer that has sufficient power and can run C.
3095 Let me re-emphasize this: when you write code in Emacs Lisp, you do not
3096 distinguish between the use of functions written in C and the use of
3097 functions written in Emacs Lisp. The difference is irrelevant. I
3098 mention the distinction only because it is interesting to know. Indeed,
3099 unless you investigate, you won't know whether an already-written
3100 function is written in Emacs Lisp or C.
3103 @section The @code{defun} Special Form
3105 @cindex Special form of @code{defun}
3107 @cindex @samp{function definition} defined
3108 In Lisp, a symbol such as @code{mark-whole-buffer} has code attached to
3109 it that tells the computer what to do when the function is called.
3110 This code is called the @dfn{function definition} and is created by
3111 evaluating a Lisp expression that starts with the symbol @code{defun}
3112 (which is an abbreviation for @emph{define function}). Because
3113 @code{defun} does not evaluate its arguments in the usual way, it is
3114 called a @dfn{special form}.
3116 In subsequent sections, we will look at function definitions from the
3117 Emacs source code, such as @code{mark-whole-buffer}. In this section,
3118 we will describe a simple function definition so you can see how it
3119 looks. This function definition uses arithmetic because it makes for a
3120 simple example. Some people dislike examples using arithmetic; however,
3121 if you are such a person, do not despair. Hardly any of the code we
3122 will study in the remainder of this introduction involves arithmetic or
3123 mathematics. The examples mostly involve text in one way or another.
3125 A function definition has up to five parts following the word
3130 The name of the symbol to which the function definition should be
3134 A list of the arguments that will be passed to the function. If no
3135 arguments will be passed to the function, this is an empty list,
3139 Documentation describing the function. (Technically optional, but
3140 strongly recommended.)
3143 Optionally, an expression to make the function interactive so you can
3144 use it by typing @kbd{M-x} and then the name of the function; or by
3145 typing an appropriate key or keychord.
3147 @cindex @samp{body} defined
3149 The code that instructs the computer what to do: the @dfn{body} of the
3150 function definition.
3153 It is helpful to think of the five parts of a function definition as
3154 being organized in a template, with slots for each part:
3158 (defun @var{function-name} (@var{arguments}@dots{})
3159 "@var{optional-documentation}@dots{}"
3160 (interactive @var{argument-passing-info}) ; @r{optional}
3165 As an example, here is the code for a function that multiplies its
3166 argument by 7. (This example is not interactive. @xref{Interactive,
3167 , Making a Function Interactive}, for that information.)
3171 (defun multiply-by-seven (number)
3172 "Multiply NUMBER by seven."
3177 This definition begins with a parenthesis and the symbol @code{defun},
3178 followed by the name of the function.
3180 @cindex @samp{argument list} defined
3181 The name of the function is followed by a list that contains the
3182 arguments that will be passed to the function. This list is called
3183 the @dfn{argument list}. In this example, the list has only one
3184 element, the symbol, @code{number}. When the function is used, the
3185 symbol will be bound to the value that is used as the argument to the
3188 Instead of choosing the word @code{number} for the name of the argument,
3189 I could have picked any other name. For example, I could have chosen
3190 the word @code{multiplicand}. I picked the word `number' because it
3191 tells what kind of value is intended for this slot; but I could just as
3192 well have chosen the word `multiplicand' to indicate the role that the
3193 value placed in this slot will play in the workings of the function. I
3194 could have called it @code{foogle}, but that would have been a bad
3195 choice because it would not tell humans what it means. The choice of
3196 name is up to the programmer and should be chosen to make the meaning of
3199 Indeed, you can choose any name you wish for a symbol in an argument
3200 list, even the name of a symbol used in some other function: the name
3201 you use in an argument list is private to that particular definition.
3202 In that definition, the name refers to a different entity than any use
3203 of the same name outside the function definition. Suppose you have a
3204 nick-name `Shorty' in your family; when your family members refer to
3205 `Shorty', they mean you. But outside your family, in a movie, for
3206 example, the name `Shorty' refers to someone else. Because a name in an
3207 argument list is private to the function definition, you can change the
3208 value of such a symbol inside the body of a function without changing
3209 its value outside the function. The effect is similar to that produced
3210 by a @code{let} expression. (@xref{let, , @code{let}}.)
3213 Note also that we discuss the word `number' in two different ways: as a
3214 symbol that appears in the code, and as the name of something that will
3215 be replaced by a something else during the evaluation of the function.
3216 In the first case, @code{number} is a symbol, not a number; it happens
3217 that within the function, it is a variable who value is the number in
3218 question, but our primary interest in it is as a symbol. On the other
3219 hand, when we are talking about the function, our interest is that we
3220 will substitute a number for the word @var{number}. To keep this
3221 distinction clear, we use different typography for the two
3222 circumstances. When we talk about this function, or about how it works,
3223 we refer to this number by writing @var{number}. In the function
3224 itself, we refer to it by writing @code{number}.
3227 The argument list is followed by the documentation string that
3228 describes the function. This is what you see when you type
3229 @w{@kbd{C-h f}} and the name of a function. Incidentally, when you
3230 write a documentation string like this, you should make the first line
3231 a complete sentence since some commands, such as @code{apropos}, print
3232 only the first line of a multi-line documentation string. Also, you
3233 should not indent the second line of a documentation string, if you
3234 have one, because that looks odd when you use @kbd{C-h f}
3235 (@code{describe-function}). The documentation string is optional, but
3236 it is so useful, it should be included in almost every function you
3239 @findex * @r{(multiplication)}
3240 The third line of the example consists of the body of the function
3241 definition. (Most functions' definitions, of course, are longer than
3242 this.) In this function, the body is the list, @code{(* 7 number)}, which
3243 says to multiply the value of @var{number} by 7. (In Emacs Lisp,
3244 @code{*} is the function for multiplication, just as @code{+} is the
3245 function for addition.)
3247 When you use the @code{multiply-by-seven} function, the argument
3248 @code{number} evaluates to the actual number you want used. Here is an
3249 example that shows how @code{multiply-by-seven} is used; but don't try
3250 to evaluate this yet!
3253 (multiply-by-seven 3)
3257 The symbol @code{number}, specified in the function definition in the
3258 next section, is given or ``bound to'' the value 3 in the actual use of
3259 the function. Note that although @code{number} was inside parentheses
3260 in the function definition, the argument passed to the
3261 @code{multiply-by-seven} function is not in parentheses. The
3262 parentheses are written in the function definition so the computer can
3263 figure out where the argument list ends and the rest of the function
3266 If you evaluate this example, you are likely to get an error message.
3267 (Go ahead, try it!) This is because we have written the function
3268 definition, but not yet told the computer about the definition---we have
3269 not yet installed (or `loaded') the function definition in Emacs.
3270 Installing a function is the process that tells the Lisp interpreter the
3271 definition of the function. Installation is described in the next
3275 @section Install a Function Definition
3276 @cindex Install a Function Definition
3277 @cindex Definition installation
3278 @cindex Function definition installation
3280 If you are reading this inside of Info in Emacs, you can try out the
3281 @code{multiply-by-seven} function by first evaluating the function
3282 definition and then evaluating @code{(multiply-by-seven 3)}. A copy of
3283 the function definition follows. Place the cursor after the last
3284 parenthesis of the function definition and type @kbd{C-x C-e}. When you
3285 do this, @code{multiply-by-seven} will appear in the echo area. (What
3286 this means is that when a function definition is evaluated, the value it
3287 returns is the name of the defined function.) At the same time, this
3288 action installs the function definition.
3292 (defun multiply-by-seven (number)
3293 "Multiply NUMBER by seven."
3299 By evaluating this @code{defun}, you have just installed
3300 @code{multiply-by-seven} in Emacs. The function is now just as much a
3301 part of Emacs as @code{forward-word} or any other editing function you
3302 use. (@code{multiply-by-seven} will stay installed until you quit
3303 Emacs. To reload code automatically whenever you start Emacs, see
3304 @ref{Permanent Installation, , Installing Code Permanently}.)
3307 * Effect of installation::
3308 * Change a defun:: How to change a function definition.
3312 @node Effect of installation
3313 @unnumberedsubsec The effect of installation
3316 You can see the effect of installing @code{multiply-by-seven} by
3317 evaluating the following sample. Place the cursor after the following
3318 expression and type @kbd{C-x C-e}. The number 21 will appear in the
3322 (multiply-by-seven 3)
3325 If you wish, you can read the documentation for the function by typing
3326 @kbd{C-h f} (@code{describe-function}) and then the name of the
3327 function, @code{multiply-by-seven}. When you do this, a
3328 @file{*Help*} window will appear on your screen that says:
3332 multiply-by-seven is a Lisp function.
3333 (multiply-by-seven NUMBER)
3335 Multiply NUMBER by seven.
3340 (To return to a single window on your screen, type @kbd{C-x 1}.)
3342 @node Change a defun
3343 @subsection Change a Function Definition
3344 @cindex Changing a function definition
3345 @cindex Function definition, how to change
3346 @cindex Definition, how to change
3348 If you want to change the code in @code{multiply-by-seven}, just rewrite
3349 it. To install the new version in place of the old one, evaluate the
3350 function definition again. This is how you modify code in Emacs. It is
3353 As an example, you can change the @code{multiply-by-seven} function to
3354 add the number to itself seven times instead of multiplying the number
3355 by seven. It produces the same answer, but by a different path. At
3356 the same time, we will add a comment to the code; a comment is text
3357 that the Lisp interpreter ignores, but that a human reader may find
3358 useful or enlightening. The comment is that this is the ``second
3363 (defun multiply-by-seven (number) ; @r{Second version.}
3364 "Multiply NUMBER by seven."
3365 (+ number number number number number number number))
3369 @cindex Comments in Lisp code
3370 The comment follows a semicolon, @samp{;}. In Lisp, everything on a
3371 line that follows a semicolon is a comment. The end of the line is the
3372 end of the comment. To stretch a comment over two or more lines, begin
3373 each line with a semicolon.
3375 @xref{Beginning a .emacs File, , Beginning a @file{.emacs}
3376 File}, and @ref{Comments, , Comments, elisp, The GNU Emacs Lisp
3377 Reference Manual}, for more about comments.
3379 You can install this version of the @code{multiply-by-seven} function by
3380 evaluating it in the same way you evaluated the first function: place
3381 the cursor after the last parenthesis and type @kbd{C-x C-e}.
3383 In summary, this is how you write code in Emacs Lisp: you write a
3384 function; install it; test it; and then make fixes or enhancements and
3388 @section Make a Function Interactive
3389 @cindex Interactive functions
3392 You make a function interactive by placing a list that begins with
3393 the special form @code{interactive} immediately after the
3394 documentation. A user can invoke an interactive function by typing
3395 @kbd{M-x} and then the name of the function; or by typing the keys to
3396 which it is bound, for example, by typing @kbd{C-n} for
3397 @code{next-line} or @kbd{C-x h} for @code{mark-whole-buffer}.
3399 Interestingly, when you call an interactive function interactively,
3400 the value returned is not automatically displayed in the echo area.
3401 This is because you often call an interactive function for its side
3402 effects, such as moving forward by a word or line, and not for the
3403 value returned. If the returned value were displayed in the echo area
3404 each time you typed a key, it would be very distracting.
3407 * Interactive multiply-by-seven:: An overview.
3408 * multiply-by-seven in detail:: The interactive version.
3412 @node Interactive multiply-by-seven
3413 @unnumberedsubsec An Interactive @code{multiply-by-seven}, An Overview
3416 Both the use of the special form @code{interactive} and one way to
3417 display a value in the echo area can be illustrated by creating an
3418 interactive version of @code{multiply-by-seven}.
3425 (defun multiply-by-seven (number) ; @r{Interactive version.}
3426 "Multiply NUMBER by seven."
3428 (message "The result is %d" (* 7 number)))
3433 You can install this code by placing your cursor after it and typing
3434 @kbd{C-x C-e}. The name of the function will appear in your echo area.
3435 Then, you can use this code by typing @kbd{C-u} and a number and then
3436 typing @kbd{M-x multiply-by-seven} and pressing @key{RET}. The phrase
3437 @samp{The result is @dots{}} followed by the product will appear in the
3440 Speaking more generally, you invoke a function like this in either of two
3445 By typing a prefix argument that contains the number to be passed, and
3446 then typing @kbd{M-x} and the name of the function, as with
3447 @kbd{C-u 3 M-x forward-sentence}; or,
3450 By typing whatever key or keychord the function is bound to, as with
3455 Both the examples just mentioned work identically to move point forward
3456 three sentences. (Since @code{multiply-by-seven} is not bound to a key,
3457 it could not be used as an example of key binding.)
3459 (@xref{Keybindings, , Some Keybindings}, to learn how to bind a command
3462 A prefix argument is passed to an interactive function by typing the
3463 @key{META} key followed by a number, for example, @kbd{M-3 M-e}, or by
3464 typing @kbd{C-u} and then a number, for example, @kbd{C-u 3 M-e} (if you
3465 type @kbd{C-u} without a number, it defaults to 4).
3467 @node multiply-by-seven in detail
3468 @subsection An Interactive @code{multiply-by-seven}
3470 Let's look at the use of the special form @code{interactive} and then at
3471 the function @code{message} in the interactive version of
3472 @code{multiply-by-seven}. You will recall that the function definition
3477 (defun multiply-by-seven (number) ; @r{Interactive version.}
3478 "Multiply NUMBER by seven."
3480 (message "The result is %d" (* 7 number)))
3484 In this function, the expression, @code{(interactive "p")}, is a list of
3485 two elements. The @code{"p"} tells Emacs to pass the prefix argument to
3486 the function and use its value for the argument of the function.
3489 The argument will be a number. This means that the symbol
3490 @code{number} will be bound to a number in the line:
3493 (message "The result is %d" (* 7 number))
3498 For example, if your prefix argument is 5, the Lisp interpreter will
3499 evaluate the line as if it were:
3502 (message "The result is %d" (* 7 5))
3506 (If you are reading this in GNU Emacs, you can evaluate this expression
3507 yourself.) First, the interpreter will evaluate the inner list, which
3508 is @code{(* 7 5)}. This returns a value of 35. Next, it
3509 will evaluate the outer list, passing the values of the second and
3510 subsequent elements of the list to the function @code{message}.
3512 As we have seen, @code{message} is an Emacs Lisp function especially
3513 designed for sending a one line message to a user. (@xref{message, ,
3514 The @code{message} function}.) In summary, the @code{message}
3515 function prints its first argument in the echo area as is, except for
3516 occurrences of @samp{%d} or @samp{%s} (and various other %-sequences
3517 which we have not mentioned). When it sees a control sequence, the
3518 function looks to the second or subsequent arguments and prints the
3519 value of the argument in the location in the string where the control
3520 sequence is located.
3522 In the interactive @code{multiply-by-seven} function, the control string
3523 is @samp{%d}, which requires a number, and the value returned by
3524 evaluating @code{(* 7 5)} is the number 35. Consequently, the number 35
3525 is printed in place of the @samp{%d} and the message is @samp{The result
3528 (Note that when you call the function @code{multiply-by-seven}, the
3529 message is printed without quotes, but when you call @code{message}, the
3530 text is printed in double quotes. This is because the value returned by
3531 @code{message} is what appears in the echo area when you evaluate an
3532 expression whose first element is @code{message}; but when embedded in a
3533 function, @code{message} prints the text as a side effect without
3536 @node Interactive Options
3537 @section Different Options for @code{interactive}
3538 @cindex Options for @code{interactive}
3539 @cindex Interactive options
3541 In the example, @code{multiply-by-seven} used @code{"p"} as the
3542 argument to @code{interactive}. This argument told Emacs to interpret
3543 your typing either @kbd{C-u} followed by a number or @key{META}
3544 followed by a number as a command to pass that number to the function
3545 as its argument. Emacs has more than twenty characters predefined for
3546 use with @code{interactive}. In almost every case, one of these
3547 options will enable you to pass the right information interactively to
3548 a function. (@xref{Interactive Codes, , Code Characters for
3549 @code{interactive}, elisp, The GNU Emacs Lisp Reference Manual}.)
3552 Consider the function @code{zap-to-char}. Its interactive expression
3556 (interactive "p\ncZap to char: ")
3559 The first part of the argument to @code{interactive} is @samp{p}, with
3560 which you are already familiar. This argument tells Emacs to
3561 interpret a `prefix', as a number to be passed to the function. You
3562 can specify a prefix either by typing @kbd{C-u} followed by a number
3563 or by typing @key{META} followed by a number. The prefix is the
3564 number of specified characters. Thus, if your prefix is three and the
3565 specified character is @samp{x}, then you will delete all the text up
3566 to and including the third next @samp{x}. If you do not set a prefix,
3567 then you delete all the text up to and including the specified
3568 character, but no more.
3570 The @samp{c} tells the function the name of the character to which to delete.
3572 More formally, a function with two or more arguments can have
3573 information passed to each argument by adding parts to the string that
3574 follows @code{interactive}. When you do this, the information is
3575 passed to each argument in the same order it is specified in the
3576 @code{interactive} list. In the string, each part is separated from
3577 the next part by a @samp{\n}, which is a newline. For example, you
3578 can follow @samp{p} with a @samp{\n} and an @samp{cZap to char:@: }.
3579 This causes Emacs to pass the value of the prefix argument (if there
3580 is one) and the character.
3582 In this case, the function definition looks like the following, where
3583 @code{arg} and @code{char} are the symbols to which @code{interactive}
3584 binds the prefix argument and the specified character:
3588 (defun @var{name-of-function} (arg char)
3589 "@var{documentation}@dots{}"
3590 (interactive "p\ncZap to char: ")
3591 @var{body-of-function}@dots{})
3596 (The space after the colon in the prompt makes it look better when you
3597 are prompted. @xref{copy-to-buffer, , The Definition of
3598 @code{copy-to-buffer}}, for an example.)
3600 When a function does not take arguments, @code{interactive} does not
3601 require any. Such a function contains the simple expression
3602 @code{(interactive)}. The @code{mark-whole-buffer} function is like
3605 Alternatively, if the special letter-codes are not right for your
3606 application, you can pass your own arguments to @code{interactive} as
3609 @xref{append-to-buffer, , The Definition of @code{append-to-buffer}},
3610 for an example. @xref{Using Interactive, , Using @code{Interactive},
3611 elisp, The GNU Emacs Lisp Reference Manual}, for a more complete
3612 explanation about this technique.
3614 @node Permanent Installation
3615 @section Install Code Permanently
3616 @cindex Install code permanently
3617 @cindex Permanent code installation
3618 @cindex Code installation
3620 When you install a function definition by evaluating it, it will stay
3621 installed until you quit Emacs. The next time you start a new session
3622 of Emacs, the function will not be installed unless you evaluate the
3623 function definition again.
3625 At some point, you may want to have code installed automatically
3626 whenever you start a new session of Emacs. There are several ways of
3631 If you have code that is just for yourself, you can put the code for the
3632 function definition in your @file{.emacs} initialization file. When you
3633 start Emacs, your @file{.emacs} file is automatically evaluated and all
3634 the function definitions within it are installed.
3635 @xref{Emacs Initialization, , Your @file{.emacs} File}.
3638 Alternatively, you can put the function definitions that you want
3639 installed in one or more files of their own and use the @code{load}
3640 function to cause Emacs to evaluate and thereby install each of the
3641 functions in the files.
3642 @xref{Loading Files, , Loading Files}.
3645 Thirdly, if you have code that your whole site will use, it is usual
3646 to put it in a file called @file{site-init.el} that is loaded when
3647 Emacs is built. This makes the code available to everyone who uses
3648 your machine. (See the @file{INSTALL} file that is part of the Emacs
3652 Finally, if you have code that everyone who uses Emacs may want, you
3653 can post it on a computer network or send a copy to the Free Software
3654 Foundation. (When you do this, please license the code and its
3655 documentation under a license that permits other people to run, copy,
3656 study, modify, and redistribute the code and which protects you from
3657 having your work taken from you.) If you send a copy of your code to
3658 the Free Software Foundation, and properly protect yourself and
3659 others, it may be included in the next release of Emacs. In large
3660 part, this is how Emacs has grown over the past years, by donations.
3666 The @code{let} expression is a special form in Lisp that you will need
3667 to use in most function definitions.
3669 @code{let} is used to attach or bind a symbol to a value in such a way
3670 that the Lisp interpreter will not confuse the variable with a
3671 variable of the same name that is not part of the function.
3673 To understand why the @code{let} special form is necessary, consider
3674 the situation in which you own a home that you generally refer to as
3675 `the house', as in the sentence, ``The house needs painting.'' If you
3676 are visiting a friend and your host refers to `the house', he is
3677 likely to be referring to @emph{his} house, not yours, that is, to a
3680 If your friend is referring to his house and you think he is referring
3681 to your house, you may be in for some confusion. The same thing could
3682 happen in Lisp if a variable that is used inside of one function has
3683 the same name as a variable that is used inside of another function,
3684 and the two are not intended to refer to the same value. The
3685 @code{let} special form prevents this kind of confusion.
3688 * Prevent confusion::
3689 * Parts of let Expression::
3690 * Sample let Expression::
3691 * Uninitialized let Variables::
3695 @node Prevent confusion
3696 @unnumberedsubsec @code{let} Prevents Confusion
3699 @cindex @samp{local variable} defined
3700 @cindex @samp{variable, local}, defined
3701 The @code{let} special form prevents confusion. @code{let} creates a
3702 name for a @dfn{local variable} that overshadows any use of the same
3703 name outside the @code{let} expression. This is like understanding
3704 that whenever your host refers to `the house', he means his house, not
3705 yours. (Symbols used in argument lists work the same way.
3706 @xref{defun, , The @code{defun} Special Form}.)
3708 Local variables created by a @code{let} expression retain their value
3709 @emph{only} within the @code{let} expression itself (and within
3710 expressions called within the @code{let} expression); the local
3711 variables have no effect outside the @code{let} expression.
3713 Another way to think about @code{let} is that it is like a @code{setq}
3714 that is temporary and local. The values set by @code{let} are
3715 automatically undone when the @code{let} is finished. The setting
3716 only affects expressions that are inside the bounds of the @code{let}
3717 expression. In computer science jargon, we would say ``the binding of
3718 a symbol is visible only in functions called in the @code{let} form;
3719 in Emacs Lisp, scoping is dynamic, not lexical.''
3721 @code{let} can create more than one variable at once. Also,
3722 @code{let} gives each variable it creates an initial value, either a
3723 value specified by you, or @code{nil}. (In the jargon, this is called
3724 `binding the variable to the value'.) After @code{let} has created
3725 and bound the variables, it executes the code in the body of the
3726 @code{let}, and returns the value of the last expression in the body,
3727 as the value of the whole @code{let} expression. (`Execute' is a jargon
3728 term that means to evaluate a list; it comes from the use of the word
3729 meaning `to give practical effect to' (@cite{Oxford English
3730 Dictionary}). Since you evaluate an expression to perform an action,
3731 `execute' has evolved as a synonym to `evaluate'.)
3733 @node Parts of let Expression
3734 @subsection The Parts of a @code{let} Expression
3735 @cindex @code{let} expression, parts of
3736 @cindex Parts of @code{let} expression
3738 @cindex @samp{varlist} defined
3739 A @code{let} expression is a list of three parts. The first part is
3740 the symbol @code{let}. The second part is a list, called a
3741 @dfn{varlist}, each element of which is either a symbol by itself or a
3742 two-element list, the first element of which is a symbol. The third
3743 part of the @code{let} expression is the body of the @code{let}. The
3744 body usually consists of one or more lists.
3747 A template for a @code{let} expression looks like this:
3750 (let @var{varlist} @var{body}@dots{})
3754 The symbols in the varlist are the variables that are given initial
3755 values by the @code{let} special form. Symbols by themselves are given
3756 the initial value of @code{nil}; and each symbol that is the first
3757 element of a two-element list is bound to the value that is returned
3758 when the Lisp interpreter evaluates the second element.
3760 Thus, a varlist might look like this: @code{(thread (needles 3))}. In
3761 this case, in a @code{let} expression, Emacs binds the symbol
3762 @code{thread} to an initial value of @code{nil}, and binds the symbol
3763 @code{needles} to an initial value of 3.
3765 When you write a @code{let} expression, what you do is put the
3766 appropriate expressions in the slots of the @code{let} expression
3769 If the varlist is composed of two-element lists, as is often the case,
3770 the template for the @code{let} expression looks like this:
3774 (let ((@var{variable} @var{value})
3775 (@var{variable} @var{value})
3781 @node Sample let Expression
3782 @subsection Sample @code{let} Expression
3783 @cindex Sample @code{let} expression
3784 @cindex @code{let} expression sample
3786 The following expression creates and gives initial values
3787 to the two variables @code{zebra} and @code{tiger}. The body of the
3788 @code{let} expression is a list which calls the @code{message} function.
3792 (let ((zebra 'stripes)
3794 (message "One kind of animal has %s and another is %s."
3799 Here, the varlist is @code{((zebra 'stripes) (tiger 'fierce))}.
3801 The two variables are @code{zebra} and @code{tiger}. Each variable is
3802 the first element of a two-element list and each value is the second
3803 element of its two-element list. In the varlist, Emacs binds the
3804 variable @code{zebra} to the value @code{stripes}@footnote{According
3805 to Jared Diamond in @cite{Guns, Germs, and Steel}, ``@dots{} zebras
3806 become impossibly dangerous as they grow older'' but the claim here is
3807 that they do not become fierce like a tiger. (1997, W. W. Norton and
3808 Co., ISBN 0-393-03894-2, page 171)}, and binds the
3809 variable @code{tiger} to the value @code{fierce}. In this example,
3810 both values are symbols preceded by a quote. The values could just as
3811 well have been another list or a string. The body of the @code{let}
3812 follows after the list holding the variables. In this example, the
3813 body is a list that uses the @code{message} function to print a string
3817 You may evaluate the example in the usual fashion, by placing the
3818 cursor after the last parenthesis and typing @kbd{C-x C-e}. When you do
3819 this, the following will appear in the echo area:
3822 "One kind of animal has stripes and another is fierce."
3825 As we have seen before, the @code{message} function prints its first
3826 argument, except for @samp{%s}. In this example, the value of the variable
3827 @code{zebra} is printed at the location of the first @samp{%s} and the
3828 value of the variable @code{tiger} is printed at the location of the
3831 @node Uninitialized let Variables
3832 @subsection Uninitialized Variables in a @code{let} Statement
3833 @cindex Uninitialized @code{let} variables
3834 @cindex @code{let} variables uninitialized
3836 If you do not bind the variables in a @code{let} statement to specific
3837 initial values, they will automatically be bound to an initial value of
3838 @code{nil}, as in the following expression:
3847 "Here are %d variables with %s, %s, and %s value."
3848 birch pine fir oak))
3853 Here, the varlist is @code{((birch 3) pine fir (oak 'some))}.
3856 If you evaluate this expression in the usual way, the following will
3857 appear in your echo area:
3860 "Here are 3 variables with nil, nil, and some value."
3864 In this example, Emacs binds the symbol @code{birch} to the number 3,
3865 binds the symbols @code{pine} and @code{fir} to @code{nil}, and binds
3866 the symbol @code{oak} to the value @code{some}.
3868 Note that in the first part of the @code{let}, the variables @code{pine}
3869 and @code{fir} stand alone as atoms that are not surrounded by
3870 parentheses; this is because they are being bound to @code{nil}, the
3871 empty list. But @code{oak} is bound to @code{some} and so is a part of
3872 the list @code{(oak 'some)}. Similarly, @code{birch} is bound to the
3873 number 3 and so is in a list with that number. (Since a number
3874 evaluates to itself, the number does not need to be quoted. Also, the
3875 number is printed in the message using a @samp{%d} rather than a
3876 @samp{%s}.) The four variables as a group are put into a list to
3877 delimit them from the body of the @code{let}.
3880 @section The @code{if} Special Form
3882 @cindex Conditional with @code{if}
3884 A third special form, in addition to @code{defun} and @code{let}, is the
3885 conditional @code{if}. This form is used to instruct the computer to
3886 make decisions. You can write function definitions without using
3887 @code{if}, but it is used often enough, and is important enough, to be
3888 included here. It is used, for example, in the code for the
3889 function @code{beginning-of-buffer}.
3891 The basic idea behind an @code{if}, is that ``@emph{if} a test is true,
3892 @emph{then} an expression is evaluated.'' If the test is not true, the
3893 expression is not evaluated. For example, you might make a decision
3894 such as, ``if it is warm and sunny, then go to the beach!''
3897 * if in more detail::
3898 * type-of-animal in detail:: An example of an @code{if} expression.
3902 @node if in more detail
3903 @unnumberedsubsec @code{if} in more detail
3906 @cindex @samp{if-part} defined
3907 @cindex @samp{then-part} defined
3908 An @code{if} expression written in Lisp does not use the word `then';
3909 the test and the action are the second and third elements of the list
3910 whose first element is @code{if}. Nonetheless, the test part of an
3911 @code{if} expression is often called the @dfn{if-part} and the second
3912 argument is often called the @dfn{then-part}.
3914 Also, when an @code{if} expression is written, the true-or-false-test
3915 is usually written on the same line as the symbol @code{if}, but the
3916 action to carry out if the test is true, the ``then-part'', is written
3917 on the second and subsequent lines. This makes the @code{if}
3918 expression easier to read.
3922 (if @var{true-or-false-test}
3923 @var{action-to-carry-out-if-test-is-true})
3928 The true-or-false-test will be an expression that
3929 is evaluated by the Lisp interpreter.
3931 Here is an example that you can evaluate in the usual manner. The test
3932 is whether the number 5 is greater than the number 4. Since it is, the
3933 message @samp{5 is greater than 4!} will be printed.
3937 (if (> 5 4) ; @r{if-part}
3938 (message "5 is greater than 4!")) ; @r{then-part}
3943 (The function @code{>} tests whether its first argument is greater than
3944 its second argument and returns true if it is.)
3945 @findex > (greater than)
3947 Of course, in actual use, the test in an @code{if} expression will not
3948 be fixed for all time as it is by the expression @code{(> 5 4)}.
3949 Instead, at least one of the variables used in the test will be bound to
3950 a value that is not known ahead of time. (If the value were known ahead
3951 of time, we would not need to run the test!)
3953 For example, the value may be bound to an argument of a function
3954 definition. In the following function definition, the character of the
3955 animal is a value that is passed to the function. If the value bound to
3956 @code{characteristic} is @code{fierce}, then the message, @samp{It's a
3957 tiger!} will be printed; otherwise, @code{nil} will be returned.
3961 (defun type-of-animal (characteristic)
3962 "Print message in echo area depending on CHARACTERISTIC.
3963 If the CHARACTERISTIC is the symbol `fierce',
3964 then warn of a tiger."
3965 (if (equal characteristic 'fierce)
3966 (message "It's a tiger!")))
3972 If you are reading this inside of GNU Emacs, you can evaluate the
3973 function definition in the usual way to install it in Emacs, and then you
3974 can evaluate the following two expressions to see the results:
3978 (type-of-animal 'fierce)
3980 (type-of-animal 'zebra)
3985 @c Following sentences rewritten to prevent overfull hbox.
3987 When you evaluate @code{(type-of-animal 'fierce)}, you will see the
3988 following message printed in the echo area: @code{"It's a tiger!"}; and
3989 when you evaluate @code{(type-of-animal 'zebra)} you will see @code{nil}
3990 printed in the echo area.
3992 @node type-of-animal in detail
3993 @subsection The @code{type-of-animal} Function in Detail
3995 Let's look at the @code{type-of-animal} function in detail.
3997 The function definition for @code{type-of-animal} was written by filling
3998 the slots of two templates, one for a function definition as a whole, and
3999 a second for an @code{if} expression.
4002 The template for every function that is not interactive is:
4006 (defun @var{name-of-function} (@var{argument-list})
4007 "@var{documentation}@dots{}"
4013 The parts of the function that match this template look like this:
4017 (defun type-of-animal (characteristic)
4018 "Print message in echo area depending on CHARACTERISTIC.
4019 If the CHARACTERISTIC is the symbol `fierce',
4020 then warn of a tiger."
4021 @var{body: the} @code{if} @var{expression})
4025 The name of function is @code{type-of-animal}; it is passed the value
4026 of one argument. The argument list is followed by a multi-line
4027 documentation string. The documentation string is included in the
4028 example because it is a good habit to write documentation string for
4029 every function definition. The body of the function definition
4030 consists of the @code{if} expression.
4033 The template for an @code{if} expression looks like this:
4037 (if @var{true-or-false-test}
4038 @var{action-to-carry-out-if-the-test-returns-true})
4043 In the @code{type-of-animal} function, the code for the @code{if}
4048 (if (equal characteristic 'fierce)
4049 (message "It's a tiger!")))
4054 Here, the true-or-false-test is the expression:
4057 (equal characteristic 'fierce)
4061 In Lisp, @code{equal} is a function that determines whether its first
4062 argument is equal to its second argument. The second argument is the
4063 quoted symbol @code{'fierce} and the first argument is the value of the
4064 symbol @code{characteristic}---in other words, the argument passed to
4067 In the first exercise of @code{type-of-animal}, the argument
4068 @code{fierce} is passed to @code{type-of-animal}. Since @code{fierce}
4069 is equal to @code{fierce}, the expression, @code{(equal characteristic
4070 'fierce)}, returns a value of true. When this happens, the @code{if}
4071 evaluates the second argument or then-part of the @code{if}:
4072 @code{(message "It's tiger!")}.
4074 On the other hand, in the second exercise of @code{type-of-animal}, the
4075 argument @code{zebra} is passed to @code{type-of-animal}. @code{zebra}
4076 is not equal to @code{fierce}, so the then-part is not evaluated and
4077 @code{nil} is returned by the @code{if} expression.
4080 @section If--then--else Expressions
4083 An @code{if} expression may have an optional third argument, called
4084 the @dfn{else-part}, for the case when the true-or-false-test returns
4085 false. When this happens, the second argument or then-part of the
4086 overall @code{if} expression is @emph{not} evaluated, but the third or
4087 else-part @emph{is} evaluated. You might think of this as the cloudy
4088 day alternative for the decision ``if it is warm and sunny, then go to
4089 the beach, else read a book!''.
4091 The word ``else'' is not written in the Lisp code; the else-part of an
4092 @code{if} expression comes after the then-part. In the written Lisp, the
4093 else-part is usually written to start on a line of its own and is
4094 indented less than the then-part:
4098 (if @var{true-or-false-test}
4099 @var{action-to-carry-out-if-the-test-returns-true}
4100 @var{action-to-carry-out-if-the-test-returns-false})
4104 For example, the following @code{if} expression prints the message @samp{4
4105 is not greater than 5!} when you evaluate it in the usual way:
4109 (if (> 4 5) ; @r{if-part}
4110 (message "4 falsely greater than 5!") ; @r{then-part}
4111 (message "4 is not greater than 5!")) ; @r{else-part}
4116 Note that the different levels of indentation make it easy to
4117 distinguish the then-part from the else-part. (GNU Emacs has several
4118 commands that automatically indent @code{if} expressions correctly.
4119 @xref{Typing Lists, , GNU Emacs Helps You Type Lists}.)
4121 We can extend the @code{type-of-animal} function to include an
4122 else-part by simply incorporating an additional part to the @code{if}
4126 You can see the consequences of doing this if you evaluate the following
4127 version of the @code{type-of-animal} function definition to install it
4128 and then evaluate the two subsequent expressions to pass different
4129 arguments to the function.
4133 (defun type-of-animal (characteristic) ; @r{Second version.}
4134 "Print message in echo area depending on CHARACTERISTIC.
4135 If the CHARACTERISTIC is the symbol `fierce',
4136 then warn of a tiger;
4137 else say it's not fierce."
4138 (if (equal characteristic 'fierce)
4139 (message "It's a tiger!")
4140 (message "It's not fierce!")))
4147 (type-of-animal 'fierce)
4149 (type-of-animal 'zebra)
4154 @c Following sentence rewritten to prevent overfull hbox.
4156 When you evaluate @code{(type-of-animal 'fierce)}, you will see the
4157 following message printed in the echo area: @code{"It's a tiger!"}; but
4158 when you evaluate @code{(type-of-animal 'zebra)}, you will see
4159 @code{"It's not fierce!"}.
4161 (Of course, if the @var{characteristic} were @code{ferocious}, the
4162 message @code{"It's not fierce!"} would be printed; and it would be
4163 misleading! When you write code, you need to take into account the
4164 possibility that some such argument will be tested by the @code{if}
4165 and write your program accordingly.)
4167 @node Truth & Falsehood
4168 @section Truth and Falsehood in Emacs Lisp
4169 @cindex Truth and falsehood in Emacs Lisp
4170 @cindex Falsehood and truth in Emacs Lisp
4173 There is an important aspect to the truth test in an @code{if}
4174 expression. So far, we have spoken of `true' and `false' as values of
4175 predicates as if they were new kinds of Emacs Lisp objects. In fact,
4176 `false' is just our old friend @code{nil}. Anything else---anything
4179 The expression that tests for truth is interpreted as @dfn{true}
4180 if the result of evaluating it is a value that is not @code{nil}. In
4181 other words, the result of the test is considered true if the value
4182 returned is a number such as 47, a string such as @code{"hello"}, or a
4183 symbol (other than @code{nil}) such as @code{flowers}, or a list (so
4184 long as it is not empty), or even a buffer!
4187 * nil explained:: @code{nil} has two meanings.
4192 @unnumberedsubsec An explanation of @code{nil}
4195 Before illustrating a test for truth, we need an explanation of @code{nil}.
4197 In Emacs Lisp, the symbol @code{nil} has two meanings. First, it means the
4198 empty list. Second, it means false and is the value returned when a
4199 true-or-false-test tests false. @code{nil} can be written as an empty
4200 list, @code{()}, or as @code{nil}. As far as the Lisp interpreter is
4201 concerned, @code{()} and @code{nil} are the same. Humans, however, tend
4202 to use @code{nil} for false and @code{()} for the empty list.
4204 In Emacs Lisp, any value that is not @code{nil}---is not the empty
4205 list---is considered true. This means that if an evaluation returns
4206 something that is not an empty list, an @code{if} expression will test
4207 true. For example, if a number is put in the slot for the test, it
4208 will be evaluated and will return itself, since that is what numbers
4209 do when evaluated. In this conditional, the @code{if} expression will
4210 test true. The expression tests false only when @code{nil}, an empty
4211 list, is returned by evaluating the expression.
4213 You can see this by evaluating the two expressions in the following examples.
4215 In the first example, the number 4 is evaluated as the test in the
4216 @code{if} expression and returns itself; consequently, the then-part
4217 of the expression is evaluated and returned: @samp{true} appears in
4218 the echo area. In the second example, the @code{nil} indicates false;
4219 consequently, the else-part of the expression is evaluated and
4220 returned: @samp{false} appears in the echo area.
4237 Incidentally, if some other useful value is not available for a test that
4238 returns true, then the Lisp interpreter will return the symbol @code{t}
4239 for true. For example, the expression @code{(> 5 4)} returns @code{t}
4240 when evaluated, as you can see by evaluating it in the usual way:
4248 On the other hand, this function returns @code{nil} if the test is false.
4254 @node save-excursion
4255 @section @code{save-excursion}
4256 @findex save-excursion
4257 @cindex Region, what it is
4258 @cindex Preserving point, mark, and buffer
4259 @cindex Point, mark, buffer preservation
4263 The @code{save-excursion} function is the fourth and final special form
4264 that we will discuss in this chapter.
4266 In Emacs Lisp programs used for editing, the @code{save-excursion}
4267 function is very common. It saves the location of point and mark,
4268 executes the body of the function, and then restores point and mark to
4269 their previous positions if their locations were changed. Its primary
4270 purpose is to keep the user from being surprised and disturbed by
4271 unexpected movement of point or mark.
4274 * Point and mark:: A review of various locations.
4275 * Template for save-excursion::
4279 @node Point and mark
4280 @unnumberedsubsec Point and Mark
4283 Before discussing @code{save-excursion}, however, it may be useful
4284 first to review what point and mark are in GNU Emacs. @dfn{Point} is
4285 the current location of the cursor. Wherever the cursor
4286 is, that is point. More precisely, on terminals where the cursor
4287 appears to be on top of a character, point is immediately before the
4288 character. In Emacs Lisp, point is an integer. The first character in
4289 a buffer is number one, the second is number two, and so on. The
4290 function @code{point} returns the current position of the cursor as a
4291 number. Each buffer has its own value for point.
4293 The @dfn{mark} is another position in the buffer; its value can be set
4294 with a command such as @kbd{C-@key{SPC}} (@code{set-mark-command}). If
4295 a mark has been set, you can use the command @kbd{C-x C-x}
4296 (@code{exchange-point-and-mark}) to cause the cursor to jump to the mark
4297 and set the mark to be the previous position of point. In addition, if
4298 you set another mark, the position of the previous mark is saved in the
4299 mark ring. Many mark positions can be saved this way. You can jump the
4300 cursor to a saved mark by typing @kbd{C-u C-@key{SPC}} one or more
4303 The part of the buffer between point and mark is called @dfn{the
4304 region}. Numerous commands work on the region, including
4305 @code{center-region}, @code{count-lines-region}, @code{kill-region}, and
4306 @code{print-region}.
4308 The @code{save-excursion} special form saves the locations of point and
4309 mark and restores those positions after the code within the body of the
4310 special form is evaluated by the Lisp interpreter. Thus, if point were
4311 in the beginning of a piece of text and some code moved point to the end
4312 of the buffer, the @code{save-excursion} would put point back to where
4313 it was before, after the expressions in the body of the function were
4316 In Emacs, a function frequently moves point as part of its internal
4317 workings even though a user would not expect this. For example,
4318 @code{count-lines-region} moves point. To prevent the user from being
4319 bothered by jumps that are both unexpected and (from the user's point of
4320 view) unnecessary, @code{save-excursion} is often used to keep point and
4321 mark in the location expected by the user. The use of
4322 @code{save-excursion} is good housekeeping.
4324 To make sure the house stays clean, @code{save-excursion} restores the
4325 values of point and mark even if something goes wrong in the code inside
4326 of it (or, to be more precise and to use the proper jargon, ``in case of
4327 abnormal exit''). This feature is very helpful.
4329 In addition to recording the values of point and mark,
4330 @code{save-excursion} keeps track of the current buffer, and restores
4331 it, too. This means you can write code that will change the buffer and
4332 have @code{save-excursion} switch you back to the original buffer.
4333 This is how @code{save-excursion} is used in @code{append-to-buffer}.
4334 (@xref{append-to-buffer, , The Definition of @code{append-to-buffer}}.)
4336 @node Template for save-excursion
4337 @subsection Template for a @code{save-excursion} Expression
4340 The template for code using @code{save-excursion} is simple:
4350 The body of the function is one or more expressions that will be
4351 evaluated in sequence by the Lisp interpreter. If there is more than
4352 one expression in the body, the value of the last one will be returned
4353 as the value of the @code{save-excursion} function. The other
4354 expressions in the body are evaluated only for their side effects; and
4355 @code{save-excursion} itself is used only for its side effect (which
4356 is restoring the positions of point and mark).
4359 In more detail, the template for a @code{save-excursion} expression
4365 @var{first-expression-in-body}
4366 @var{second-expression-in-body}
4367 @var{third-expression-in-body}
4369 @var{last-expression-in-body})
4374 An expression, of course, may be a symbol on its own or a list.
4376 In Emacs Lisp code, a @code{save-excursion} expression often occurs
4377 within the body of a @code{let} expression. It looks like this:
4390 In the last few chapters we have introduced a fair number of functions
4391 and special forms. Here they are described in brief, along with a few
4392 similar functions that have not been mentioned yet.
4395 @item eval-last-sexp
4396 Evaluate the last symbolic expression before the current location of
4397 point. The value is printed in the echo area unless the function is
4398 invoked with an argument; in that case, the output is printed in the
4399 current buffer. This command is normally bound to @kbd{C-x C-e}.
4402 Define function. This special form has up to five parts: the name,
4403 a template for the arguments that will be passed to the function,
4404 documentation, an optional interactive declaration, and the body of the
4408 For example, in an early version of Emacs, the function definition was
4409 as follows. (It is slightly more complex now that it seeks the first
4410 non-whitespace character rather than the first visible character.)
4414 (defun back-to-indentation ()
4415 "Move point to first visible character on line."
4417 (beginning-of-line 1)
4418 (skip-chars-forward " \t"))
4425 (defun backward-to-indentation (&optional arg)
4426 "Move backward ARG lines and position at first nonblank character."
4428 (forward-line (- (or arg 1)))
4429 (skip-chars-forward " \t"))
4431 (defun back-to-indentation ()
4432 "Move point to the first non-whitespace character on this line."
4434 (beginning-of-line 1)
4435 (skip-syntax-forward " " (line-end-position))
4436 ;; Move back over chars that have whitespace syntax but have the p flag.
4437 (backward-prefix-chars))
4441 Declare to the interpreter that the function can be used
4442 interactively. This special form may be followed by a string with one
4443 or more parts that pass the information to the arguments of the
4444 function, in sequence. These parts may also tell the interpreter to
4445 prompt for information. Parts of the string are separated by
4446 newlines, @samp{\n}.
4449 Common code characters are:
4453 The name of an existing buffer.
4456 The name of an existing file.
4459 The numeric prefix argument. (Note that this `p' is lower case.)
4462 Point and the mark, as two numeric arguments, smallest first. This
4463 is the only code letter that specifies two successive arguments
4467 @xref{Interactive Codes, , Code Characters for @samp{interactive},
4468 elisp, The GNU Emacs Lisp Reference Manual}, for a complete list of
4472 Declare that a list of variables is for use within the body of the
4473 @code{let} and give them an initial value, either @code{nil} or a
4474 specified value; then evaluate the rest of the expressions in the body
4475 of the @code{let} and return the value of the last one. Inside the
4476 body of the @code{let}, the Lisp interpreter does not see the values of
4477 the variables of the same names that are bound outside of the
4485 (let ((foo (buffer-name))
4486 (bar (buffer-size)))
4488 "This buffer is %s and has %d characters."
4493 @item save-excursion
4494 Record the values of point and mark and the current buffer before
4495 evaluating the body of this special form. Restore the values of point
4496 and mark and buffer afterward.
4503 (message "We are %d characters into this buffer."
4506 (goto-char (point-min)) (point))))
4511 Evaluate the first argument to the function; if it is true, evaluate
4512 the second argument; else evaluate the third argument, if there is one.
4514 The @code{if} special form is called a @dfn{conditional}. There are
4515 other conditionals in Emacs Lisp, but @code{if} is perhaps the most
4523 (if (= 22 emacs-major-version)
4524 (message "This is version 22 Emacs")
4525 (message "This is not version 22 Emacs"))
4534 The @code{<} function tests whether its first argument is smaller than
4535 its second argument. A corresponding function, @code{>}, tests whether
4536 the first argument is greater than the second. Likewise, @code{<=}
4537 tests whether the first argument is less than or equal to the second and
4538 @code{>=} tests whether the first argument is greater than or equal to
4539 the second. In all cases, both arguments must be numbers or markers
4540 (markers indicate positions in buffers).
4544 The @code{=} function tests whether two arguments, both numbers or
4550 Test whether two objects are the same. @code{equal} uses one meaning
4551 of the word `same' and @code{eq} uses another: @code{equal} returns
4552 true if the two objects have a similar structure and contents, such as
4553 two copies of the same book. On the other hand, @code{eq}, returns
4554 true if both arguments are actually the same object.
4563 The @code{string-lessp} function tests whether its first argument is
4564 smaller than the second argument. A shorter, alternative name for the
4565 same function (a @code{defalias}) is @code{string<}.
4567 The arguments to @code{string-lessp} must be strings or symbols; the
4568 ordering is lexicographic, so case is significant. The print names of
4569 symbols are used instead of the symbols themselves.
4571 @cindex @samp{empty string} defined
4572 An empty string, @samp{""}, a string with no characters in it, is
4573 smaller than any string of characters.
4575 @code{string-equal} provides the corresponding test for equality. Its
4576 shorter, alternative name is @code{string=}. There are no string test
4577 functions that correspond to @var{>}, @code{>=}, or @code{<=}.
4580 Print a message in the echo area. The first argument is a string that
4581 can contain @samp{%s}, @samp{%d}, or @samp{%c} to print the value of
4582 arguments that follow the string. The argument used by @samp{%s} must
4583 be a string or a symbol; the argument used by @samp{%d} must be a
4584 number. The argument used by @samp{%c} must be an @sc{ascii} code
4585 number; it will be printed as the character with that @sc{ascii} code.
4586 (Various other %-sequences have not been mentioned.)
4590 The @code{setq} function sets the value of its first argument to the
4591 value of the second argument. The first argument is automatically
4592 quoted by @code{setq}. It does the same for succeeding pairs of
4593 arguments. Another function, @code{set}, takes only two arguments and
4594 evaluates both of them before setting the value returned by its first
4595 argument to the value returned by its second argument.
4598 Without an argument, return the name of the buffer, as a string.
4600 @item buffer-file-name
4601 Without an argument, return the name of the file the buffer is
4604 @item current-buffer
4605 Return the buffer in which Emacs is active; it may not be
4606 the buffer that is visible on the screen.
4609 Return the most recently selected buffer (other than the buffer passed
4610 to @code{other-buffer} as an argument and other than the current
4613 @item switch-to-buffer
4614 Select a buffer for Emacs to be active in and display it in the current
4615 window so users can look at it. Usually bound to @kbd{C-x b}.
4618 Switch Emacs's attention to a buffer on which programs will run. Don't
4619 alter what the window is showing.
4622 Return the number of characters in the current buffer.
4625 Return the value of the current position of the cursor, as an
4626 integer counting the number of characters from the beginning of the
4630 Return the minimum permissible value of point in
4631 the current buffer. This is 1, unless narrowing is in effect.
4634 Return the value of the maximum permissible value of point in the
4635 current buffer. This is the end of the buffer, unless narrowing is in
4640 @node defun Exercises
4645 Write a non-interactive function that doubles the value of its
4646 argument, a number. Make that function interactive.
4649 Write a function that tests whether the current value of
4650 @code{fill-column} is greater than the argument passed to the function,
4651 and if so, prints an appropriate message.
4654 @node Buffer Walk Through
4655 @chapter A Few Buffer--Related Functions
4657 In this chapter we study in detail several of the functions used in GNU
4658 Emacs. This is called a ``walk-through''. These functions are used as
4659 examples of Lisp code, but are not imaginary examples; with the
4660 exception of the first, simplified function definition, these functions
4661 show the actual code used in GNU Emacs. You can learn a great deal from
4662 these definitions. The functions described here are all related to
4663 buffers. Later, we will study other functions.
4666 * Finding More:: How to find more information.
4667 * simplified-beginning-of-buffer:: Shows @code{goto-char},
4668 @code{point-min}, and @code{push-mark}.
4669 * mark-whole-buffer:: Almost the same as @code{beginning-of-buffer}.
4670 * append-to-buffer:: Uses @code{save-excursion} and
4671 @code{insert-buffer-substring}.
4672 * Buffer Related Review:: Review.
4673 * Buffer Exercises::
4677 @section Finding More Information
4679 @findex describe-function, @r{introduced}
4680 @cindex Find function documentation
4681 In this walk-through, I will describe each new function as we come to
4682 it, sometimes in detail and sometimes briefly. If you are interested,
4683 you can get the full documentation of any Emacs Lisp function at any
4684 time by typing @kbd{C-h f} and then the name of the function (and then
4685 @key{RET}). Similarly, you can get the full documentation for a
4686 variable by typing @kbd{C-h v} and then the name of the variable (and
4689 @cindex Find source of function
4690 @c In version 22, tells location both of C and of Emacs Lisp
4691 Also, @code{describe-function} will tell you the location of the
4692 function definition.
4694 Put point into the name of the file that contains the function and
4695 press the @key{RET} key. In this case, @key{RET} means
4696 @code{push-button} rather than `return' or `enter'. Emacs will take
4697 you directly to the function definition.
4702 If you move point over the file name and press
4703 the @key{RET} key, which in this case means @code{help-follow} rather
4704 than `return' or `enter', Emacs will take you directly to the function
4708 More generally, if you want to see a function in its original source
4709 file, you can use the @code{find-tag} function to jump to it.
4710 @code{find-tag} works with a wide variety of languages, not just
4711 Lisp, and C, and it works with non-programming text as well. For
4712 example, @code{find-tag} will jump to the various nodes in the
4713 Texinfo source file of this document.
4714 The @code{find-tag} function depends on `tags tables' that record
4715 the locations of the functions, variables, and other items to which
4716 @code{find-tag} jumps.
4718 To use the @code{find-tag} command, type @kbd{M-.} (i.e., press the
4719 period key while holding down the @key{META} key, or else type the
4720 @key{ESC} key and then type the period key), and then, at the prompt,
4721 type in the name of the function whose source code you want to see,
4722 such as @code{mark-whole-buffer}, and then type @key{RET}. Emacs will
4723 switch buffers and display the source code for the function on your
4724 screen. To switch back to your current buffer, type @kbd{C-x b
4725 @key{RET}}. (On some keyboards, the @key{META} key is labeled
4728 @c !!! 22.1.1 tags table location in this paragraph
4729 @cindex TAGS table, specifying
4731 Depending on how the initial default values of your copy of Emacs are
4732 set, you may also need to specify the location of your `tags table',
4733 which is a file called @file{TAGS}. For example, if you are
4734 interested in Emacs sources, the tags table you will most likely want,
4735 if it has already been created for you, will be in a subdirectory of
4736 the @file{/usr/local/share/emacs/} directory; thus you would use the
4737 @code{M-x visit-tags-table} command and specify a pathname such as
4738 @file{/usr/local/share/emacs/22.1.1/lisp/TAGS}. If the tags table
4739 has not already been created, you will have to create it yourself. It
4740 will be in a file such as @file{/usr/local/src/emacs/src/TAGS}.
4743 To create a @file{TAGS} file in a specific directory, switch to that
4744 directory in Emacs using @kbd{M-x cd} command, or list the directory
4745 with @kbd{C-x d} (@code{dired}). Then run the compile command, with
4746 @w{@code{etags *.el}} as the command to execute:
4749 M-x compile RET etags *.el RET
4752 For more information, see @ref{etags, , Create Your Own @file{TAGS} File}.
4754 After you become more familiar with Emacs Lisp, you will find that you will
4755 frequently use @code{find-tag} to navigate your way around source code;
4756 and you will create your own @file{TAGS} tables.
4758 @cindex Library, as term for `file'
4759 Incidentally, the files that contain Lisp code are conventionally
4760 called @dfn{libraries}. The metaphor is derived from that of a
4761 specialized library, such as a law library or an engineering library,
4762 rather than a general library. Each library, or file, contains
4763 functions that relate to a particular topic or activity, such as
4764 @file{abbrev.el} for handling abbreviations and other typing
4765 shortcuts, and @file{help.el} for on-line help. (Sometimes several
4766 libraries provide code for a single activity, as the various
4767 @file{rmail@dots{}} files provide code for reading electronic mail.)
4768 In @cite{The GNU Emacs Manual}, you will see sentences such as ``The
4769 @kbd{C-h p} command lets you search the standard Emacs Lisp libraries
4770 by topic keywords.''
4772 @node simplified-beginning-of-buffer
4773 @section A Simplified @code{beginning-of-buffer} Definition
4774 @findex simplified-beginning-of-buffer
4776 The @code{beginning-of-buffer} command is a good function to start with
4777 since you are likely to be familiar with it and it is easy to
4778 understand. Used as an interactive command, @code{beginning-of-buffer}
4779 moves the cursor to the beginning of the buffer, leaving the mark at the
4780 previous position. It is generally bound to @kbd{M-<}.
4782 In this section, we will discuss a shortened version of the function
4783 that shows how it is most frequently used. This shortened function
4784 works as written, but it does not contain the code for a complex option.
4785 In another section, we will describe the entire function.
4786 (@xref{beginning-of-buffer, , Complete Definition of
4787 @code{beginning-of-buffer}}.)
4789 Before looking at the code, let's consider what the function
4790 definition has to contain: it must include an expression that makes
4791 the function interactive so it can be called by typing @kbd{M-x
4792 beginning-of-buffer} or by typing a keychord such as @kbd{M-<}; it
4793 must include code to leave a mark at the original position in the
4794 buffer; and it must include code to move the cursor to the beginning
4798 Here is the complete text of the shortened version of the function:
4802 (defun simplified-beginning-of-buffer ()
4803 "Move point to the beginning of the buffer;
4804 leave mark at previous position."
4807 (goto-char (point-min)))
4811 Like all function definitions, this definition has five parts following
4812 the special form @code{defun}:
4816 The name: in this example, @code{simplified-beginning-of-buffer}.
4819 A list of the arguments: in this example, an empty list, @code{()},
4822 The documentation string.
4825 The interactive expression.
4832 In this function definition, the argument list is empty; this means that
4833 this function does not require any arguments. (When we look at the
4834 definition for the complete function, we will see that it may be passed
4835 an optional argument.)
4837 The interactive expression tells Emacs that the function is intended to
4838 be used interactively. In this example, @code{interactive} does not have
4839 an argument because @code{simplified-beginning-of-buffer} does not
4843 The body of the function consists of the two lines:
4848 (goto-char (point-min))
4852 The first of these lines is the expression, @code{(push-mark)}. When
4853 this expression is evaluated by the Lisp interpreter, it sets a mark at
4854 the current position of the cursor, wherever that may be. The position
4855 of this mark is saved in the mark ring.
4857 The next line is @code{(goto-char (point-min))}. This expression
4858 jumps the cursor to the minimum point in the buffer, that is, to the
4859 beginning of the buffer (or to the beginning of the accessible portion
4860 of the buffer if it is narrowed. @xref{Narrowing & Widening, ,
4861 Narrowing and Widening}.)
4863 The @code{push-mark} command sets a mark at the place where the cursor
4864 was located before it was moved to the beginning of the buffer by the
4865 @code{(goto-char (point-min))} expression. Consequently, you can, if
4866 you wish, go back to where you were originally by typing @kbd{C-x C-x}.
4868 That is all there is to the function definition!
4870 @findex describe-function
4871 When you are reading code such as this and come upon an unfamiliar
4872 function, such as @code{goto-char}, you can find out what it does by
4873 using the @code{describe-function} command. To use this command, type
4874 @kbd{C-h f} and then type in the name of the function and press
4875 @key{RET}. The @code{describe-function} command will print the
4876 function's documentation string in a @file{*Help*} window. For
4877 example, the documentation for @code{goto-char} is:
4881 Set point to POSITION, a number or marker.
4882 Beginning of buffer is position (point-min), end is (point-max).
4887 The function's one argument is the desired position.
4890 (The prompt for @code{describe-function} will offer you the symbol
4891 under or preceding the cursor, so you can save typing by positioning
4892 the cursor right over or after the function and then typing @kbd{C-h f
4895 The @code{end-of-buffer} function definition is written in the same way as
4896 the @code{beginning-of-buffer} definition except that the body of the
4897 function contains the expression @code{(goto-char (point-max))} in place
4898 of @code{(goto-char (point-min))}.
4900 @node mark-whole-buffer
4901 @section The Definition of @code{mark-whole-buffer}
4902 @findex mark-whole-buffer
4904 The @code{mark-whole-buffer} function is no harder to understand than the
4905 @code{simplified-beginning-of-buffer} function. In this case, however,
4906 we will look at the complete function, not a shortened version.
4908 The @code{mark-whole-buffer} function is not as commonly used as the
4909 @code{beginning-of-buffer} function, but is useful nonetheless: it
4910 marks a whole buffer as a region by putting point at the beginning and
4911 a mark at the end of the buffer. It is generally bound to @kbd{C-x
4915 * mark-whole-buffer overview::
4916 * Body of mark-whole-buffer:: Only three lines of code.
4920 @node mark-whole-buffer overview
4921 @unnumberedsubsec An overview of @code{mark-whole-buffer}
4925 In GNU Emacs 22, the code for the complete function looks like this:
4929 (defun mark-whole-buffer ()
4930 "Put point at beginning and mark at end of buffer.
4931 You probably should not use this function in Lisp programs;
4932 it is usually a mistake for a Lisp function to use any subroutine
4933 that uses or sets the mark."
4936 (push-mark (point-max) nil t)
4937 (goto-char (point-min)))
4942 Like all other functions, the @code{mark-whole-buffer} function fits
4943 into the template for a function definition. The template looks like
4948 (defun @var{name-of-function} (@var{argument-list})
4949 "@var{documentation}@dots{}"
4950 (@var{interactive-expression}@dots{})
4955 Here is how the function works: the name of the function is
4956 @code{mark-whole-buffer}; it is followed by an empty argument list,
4957 @samp{()}, which means that the function does not require arguments.
4958 The documentation comes next.
4960 The next line is an @code{(interactive)} expression that tells Emacs
4961 that the function will be used interactively. These details are similar
4962 to the @code{simplified-beginning-of-buffer} function described in the
4966 @node Body of mark-whole-buffer
4967 @subsection Body of @code{mark-whole-buffer}
4969 The body of the @code{mark-whole-buffer} function consists of three
4976 (push-mark (point-max) nil t)
4977 (goto-char (point-min))
4981 The first of these lines is the expression, @code{(push-mark (point))}.
4983 This line does exactly the same job as the first line of the body of
4984 the @code{simplified-beginning-of-buffer} function, which is written
4985 @code{(push-mark)}. In both cases, the Lisp interpreter sets a mark
4986 at the current position of the cursor.
4988 I don't know why the expression in @code{mark-whole-buffer} is written
4989 @code{(push-mark (point))} and the expression in
4990 @code{beginning-of-buffer} is written @code{(push-mark)}. Perhaps
4991 whoever wrote the code did not know that the arguments for
4992 @code{push-mark} are optional and that if @code{push-mark} is not
4993 passed an argument, the function automatically sets mark at the
4994 location of point by default. Or perhaps the expression was written
4995 so as to parallel the structure of the next line. In any case, the
4996 line causes Emacs to determine the position of point and set a mark
4999 In earlier versions of GNU Emacs, the next line of
5000 @code{mark-whole-buffer} was @code{(push-mark (point-max))}. This
5001 expression sets a mark at the point in the buffer that has the highest
5002 number. This will be the end of the buffer (or, if the buffer is
5003 narrowed, the end of the accessible portion of the buffer.
5004 @xref{Narrowing & Widening, , Narrowing and Widening}, for more about
5005 narrowing.) After this mark has been set, the previous mark, the one
5006 set at point, is no longer set, but Emacs remembers its position, just
5007 as all other recent marks are always remembered. This means that you
5008 can, if you wish, go back to that position by typing @kbd{C-u
5012 In GNU Emacs 22, the @code{(point-max)} is slightly more complicated.
5016 (push-mark (point-max) nil t)
5020 The expression works nearly the same as before. It sets a mark at the
5021 highest numbered place in the buffer that it can. However, in this
5022 version, @code{push-mark} has two additional arguments. The second
5023 argument to @code{push-mark} is @code{nil}. This tells the function
5024 it @emph{should} display a message that says `Mark set' when it pushes
5025 the mark. The third argument is @code{t}. This tells
5026 @code{push-mark} to activate the mark when Transient Mark mode is
5027 turned on. Transient Mark mode highlights the currently active
5028 region. It is often turned off.
5030 Finally, the last line of the function is @code{(goto-char
5031 (point-min)))}. This is written exactly the same way as it is written
5032 in @code{beginning-of-buffer}. The expression moves the cursor to
5033 the minimum point in the buffer, that is, to the beginning of the buffer
5034 (or to the beginning of the accessible portion of the buffer). As a
5035 result of this, point is placed at the beginning of the buffer and mark
5036 is set at the end of the buffer. The whole buffer is, therefore, the
5039 @node append-to-buffer
5040 @section The Definition of @code{append-to-buffer}
5041 @findex append-to-buffer
5043 The @code{append-to-buffer} command is more complex than the
5044 @code{mark-whole-buffer} command. What it does is copy the region
5045 (that is, the part of the buffer between point and mark) from the
5046 current buffer to a specified buffer.
5049 * append-to-buffer overview::
5050 * append interactive:: A two part interactive expression.
5051 * append-to-buffer body:: Incorporates a @code{let} expression.
5052 * append save-excursion:: How the @code{save-excursion} works.
5056 @node append-to-buffer overview
5057 @unnumberedsubsec An Overview of @code{append-to-buffer}
5060 @findex insert-buffer-substring
5061 The @code{append-to-buffer} command uses the
5062 @code{insert-buffer-substring} function to copy the region.
5063 @code{insert-buffer-substring} is described by its name: it takes a
5064 string of characters from part of a buffer, a ``substring'', and
5065 inserts them into another buffer.
5067 Most of @code{append-to-buffer} is
5068 concerned with setting up the conditions for
5069 @code{insert-buffer-substring} to work: the code must specify both the
5070 buffer to which the text will go, the window it comes from and goes
5071 to, and the region that will be copied.
5074 Here is the complete text of the function:
5078 (defun append-to-buffer (buffer start end)
5079 "Append to specified buffer the text of the region.
5080 It is inserted into that buffer before its point.
5084 When calling from a program, give three arguments:
5085 BUFFER (or buffer name), START and END.
5086 START and END specify the portion of the current buffer to be copied."
5088 (list (read-buffer "Append to buffer: " (other-buffer
5089 (current-buffer) t))
5090 (region-beginning) (region-end)))
5093 (let ((oldbuf (current-buffer)))
5095 (let* ((append-to (get-buffer-create buffer))
5096 (windows (get-buffer-window-list append-to t t))
5098 (set-buffer append-to)
5099 (setq point (point))
5100 (barf-if-buffer-read-only)
5101 (insert-buffer-substring oldbuf start end)
5102 (dolist (window windows)
5103 (when (= (window-point window) point)
5104 (set-window-point window (point))))))))
5108 The function can be understood by looking at it as a series of
5109 filled-in templates.
5111 The outermost template is for the function definition. In this
5112 function, it looks like this (with several slots filled in):
5116 (defun append-to-buffer (buffer start end)
5117 "@var{documentation}@dots{}"
5118 (interactive @dots{})
5123 The first line of the function includes its name and three arguments.
5124 The arguments are the @code{buffer} to which the text will be copied, and
5125 the @code{start} and @code{end} of the region in the current buffer that
5128 The next part of the function is the documentation, which is clear and
5129 complete. As is conventional, the three arguments are written in
5130 upper case so you will notice them easily. Even better, they are
5131 described in the same order as in the argument list.
5133 Note that the documentation distinguishes between a buffer and its
5134 name. (The function can handle either.)
5136 @node append interactive
5137 @subsection The @code{append-to-buffer} Interactive Expression
5139 Since the @code{append-to-buffer} function will be used interactively,
5140 the function must have an @code{interactive} expression. (For a
5141 review of @code{interactive}, see @ref{Interactive, , Making a
5142 Function Interactive}.) The expression reads as follows:
5148 "Append to buffer: "
5149 (other-buffer (current-buffer) t))
5156 This expression is not one with letters standing for parts, as
5157 described earlier. Instead, it starts a list with these parts:
5159 The first part of the list is an expression to read the name of a
5160 buffer and return it as a string. That is @code{read-buffer}. The
5161 function requires a prompt as its first argument, @samp{"Append to
5162 buffer: "}. Its second argument tells the command what value to
5163 provide if you don't specify anything.
5165 In this case that second argument is an expression containing the
5166 function @code{other-buffer}, an exception, and a @samp{t}, standing
5169 The first argument to @code{other-buffer}, the exception, is yet
5170 another function, @code{current-buffer}. That is not going to be
5171 returned. The second argument is the symbol for true, @code{t}. that
5172 tells @code{other-buffer} that it may show visible buffers (except in
5173 this case, it will not show the current buffer, which makes sense).
5176 The expression looks like this:
5179 (other-buffer (current-buffer) t)
5182 The second and third arguments to the @code{list} expression are
5183 @code{(region-beginning)} and @code{(region-end)}. These two
5184 functions specify the beginning and end of the text to be appended.
5187 Originally, the command used the letters @samp{B} and @samp{r}.
5188 The whole @code{interactive} expression looked like this:
5191 (interactive "BAppend to buffer:@: \nr")
5195 But when that was done, the default value of the buffer switched to
5196 was invisible. That was not wanted.
5198 (The prompt was separated from the second argument with a newline,
5199 @samp{\n}. It was followed by an @samp{r} that told Emacs to bind the
5200 two arguments that follow the symbol @code{buffer} in the function's
5201 argument list (that is, @code{start} and @code{end}) to the values of
5202 point and mark. That argument worked fine.)
5204 @node append-to-buffer body
5205 @subsection The Body of @code{append-to-buffer}
5208 in GNU Emacs 22 in /usr/local/src/emacs/lisp/simple.el
5210 (defun append-to-buffer (buffer start end)
5211 "Append to specified buffer the text of the region.
5212 It is inserted into that buffer before its point.
5214 When calling from a program, give three arguments:
5215 BUFFER (or buffer name), START and END.
5216 START and END specify the portion of the current buffer to be copied."
5218 (list (read-buffer "Append to buffer: " (other-buffer (current-buffer) t))
5219 (region-beginning) (region-end)))
5220 (let ((oldbuf (current-buffer)))
5222 (let* ((append-to (get-buffer-create buffer))
5223 (windows (get-buffer-window-list append-to t t))
5225 (set-buffer append-to)
5226 (setq point (point))
5227 (barf-if-buffer-read-only)
5228 (insert-buffer-substring oldbuf start end)
5229 (dolist (window windows)
5230 (when (= (window-point window) point)
5231 (set-window-point window (point))))))))
5234 The body of the @code{append-to-buffer} function begins with @code{let}.
5236 As we have seen before (@pxref{let, , @code{let}}), the purpose of a
5237 @code{let} expression is to create and give initial values to one or
5238 more variables that will only be used within the body of the
5239 @code{let}. This means that such a variable will not be confused with
5240 any variable of the same name outside the @code{let} expression.
5242 We can see how the @code{let} expression fits into the function as a
5243 whole by showing a template for @code{append-to-buffer} with the
5244 @code{let} expression in outline:
5248 (defun append-to-buffer (buffer start end)
5249 "@var{documentation}@dots{}"
5250 (interactive @dots{})
5251 (let ((@var{variable} @var{value}))
5256 The @code{let} expression has three elements:
5260 The symbol @code{let};
5263 A varlist containing, in this case, a single two-element list,
5264 @code{(@var{variable} @var{value})};
5267 The body of the @code{let} expression.
5271 In the @code{append-to-buffer} function, the varlist looks like this:
5274 (oldbuf (current-buffer))
5278 In this part of the @code{let} expression, the one variable,
5279 @code{oldbuf}, is bound to the value returned by the
5280 @code{(current-buffer)} expression. The variable, @code{oldbuf}, is
5281 used to keep track of the buffer in which you are working and from
5282 which you will copy.
5284 The element or elements of a varlist are surrounded by a set of
5285 parentheses so the Lisp interpreter can distinguish the varlist from
5286 the body of the @code{let}. As a consequence, the two-element list
5287 within the varlist is surrounded by a circumscribing set of parentheses.
5288 The line looks like this:
5292 (let ((oldbuf (current-buffer)))
5298 The two parentheses before @code{oldbuf} might surprise you if you did
5299 not realize that the first parenthesis before @code{oldbuf} marks the
5300 boundary of the varlist and the second parenthesis marks the beginning
5301 of the two-element list, @code{(oldbuf (current-buffer))}.
5303 @node append save-excursion
5304 @subsection @code{save-excursion} in @code{append-to-buffer}
5306 The body of the @code{let} expression in @code{append-to-buffer}
5307 consists of a @code{save-excursion} expression.
5309 The @code{save-excursion} function saves the locations of point and
5310 mark, and restores them to those positions after the expressions in the
5311 body of the @code{save-excursion} complete execution. In addition,
5312 @code{save-excursion} keeps track of the original buffer, and
5313 restores it. This is how @code{save-excursion} is used in
5314 @code{append-to-buffer}.
5317 @cindex Indentation for formatting
5318 @cindex Formatting convention
5319 Incidentally, it is worth noting here that a Lisp function is normally
5320 formatted so that everything that is enclosed in a multi-line spread is
5321 indented more to the right than the first symbol. In this function
5322 definition, the @code{let} is indented more than the @code{defun}, and
5323 the @code{save-excursion} is indented more than the @code{let}, like
5339 This formatting convention makes it easy to see that the lines in
5340 the body of the @code{save-excursion} are enclosed by the parentheses
5341 associated with @code{save-excursion}, just as the
5342 @code{save-excursion} itself is enclosed by the parentheses associated
5343 with the @code{let}:
5347 (let ((oldbuf (current-buffer)))
5350 (set-buffer @dots{})
5351 (insert-buffer-substring oldbuf start end)
5357 The use of the @code{save-excursion} function can be viewed as a process
5358 of filling in the slots of a template:
5363 @var{first-expression-in-body}
5364 @var{second-expression-in-body}
5366 @var{last-expression-in-body})
5372 In this function, the body of the @code{save-excursion} contains only
5373 one expression, the @code{let*} expression. You know about a
5374 @code{let} function. The @code{let*} function is different. It has a
5375 @samp{*} in its name. It enables Emacs to set each variable in its
5376 varlist in sequence, one after another.
5378 Its critical feature is that variables later in the varlist can make
5379 use of the values to which Emacs set variables earlier in the varlist.
5380 @xref{fwd-para let, , The @code{let*} expression}.
5382 We will skip functions like @code{let*} and focus on two: the
5383 @code{set-buffer} function and the @code{insert-buffer-substring}
5387 In the old days, the @code{set-buffer} expression was simply
5390 (set-buffer (get-buffer-create buffer))
5398 (set-buffer append-to)
5402 @code{append-to} is bound to @code{(get-buffer-create buffer)} earlier
5403 on in the @code{let*} expression. That extra binding would not be
5404 necessary except for that @code{append-to} is used later in the
5405 varlist as an argument to @code{get-buffer-window-list}.
5410 (let ((oldbuf (current-buffer)))
5412 (let* ((append-to (get-buffer-create buffer))
5413 (windows (get-buffer-window-list append-to t t))
5415 (set-buffer append-to)
5416 (setq point (point))
5417 (barf-if-buffer-read-only)
5418 (insert-buffer-substring oldbuf start end)
5419 (dolist (window windows)
5420 (when (= (window-point window) point)
5421 (set-window-point window (point))))))))
5424 The @code{append-to-buffer} function definition inserts text from the
5425 buffer in which you are currently to a named buffer. It happens that
5426 @code{insert-buffer-substring} copies text from another buffer to the
5427 current buffer, just the reverse---that is why the
5428 @code{append-to-buffer} definition starts out with a @code{let} that
5429 binds the local symbol @code{oldbuf} to the value returned by
5430 @code{current-buffer}.
5433 The @code{insert-buffer-substring} expression looks like this:
5436 (insert-buffer-substring oldbuf start end)
5440 The @code{insert-buffer-substring} function copies a string
5441 @emph{from} the buffer specified as its first argument and inserts the
5442 string into the present buffer. In this case, the argument to
5443 @code{insert-buffer-substring} is the value of the variable created
5444 and bound by the @code{let}, namely the value of @code{oldbuf}, which
5445 was the current buffer when you gave the @code{append-to-buffer}
5448 After @code{insert-buffer-substring} has done its work,
5449 @code{save-excursion} will restore the action to the original buffer
5450 and @code{append-to-buffer} will have done its job.
5453 Written in skeletal form, the workings of the body look like this:
5457 (let (@var{bind-}@code{oldbuf}@var{-to-value-of-}@code{current-buffer})
5458 (save-excursion ; @r{Keep track of buffer.}
5460 @var{insert-substring-from-}@code{oldbuf}@var{-into-buffer})
5462 @var{change-back-to-original-buffer-when-finished}
5463 @var{let-the-local-meaning-of-}@code{oldbuf}@var{-disappear-when-finished}
5467 In summary, @code{append-to-buffer} works as follows: it saves the
5468 value of the current buffer in the variable called @code{oldbuf}. It
5469 gets the new buffer (creating one if need be) and switches Emacs's
5470 attention to it. Using the value of @code{oldbuf}, it inserts the
5471 region of text from the old buffer into the new buffer; and then using
5472 @code{save-excursion}, it brings you back to your original buffer.
5474 In looking at @code{append-to-buffer}, you have explored a fairly
5475 complex function. It shows how to use @code{let} and
5476 @code{save-excursion}, and how to change to and come back from another
5477 buffer. Many function definitions use @code{let},
5478 @code{save-excursion}, and @code{set-buffer} this way.
5480 @node Buffer Related Review
5483 Here is a brief summary of the various functions discussed in this chapter.
5486 @item describe-function
5487 @itemx describe-variable
5488 Print the documentation for a function or variable.
5489 Conventionally bound to @kbd{C-h f} and @kbd{C-h v}.
5492 Find the file containing the source for a function or variable and
5493 switch buffers to it, positioning point at the beginning of the item.
5494 Conventionally bound to @kbd{M-.} (that's a period following the
5497 @item save-excursion
5498 Save the location of point and mark and restore their values after the
5499 arguments to @code{save-excursion} have been evaluated. Also, remember
5500 the current buffer and return to it.
5503 Set mark at a location and record the value of the previous mark on the
5504 mark ring. The mark is a location in the buffer that will keep its
5505 relative position even if text is added to or removed from the buffer.
5508 Set point to the location specified by the value of the argument, which
5509 can be a number, a marker, or an expression that returns the number of
5510 a position, such as @code{(point-min)}.
5512 @item insert-buffer-substring
5513 Copy a region of text from a buffer that is passed to the function as
5514 an argument and insert the region into the current buffer.
5516 @item mark-whole-buffer
5517 Mark the whole buffer as a region. Normally bound to @kbd{C-x h}.
5520 Switch the attention of Emacs to another buffer, but do not change the
5521 window being displayed. Used when the program rather than a human is
5522 to work on a different buffer.
5524 @item get-buffer-create
5526 Find a named buffer or create one if a buffer of that name does not
5527 exist. The @code{get-buffer} function returns @code{nil} if the named
5528 buffer does not exist.
5532 @node Buffer Exercises
5537 Write your own @code{simplified-end-of-buffer} function definition;
5538 then test it to see whether it works.
5541 Use @code{if} and @code{get-buffer} to write a function that prints a
5542 message telling you whether a buffer exists.
5545 Using @code{find-tag}, find the source for the @code{copy-to-buffer}
5550 @chapter A Few More Complex Functions
5552 In this chapter, we build on what we have learned in previous chapters
5553 by looking at more complex functions. The @code{copy-to-buffer}
5554 function illustrates use of two @code{save-excursion} expressions in
5555 one definition, while the @code{insert-buffer} function illustrates
5556 use of an asterisk in an @code{interactive} expression, use of
5557 @code{or}, and the important distinction between a name and the object
5558 to which the name refers.
5561 * copy-to-buffer:: With @code{set-buffer}, @code{get-buffer-create}.
5562 * insert-buffer:: Read-only, and with @code{or}.
5563 * beginning-of-buffer:: Shows @code{goto-char},
5564 @code{point-min}, and @code{push-mark}.
5565 * Second Buffer Related Review::
5566 * optional Exercise::
5569 @node copy-to-buffer
5570 @section The Definition of @code{copy-to-buffer}
5571 @findex copy-to-buffer
5573 After understanding how @code{append-to-buffer} works, it is easy to
5574 understand @code{copy-to-buffer}. This function copies text into a
5575 buffer, but instead of adding to the second buffer, it replaces all the
5576 previous text in the second buffer.
5579 The body of @code{copy-to-buffer} looks like this,
5584 (interactive "BCopy to buffer: \nr")
5585 (let ((oldbuf (current-buffer)))
5586 (with-current-buffer (get-buffer-create buffer)
5587 (barf-if-buffer-read-only)
5590 (insert-buffer-substring oldbuf start end)))))
5594 The @code{copy-to-buffer} function has a simpler @code{interactive}
5595 expression than @code{append-to-buffer}.
5598 The definition then says
5601 (with-current-buffer (get-buffer-create buffer) @dots{}
5604 First, look at the earliest inner expression; that is evaluated first.
5605 That expression starts with @code{get-buffer-create buffer}. The
5606 function tells the computer to use the buffer with the name specified
5607 as the one to which you are copying, or if such a buffer does not
5608 exist, to create it. Then, the @code{with-current-buffer} function
5609 evaluates its body with that buffer temporarily current.
5611 (This demonstrates another way to shift the computer's attention but
5612 not the user's. The @code{append-to-buffer} function showed how to do
5613 the same with @code{save-excursion} and @code{set-buffer}.
5614 @code{with-current-buffer} is a newer, and arguably easier,
5617 The @code{barf-if-buffer-read-only} function sends you an error
5618 message saying the buffer is read-only if you cannot modify it.
5620 The next line has the @code{erase-buffer} function as its sole
5621 contents. That function erases the buffer.
5623 Finally, the last two lines contain the @code{save-excursion}
5624 expression with @code{insert-buffer-substring} as its body.
5625 The @code{insert-buffer-substring} expression copies the text from
5626 the buffer you are in (and you have not seen the computer shift its
5627 attention, so you don't know that that buffer is now called
5630 Incidentally, this is what is meant by `replacement'. To replace text,
5631 Emacs erases the previous text and then inserts new text.
5634 In outline, the body of @code{copy-to-buffer} looks like this:
5638 (let (@var{bind-}@code{oldbuf}@var{-to-value-of-}@code{current-buffer})
5639 (@var{with-the-buffer-you-are-copying-to}
5640 (@var{but-do-not-erase-or-copy-to-a-read-only-buffer})
5643 @var{insert-substring-from-}@code{oldbuf}@var{-into-buffer})))
5648 @section The Definition of @code{insert-buffer}
5649 @findex insert-buffer
5651 @code{insert-buffer} is yet another buffer-related function. This
5652 command copies another buffer @emph{into} the current buffer. It is the
5653 reverse of @code{append-to-buffer} or @code{copy-to-buffer}, since they
5654 copy a region of text @emph{from} the current buffer to another buffer.
5656 Here is a discussion based on the original code. The code was
5657 simplified in 2003 and is harder to understand.
5659 (@xref{New insert-buffer, , New Body for @code{insert-buffer}}, to see
5660 a discussion of the new body.)
5662 In addition, this code illustrates the use of @code{interactive} with a
5663 buffer that might be @dfn{read-only} and the important distinction
5664 between the name of an object and the object actually referred to.
5667 * insert-buffer code::
5668 * insert-buffer interactive:: When you can read, but not write.
5669 * insert-buffer body:: The body has an @code{or} and a @code{let}.
5670 * if & or:: Using an @code{if} instead of an @code{or}.
5671 * Insert or:: How the @code{or} expression works.
5672 * Insert let:: Two @code{save-excursion} expressions.
5673 * New insert-buffer::
5677 @node insert-buffer code
5678 @unnumberedsubsec The Code for @code{insert-buffer}
5682 Here is the earlier code:
5686 (defun insert-buffer (buffer)
5687 "Insert after point the contents of BUFFER.
5688 Puts mark after the inserted text.
5689 BUFFER may be a buffer or a buffer name."
5690 (interactive "*bInsert buffer:@: ")
5693 (or (bufferp buffer)
5694 (setq buffer (get-buffer buffer)))
5695 (let (start end newmark)
5699 (setq start (point-min) end (point-max)))
5702 (insert-buffer-substring buffer start end)
5703 (setq newmark (point)))
5704 (push-mark newmark)))
5709 As with other function definitions, you can use a template to see an
5710 outline of the function:
5714 (defun insert-buffer (buffer)
5715 "@var{documentation}@dots{}"
5716 (interactive "*bInsert buffer:@: ")
5721 @node insert-buffer interactive
5722 @subsection The Interactive Expression in @code{insert-buffer}
5723 @findex interactive, @r{example use of}
5725 In @code{insert-buffer}, the argument to the @code{interactive}
5726 declaration has two parts, an asterisk, @samp{*}, and @samp{bInsert
5730 * Read-only buffer:: When a buffer cannot be modified.
5731 * b for interactive:: An existing buffer or else its name.
5734 @node Read-only buffer
5735 @unnumberedsubsubsec A Read-only Buffer
5736 @cindex Read-only buffer
5737 @cindex Asterisk for read-only buffer
5738 @findex * @r{for read-only buffer}
5740 The asterisk is for the situation when the current buffer is a
5741 read-only buffer---a buffer that cannot be modified. If
5742 @code{insert-buffer} is called when the current buffer is read-only, a
5743 message to this effect is printed in the echo area and the terminal
5744 may beep or blink at you; you will not be permitted to insert anything
5745 into current buffer. The asterisk does not need to be followed by a
5746 newline to separate it from the next argument.
5748 @node b for interactive
5749 @unnumberedsubsubsec @samp{b} in an Interactive Expression
5751 The next argument in the interactive expression starts with a lower
5752 case @samp{b}. (This is different from the code for
5753 @code{append-to-buffer}, which uses an upper-case @samp{B}.
5754 @xref{append-to-buffer, , The Definition of @code{append-to-buffer}}.)
5755 The lower-case @samp{b} tells the Lisp interpreter that the argument
5756 for @code{insert-buffer} should be an existing buffer or else its
5757 name. (The upper-case @samp{B} option provides for the possibility
5758 that the buffer does not exist.) Emacs will prompt you for the name
5759 of the buffer, offering you a default buffer, with name completion
5760 enabled. If the buffer does not exist, you receive a message that
5761 says ``No match''; your terminal may beep at you as well.
5763 The new and simplified code generates a list for @code{interactive}.
5764 It uses the @code{barf-if-buffer-read-only} and @code{read-buffer}
5765 functions with which we are already familiar and the @code{progn}
5766 special form with which we are not. (It will be described later.)
5768 @node insert-buffer body
5769 @subsection The Body of the @code{insert-buffer} Function
5771 The body of the @code{insert-buffer} function has two major parts: an
5772 @code{or} expression and a @code{let} expression. The purpose of the
5773 @code{or} expression is to ensure that the argument @code{buffer} is
5774 bound to a buffer and not just the name of a buffer. The body of the
5775 @code{let} expression contains the code which copies the other buffer
5776 into the current buffer.
5779 In outline, the two expressions fit into the @code{insert-buffer}
5784 (defun insert-buffer (buffer)
5785 "@var{documentation}@dots{}"
5786 (interactive "*bInsert buffer:@: ")
5791 (let (@var{varlist})
5792 @var{body-of-}@code{let}@dots{} )
5796 To understand how the @code{or} expression ensures that the argument
5797 @code{buffer} is bound to a buffer and not to the name of a buffer, it
5798 is first necessary to understand the @code{or} function.
5800 Before doing this, let me rewrite this part of the function using
5801 @code{if} so that you can see what is done in a manner that will be familiar.
5804 @subsection @code{insert-buffer} With an @code{if} Instead of an @code{or}
5806 The job to be done is to make sure the value of @code{buffer} is a
5807 buffer itself and not the name of a buffer. If the value is the name,
5808 then the buffer itself must be got.
5810 You can imagine yourself at a conference where an usher is wandering
5811 around holding a list with your name on it and looking for you: the
5812 usher is ``bound'' to your name, not to you; but when the usher finds
5813 you and takes your arm, the usher becomes ``bound'' to you.
5816 In Lisp, you might describe this situation like this:
5820 (if (not (holding-on-to-guest))
5821 (find-and-take-arm-of-guest))
5825 We want to do the same thing with a buffer---if we do not have the
5826 buffer itself, we want to get it.
5829 Using a predicate called @code{bufferp} that tells us whether we have a
5830 buffer (rather than its name), we can write the code like this:
5834 (if (not (bufferp buffer)) ; @r{if-part}
5835 (setq buffer (get-buffer buffer))) ; @r{then-part}
5840 Here, the true-or-false-test of the @code{if} expression is
5841 @w{@code{(not (bufferp buffer))}}; and the then-part is the expression
5842 @w{@code{(setq buffer (get-buffer buffer))}}.
5844 In the test, the function @code{bufferp} returns true if its argument is
5845 a buffer---but false if its argument is the name of the buffer. (The
5846 last character of the function name @code{bufferp} is the character
5847 @samp{p}; as we saw earlier, such use of @samp{p} is a convention that
5848 indicates that the function is a predicate, which is a term that means
5849 that the function will determine whether some property is true or false.
5850 @xref{Wrong Type of Argument, , Using the Wrong Type Object as an
5854 The function @code{not} precedes the expression @code{(bufferp buffer)},
5855 so the true-or-false-test looks like this:
5858 (not (bufferp buffer))
5862 @code{not} is a function that returns true if its argument is false
5863 and false if its argument is true. So if @code{(bufferp buffer)}
5864 returns true, the @code{not} expression returns false and vice-verse:
5865 what is ``not true'' is false and what is ``not false'' is true.
5867 Using this test, the @code{if} expression works as follows: when the
5868 value of the variable @code{buffer} is actually a buffer rather than
5869 its name, the true-or-false-test returns false and the @code{if}
5870 expression does not evaluate the then-part. This is fine, since we do
5871 not need to do anything to the variable @code{buffer} if it really is
5874 On the other hand, when the value of @code{buffer} is not a buffer
5875 itself, but the name of a buffer, the true-or-false-test returns true
5876 and the then-part of the expression is evaluated. In this case, the
5877 then-part is @code{(setq buffer (get-buffer buffer))}. This
5878 expression uses the @code{get-buffer} function to return an actual
5879 buffer itself, given its name. The @code{setq} then sets the variable
5880 @code{buffer} to the value of the buffer itself, replacing its previous
5881 value (which was the name of the buffer).
5884 @subsection The @code{or} in the Body
5886 The purpose of the @code{or} expression in the @code{insert-buffer}
5887 function is to ensure that the argument @code{buffer} is bound to a
5888 buffer and not just to the name of a buffer. The previous section shows
5889 how the job could have been done using an @code{if} expression.
5890 However, the @code{insert-buffer} function actually uses @code{or}.
5891 To understand this, it is necessary to understand how @code{or} works.
5894 An @code{or} function can have any number of arguments. It evaluates
5895 each argument in turn and returns the value of the first of its
5896 arguments that is not @code{nil}. Also, and this is a crucial feature
5897 of @code{or}, it does not evaluate any subsequent arguments after
5898 returning the first non-@code{nil} value.
5901 The @code{or} expression looks like this:
5905 (or (bufferp buffer)
5906 (setq buffer (get-buffer buffer)))
5911 The first argument to @code{or} is the expression @code{(bufferp buffer)}.
5912 This expression returns true (a non-@code{nil} value) if the buffer is
5913 actually a buffer, and not just the name of a buffer. In the @code{or}
5914 expression, if this is the case, the @code{or} expression returns this
5915 true value and does not evaluate the next expression---and this is fine
5916 with us, since we do not want to do anything to the value of
5917 @code{buffer} if it really is a buffer.
5919 On the other hand, if the value of @code{(bufferp buffer)} is @code{nil},
5920 which it will be if the value of @code{buffer} is the name of a buffer,
5921 the Lisp interpreter evaluates the next element of the @code{or}
5922 expression. This is the expression @code{(setq buffer (get-buffer
5923 buffer))}. This expression returns a non-@code{nil} value, which
5924 is the value to which it sets the variable @code{buffer}---and this
5925 value is a buffer itself, not the name of a buffer.
5927 The result of all this is that the symbol @code{buffer} is always
5928 bound to a buffer itself rather than to the name of a buffer. All
5929 this is necessary because the @code{set-buffer} function in a
5930 following line only works with a buffer itself, not with the name to a
5934 Incidentally, using @code{or}, the situation with the usher would be
5938 (or (holding-on-to-guest) (find-and-take-arm-of-guest))
5942 @subsection The @code{let} Expression in @code{insert-buffer}
5944 After ensuring that the variable @code{buffer} refers to a buffer itself
5945 and not just to the name of a buffer, the @code{insert-buffer function}
5946 continues with a @code{let} expression. This specifies three local
5947 variables, @code{start}, @code{end}, and @code{newmark} and binds them
5948 to the initial value @code{nil}. These variables are used inside the
5949 remainder of the @code{let} and temporarily hide any other occurrence of
5950 variables of the same name in Emacs until the end of the @code{let}.
5953 The body of the @code{let} contains two @code{save-excursion}
5954 expressions. First, we will look at the inner @code{save-excursion}
5955 expression in detail. The expression looks like this:
5961 (setq start (point-min) end (point-max)))
5966 The expression @code{(set-buffer buffer)} changes Emacs's attention
5967 from the current buffer to the one from which the text will copied.
5968 In that buffer, the variables @code{start} and @code{end} are set to
5969 the beginning and end of the buffer, using the commands
5970 @code{point-min} and @code{point-max}. Note that we have here an
5971 illustration of how @code{setq} is able to set two variables in the
5972 same expression. The first argument of @code{setq} is set to the
5973 value of its second, and its third argument is set to the value of its
5976 After the body of the inner @code{save-excursion} is evaluated, the
5977 @code{save-excursion} restores the original buffer, but @code{start} and
5978 @code{end} remain set to the values of the beginning and end of the
5979 buffer from which the text will be copied.
5982 The outer @code{save-excursion} expression looks like this:
5987 (@var{inner-}@code{save-excursion}@var{-expression}
5988 (@var{go-to-new-buffer-and-set-}@code{start}@var{-and-}@code{end})
5989 (insert-buffer-substring buffer start end)
5990 (setq newmark (point)))
5995 The @code{insert-buffer-substring} function copies the text
5996 @emph{into} the current buffer @emph{from} the region indicated by
5997 @code{start} and @code{end} in @code{buffer}. Since the whole of the
5998 second buffer lies between @code{start} and @code{end}, the whole of
5999 the second buffer is copied into the buffer you are editing. Next,
6000 the value of point, which will be at the end of the inserted text, is
6001 recorded in the variable @code{newmark}.
6003 After the body of the outer @code{save-excursion} is evaluated, point
6004 and mark are relocated to their original places.
6006 However, it is convenient to locate a mark at the end of the newly
6007 inserted text and locate point at its beginning. The @code{newmark}
6008 variable records the end of the inserted text. In the last line of
6009 the @code{let} expression, the @code{(push-mark newmark)} expression
6010 function sets a mark to this location. (The previous location of the
6011 mark is still accessible; it is recorded on the mark ring and you can
6012 go back to it with @kbd{C-u C-@key{SPC}}.) Meanwhile, point is
6013 located at the beginning of the inserted text, which is where it was
6014 before you called the insert function, the position of which was saved
6015 by the first @code{save-excursion}.
6018 The whole @code{let} expression looks like this:
6022 (let (start end newmark)
6026 (setq start (point-min) end (point-max)))
6027 (insert-buffer-substring buffer start end)
6028 (setq newmark (point)))
6029 (push-mark newmark))
6033 Like the @code{append-to-buffer} function, the @code{insert-buffer}
6034 function uses @code{let}, @code{save-excursion}, and
6035 @code{set-buffer}. In addition, the function illustrates one way to
6036 use @code{or}. All these functions are building blocks that we will
6037 find and use again and again.
6039 @node New insert-buffer
6040 @subsection New Body for @code{insert-buffer}
6041 @findex insert-buffer, new version body
6042 @findex new version body for insert-buffer
6044 The body in the GNU Emacs 22 version is more confusing than the original.
6047 It consists of two expressions,
6053 (insert-buffer-substring (get-buffer buffer))
6061 except, and this is what confuses novices, very important work is done
6062 inside the @code{push-mark} expression.
6064 The @code{get-buffer} function returns a buffer with the name
6065 provided. You will note that the function is @emph{not} called
6066 @code{get-buffer-create}; it does not create a buffer if one does not
6067 already exist. The buffer returned by @code{get-buffer}, an existing
6068 buffer, is passed to @code{insert-buffer-substring}, which inserts the
6069 whole of the buffer (since you did not specify anything else).
6071 The location into which the buffer is inserted is recorded by
6072 @code{push-mark}. Then the function returns @code{nil}, the value of
6073 its last command. Put another way, the @code{insert-buffer} function
6074 exists only to produce a side effect, inserting another buffer, not to
6077 @node beginning-of-buffer
6078 @section Complete Definition of @code{beginning-of-buffer}
6079 @findex beginning-of-buffer
6081 The basic structure of the @code{beginning-of-buffer} function has
6082 already been discussed. (@xref{simplified-beginning-of-buffer, , A
6083 Simplified @code{beginning-of-buffer} Definition}.)
6084 This section describes the complex part of the definition.
6086 As previously described, when invoked without an argument,
6087 @code{beginning-of-buffer} moves the cursor to the beginning of the
6088 buffer (in truth, the beginning of the accessible portion of the
6089 buffer), leaving the mark at the previous position. However, when the
6090 command is invoked with a number between one and ten, the function
6091 considers that number to be a fraction of the length of the buffer,
6092 measured in tenths, and Emacs moves the cursor that fraction of the
6093 way from the beginning of the buffer. Thus, you can either call this
6094 function with the key command @kbd{M-<}, which will move the cursor to
6095 the beginning of the buffer, or with a key command such as @kbd{C-u 7
6096 M-<} which will move the cursor to a point 70% of the way through the
6097 buffer. If a number bigger than ten is used for the argument, it
6098 moves to the end of the buffer.
6100 The @code{beginning-of-buffer} function can be called with or without an
6101 argument. The use of the argument is optional.
6104 * Optional Arguments::
6105 * beginning-of-buffer opt arg:: Example with optional argument.
6106 * beginning-of-buffer complete::
6109 @node Optional Arguments
6110 @subsection Optional Arguments
6112 Unless told otherwise, Lisp expects that a function with an argument in
6113 its function definition will be called with a value for that argument.
6114 If that does not happen, you get an error and a message that says
6115 @samp{Wrong number of arguments}.
6117 @cindex Optional arguments
6120 However, optional arguments are a feature of Lisp: a particular
6121 @dfn{keyword} is used to tell the Lisp interpreter that an argument is
6122 optional. The keyword is @code{&optional}. (The @samp{&} in front of
6123 @samp{optional} is part of the keyword.) In a function definition, if
6124 an argument follows the keyword @code{&optional}, no value need be
6125 passed to that argument when the function is called.
6128 The first line of the function definition of @code{beginning-of-buffer}
6129 therefore looks like this:
6132 (defun beginning-of-buffer (&optional arg)
6136 In outline, the whole function looks like this:
6140 (defun beginning-of-buffer (&optional arg)
6141 "@var{documentation}@dots{}"
6143 (or (@var{is-the-argument-a-cons-cell} arg)
6144 (and @var{are-both-transient-mark-mode-and-mark-active-true})
6146 (let (@var{determine-size-and-set-it})
6148 (@var{if-there-is-an-argument}
6149 @var{figure-out-where-to-go}
6156 The function is similar to the @code{simplified-beginning-of-buffer}
6157 function except that the @code{interactive} expression has @code{"P"}
6158 as an argument and the @code{goto-char} function is followed by an
6159 if-then-else expression that figures out where to put the cursor if
6160 there is an argument that is not a cons cell.
6162 (Since I do not explain a cons cell for many more chapters, please
6163 consider ignoring the function @code{consp}. @xref{List
6164 Implementation, , How Lists are Implemented}, and @ref{Cons Cell Type,
6165 , Cons Cell and List Types, elisp, The GNU Emacs Lisp Reference
6168 The @code{"P"} in the @code{interactive} expression tells Emacs to
6169 pass a prefix argument, if there is one, to the function in raw form.
6170 A prefix argument is made by typing the @key{META} key followed by a
6171 number, or by typing @kbd{C-u} and then a number. (If you don't type
6172 a number, @kbd{C-u} defaults to a cons cell with a 4. A lowercase
6173 @code{"p"} in the @code{interactive} expression causes the function to
6174 convert a prefix arg to a number.)
6176 The true-or-false-test of the @code{if} expression looks complex, but
6177 it is not: it checks whether @code{arg} has a value that is not
6178 @code{nil} and whether it is a cons cell. (That is what @code{consp}
6179 does; it checks whether its argument is a cons cell.) If @code{arg}
6180 has a value that is not @code{nil} (and is not a cons cell), which
6181 will be the case if @code{beginning-of-buffer} is called with a
6182 numeric argument, then this true-or-false-test will return true and
6183 the then-part of the @code{if} expression will be evaluated. On the
6184 other hand, if @code{beginning-of-buffer} is not called with an
6185 argument, the value of @code{arg} will be @code{nil} and the else-part
6186 of the @code{if} expression will be evaluated. The else-part is
6187 simply @code{point-min}, and when this is the outcome, the whole
6188 @code{goto-char} expression is @code{(goto-char (point-min))}, which
6189 is how we saw the @code{beginning-of-buffer} function in its
6192 @node beginning-of-buffer opt arg
6193 @subsection @code{beginning-of-buffer} with an Argument
6195 When @code{beginning-of-buffer} is called with an argument, an
6196 expression is evaluated which calculates what value to pass to
6197 @code{goto-char}. This expression is rather complicated at first sight.
6198 It includes an inner @code{if} expression and much arithmetic. It looks
6203 (if (> (buffer-size) 10000)
6204 ;; @r{Avoid overflow for large buffer sizes!}
6205 (* (prefix-numeric-value arg)
6210 size (prefix-numeric-value arg))) 10)))
6215 * Disentangle beginning-of-buffer::
6216 * Large buffer case::
6217 * Small buffer case::
6221 @node Disentangle beginning-of-buffer
6222 @unnumberedsubsubsec Disentangle @code{beginning-of-buffer}
6225 Like other complex-looking expressions, the conditional expression
6226 within @code{beginning-of-buffer} can be disentangled by looking at it
6227 as parts of a template, in this case, the template for an if-then-else
6228 expression. In skeletal form, the expression looks like this:
6232 (if (@var{buffer-is-large}
6233 @var{divide-buffer-size-by-10-and-multiply-by-arg}
6234 @var{else-use-alternate-calculation}
6238 The true-or-false-test of this inner @code{if} expression checks the
6239 size of the buffer. The reason for this is that the old version 18
6240 Emacs used numbers that are no bigger than eight million or so and in
6241 the computation that followed, the programmer feared that Emacs might
6242 try to use over-large numbers if the buffer were large. The term
6243 `overflow', mentioned in the comment, means numbers that are over
6244 large. More recent versions of Emacs use larger numbers, but this
6245 code has not been touched, if only because people now look at buffers
6246 that are far, far larger than ever before.
6248 There are two cases: if the buffer is large and if it is not.
6250 @node Large buffer case
6251 @unnumberedsubsubsec What happens in a large buffer
6253 In @code{beginning-of-buffer}, the inner @code{if} expression tests
6254 whether the size of the buffer is greater than 10,000 characters. To do
6255 this, it uses the @code{>} function and the computation of @code{size}
6256 that comes from the let expression.
6258 In the old days, the function @code{buffer-size} was used. Not only
6259 was that function called several times, it gave the size of the whole
6260 buffer, not the accessible part. The computation makes much more
6261 sense when it handles just the accessible part. (@xref{Narrowing &
6262 Widening, , Narrowing and Widening}, for more information on focusing
6263 attention to an `accessible' part.)
6266 The line looks like this:
6274 When the buffer is large, the then-part of the @code{if} expression is
6275 evaluated. It reads like this (after formatting for easy reading):
6280 (prefix-numeric-value arg)
6286 This expression is a multiplication, with two arguments to the function
6289 The first argument is @code{(prefix-numeric-value arg)}. When
6290 @code{"P"} is used as the argument for @code{interactive}, the value
6291 passed to the function as its argument is passed a ``raw prefix
6292 argument'', and not a number. (It is a number in a list.) To perform
6293 the arithmetic, a conversion is necessary, and
6294 @code{prefix-numeric-value} does the job.
6296 @findex / @r{(division)}
6298 The second argument is @code{(/ size 10)}. This expression divides
6299 the numeric value by ten---the numeric value of the size of the
6300 accessible portion of the buffer. This produces a number that tells
6301 how many characters make up one tenth of the buffer size. (In Lisp,
6302 @code{/} is used for division, just as @code{*} is used for
6306 In the multiplication expression as a whole, this amount is multiplied
6307 by the value of the prefix argument---the multiplication looks like this:
6311 (* @var{numeric-value-of-prefix-arg}
6312 @var{number-of-characters-in-one-tenth-of-the-accessible-buffer})
6317 If, for example, the prefix argument is @samp{7}, the one-tenth value
6318 will be multiplied by 7 to give a position 70% of the way through.
6321 The result of all this is that if the accessible portion of the buffer
6322 is large, the @code{goto-char} expression reads like this:
6326 (goto-char (* (prefix-numeric-value arg)
6331 This puts the cursor where we want it.
6333 @node Small buffer case
6334 @unnumberedsubsubsec What happens in a small buffer
6336 If the buffer contains fewer than 10,000 characters, a slightly
6337 different computation is performed. You might think this is not
6338 necessary, since the first computation could do the job. However, in
6339 a small buffer, the first method may not put the cursor on exactly the
6340 desired line; the second method does a better job.
6343 The code looks like this:
6345 @c Keep this on one line.
6347 (/ (+ 10 (* size (prefix-numeric-value arg))) 10))
6352 This is code in which you figure out what happens by discovering how the
6353 functions are embedded in parentheses. It is easier to read if you
6354 reformat it with each expression indented more deeply than its
6355 enclosing expression:
6363 (prefix-numeric-value arg)))
6370 Looking at parentheses, we see that the innermost operation is
6371 @code{(prefix-numeric-value arg)}, which converts the raw argument to
6372 a number. In the following expression, this number is multiplied by
6373 the size of the accessible portion of the buffer:
6376 (* size (prefix-numeric-value arg))
6380 This multiplication creates a number that may be larger than the size of
6381 the buffer---seven times larger if the argument is 7, for example. Ten
6382 is then added to this number and finally the large number is divided by
6383 ten to provide a value that is one character larger than the percentage
6384 position in the buffer.
6386 The number that results from all this is passed to @code{goto-char} and
6387 the cursor is moved to that point.
6390 @node beginning-of-buffer complete
6391 @subsection The Complete @code{beginning-of-buffer}
6394 Here is the complete text of the @code{beginning-of-buffer} function:
6400 (defun beginning-of-buffer (&optional arg)
6401 "Move point to the beginning of the buffer;
6402 leave mark at previous position.
6403 With \\[universal-argument] prefix,
6404 do not set mark at previous position.
6406 put point N/10 of the way from the beginning.
6408 If the buffer is narrowed,
6409 this command uses the beginning and size
6410 of the accessible part of the buffer.
6414 Don't use this command in Lisp programs!
6415 \(goto-char (point-min)) is faster
6416 and avoids clobbering the mark."
6419 (and transient-mark-mode mark-active)
6423 (let ((size (- (point-max) (point-min))))
6424 (goto-char (if (and arg (not (consp arg)))
6427 ;; Avoid overflow for large buffer sizes!
6428 (* (prefix-numeric-value arg)
6430 (/ (+ 10 (* size (prefix-numeric-value arg)))
6433 (if arg (forward-line 1)))
6438 From before GNU Emacs 22
6441 (defun beginning-of-buffer (&optional arg)
6442 "Move point to the beginning of the buffer;
6443 leave mark at previous position.
6444 With arg N, put point N/10 of the way
6445 from the true beginning.
6448 Don't use this in Lisp programs!
6449 \(goto-char (point-min)) is faster
6450 and does not set the mark."
6457 (if (> (buffer-size) 10000)
6458 ;; @r{Avoid overflow for large buffer sizes!}
6459 (* (prefix-numeric-value arg)
6460 (/ (buffer-size) 10))
6463 (/ (+ 10 (* (buffer-size)
6464 (prefix-numeric-value arg)))
6467 (if arg (forward-line 1)))
6473 Except for two small points, the previous discussion shows how this
6474 function works. The first point deals with a detail in the
6475 documentation string, and the second point concerns the last line of
6479 In the documentation string, there is reference to an expression:
6482 \\[universal-argument]
6486 A @samp{\\} is used before the first square bracket of this
6487 expression. This @samp{\\} tells the Lisp interpreter to substitute
6488 whatever key is currently bound to the @samp{[@dots{}]}. In the case
6489 of @code{universal-argument}, that is usually @kbd{C-u}, but it might
6490 be different. (@xref{Documentation Tips, , Tips for Documentation
6491 Strings, elisp, The GNU Emacs Lisp Reference Manual}, for more
6495 Finally, the last line of the @code{beginning-of-buffer} command says
6496 to move point to the beginning of the next line if the command is
6497 invoked with an argument:
6500 (if arg (forward-line 1)))
6504 This puts the cursor at the beginning of the first line after the
6505 appropriate tenths position in the buffer. This is a flourish that
6506 means that the cursor is always located @emph{at least} the requested
6507 tenths of the way through the buffer, which is a nicety that is,
6508 perhaps, not necessary, but which, if it did not occur, would be sure
6511 On the other hand, it also means that if you specify the command with
6512 a @kbd{C-u}, but without a number, that is to say, if the `raw prefix
6513 argument' is simply a cons cell, then the command puts you at the
6514 beginning of the second line @dots{} I don't know whether this is
6515 intended or whether no one has dealt with the code to avoid this
6518 @node Second Buffer Related Review
6521 Here is a brief summary of some of the topics covered in this chapter.
6525 Evaluate each argument in sequence, and return the value of the first
6526 argument that is not @code{nil}; if none return a value that is not
6527 @code{nil}, return @code{nil}. In brief, return the first true value
6528 of the arguments; return a true value if one @emph{or} any of the
6532 Evaluate each argument in sequence, and if any are @code{nil}, return
6533 @code{nil}; if none are @code{nil}, return the value of the last
6534 argument. In brief, return a true value only if all the arguments are
6535 true; return a true value if one @emph{and} each of the others is
6539 A keyword used to indicate that an argument to a function definition
6540 is optional; this means that the function can be evaluated without the
6541 argument, if desired.
6543 @item prefix-numeric-value
6544 Convert the `raw prefix argument' produced by @code{(interactive
6545 "P")} to a numeric value.
6548 Move point forward to the beginning of the next line, or if the argument
6549 is greater than one, forward that many lines. If it can't move as far
6550 forward as it is supposed to, @code{forward-line} goes forward as far as
6551 it can and then returns a count of the number of additional lines it was
6552 supposed to move but couldn't.
6555 Delete the entire contents of the current buffer.
6558 Return @code{t} if its argument is a buffer; otherwise return @code{nil}.
6561 @node optional Exercise
6562 @section @code{optional} Argument Exercise
6564 Write an interactive function with an optional argument that tests
6565 whether its argument, a number, is greater than or equal to, or else,
6566 less than the value of @code{fill-column}, and tells you which, in a
6567 message. However, if you do not pass an argument to the function, use
6568 56 as a default value.
6570 @node Narrowing & Widening
6571 @chapter Narrowing and Widening
6572 @cindex Focusing attention (narrowing)
6576 Narrowing is a feature of Emacs that makes it possible for you to focus
6577 on a specific part of a buffer, and work without accidentally changing
6578 other parts. Narrowing is normally disabled since it can confuse
6582 * Narrowing advantages:: The advantages of narrowing
6583 * save-restriction:: The @code{save-restriction} special form.
6584 * what-line:: The number of the line that point is on.
6589 @node Narrowing advantages
6590 @unnumberedsec The Advantages of Narrowing
6593 With narrowing, the rest of a buffer is made invisible, as if it weren't
6594 there. This is an advantage if, for example, you want to replace a word
6595 in one part of a buffer but not in another: you narrow to the part you want
6596 and the replacement is carried out only in that section, not in the rest
6597 of the buffer. Searches will only work within a narrowed region, not
6598 outside of one, so if you are fixing a part of a document, you can keep
6599 yourself from accidentally finding parts you do not need to fix by
6600 narrowing just to the region you want.
6601 (The key binding for @code{narrow-to-region} is @kbd{C-x n n}.)
6603 However, narrowing does make the rest of the buffer invisible, which
6604 can scare people who inadvertently invoke narrowing and think they
6605 have deleted a part of their file. Moreover, the @code{undo} command
6606 (which is usually bound to @kbd{C-x u}) does not turn off narrowing
6607 (nor should it), so people can become quite desperate if they do not
6608 know that they can return the rest of a buffer to visibility with the
6609 @code{widen} command.
6610 (The key binding for @code{widen} is @kbd{C-x n w}.)
6612 Narrowing is just as useful to the Lisp interpreter as to a human.
6613 Often, an Emacs Lisp function is designed to work on just part of a
6614 buffer; or conversely, an Emacs Lisp function needs to work on all of a
6615 buffer that has been narrowed. The @code{what-line} function, for
6616 example, removes the narrowing from a buffer, if it has any narrowing
6617 and when it has finished its job, restores the narrowing to what it was.
6618 On the other hand, the @code{count-lines} function
6619 uses narrowing to restrict itself to just that portion
6620 of the buffer in which it is interested and then restores the previous
6623 @node save-restriction
6624 @section The @code{save-restriction} Special Form
6625 @findex save-restriction
6627 In Emacs Lisp, you can use the @code{save-restriction} special form to
6628 keep track of whatever narrowing is in effect, if any. When the Lisp
6629 interpreter meets with @code{save-restriction}, it executes the code
6630 in the body of the @code{save-restriction} expression, and then undoes
6631 any changes to narrowing that the code caused. If, for example, the
6632 buffer is narrowed and the code that follows @code{save-restriction}
6633 gets rid of the narrowing, @code{save-restriction} returns the buffer
6634 to its narrowed region afterwards. In the @code{what-line} command,
6635 any narrowing the buffer may have is undone by the @code{widen}
6636 command that immediately follows the @code{save-restriction} command.
6637 Any original narrowing is restored just before the completion of the
6641 The template for a @code{save-restriction} expression is simple:
6651 The body of the @code{save-restriction} is one or more expressions that
6652 will be evaluated in sequence by the Lisp interpreter.
6654 Finally, a point to note: when you use both @code{save-excursion} and
6655 @code{save-restriction}, one right after the other, you should use
6656 @code{save-excursion} outermost. If you write them in reverse order,
6657 you may fail to record narrowing in the buffer to which Emacs switches
6658 after calling @code{save-excursion}. Thus, when written together,
6659 @code{save-excursion} and @code{save-restriction} should be written
6670 In other circumstances, when not written together, the
6671 @code{save-excursion} and @code{save-restriction} special forms must
6672 be written in the order appropriate to the function.
6688 /usr/local/src/emacs/lisp/simple.el
6691 "Print the current buffer line number and narrowed line number of point."
6693 (let ((start (point-min))
6694 (n (line-number-at-pos)))
6696 (message "Line %d" n)
6700 (message "line %d (narrowed line %d)"
6701 (+ n (line-number-at-pos start) -1) n))))))
6703 (defun line-number-at-pos (&optional pos)
6704 "Return (narrowed) buffer line number at position POS.
6705 If POS is nil, use current buffer location.
6706 Counting starts at (point-min), so the value refers
6707 to the contents of the accessible portion of the buffer."
6708 (let ((opoint (or pos (point))) start)
6710 (goto-char (point-min))
6711 (setq start (point))
6714 (1+ (count-lines start (point))))))
6716 (defun count-lines (start end)
6717 "Return number of lines between START and END.
6718 This is usually the number of newlines between them,
6719 but can be one more if START is not equal to END
6720 and the greater of them is not at the start of a line."
6723 (narrow-to-region start end)
6724 (goto-char (point-min))
6725 (if (eq selective-display t)
6728 (while (re-search-forward "[\n\C-m]" nil t 40)
6729 (setq done (+ 40 done)))
6730 (while (re-search-forward "[\n\C-m]" nil t 1)
6731 (setq done (+ 1 done)))
6732 (goto-char (point-max))
6733 (if (and (/= start end)
6737 (- (buffer-size) (forward-line (buffer-size)))))))
6741 @section @code{what-line}
6743 @cindex Widening, example of
6745 The @code{what-line} command tells you the number of the line in which
6746 the cursor is located. The function illustrates the use of the
6747 @code{save-restriction} and @code{save-excursion} commands. Here is the
6748 original text of the function:
6753 "Print the current line number (in the buffer) of point."
6760 (1+ (count-lines 1 (point)))))))
6764 (In recent versions of GNU Emacs, the @code{what-line} function has
6765 been expanded to tell you your line number in a narrowed buffer as
6766 well as your line number in a widened buffer. The recent version is
6767 more complex than the version shown here. If you feel adventurous,
6768 you might want to look at it after figuring out how this version
6769 works. You will probably need to use @kbd{C-h f}
6770 (@code{describe-function}). The newer version uses a conditional to
6771 determine whether the buffer has been narrowed.
6773 (Also, it uses @code{line-number-at-pos}, which among other simple
6774 expressions, such as @code{(goto-char (point-min))}, moves point to
6775 the beginning of the current line with @code{(forward-line 0)} rather
6776 than @code{beginning-of-line}.)
6778 The @code{what-line} function as shown here has a documentation line
6779 and is interactive, as you would expect. The next two lines use the
6780 functions @code{save-restriction} and @code{widen}.
6782 The @code{save-restriction} special form notes whatever narrowing is in
6783 effect, if any, in the current buffer and restores that narrowing after
6784 the code in the body of the @code{save-restriction} has been evaluated.
6786 The @code{save-restriction} special form is followed by @code{widen}.
6787 This function undoes any narrowing the current buffer may have had
6788 when @code{what-line} was called. (The narrowing that was there is
6789 the narrowing that @code{save-restriction} remembers.) This widening
6790 makes it possible for the line counting commands to count from the
6791 beginning of the buffer. Otherwise, they would have been limited to
6792 counting within the accessible region. Any original narrowing is
6793 restored just before the completion of the function by the
6794 @code{save-restriction} special form.
6796 The call to @code{widen} is followed by @code{save-excursion}, which
6797 saves the location of the cursor (i.e., of point) and of the mark, and
6798 restores them after the code in the body of the @code{save-excursion}
6799 uses the @code{beginning-of-line} function to move point.
6801 (Note that the @code{(widen)} expression comes between the
6802 @code{save-restriction} and @code{save-excursion} special forms. When
6803 you write the two @code{save- @dots{}} expressions in sequence, write
6804 @code{save-excursion} outermost.)
6807 The last two lines of the @code{what-line} function are functions to
6808 count the number of lines in the buffer and then print the number in the
6814 (1+ (count-lines 1 (point)))))))
6818 The @code{message} function prints a one-line message at the bottom of
6819 the Emacs screen. The first argument is inside of quotation marks and
6820 is printed as a string of characters. However, it may contain a
6821 @samp{%d} expression to print a following argument. @samp{%d} prints
6822 the argument as a decimal, so the message will say something such as
6826 The number that is printed in place of the @samp{%d} is computed by the
6827 last line of the function:
6830 (1+ (count-lines 1 (point)))
6836 (defun count-lines (start end)
6837 "Return number of lines between START and END.
6838 This is usually the number of newlines between them,
6839 but can be one more if START is not equal to END
6840 and the greater of them is not at the start of a line."
6843 (narrow-to-region start end)
6844 (goto-char (point-min))
6845 (if (eq selective-display t)
6848 (while (re-search-forward "[\n\C-m]" nil t 40)
6849 (setq done (+ 40 done)))
6850 (while (re-search-forward "[\n\C-m]" nil t 1)
6851 (setq done (+ 1 done)))
6852 (goto-char (point-max))
6853 (if (and (/= start end)
6857 (- (buffer-size) (forward-line (buffer-size)))))))
6861 What this does is count the lines from the first position of the
6862 buffer, indicated by the @code{1}, up to @code{(point)}, and then add
6863 one to that number. (The @code{1+} function adds one to its
6864 argument.) We add one to it because line 2 has only one line before
6865 it, and @code{count-lines} counts only the lines @emph{before} the
6868 After @code{count-lines} has done its job, and the message has been
6869 printed in the echo area, the @code{save-excursion} restores point and
6870 mark to their original positions; and @code{save-restriction} restores
6871 the original narrowing, if any.
6873 @node narrow Exercise
6874 @section Exercise with Narrowing
6876 Write a function that will display the first 60 characters of the
6877 current buffer, even if you have narrowed the buffer to its latter
6878 half so that the first line is inaccessible. Restore point, mark, and
6879 narrowing. For this exercise, you need to use a whole potpourri of
6880 functions, including @code{save-restriction}, @code{widen},
6881 @code{goto-char}, @code{point-min}, @code{message}, and
6882 @code{buffer-substring}.
6884 @cindex Properties, mention of @code{buffer-substring-no-properties}
6885 (@code{buffer-substring} is a previously unmentioned function you will
6886 have to investigate yourself; or perhaps you will have to use
6887 @code{buffer-substring-no-properties} or
6888 @code{filter-buffer-substring} @dots{}, yet other functions. Text
6889 properties are a feature otherwise not discussed here. @xref{Text
6890 Properties, , Text Properties, elisp, The GNU Emacs Lisp Reference
6893 Additionally, do you really need @code{goto-char} or @code{point-min}?
6894 Or can you write the function without them?
6896 @node car cdr & cons
6897 @chapter @code{car}, @code{cdr}, @code{cons}: Fundamental Functions
6898 @findex car, @r{introduced}
6899 @findex cdr, @r{introduced}
6901 In Lisp, @code{car}, @code{cdr}, and @code{cons} are fundamental
6902 functions. The @code{cons} function is used to construct lists, and
6903 the @code{car} and @code{cdr} functions are used to take them apart.
6905 In the walk through of the @code{copy-region-as-kill} function, we
6906 will see @code{cons} as well as two variants on @code{cdr},
6907 namely, @code{setcdr} and @code{nthcdr}. (@xref{copy-region-as-kill}.)
6910 * Strange Names:: An historical aside: why the strange names?
6911 * car & cdr:: Functions for extracting part of a list.
6912 * cons:: Constructing a list.
6913 * nthcdr:: Calling @code{cdr} repeatedly.
6915 * setcar:: Changing the first element of a list.
6916 * setcdr:: Changing the rest of a list.
6922 @unnumberedsec Strange Names
6925 The name of the @code{cons} function is not unreasonable: it is an
6926 abbreviation of the word `construct'. The origins of the names for
6927 @code{car} and @code{cdr}, on the other hand, are esoteric: @code{car}
6928 is an acronym from the phrase `Contents of the Address part of the
6929 Register'; and @code{cdr} (pronounced `could-er') is an acronym from
6930 the phrase `Contents of the Decrement part of the Register'. These
6931 phrases refer to specific pieces of hardware on the very early
6932 computer on which the original Lisp was developed. Besides being
6933 obsolete, the phrases have been completely irrelevant for more than 25
6934 years to anyone thinking about Lisp. Nonetheless, although a few
6935 brave scholars have begun to use more reasonable names for these
6936 functions, the old terms are still in use. In particular, since the
6937 terms are used in the Emacs Lisp source code, we will use them in this
6941 @section @code{car} and @code{cdr}
6943 The @sc{car} of a list is, quite simply, the first item in the list.
6944 Thus the @sc{car} of the list @code{(rose violet daisy buttercup)} is
6948 If you are reading this in Info in GNU Emacs, you can see this by
6949 evaluating the following:
6952 (car '(rose violet daisy buttercup))
6956 After evaluating the expression, @code{rose} will appear in the echo
6959 Clearly, a more reasonable name for the @code{car} function would be
6960 @code{first} and this is often suggested.
6962 @code{car} does not remove the first item from the list; it only reports
6963 what it is. After @code{car} has been applied to a list, the list is
6964 still the same as it was. In the jargon, @code{car} is
6965 `non-destructive'. This feature turns out to be important.
6967 The @sc{cdr} of a list is the rest of the list, that is, the
6968 @code{cdr} function returns the part of the list that follows the
6969 first item. Thus, while the @sc{car} of the list @code{'(rose violet
6970 daisy buttercup)} is @code{rose}, the rest of the list, the value
6971 returned by the @code{cdr} function, is @code{(violet daisy
6975 You can see this by evaluating the following in the usual way:
6978 (cdr '(rose violet daisy buttercup))
6982 When you evaluate this, @code{(violet daisy buttercup)} will appear in
6985 Like @code{car}, @code{cdr} does not remove any elements from the
6986 list---it just returns a report of what the second and subsequent
6989 Incidentally, in the example, the list of flowers is quoted. If it were
6990 not, the Lisp interpreter would try to evaluate the list by calling
6991 @code{rose} as a function. In this example, we do not want to do that.
6993 Clearly, a more reasonable name for @code{cdr} would be @code{rest}.
6995 (There is a lesson here: when you name new functions, consider very
6996 carefully what you are doing, since you may be stuck with the names
6997 for far longer than you expect. The reason this document perpetuates
6998 these names is that the Emacs Lisp source code uses them, and if I did
6999 not use them, you would have a hard time reading the code; but do,
7000 please, try to avoid using these terms yourself. The people who come
7001 after you will be grateful to you.)
7003 When @code{car} and @code{cdr} are applied to a list made up of symbols,
7004 such as the list @code{(pine fir oak maple)}, the element of the list
7005 returned by the function @code{car} is the symbol @code{pine} without
7006 any parentheses around it. @code{pine} is the first element in the
7007 list. However, the @sc{cdr} of the list is a list itself, @code{(fir
7008 oak maple)}, as you can see by evaluating the following expressions in
7013 (car '(pine fir oak maple))
7015 (cdr '(pine fir oak maple))
7019 On the other hand, in a list of lists, the first element is itself a
7020 list. @code{car} returns this first element as a list. For example,
7021 the following list contains three sub-lists, a list of carnivores, a
7022 list of herbivores and a list of sea mammals:
7026 (car '((lion tiger cheetah)
7027 (gazelle antelope zebra)
7028 (whale dolphin seal)))
7033 In this example, the first element or @sc{car} of the list is the list of
7034 carnivores, @code{(lion tiger cheetah)}, and the rest of the list is
7035 @code{((gazelle antelope zebra) (whale dolphin seal))}.
7039 (cdr '((lion tiger cheetah)
7040 (gazelle antelope zebra)
7041 (whale dolphin seal)))
7045 It is worth saying again that @code{car} and @code{cdr} are
7046 non-destructive---that is, they do not modify or change lists to which
7047 they are applied. This is very important for how they are used.
7049 Also, in the first chapter, in the discussion about atoms, I said that
7050 in Lisp, ``certain kinds of atom, such as an array, can be separated
7051 into parts; but the mechanism for doing this is different from the
7052 mechanism for splitting a list. As far as Lisp is concerned, the
7053 atoms of a list are unsplittable.'' (@xref{Lisp Atoms}.) The
7054 @code{car} and @code{cdr} functions are used for splitting lists and
7055 are considered fundamental to Lisp. Since they cannot split or gain
7056 access to the parts of an array, an array is considered an atom.
7057 Conversely, the other fundamental function, @code{cons}, can put
7058 together or construct a list, but not an array. (Arrays are handled
7059 by array-specific functions. @xref{Arrays, , Arrays, elisp, The GNU
7060 Emacs Lisp Reference Manual}.)
7063 @section @code{cons}
7064 @findex cons, @r{introduced}
7066 The @code{cons} function constructs lists; it is the inverse of
7067 @code{car} and @code{cdr}. For example, @code{cons} can be used to make
7068 a four element list from the three element list, @code{(fir oak maple)}:
7071 (cons 'pine '(fir oak maple))
7076 After evaluating this list, you will see
7079 (pine fir oak maple)
7083 appear in the echo area. @code{cons} causes the creation of a new
7084 list in which the element is followed by the elements of the original
7087 We often say that `@code{cons} puts a new element at the beginning of
7088 a list; it attaches or pushes elements onto the list', but this
7089 phrasing can be misleading, since @code{cons} does not change an
7090 existing list, but creates a new one.
7092 Like @code{car} and @code{cdr}, @code{cons} is non-destructive.
7096 * length:: How to find the length of a list.
7101 @unnumberedsubsec Build a list
7104 @code{cons} must have a list to attach to.@footnote{Actually, you can
7105 @code{cons} an element to an atom to produce a dotted pair. Dotted
7106 pairs are not discussed here; see @ref{Dotted Pair Notation, , Dotted
7107 Pair Notation, elisp, The GNU Emacs Lisp Reference Manual}.} You
7108 cannot start from absolutely nothing. If you are building a list, you
7109 need to provide at least an empty list at the beginning. Here is a
7110 series of @code{cons} expressions that build up a list of flowers. If
7111 you are reading this in Info in GNU Emacs, you can evaluate each of
7112 the expressions in the usual way; the value is printed in this text
7113 after @samp{@result{}}, which you may read as `evaluates to'.
7117 (cons 'buttercup ())
7118 @result{} (buttercup)
7122 (cons 'daisy '(buttercup))
7123 @result{} (daisy buttercup)
7127 (cons 'violet '(daisy buttercup))
7128 @result{} (violet daisy buttercup)
7132 (cons 'rose '(violet daisy buttercup))
7133 @result{} (rose violet daisy buttercup)
7138 In the first example, the empty list is shown as @code{()} and a list
7139 made up of @code{buttercup} followed by the empty list is constructed.
7140 As you can see, the empty list is not shown in the list that was
7141 constructed. All that you see is @code{(buttercup)}. The empty list is
7142 not counted as an element of a list because there is nothing in an empty
7143 list. Generally speaking, an empty list is invisible.
7145 The second example, @code{(cons 'daisy '(buttercup))} constructs a new,
7146 two element list by putting @code{daisy} in front of @code{buttercup};
7147 and the third example constructs a three element list by putting
7148 @code{violet} in front of @code{daisy} and @code{buttercup}.
7151 @subsection Find the Length of a List: @code{length}
7154 You can find out how many elements there are in a list by using the Lisp
7155 function @code{length}, as in the following examples:
7159 (length '(buttercup))
7164 (length '(daisy buttercup))
7169 (length (cons 'violet '(daisy buttercup)))
7175 In the third example, the @code{cons} function is used to construct a
7176 three element list which is then passed to the @code{length} function as
7180 We can also use @code{length} to count the number of elements in an
7191 As you would expect, the number of elements in an empty list is zero.
7193 An interesting experiment is to find out what happens if you try to find
7194 the length of no list at all; that is, if you try to call @code{length}
7195 without giving it an argument, not even an empty list:
7203 What you see, if you evaluate this, is the error message
7206 Lisp error: (wrong-number-of-arguments length 0)
7210 This means that the function receives the wrong number of
7211 arguments, zero, when it expects some other number of arguments. In
7212 this case, one argument is expected, the argument being a list whose
7213 length the function is measuring. (Note that @emph{one} list is
7214 @emph{one} argument, even if the list has many elements inside it.)
7216 The part of the error message that says @samp{length} is the name of
7220 @code{length} is still a subroutine, but you need C-h f to discover that.
7222 In an earlier version:
7223 This is written with a special notation, @samp{#<subr},
7224 that indicates that the function @code{length} is one of the primitive
7225 functions written in C rather than in Emacs Lisp. (@samp{subr} is an
7226 abbreviation for `subroutine'.) @xref{What Is a Function, , What Is a
7227 Function?, elisp , The GNU Emacs Lisp Reference Manual}, for more
7232 @section @code{nthcdr}
7235 The @code{nthcdr} function is associated with the @code{cdr} function.
7236 What it does is take the @sc{cdr} of a list repeatedly.
7238 If you take the @sc{cdr} of the list @code{(pine fir
7239 oak maple)}, you will be returned the list @code{(fir oak maple)}. If you
7240 repeat this on what was returned, you will be returned the list
7241 @code{(oak maple)}. (Of course, repeated @sc{cdr}ing on the original
7242 list will just give you the original @sc{cdr} since the function does
7243 not change the list. You need to evaluate the @sc{cdr} of the
7244 @sc{cdr} and so on.) If you continue this, eventually you will be
7245 returned an empty list, which in this case, instead of being shown as
7246 @code{()} is shown as @code{nil}.
7249 For review, here is a series of repeated @sc{cdr}s, the text following
7250 the @samp{@result{}} shows what is returned.
7254 (cdr '(pine fir oak maple))
7255 @result{}(fir oak maple)
7259 (cdr '(fir oak maple))
7260 @result{} (oak maple)
7285 You can also do several @sc{cdr}s without printing the values in
7290 (cdr (cdr '(pine fir oak maple)))
7291 @result{} (oak maple)
7296 In this example, the Lisp interpreter evaluates the innermost list first.
7297 The innermost list is quoted, so it just passes the list as it is to the
7298 innermost @code{cdr}. This @code{cdr} passes a list made up of the
7299 second and subsequent elements of the list to the outermost @code{cdr},
7300 which produces a list composed of the third and subsequent elements of
7301 the original list. In this example, the @code{cdr} function is repeated
7302 and returns a list that consists of the original list without its
7305 The @code{nthcdr} function does the same as repeating the call to
7306 @code{cdr}. In the following example, the argument 2 is passed to the
7307 function @code{nthcdr}, along with the list, and the value returned is
7308 the list without its first two items, which is exactly the same
7309 as repeating @code{cdr} twice on the list:
7313 (nthcdr 2 '(pine fir oak maple))
7314 @result{} (oak maple)
7319 Using the original four element list, we can see what happens when
7320 various numeric arguments are passed to @code{nthcdr}, including 0, 1,
7325 ;; @r{Leave the list as it was.}
7326 (nthcdr 0 '(pine fir oak maple))
7327 @result{} (pine fir oak maple)
7331 ;; @r{Return a copy without the first element.}
7332 (nthcdr 1 '(pine fir oak maple))
7333 @result{} (fir oak maple)
7337 ;; @r{Return a copy of the list without three elements.}
7338 (nthcdr 3 '(pine fir oak maple))
7343 ;; @r{Return a copy lacking all four elements.}
7344 (nthcdr 4 '(pine fir oak maple))
7349 ;; @r{Return a copy lacking all elements.}
7350 (nthcdr 5 '(pine fir oak maple))
7359 The @code{nthcdr} function takes the @sc{cdr} of a list repeatedly.
7360 The @code{nth} function takes the @sc{car} of the result returned by
7361 @code{nthcdr}. It returns the Nth element of the list.
7364 Thus, if it were not defined in C for speed, the definition of
7365 @code{nth} would be:
7370 "Returns the Nth element of LIST.
7371 N counts from zero. If LIST is not that long, nil is returned."
7372 (car (nthcdr n list)))
7377 (Originally, @code{nth} was defined in Emacs Lisp in @file{subr.el},
7378 but its definition was redone in C in the 1980s.)
7380 The @code{nth} function returns a single element of a list.
7381 This can be very convenient.
7383 Note that the elements are numbered from zero, not one. That is to
7384 say, the first element of a list, its @sc{car} is the zeroth element.
7385 This is called `zero-based' counting and often bothers people who
7386 are accustomed to the first element in a list being number one, which
7394 (nth 0 '("one" "two" "three"))
7397 (nth 1 '("one" "two" "three"))
7402 It is worth mentioning that @code{nth}, like @code{nthcdr} and
7403 @code{cdr}, does not change the original list---the function is
7404 non-destructive. This is in sharp contrast to the @code{setcar} and
7405 @code{setcdr} functions.
7408 @section @code{setcar}
7411 As you might guess from their names, the @code{setcar} and @code{setcdr}
7412 functions set the @sc{car} or the @sc{cdr} of a list to a new value.
7413 They actually change the original list, unlike @code{car} and @code{cdr}
7414 which leave the original list as it was. One way to find out how this
7415 works is to experiment. We will start with the @code{setcar} function.
7418 First, we can make a list and then set the value of a variable to the
7419 list, using the @code{setq} function. Here is a list of animals:
7422 (setq animals '(antelope giraffe lion tiger))
7426 If you are reading this in Info inside of GNU Emacs, you can evaluate
7427 this expression in the usual fashion, by positioning the cursor after
7428 the expression and typing @kbd{C-x C-e}. (I'm doing this right here
7429 as I write this. This is one of the advantages of having the
7430 interpreter built into the computing environment. Incidentally, when
7431 there is nothing on the line after the final parentheses, such as a
7432 comment, point can be on the next line. Thus, if your cursor is in
7433 the first column of the next line, you do not need to move it.
7434 Indeed, Emacs permits any amount of white space after the final
7438 When we evaluate the variable @code{animals}, we see that it is bound to
7439 the list @code{(antelope giraffe lion tiger)}:
7444 @result{} (antelope giraffe lion tiger)
7449 Put another way, the variable @code{animals} points to the list
7450 @code{(antelope giraffe lion tiger)}.
7452 Next, evaluate the function @code{setcar} while passing it two
7453 arguments, the variable @code{animals} and the quoted symbol
7454 @code{hippopotamus}; this is done by writing the three element list
7455 @code{(setcar animals 'hippopotamus)} and then evaluating it in the
7459 (setcar animals 'hippopotamus)
7464 After evaluating this expression, evaluate the variable @code{animals}
7465 again. You will see that the list of animals has changed:
7470 @result{} (hippopotamus giraffe lion tiger)
7475 The first element on the list, @code{antelope} is replaced by
7476 @code{hippopotamus}.
7478 So we can see that @code{setcar} did not add a new element to the list
7479 as @code{cons} would have; it replaced @code{antelope} with
7480 @code{hippopotamus}; it @emph{changed} the list.
7483 @section @code{setcdr}
7486 The @code{setcdr} function is similar to the @code{setcar} function,
7487 except that the function replaces the second and subsequent elements of
7488 a list rather than the first element.
7490 (To see how to change the last element of a list, look ahead to
7491 @ref{kill-new function, , The @code{kill-new} function}, which uses
7492 the @code{nthcdr} and @code{setcdr} functions.)
7495 To see how this works, set the value of the variable to a list of
7496 domesticated animals by evaluating the following expression:
7499 (setq domesticated-animals '(horse cow sheep goat))
7504 If you now evaluate the list, you will be returned the list
7505 @code{(horse cow sheep goat)}:
7509 domesticated-animals
7510 @result{} (horse cow sheep goat)
7515 Next, evaluate @code{setcdr} with two arguments, the name of the
7516 variable which has a list as its value, and the list to which the
7517 @sc{cdr} of the first list will be set;
7520 (setcdr domesticated-animals '(cat dog))
7524 If you evaluate this expression, the list @code{(cat dog)} will appear
7525 in the echo area. This is the value returned by the function. The
7526 result we are interested in is the ``side effect'', which we can see by
7527 evaluating the variable @code{domesticated-animals}:
7531 domesticated-animals
7532 @result{} (horse cat dog)
7537 Indeed, the list is changed from @code{(horse cow sheep goat)} to
7538 @code{(horse cat dog)}. The @sc{cdr} of the list is changed from
7539 @code{(cow sheep goat)} to @code{(cat dog)}.
7544 Construct a list of four birds by evaluating several expressions with
7545 @code{cons}. Find out what happens when you @code{cons} a list onto
7546 itself. Replace the first element of the list of four birds with a
7547 fish. Replace the rest of that list with a list of other fish.
7549 @node Cutting & Storing Text
7550 @chapter Cutting and Storing Text
7551 @cindex Cutting and storing text
7552 @cindex Storing and cutting text
7553 @cindex Killing text
7554 @cindex Clipping text
7555 @cindex Erasing text
7556 @cindex Deleting text
7558 Whenever you cut or clip text out of a buffer with a `kill' command in
7559 GNU Emacs, it is stored in a list and you can bring it back with a
7562 (The use of the word `kill' in Emacs for processes which specifically
7563 @emph{do not} destroy the values of the entities is an unfortunate
7564 historical accident. A much more appropriate word would be `clip' since
7565 that is what the kill commands do; they clip text out of a buffer and
7566 put it into storage from which it can be brought back. I have often
7567 been tempted to replace globally all occurrences of `kill' in the Emacs
7568 sources with `clip' and all occurrences of `killed' with `clipped'.)
7571 * Storing Text:: Text is stored in a list.
7572 * zap-to-char:: Cutting out text up to a character.
7573 * kill-region:: Cutting text out of a region.
7574 * copy-region-as-kill:: A definition for copying text.
7575 * Digression into C:: Minor note on C programming language macros.
7576 * defvar:: How to give a variable an initial value.
7577 * cons & search-fwd Review::
7578 * search Exercises::
7583 @unnumberedsec Storing Text in a List
7586 When text is cut out of a buffer, it is stored on a list. Successive
7587 pieces of text are stored on the list successively, so the list might
7591 ("a piece of text" "previous piece")
7596 The function @code{cons} can be used to create a new list from a piece
7597 of text (an `atom', to use the jargon) and an existing list, like
7602 (cons "another piece"
7603 '("a piece of text" "previous piece"))
7609 If you evaluate this expression, a list of three elements will appear in
7613 ("another piece" "a piece of text" "previous piece")
7616 With the @code{car} and @code{nthcdr} functions, you can retrieve
7617 whichever piece of text you want. For example, in the following code,
7618 @code{nthcdr 1 @dots{}} returns the list with the first item removed;
7619 and the @code{car} returns the first element of that remainder---the
7620 second element of the original list:
7624 (car (nthcdr 1 '("another piece"
7627 @result{} "a piece of text"
7631 The actual functions in Emacs are more complex than this, of course.
7632 The code for cutting and retrieving text has to be written so that
7633 Emacs can figure out which element in the list you want---the first,
7634 second, third, or whatever. In addition, when you get to the end of
7635 the list, Emacs should give you the first element of the list, rather
7636 than nothing at all.
7638 The list that holds the pieces of text is called the @dfn{kill ring}.
7639 This chapter leads up to a description of the kill ring and how it is
7640 used by first tracing how the @code{zap-to-char} function works. This
7641 function uses (or `calls') a function that invokes a function that
7642 manipulates the kill ring. Thus, before reaching the mountains, we
7643 climb the foothills.
7645 A subsequent chapter describes how text that is cut from the buffer is
7646 retrieved. @xref{Yanking, , Yanking Text Back}.
7649 @section @code{zap-to-char}
7652 @c FIXME remove obsolete stuff
7653 The @code{zap-to-char} function changed little between GNU Emacs
7654 version 19 and GNU Emacs version 22. However, @code{zap-to-char}
7655 calls another function, @code{kill-region}, which enjoyed a major
7658 The @code{kill-region} function in Emacs 19 is complex, but does not
7659 use code that is important at this time. We will skip it.
7661 The @code{kill-region} function in Emacs 22 is easier to read than the
7662 same function in Emacs 19 and introduces a very important concept,
7663 that of error handling. We will walk through the function.
7665 But first, let us look at the interactive @code{zap-to-char} function.
7668 * Complete zap-to-char:: The complete implementation.
7669 * zap-to-char interactive:: A three part interactive expression.
7670 * zap-to-char body:: A short overview.
7671 * search-forward:: How to search for a string.
7672 * progn:: The @code{progn} special form.
7673 * Summing up zap-to-char:: Using @code{point} and @code{search-forward}.
7677 @node Complete zap-to-char
7678 @unnumberedsubsec The Complete @code{zap-to-char} Implementation
7681 The @code{zap-to-char} function removes the text in the region between
7682 the location of the cursor (i.e., of point) up to and including the
7683 next occurrence of a specified character. The text that
7684 @code{zap-to-char} removes is put in the kill ring; and it can be
7685 retrieved from the kill ring by typing @kbd{C-y} (@code{yank}). If
7686 the command is given an argument, it removes text through that number
7687 of occurrences. Thus, if the cursor were at the beginning of this
7688 sentence and the character were @samp{s}, @samp{Thus} would be
7689 removed. If the argument were two, @samp{Thus, if the curs} would be
7690 removed, up to and including the @samp{s} in @samp{cursor}.
7692 If the specified character is not found, @code{zap-to-char} will say
7693 ``Search failed'', tell you the character you typed, and not remove
7696 In order to determine how much text to remove, @code{zap-to-char} uses
7697 a search function. Searches are used extensively in code that
7698 manipulates text, and we will focus attention on them as well as on the
7702 @c GNU Emacs version 19
7703 (defun zap-to-char (arg char) ; version 19 implementation
7704 "Kill up to and including ARG'th occurrence of CHAR.
7705 Goes backward if ARG is negative; error if CHAR not found."
7706 (interactive "*p\ncZap to char: ")
7707 (kill-region (point)
7710 (char-to-string char) nil nil arg)
7715 Here is the complete text of the version 22 implementation of the function:
7720 (defun zap-to-char (arg char)
7721 "Kill up to and including ARG'th occurrence of CHAR.
7722 Case is ignored if `case-fold-search' is non-nil in the current buffer.
7723 Goes backward if ARG is negative; error if CHAR not found."
7724 (interactive "p\ncZap to char: ")
7725 (if (char-table-p translation-table-for-input)
7726 (setq char (or (aref translation-table-for-input char) char)))
7727 (kill-region (point) (progn
7728 (search-forward (char-to-string char)
7734 The documentation is thorough. You do need to know the jargon meaning
7737 @node zap-to-char interactive
7738 @subsection The @code{interactive} Expression
7741 The interactive expression in the @code{zap-to-char} command looks like
7745 (interactive "p\ncZap to char: ")
7748 The part within quotation marks, @code{"p\ncZap to char:@: "}, specifies
7749 two different things. First, and most simply, is the @samp{p}.
7750 This part is separated from the next part by a newline, @samp{\n}.
7751 The @samp{p} means that the first argument to the function will be
7752 passed the value of a `processed prefix'. The prefix argument is
7753 passed by typing @kbd{C-u} and a number, or @kbd{M-} and a number. If
7754 the function is called interactively without a prefix, 1 is passed to
7757 The second part of @code{"p\ncZap to char:@: "} is
7758 @samp{cZap to char:@: }. In this part, the lower case @samp{c}
7759 indicates that @code{interactive} expects a prompt and that the
7760 argument will be a character. The prompt follows the @samp{c} and is
7761 the string @samp{Zap to char:@: } (with a space after the colon to
7764 What all this does is prepare the arguments to @code{zap-to-char} so they
7765 are of the right type, and give the user a prompt.
7767 In a read-only buffer, the @code{zap-to-char} function copies the text
7768 to the kill ring, but does not remove it. The echo area displays a
7769 message saying that the buffer is read-only. Also, the terminal may
7770 beep or blink at you.
7772 @node zap-to-char body
7773 @subsection The Body of @code{zap-to-char}
7775 The body of the @code{zap-to-char} function contains the code that
7776 kills (that is, removes) the text in the region from the current
7777 position of the cursor up to and including the specified character.
7779 The first part of the code looks like this:
7782 (if (char-table-p translation-table-for-input)
7783 (setq char (or (aref translation-table-for-input char) char)))
7784 (kill-region (point) (progn
7785 (search-forward (char-to-string char) nil nil arg)
7790 @code{char-table-p} is an hitherto unseen function. It determines
7791 whether its argument is a character table. When it is, it sets the
7792 character passed to @code{zap-to-char} to one of them, if that
7793 character exists, or to the character itself. (This becomes important
7794 for certain characters in non-European languages. The @code{aref}
7795 function extracts an element from an array. It is an array-specific
7796 function that is not described in this document. @xref{Arrays, ,
7797 Arrays, elisp, The GNU Emacs Lisp Reference Manual}.)
7800 @code{(point)} is the current position of the cursor.
7802 The next part of the code is an expression using @code{progn}. The body
7803 of the @code{progn} consists of calls to @code{search-forward} and
7806 It is easier to understand how @code{progn} works after learning about
7807 @code{search-forward}, so we will look at @code{search-forward} and
7808 then at @code{progn}.
7810 @node search-forward
7811 @subsection The @code{search-forward} Function
7812 @findex search-forward
7814 The @code{search-forward} function is used to locate the
7815 zapped-for-character in @code{zap-to-char}. If the search is
7816 successful, @code{search-forward} leaves point immediately after the
7817 last character in the target string. (In @code{zap-to-char}, the
7818 target string is just one character long. @code{zap-to-char} uses the
7819 function @code{char-to-string} to ensure that the computer treats that
7820 character as a string.) If the search is backwards,
7821 @code{search-forward} leaves point just before the first character in
7822 the target. Also, @code{search-forward} returns @code{t} for true.
7823 (Moving point is therefore a `side effect'.)
7826 In @code{zap-to-char}, the @code{search-forward} function looks like this:
7829 (search-forward (char-to-string char) nil nil arg)
7832 The @code{search-forward} function takes four arguments:
7836 The first argument is the target, what is searched for. This must be a
7837 string, such as @samp{"z"}.
7839 As it happens, the argument passed to @code{zap-to-char} is a single
7840 character. Because of the way computers are built, the Lisp
7841 interpreter may treat a single character as being different from a
7842 string of characters. Inside the computer, a single character has a
7843 different electronic format than a string of one character. (A single
7844 character can often be recorded in the computer using exactly one
7845 byte; but a string may be longer, and the computer needs to be ready
7846 for this.) Since the @code{search-forward} function searches for a
7847 string, the character that the @code{zap-to-char} function receives as
7848 its argument must be converted inside the computer from one format to
7849 the other; otherwise the @code{search-forward} function will fail.
7850 The @code{char-to-string} function is used to make this conversion.
7853 The second argument bounds the search; it is specified as a position in
7854 the buffer. In this case, the search can go to the end of the buffer,
7855 so no bound is set and the second argument is @code{nil}.
7858 The third argument tells the function what it should do if the search
7859 fails---it can signal an error (and print a message) or it can return
7860 @code{nil}. A @code{nil} as the third argument causes the function to
7861 signal an error when the search fails.
7864 The fourth argument to @code{search-forward} is the repeat count---how
7865 many occurrences of the string to look for. This argument is optional
7866 and if the function is called without a repeat count, this argument is
7867 passed the value 1. If this argument is negative, the search goes
7872 In template form, a @code{search-forward} expression looks like this:
7876 (search-forward "@var{target-string}"
7877 @var{limit-of-search}
7878 @var{what-to-do-if-search-fails}
7883 We will look at @code{progn} next.
7886 @subsection The @code{progn} Special Form
7889 @code{progn} is a special form that causes each of its arguments to be
7890 evaluated in sequence and then returns the value of the last one. The
7891 preceding expressions are evaluated only for the side effects they
7892 perform. The values produced by them are discarded.
7895 The template for a @code{progn} expression is very simple:
7904 In @code{zap-to-char}, the @code{progn} expression has to do two things:
7905 put point in exactly the right position; and return the location of
7906 point so that @code{kill-region} will know how far to kill to.
7908 The first argument to the @code{progn} is @code{search-forward}. When
7909 @code{search-forward} finds the string, the function leaves point
7910 immediately after the last character in the target string. (In this
7911 case the target string is just one character long.) If the search is
7912 backwards, @code{search-forward} leaves point just before the first
7913 character in the target. The movement of point is a side effect.
7915 The second and last argument to @code{progn} is the expression
7916 @code{(point)}. This expression returns the value of point, which in
7917 this case will be the location to which it has been moved by
7918 @code{search-forward}. (In the source, a line that tells the function
7919 to go to the previous character, if it is going forward, was commented
7920 out in 1999; I don't remember whether that feature or mis-feature was
7921 ever a part of the distributed source.) The value of @code{point} is
7922 returned by the @code{progn} expression and is passed to
7923 @code{kill-region} as @code{kill-region}'s second argument.
7925 @node Summing up zap-to-char
7926 @subsection Summing up @code{zap-to-char}
7928 Now that we have seen how @code{search-forward} and @code{progn} work,
7929 we can see how the @code{zap-to-char} function works as a whole.
7931 The first argument to @code{kill-region} is the position of the cursor
7932 when the @code{zap-to-char} command is given---the value of point at
7933 that time. Within the @code{progn}, the search function then moves
7934 point to just after the zapped-to-character and @code{point} returns the
7935 value of this location. The @code{kill-region} function puts together
7936 these two values of point, the first one as the beginning of the region
7937 and the second one as the end of the region, and removes the region.
7939 The @code{progn} special form is necessary because the
7940 @code{kill-region} command takes two arguments; and it would fail if
7941 @code{search-forward} and @code{point} expressions were written in
7942 sequence as two additional arguments. The @code{progn} expression is
7943 a single argument to @code{kill-region} and returns the one value that
7944 @code{kill-region} needs for its second argument.
7947 @section @code{kill-region}
7950 The @code{zap-to-char} function uses the @code{kill-region} function.
7951 This function clips text from a region and copies that text to
7952 the kill ring, from which it may be retrieved.
7957 (defun kill-region (beg end &optional yank-handler)
7958 "Kill (\"cut\") text between point and mark.
7959 This deletes the text from the buffer and saves it in the kill ring.
7960 The command \\[yank] can retrieve it from there.
7961 \(If you want to kill and then yank immediately, use \\[kill-ring-save].)
7963 If you want to append the killed region to the last killed text,
7964 use \\[append-next-kill] before \\[kill-region].
7966 If the buffer is read-only, Emacs will beep and refrain from deleting
7967 the text, but put the text in the kill ring anyway. This means that
7968 you can use the killing commands to copy text from a read-only buffer.
7970 This is the primitive for programs to kill text (as opposed to deleting it).
7971 Supply two arguments, character positions indicating the stretch of text
7973 Any command that calls this function is a \"kill command\".
7974 If the previous command was also a kill command,
7975 the text killed this time appends to the text killed last time
7976 to make one entry in the kill ring.
7978 In Lisp code, optional third arg YANK-HANDLER, if non-nil,
7979 specifies the yank-handler text property to be set on the killed
7980 text. See `insert-for-yank'."
7981 ;; Pass point first, then mark, because the order matters
7982 ;; when calling kill-append.
7983 (interactive (list (point) (mark)))
7984 (unless (and beg end)
7985 (error "The mark is not set now, so there is no region"))
7987 (let ((string (filter-buffer-substring beg end t)))
7988 (when string ;STRING is nil if BEG = END
7989 ;; Add that string to the kill ring, one way or another.
7990 (if (eq last-command 'kill-region)
7991 (kill-append string (< end beg) yank-handler)
7992 (kill-new string nil yank-handler)))
7993 (when (or string (eq last-command 'kill-region))
7994 (setq this-command 'kill-region))
7996 ((buffer-read-only text-read-only)
7997 ;; The code above failed because the buffer, or some of the characters
7998 ;; in the region, are read-only.
7999 ;; We should beep, in case the user just isn't aware of this.
8000 ;; However, there's no harm in putting
8001 ;; the region's text in the kill ring, anyway.
8002 (copy-region-as-kill beg end)
8003 ;; Set this-command now, so it will be set even if we get an error.
8004 (setq this-command 'kill-region)
8005 ;; This should barf, if appropriate, and give us the correct error.
8006 (if kill-read-only-ok
8007 (progn (message "Read only text copied to kill ring") nil)
8008 ;; Signal an error if the buffer is read-only.
8009 (barf-if-buffer-read-only)
8010 ;; If the buffer isn't read-only, the text is.
8011 (signal 'text-read-only (list (current-buffer)))))))
8014 The Emacs 22 version of that function uses @code{condition-case} and
8015 @code{copy-region-as-kill}, both of which we will explain.
8016 @code{condition-case} is an important special form.
8018 In essence, the @code{kill-region} function calls
8019 @code{condition-case}, which takes three arguments. In this function,
8020 the first argument does nothing. The second argument contains the
8021 code that does the work when all goes well. The third argument
8022 contains the code that is called in the event of an error.
8025 * Complete kill-region:: The function definition.
8026 * condition-case:: Dealing with a problem.
8031 @node Complete kill-region
8032 @unnumberedsubsec The Complete @code{kill-region} Definition
8036 We will go through the @code{condition-case} code in a moment. First,
8037 let us look at the definition of @code{kill-region}, with comments
8043 (defun kill-region (beg end)
8044 "Kill (\"cut\") text between point and mark.
8045 This deletes the text from the buffer and saves it in the kill ring.
8046 The command \\[yank] can retrieve it from there. @dots{} "
8050 ;; @bullet{} Since order matters, pass point first.
8051 (interactive (list (point) (mark)))
8052 ;; @bullet{} And tell us if we cannot cut the text.
8053 ;; `unless' is an `if' without a then-part.
8054 (unless (and beg end)
8055 (error "The mark is not set now, so there is no region"))
8059 ;; @bullet{} `condition-case' takes three arguments.
8060 ;; If the first argument is nil, as it is here,
8061 ;; information about the error signal is not
8062 ;; stored for use by another function.
8067 ;; @bullet{} The second argument to `condition-case' tells the
8068 ;; Lisp interpreter what to do when all goes well.
8072 ;; It starts with a `let' function that extracts the string
8073 ;; and tests whether it exists. If so (that is what the
8074 ;; `when' checks), it calls an `if' function that determines
8075 ;; whether the previous command was another call to
8076 ;; `kill-region'; if it was, then the new text is appended to
8077 ;; the previous text; if not, then a different function,
8078 ;; `kill-new', is called.
8082 ;; The `kill-append' function concatenates the new string and
8083 ;; the old. The `kill-new' function inserts text into a new
8084 ;; item in the kill ring.
8088 ;; `when' is an `if' without an else-part. The second `when'
8089 ;; again checks whether the current string exists; in
8090 ;; addition, it checks whether the previous command was
8091 ;; another call to `kill-region'. If one or the other
8092 ;; condition is true, then it sets the current command to
8093 ;; be `kill-region'.
8096 (let ((string (filter-buffer-substring beg end t)))
8097 (when string ;STRING is nil if BEG = END
8098 ;; Add that string to the kill ring, one way or another.
8099 (if (eq last-command 'kill-region)
8102 ;; @minus{} `yank-handler' is an optional argument to
8103 ;; `kill-region' that tells the `kill-append' and
8104 ;; `kill-new' functions how deal with properties
8105 ;; added to the text, such as `bold' or `italics'.
8106 (kill-append string (< end beg) yank-handler)
8107 (kill-new string nil yank-handler)))
8108 (when (or string (eq last-command 'kill-region))
8109 (setq this-command 'kill-region))
8114 ;; @bullet{} The third argument to `condition-case' tells the interpreter
8115 ;; what to do with an error.
8118 ;; The third argument has a conditions part and a body part.
8119 ;; If the conditions are met (in this case,
8120 ;; if text or buffer are read-only)
8121 ;; then the body is executed.
8124 ;; The first part of the third argument is the following:
8125 ((buffer-read-only text-read-only) ;; the if-part
8126 ;; @dots{} the then-part
8127 (copy-region-as-kill beg end)
8130 ;; Next, also as part of the then-part, set this-command, so
8131 ;; it will be set in an error
8132 (setq this-command 'kill-region)
8133 ;; Finally, in the then-part, send a message if you may copy
8134 ;; the text to the kill ring without signaling an error, but
8135 ;; don't if you may not.
8138 (if kill-read-only-ok
8139 (progn (message "Read only text copied to kill ring") nil)
8140 (barf-if-buffer-read-only)
8141 ;; If the buffer isn't read-only, the text is.
8142 (signal 'text-read-only (list (current-buffer)))))
8150 (defun kill-region (beg end)
8151 "Kill between point and mark.
8152 The text is deleted but saved in the kill ring."
8157 ;; 1. `condition-case' takes three arguments.
8158 ;; If the first argument is nil, as it is here,
8159 ;; information about the error signal is not
8160 ;; stored for use by another function.
8165 ;; 2. The second argument to `condition-case'
8166 ;; tells the Lisp interpreter what to do when all goes well.
8170 ;; The `delete-and-extract-region' function usually does the
8171 ;; work. If the beginning and ending of the region are both
8172 ;; the same, then the variable `string' will be empty, or nil
8173 (let ((string (delete-and-extract-region beg end)))
8177 ;; `when' is an `if' clause that cannot take an `else-part'.
8178 ;; Emacs normally sets the value of `last-command' to the
8179 ;; previous command.
8182 ;; `kill-append' concatenates the new string and the old.
8183 ;; `kill-new' inserts text into a new item in the kill ring.
8185 (if (eq last-command 'kill-region)
8186 ;; if true, prepend string
8187 (kill-append string (< end beg))
8189 (setq this-command 'kill-region))
8193 ;; 3. The third argument to `condition-case' tells the interpreter
8194 ;; what to do with an error.
8197 ;; The third argument has a conditions part and a body part.
8198 ;; If the conditions are met (in this case,
8199 ;; if text or buffer are read-only)
8200 ;; then the body is executed.
8203 ((buffer-read-only text-read-only) ;; this is the if-part
8205 (copy-region-as-kill beg end)
8208 (if kill-read-only-ok ;; usually this variable is nil
8209 (message "Read only text copied to kill ring")
8210 ;; or else, signal an error if the buffer is read-only;
8211 (barf-if-buffer-read-only)
8212 ;; and, in any case, signal that the text is read-only.
8213 (signal 'text-read-only (list (current-buffer)))))))
8218 @node condition-case
8219 @subsection @code{condition-case}
8220 @findex condition-case
8222 As we have seen earlier (@pxref{Making Errors, , Generate an Error
8223 Message}), when the Emacs Lisp interpreter has trouble evaluating an
8224 expression, it provides you with help; in the jargon, this is called
8225 ``signaling an error''. Usually, the computer stops the program and
8226 shows you a message.
8228 However, some programs undertake complicated actions. They should not
8229 simply stop on an error. In the @code{kill-region} function, the most
8230 likely error is that you will try to kill text that is read-only and
8231 cannot be removed. So the @code{kill-region} function contains code
8232 to handle this circumstance. This code, which makes up the body of
8233 the @code{kill-region} function, is inside of a @code{condition-case}
8237 The template for @code{condition-case} looks like this:
8244 @var{error-handler}@dots{})
8248 The second argument, @var{bodyform}, is straightforward. The
8249 @code{condition-case} special form causes the Lisp interpreter to
8250 evaluate the code in @var{bodyform}. If no error occurs, the special
8251 form returns the code's value and produces the side-effects, if any.
8253 In short, the @var{bodyform} part of a @code{condition-case}
8254 expression determines what should happen when everything works
8257 However, if an error occurs, among its other actions, the function
8258 generating the error signal will define one or more error condition
8261 An error handler is the third argument to @code{condition case}.
8262 An error handler has two parts, a @var{condition-name} and a
8263 @var{body}. If the @var{condition-name} part of an error handler
8264 matches a condition name generated by an error, then the @var{body}
8265 part of the error handler is run.
8267 As you will expect, the @var{condition-name} part of an error handler
8268 may be either a single condition name or a list of condition names.
8270 Also, a complete @code{condition-case} expression may contain more
8271 than one error handler. When an error occurs, the first applicable
8274 Lastly, the first argument to the @code{condition-case} expression,
8275 the @var{var} argument, is sometimes bound to a variable that
8276 contains information about the error. However, if that argument is
8277 nil, as is the case in @code{kill-region}, that information is
8281 In brief, in the @code{kill-region} function, the code
8282 @code{condition-case} works like this:
8286 @var{If no errors}, @var{run only this code}
8287 @var{but}, @var{if errors}, @var{run this other code}.
8294 copy-region-as-kill is short, 12 lines, and uses
8295 filter-buffer-substring, which is longer, 39 lines
8296 and has delete-and-extract-region in it.
8297 delete-and-extract-region is written in C.
8299 see Initializing a Variable with @code{defvar}
8301 Initializing a Variable with @code{defvar} includes line 8350
8305 @subsection Lisp macro
8309 The part of the @code{condition-case} expression that is evaluated in
8310 the expectation that all goes well has a @code{when}. The code uses
8311 @code{when} to determine whether the @code{string} variable points to
8314 A @code{when} expression is simply a programmers' convenience. It is
8315 an @code{if} without the possibility of an else clause. In your mind,
8316 you can replace @code{when} with @code{if} and understand what goes
8317 on. That is what the Lisp interpreter does.
8319 Technically speaking, @code{when} is a Lisp macro. A Lisp @dfn{macro}
8320 enables you to define new control constructs and other language
8321 features. It tells the interpreter how to compute another Lisp
8322 expression which will in turn compute the value. In this case, the
8323 `other expression' is an @code{if} expression.
8325 The @code{kill-region} function definition also has an @code{unless}
8326 macro; it is the converse of @code{when}. The @code{unless} macro is
8327 an @code{if} without a then clause
8329 For more about Lisp macros, see @ref{Macros, , Macros, elisp, The GNU
8330 Emacs Lisp Reference Manual}. The C programming language also
8331 provides macros. These are different, but also useful.
8334 We will briefly look at C macros in
8335 @ref{Digression into C}.
8339 Regarding the @code{when} macro, in the @code{condition-case}
8340 expression, when the string has content, then another conditional
8341 expression is executed. This is an @code{if} with both a then-part
8346 (if (eq last-command 'kill-region)
8347 (kill-append string (< end beg) yank-handler)
8348 (kill-new string nil yank-handler))
8352 The then-part is evaluated if the previous command was another call to
8353 @code{kill-region}; if not, the else-part is evaluated.
8355 @code{yank-handler} is an optional argument to @code{kill-region} that
8356 tells the @code{kill-append} and @code{kill-new} functions how deal
8357 with properties added to the text, such as `bold' or `italics'.
8359 @code{last-command} is a variable that comes with Emacs that we have
8360 not seen before. Normally, whenever a function is executed, Emacs
8361 sets the value of @code{last-command} to the previous command.
8364 In this segment of the definition, the @code{if} expression checks
8365 whether the previous command was @code{kill-region}. If it was,
8368 (kill-append string (< end beg) yank-handler)
8372 concatenates a copy of the newly clipped text to the just previously
8373 clipped text in the kill ring.
8375 @node copy-region-as-kill
8376 @section @code{copy-region-as-kill}
8377 @findex copy-region-as-kill
8380 The @code{copy-region-as-kill} function copies a region of text from a
8381 buffer and (via either @code{kill-append} or @code{kill-new}) saves it
8382 in the @code{kill-ring}.
8384 If you call @code{copy-region-as-kill} immediately after a
8385 @code{kill-region} command, Emacs appends the newly copied text to the
8386 previously copied text. This means that if you yank back the text, you
8387 get it all, from both this and the previous operation. On the other
8388 hand, if some other command precedes the @code{copy-region-as-kill},
8389 the function copies the text into a separate entry in the kill ring.
8392 * Complete copy-region-as-kill:: The complete function definition.
8393 * copy-region-as-kill body:: The body of @code{copy-region-as-kill}.
8397 @node Complete copy-region-as-kill
8398 @unnumberedsubsec The complete @code{copy-region-as-kill} function definition
8402 Here is the complete text of the version 22 @code{copy-region-as-kill}
8407 (defun copy-region-as-kill (beg end)
8408 "Save the region as if killed, but don't kill it.
8409 In Transient Mark mode, deactivate the mark.
8410 If `interprogram-cut-function' is non-nil, also save the text for a window
8411 system cut and paste."
8415 (if (eq last-command 'kill-region)
8416 (kill-append (filter-buffer-substring beg end) (< end beg))
8417 (kill-new (filter-buffer-substring beg end)))
8420 (if transient-mark-mode
8421 (setq deactivate-mark t))
8427 As usual, this function can be divided into its component parts:
8431 (defun copy-region-as-kill (@var{argument-list})
8432 "@var{documentation}@dots{}"
8438 The arguments are @code{beg} and @code{end} and the function is
8439 interactive with @code{"r"}, so the two arguments must refer to the
8440 beginning and end of the region. If you have been reading though this
8441 document from the beginning, understanding these parts of a function is
8442 almost becoming routine.
8444 The documentation is somewhat confusing unless you remember that the
8445 word `kill' has a meaning different from usual. The `Transient Mark'
8446 and @code{interprogram-cut-function} comments explain certain
8449 After you once set a mark, a buffer always contains a region. If you
8450 wish, you can use Transient Mark mode to highlight the region
8451 temporarily. (No one wants to highlight the region all the time, so
8452 Transient Mark mode highlights it only at appropriate times. Many
8453 people turn off Transient Mark mode, so the region is never
8456 Also, a windowing system allows you to copy, cut, and paste among
8457 different programs. In the X windowing system, for example, the
8458 @code{interprogram-cut-function} function is @code{x-select-text},
8459 which works with the windowing system's equivalent of the Emacs kill
8462 The body of the @code{copy-region-as-kill} function starts with an
8463 @code{if} clause. What this clause does is distinguish between two
8464 different situations: whether or not this command is executed
8465 immediately after a previous @code{kill-region} command. In the first
8466 case, the new region is appended to the previously copied text.
8467 Otherwise, it is inserted into the beginning of the kill ring as a
8468 separate piece of text from the previous piece.
8470 The last two lines of the function prevent the region from lighting up
8471 if Transient Mark mode is turned on.
8473 The body of @code{copy-region-as-kill} merits discussion in detail.
8475 @node copy-region-as-kill body
8476 @subsection The Body of @code{copy-region-as-kill}
8478 The @code{copy-region-as-kill} function works in much the same way as
8479 the @code{kill-region} function. Both are written so that two or more
8480 kills in a row combine their text into a single entry. If you yank
8481 back the text from the kill ring, you get it all in one piece.
8482 Moreover, kills that kill forward from the current position of the
8483 cursor are added to the end of the previously copied text and commands
8484 that copy text backwards add it to the beginning of the previously
8485 copied text. This way, the words in the text stay in the proper
8488 Like @code{kill-region}, the @code{copy-region-as-kill} function makes
8489 use of the @code{last-command} variable that keeps track of the
8490 previous Emacs command.
8493 * last-command & this-command::
8494 * kill-append function::
8495 * kill-new function::
8499 @node last-command & this-command
8500 @unnumberedsubsubsec @code{last-command} and @code{this-command}
8503 Normally, whenever a function is executed, Emacs sets the value of
8504 @code{this-command} to the function being executed (which in this case
8505 would be @code{copy-region-as-kill}). At the same time, Emacs sets
8506 the value of @code{last-command} to the previous value of
8507 @code{this-command}.
8509 In the first part of the body of the @code{copy-region-as-kill}
8510 function, an @code{if} expression determines whether the value of
8511 @code{last-command} is @code{kill-region}. If so, the then-part of
8512 the @code{if} expression is evaluated; it uses the @code{kill-append}
8513 function to concatenate the text copied at this call to the function
8514 with the text already in the first element (the @sc{car}) of the kill
8515 ring. On the other hand, if the value of @code{last-command} is not
8516 @code{kill-region}, then the @code{copy-region-as-kill} function
8517 attaches a new element to the kill ring using the @code{kill-new}
8521 The @code{if} expression reads as follows; it uses @code{eq}:
8525 (if (eq last-command 'kill-region)
8527 (kill-append (filter-buffer-substring beg end) (< end beg))
8529 (kill-new (filter-buffer-substring beg end)))
8533 @findex filter-buffer-substring
8534 (The @code{filter-buffer-substring} function returns a filtered
8535 substring of the buffer, if any. Optionally---the arguments are not
8536 here, so neither is done---the function may delete the initial text or
8537 return the text without its properties; this function is a replacement
8538 for the older @code{buffer-substring} function, which came before text
8539 properties were implemented.)
8541 @findex eq @r{(example of use)}
8543 The @code{eq} function tests whether its first argument is the same Lisp
8544 object as its second argument. The @code{eq} function is similar to the
8545 @code{equal} function in that it is used to test for equality, but
8546 differs in that it determines whether two representations are actually
8547 the same object inside the computer, but with different names.
8548 @code{equal} determines whether the structure and contents of two
8549 expressions are the same.
8551 If the previous command was @code{kill-region}, then the Emacs Lisp
8552 interpreter calls the @code{kill-append} function
8554 @node kill-append function
8555 @unnumberedsubsubsec The @code{kill-append} function
8559 The @code{kill-append} function looks like this:
8564 (defun kill-append (string before-p &optional yank-handler)
8565 "Append STRING to the end of the latest kill in the kill ring.
8566 If BEFORE-P is non-nil, prepend STRING to the kill.
8568 (let* ((cur (car kill-ring)))
8569 (kill-new (if before-p (concat string cur) (concat cur string))
8570 (or (= (length cur) 0)
8572 (get-text-property 0 'yank-handler cur)))
8579 (defun kill-append (string before-p)
8580 "Append STRING to the end of the latest kill in the kill ring.
8581 If BEFORE-P is non-nil, prepend STRING to the kill.
8582 If `interprogram-cut-function' is set, pass the resulting kill to
8584 (kill-new (if before-p
8585 (concat string (car kill-ring))
8586 (concat (car kill-ring) string))
8591 The @code{kill-append} function is fairly straightforward. It uses
8592 the @code{kill-new} function, which we will discuss in more detail in
8595 (Also, the function provides an optional argument called
8596 @code{yank-handler}; when invoked, this argument tells the function
8597 how to deal with properties added to the text, such as `bold' or
8600 @c !!! bug in GNU Emacs 22 version of kill-append ?
8601 It has a @code{let*} function to set the value of the first element of
8602 the kill ring to @code{cur}. (I do not know why the function does not
8603 use @code{let} instead; only one value is set in the expression.
8604 Perhaps this is a bug that produces no problems?)
8606 Consider the conditional that is one of the two arguments to
8607 @code{kill-new}. It uses @code{concat} to concatenate the new text to
8608 the @sc{car} of the kill ring. Whether it prepends or appends the
8609 text depends on the results of an @code{if} expression:
8613 (if before-p ; @r{if-part}
8614 (concat string cur) ; @r{then-part}
8615 (concat cur string)) ; @r{else-part}
8620 If the region being killed is before the region that was killed in the
8621 last command, then it should be prepended before the material that was
8622 saved in the previous kill; and conversely, if the killed text follows
8623 what was just killed, it should be appended after the previous text.
8624 The @code{if} expression depends on the predicate @code{before-p} to
8625 decide whether the newly saved text should be put before or after the
8626 previously saved text.
8628 The symbol @code{before-p} is the name of one of the arguments to
8629 @code{kill-append}. When the @code{kill-append} function is
8630 evaluated, it is bound to the value returned by evaluating the actual
8631 argument. In this case, this is the expression @code{(< end beg)}.
8632 This expression does not directly determine whether the killed text in
8633 this command is located before or after the kill text of the last
8634 command; what it does is determine whether the value of the variable
8635 @code{end} is less than the value of the variable @code{beg}. If it
8636 is, it means that the user is most likely heading towards the
8637 beginning of the buffer. Also, the result of evaluating the predicate
8638 expression, @code{(< end beg)}, will be true and the text will be
8639 prepended before the previous text. On the other hand, if the value of
8640 the variable @code{end} is greater than the value of the variable
8641 @code{beg}, the text will be appended after the previous text.
8644 When the newly saved text will be prepended, then the string with the new
8645 text will be concatenated before the old text:
8653 But if the text will be appended, it will be concatenated
8657 (concat cur string))
8660 To understand how this works, we first need to review the
8661 @code{concat} function. The @code{concat} function links together or
8662 unites two strings of text. The result is a string. For example:
8666 (concat "abc" "def")
8672 (car '("first element" "second element")))
8673 @result{} "new first element"
8676 '("first element" "second element")) " modified")
8677 @result{} "first element modified"
8681 We can now make sense of @code{kill-append}: it modifies the contents
8682 of the kill ring. The kill ring is a list, each element of which is
8683 saved text. The @code{kill-append} function uses the @code{kill-new}
8684 function which in turn uses the @code{setcar} function.
8686 @node kill-new function
8687 @unnumberedsubsubsec The @code{kill-new} function
8690 @c in GNU Emacs 22, additional documentation to kill-new:
8692 Optional third arguments YANK-HANDLER controls how the STRING is later
8693 inserted into a buffer; see `insert-for-yank' for details.
8694 When a yank handler is specified, STRING must be non-empty (the yank
8695 handler, if non-nil, is stored as a `yank-handler' text property on STRING).
8697 When the yank handler has a non-nil PARAM element, the original STRING
8698 argument is not used by `insert-for-yank'. However, since Lisp code
8699 may access and use elements from the kill ring directly, the STRING
8700 argument should still be a \"useful\" string for such uses."
8703 The @code{kill-new} function looks like this:
8707 (defun kill-new (string &optional replace yank-handler)
8708 "Make STRING the latest kill in the kill ring.
8709 Set `kill-ring-yank-pointer' to point to it.
8711 If `interprogram-cut-function' is non-nil, apply it to STRING.
8712 Optional second argument REPLACE non-nil means that STRING will replace
8713 the front of the kill ring, rather than being added to the list.
8717 (if (> (length string) 0)
8719 (put-text-property 0 (length string)
8720 'yank-handler yank-handler string))
8722 (signal 'args-out-of-range
8723 (list string "yank-handler specified for empty string"))))
8726 (if (fboundp 'menu-bar-update-yank-menu)
8727 (menu-bar-update-yank-menu string (and replace (car kill-ring))))
8730 (if (and replace kill-ring)
8731 (setcar kill-ring string)
8732 (push string kill-ring)
8733 (if (> (length kill-ring) kill-ring-max)
8734 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)))
8737 (setq kill-ring-yank-pointer kill-ring)
8738 (if interprogram-cut-function
8739 (funcall interprogram-cut-function string (not replace))))
8744 (defun kill-new (string &optional replace)
8745 "Make STRING the latest kill in the kill ring.
8746 Set the kill-ring-yank pointer to point to it.
8747 If `interprogram-cut-function' is non-nil, apply it to STRING.
8748 Optional second argument REPLACE non-nil means that STRING will replace
8749 the front of the kill ring, rather than being added to the list."
8750 (and (fboundp 'menu-bar-update-yank-menu)
8751 (menu-bar-update-yank-menu string (and replace (car kill-ring))))
8752 (if (and replace kill-ring)
8753 (setcar kill-ring string)
8754 (setq kill-ring (cons string kill-ring))
8755 (if (> (length kill-ring) kill-ring-max)
8756 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)))
8757 (setq kill-ring-yank-pointer kill-ring)
8758 (if interprogram-cut-function
8759 (funcall interprogram-cut-function string (not replace))))
8762 (Notice that the function is not interactive.)
8764 As usual, we can look at this function in parts.
8766 The function definition has an optional @code{yank-handler} argument,
8767 which when invoked tells the function how to deal with properties
8768 added to the text, such as `bold' or `italics'. We will skip that.
8771 The first line of the documentation makes sense:
8774 Make STRING the latest kill in the kill ring.
8778 Let's skip over the rest of the documentation for the moment.
8781 Also, let's skip over the initial @code{if} expression and those lines
8782 of code involving @code{menu-bar-update-yank-menu}. We will explain
8786 The critical lines are these:
8790 (if (and replace kill-ring)
8792 (setcar kill-ring string)
8796 (push string kill-ring)
8799 (setq kill-ring (cons string kill-ring))
8800 (if (> (length kill-ring) kill-ring-max)
8801 ;; @r{avoid overly long kill ring}
8802 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)))
8805 (setq kill-ring-yank-pointer kill-ring)
8806 (if interprogram-cut-function
8807 (funcall interprogram-cut-function string (not replace))))
8811 The conditional test is @w{@code{(and replace kill-ring)}}.
8812 This will be true when two conditions are met: the kill ring has
8813 something in it, and the @code{replace} variable is true.
8816 When the @code{kill-append} function sets @code{replace} to be true
8817 and when the kill ring has at least one item in it, the @code{setcar}
8818 expression is executed:
8821 (setcar kill-ring string)
8824 The @code{setcar} function actually changes the first element of the
8825 @code{kill-ring} list to the value of @code{string}. It replaces the
8829 On the other hand, if the kill ring is empty, or replace is false, the
8830 else-part of the condition is executed:
8833 (push string kill-ring)
8838 @code{push} puts its first argument onto the second. It is similar to
8842 (setq kill-ring (cons string kill-ring))
8850 (add-to-list kill-ring string)
8854 When it is false, the expression first constructs a new version of the
8855 kill ring by prepending @code{string} to the existing kill ring as a
8856 new element (that is what the @code{push} does). Then it executes a
8857 second @code{if} clause. This second @code{if} clause keeps the kill
8858 ring from growing too long.
8860 Let's look at these two expressions in order.
8862 The @code{push} line of the else-part sets the new value of the kill
8863 ring to what results from adding the string being killed to the old
8866 We can see how this works with an example.
8872 (setq example-list '("here is a clause" "another clause"))
8877 After evaluating this expression with @kbd{C-x C-e}, you can evaluate
8878 @code{example-list} and see what it returns:
8883 @result{} ("here is a clause" "another clause")
8889 Now, we can add a new element on to this list by evaluating the
8890 following expression:
8891 @findex push, @r{example}
8894 (push "a third clause" example-list)
8899 When we evaluate @code{example-list}, we find its value is:
8904 @result{} ("a third clause" "here is a clause" "another clause")
8909 Thus, the third clause is added to the list by @code{push}.
8912 Now for the second part of the @code{if} clause. This expression
8913 keeps the kill ring from growing too long. It looks like this:
8917 (if (> (length kill-ring) kill-ring-max)
8918 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil))
8922 The code checks whether the length of the kill ring is greater than
8923 the maximum permitted length. This is the value of
8924 @code{kill-ring-max} (which is 60, by default). If the length of the
8925 kill ring is too long, then this code sets the last element of the
8926 kill ring to @code{nil}. It does this by using two functions,
8927 @code{nthcdr} and @code{setcdr}.
8929 We looked at @code{setcdr} earlier (@pxref{setcdr, , @code{setcdr}}).
8930 It sets the @sc{cdr} of a list, just as @code{setcar} sets the
8931 @sc{car} of a list. In this case, however, @code{setcdr} will not be
8932 setting the @sc{cdr} of the whole kill ring; the @code{nthcdr}
8933 function is used to cause it to set the @sc{cdr} of the next to last
8934 element of the kill ring---this means that since the @sc{cdr} of the
8935 next to last element is the last element of the kill ring, it will set
8936 the last element of the kill ring.
8938 @findex nthcdr, @r{example}
8939 The @code{nthcdr} function works by repeatedly taking the @sc{cdr} of a
8940 list---it takes the @sc{cdr} of the @sc{cdr} of the @sc{cdr}
8941 @dots{} It does this @var{N} times and returns the results.
8942 (@xref{nthcdr, , @code{nthcdr}}.)
8944 @findex setcdr, @r{example}
8945 Thus, if we had a four element list that was supposed to be three
8946 elements long, we could set the @sc{cdr} of the next to last element
8947 to @code{nil}, and thereby shorten the list. (If you set the last
8948 element to some other value than @code{nil}, which you could do, then
8949 you would not have shortened the list. @xref{setcdr, ,
8952 You can see shortening by evaluating the following three expressions
8953 in turn. First set the value of @code{trees} to @code{(maple oak pine
8954 birch)}, then set the @sc{cdr} of its second @sc{cdr} to @code{nil}
8955 and then find the value of @code{trees}:
8959 (setq trees '(maple oak pine birch))
8960 @result{} (maple oak pine birch)
8964 (setcdr (nthcdr 2 trees) nil)
8968 @result{} (maple oak pine)
8973 (The value returned by the @code{setcdr} expression is @code{nil} since
8974 that is what the @sc{cdr} is set to.)
8976 To repeat, in @code{kill-new}, the @code{nthcdr} function takes the
8977 @sc{cdr} a number of times that is one less than the maximum permitted
8978 size of the kill ring and @code{setcdr} sets the @sc{cdr} of that
8979 element (which will be the rest of the elements in the kill ring) to
8980 @code{nil}. This prevents the kill ring from growing too long.
8983 The next to last expression in the @code{kill-new} function is
8986 (setq kill-ring-yank-pointer kill-ring)
8989 The @code{kill-ring-yank-pointer} is a global variable that is set to be
8990 the @code{kill-ring}.
8992 Even though the @code{kill-ring-yank-pointer} is called a
8993 @samp{pointer}, it is a variable just like the kill ring. However, the
8994 name has been chosen to help humans understand how the variable is used.
8997 Now, to return to an early expression in the body of the function:
9001 (if (fboundp 'menu-bar-update-yank-menu)
9002 (menu-bar-update-yank-menu string (and replace (car kill-ring))))
9007 It starts with an @code{if} expression
9009 In this case, the expression tests first to see whether
9010 @code{menu-bar-update-yank-menu} exists as a function, and if so,
9011 calls it. The @code{fboundp} function returns true if the symbol it
9012 is testing has a function definition that `is not void'. If the
9013 symbol's function definition were void, we would receive an error
9014 message, as we did when we created errors intentionally (@pxref{Making
9015 Errors, , Generate an Error Message}).
9018 The then-part contains an expression whose first element is the
9019 function @code{and}.
9022 The @code{and} special form evaluates each of its arguments until one
9023 of the arguments returns a value of @code{nil}, in which case the
9024 @code{and} expression returns @code{nil}; however, if none of the
9025 arguments returns a value of @code{nil}, the value resulting from
9026 evaluating the last argument is returned. (Since such a value is not
9027 @code{nil}, it is considered true in Emacs Lisp.) In other words, an
9028 @code{and} expression returns a true value only if all its arguments
9029 are true. (@xref{Second Buffer Related Review}.)
9031 The expression determines whether the second argument to
9032 @code{menu-bar-update-yank-menu} is true or not.
9034 ;; If we're supposed to be extending an existing string, and that
9035 ;; string really is at the front of the menu, then update it in place.
9038 @code{menu-bar-update-yank-menu} is one of the functions that make it
9039 possible to use the `Select and Paste' menu in the Edit item of a menu
9040 bar; using a mouse, you can look at the various pieces of text you
9041 have saved and select one piece to paste.
9043 The last expression in the @code{kill-new} function adds the newly
9044 copied string to whatever facility exists for copying and pasting
9045 among different programs running in a windowing system. In the X
9046 Windowing system, for example, the @code{x-select-text} function takes
9047 the string and stores it in memory operated by X@. You can paste the
9048 string in another program, such as an Xterm.
9051 The expression looks like this:
9055 (if interprogram-cut-function
9056 (funcall interprogram-cut-function string (not replace))))
9060 If an @code{interprogram-cut-function} exists, then Emacs executes
9061 @code{funcall}, which in turn calls its first argument as a function
9062 and passes the remaining arguments to it. (Incidentally, as far as I
9063 can see, this @code{if} expression could be replaced by an @code{and}
9064 expression similar to the one in the first part of the function.)
9066 We are not going to discuss windowing systems and other programs
9067 further, but merely note that this is a mechanism that enables GNU
9068 Emacs to work easily and well with other programs.
9070 This code for placing text in the kill ring, either concatenated with
9071 an existing element or as a new element, leads us to the code for
9072 bringing back text that has been cut out of the buffer---the yank
9073 commands. However, before discussing the yank commands, it is better
9074 to learn how lists are implemented in a computer. This will make
9075 clear such mysteries as the use of the term `pointer'. But before
9076 that, we will digress into C.
9079 @c is this true in Emacs 22? Does not seems to be
9081 (If the @w{@code{(< end beg))}}
9082 expression is true, @code{kill-append} prepends the string to the just
9083 previously clipped text. For a detailed discussion, see
9084 @ref{kill-append function, , The @code{kill-append} function}.)
9086 If you then yank back the text, i.e., `paste' it, you get both
9087 pieces of text at once. That way, if you delete two words in a row,
9088 and then yank them back, you get both words, in their proper order,
9089 with one yank. (The @w{@code{(< end beg))}} expression makes sure the
9092 On the other hand, if the previous command is not @code{kill-region},
9093 then the @code{kill-new} function is called, which adds the text to
9094 the kill ring as the latest item, and sets the
9095 @code{kill-ring-yank-pointer} variable to point to it.
9099 @c Evidently, changed for Emacs 22. The zap-to-char command does not
9100 @c use the delete-and-extract-region function
9102 2006 Oct 26, the Digression into C is now OK but should come after
9103 copy-region-as-kill and filter-buffer-substring
9107 copy-region-as-kill is short, 12 lines, and uses
9108 filter-buffer-substring, which is longer, 39 lines
9109 and has delete-and-extract-region in it.
9110 delete-and-extract-region is written in C.
9112 see Initializing a Variable with @code{defvar}
9115 @node Digression into C
9116 @section Digression into C
9117 @findex delete-and-extract-region
9118 @cindex C, a digression into
9119 @cindex Digression into C
9121 The @code{copy-region-as-kill} function (@pxref{copy-region-as-kill, ,
9122 @code{copy-region-as-kill}}) uses the @code{filter-buffer-substring}
9123 function, which in turn uses the @code{delete-and-extract-region}
9124 function. It removes the contents of a region and you cannot get them
9127 Unlike the other code discussed here, the
9128 @code{delete-and-extract-region} function is not written in Emacs
9129 Lisp; it is written in C and is one of the primitives of the GNU Emacs
9130 system. Since it is very simple, I will digress briefly from Lisp and
9133 @c GNU Emacs 24 in src/editfns.c
9134 @c the DEFUN for delete-and-extract-region
9137 Like many of the other Emacs primitives,
9138 @code{delete-and-extract-region} is written as an instance of a C
9139 macro, a macro being a template for code. The complete macro looks
9144 DEFUN ("delete-and-extract-region", Fdelete_and_extract_region,
9145 Sdelete_and_extract_region, 2, 2, 0,
9146 doc: /* Delete the text between START and END and return it. */)
9147 (Lisp_Object start, Lisp_Object end)
9149 validate_region (&start, &end);
9150 if (XINT (start) == XINT (end))
9151 return empty_unibyte_string;
9152 return del_range_1 (XINT (start), XINT (end), 1, 1);
9157 Without going into the details of the macro writing process, let me
9158 point out that this macro starts with the word @code{DEFUN}. The word
9159 @code{DEFUN} was chosen since the code serves the same purpose as
9160 @code{defun} does in Lisp. (The @code{DEFUN} C macro is defined in
9161 @file{emacs/src/lisp.h}.)
9163 The word @code{DEFUN} is followed by seven parts inside of
9168 The first part is the name given to the function in Lisp,
9169 @code{delete-and-extract-region}.
9172 The second part is the name of the function in C,
9173 @code{Fdelete_and_extract_region}. By convention, it starts with
9174 @samp{F}. Since C does not use hyphens in names, underscores are used
9178 The third part is the name for the C constant structure that records
9179 information on this function for internal use. It is the name of the
9180 function in C but begins with an @samp{S} instead of an @samp{F}.
9183 The fourth and fifth parts specify the minimum and maximum number of
9184 arguments the function can have. This function demands exactly 2
9188 The sixth part is nearly like the argument that follows the
9189 @code{interactive} declaration in a function written in Lisp: a letter
9190 followed, perhaps, by a prompt. The only difference from the Lisp is
9191 when the macro is called with no arguments. Then you write a @code{0}
9192 (which is a `null string'), as in this macro.
9194 If you were to specify arguments, you would place them between
9195 quotation marks. The C macro for @code{goto-char} includes
9196 @code{"NGoto char: "} in this position to indicate that the function
9197 expects a raw prefix, in this case, a numerical location in a buffer,
9198 and provides a prompt.
9201 The seventh part is a documentation string, just like the one for a
9202 function written in Emacs Lisp. This is written as a C comment. (When
9203 you build Emacs, the program @command{lib-src/make-docfile} extracts
9204 these comments and uses them to make the ``real'' documentation.)
9208 In a C macro, the formal parameters come next, with a statement of
9209 what kind of object they are, followed by what might be called the `body'
9210 of the macro. For @code{delete-and-extract-region} the `body'
9211 consists of the following four lines:
9215 validate_region (&start, &end);
9216 if (XINT (start) == XINT (end))
9217 return empty_unibyte_string;
9218 return del_range_1 (XINT (start), XINT (end), 1, 1);
9222 The @code{validate_region} function checks whether the values
9223 passed as the beginning and end of the region are the proper type and
9224 are within range. If the beginning and end positions are the same,
9225 then return an empty string.
9227 The @code{del_range_1} function actually deletes the text. It is a
9228 complex function we will not look into. It updates the buffer and
9229 does other things. However, it is worth looking at the two arguments
9230 passed to @code{del_range}. These are @w{@code{XINT (start)}} and
9231 @w{@code{XINT (end)}}.
9233 As far as the C language is concerned, @code{start} and @code{end} are
9234 two integers that mark the beginning and end of the region to be
9235 deleted@footnote{More precisely, and requiring more expert knowledge
9236 to understand, the two integers are of type `Lisp_Object', which can
9237 also be a C union instead of an integer type.}.
9239 In early versions of Emacs, these two numbers were thirty-two bits
9240 long, but the code is slowly being generalized to handle other
9241 lengths. Three of the available bits are used to specify the type of
9242 information; the remaining bits are used as `content'.
9244 @samp{XINT} is a C macro that extracts the relevant number from the
9245 longer collection of bits; the three other bits are discarded.
9248 The command in @code{delete-and-extract-region} looks like this:
9251 del_range_1 (XINT (start), XINT (end), 1, 1);
9255 It deletes the region between the beginning position, @code{start},
9256 and the ending position, @code{end}.
9258 From the point of view of the person writing Lisp, Emacs is all very
9259 simple; but hidden underneath is a great deal of complexity to make it
9263 @section Initializing a Variable with @code{defvar}
9265 @cindex Initializing a variable
9266 @cindex Variable initialization
9271 copy-region-as-kill is short, 12 lines, and uses
9272 filter-buffer-substring, which is longer, 39 lines
9273 and has delete-and-extract-region in it.
9274 delete-and-extract-region is written in C.
9276 see Initializing a Variable with @code{defvar}
9280 The @code{copy-region-as-kill} function is written in Emacs Lisp. Two
9281 functions within it, @code{kill-append} and @code{kill-new}, copy a
9282 region in a buffer and save it in a variable called the
9283 @code{kill-ring}. This section describes how the @code{kill-ring}
9284 variable is created and initialized using the @code{defvar} special
9287 (Again we note that the term @code{kill-ring} is a misnomer. The text
9288 that is clipped out of the buffer can be brought back; it is not a ring
9289 of corpses, but a ring of resurrectable text.)
9291 In Emacs Lisp, a variable such as the @code{kill-ring} is created and
9292 given an initial value by using the @code{defvar} special form. The
9293 name comes from ``define variable''.
9295 The @code{defvar} special form is similar to @code{setq} in that it sets
9296 the value of a variable. It is unlike @code{setq} in two ways: first,
9297 it only sets the value of the variable if the variable does not already
9298 have a value. If the variable already has a value, @code{defvar} does
9299 not override the existing value. Second, @code{defvar} has a
9300 documentation string.
9302 (Another special form, @code{defcustom}, is designed for variables
9303 that people customize. It has more features than @code{defvar}.
9304 (@xref{defcustom, , Setting Variables with @code{defcustom}}.)
9307 * See variable current value::
9308 * defvar and asterisk::
9312 @node See variable current value
9313 @unnumberedsubsec Seeing the Current Value of a Variable
9316 You can see the current value of a variable, any variable, by using
9317 the @code{describe-variable} function, which is usually invoked by
9318 typing @kbd{C-h v}. If you type @kbd{C-h v} and then @code{kill-ring}
9319 (followed by @key{RET}) when prompted, you will see what is in your
9320 current kill ring---this may be quite a lot! Conversely, if you have
9321 been doing nothing this Emacs session except read this document, you
9322 may have nothing in it. Also, you will see the documentation for
9328 List of killed text sequences.
9329 Since the kill ring is supposed to interact nicely with cut-and-paste
9330 facilities offered by window systems, use of this variable should
9333 interact nicely with `interprogram-cut-function' and
9334 `interprogram-paste-function'. The functions `kill-new',
9335 `kill-append', and `current-kill' are supposed to implement this
9336 interaction; you may want to use them instead of manipulating the kill
9342 The kill ring is defined by a @code{defvar} in the following way:
9346 (defvar kill-ring nil
9347 "List of killed text sequences.
9353 In this variable definition, the variable is given an initial value of
9354 @code{nil}, which makes sense, since if you have saved nothing, you want
9355 nothing back if you give a @code{yank} command. The documentation
9356 string is written just like the documentation string of a @code{defun}.
9357 As with the documentation string of the @code{defun}, the first line of
9358 the documentation should be a complete sentence, since some commands,
9359 like @code{apropos}, print only the first line of documentation.
9360 Succeeding lines should not be indented; otherwise they look odd when
9361 you use @kbd{C-h v} (@code{describe-variable}).
9363 @node defvar and asterisk
9364 @subsection @code{defvar} and an asterisk
9365 @findex defvar @r{for a user customizable variable}
9366 @findex defvar @r{with an asterisk}
9368 In the past, Emacs used the @code{defvar} special form both for
9369 internal variables that you would not expect a user to change and for
9370 variables that you do expect a user to change. Although you can still
9371 use @code{defvar} for user customizable variables, please use
9372 @code{defcustom} instead, since that special form provides a path into
9373 the Customization commands. (@xref{defcustom, , Specifying Variables
9374 using @code{defcustom}}.)
9376 When you specified a variable using the @code{defvar} special form,
9377 you could distinguish a variable that a user might want to change from
9378 others by typing an asterisk, @samp{*}, in the first column of its
9379 documentation string. For example:
9383 (defvar shell-command-default-error-buffer nil
9384 "*Buffer name for `shell-command' @dots{} error output.
9389 @findex set-variable
9391 You could (and still can) use the @code{set-variable} command to
9392 change the value of @code{shell-command-default-error-buffer}
9393 temporarily. However, options set using @code{set-variable} are set
9394 only for the duration of your editing session. The new values are not
9395 saved between sessions. Each time Emacs starts, it reads the original
9396 value, unless you change the value within your @file{.emacs} file,
9397 either by setting it manually or by using @code{customize}.
9398 @xref{Emacs Initialization, , Your @file{.emacs} File}.
9400 For me, the major use of the @code{set-variable} command is to suggest
9401 variables that I might want to set in my @file{.emacs} file. There
9402 are now more than 700 such variables, far too many to remember
9403 readily. Fortunately, you can press @key{TAB} after calling the
9404 @code{M-x set-variable} command to see the list of variables.
9405 (@xref{Examining, , Examining and Setting Variables, emacs,
9406 The GNU Emacs Manual}.)
9409 @node cons & search-fwd Review
9412 Here is a brief summary of some recently introduced functions.
9417 @code{car} returns the first element of a list; @code{cdr} returns the
9418 second and subsequent elements of a list.
9425 (car '(1 2 3 4 5 6 7))
9427 (cdr '(1 2 3 4 5 6 7))
9428 @result{} (2 3 4 5 6 7)
9433 @code{cons} constructs a list by prepending its first argument to its
9447 @code{funcall} evaluates its first argument as a function. It passes
9448 its remaining arguments to its first argument.
9451 Return the result of taking @sc{cdr} `n' times on a list.
9459 The `rest of the rest', as it were.
9466 (nthcdr 3 '(1 2 3 4 5 6 7))
9473 @code{setcar} changes the first element of a list; @code{setcdr}
9474 changes the second and subsequent elements of a list.
9481 (setq triple '(1 2 3))
9488 (setcdr triple '("foo" "bar"))
9491 @result{} (37 "foo" "bar")
9496 Evaluate each argument in sequence and then return the value of the
9509 @item save-restriction
9510 Record whatever narrowing is in effect in the current buffer, if any,
9511 and restore that narrowing after evaluating the arguments.
9513 @item search-forward
9514 Search for a string, and if the string is found, move point. With a
9515 regular expression, use the similar @code{re-search-forward}.
9516 (@xref{Regexp Search, , Regular Expression Searches}, for an
9517 explanation of regular expression patterns and searches.)
9521 @code{search-forward} and @code{re-search-forward} take four
9526 The string or regular expression to search for.
9529 Optionally, the limit of the search.
9532 Optionally, what to do if the search fails, return @code{nil} or an
9536 Optionally, how many times to repeat the search; if negative, the
9537 search goes backwards.
9541 @itemx delete-and-extract-region
9542 @itemx copy-region-as-kill
9544 @code{kill-region} cuts the text between point and mark from the
9545 buffer and stores that text in the kill ring, so you can get it back
9548 @code{copy-region-as-kill} copies the text between point and mark into
9549 the kill ring, from which you can get it by yanking. The function
9550 does not cut or remove the text from the buffer.
9553 @code{delete-and-extract-region} removes the text between point and
9554 mark from the buffer and throws it away. You cannot get it back.
9555 (This is not an interactive command.)
9558 @node search Exercises
9559 @section Searching Exercises
9563 Write an interactive function that searches for a string. If the
9564 search finds the string, leave point after it and display a message
9565 that says ``Found!''. (Do not use @code{search-forward} for the name
9566 of this function; if you do, you will overwrite the existing version of
9567 @code{search-forward} that comes with Emacs. Use a name such as
9568 @code{test-search} instead.)
9571 Write a function that prints the third element of the kill ring in the
9572 echo area, if any; if the kill ring does not contain a third element,
9573 print an appropriate message.
9576 @node List Implementation
9577 @chapter How Lists are Implemented
9578 @cindex Lists in a computer
9580 In Lisp, atoms are recorded in a straightforward fashion; if the
9581 implementation is not straightforward in practice, it is, nonetheless,
9582 straightforward in theory. The atom @samp{rose}, for example, is
9583 recorded as the four contiguous letters @samp{r}, @samp{o}, @samp{s},
9584 @samp{e}. A list, on the other hand, is kept differently. The mechanism
9585 is equally simple, but it takes a moment to get used to the idea. A
9586 list is kept using a series of pairs of pointers. In the series, the
9587 first pointer in each pair points to an atom or to another list, and the
9588 second pointer in each pair points to the next pair, or to the symbol
9589 @code{nil}, which marks the end of the list.
9591 A pointer itself is quite simply the electronic address of what is
9592 pointed to. Hence, a list is kept as a series of electronic addresses.
9595 * Lists diagrammed::
9596 * Symbols as Chest:: Exploring a powerful metaphor.
9601 @node Lists diagrammed
9602 @unnumberedsec Lists diagrammed
9605 For example, the list @code{(rose violet buttercup)} has three elements,
9606 @samp{rose}, @samp{violet}, and @samp{buttercup}. In the computer, the
9607 electronic address of @samp{rose} is recorded in a segment of computer
9608 memory along with the address that gives the electronic address of where
9609 the atom @samp{violet} is located; and that address (the one that tells
9610 where @samp{violet} is located) is kept along with an address that tells
9611 where the address for the atom @samp{buttercup} is located.
9614 This sounds more complicated than it is and is easier seen in a diagram:
9616 @c clear print-postscript-figures
9617 @c !!! cons-cell-diagram #1
9621 ___ ___ ___ ___ ___ ___
9622 |___|___|--> |___|___|--> |___|___|--> nil
9625 --> rose --> violet --> buttercup
9629 @ifset print-postscript-figures
9632 @center @image{cons-1}
9633 %%%% old method of including an image
9634 % \input /usr/local/lib/tex/inputs/psfig.tex
9635 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-1.eps}}
9640 @ifclear print-postscript-figures
9644 ___ ___ ___ ___ ___ ___
9645 |___|___|--> |___|___|--> |___|___|--> nil
9648 --> rose --> violet --> buttercup
9655 In the diagram, each box represents a word of computer memory that
9656 holds a Lisp object, usually in the form of a memory address. The boxes,
9657 i.e., the addresses, are in pairs. Each arrow points to what the address
9658 is the address of, either an atom or another pair of addresses. The
9659 first box is the electronic address of @samp{rose} and the arrow points
9660 to @samp{rose}; the second box is the address of the next pair of boxes,
9661 the first part of which is the address of @samp{violet} and the second
9662 part of which is the address of the next pair. The very last box
9663 points to the symbol @code{nil}, which marks the end of the list.
9666 When a variable is set to a list with a function such as @code{setq},
9667 it stores the address of the first box in the variable. Thus,
9668 evaluation of the expression
9671 (setq bouquet '(rose violet buttercup))
9676 creates a situation like this:
9678 @c cons-cell-diagram #2
9684 | ___ ___ ___ ___ ___ ___
9685 --> |___|___|--> |___|___|--> |___|___|--> nil
9688 --> rose --> violet --> buttercup
9692 @ifset print-postscript-figures
9695 @center @image{cons-2}
9696 %%%% old method of including an image
9697 % \input /usr/local/lib/tex/inputs/psfig.tex
9698 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-2.eps}}
9703 @ifclear print-postscript-figures
9709 | ___ ___ ___ ___ ___ ___
9710 --> |___|___|--> |___|___|--> |___|___|--> nil
9713 --> rose --> violet --> buttercup
9720 In this example, the symbol @code{bouquet} holds the address of the first
9724 This same list can be illustrated in a different sort of box notation
9727 @c cons-cell-diagram #2a
9733 | -------------- --------------- ----------------
9734 | | car | cdr | | car | cdr | | car | cdr |
9735 -->| rose | o------->| violet | o------->| butter- | nil |
9736 | | | | | | | cup | |
9737 -------------- --------------- ----------------
9741 @ifset print-postscript-figures
9744 @center @image{cons-2a}
9745 %%%% old method of including an image
9746 % \input /usr/local/lib/tex/inputs/psfig.tex
9747 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-2a.eps}}
9752 @ifclear print-postscript-figures
9758 | -------------- --------------- ----------------
9759 | | car | cdr | | car | cdr | | car | cdr |
9760 -->| rose | o------->| violet | o------->| butter- | nil |
9761 | | | | | | | cup | |
9762 -------------- --------------- ----------------
9768 (Symbols consist of more than pairs of addresses, but the structure of
9769 a symbol is made up of addresses. Indeed, the symbol @code{bouquet}
9770 consists of a group of address-boxes, one of which is the address of
9771 the printed word @samp{bouquet}, a second of which is the address of a
9772 function definition attached to the symbol, if any, a third of which
9773 is the address of the first pair of address-boxes for the list
9774 @code{(rose violet buttercup)}, and so on. Here we are showing that
9775 the symbol's third address-box points to the first pair of
9776 address-boxes for the list.)
9778 If a symbol is set to the @sc{cdr} of a list, the list itself is not
9779 changed; the symbol simply has an address further down the list. (In
9780 the jargon, @sc{car} and @sc{cdr} are `non-destructive'.) Thus,
9781 evaluation of the following expression
9784 (setq flowers (cdr bouquet))
9791 @c cons-cell-diagram #3
9798 | ___ ___ | ___ ___ ___ ___
9799 --> | | | --> | | | | | |
9800 |___|___|----> |___|___|--> |___|___|--> nil
9803 --> rose --> violet --> buttercup
9808 @ifset print-postscript-figures
9811 @center @image{cons-3}
9812 %%%% old method of including an image
9813 % \input /usr/local/lib/tex/inputs/psfig.tex
9814 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-3.eps}}
9819 @ifclear print-postscript-figures
9826 | ___ ___ | ___ ___ ___ ___
9827 --> | | | --> | | | | | |
9828 |___|___|----> |___|___|--> |___|___|--> nil
9831 --> rose --> violet --> buttercup
9839 The value of @code{flowers} is @code{(violet buttercup)}, which is
9840 to say, the symbol @code{flowers} holds the address of the pair of
9841 address-boxes, the first of which holds the address of @code{violet},
9842 and the second of which holds the address of @code{buttercup}.
9844 A pair of address-boxes is called a @dfn{cons cell} or @dfn{dotted
9845 pair}. @xref{Cons Cell Type, , Cons Cell and List Types, elisp, The GNU Emacs Lisp
9846 Reference Manual}, and @ref{Dotted Pair Notation, , Dotted Pair
9847 Notation, elisp, The GNU Emacs Lisp Reference Manual}, for more
9848 information about cons cells and dotted pairs.
9851 The function @code{cons} adds a new pair of addresses to the front of
9852 a series of addresses like that shown above. For example, evaluating
9856 (setq bouquet (cons 'lily bouquet))
9863 @c cons-cell-diagram #4
9870 | ___ ___ ___ ___ | ___ ___ ___ ___
9871 --> | | | | | | --> | | | | | |
9872 |___|___|----> |___|___|----> |___|___|---->|___|___|--> nil
9875 --> lily --> rose --> violet --> buttercup
9880 @ifset print-postscript-figures
9883 @center @image{cons-4}
9884 %%%% old method of including an image
9885 % \input /usr/local/lib/tex/inputs/psfig.tex
9886 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-4.eps}}
9891 @ifclear print-postscript-figures
9898 | ___ ___ ___ ___ | ___ ___ ___ ___
9899 --> | | | | | | --> | | | | | |
9900 |___|___|----> |___|___|----> |___|___|---->|___|___|--> nil
9903 --> lily --> rose --> violet --> buttercup
9912 However, this does not change the value of the symbol
9913 @code{flowers}, as you can see by evaluating the following,
9916 (eq (cdr (cdr bouquet)) flowers)
9920 which returns @code{t} for true.
9922 Until it is reset, @code{flowers} still has the value
9923 @code{(violet buttercup)}; that is, it has the address of the cons
9924 cell whose first address is of @code{violet}. Also, this does not
9925 alter any of the pre-existing cons cells; they are all still there.
9927 Thus, in Lisp, to get the @sc{cdr} of a list, you just get the address
9928 of the next cons cell in the series; to get the @sc{car} of a list,
9929 you get the address of the first element of the list; to @code{cons} a
9930 new element on a list, you add a new cons cell to the front of the list.
9931 That is all there is to it! The underlying structure of Lisp is
9934 And what does the last address in a series of cons cells refer to? It
9935 is the address of the empty list, of @code{nil}.
9937 In summary, when a Lisp variable is set to a value, it is provided with
9938 the address of the list to which the variable refers.
9940 @node Symbols as Chest
9941 @section Symbols as a Chest of Drawers
9942 @cindex Symbols as a Chest of Drawers
9943 @cindex Chest of Drawers, metaphor for a symbol
9944 @cindex Drawers, Chest of, metaphor for a symbol
9946 In an earlier section, I suggested that you might imagine a symbol as
9947 being a chest of drawers. The function definition is put in one
9948 drawer, the value in another, and so on. What is put in the drawer
9949 holding the value can be changed without affecting the contents of the
9950 drawer holding the function definition, and vice-verse.
9952 Actually, what is put in each drawer is the address of the value or
9953 function definition. It is as if you found an old chest in the attic,
9954 and in one of its drawers you found a map giving you directions to
9955 where the buried treasure lies.
9957 (In addition to its name, symbol definition, and variable value, a
9958 symbol has a `drawer' for a @dfn{property list} which can be used to
9959 record other information. Property lists are not discussed here; see
9960 @ref{Property Lists, , Property Lists, elisp, The GNU Emacs Lisp
9964 Here is a fanciful representation:
9966 @c chest-of-drawers diagram
9971 Chest of Drawers Contents of Drawers
9975 ---------------------
9976 | directions to | [map to]
9977 | symbol name | bouquet
9979 +---------------------+
9981 | symbol definition | [none]
9983 +---------------------+
9984 | directions to | [map to]
9985 | variable value | (rose violet buttercup)
9987 +---------------------+
9989 | property list | [not described here]
9991 +---------------------+
9997 @ifset print-postscript-figures
10000 @center @image{drawers}
10001 %%%% old method of including an image
10002 % \input /usr/local/lib/tex/inputs/psfig.tex
10003 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/drawers.eps}}
10008 @ifclear print-postscript-figures
10013 Chest of Drawers Contents of Drawers
10017 ---------------------
10018 | directions to | [map to]
10019 | symbol name | bouquet
10021 +---------------------+
10023 | symbol definition | [none]
10025 +---------------------+
10026 | directions to | [map to]
10027 | variable value | (rose violet buttercup)
10029 +---------------------+
10031 | property list | [not described here]
10033 +---------------------+
10041 @node List Exercise
10044 Set @code{flowers} to @code{violet} and @code{buttercup}. Cons two
10045 more flowers on to this list and set this new list to
10046 @code{more-flowers}. Set the @sc{car} of @code{flowers} to a fish.
10047 What does the @code{more-flowers} list now contain?
10050 @chapter Yanking Text Back
10052 @cindex Text retrieval
10053 @cindex Retrieving text
10054 @cindex Pasting text
10056 Whenever you cut text out of a buffer with a `kill' command in GNU Emacs,
10057 you can bring it back with a `yank' command. The text that is cut out of
10058 the buffer is put in the kill ring and the yank commands insert the
10059 appropriate contents of the kill ring back into a buffer (not necessarily
10060 the original buffer).
10062 A simple @kbd{C-y} (@code{yank}) command inserts the first item from
10063 the kill ring into the current buffer. If the @kbd{C-y} command is
10064 followed immediately by @kbd{M-y}, the first element is replaced by
10065 the second element. Successive @kbd{M-y} commands replace the second
10066 element with the third, fourth, or fifth element, and so on. When the
10067 last element in the kill ring is reached, it is replaced by the first
10068 element and the cycle is repeated. (Thus the kill ring is called a
10069 `ring' rather than just a `list'. However, the actual data structure
10070 that holds the text is a list.
10071 @xref{Kill Ring, , Handling the Kill Ring}, for the details of how the
10072 list is handled as a ring.)
10075 * Kill Ring Overview::
10076 * kill-ring-yank-pointer:: The kill ring is a list.
10077 * yank nthcdr Exercises:: The @code{kill-ring-yank-pointer} variable.
10080 @node Kill Ring Overview
10081 @section Kill Ring Overview
10082 @cindex Kill ring overview
10084 The kill ring is a list of textual strings. This is what it looks like:
10087 ("some text" "a different piece of text" "yet more text")
10090 If this were the contents of my kill ring and I pressed @kbd{C-y}, the
10091 string of characters saying @samp{some text} would be inserted in this
10092 buffer where my cursor is located.
10094 The @code{yank} command is also used for duplicating text by copying it.
10095 The copied text is not cut from the buffer, but a copy of it is put on the
10096 kill ring and is inserted by yanking it back.
10098 Three functions are used for bringing text back from the kill ring:
10099 @code{yank}, which is usually bound to @kbd{C-y}; @code{yank-pop},
10100 which is usually bound to @kbd{M-y}; and @code{rotate-yank-pointer},
10101 which is used by the two other functions.
10103 These functions refer to the kill ring through a variable called the
10104 @code{kill-ring-yank-pointer}. Indeed, the insertion code for both the
10105 @code{yank} and @code{yank-pop} functions is:
10108 (insert (car kill-ring-yank-pointer))
10112 (Well, no more. In GNU Emacs 22, the function has been replaced by
10113 @code{insert-for-yank} which calls @code{insert-for-yank-1}
10114 repetitively for each @code{yank-handler} segment. In turn,
10115 @code{insert-for-yank-1} strips text properties from the inserted text
10116 according to @code{yank-excluded-properties}. Otherwise, it is just
10117 like @code{insert}. We will stick with plain @code{insert} since it
10118 is easier to understand.)
10120 To begin to understand how @code{yank} and @code{yank-pop} work, it is
10121 first necessary to look at the @code{kill-ring-yank-pointer} variable.
10123 @node kill-ring-yank-pointer
10124 @section The @code{kill-ring-yank-pointer} Variable
10126 @code{kill-ring-yank-pointer} is a variable, just as @code{kill-ring} is
10127 a variable. It points to something by being bound to the value of what
10128 it points to, like any other Lisp variable.
10131 Thus, if the value of the kill ring is:
10134 ("some text" "a different piece of text" "yet more text")
10139 and the @code{kill-ring-yank-pointer} points to the second clause, the
10140 value of @code{kill-ring-yank-pointer} is:
10143 ("a different piece of text" "yet more text")
10146 As explained in the previous chapter (@pxref{List Implementation}), the
10147 computer does not keep two different copies of the text being pointed to
10148 by both the @code{kill-ring} and the @code{kill-ring-yank-pointer}. The
10149 words ``a different piece of text'' and ``yet more text'' are not
10150 duplicated. Instead, the two Lisp variables point to the same pieces of
10151 text. Here is a diagram:
10153 @c cons-cell-diagram #5
10157 kill-ring kill-ring-yank-pointer
10159 | ___ ___ | ___ ___ ___ ___
10160 ---> | | | --> | | | | | |
10161 |___|___|----> |___|___|--> |___|___|--> nil
10164 | | --> "yet more text"
10166 | --> "a different piece of text"
10173 @ifset print-postscript-figures
10176 @center @image{cons-5}
10177 %%%% old method of including an image
10178 % \input /usr/local/lib/tex/inputs/psfig.tex
10179 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-5.eps}}
10184 @ifclear print-postscript-figures
10188 kill-ring kill-ring-yank-pointer
10190 | ___ ___ | ___ ___ ___ ___
10191 ---> | | | --> | | | | | |
10192 |___|___|----> |___|___|--> |___|___|--> nil
10195 | | --> "yet more text"
10197 | --> "a different piece of text
10206 Both the variable @code{kill-ring} and the variable
10207 @code{kill-ring-yank-pointer} are pointers. But the kill ring itself is
10208 usually described as if it were actually what it is composed of. The
10209 @code{kill-ring} is spoken of as if it were the list rather than that it
10210 points to the list. Conversely, the @code{kill-ring-yank-pointer} is
10211 spoken of as pointing to a list.
10213 These two ways of talking about the same thing sound confusing at first but
10214 make sense on reflection. The kill ring is generally thought of as the
10215 complete structure of data that holds the information of what has recently
10216 been cut out of the Emacs buffers. The @code{kill-ring-yank-pointer}
10217 on the other hand, serves to indicate---that is, to `point to'---that part
10218 of the kill ring of which the first element (the @sc{car}) will be
10222 In GNU Emacs 22, the @code{kill-new} function calls
10224 @code{(setq kill-ring-yank-pointer kill-ring)}
10226 (defun rotate-yank-pointer (arg)
10227 "Rotate the yanking point in the kill ring.
10228 With argument, rotate that many kills forward (or backward, if negative)."
10230 (current-kill arg))
10232 (defun current-kill (n &optional do-not-move)
10233 "Rotate the yanking point by N places, and then return that kill.
10234 If N is zero, `interprogram-paste-function' is set, and calling it
10235 returns a string, then that string is added to the front of the
10236 kill ring and returned as the latest kill.
10237 If optional arg DO-NOT-MOVE is non-nil, then don't actually move the
10238 yanking point; just return the Nth kill forward."
10239 (let ((interprogram-paste (and (= n 0)
10240 interprogram-paste-function
10241 (funcall interprogram-paste-function))))
10242 (if interprogram-paste
10244 ;; Disable the interprogram cut function when we add the new
10245 ;; text to the kill ring, so Emacs doesn't try to own the
10246 ;; selection, with identical text.
10247 (let ((interprogram-cut-function nil))
10248 (kill-new interprogram-paste))
10249 interprogram-paste)
10250 (or kill-ring (error "Kill ring is empty"))
10251 (let ((ARGth-kill-element
10252 (nthcdr (mod (- n (length kill-ring-yank-pointer))
10253 (length kill-ring))
10256 (setq kill-ring-yank-pointer ARGth-kill-element))
10257 (car ARGth-kill-element)))))
10262 @node yank nthcdr Exercises
10263 @section Exercises with @code{yank} and @code{nthcdr}
10267 Using @kbd{C-h v} (@code{describe-variable}), look at the value of
10268 your kill ring. Add several items to your kill ring; look at its
10269 value again. Using @kbd{M-y} (@code{yank-pop)}, move all the way
10270 around the kill ring. How many items were in your kill ring? Find
10271 the value of @code{kill-ring-max}. Was your kill ring full, or could
10272 you have kept more blocks of text within it?
10275 Using @code{nthcdr} and @code{car}, construct a series of expressions
10276 to return the first, second, third, and fourth elements of a list.
10279 @node Loops & Recursion
10280 @chapter Loops and Recursion
10281 @cindex Loops and recursion
10282 @cindex Recursion and loops
10283 @cindex Repetition (loops)
10285 Emacs Lisp has two primary ways to cause an expression, or a series of
10286 expressions, to be evaluated repeatedly: one uses a @code{while}
10287 loop, and the other uses @dfn{recursion}.
10289 Repetition can be very valuable. For example, to move forward four
10290 sentences, you need only write a program that will move forward one
10291 sentence and then repeat the process four times. Since a computer does
10292 not get bored or tired, such repetitive action does not have the
10293 deleterious effects that excessive or the wrong kinds of repetition can
10296 People mostly write Emacs Lisp functions using @code{while} loops and
10297 their kin; but you can use recursion, which provides a very powerful
10298 way to think about and then to solve problems@footnote{You can write
10299 recursive functions to be frugal or wasteful of mental or computer
10300 resources; as it happens, methods that people find easy---that are
10301 frugal of `mental resources'---sometimes use considerable computer
10302 resources. Emacs was designed to run on machines that we now consider
10303 limited and its default settings are conservative. You may want to
10304 increase the values of @code{max-specpdl-size} and
10305 @code{max-lisp-eval-depth}. In my @file{.emacs} file, I set them to
10306 15 and 30 times their default value.}.
10309 * while:: Causing a stretch of code to repeat.
10311 * Recursion:: Causing a function to call itself.
10312 * Looping exercise::
10316 @section @code{while}
10320 The @code{while} special form tests whether the value returned by
10321 evaluating its first argument is true or false. This is similar to what
10322 the Lisp interpreter does with an @code{if}; what the interpreter does
10323 next, however, is different.
10325 In a @code{while} expression, if the value returned by evaluating the
10326 first argument is false, the Lisp interpreter skips the rest of the
10327 expression (the @dfn{body} of the expression) and does not evaluate it.
10328 However, if the value is true, the Lisp interpreter evaluates the body
10329 of the expression and then again tests whether the first argument to
10330 @code{while} is true or false. If the value returned by evaluating the
10331 first argument is again true, the Lisp interpreter again evaluates the
10332 body of the expression.
10335 The template for a @code{while} expression looks like this:
10339 (while @var{true-or-false-test}
10345 * Looping with while:: Repeat so long as test returns true.
10346 * Loop Example:: A @code{while} loop that uses a list.
10347 * print-elements-of-list:: Uses @code{while}, @code{car}, @code{cdr}.
10348 * Incrementing Loop:: A loop with an incrementing counter.
10349 * Incrementing Loop Details::
10350 * Decrementing Loop:: A loop with a decrementing counter.
10354 @node Looping with while
10355 @unnumberedsubsec Looping with @code{while}
10358 So long as the true-or-false-test of the @code{while} expression
10359 returns a true value when it is evaluated, the body is repeatedly
10360 evaluated. This process is called a loop since the Lisp interpreter
10361 repeats the same thing again and again, like an airplane doing a loop.
10362 When the result of evaluating the true-or-false-test is false, the
10363 Lisp interpreter does not evaluate the rest of the @code{while}
10364 expression and `exits the loop'.
10366 Clearly, if the value returned by evaluating the first argument to
10367 @code{while} is always true, the body following will be evaluated
10368 again and again @dots{} and again @dots{} forever. Conversely, if the
10369 value returned is never true, the expressions in the body will never
10370 be evaluated. The craft of writing a @code{while} loop consists of
10371 choosing a mechanism such that the true-or-false-test returns true
10372 just the number of times that you want the subsequent expressions to
10373 be evaluated, and then have the test return false.
10375 The value returned by evaluating a @code{while} is the value of the
10376 true-or-false-test. An interesting consequence of this is that a
10377 @code{while} loop that evaluates without error will return @code{nil}
10378 or false regardless of whether it has looped 1 or 100 times or none at
10379 all. A @code{while} expression that evaluates successfully never
10380 returns a true value! What this means is that @code{while} is always
10381 evaluated for its side effects, which is to say, the consequences of
10382 evaluating the expressions within the body of the @code{while} loop.
10383 This makes sense. It is not the mere act of looping that is desired,
10384 but the consequences of what happens when the expressions in the loop
10385 are repeatedly evaluated.
10388 @subsection A @code{while} Loop and a List
10390 A common way to control a @code{while} loop is to test whether a list
10391 has any elements. If it does, the loop is repeated; but if it does not,
10392 the repetition is ended. Since this is an important technique, we will
10393 create a short example to illustrate it.
10395 A simple way to test whether a list has elements is to evaluate the
10396 list: if it has no elements, it is an empty list and will return the
10397 empty list, @code{()}, which is a synonym for @code{nil} or false. On
10398 the other hand, a list with elements will return those elements when it
10399 is evaluated. Since Emacs Lisp considers as true any value that is not
10400 @code{nil}, a list that returns elements will test true in a
10404 For example, you can set the variable @code{empty-list} to @code{nil} by
10405 evaluating the following @code{setq} expression:
10408 (setq empty-list ())
10412 After evaluating the @code{setq} expression, you can evaluate the
10413 variable @code{empty-list} in the usual way, by placing the cursor after
10414 the symbol and typing @kbd{C-x C-e}; @code{nil} will appear in your
10421 On the other hand, if you set a variable to be a list with elements, the
10422 list will appear when you evaluate the variable, as you can see by
10423 evaluating the following two expressions:
10427 (setq animals '(gazelle giraffe lion tiger))
10433 Thus, to create a @code{while} loop that tests whether there are any
10434 items in the list @code{animals}, the first part of the loop will be
10445 When the @code{while} tests its first argument, the variable
10446 @code{animals} is evaluated. It returns a list. So long as the list
10447 has elements, the @code{while} considers the results of the test to be
10448 true; but when the list is empty, it considers the results of the test
10451 To prevent the @code{while} loop from running forever, some mechanism
10452 needs to be provided to empty the list eventually. An oft-used
10453 technique is to have one of the subsequent forms in the @code{while}
10454 expression set the value of the list to be the @sc{cdr} of the list.
10455 Each time the @code{cdr} function is evaluated, the list will be made
10456 shorter, until eventually only the empty list will be left. At this
10457 point, the test of the @code{while} loop will return false, and the
10458 arguments to the @code{while} will no longer be evaluated.
10460 For example, the list of animals bound to the variable @code{animals}
10461 can be set to be the @sc{cdr} of the original list with the
10462 following expression:
10465 (setq animals (cdr animals))
10469 If you have evaluated the previous expressions and then evaluate this
10470 expression, you will see @code{(giraffe lion tiger)} appear in the echo
10471 area. If you evaluate the expression again, @code{(lion tiger)} will
10472 appear in the echo area. If you evaluate it again and yet again,
10473 @code{(tiger)} appears and then the empty list, shown by @code{nil}.
10475 A template for a @code{while} loop that uses the @code{cdr} function
10476 repeatedly to cause the true-or-false-test eventually to test false
10481 (while @var{test-whether-list-is-empty}
10483 @var{set-list-to-cdr-of-list})
10487 This test and use of @code{cdr} can be put together in a function that
10488 goes through a list and prints each element of the list on a line of its
10491 @node print-elements-of-list
10492 @subsection An Example: @code{print-elements-of-list}
10493 @findex print-elements-of-list
10495 The @code{print-elements-of-list} function illustrates a @code{while}
10498 @cindex @file{*scratch*} buffer
10499 The function requires several lines for its output. If you are
10500 reading this in a recent instance of GNU Emacs,
10501 @c GNU Emacs 21, GNU Emacs 22, or a later version,
10502 you can evaluate the following expression inside of Info, as usual.
10504 If you are using an earlier version of Emacs, you need to copy the
10505 necessary expressions to your @file{*scratch*} buffer and evaluate
10506 them there. This is because the echo area had only one line in the
10509 You can copy the expressions by marking the beginning of the region
10510 with @kbd{C-@key{SPC}} (@code{set-mark-command}), moving the cursor to
10511 the end of the region and then copying the region using @kbd{M-w}
10512 (@code{kill-ring-save}, which calls @code{copy-region-as-kill} and
10513 then provides visual feedback). In the @file{*scratch*}
10514 buffer, you can yank the expressions back by typing @kbd{C-y}
10517 After you have copied the expressions to the @file{*scratch*} buffer,
10518 evaluate each expression in turn. Be sure to evaluate the last
10519 expression, @code{(print-elements-of-list animals)}, by typing
10520 @kbd{C-u C-x C-e}, that is, by giving an argument to
10521 @code{eval-last-sexp}. This will cause the result of the evaluation
10522 to be printed in the @file{*scratch*} buffer instead of being printed
10523 in the echo area. (Otherwise you will see something like this in your
10524 echo area: @code{^Jgazelle^J^Jgiraffe^J^Jlion^J^Jtiger^Jnil}, in which
10525 each @samp{^J} stands for a `newline'.)
10528 In a recent instance of GNU Emacs, you can evaluate these expressions
10529 directly in the Info buffer, and the echo area will grow to show the
10534 (setq animals '(gazelle giraffe lion tiger))
10536 (defun print-elements-of-list (list)
10537 "Print each element of LIST on a line of its own."
10540 (setq list (cdr list))))
10542 (print-elements-of-list animals)
10548 When you evaluate the three expressions in sequence, you will see
10564 Each element of the list is printed on a line of its own (that is what
10565 the function @code{print} does) and then the value returned by the
10566 function is printed. Since the last expression in the function is the
10567 @code{while} loop, and since @code{while} loops always return
10568 @code{nil}, a @code{nil} is printed after the last element of the list.
10570 @node Incrementing Loop
10571 @subsection A Loop with an Incrementing Counter
10573 A loop is not useful unless it stops when it ought. Besides
10574 controlling a loop with a list, a common way of stopping a loop is to
10575 write the first argument as a test that returns false when the correct
10576 number of repetitions are complete. This means that the loop must
10577 have a counter---an expression that counts how many times the loop
10581 @node Incrementing Loop Details
10582 @unnumberedsubsec Details of an Incrementing Loop
10585 The test for a loop with an incrementing counter can be an expression
10586 such as @code{(< count desired-number)} which returns @code{t} for
10587 true if the value of @code{count} is less than the
10588 @code{desired-number} of repetitions and @code{nil} for false if the
10589 value of @code{count} is equal to or is greater than the
10590 @code{desired-number}. The expression that increments the count can
10591 be a simple @code{setq} such as @code{(setq count (1+ count))}, where
10592 @code{1+} is a built-in function in Emacs Lisp that adds 1 to its
10593 argument. (The expression @w{@code{(1+ count)}} has the same result
10594 as @w{@code{(+ count 1)}}, but is easier for a human to read.)
10597 The template for a @code{while} loop controlled by an incrementing
10598 counter looks like this:
10602 @var{set-count-to-initial-value}
10603 (while (< count desired-number) ; @r{true-or-false-test}
10605 (setq count (1+ count))) ; @r{incrementer}
10610 Note that you need to set the initial value of @code{count}; usually it
10614 * Incrementing Example:: Counting pebbles in a triangle.
10615 * Inc Example parts:: The parts of the function definition.
10616 * Inc Example altogether:: Putting the function definition together.
10619 @node Incrementing Example
10620 @unnumberedsubsubsec Example with incrementing counter
10622 Suppose you are playing on the beach and decide to make a triangle of
10623 pebbles, putting one pebble in the first row, two in the second row,
10624 three in the third row and so on, like this:
10642 @bullet{} @bullet{}
10643 @bullet{} @bullet{} @bullet{}
10644 @bullet{} @bullet{} @bullet{} @bullet{}
10651 (About 2500 years ago, Pythagoras and others developed the beginnings of
10652 number theory by considering questions such as this.)
10654 Suppose you want to know how many pebbles you will need to make a
10655 triangle with 7 rows?
10657 Clearly, what you need to do is add up the numbers from 1 to 7. There
10658 are two ways to do this; start with the smallest number, one, and add up
10659 the list in sequence, 1, 2, 3, 4 and so on; or start with the largest
10660 number and add the list going down: 7, 6, 5, 4 and so on. Because both
10661 mechanisms illustrate common ways of writing @code{while} loops, we will
10662 create two examples, one counting up and the other counting down. In
10663 this first example, we will start with 1 and add 2, 3, 4 and so on.
10665 If you are just adding up a short list of numbers, the easiest way to do
10666 it is to add up all the numbers at once. However, if you do not know
10667 ahead of time how many numbers your list will have, or if you want to be
10668 prepared for a very long list, then you need to design your addition so
10669 that what you do is repeat a simple process many times instead of doing
10670 a more complex process once.
10672 For example, instead of adding up all the pebbles all at once, what you
10673 can do is add the number of pebbles in the first row, 1, to the number
10674 in the second row, 2, and then add the total of those two rows to the
10675 third row, 3. Then you can add the number in the fourth row, 4, to the
10676 total of the first three rows; and so on.
10678 The critical characteristic of the process is that each repetitive
10679 action is simple. In this case, at each step we add only two numbers,
10680 the number of pebbles in the row and the total already found. This
10681 process of adding two numbers is repeated again and again until the last
10682 row has been added to the total of all the preceding rows. In a more
10683 complex loop the repetitive action might not be so simple, but it will
10684 be simpler than doing everything all at once.
10686 @node Inc Example parts
10687 @unnumberedsubsubsec The parts of the function definition
10689 The preceding analysis gives us the bones of our function definition:
10690 first, we will need a variable that we can call @code{total} that will
10691 be the total number of pebbles. This will be the value returned by
10694 Second, we know that the function will require an argument: this
10695 argument will be the total number of rows in the triangle. It can be
10696 called @code{number-of-rows}.
10698 Finally, we need a variable to use as a counter. We could call this
10699 variable @code{counter}, but a better name is @code{row-number}. That
10700 is because what the counter does in this function is count rows, and a
10701 program should be written to be as understandable as possible.
10703 When the Lisp interpreter first starts evaluating the expressions in the
10704 function, the value of @code{total} should be set to zero, since we have
10705 not added anything to it. Then the function should add the number of
10706 pebbles in the first row to the total, and then add the number of
10707 pebbles in the second to the total, and then add the number of
10708 pebbles in the third row to the total, and so on, until there are no
10709 more rows left to add.
10711 Both @code{total} and @code{row-number} are used only inside the
10712 function, so they can be declared as local variables with @code{let}
10713 and given initial values. Clearly, the initial value for @code{total}
10714 should be 0. The initial value of @code{row-number} should be 1,
10715 since we start with the first row. This means that the @code{let}
10716 statement will look like this:
10726 After the internal variables are declared and bound to their initial
10727 values, we can begin the @code{while} loop. The expression that serves
10728 as the test should return a value of @code{t} for true so long as the
10729 @code{row-number} is less than or equal to the @code{number-of-rows}.
10730 (If the expression tests true only so long as the row number is less
10731 than the number of rows in the triangle, the last row will never be
10732 added to the total; hence the row number has to be either less than or
10733 equal to the number of rows.)
10736 @findex <= @r{(less than or equal)}
10737 Lisp provides the @code{<=} function that returns true if the value of
10738 its first argument is less than or equal to the value of its second
10739 argument and false otherwise. So the expression that the @code{while}
10740 will evaluate as its test should look like this:
10743 (<= row-number number-of-rows)
10746 The total number of pebbles can be found by repeatedly adding the number
10747 of pebbles in a row to the total already found. Since the number of
10748 pebbles in the row is equal to the row number, the total can be found by
10749 adding the row number to the total. (Clearly, in a more complex
10750 situation, the number of pebbles in the row might be related to the row
10751 number in a more complicated way; if this were the case, the row number
10752 would be replaced by the appropriate expression.)
10755 (setq total (+ total row-number))
10759 What this does is set the new value of @code{total} to be equal to the
10760 sum of adding the number of pebbles in the row to the previous total.
10762 After setting the value of @code{total}, the conditions need to be
10763 established for the next repetition of the loop, if there is one. This
10764 is done by incrementing the value of the @code{row-number} variable,
10765 which serves as a counter. After the @code{row-number} variable has
10766 been incremented, the true-or-false-test at the beginning of the
10767 @code{while} loop tests whether its value is still less than or equal to
10768 the value of the @code{number-of-rows} and if it is, adds the new value
10769 of the @code{row-number} variable to the @code{total} of the previous
10770 repetition of the loop.
10773 The built-in Emacs Lisp function @code{1+} adds 1 to a number, so the
10774 @code{row-number} variable can be incremented with this expression:
10777 (setq row-number (1+ row-number))
10780 @node Inc Example altogether
10781 @unnumberedsubsubsec Putting the function definition together
10783 We have created the parts for the function definition; now we need to
10787 First, the contents of the @code{while} expression:
10791 (while (<= row-number number-of-rows) ; @r{true-or-false-test}
10792 (setq total (+ total row-number))
10793 (setq row-number (1+ row-number))) ; @r{incrementer}
10797 Along with the @code{let} expression varlist, this very nearly
10798 completes the body of the function definition. However, it requires
10799 one final element, the need for which is somewhat subtle.
10801 The final touch is to place the variable @code{total} on a line by
10802 itself after the @code{while} expression. Otherwise, the value returned
10803 by the whole function is the value of the last expression that is
10804 evaluated in the body of the @code{let}, and this is the value
10805 returned by the @code{while}, which is always @code{nil}.
10807 This may not be evident at first sight. It almost looks as if the
10808 incrementing expression is the last expression of the whole function.
10809 But that expression is part of the body of the @code{while}; it is the
10810 last element of the list that starts with the symbol @code{while}.
10811 Moreover, the whole of the @code{while} loop is a list within the body
10815 In outline, the function will look like this:
10819 (defun @var{name-of-function} (@var{argument-list})
10820 "@var{documentation}@dots{}"
10821 (let (@var{varlist})
10822 (while (@var{true-or-false-test})
10823 @var{body-of-while}@dots{} )
10824 @dots{} )) ; @r{Need final expression here.}
10828 The result of evaluating the @code{let} is what is going to be returned
10829 by the @code{defun} since the @code{let} is not embedded within any
10830 containing list, except for the @code{defun} as a whole. However, if
10831 the @code{while} is the last element of the @code{let} expression, the
10832 function will always return @code{nil}. This is not what we want!
10833 Instead, what we want is the value of the variable @code{total}. This
10834 is returned by simply placing the symbol as the last element of the list
10835 starting with @code{let}. It gets evaluated after the preceding
10836 elements of the list are evaluated, which means it gets evaluated after
10837 it has been assigned the correct value for the total.
10839 It may be easier to see this by printing the list starting with
10840 @code{let} all on one line. This format makes it evident that the
10841 @var{varlist} and @code{while} expressions are the second and third
10842 elements of the list starting with @code{let}, and the @code{total} is
10847 (let (@var{varlist}) (while (@var{true-or-false-test}) @var{body-of-while}@dots{} ) total)
10852 Putting everything together, the @code{triangle} function definition
10857 (defun triangle (number-of-rows) ; @r{Version with}
10858 ; @r{ incrementing counter.}
10859 "Add up the number of pebbles in a triangle.
10860 The first row has one pebble, the second row two pebbles,
10861 the third row three pebbles, and so on.
10862 The argument is NUMBER-OF-ROWS."
10867 (while (<= row-number number-of-rows)
10868 (setq total (+ total row-number))
10869 (setq row-number (1+ row-number)))
10875 After you have installed @code{triangle} by evaluating the function, you
10876 can try it out. Here are two examples:
10887 The sum of the first four numbers is 10 and the sum of the first seven
10890 @node Decrementing Loop
10891 @subsection Loop with a Decrementing Counter
10893 Another common way to write a @code{while} loop is to write the test
10894 so that it determines whether a counter is greater than zero. So long
10895 as the counter is greater than zero, the loop is repeated. But when
10896 the counter is equal to or less than zero, the loop is stopped. For
10897 this to work, the counter has to start out greater than zero and then
10898 be made smaller and smaller by a form that is evaluated
10901 The test will be an expression such as @code{(> counter 0)} which
10902 returns @code{t} for true if the value of @code{counter} is greater
10903 than zero, and @code{nil} for false if the value of @code{counter} is
10904 equal to or less than zero. The expression that makes the number
10905 smaller and smaller can be a simple @code{setq} such as @code{(setq
10906 counter (1- counter))}, where @code{1-} is a built-in function in
10907 Emacs Lisp that subtracts 1 from its argument.
10910 The template for a decrementing @code{while} loop looks like this:
10914 (while (> counter 0) ; @r{true-or-false-test}
10916 (setq counter (1- counter))) ; @r{decrementer}
10921 * Decrementing Example:: More pebbles on the beach.
10922 * Dec Example parts:: The parts of the function definition.
10923 * Dec Example altogether:: Putting the function definition together.
10926 @node Decrementing Example
10927 @unnumberedsubsubsec Example with decrementing counter
10929 To illustrate a loop with a decrementing counter, we will rewrite the
10930 @code{triangle} function so the counter decreases to zero.
10932 This is the reverse of the earlier version of the function. In this
10933 case, to find out how many pebbles are needed to make a triangle with
10934 3 rows, add the number of pebbles in the third row, 3, to the number
10935 in the preceding row, 2, and then add the total of those two rows to
10936 the row that precedes them, which is 1.
10938 Likewise, to find the number of pebbles in a triangle with 7 rows, add
10939 the number of pebbles in the seventh row, 7, to the number in the
10940 preceding row, which is 6, and then add the total of those two rows to
10941 the row that precedes them, which is 5, and so on. As in the previous
10942 example, each addition only involves adding two numbers, the total of
10943 the rows already added up and the number of pebbles in the row that is
10944 being added to the total. This process of adding two numbers is
10945 repeated again and again until there are no more pebbles to add.
10947 We know how many pebbles to start with: the number of pebbles in the
10948 last row is equal to the number of rows. If the triangle has seven
10949 rows, the number of pebbles in the last row is 7. Likewise, we know how
10950 many pebbles are in the preceding row: it is one less than the number in
10953 @node Dec Example parts
10954 @unnumberedsubsubsec The parts of the function definition
10956 We start with three variables: the total number of rows in the
10957 triangle; the number of pebbles in a row; and the total number of
10958 pebbles, which is what we want to calculate. These variables can be
10959 named @code{number-of-rows}, @code{number-of-pebbles-in-row}, and
10960 @code{total}, respectively.
10962 Both @code{total} and @code{number-of-pebbles-in-row} are used only
10963 inside the function and are declared with @code{let}. The initial
10964 value of @code{total} should, of course, be zero. However, the
10965 initial value of @code{number-of-pebbles-in-row} should be equal to
10966 the number of rows in the triangle, since the addition will start with
10970 This means that the beginning of the @code{let} expression will look
10976 (number-of-pebbles-in-row number-of-rows))
10981 The total number of pebbles can be found by repeatedly adding the number
10982 of pebbles in a row to the total already found, that is, by repeatedly
10983 evaluating the following expression:
10986 (setq total (+ total number-of-pebbles-in-row))
10990 After the @code{number-of-pebbles-in-row} is added to the @code{total},
10991 the @code{number-of-pebbles-in-row} should be decremented by one, since
10992 the next time the loop repeats, the preceding row will be
10993 added to the total.
10995 The number of pebbles in a preceding row is one less than the number of
10996 pebbles in a row, so the built-in Emacs Lisp function @code{1-} can be
10997 used to compute the number of pebbles in the preceding row. This can be
10998 done with the following expression:
11002 (setq number-of-pebbles-in-row
11003 (1- number-of-pebbles-in-row))
11007 Finally, we know that the @code{while} loop should stop making repeated
11008 additions when there are no pebbles in a row. So the test for
11009 the @code{while} loop is simply:
11012 (while (> number-of-pebbles-in-row 0)
11015 @node Dec Example altogether
11016 @unnumberedsubsubsec Putting the function definition together
11018 We can put these expressions together to create a function definition
11019 that works. However, on examination, we find that one of the local
11020 variables is unneeded!
11023 The function definition looks like this:
11027 ;;; @r{First subtractive version.}
11028 (defun triangle (number-of-rows)
11029 "Add up the number of pebbles in a triangle."
11031 (number-of-pebbles-in-row number-of-rows))
11032 (while (> number-of-pebbles-in-row 0)
11033 (setq total (+ total number-of-pebbles-in-row))
11034 (setq number-of-pebbles-in-row
11035 (1- number-of-pebbles-in-row)))
11040 As written, this function works.
11042 However, we do not need @code{number-of-pebbles-in-row}.
11044 @cindex Argument as local variable
11045 When the @code{triangle} function is evaluated, the symbol
11046 @code{number-of-rows} will be bound to a number, giving it an initial
11047 value. That number can be changed in the body of the function as if
11048 it were a local variable, without any fear that such a change will
11049 effect the value of the variable outside of the function. This is a
11050 very useful characteristic of Lisp; it means that the variable
11051 @code{number-of-rows} can be used anywhere in the function where
11052 @code{number-of-pebbles-in-row} is used.
11055 Here is a second version of the function written a bit more cleanly:
11059 (defun triangle (number) ; @r{Second version.}
11060 "Return sum of numbers 1 through NUMBER inclusive."
11062 (while (> number 0)
11063 (setq total (+ total number))
11064 (setq number (1- number)))
11069 In brief, a properly written @code{while} loop will consist of three parts:
11073 A test that will return false after the loop has repeated itself the
11074 correct number of times.
11077 An expression the evaluation of which will return the value desired
11078 after being repeatedly evaluated.
11081 An expression to change the value passed to the true-or-false-test so
11082 that the test returns false after the loop has repeated itself the right
11086 @node dolist dotimes
11087 @section Save your time: @code{dolist} and @code{dotimes}
11089 In addition to @code{while}, both @code{dolist} and @code{dotimes}
11090 provide for looping. Sometimes these are quicker to write than the
11091 equivalent @code{while} loop. Both are Lisp macros. (@xref{Macros, ,
11092 Macros, elisp, The GNU Emacs Lisp Reference Manual}. )
11094 @code{dolist} works like a @code{while} loop that `@sc{cdr}s down a
11095 list': @code{dolist} automatically shortens the list each time it
11096 loops---takes the @sc{cdr} of the list---and binds the @sc{car} of
11097 each shorter version of the list to the first of its arguments.
11099 @code{dotimes} loops a specific number of times: you specify the number.
11107 @unnumberedsubsec The @code{dolist} Macro
11110 Suppose, for example, you want to reverse a list, so that
11111 ``first'' ``second'' ``third'' becomes ``third'' ``second'' ``first''.
11114 In practice, you would use the @code{reverse} function, like this:
11118 (setq animals '(gazelle giraffe lion tiger))
11126 Here is how you could reverse the list using a @code{while} loop:
11130 (setq animals '(gazelle giraffe lion tiger))
11132 (defun reverse-list-with-while (list)
11133 "Using while, reverse the order of LIST."
11134 (let (value) ; make sure list starts empty
11136 (setq value (cons (car list) value))
11137 (setq list (cdr list)))
11140 (reverse-list-with-while animals)
11146 And here is how you could use the @code{dolist} macro:
11150 (setq animals '(gazelle giraffe lion tiger))
11152 (defun reverse-list-with-dolist (list)
11153 "Using dolist, reverse the order of LIST."
11154 (let (value) ; make sure list starts empty
11155 (dolist (element list value)
11156 (setq value (cons element value)))))
11158 (reverse-list-with-dolist animals)
11164 In Info, you can place your cursor after the closing parenthesis of
11165 each expression and type @kbd{C-x C-e}; in each case, you should see
11168 (tiger lion giraffe gazelle)
11174 For this example, the existing @code{reverse} function is obviously best.
11175 The @code{while} loop is just like our first example (@pxref{Loop
11176 Example, , A @code{while} Loop and a List}). The @code{while} first
11177 checks whether the list has elements; if so, it constructs a new list
11178 by adding the first element of the list to the existing list (which in
11179 the first iteration of the loop is @code{nil}). Since the second
11180 element is prepended in front of the first element, and the third
11181 element is prepended in front of the second element, the list is reversed.
11183 In the expression using a @code{while} loop,
11184 the @w{@code{(setq list (cdr list))}}
11185 expression shortens the list, so the @code{while} loop eventually
11186 stops. In addition, it provides the @code{cons} expression with a new
11187 first element by creating a new and shorter list at each repetition of
11190 The @code{dolist} expression does very much the same as the
11191 @code{while} expression, except that the @code{dolist} macro does some
11192 of the work you have to do when writing a @code{while} expression.
11194 Like a @code{while} loop, a @code{dolist} loops. What is different is
11195 that it automatically shortens the list each time it loops---it
11196 `@sc{cdr}s down the list' on its own---and it automatically binds
11197 the @sc{car} of each shorter version of the list to the first of its
11200 In the example, the @sc{car} of each shorter version of the list is
11201 referred to using the symbol @samp{element}, the list itself is called
11202 @samp{list}, and the value returned is called @samp{value}. The
11203 remainder of the @code{dolist} expression is the body.
11205 The @code{dolist} expression binds the @sc{car} of each shorter
11206 version of the list to @code{element} and then evaluates the body of
11207 the expression; and repeats the loop. The result is returned in
11211 @unnumberedsubsec The @code{dotimes} Macro
11214 The @code{dotimes} macro is similar to @code{dolist}, except that it
11215 loops a specific number of times.
11217 The first argument to @code{dotimes} is assigned the numbers 0, 1, 2
11218 and so forth each time around the loop, and the value of the third
11219 argument is returned. You need to provide the value of the second
11220 argument, which is how many times the macro loops.
11223 For example, the following binds the numbers from 0 up to, but not
11224 including, the number 3 to the first argument, @var{number}, and then
11225 constructs a list of the three numbers. (The first number is 0, the
11226 second number is 1, and the third number is 2; this makes a total of
11227 three numbers in all, starting with zero as the first number.)
11231 (let (value) ; otherwise a value is a void variable
11232 (dotimes (number 3 value)
11233 (setq value (cons number value))))
11240 @code{dotimes} returns @code{value}, so the way to use
11241 @code{dotimes} is to operate on some expression @var{number} number of
11242 times and then return the result, either as a list or an atom.
11245 Here is an example of a @code{defun} that uses @code{dotimes} to add
11246 up the number of pebbles in a triangle.
11250 (defun triangle-using-dotimes (number-of-rows)
11251 "Using dotimes, add up the number of pebbles in a triangle."
11252 (let ((total 0)) ; otherwise a total is a void variable
11253 (dotimes (number number-of-rows total)
11254 (setq total (+ total (1+ number))))))
11256 (triangle-using-dotimes 4)
11264 A recursive function contains code that tells the Lisp interpreter to
11265 call a program that runs exactly like itself, but with slightly
11266 different arguments. The code runs exactly the same because it has
11267 the same name. However, even though the program has the same name, it
11268 is not the same entity. It is different. In the jargon, it is a
11269 different `instance'.
11271 Eventually, if the program is written correctly, the `slightly
11272 different arguments' will become sufficiently different from the first
11273 arguments that the final instance will stop.
11276 * Building Robots:: Same model, different serial number ...
11277 * Recursive Definition Parts:: Walk until you stop ...
11278 * Recursion with list:: Using a list as the test whether to recurse.
11279 * Recursive triangle function::
11280 * Recursion with cond::
11281 * Recursive Patterns:: Often used templates.
11282 * No Deferment:: Don't store up work ...
11283 * No deferment solution::
11286 @node Building Robots
11287 @subsection Building Robots: Extending the Metaphor
11288 @cindex Building robots
11289 @cindex Robots, building
11291 It is sometimes helpful to think of a running program as a robot that
11292 does a job. In doing its job, a recursive function calls on a second
11293 robot to help it. The second robot is identical to the first in every
11294 way, except that the second robot helps the first and has been
11295 passed different arguments than the first.
11297 In a recursive function, the second robot may call a third; and the
11298 third may call a fourth, and so on. Each of these is a different
11299 entity; but all are clones.
11301 Since each robot has slightly different instructions---the arguments
11302 will differ from one robot to the next---the last robot should know
11305 Let's expand on the metaphor in which a computer program is a robot.
11307 A function definition provides the blueprints for a robot. When you
11308 install a function definition, that is, when you evaluate a
11309 @code{defun} special form, you install the necessary equipment to
11310 build robots. It is as if you were in a factory, setting up an
11311 assembly line. Robots with the same name are built according to the
11312 same blueprints. So they have, as it were, the same `model number',
11313 but a different `serial number'.
11315 We often say that a recursive function `calls itself'. What we mean
11316 is that the instructions in a recursive function cause the Lisp
11317 interpreter to run a different function that has the same name and
11318 does the same job as the first, but with different arguments.
11320 It is important that the arguments differ from one instance to the
11321 next; otherwise, the process will never stop.
11323 @node Recursive Definition Parts
11324 @subsection The Parts of a Recursive Definition
11325 @cindex Parts of a Recursive Definition
11326 @cindex Recursive Definition Parts
11328 A recursive function typically contains a conditional expression which
11333 A true-or-false-test that determines whether the function is called
11334 again, here called the @dfn{do-again-test}.
11337 The name of the function. When this name is called, a new instance of
11338 the function---a new robot, as it were---is created and told what to do.
11341 An expression that returns a different value each time the function is
11342 called, here called the @dfn{next-step-expression}. Consequently, the
11343 argument (or arguments) passed to the new instance of the function
11344 will be different from that passed to the previous instance. This
11345 causes the conditional expression, the @dfn{do-again-test}, to test
11346 false after the correct number of repetitions.
11349 Recursive functions can be much simpler than any other kind of
11350 function. Indeed, when people first start to use them, they often look
11351 so mysteriously simple as to be incomprehensible. Like riding a
11352 bicycle, reading a recursive function definition takes a certain knack
11353 which is hard at first but then seems simple.
11356 There are several different common recursive patterns. A very simple
11357 pattern looks like this:
11361 (defun @var{name-of-recursive-function} (@var{argument-list})
11362 "@var{documentation}@dots{}"
11363 (if @var{do-again-test}
11365 (@var{name-of-recursive-function}
11366 @var{next-step-expression})))
11370 Each time a recursive function is evaluated, a new instance of it is
11371 created and told what to do. The arguments tell the instance what to do.
11373 An argument is bound to the value of the next-step-expression. Each
11374 instance runs with a different value of the next-step-expression.
11376 The value in the next-step-expression is used in the do-again-test.
11378 The value returned by the next-step-expression is passed to the new
11379 instance of the function, which evaluates it (or some
11380 transmogrification of it) to determine whether to continue or stop.
11381 The next-step-expression is designed so that the do-again-test returns
11382 false when the function should no longer be repeated.
11384 The do-again-test is sometimes called the @dfn{stop condition},
11385 since it stops the repetitions when it tests false.
11387 @node Recursion with list
11388 @subsection Recursion with a List
11390 The example of a @code{while} loop that printed the elements of a list
11391 of numbers can be written recursively. Here is the code, including
11392 an expression to set the value of the variable @code{animals} to a list.
11394 If you are reading this in Info in Emacs, you can evaluate this
11395 expression directly in Info. Otherwise, you must copy the example
11396 to the @file{*scratch*} buffer and evaluate each expression there.
11397 Use @kbd{C-u C-x C-e} to evaluate the
11398 @code{(print-elements-recursively animals)} expression so that the
11399 results are printed in the buffer; otherwise the Lisp interpreter will
11400 try to squeeze the results into the one line of the echo area.
11402 Also, place your cursor immediately after the last closing parenthesis
11403 of the @code{print-elements-recursively} function, before the comment.
11404 Otherwise, the Lisp interpreter will try to evaluate the comment.
11406 @findex print-elements-recursively
11409 (setq animals '(gazelle giraffe lion tiger))
11411 (defun print-elements-recursively (list)
11412 "Print each element of LIST on a line of its own.
11414 (when list ; @r{do-again-test}
11415 (print (car list)) ; @r{body}
11416 (print-elements-recursively ; @r{recursive call}
11417 (cdr list)))) ; @r{next-step-expression}
11419 (print-elements-recursively animals)
11423 The @code{print-elements-recursively} function first tests whether
11424 there is any content in the list; if there is, the function prints the
11425 first element of the list, the @sc{car} of the list. Then the
11426 function `invokes itself', but gives itself as its argument, not the
11427 whole list, but the second and subsequent elements of the list, the
11428 @sc{cdr} of the list.
11430 Put another way, if the list is not empty, the function invokes
11431 another instance of code that is similar to the initial code, but is a
11432 different thread of execution, with different arguments than the first
11435 Put in yet another way, if the list is not empty, the first robot
11436 assembles a second robot and tells it what to do; the second robot is
11437 a different individual from the first, but is the same model.
11439 When the second evaluation occurs, the @code{when} expression is
11440 evaluated and if true, prints the first element of the list it
11441 receives as its argument (which is the second element of the original
11442 list). Then the function `calls itself' with the @sc{cdr} of the list
11443 it is invoked with, which (the second time around) is the @sc{cdr} of
11444 the @sc{cdr} of the original list.
11446 Note that although we say that the function `calls itself', what we
11447 mean is that the Lisp interpreter assembles and instructs a new
11448 instance of the program. The new instance is a clone of the first,
11449 but is a separate individual.
11451 Each time the function `invokes itself', it invokes itself on a
11452 shorter version of the original list. It creates a new instance that
11453 works on a shorter list.
11455 Eventually, the function invokes itself on an empty list. It creates
11456 a new instance whose argument is @code{nil}. The conditional expression
11457 tests the value of @code{list}. Since the value of @code{list} is
11458 @code{nil}, the @code{when} expression tests false so the then-part is
11459 not evaluated. The function as a whole then returns @code{nil}.
11462 When you evaluate the expression @code{(print-elements-recursively
11463 animals)} in the @file{*scratch*} buffer, you see this result:
11479 @node Recursive triangle function
11480 @subsection Recursion in Place of a Counter
11481 @findex triangle-recursively
11484 The @code{triangle} function described in a previous section can also
11485 be written recursively. It looks like this:
11489 (defun triangle-recursively (number)
11490 "Return the sum of the numbers 1 through NUMBER inclusive.
11492 (if (= number 1) ; @r{do-again-test}
11494 (+ number ; @r{else-part}
11495 (triangle-recursively ; @r{recursive call}
11496 (1- number))))) ; @r{next-step-expression}
11498 (triangle-recursively 7)
11503 You can install this function by evaluating it and then try it by
11504 evaluating @code{(triangle-recursively 7)}. (Remember to put your
11505 cursor immediately after the last parenthesis of the function
11506 definition, before the comment.) The function evaluates to 28.
11508 To understand how this function works, let's consider what happens in the
11509 various cases when the function is passed 1, 2, 3, or 4 as the value of
11513 * Recursive Example arg of 1 or 2::
11514 * Recursive Example arg of 3 or 4::
11518 @node Recursive Example arg of 1 or 2
11519 @unnumberedsubsubsec An argument of 1 or 2
11522 First, what happens if the value of the argument is 1?
11524 The function has an @code{if} expression after the documentation
11525 string. It tests whether the value of @code{number} is equal to 1; if
11526 so, Emacs evaluates the then-part of the @code{if} expression, which
11527 returns the number 1 as the value of the function. (A triangle with
11528 one row has one pebble in it.)
11530 Suppose, however, that the value of the argument is 2. In this case,
11531 Emacs evaluates the else-part of the @code{if} expression.
11534 The else-part consists of an addition, the recursive call to
11535 @code{triangle-recursively} and a decrementing action; and it looks like
11539 (+ number (triangle-recursively (1- number)))
11542 When Emacs evaluates this expression, the innermost expression is
11543 evaluated first; then the other parts in sequence. Here are the steps
11547 @item Step 1 @w{ } Evaluate the innermost expression.
11549 The innermost expression is @code{(1- number)} so Emacs decrements the
11550 value of @code{number} from 2 to 1.
11552 @item Step 2 @w{ } Evaluate the @code{triangle-recursively} function.
11554 The Lisp interpreter creates an individual instance of
11555 @code{triangle-recursively}. It does not matter that this function is
11556 contained within itself. Emacs passes the result Step 1 as the
11557 argument used by this instance of the @code{triangle-recursively}
11560 In this case, Emacs evaluates @code{triangle-recursively} with an
11561 argument of 1. This means that this evaluation of
11562 @code{triangle-recursively} returns 1.
11564 @item Step 3 @w{ } Evaluate the value of @code{number}.
11566 The variable @code{number} is the second element of the list that
11567 starts with @code{+}; its value is 2.
11569 @item Step 4 @w{ } Evaluate the @code{+} expression.
11571 The @code{+} expression receives two arguments, the first
11572 from the evaluation of @code{number} (Step 3) and the second from the
11573 evaluation of @code{triangle-recursively} (Step 2).
11575 The result of the addition is the sum of 2 plus 1, and the number 3 is
11576 returned, which is correct. A triangle with two rows has three
11580 @node Recursive Example arg of 3 or 4
11581 @unnumberedsubsubsec An argument of 3 or 4
11583 Suppose that @code{triangle-recursively} is called with an argument of
11587 @item Step 1 @w{ } Evaluate the do-again-test.
11589 The @code{if} expression is evaluated first. This is the do-again
11590 test and returns false, so the else-part of the @code{if} expression
11591 is evaluated. (Note that in this example, the do-again-test causes
11592 the function to call itself when it tests false, not when it tests
11595 @item Step 2 @w{ } Evaluate the innermost expression of the else-part.
11597 The innermost expression of the else-part is evaluated, which decrements
11598 3 to 2. This is the next-step-expression.
11600 @item Step 3 @w{ } Evaluate the @code{triangle-recursively} function.
11602 The number 2 is passed to the @code{triangle-recursively} function.
11604 We already know what happens when Emacs evaluates @code{triangle-recursively} with
11605 an argument of 2. After going through the sequence of actions described
11606 earlier, it returns a value of 3. So that is what will happen here.
11608 @item Step 4 @w{ } Evaluate the addition.
11610 3 will be passed as an argument to the addition and will be added to the
11611 number with which the function was called, which is 3.
11615 The value returned by the function as a whole will be 6.
11617 Now that we know what will happen when @code{triangle-recursively} is
11618 called with an argument of 3, it is evident what will happen if it is
11619 called with an argument of 4:
11623 In the recursive call, the evaluation of
11626 (triangle-recursively (1- 4))
11631 will return the value of evaluating
11634 (triangle-recursively 3)
11638 which is 6 and this value will be added to 4 by the addition in the
11643 The value returned by the function as a whole will be 10.
11645 Each time @code{triangle-recursively} is evaluated, it evaluates a
11646 version of itself---a different instance of itself---with a smaller
11647 argument, until the argument is small enough so that it does not
11650 Note that this particular design for a recursive function
11651 requires that operations be deferred.
11653 Before @code{(triangle-recursively 7)} can calculate its answer, it
11654 must call @code{(triangle-recursively 6)}; and before
11655 @code{(triangle-recursively 6)} can calculate its answer, it must call
11656 @code{(triangle-recursively 5)}; and so on. That is to say, the
11657 calculation that @code{(triangle-recursively 7)} makes must be
11658 deferred until @code{(triangle-recursively 6)} makes its calculation;
11659 and @code{(triangle-recursively 6)} must defer until
11660 @code{(triangle-recursively 5)} completes; and so on.
11662 If each of these instances of @code{triangle-recursively} are thought
11663 of as different robots, the first robot must wait for the second to
11664 complete its job, which must wait until the third completes, and so
11667 There is a way around this kind of waiting, which we will discuss in
11668 @ref{No Deferment, , Recursion without Deferments}.
11670 @node Recursion with cond
11671 @subsection Recursion Example Using @code{cond}
11674 The version of @code{triangle-recursively} described earlier is written
11675 with the @code{if} special form. It can also be written using another
11676 special form called @code{cond}. The name of the special form
11677 @code{cond} is an abbreviation of the word @samp{conditional}.
11679 Although the @code{cond} special form is not used as often in the
11680 Emacs Lisp sources as @code{if}, it is used often enough to justify
11684 The template for a @code{cond} expression looks like this:
11694 where the @var{body} is a series of lists.
11697 Written out more fully, the template looks like this:
11702 (@var{first-true-or-false-test} @var{first-consequent})
11703 (@var{second-true-or-false-test} @var{second-consequent})
11704 (@var{third-true-or-false-test} @var{third-consequent})
11709 When the Lisp interpreter evaluates the @code{cond} expression, it
11710 evaluates the first element (the @sc{car} or true-or-false-test) of
11711 the first expression in a series of expressions within the body of the
11714 If the true-or-false-test returns @code{nil} the rest of that
11715 expression, the consequent, is skipped and the true-or-false-test of the
11716 next expression is evaluated. When an expression is found whose
11717 true-or-false-test returns a value that is not @code{nil}, the
11718 consequent of that expression is evaluated. The consequent can be one
11719 or more expressions. If the consequent consists of more than one
11720 expression, the expressions are evaluated in sequence and the value of
11721 the last one is returned. If the expression does not have a consequent,
11722 the value of the true-or-false-test is returned.
11724 If none of the true-or-false-tests test true, the @code{cond} expression
11725 returns @code{nil}.
11728 Written using @code{cond}, the @code{triangle} function looks like this:
11732 (defun triangle-using-cond (number)
11733 (cond ((<= number 0) 0)
11736 (+ number (triangle-using-cond (1- number))))))
11741 In this example, the @code{cond} returns 0 if the number is less than or
11742 equal to 0, it returns 1 if the number is 1 and it evaluates @code{(+
11743 number (triangle-using-cond (1- number)))} if the number is greater than
11746 @node Recursive Patterns
11747 @subsection Recursive Patterns
11748 @cindex Recursive Patterns
11750 Here are three common recursive patterns. Each involves a list.
11751 Recursion does not need to involve lists, but Lisp is designed for lists
11752 and this provides a sense of its primal capabilities.
11761 @unnumberedsubsubsec Recursive Pattern: @emph{every}
11762 @cindex Every, type of recursive pattern
11763 @cindex Recursive pattern: every
11765 In the @code{every} recursive pattern, an action is performed on every
11769 The basic pattern is:
11773 If a list be empty, return @code{nil}.
11775 Else, act on the beginning of the list (the @sc{car} of the list)
11778 through a recursive call by the function on the rest (the
11779 @sc{cdr}) of the list,
11781 and, optionally, combine the acted-on element, using @code{cons},
11782 with the results of acting on the rest.
11791 (defun square-each (numbers-list)
11792 "Square each of a NUMBERS LIST, recursively."
11793 (if (not numbers-list) ; do-again-test
11796 (* (car numbers-list) (car numbers-list))
11797 (square-each (cdr numbers-list))))) ; next-step-expression
11801 (square-each '(1 2 3))
11808 If @code{numbers-list} is empty, do nothing. But if it has content,
11809 construct a list combining the square of the first number in the list
11810 with the result of the recursive call.
11812 (The example follows the pattern exactly: @code{nil} is returned if
11813 the numbers' list is empty. In practice, you would write the
11814 conditional so it carries out the action when the numbers' list is not
11817 The @code{print-elements-recursively} function (@pxref{Recursion with
11818 list, , Recursion with a List}) is another example of an @code{every}
11819 pattern, except in this case, rather than bring the results together
11820 using @code{cons}, we print each element of output.
11823 The @code{print-elements-recursively} function looks like this:
11827 (setq animals '(gazelle giraffe lion tiger))
11831 (defun print-elements-recursively (list)
11832 "Print each element of LIST on a line of its own.
11834 (when list ; @r{do-again-test}
11835 (print (car list)) ; @r{body}
11836 (print-elements-recursively ; @r{recursive call}
11837 (cdr list)))) ; @r{next-step-expression}
11839 (print-elements-recursively animals)
11844 The pattern for @code{print-elements-recursively} is:
11848 When the list is empty, do nothing.
11850 But when the list has at least one element,
11853 act on the beginning of the list (the @sc{car} of the list),
11855 and make a recursive call on the rest (the @sc{cdr}) of the list.
11860 @unnumberedsubsubsec Recursive Pattern: @emph{accumulate}
11861 @cindex Accumulate, type of recursive pattern
11862 @cindex Recursive pattern: accumulate
11864 Another recursive pattern is called the @code{accumulate} pattern. In
11865 the @code{accumulate} recursive pattern, an action is performed on
11866 every element of a list and the result of that action is accumulated
11867 with the results of performing the action on the other elements.
11869 This is very like the `every' pattern using @code{cons}, except that
11870 @code{cons} is not used, but some other combiner.
11877 If a list be empty, return zero or some other constant.
11879 Else, act on the beginning of the list (the @sc{car} of the list),
11882 and combine that acted-on element, using @code{+} or
11883 some other combining function, with
11885 a recursive call by the function on the rest (the @sc{cdr}) of the list.
11890 Here is an example:
11894 (defun add-elements (numbers-list)
11895 "Add the elements of NUMBERS-LIST together."
11896 (if (not numbers-list)
11898 (+ (car numbers-list) (add-elements (cdr numbers-list)))))
11902 (add-elements '(1 2 3 4))
11907 @xref{Files List, , Making a List of Files}, for an example of the
11908 accumulate pattern.
11911 @unnumberedsubsubsec Recursive Pattern: @emph{keep}
11912 @cindex Keep, type of recursive pattern
11913 @cindex Recursive pattern: keep
11915 A third recursive pattern is called the @code{keep} pattern.
11916 In the @code{keep} recursive pattern, each element of a list is tested;
11917 the element is acted on and the results are kept only if the element
11920 Again, this is very like the `every' pattern, except the element is
11921 skipped unless it meets a criterion.
11924 The pattern has three parts:
11928 If a list be empty, return @code{nil}.
11930 Else, if the beginning of the list (the @sc{car} of the list) passes
11934 act on that element and combine it, using @code{cons} with
11936 a recursive call by the function on the rest (the @sc{cdr}) of the list.
11939 Otherwise, if the beginning of the list (the @sc{car} of the list) fails
11943 skip on that element,
11945 and, recursively call the function on the rest (the @sc{cdr}) of the list.
11950 Here is an example that uses @code{cond}:
11954 (defun keep-three-letter-words (word-list)
11955 "Keep three letter words in WORD-LIST."
11957 ;; First do-again-test: stop-condition
11958 ((not word-list) nil)
11960 ;; Second do-again-test: when to act
11961 ((eq 3 (length (symbol-name (car word-list))))
11962 ;; combine acted-on element with recursive call on shorter list
11963 (cons (car word-list) (keep-three-letter-words (cdr word-list))))
11965 ;; Third do-again-test: when to skip element;
11966 ;; recursively call shorter list with next-step expression
11967 (t (keep-three-letter-words (cdr word-list)))))
11971 (keep-three-letter-words '(one two three four five six))
11972 @result{} (one two six)
11976 It goes without saying that you need not use @code{nil} as the test for
11977 when to stop; and you can, of course, combine these patterns.
11980 @subsection Recursion without Deferments
11981 @cindex Deferment in recursion
11982 @cindex Recursion without Deferments
11984 Let's consider again what happens with the @code{triangle-recursively}
11985 function. We will find that the intermediate calculations are
11986 deferred until all can be done.
11989 Here is the function definition:
11993 (defun triangle-recursively (number)
11994 "Return the sum of the numbers 1 through NUMBER inclusive.
11996 (if (= number 1) ; @r{do-again-test}
11998 (+ number ; @r{else-part}
11999 (triangle-recursively ; @r{recursive call}
12000 (1- number))))) ; @r{next-step-expression}
12004 What happens when we call this function with a argument of 7?
12006 The first instance of the @code{triangle-recursively} function adds
12007 the number 7 to the value returned by a second instance of
12008 @code{triangle-recursively}, an instance that has been passed an
12009 argument of 6. That is to say, the first calculation is:
12012 (+ 7 (triangle-recursively 6))
12016 The first instance of @code{triangle-recursively}---you may want to
12017 think of it as a little robot---cannot complete its job. It must hand
12018 off the calculation for @code{(triangle-recursively 6)} to a second
12019 instance of the program, to a second robot. This second individual is
12020 completely different from the first one; it is, in the jargon, a
12021 `different instantiation'. Or, put another way, it is a different
12022 robot. It is the same model as the first; it calculates triangle
12023 numbers recursively; but it has a different serial number.
12025 And what does @code{(triangle-recursively 6)} return? It returns the
12026 number 6 added to the value returned by evaluating
12027 @code{triangle-recursively} with an argument of 5. Using the robot
12028 metaphor, it asks yet another robot to help it.
12034 (+ 7 6 (triangle-recursively 5))
12038 And what happens next?
12041 (+ 7 6 5 (triangle-recursively 4))
12044 Each time @code{triangle-recursively} is called, except for the last
12045 time, it creates another instance of the program---another robot---and
12046 asks it to make a calculation.
12049 Eventually, the full addition is set up and performed:
12055 This design for the function defers the calculation of the first step
12056 until the second can be done, and defers that until the third can be
12057 done, and so on. Each deferment means the computer must remember what
12058 is being waited on. This is not a problem when there are only a few
12059 steps, as in this example. But it can be a problem when there are
12062 @node No deferment solution
12063 @subsection No Deferment Solution
12064 @cindex No deferment solution
12065 @cindex Defermentless solution
12066 @cindex Solution without deferment
12068 The solution to the problem of deferred operations is to write in a
12069 manner that does not defer operations@footnote{The phrase @dfn{tail
12070 recursive} is used to describe such a process, one that uses
12071 `constant space'.}. This requires
12072 writing to a different pattern, often one that involves writing two
12073 function definitions, an `initialization' function and a `helper'
12076 The `initialization' function sets up the job; the `helper' function
12080 Here are the two function definitions for adding up numbers. They are
12081 so simple, I find them hard to understand.
12085 (defun triangle-initialization (number)
12086 "Return the sum of the numbers 1 through NUMBER inclusive.
12087 This is the `initialization' component of a two function
12088 duo that uses recursion."
12089 (triangle-recursive-helper 0 0 number))
12095 (defun triangle-recursive-helper (sum counter number)
12096 "Return SUM, using COUNTER, through NUMBER inclusive.
12097 This is the `helper' component of a two function duo
12098 that uses recursion."
12099 (if (> counter number)
12101 (triangle-recursive-helper (+ sum counter) ; @r{sum}
12102 (1+ counter) ; @r{counter}
12103 number))) ; @r{number}
12108 Install both function definitions by evaluating them, then call
12109 @code{triangle-initialization} with 2 rows:
12113 (triangle-initialization 2)
12118 The `initialization' function calls the first instance of the `helper'
12119 function with three arguments: zero, zero, and a number which is the
12120 number of rows in the triangle.
12122 The first two arguments passed to the `helper' function are
12123 initialization values. These values are changed when
12124 @code{triangle-recursive-helper} invokes new instances.@footnote{The
12125 jargon is mildly confusing: @code{triangle-recursive-helper} uses a
12126 process that is iterative in a procedure that is recursive. The
12127 process is called iterative because the computer need only record the
12128 three values, @code{sum}, @code{counter}, and @code{number}; the
12129 procedure is recursive because the function `calls itself'. On the
12130 other hand, both the process and the procedure used by
12131 @code{triangle-recursively} are called recursive. The word
12132 `recursive' has different meanings in the two contexts.}
12134 Let's see what happens when we have a triangle that has one row. (This
12135 triangle will have one pebble in it!)
12138 @code{triangle-initialization} will call its helper with
12139 the arguments @w{@code{0 0 1}}. That function will run the conditional
12140 test whether @code{(> counter number)}:
12148 and find that the result is false, so it will invoke
12149 the else-part of the @code{if} clause:
12153 (triangle-recursive-helper
12154 (+ sum counter) ; @r{sum plus counter} @result{} @r{sum}
12155 (1+ counter) ; @r{increment counter} @result{} @r{counter}
12156 number) ; @r{number stays the same}
12162 which will first compute:
12166 (triangle-recursive-helper (+ 0 0) ; @r{sum}
12167 (1+ 0) ; @r{counter}
12171 (triangle-recursive-helper 0 1 1)
12175 Again, @code{(> counter number)} will be false, so again, the Lisp
12176 interpreter will evaluate @code{triangle-recursive-helper}, creating a
12177 new instance with new arguments.
12180 This new instance will be;
12184 (triangle-recursive-helper
12185 (+ sum counter) ; @r{sum plus counter} @result{} @r{sum}
12186 (1+ counter) ; @r{increment counter} @result{} @r{counter}
12187 number) ; @r{number stays the same}
12191 (triangle-recursive-helper 1 2 1)
12195 In this case, the @code{(> counter number)} test will be true! So the
12196 instance will return the value of the sum, which will be 1, as
12199 Now, let's pass @code{triangle-initialization} an argument
12200 of 2, to find out how many pebbles there are in a triangle with two rows.
12202 That function calls @code{(triangle-recursive-helper 0 0 2)}.
12205 In stages, the instances called will be:
12209 @r{sum counter number}
12210 (triangle-recursive-helper 0 1 2)
12212 (triangle-recursive-helper 1 2 2)
12214 (triangle-recursive-helper 3 3 2)
12218 When the last instance is called, the @code{(> counter number)} test
12219 will be true, so the instance will return the value of @code{sum},
12222 This kind of pattern helps when you are writing functions that can use
12223 many resources in a computer.
12226 @node Looping exercise
12227 @section Looping Exercise
12231 Write a function similar to @code{triangle} in which each row has a
12232 value which is the square of the row number. Use a @code{while} loop.
12235 Write a function similar to @code{triangle} that multiplies instead of
12239 Rewrite these two functions recursively. Rewrite these functions
12242 @c comma in printed title causes problem in Info cross reference
12244 Write a function for Texinfo mode that creates an index entry at the
12245 beginning of a paragraph for every @samp{@@dfn} within the paragraph.
12246 (In a Texinfo file, @samp{@@dfn} marks a definition. This book is
12247 written in Texinfo.)
12249 Many of the functions you will need are described in two of the
12250 previous chapters, @ref{Cutting & Storing Text, , Cutting and Storing
12251 Text}, and @ref{Yanking, , Yanking Text Back}. If you use
12252 @code{forward-paragraph} to put the index entry at the beginning of
12253 the paragraph, you will have to use @w{@kbd{C-h f}}
12254 (@code{describe-function}) to find out how to make the command go
12257 For more information, see
12259 @ref{Indicating, , Indicating Definitions, texinfo}.
12262 @ref{Indicating, , Indicating, texinfo, Texinfo Manual}, which goes to
12263 a Texinfo manual in the current directory. Or, if you are on the
12265 @uref{http://www.gnu.org/software/texinfo/manual/texinfo/}
12268 ``Indicating Definitions, Commands, etc.'' in @cite{Texinfo, The GNU
12269 Documentation Format}.
12273 @node Regexp Search
12274 @chapter Regular Expression Searches
12275 @cindex Searches, illustrating
12276 @cindex Regular expression searches
12277 @cindex Patterns, searching for
12278 @cindex Motion by sentence and paragraph
12279 @cindex Sentences, movement by
12280 @cindex Paragraphs, movement by
12282 Regular expression searches are used extensively in GNU Emacs. The
12283 two functions, @code{forward-sentence} and @code{forward-paragraph},
12284 illustrate these searches well. They use regular expressions to find
12285 where to move point. The phrase `regular expression' is often written
12288 Regular expression searches are described in @ref{Regexp Search, ,
12289 Regular Expression Search, emacs, The GNU Emacs Manual}, as well as in
12290 @ref{Regular Expressions, , , elisp, The GNU Emacs Lisp Reference
12291 Manual}. In writing this chapter, I am presuming that you have at
12292 least a mild acquaintance with them. The major point to remember is
12293 that regular expressions permit you to search for patterns as well as
12294 for literal strings of characters. For example, the code in
12295 @code{forward-sentence} searches for the pattern of possible
12296 characters that could mark the end of a sentence, and moves point to
12299 Before looking at the code for the @code{forward-sentence} function, it
12300 is worth considering what the pattern that marks the end of a sentence
12301 must be. The pattern is discussed in the next section; following that
12302 is a description of the regular expression search function,
12303 @code{re-search-forward}. The @code{forward-sentence} function
12304 is described in the section following. Finally, the
12305 @code{forward-paragraph} function is described in the last section of
12306 this chapter. @code{forward-paragraph} is a complex function that
12307 introduces several new features.
12310 * sentence-end:: The regular expression for @code{sentence-end}.
12311 * re-search-forward:: Very similar to @code{search-forward}.
12312 * forward-sentence:: A straightforward example of regexp search.
12313 * forward-paragraph:: A somewhat complex example.
12314 * etags:: How to create your own @file{TAGS} table.
12316 * re-search Exercises::
12320 @section The Regular Expression for @code{sentence-end}
12321 @findex sentence-end
12323 The symbol @code{sentence-end} is bound to the pattern that marks the
12324 end of a sentence. What should this regular expression be?
12326 Clearly, a sentence may be ended by a period, a question mark, or an
12327 exclamation mark. Indeed, in English, only clauses that end with one
12328 of those three characters should be considered the end of a sentence.
12329 This means that the pattern should include the character set:
12335 However, we do not want @code{forward-sentence} merely to jump to a
12336 period, a question mark, or an exclamation mark, because such a character
12337 might be used in the middle of a sentence. A period, for example, is
12338 used after abbreviations. So other information is needed.
12340 According to convention, you type two spaces after every sentence, but
12341 only one space after a period, a question mark, or an exclamation mark in
12342 the body of a sentence. So a period, a question mark, or an exclamation
12343 mark followed by two spaces is a good indicator of an end of sentence.
12344 However, in a file, the two spaces may instead be a tab or the end of a
12345 line. This means that the regular expression should include these three
12346 items as alternatives.
12349 This group of alternatives will look like this:
12360 Here, @samp{$} indicates the end of the line, and I have pointed out
12361 where the tab and two spaces are inserted in the expression. Both are
12362 inserted by putting the actual characters into the expression.
12364 Two backslashes, @samp{\\}, are required before the parentheses and
12365 vertical bars: the first backslash quotes the following backslash in
12366 Emacs; and the second indicates that the following character, the
12367 parenthesis or the vertical bar, is special.
12370 Also, a sentence may be followed by one or more carriage returns, like
12381 Like tabs and spaces, a carriage return is inserted into a regular
12382 expression by inserting it literally. The asterisk indicates that the
12383 @key{RET} is repeated zero or more times.
12385 But a sentence end does not consist only of a period, a question mark or
12386 an exclamation mark followed by appropriate space: a closing quotation
12387 mark or a closing brace of some kind may precede the space. Indeed more
12388 than one such mark or brace may precede the space. These require a
12389 expression that looks like this:
12395 In this expression, the first @samp{]} is the first character in the
12396 expression; the second character is @samp{"}, which is preceded by a
12397 @samp{\} to tell Emacs the @samp{"} is @emph{not} special. The last
12398 three characters are @samp{'}, @samp{)}, and @samp{@}}.
12400 All this suggests what the regular expression pattern for matching the
12401 end of a sentence should be; and, indeed, if we evaluate
12402 @code{sentence-end} we find that it returns the following value:
12407 @result{} "[.?!][]\"')@}]*\\($\\| \\| \\)[
12413 (Well, not in GNU Emacs 22; that is because of an effort to make the
12414 process simpler and to handle more glyphs and languages. When the
12415 value of @code{sentence-end} is @code{nil}, then use the value defined
12416 by the function @code{sentence-end}. (Here is a use of the difference
12417 between a value and a function in Emacs Lisp.) The function returns a
12418 value constructed from the variables @code{sentence-end-base},
12419 @code{sentence-end-double-space}, @code{sentence-end-without-period},
12420 and @code{sentence-end-without-space}. The critical variable is
12421 @code{sentence-end-base}; its global value is similar to the one
12422 described above but it also contains two additional quotation marks.
12423 These have differing degrees of curliness. The
12424 @code{sentence-end-without-period} variable, when true, tells Emacs
12425 that a sentence may end without a period, such as text in Thai.)
12429 (Note that here the @key{TAB}, two spaces, and @key{RET} are shown
12430 literally in the pattern.)
12432 This regular expression can be deciphered as follows:
12436 The first part of the pattern is the three characters, a period, a question
12437 mark and an exclamation mark, within square brackets. The pattern must
12438 begin with one or other of these characters.
12441 The second part of the pattern is the group of closing braces and
12442 quotation marks, which can appear zero or more times. These may follow
12443 the period, question mark or exclamation mark. In a regular expression,
12444 the backslash, @samp{\}, followed by the double quotation mark,
12445 @samp{"}, indicates the class of string-quote characters. Usually, the
12446 double quotation mark is the only character in this class. The
12447 asterisk, @samp{*}, indicates that the items in the previous group (the
12448 group surrounded by square brackets, @samp{[]}) may be repeated zero or
12451 @item \\($\\| \\| \\)
12452 The third part of the pattern is one or other of: either the end of a
12453 line, or two blank spaces, or a tab. The double back-slashes are used
12454 to prevent Emacs from reading the parentheses and vertical bars as part
12455 of the search pattern; the parentheses are used to mark the group and
12456 the vertical bars are used to indicated that the patterns to either side
12457 of them are alternatives. The dollar sign is used to indicate the end
12458 of a line and both the two spaces and the tab are each inserted as is to
12459 indicate what they are.
12462 Finally, the last part of the pattern indicates that the end of the line
12463 or the whitespace following the period, question mark or exclamation
12464 mark may, but need not, be followed by one or more carriage returns. In
12465 the pattern, the carriage return is inserted as an actual carriage
12466 return between square brackets but here it is shown as @key{RET}.
12470 @node re-search-forward
12471 @section The @code{re-search-forward} Function
12472 @findex re-search-forward
12474 The @code{re-search-forward} function is very like the
12475 @code{search-forward} function. (@xref{search-forward, , The
12476 @code{search-forward} Function}.)
12478 @code{re-search-forward} searches for a regular expression. If the
12479 search is successful, it leaves point immediately after the last
12480 character in the target. If the search is backwards, it leaves point
12481 just before the first character in the target. You may tell
12482 @code{re-search-forward} to return @code{t} for true. (Moving point
12483 is therefore a `side effect'.)
12485 Like @code{search-forward}, the @code{re-search-forward} function takes
12490 The first argument is the regular expression that the function searches
12491 for. The regular expression will be a string between quotation marks.
12494 The optional second argument limits how far the function will search; it is a
12495 bound, which is specified as a position in the buffer.
12498 The optional third argument specifies how the function responds to
12499 failure: @code{nil} as the third argument causes the function to
12500 signal an error (and print a message) when the search fails; any other
12501 value causes it to return @code{nil} if the search fails and @code{t}
12502 if the search succeeds.
12505 The optional fourth argument is the repeat count. A negative repeat
12506 count causes @code{re-search-forward} to search backwards.
12510 The template for @code{re-search-forward} looks like this:
12514 (re-search-forward "@var{regular-expression}"
12515 @var{limit-of-search}
12516 @var{what-to-do-if-search-fails}
12517 @var{repeat-count})
12521 The second, third, and fourth arguments are optional. However, if you
12522 want to pass a value to either or both of the last two arguments, you
12523 must also pass a value to all the preceding arguments. Otherwise, the
12524 Lisp interpreter will mistake which argument you are passing the value
12528 In the @code{forward-sentence} function, the regular expression will be
12529 the value of the variable @code{sentence-end}. In simple form, that is:
12533 "[.?!][]\"')@}]*\\($\\| \\| \\)[
12539 The limit of the search will be the end of the paragraph (since a
12540 sentence cannot go beyond a paragraph). If the search fails, the
12541 function will return @code{nil}; and the repeat count will be provided
12542 by the argument to the @code{forward-sentence} function.
12544 @node forward-sentence
12545 @section @code{forward-sentence}
12546 @findex forward-sentence
12548 The command to move the cursor forward a sentence is a straightforward
12549 illustration of how to use regular expression searches in Emacs Lisp.
12550 Indeed, the function looks longer and more complicated than it is; this
12551 is because the function is designed to go backwards as well as forwards;
12552 and, optionally, over more than one sentence. The function is usually
12553 bound to the key command @kbd{M-e}.
12556 * Complete forward-sentence::
12557 * fwd-sentence while loops:: Two @code{while} loops.
12558 * fwd-sentence re-search:: A regular expression search.
12562 @node Complete forward-sentence
12563 @unnumberedsubsec Complete @code{forward-sentence} function definition
12567 Here is the code for @code{forward-sentence}:
12572 (defun forward-sentence (&optional arg)
12573 "Move forward to next `sentence-end'. With argument, repeat.
12574 With negative argument, move backward repeatedly to `sentence-beginning'.
12576 The variable `sentence-end' is a regular expression that matches ends of
12577 sentences. Also, every paragraph boundary terminates sentences as well."
12581 (or arg (setq arg 1))
12582 (let ((opoint (point))
12583 (sentence-end (sentence-end)))
12585 (let ((pos (point))
12586 (par-beg (save-excursion (start-of-paragraph-text) (point))))
12587 (if (and (re-search-backward sentence-end par-beg t)
12588 (or (< (match-end 0) pos)
12589 (re-search-backward sentence-end par-beg t)))
12590 (goto-char (match-end 0))
12591 (goto-char par-beg)))
12592 (setq arg (1+ arg)))
12596 (let ((par-end (save-excursion (end-of-paragraph-text) (point))))
12597 (if (re-search-forward sentence-end par-end t)
12598 (skip-chars-backward " \t\n")
12599 (goto-char par-end)))
12600 (setq arg (1- arg)))
12601 (constrain-to-field nil opoint t)))
12609 (defun forward-sentence (&optional arg)
12610 "Move forward to next sentence-end. With argument, repeat.
12611 With negative argument, move backward repeatedly to sentence-beginning.
12612 Sentence ends are identified by the value of sentence-end
12613 treated as a regular expression. Also, every paragraph boundary
12614 terminates sentences as well."
12618 (or arg (setq arg 1))
12621 (save-excursion (start-of-paragraph-text) (point))))
12622 (if (re-search-backward
12623 (concat sentence-end "[^ \t\n]") par-beg t)
12624 (goto-char (1- (match-end 0)))
12625 (goto-char par-beg)))
12626 (setq arg (1+ arg)))
12629 (save-excursion (end-of-paragraph-text) (point))))
12630 (if (re-search-forward sentence-end par-end t)
12631 (skip-chars-backward " \t\n")
12632 (goto-char par-end)))
12633 (setq arg (1- arg))))
12638 The function looks long at first sight and it is best to look at its
12639 skeleton first, and then its muscle. The way to see the skeleton is to
12640 look at the expressions that start in the left-most columns:
12644 (defun forward-sentence (&optional arg)
12645 "@var{documentation}@dots{}"
12647 (or arg (setq arg 1))
12648 (let ((opoint (point)) (sentence-end (sentence-end)))
12650 (let ((pos (point))
12651 (par-beg (save-excursion (start-of-paragraph-text) (point))))
12652 @var{rest-of-body-of-while-loop-when-going-backwards}
12654 (let ((par-end (save-excursion (end-of-paragraph-text) (point))))
12655 @var{rest-of-body-of-while-loop-when-going-forwards}
12656 @var{handle-forms-and-equivalent}
12660 This looks much simpler! The function definition consists of
12661 documentation, an @code{interactive} expression, an @code{or}
12662 expression, a @code{let} expression, and @code{while} loops.
12664 Let's look at each of these parts in turn.
12666 We note that the documentation is thorough and understandable.
12668 The function has an @code{interactive "p"} declaration. This means
12669 that the processed prefix argument, if any, is passed to the
12670 function as its argument. (This will be a number.) If the function
12671 is not passed an argument (it is optional) then the argument
12672 @code{arg} will be bound to 1.
12674 When @code{forward-sentence} is called non-interactively without an
12675 argument, @code{arg} is bound to @code{nil}. The @code{or} expression
12676 handles this. What it does is either leave the value of @code{arg} as
12677 it is, but only if @code{arg} is bound to a value; or it sets the
12678 value of @code{arg} to 1, in the case when @code{arg} is bound to
12681 Next is a @code{let}. That specifies the values of two local
12682 variables, @code{point} and @code{sentence-end}. The local value of
12683 point, from before the search, is used in the
12684 @code{constrain-to-field} function which handles forms and
12685 equivalents. The @code{sentence-end} variable is set by the
12686 @code{sentence-end} function.
12688 @node fwd-sentence while loops
12689 @unnumberedsubsec The @code{while} loops
12691 Two @code{while} loops follow. The first @code{while} has a
12692 true-or-false-test that tests true if the prefix argument for
12693 @code{forward-sentence} is a negative number. This is for going
12694 backwards. The body of this loop is similar to the body of the second
12695 @code{while} clause, but it is not exactly the same. We will skip
12696 this @code{while} loop and concentrate on the second @code{while}
12700 The second @code{while} loop is for moving point forward. Its skeleton
12705 (while (> arg 0) ; @r{true-or-false-test}
12707 (if (@var{true-or-false-test})
12710 (setq arg (1- arg)))) ; @code{while} @r{loop decrementer}
12714 The @code{while} loop is of the decrementing kind.
12715 (@xref{Decrementing Loop, , A Loop with a Decrementing Counter}.) It
12716 has a true-or-false-test that tests true so long as the counter (in
12717 this case, the variable @code{arg}) is greater than zero; and it has a
12718 decrementer that subtracts 1 from the value of the counter every time
12721 If no prefix argument is given to @code{forward-sentence}, which is
12722 the most common way the command is used, this @code{while} loop will
12723 run once, since the value of @code{arg} will be 1.
12725 The body of the @code{while} loop consists of a @code{let} expression,
12726 which creates and binds a local variable, and has, as its body, an
12727 @code{if} expression.
12730 The body of the @code{while} loop looks like this:
12735 (save-excursion (end-of-paragraph-text) (point))))
12736 (if (re-search-forward sentence-end par-end t)
12737 (skip-chars-backward " \t\n")
12738 (goto-char par-end)))
12742 The @code{let} expression creates and binds the local variable
12743 @code{par-end}. As we shall see, this local variable is designed to
12744 provide a bound or limit to the regular expression search. If the
12745 search fails to find a proper sentence ending in the paragraph, it will
12746 stop on reaching the end of the paragraph.
12748 But first, let us examine how @code{par-end} is bound to the value of
12749 the end of the paragraph. What happens is that the @code{let} sets the
12750 value of @code{par-end} to the value returned when the Lisp interpreter
12751 evaluates the expression
12755 (save-excursion (end-of-paragraph-text) (point))
12760 In this expression, @code{(end-of-paragraph-text)} moves point to the
12761 end of the paragraph, @code{(point)} returns the value of point, and then
12762 @code{save-excursion} restores point to its original position. Thus,
12763 the @code{let} binds @code{par-end} to the value returned by the
12764 @code{save-excursion} expression, which is the position of the end of
12765 the paragraph. (The @code{end-of-paragraph-text} function uses
12766 @code{forward-paragraph}, which we will discuss shortly.)
12769 Emacs next evaluates the body of the @code{let}, which is an @code{if}
12770 expression that looks like this:
12774 (if (re-search-forward sentence-end par-end t) ; @r{if-part}
12775 (skip-chars-backward " \t\n") ; @r{then-part}
12776 (goto-char par-end))) ; @r{else-part}
12780 The @code{if} tests whether its first argument is true and if so,
12781 evaluates its then-part; otherwise, the Emacs Lisp interpreter
12782 evaluates the else-part. The true-or-false-test of the @code{if}
12783 expression is the regular expression search.
12785 It may seem odd to have what looks like the `real work' of
12786 the @code{forward-sentence} function buried here, but this is a common
12787 way this kind of operation is carried out in Lisp.
12789 @node fwd-sentence re-search
12790 @unnumberedsubsec The regular expression search
12792 The @code{re-search-forward} function searches for the end of the
12793 sentence, that is, for the pattern defined by the @code{sentence-end}
12794 regular expression. If the pattern is found---if the end of the sentence is
12795 found---then the @code{re-search-forward} function does two things:
12799 The @code{re-search-forward} function carries out a side effect, which
12800 is to move point to the end of the occurrence found.
12803 The @code{re-search-forward} function returns a value of true. This is
12804 the value received by the @code{if}, and means that the search was
12809 The side effect, the movement of point, is completed before the
12810 @code{if} function is handed the value returned by the successful
12811 conclusion of the search.
12813 When the @code{if} function receives the value of true from a successful
12814 call to @code{re-search-forward}, the @code{if} evaluates the then-part,
12815 which is the expression @code{(skip-chars-backward " \t\n")}. This
12816 expression moves backwards over any blank spaces, tabs or carriage
12817 returns until a printed character is found and then leaves point after
12818 the character. Since point has already been moved to the end of the
12819 pattern that marks the end of the sentence, this action leaves point
12820 right after the closing printed character of the sentence, which is
12823 On the other hand, if the @code{re-search-forward} function fails to
12824 find a pattern marking the end of the sentence, the function returns
12825 false. The false then causes the @code{if} to evaluate its third
12826 argument, which is @code{(goto-char par-end)}: it moves point to the
12827 end of the paragraph.
12829 (And if the text is in a form or equivalent, and point may not move
12830 fully, then the @code{constrain-to-field} function comes into play.)
12832 Regular expression searches are exceptionally useful and the pattern
12833 illustrated by @code{re-search-forward}, in which the search is the
12834 test of an @code{if} expression, is handy. You will see or write code
12835 incorporating this pattern often.
12837 @node forward-paragraph
12838 @section @code{forward-paragraph}: a Goldmine of Functions
12839 @findex forward-paragraph
12843 (defun forward-paragraph (&optional arg)
12844 "Move forward to end of paragraph.
12845 With argument ARG, do it ARG times;
12846 a negative argument ARG = -N means move backward N paragraphs.
12848 A line which `paragraph-start' matches either separates paragraphs
12849 \(if `paragraph-separate' matches it also) or is the first line of a paragraph.
12850 A paragraph end is the beginning of a line which is not part of the paragraph
12851 to which the end of the previous line belongs, or the end of the buffer.
12852 Returns the count of paragraphs left to move."
12854 (or arg (setq arg 1))
12855 (let* ((opoint (point))
12856 (fill-prefix-regexp
12857 (and fill-prefix (not (equal fill-prefix ""))
12858 (not paragraph-ignore-fill-prefix)
12859 (regexp-quote fill-prefix)))
12860 ;; Remove ^ from paragraph-start and paragraph-sep if they are there.
12861 ;; These regexps shouldn't be anchored, because we look for them
12862 ;; starting at the left-margin. This allows paragraph commands to
12863 ;; work normally with indented text.
12864 ;; This hack will not find problem cases like "whatever\\|^something".
12865 (parstart (if (and (not (equal "" paragraph-start))
12866 (equal ?^ (aref paragraph-start 0)))
12867 (substring paragraph-start 1)
12869 (parsep (if (and (not (equal "" paragraph-separate))
12870 (equal ?^ (aref paragraph-separate 0)))
12871 (substring paragraph-separate 1)
12872 paragraph-separate))
12874 (if fill-prefix-regexp
12875 (concat parsep "\\|"
12876 fill-prefix-regexp "[ \t]*$")
12878 ;; This is used for searching.
12879 (sp-parstart (concat "^[ \t]*\\(?:" parstart "\\|" parsep "\\)"))
12881 (while (and (< arg 0) (not (bobp)))
12882 (if (and (not (looking-at parsep))
12883 (re-search-backward "^\n" (max (1- (point)) (point-min)) t)
12884 (looking-at parsep))
12885 (setq arg (1+ arg))
12886 (setq start (point))
12887 ;; Move back over paragraph-separating lines.
12888 (forward-char -1) (beginning-of-line)
12889 (while (and (not (bobp))
12890 (progn (move-to-left-margin)
12891 (looking-at parsep)))
12895 (setq arg (1+ arg))
12896 ;; Go to end of the previous (non-separating) line.
12898 ;; Search back for line that starts or separates paragraphs.
12899 (if (if fill-prefix-regexp
12900 ;; There is a fill prefix; it overrides parstart.
12901 (let (multiple-lines)
12902 (while (and (progn (beginning-of-line) (not (bobp)))
12903 (progn (move-to-left-margin)
12904 (not (looking-at parsep)))
12905 (looking-at fill-prefix-regexp))
12906 (unless (= (point) start)
12907 (setq multiple-lines t))
12909 (move-to-left-margin)
12910 ;; This deleted code caused a long hanging-indent line
12911 ;; not to be filled together with the following lines.
12912 ;; ;; Don't move back over a line before the paragraph
12913 ;; ;; which doesn't start with fill-prefix
12914 ;; ;; unless that is the only line we've moved over.
12915 ;; (and (not (looking-at fill-prefix-regexp))
12917 ;; (forward-line 1))
12919 (while (and (re-search-backward sp-parstart nil 1)
12920 (setq found-start t)
12921 ;; Found a candidate, but need to check if it is a
12923 (progn (setq start (point))
12924 (move-to-left-margin)
12925 (not (looking-at parsep)))
12926 (not (and (looking-at parstart)
12927 (or (not use-hard-newlines)
12930 (1- start) 'hard)))))
12931 (setq found-start nil)
12936 ;; Move forward over paragraph separators.
12937 ;; We know this cannot reach the place we started
12938 ;; because we know we moved back over a non-separator.
12939 (while (and (not (eobp))
12940 (progn (move-to-left-margin)
12941 (looking-at parsep)))
12943 ;; If line before paragraph is just margin, back up to there.
12945 (if (> (current-column) (current-left-margin))
12947 (skip-chars-backward " \t")
12949 (forward-line 1))))
12950 ;; No starter or separator line => use buffer beg.
12951 (goto-char (point-min))))))
12953 (while (and (> arg 0) (not (eobp)))
12954 ;; Move forward over separator lines...
12955 (while (and (not (eobp))
12956 (progn (move-to-left-margin) (not (eobp)))
12957 (looking-at parsep))
12959 (unless (eobp) (setq arg (1- arg)))
12960 ;; ... and one more line.
12962 (if fill-prefix-regexp
12963 ;; There is a fill prefix; it overrides parstart.
12964 (while (and (not (eobp))
12965 (progn (move-to-left-margin) (not (eobp)))
12966 (not (looking-at parsep))
12967 (looking-at fill-prefix-regexp))
12969 (while (and (re-search-forward sp-parstart nil 1)
12970 (progn (setq start (match-beginning 0))
12973 (progn (move-to-left-margin)
12974 (not (looking-at parsep)))
12975 (or (not (looking-at parstart))
12976 (and use-hard-newlines
12977 (not (get-text-property (1- start) 'hard)))))
12979 (if (< (point) (point-max))
12980 (goto-char start))))
12981 (constrain-to-field nil opoint t)
12982 ;; Return the number of steps that could not be done.
12986 The @code{forward-paragraph} function moves point forward to the end
12987 of the paragraph. It is usually bound to @kbd{M-@}} and makes use of a
12988 number of functions that are important in themselves, including
12989 @code{let*}, @code{match-beginning}, and @code{looking-at}.
12991 The function definition for @code{forward-paragraph} is considerably
12992 longer than the function definition for @code{forward-sentence}
12993 because it works with a paragraph, each line of which may begin with a
12996 A fill prefix consists of a string of characters that are repeated at
12997 the beginning of each line. For example, in Lisp code, it is a
12998 convention to start each line of a paragraph-long comment with
12999 @samp{;;; }. In Text mode, four blank spaces make up another common
13000 fill prefix, creating an indented paragraph. (@xref{Fill Prefix, , ,
13001 emacs, The GNU Emacs Manual}, for more information about fill
13004 The existence of a fill prefix means that in addition to being able to
13005 find the end of a paragraph whose lines begin on the left-most
13006 column, the @code{forward-paragraph} function must be able to find the
13007 end of a paragraph when all or many of the lines in the buffer begin
13008 with the fill prefix.
13010 Moreover, it is sometimes practical to ignore a fill prefix that
13011 exists, especially when blank lines separate paragraphs.
13012 This is an added complication.
13015 * forward-paragraph in brief:: Key parts of the function definition.
13016 * fwd-para let:: The @code{let*} expression.
13017 * fwd-para while:: The forward motion @code{while} loop.
13021 @node forward-paragraph in brief
13022 @unnumberedsubsec Shortened @code{forward-paragraph} function definition
13025 Rather than print all of the @code{forward-paragraph} function, we
13026 will only print parts of it. Read without preparation, the function
13030 In outline, the function looks like this:
13034 (defun forward-paragraph (&optional arg)
13035 "@var{documentation}@dots{}"
13037 (or arg (setq arg 1))
13040 (while (and (< arg 0) (not (bobp))) ; @r{backward-moving-code}
13042 (while (and (> arg 0) (not (eobp))) ; @r{forward-moving-code}
13047 The first parts of the function are routine: the function's argument
13048 list consists of one optional argument. Documentation follows.
13050 The lower case @samp{p} in the @code{interactive} declaration means
13051 that the processed prefix argument, if any, is passed to the function.
13052 This will be a number, and is the repeat count of how many paragraphs
13053 point will move. The @code{or} expression in the next line handles
13054 the common case when no argument is passed to the function, which occurs
13055 if the function is called from other code rather than interactively.
13056 This case was described earlier. (@xref{forward-sentence, The
13057 @code{forward-sentence} function}.) Now we reach the end of the
13058 familiar part of this function.
13061 @unnumberedsubsec The @code{let*} expression
13063 The next line of the @code{forward-paragraph} function begins a
13064 @code{let*} expression. This is a different than @code{let}. The
13065 symbol is @code{let*} not @code{let}.
13067 The @code{let*} special form is like @code{let} except that Emacs sets
13068 each variable in sequence, one after another, and variables in the
13069 latter part of the varlist can make use of the values to which Emacs
13070 set variables in the earlier part of the varlist.
13073 ( refappend save-excursion, , code save-excursion in code append-to-buffer .)
13076 (@ref{append save-excursion, , @code{save-excursion} in @code{append-to-buffer}}.)
13078 In the @code{let*} expression in this function, Emacs binds a total of
13079 seven variables: @code{opoint}, @code{fill-prefix-regexp},
13080 @code{parstart}, @code{parsep}, @code{sp-parstart}, @code{start}, and
13081 @code{found-start}.
13083 The variable @code{parsep} appears twice, first, to remove instances
13084 of @samp{^}, and second, to handle fill prefixes.
13086 The variable @code{opoint} is just the value of @code{point}. As you
13087 can guess, it is used in a @code{constrain-to-field} expression, just
13088 as in @code{forward-sentence}.
13090 The variable @code{fill-prefix-regexp} is set to the value returned by
13091 evaluating the following list:
13096 (not (equal fill-prefix ""))
13097 (not paragraph-ignore-fill-prefix)
13098 (regexp-quote fill-prefix))
13103 This is an expression whose first element is the @code{and} special form.
13105 As we learned earlier (@pxref{kill-new function, , The @code{kill-new}
13106 function}), the @code{and} special form evaluates each of its
13107 arguments until one of the arguments returns a value of @code{nil}, in
13108 which case the @code{and} expression returns @code{nil}; however, if
13109 none of the arguments returns a value of @code{nil}, the value
13110 resulting from evaluating the last argument is returned. (Since such
13111 a value is not @code{nil}, it is considered true in Lisp.) In other
13112 words, an @code{and} expression returns a true value only if all its
13113 arguments are true.
13116 In this case, the variable @code{fill-prefix-regexp} is bound to a
13117 non-@code{nil} value only if the following four expressions produce a
13118 true (i.e., a non-@code{nil}) value when they are evaluated; otherwise,
13119 @code{fill-prefix-regexp} is bound to @code{nil}.
13123 When this variable is evaluated, the value of the fill prefix, if any,
13124 is returned. If there is no fill prefix, this variable returns
13127 @item (not (equal fill-prefix "")
13128 This expression checks whether an existing fill prefix is an empty
13129 string, that is, a string with no characters in it. An empty string is
13130 not a useful fill prefix.
13132 @item (not paragraph-ignore-fill-prefix)
13133 This expression returns @code{nil} if the variable
13134 @code{paragraph-ignore-fill-prefix} has been turned on by being set to a
13135 true value such as @code{t}.
13137 @item (regexp-quote fill-prefix)
13138 This is the last argument to the @code{and} special form. If all the
13139 arguments to the @code{and} are true, the value resulting from
13140 evaluating this expression will be returned by the @code{and} expression
13141 and bound to the variable @code{fill-prefix-regexp},
13144 @findex regexp-quote
13146 The result of evaluating this @code{and} expression successfully is that
13147 @code{fill-prefix-regexp} will be bound to the value of
13148 @code{fill-prefix} as modified by the @code{regexp-quote} function.
13149 What @code{regexp-quote} does is read a string and return a regular
13150 expression that will exactly match the string and match nothing else.
13151 This means that @code{fill-prefix-regexp} will be set to a value that
13152 will exactly match the fill prefix if the fill prefix exists.
13153 Otherwise, the variable will be set to @code{nil}.
13155 The next two local variables in the @code{let*} expression are
13156 designed to remove instances of @samp{^} from @code{parstart} and
13157 @code{parsep}, the local variables which indicate the paragraph start
13158 and the paragraph separator. The next expression sets @code{parsep}
13159 again. That is to handle fill prefixes.
13161 This is the setting that requires the definition call @code{let*}
13162 rather than @code{let}. The true-or-false-test for the @code{if}
13163 depends on whether the variable @code{fill-prefix-regexp} evaluates to
13164 @code{nil} or some other value.
13166 If @code{fill-prefix-regexp} does not have a value, Emacs evaluates
13167 the else-part of the @code{if} expression and binds @code{parsep} to
13168 its local value. (@code{parsep} is a regular expression that matches
13169 what separates paragraphs.)
13171 But if @code{fill-prefix-regexp} does have a value, Emacs evaluates
13172 the then-part of the @code{if} expression and binds @code{parsep} to a
13173 regular expression that includes the @code{fill-prefix-regexp} as part
13176 Specifically, @code{parsep} is set to the original value of the
13177 paragraph separate regular expression concatenated with an alternative
13178 expression that consists of the @code{fill-prefix-regexp} followed by
13179 optional whitespace to the end of the line. The whitespace is defined
13180 by @w{@code{"[ \t]*$"}}.) The @samp{\\|} defines this portion of the
13181 regexp as an alternative to @code{parsep}.
13183 According to a comment in the code, the next local variable,
13184 @code{sp-parstart}, is used for searching, and then the final two,
13185 @code{start} and @code{found-start}, are set to @code{nil}.
13187 Now we get into the body of the @code{let*}. The first part of the body
13188 of the @code{let*} deals with the case when the function is given a
13189 negative argument and is therefore moving backwards. We will skip this
13192 @node fwd-para while
13193 @unnumberedsubsec The forward motion @code{while} loop
13195 The second part of the body of the @code{let*} deals with forward
13196 motion. It is a @code{while} loop that repeats itself so long as the
13197 value of @code{arg} is greater than zero. In the most common use of
13198 the function, the value of the argument is 1, so the body of the
13199 @code{while} loop is evaluated exactly once, and the cursor moves
13200 forward one paragraph.
13203 (while (and (> arg 0) (not (eobp)))
13205 ;; Move forward over separator lines...
13206 (while (and (not (eobp))
13207 (progn (move-to-left-margin) (not (eobp)))
13208 (looking-at parsep))
13210 (unless (eobp) (setq arg (1- arg)))
13211 ;; ... and one more line.
13214 (if fill-prefix-regexp
13215 ;; There is a fill prefix; it overrides parstart.
13216 (while (and (not (eobp))
13217 (progn (move-to-left-margin) (not (eobp)))
13218 (not (looking-at parsep))
13219 (looking-at fill-prefix-regexp))
13222 (while (and (re-search-forward sp-parstart nil 1)
13223 (progn (setq start (match-beginning 0))
13226 (progn (move-to-left-margin)
13227 (not (looking-at parsep)))
13228 (or (not (looking-at parstart))
13229 (and use-hard-newlines
13230 (not (get-text-property (1- start) 'hard)))))
13233 (if (< (point) (point-max))
13234 (goto-char start))))
13237 This part handles three situations: when point is between paragraphs,
13238 when there is a fill prefix and when there is no fill prefix.
13241 The @code{while} loop looks like this:
13245 ;; @r{going forwards and not at the end of the buffer}
13246 (while (and (> arg 0) (not (eobp)))
13248 ;; @r{between paragraphs}
13249 ;; Move forward over separator lines...
13250 (while (and (not (eobp))
13251 (progn (move-to-left-margin) (not (eobp)))
13252 (looking-at parsep))
13254 ;; @r{This decrements the loop}
13255 (unless (eobp) (setq arg (1- arg)))
13256 ;; ... and one more line.
13261 (if fill-prefix-regexp
13262 ;; There is a fill prefix; it overrides parstart;
13263 ;; we go forward line by line
13264 (while (and (not (eobp))
13265 (progn (move-to-left-margin) (not (eobp)))
13266 (not (looking-at parsep))
13267 (looking-at fill-prefix-regexp))
13272 ;; There is no fill prefix;
13273 ;; we go forward character by character
13274 (while (and (re-search-forward sp-parstart nil 1)
13275 (progn (setq start (match-beginning 0))
13278 (progn (move-to-left-margin)
13279 (not (looking-at parsep)))
13280 (or (not (looking-at parstart))
13281 (and use-hard-newlines
13282 (not (get-text-property (1- start) 'hard)))))
13287 ;; and if there is no fill prefix and if we are not at the end,
13288 ;; go to whatever was found in the regular expression search
13290 (if (< (point) (point-max))
13291 (goto-char start))))
13296 We can see that this is a decrementing counter @code{while} loop,
13297 using the expression @code{(setq arg (1- arg))} as the decrementer.
13298 That expression is not far from the @code{while}, but is hidden in
13299 another Lisp macro, an @code{unless} macro. Unless we are at the end
13300 of the buffer---that is what the @code{eobp} function determines; it
13301 is an abbreviation of @samp{End Of Buffer P}---we decrease the value
13302 of @code{arg} by one.
13304 (If we are at the end of the buffer, we cannot go forward any more and
13305 the next loop of the @code{while} expression will test false since the
13306 test is an @code{and} with @code{(not (eobp))}. The @code{not}
13307 function means exactly as you expect; it is another name for
13308 @code{null}, a function that returns true when its argument is false.)
13310 Interestingly, the loop count is not decremented until we leave the
13311 space between paragraphs, unless we come to the end of buffer or stop
13312 seeing the local value of the paragraph separator.
13314 That second @code{while} also has a @code{(move-to-left-margin)}
13315 expression. The function is self-explanatory. It is inside a
13316 @code{progn} expression and not the last element of its body, so it is
13317 only invoked for its side effect, which is to move point to the left
13318 margin of the current line.
13321 The @code{looking-at} function is also self-explanatory; it returns
13322 true if the text after point matches the regular expression given as
13325 The rest of the body of the loop looks difficult at first, but makes
13326 sense as you come to understand it.
13329 First consider what happens if there is a fill prefix:
13333 (if fill-prefix-regexp
13334 ;; There is a fill prefix; it overrides parstart;
13335 ;; we go forward line by line
13336 (while (and (not (eobp))
13337 (progn (move-to-left-margin) (not (eobp)))
13338 (not (looking-at parsep))
13339 (looking-at fill-prefix-regexp))
13345 This expression moves point forward line by line so long
13346 as four conditions are true:
13350 Point is not at the end of the buffer.
13353 We can move to the left margin of the text and are
13354 not at the end of the buffer.
13357 The text following point does not separate paragraphs.
13360 The pattern following point is the fill prefix regular expression.
13363 The last condition may be puzzling, until you remember that point was
13364 moved to the beginning of the line early in the @code{forward-paragraph}
13365 function. This means that if the text has a fill prefix, the
13366 @code{looking-at} function will see it.
13369 Consider what happens when there is no fill prefix.
13373 (while (and (re-search-forward sp-parstart nil 1)
13374 (progn (setq start (match-beginning 0))
13377 (progn (move-to-left-margin)
13378 (not (looking-at parsep)))
13379 (or (not (looking-at parstart))
13380 (and use-hard-newlines
13381 (not (get-text-property (1- start) 'hard)))))
13387 This @code{while} loop has us searching forward for
13388 @code{sp-parstart}, which is the combination of possible whitespace
13389 with a the local value of the start of a paragraph or of a paragraph
13390 separator. (The latter two are within an expression starting
13391 @code{\(?:} so that they are not referenced by the
13392 @code{match-beginning} function.)
13395 The two expressions,
13399 (setq start (match-beginning 0))
13405 mean go to the start of the text matched by the regular expression
13408 The @code{(match-beginning 0)} expression is new. It returns a number
13409 specifying the location of the start of the text that was matched by
13412 The @code{match-beginning} function is used here because of a
13413 characteristic of a forward search: a successful forward search,
13414 regardless of whether it is a plain search or a regular expression
13415 search, moves point to the end of the text that is found. In this
13416 case, a successful search moves point to the end of the pattern for
13417 @code{sp-parstart}.
13419 However, we want to put point at the end of the current paragraph, not
13420 somewhere else. Indeed, since the search possibly includes the
13421 paragraph separator, point may end up at the beginning of the next one
13422 unless we use an expression that includes @code{match-beginning}.
13424 @findex match-beginning
13425 When given an argument of 0, @code{match-beginning} returns the
13426 position that is the start of the text matched by the most recent
13427 search. In this case, the most recent search looks for
13428 @code{sp-parstart}. The @code{(match-beginning 0)} expression returns
13429 the beginning position of that pattern, rather than the end position
13432 (Incidentally, when passed a positive number as an argument, the
13433 @code{match-beginning} function returns the location of point at that
13434 parenthesized expression in the last search unless that parenthesized
13435 expression begins with @code{\(?:}. I don't know why @code{\(?:}
13436 appears here since the argument is 0.)
13439 The last expression when there is no fill prefix is
13443 (if (< (point) (point-max))
13444 (goto-char start))))
13449 This says that if there is no fill prefix and if we are not at the
13450 end, point should move to the beginning of whatever was found by the
13451 regular expression search for @code{sp-parstart}.
13453 The full definition for the @code{forward-paragraph} function not only
13454 includes code for going forwards, but also code for going backwards.
13456 If you are reading this inside of GNU Emacs and you want to see the
13457 whole function, you can type @kbd{C-h f} (@code{describe-function})
13458 and the name of the function. This gives you the function
13459 documentation and the name of the library containing the function's
13460 source. Place point over the name of the library and press the RET
13461 key; you will be taken directly to the source. (Be sure to install
13462 your sources! Without them, you are like a person who tries to drive
13463 a car with his eyes shut!)
13466 @section Create Your Own @file{TAGS} File
13468 @cindex @file{TAGS} file, create own
13470 Besides @kbd{C-h f} (@code{describe-function}), another way to see the
13471 source of a function is to type @kbd{M-.} (@code{find-tag}) and the
13472 name of the function when prompted for it. This is a good habit to
13473 get into. The @kbd{M-.} (@code{find-tag}) command takes you directly
13474 to the source for a function, variable, or node. The function depends
13475 on tags tables to tell it where to go.
13477 If the @code{find-tag} function first asks you for the name of a
13478 @file{TAGS} table, give it the name of a @file{TAGS} file such as
13479 @file{/usr/local/src/emacs/src/TAGS}. (The exact path to your
13480 @file{TAGS} file depends on how your copy of Emacs was installed. I
13481 just told you the location that provides both my C and my Emacs Lisp
13484 You can also create your own @file{TAGS} file for directories that
13487 You often need to build and install tags tables yourself. They are
13488 not built automatically. A tags table is called a @file{TAGS} file;
13489 the name is in upper case letters.
13491 You can create a @file{TAGS} file by calling the @code{etags} program
13492 that comes as a part of the Emacs distribution. Usually, @code{etags}
13493 is compiled and installed when Emacs is built. (@code{etags} is not
13494 an Emacs Lisp function or a part of Emacs; it is a C program.)
13497 To create a @file{TAGS} file, first switch to the directory in which
13498 you want to create the file. In Emacs you can do this with the
13499 @kbd{M-x cd} command, or by visiting a file in the directory, or by
13500 listing the directory with @kbd{C-x d} (@code{dired}). Then run the
13501 compile command, with @w{@code{etags *.el}} as the command to execute
13504 M-x compile RET etags *.el RET
13508 to create a @file{TAGS} file for Emacs Lisp.
13510 For example, if you have a large number of files in your
13511 @file{~/emacs} directory, as I do---I have 137 @file{.el} files in it,
13512 of which I load 12---you can create a @file{TAGS} file for the Emacs
13513 Lisp files in that directory.
13516 The @code{etags} program takes all the usual shell `wildcards'. For
13517 example, if you have two directories for which you want a single
13518 @file{TAGS} file, type @w{@code{etags *.el ../elisp/*.el}}, where
13519 @file{../elisp/} is the second directory:
13522 M-x compile RET etags *.el ../elisp/*.el RET
13529 M-x compile RET etags --help RET
13533 to see a list of the options accepted by @code{etags} as well as a
13534 list of supported languages.
13536 The @code{etags} program handles more than 20 languages, including
13537 Emacs Lisp, Common Lisp, Scheme, C, C++, Ada, Fortran, HTML, Java,
13538 LaTeX, Pascal, Perl, PostScript, Python, TeX, Texinfo, makefiles, and
13539 most assemblers. The program has no switches for specifying the
13540 language; it recognizes the language in an input file according to its
13541 file name and contents.
13543 @file{etags} is very helpful when you are writing code yourself and
13544 want to refer back to functions you have already written. Just run
13545 @code{etags} again at intervals as you write new functions, so they
13546 become part of the @file{TAGS} file.
13548 If you think an appropriate @file{TAGS} file already exists for what
13549 you want, but do not know where it is, you can use the @code{locate}
13550 program to attempt to find it.
13552 Type @w{@kbd{M-x locate @key{RET} TAGS @key{RET}}} and Emacs will list
13553 for you the full path names of all your @file{TAGS} files. On my
13554 system, this command lists 34 @file{TAGS} files. On the other hand, a
13555 `plain vanilla' system I recently installed did not contain any
13558 If the tags table you want has been created, you can use the @code{M-x
13559 visit-tags-table} command to specify it. Otherwise, you will need to
13560 create the tag table yourself and then use @code{M-x
13563 @subsubheading Building Tags in the Emacs sources
13564 @cindex Building Tags in the Emacs sources
13565 @cindex Tags in the Emacs sources
13568 The GNU Emacs sources come with a @file{Makefile} that contains a
13569 sophisticated @code{etags} command that creates, collects, and merges
13570 tags tables from all over the Emacs sources and puts the information
13571 into one @file{TAGS} file in the @file{src/} directory. (The
13572 @file{src/} directory is below the top level of your Emacs directory.)
13575 To build this @file{TAGS} file, go to the top level of your Emacs
13576 source directory and run the compile command @code{make tags}:
13579 M-x compile RET make tags RET
13583 (The @code{make tags} command works well with the GNU Emacs sources,
13584 as well as with some other source packages.)
13586 For more information, see @ref{Tags, , Tag Tables, emacs, The GNU Emacs
13589 @node Regexp Review
13592 Here is a brief summary of some recently introduced functions.
13596 Repeatedly evaluate the body of the expression so long as the first
13597 element of the body tests true. Then return @code{nil}. (The
13598 expression is evaluated only for its side effects.)
13607 (insert (format "foo is %d.\n" foo))
13608 (setq foo (1- foo))))
13610 @result{} foo is 2.
13617 (The @code{insert} function inserts its arguments at point; the
13618 @code{format} function returns a string formatted from its arguments
13619 the way @code{message} formats its arguments; @code{\n} produces a new
13622 @item re-search-forward
13623 Search for a pattern, and if the pattern is found, move point to rest
13627 Takes four arguments, like @code{search-forward}:
13631 A regular expression that specifies the pattern to search for.
13632 (Remember to put quotation marks around this argument!)
13635 Optionally, the limit of the search.
13638 Optionally, what to do if the search fails, return @code{nil} or an
13642 Optionally, how many times to repeat the search; if negative, the
13643 search goes backwards.
13647 Bind some variables locally to particular values,
13648 and then evaluate the remaining arguments, returning the value of the
13649 last one. While binding the local variables, use the local values of
13650 variables bound earlier, if any.
13659 (message "`bar' is %d." bar))
13660 @result{} `bar' is 21.
13664 @item match-beginning
13665 Return the position of the start of the text found by the last regular
13669 Return @code{t} for true if the text after point matches the argument,
13670 which should be a regular expression.
13673 Return @code{t} for true if point is at the end of the accessible part
13674 of a buffer. The end of the accessible part is the end of the buffer
13675 if the buffer is not narrowed; it is the end of the narrowed part if
13676 the buffer is narrowed.
13680 @node re-search Exercises
13681 @section Exercises with @code{re-search-forward}
13685 Write a function to search for a regular expression that matches two
13686 or more blank lines in sequence.
13689 Write a function to search for duplicated words, such as `the the'.
13690 @xref{Regexps, , Syntax of Regular Expressions, emacs, The GNU Emacs
13691 Manual}, for information on how to write a regexp (a regular
13692 expression) to match a string that is composed of two identical
13693 halves. You can devise several regexps; some are better than others.
13694 The function I use is described in an appendix, along with several
13695 regexps. @xref{the-the, , @code{the-the} Duplicated Words Function}.
13698 @node Counting Words
13699 @chapter Counting: Repetition and Regexps
13700 @cindex Repetition for word counting
13701 @cindex Regular expressions for word counting
13703 Repetition and regular expression searches are powerful tools that you
13704 often use when you write code in Emacs Lisp. This chapter illustrates
13705 the use of regular expression searches through the construction of
13706 word count commands using @code{while} loops and recursion.
13709 * Why Count Words::
13710 * @value{COUNT-WORDS}:: Use a regexp, but find a problem.
13711 * recursive-count-words:: Start with case of no words in region.
13712 * Counting Exercise::
13716 @node Why Count Words
13717 @unnumberedsec Counting words
13720 The standard Emacs distribution contains functions for counting the
13721 number of lines and words within a region.
13723 Certain types of writing ask you to count words. Thus, if you write
13724 an essay, you may be limited to 800 words; if you write a novel, you
13725 may discipline yourself to write 1000 words a day. It seems odd, but
13726 for a long time, Emacs lacked a word count command. Perhaps people used
13727 Emacs mostly for code or types of documentation that did not require
13728 word counts; or perhaps they restricted themselves to the operating
13729 system word count command, @code{wc}. Alternatively, people may have
13730 followed the publishers' convention and computed a word count by
13731 dividing the number of characters in a document by five.
13733 There are many ways to implement a command to count words. Here are
13734 some examples, which you may wish to compare with the standard Emacs
13735 command, @code{count-words-region}.
13737 @node @value{COUNT-WORDS}
13738 @section The @code{@value{COUNT-WORDS}} Function
13739 @findex @value{COUNT-WORDS}
13741 A word count command could count words in a line, paragraph, region,
13742 or buffer. What should the command cover? You could design the
13743 command to count the number of words in a complete buffer. However,
13744 the Emacs tradition encourages flexibility---you may want to count
13745 words in just a section, rather than all of a buffer. So it makes
13746 more sense to design the command to count the number of words in a
13747 region. Once you have a command to count words in a region, you can,
13748 if you wish, count words in a whole buffer by marking it with
13749 @w{@kbd{C-x h}} (@code{mark-whole-buffer}).
13751 Clearly, counting words is a repetitive act: starting from the
13752 beginning of the region, you count the first word, then the second
13753 word, then the third word, and so on, until you reach the end of the
13754 region. This means that word counting is ideally suited to recursion
13755 or to a @code{while} loop.
13758 * Design @value{COUNT-WORDS}:: The definition using a @code{while} loop.
13759 * Whitespace Bug:: The Whitespace Bug in @code{@value{COUNT-WORDS}}.
13763 @node Design @value{COUNT-WORDS}
13764 @unnumberedsubsec Designing @code{@value{COUNT-WORDS}}
13767 First, we will implement the word count command with a @code{while}
13768 loop, then with recursion. The command will, of course, be
13772 The template for an interactive function definition is, as always:
13776 (defun @var{name-of-function} (@var{argument-list})
13777 "@var{documentation}@dots{}"
13778 (@var{interactive-expression}@dots{})
13783 What we need to do is fill in the slots.
13785 The name of the function should be self-explanatory and similar to the
13786 existing @code{count-lines-region} name. This makes the name easier
13787 to remember. @code{count-words-region} is the obvious choice. Since
13788 that name is now used for the standard Emacs command to count words, we
13789 will name our implementation @code{@value{COUNT-WORDS}}.
13791 The function counts words within a region. This means that the
13792 argument list must contain symbols that are bound to the two
13793 positions, the beginning and end of the region. These two positions
13794 can be called @samp{beginning} and @samp{end} respectively. The first
13795 line of the documentation should be a single sentence, since that is
13796 all that is printed as documentation by a command such as
13797 @code{apropos}. The interactive expression will be of the form
13798 @samp{(interactive "r")}, since that will cause Emacs to pass the
13799 beginning and end of the region to the function's argument list. All
13802 The body of the function needs to be written to do three tasks:
13803 first, to set up conditions under which the @code{while} loop can
13804 count words, second, to run the @code{while} loop, and third, to send
13805 a message to the user.
13807 When a user calls @code{@value{COUNT-WORDS}}, point may be at the
13808 beginning or the end of the region. However, the counting process
13809 must start at the beginning of the region. This means we will want
13810 to put point there if it is not already there. Executing
13811 @code{(goto-char beginning)} ensures this. Of course, we will want to
13812 return point to its expected position when the function finishes its
13813 work. For this reason, the body must be enclosed in a
13814 @code{save-excursion} expression.
13816 The central part of the body of the function consists of a
13817 @code{while} loop in which one expression jumps point forward word by
13818 word, and another expression counts those jumps. The true-or-false-test
13819 of the @code{while} loop should test true so long as point should jump
13820 forward, and false when point is at the end of the region.
13822 We could use @code{(forward-word 1)} as the expression for moving point
13823 forward word by word, but it is easier to see what Emacs identifies as a
13824 `word' if we use a regular expression search.
13826 A regular expression search that finds the pattern for which it is
13827 searching leaves point after the last character matched. This means
13828 that a succession of successful word searches will move point forward
13831 As a practical matter, we want the regular expression search to jump
13832 over whitespace and punctuation between words as well as over the
13833 words themselves. A regexp that refuses to jump over interword
13834 whitespace would never jump more than one word! This means that
13835 the regexp should include the whitespace and punctuation that follows
13836 a word, if any, as well as the word itself. (A word may end a buffer
13837 and not have any following whitespace or punctuation, so that part of
13838 the regexp must be optional.)
13840 Thus, what we want for the regexp is a pattern defining one or more
13841 word constituent characters followed, optionally, by one or more
13842 characters that are not word constituents. The regular expression for
13850 The buffer's syntax table determines which characters are and are not
13851 word constituents. For more information about syntax,
13852 @pxref{Syntax Tables, , Syntax Tables, elisp, The GNU Emacs Lisp
13856 The search expression looks like this:
13859 (re-search-forward "\\w+\\W*")
13863 (Note that paired backslashes precede the @samp{w} and @samp{W}. A
13864 single backslash has special meaning to the Emacs Lisp interpreter.
13865 It indicates that the following character is interpreted differently
13866 than usual. For example, the two characters, @samp{\n}, stand for
13867 @samp{newline}, rather than for a backslash followed by @samp{n}. Two
13868 backslashes in a row stand for an ordinary, `unspecial' backslash, so
13869 Emacs Lisp interpreter ends of seeing a single backslash followed by a
13870 letter. So it discovers the letter is special.)
13872 We need a counter to count how many words there are; this variable
13873 must first be set to 0 and then incremented each time Emacs goes
13874 around the @code{while} loop. The incrementing expression is simply:
13877 (setq count (1+ count))
13880 Finally, we want to tell the user how many words there are in the
13881 region. The @code{message} function is intended for presenting this
13882 kind of information to the user. The message has to be phrased so
13883 that it reads properly regardless of how many words there are in the
13884 region: we don't want to say that ``there are 1 words in the region''.
13885 The conflict between singular and plural is ungrammatical. We can
13886 solve this problem by using a conditional expression that evaluates
13887 different messages depending on the number of words in the region.
13888 There are three possibilities: no words in the region, one word in the
13889 region, and more than one word. This means that the @code{cond}
13890 special form is appropriate.
13893 All this leads to the following function definition:
13897 ;;; @r{First version; has bugs!}
13898 (defun @value{COUNT-WORDS} (beginning end)
13899 "Print number of words in the region.
13900 Words are defined as at least one word-constituent
13901 character followed by at least one character that
13902 is not a word-constituent. The buffer's syntax
13903 table determines which characters these are."
13905 (message "Counting words in region ... ")
13909 ;;; @r{1. Set up appropriate conditions.}
13911 (goto-char beginning)
13916 ;;; @r{2. Run the} while @r{loop.}
13917 (while (< (point) end)
13918 (re-search-forward "\\w+\\W*")
13919 (setq count (1+ count)))
13923 ;;; @r{3. Send a message to the user.}
13924 (cond ((zerop count)
13926 "The region does NOT have any words."))
13929 "The region has 1 word."))
13932 "The region has %d words." count))))))
13937 As written, the function works, but not in all circumstances.
13939 @node Whitespace Bug
13940 @subsection The Whitespace Bug in @code{@value{COUNT-WORDS}}
13942 The @code{@value{COUNT-WORDS}} command described in the preceding
13943 section has two bugs, or rather, one bug with two manifestations.
13944 First, if you mark a region containing only whitespace in the middle
13945 of some text, the @code{@value{COUNT-WORDS}} command tells you that the
13946 region contains one word! Second, if you mark a region containing
13947 only whitespace at the end of the buffer or the accessible portion of
13948 a narrowed buffer, the command displays an error message that looks
13952 Search failed: "\\w+\\W*"
13955 If you are reading this in Info in GNU Emacs, you can test for these
13958 First, evaluate the function in the usual manner to install it.
13960 Here is a copy of the definition. Place your cursor after the closing
13961 parenthesis and type @kbd{C-x C-e} to install it.
13965 ;; @r{First version; has bugs!}
13966 (defun @value{COUNT-WORDS} (beginning end)
13967 "Print number of words in the region.
13968 Words are defined as at least one word-constituent character followed
13969 by at least one character that is not a word-constituent. The buffer's
13970 syntax table determines which characters these are."
13974 (message "Counting words in region ... ")
13978 ;;; @r{1. Set up appropriate conditions.}
13980 (goto-char beginning)
13985 ;;; @r{2. Run the} while @r{loop.}
13986 (while (< (point) end)
13987 (re-search-forward "\\w+\\W*")
13988 (setq count (1+ count)))
13992 ;;; @r{3. Send a message to the user.}
13993 (cond ((zerop count)
13994 (message "The region does NOT have any words."))
13995 ((= 1 count) (message "The region has 1 word."))
13996 (t (message "The region has %d words." count))))))
14002 If you wish, you can also install this keybinding by evaluating it:
14005 (global-set-key "\C-c=" '@value{COUNT-WORDS})
14008 To conduct the first test, set mark and point to the beginning and end
14009 of the following line and then type @kbd{C-c =} (or @kbd{M-x
14010 @value{COUNT-WORDS}} if you have not bound @kbd{C-c =}):
14017 Emacs will tell you, correctly, that the region has three words.
14019 Repeat the test, but place mark at the beginning of the line and place
14020 point just @emph{before} the word @samp{one}. Again type the command
14021 @kbd{C-c =} (or @kbd{M-x @value{COUNT-WORDS}}). Emacs should tell you
14022 that the region has no words, since it is composed only of the
14023 whitespace at the beginning of the line. But instead Emacs tells you
14024 that the region has one word!
14026 For the third test, copy the sample line to the end of the
14027 @file{*scratch*} buffer and then type several spaces at the end of the
14028 line. Place mark right after the word @samp{three} and point at the
14029 end of line. (The end of the line will be the end of the buffer.)
14030 Type @kbd{C-c =} (or @kbd{M-x @value{COUNT-WORDS}}) as you did before.
14031 Again, Emacs should tell you that the region has no words, since it is
14032 composed only of the whitespace at the end of the line. Instead,
14033 Emacs displays an error message saying @samp{Search failed}.
14035 The two bugs stem from the same problem.
14037 Consider the first manifestation of the bug, in which the command
14038 tells you that the whitespace at the beginning of the line contains
14039 one word. What happens is this: The @code{M-x @value{COUNT-WORDS}}
14040 command moves point to the beginning of the region. The @code{while}
14041 tests whether the value of point is smaller than the value of
14042 @code{end}, which it is. Consequently, the regular expression search
14043 looks for and finds the first word. It leaves point after the word.
14044 @code{count} is set to one. The @code{while} loop repeats; but this
14045 time the value of point is larger than the value of @code{end}, the
14046 loop is exited; and the function displays a message saying the number
14047 of words in the region is one. In brief, the regular expression
14048 search looks for and finds the word even though it is outside
14051 In the second manifestation of the bug, the region is whitespace at
14052 the end of the buffer. Emacs says @samp{Search failed}. What happens
14053 is that the true-or-false-test in the @code{while} loop tests true, so
14054 the search expression is executed. But since there are no more words
14055 in the buffer, the search fails.
14057 In both manifestations of the bug, the search extends or attempts to
14058 extend outside of the region.
14060 The solution is to limit the search to the region---this is a fairly
14061 simple action, but as you may have come to expect, it is not quite as
14062 simple as you might think.
14064 As we have seen, the @code{re-search-forward} function takes a search
14065 pattern as its first argument. But in addition to this first,
14066 mandatory argument, it accepts three optional arguments. The optional
14067 second argument bounds the search. The optional third argument, if
14068 @code{t}, causes the function to return @code{nil} rather than signal
14069 an error if the search fails. The optional fourth argument is a
14070 repeat count. (In Emacs, you can see a function's documentation by
14071 typing @kbd{C-h f}, the name of the function, and then @key{RET}.)
14073 In the @code{@value{COUNT-WORDS}} definition, the value of the end of
14074 the region is held by the variable @code{end} which is passed as an
14075 argument to the function. Thus, we can add @code{end} as an argument
14076 to the regular expression search expression:
14079 (re-search-forward "\\w+\\W*" end)
14082 However, if you make only this change to the @code{@value{COUNT-WORDS}}
14083 definition and then test the new version of the definition on a
14084 stretch of whitespace, you will receive an error message saying
14085 @samp{Search failed}.
14087 What happens is this: the search is limited to the region, and fails
14088 as you expect because there are no word-constituent characters in the
14089 region. Since it fails, we receive an error message. But we do not
14090 want to receive an error message in this case; we want to receive the
14091 message that "The region does NOT have any words."
14093 The solution to this problem is to provide @code{re-search-forward}
14094 with a third argument of @code{t}, which causes the function to return
14095 @code{nil} rather than signal an error if the search fails.
14097 However, if you make this change and try it, you will see the message
14098 ``Counting words in region ... '' and @dots{} you will keep on seeing
14099 that message @dots{}, until you type @kbd{C-g} (@code{keyboard-quit}).
14101 Here is what happens: the search is limited to the region, as before,
14102 and it fails because there are no word-constituent characters in the
14103 region, as expected. Consequently, the @code{re-search-forward}
14104 expression returns @code{nil}. It does nothing else. In particular,
14105 it does not move point, which it does as a side effect if it finds the
14106 search target. After the @code{re-search-forward} expression returns
14107 @code{nil}, the next expression in the @code{while} loop is evaluated.
14108 This expression increments the count. Then the loop repeats. The
14109 true-or-false-test tests true because the value of point is still less
14110 than the value of end, since the @code{re-search-forward} expression
14111 did not move point. @dots{} and the cycle repeats @dots{}
14113 The @code{@value{COUNT-WORDS}} definition requires yet another
14114 modification, to cause the true-or-false-test of the @code{while} loop
14115 to test false if the search fails. Put another way, there are two
14116 conditions that must be satisfied in the true-or-false-test before the
14117 word count variable is incremented: point must still be within the
14118 region and the search expression must have found a word to count.
14120 Since both the first condition and the second condition must be true
14121 together, the two expressions, the region test and the search
14122 expression, can be joined with an @code{and} special form and embedded in
14123 the @code{while} loop as the true-or-false-test, like this:
14126 (and (< (point) end) (re-search-forward "\\w+\\W*" end t))
14129 @c colon in printed section title causes problem in Info cross reference
14130 @c also trouble with an overfull hbox
14133 (For information about @code{and}, see
14134 @ref{kill-new function, , The @code{kill-new} function}.)
14138 (@xref{kill-new function, , The @code{kill-new} function}, for
14139 information about @code{and}.)
14142 The @code{re-search-forward} expression returns @code{t} if the search
14143 succeeds and as a side effect moves point. Consequently, as words are
14144 found, point is moved through the region. When the search expression
14145 fails to find another word, or when point reaches the end of the
14146 region, the true-or-false-test tests false, the @code{while} loop
14147 exits, and the @code{@value{COUNT-WORDS}} function displays one or
14148 other of its messages.
14150 After incorporating these final changes, the @code{@value{COUNT-WORDS}}
14151 works without bugs (or at least, without bugs that I have found!).
14152 Here is what it looks like:
14156 ;;; @r{Final version:} @code{while}
14157 (defun @value{COUNT-WORDS} (beginning end)
14158 "Print number of words in the region."
14160 (message "Counting words in region ... ")
14164 ;;; @r{1. Set up appropriate conditions.}
14167 (goto-char beginning)
14171 ;;; @r{2. Run the} while @r{loop.}
14172 (while (and (< (point) end)
14173 (re-search-forward "\\w+\\W*" end t))
14174 (setq count (1+ count)))
14178 ;;; @r{3. Send a message to the user.}
14179 (cond ((zerop count)
14181 "The region does NOT have any words."))
14184 "The region has 1 word."))
14187 "The region has %d words." count))))))
14191 @node recursive-count-words
14192 @section Count Words Recursively
14193 @cindex Count words recursively
14194 @cindex Recursively counting words
14195 @cindex Words, counted recursively
14197 You can write the function for counting words recursively as well as
14198 with a @code{while} loop. Let's see how this is done.
14200 First, we need to recognize that the @code{@value{COUNT-WORDS}}
14201 function has three jobs: it sets up the appropriate conditions for
14202 counting to occur; it counts the words in the region; and it sends a
14203 message to the user telling how many words there are.
14205 If we write a single recursive function to do everything, we will
14206 receive a message for every recursive call. If the region contains 13
14207 words, we will receive thirteen messages, one right after the other.
14208 We don't want this! Instead, we must write two functions to do the
14209 job, one of which (the recursive function) will be used inside of the
14210 other. One function will set up the conditions and display the
14211 message; the other will return the word count.
14213 Let us start with the function that causes the message to be displayed.
14214 We can continue to call this @code{@value{COUNT-WORDS}}.
14216 This is the function that the user will call. It will be interactive.
14217 Indeed, it will be similar to our previous versions of this
14218 function, except that it will call @code{recursive-count-words} to
14219 determine how many words are in the region.
14222 We can readily construct a template for this function, based on our
14227 ;; @r{Recursive version; uses regular expression search}
14228 (defun @value{COUNT-WORDS} (beginning end)
14229 "@var{documentation}@dots{}"
14230 (@var{interactive-expression}@dots{})
14234 ;;; @r{1. Set up appropriate conditions.}
14235 (@var{explanatory message})
14236 (@var{set-up functions}@dots{}
14240 ;;; @r{2. Count the words.}
14241 @var{recursive call}
14245 ;;; @r{3. Send a message to the user.}
14246 @var{message providing word count}))
14250 The definition looks straightforward, except that somehow the count
14251 returned by the recursive call must be passed to the message
14252 displaying the word count. A little thought suggests that this can be
14253 done by making use of a @code{let} expression: we can bind a variable
14254 in the varlist of a @code{let} expression to the number of words in
14255 the region, as returned by the recursive call; and then the
14256 @code{cond} expression, using binding, can display the value to the
14259 Often, one thinks of the binding within a @code{let} expression as
14260 somehow secondary to the `primary' work of a function. But in this
14261 case, what you might consider the `primary' job of the function,
14262 counting words, is done within the @code{let} expression.
14265 Using @code{let}, the function definition looks like this:
14269 (defun @value{COUNT-WORDS} (beginning end)
14270 "Print number of words in the region."
14275 ;;; @r{1. Set up appropriate conditions.}
14276 (message "Counting words in region ... ")
14278 (goto-char beginning)
14282 ;;; @r{2. Count the words.}
14283 (let ((count (recursive-count-words end)))
14287 ;;; @r{3. Send a message to the user.}
14288 (cond ((zerop count)
14290 "The region does NOT have any words."))
14293 "The region has 1 word."))
14296 "The region has %d words." count))))))
14300 Next, we need to write the recursive counting function.
14302 A recursive function has at least three parts: the `do-again-test', the
14303 `next-step-expression', and the recursive call.
14305 The do-again-test determines whether the function will or will not be
14306 called again. Since we are counting words in a region and can use a
14307 function that moves point forward for every word, the do-again-test
14308 can check whether point is still within the region. The do-again-test
14309 should find the value of point and determine whether point is before,
14310 at, or after the value of the end of the region. We can use the
14311 @code{point} function to locate point. Clearly, we must pass the
14312 value of the end of the region to the recursive counting function as an
14315 In addition, the do-again-test should also test whether the search finds a
14316 word. If it does not, the function should not call itself again.
14318 The next-step-expression changes a value so that when the recursive
14319 function is supposed to stop calling itself, it stops. More
14320 precisely, the next-step-expression changes a value so that at the
14321 right time, the do-again-test stops the recursive function from
14322 calling itself again. In this case, the next-step-expression can be
14323 the expression that moves point forward, word by word.
14325 The third part of a recursive function is the recursive call.
14327 Somewhere, also, we also need a part that does the `work' of the
14328 function, a part that does the counting. A vital part!
14331 But already, we have an outline of the recursive counting function:
14335 (defun recursive-count-words (region-end)
14336 "@var{documentation}@dots{}"
14337 @var{do-again-test}
14338 @var{next-step-expression}
14339 @var{recursive call})
14343 Now we need to fill in the slots. Let's start with the simplest cases
14344 first: if point is at or beyond the end of the region, there cannot
14345 be any words in the region, so the function should return zero.
14346 Likewise, if the search fails, there are no words to count, so the
14347 function should return zero.
14349 On the other hand, if point is within the region and the search
14350 succeeds, the function should call itself again.
14353 Thus, the do-again-test should look like this:
14357 (and (< (point) region-end)
14358 (re-search-forward "\\w+\\W*" region-end t))
14362 Note that the search expression is part of the do-again-test---the
14363 function returns @code{t} if its search succeeds and @code{nil} if it
14364 fails. (@xref{Whitespace Bug, , The Whitespace Bug in
14365 @code{@value{COUNT-WORDS}}}, for an explanation of how
14366 @code{re-search-forward} works.)
14368 The do-again-test is the true-or-false test of an @code{if} clause.
14369 Clearly, if the do-again-test succeeds, the then-part of the @code{if}
14370 clause should call the function again; but if it fails, the else-part
14371 should return zero since either point is outside the region or the
14372 search failed because there were no words to find.
14374 But before considering the recursive call, we need to consider the
14375 next-step-expression. What is it? Interestingly, it is the search
14376 part of the do-again-test.
14378 In addition to returning @code{t} or @code{nil} for the
14379 do-again-test, @code{re-search-forward} moves point forward as a side
14380 effect of a successful search. This is the action that changes the
14381 value of point so that the recursive function stops calling itself
14382 when point completes its movement through the region. Consequently,
14383 the @code{re-search-forward} expression is the next-step-expression.
14386 In outline, then, the body of the @code{recursive-count-words}
14387 function looks like this:
14391 (if @var{do-again-test-and-next-step-combined}
14393 @var{recursive-call-returning-count}
14399 How to incorporate the mechanism that counts?
14401 If you are not used to writing recursive functions, a question like
14402 this can be troublesome. But it can and should be approached
14405 We know that the counting mechanism should be associated in some way
14406 with the recursive call. Indeed, since the next-step-expression moves
14407 point forward by one word, and since a recursive call is made for
14408 each word, the counting mechanism must be an expression that adds one
14409 to the value returned by a call to @code{recursive-count-words}.
14412 Consider several cases:
14416 If there are two words in the region, the function should return
14417 a value resulting from adding one to the value returned when it counts
14418 the first word, plus the number returned when it counts the remaining
14419 words in the region, which in this case is one.
14422 If there is one word in the region, the function should return
14423 a value resulting from adding one to the value returned when it counts
14424 that word, plus the number returned when it counts the remaining
14425 words in the region, which in this case is zero.
14428 If there are no words in the region, the function should return zero.
14431 From the sketch we can see that the else-part of the @code{if} returns
14432 zero for the case of no words. This means that the then-part of the
14433 @code{if} must return a value resulting from adding one to the value
14434 returned from a count of the remaining words.
14437 The expression will look like this, where @code{1+} is a function that
14438 adds one to its argument.
14441 (1+ (recursive-count-words region-end))
14445 The whole @code{recursive-count-words} function will then look like
14450 (defun recursive-count-words (region-end)
14451 "@var{documentation}@dots{}"
14453 ;;; @r{1. do-again-test}
14454 (if (and (< (point) region-end)
14455 (re-search-forward "\\w+\\W*" region-end t))
14459 ;;; @r{2. then-part: the recursive call}
14460 (1+ (recursive-count-words region-end))
14462 ;;; @r{3. else-part}
14468 Let's examine how this works:
14470 If there are no words in the region, the else part of the @code{if}
14471 expression is evaluated and consequently the function returns zero.
14473 If there is one word in the region, the value of point is less than
14474 the value of @code{region-end} and the search succeeds. In this case,
14475 the true-or-false-test of the @code{if} expression tests true, and the
14476 then-part of the @code{if} expression is evaluated. The counting
14477 expression is evaluated. This expression returns a value (which will
14478 be the value returned by the whole function) that is the sum of one
14479 added to the value returned by a recursive call.
14481 Meanwhile, the next-step-expression has caused point to jump over the
14482 first (and in this case only) word in the region. This means that
14483 when @code{(recursive-count-words region-end)} is evaluated a second
14484 time, as a result of the recursive call, the value of point will be
14485 equal to or greater than the value of region end. So this time,
14486 @code{recursive-count-words} will return zero. The zero will be added
14487 to one, and the original evaluation of @code{recursive-count-words}
14488 will return one plus zero, which is one, which is the correct amount.
14490 Clearly, if there are two words in the region, the first call to
14491 @code{recursive-count-words} returns one added to the value returned
14492 by calling @code{recursive-count-words} on a region containing the
14493 remaining word---that is, it adds one to one, producing two, which is
14494 the correct amount.
14496 Similarly, if there are three words in the region, the first call to
14497 @code{recursive-count-words} returns one added to the value returned
14498 by calling @code{recursive-count-words} on a region containing the
14499 remaining two words---and so on and so on.
14503 With full documentation the two functions look like this:
14507 The recursive function:
14509 @findex recursive-count-words
14512 (defun recursive-count-words (region-end)
14513 "Number of words between point and REGION-END."
14517 ;;; @r{1. do-again-test}
14518 (if (and (< (point) region-end)
14519 (re-search-forward "\\w+\\W*" region-end t))
14523 ;;; @r{2. then-part: the recursive call}
14524 (1+ (recursive-count-words region-end))
14526 ;;; @r{3. else-part}
14537 ;;; @r{Recursive version}
14538 (defun @value{COUNT-WORDS} (beginning end)
14539 "Print number of words in the region.
14543 Words are defined as at least one word-constituent
14544 character followed by at least one character that is
14545 not a word-constituent. The buffer's syntax table
14546 determines which characters these are."
14550 (message "Counting words in region ... ")
14552 (goto-char beginning)
14553 (let ((count (recursive-count-words end)))
14556 (cond ((zerop count)
14558 "The region does NOT have any words."))
14562 (message "The region has 1 word."))
14565 "The region has %d words." count))))))
14569 @node Counting Exercise
14570 @section Exercise: Counting Punctuation
14572 Using a @code{while} loop, write a function to count the number of
14573 punctuation marks in a region---period, comma, semicolon, colon,
14574 exclamation mark, and question mark. Do the same using recursion.
14576 @node Words in a defun
14577 @chapter Counting Words in a @code{defun}
14578 @cindex Counting words in a @code{defun}
14579 @cindex Word counting in a @code{defun}
14581 Our next project is to count the number of words in a function
14582 definition. Clearly, this can be done using some variant of
14583 @code{@value{COUNT-WORDS}}. @xref{Counting Words, , Counting Words:
14584 Repetition and Regexps}. If we are just going to count the words in
14585 one definition, it is easy enough to mark the definition with the
14586 @kbd{C-M-h} (@code{mark-defun}) command, and then call
14587 @code{@value{COUNT-WORDS}}.
14589 However, I am more ambitious: I want to count the words and symbols in
14590 every definition in the Emacs sources and then print a graph that
14591 shows how many functions there are of each length: how many contain 40
14592 to 49 words or symbols, how many contain 50 to 59 words or symbols,
14593 and so on. I have often been curious how long a typical function is,
14594 and this will tell.
14597 * Divide and Conquer::
14598 * Words and Symbols:: What to count?
14599 * Syntax:: What constitutes a word or symbol?
14600 * count-words-in-defun:: Very like @code{@value{COUNT-WORDS}}.
14601 * Several defuns:: Counting several defuns in a file.
14602 * Find a File:: Do you want to look at a file?
14603 * lengths-list-file:: A list of the lengths of many definitions.
14604 * Several files:: Counting in definitions in different files.
14605 * Several files recursively:: Recursively counting in different files.
14606 * Prepare the data:: Prepare the data for display in a graph.
14610 @node Divide and Conquer
14611 @unnumberedsec Divide and Conquer
14614 Described in one phrase, the histogram project is daunting; but
14615 divided into numerous small steps, each of which we can take one at a
14616 time, the project becomes less fearsome. Let us consider what the
14621 First, write a function to count the words in one definition. This
14622 includes the problem of handling symbols as well as words.
14625 Second, write a function to list the numbers of words in each function
14626 in a file. This function can use the @code{count-words-in-defun}
14630 Third, write a function to list the numbers of words in each function
14631 in each of several files. This entails automatically finding the
14632 various files, switching to them, and counting the words in the
14633 definitions within them.
14636 Fourth, write a function to convert the list of numbers that we
14637 created in step three to a form that will be suitable for printing as
14641 Fifth, write a function to print the results as a graph.
14644 This is quite a project! But if we take each step slowly, it will not
14647 @node Words and Symbols
14648 @section What to Count?
14649 @cindex Words and symbols in defun
14651 When we first start thinking about how to count the words in a
14652 function definition, the first question is (or ought to be) what are
14653 we going to count? When we speak of `words' with respect to a Lisp
14654 function definition, we are actually speaking, in large part, of
14655 `symbols'. For example, the following @code{multiply-by-seven}
14656 function contains the five symbols @code{defun},
14657 @code{multiply-by-seven}, @code{number}, @code{*}, and @code{7}. In
14658 addition, in the documentation string, it contains the four words
14659 @samp{Multiply}, @samp{NUMBER}, @samp{by}, and @samp{seven}. The
14660 symbol @samp{number} is repeated, so the definition contains a total
14661 of ten words and symbols.
14665 (defun multiply-by-seven (number)
14666 "Multiply NUMBER by seven."
14672 However, if we mark the @code{multiply-by-seven} definition with
14673 @kbd{C-M-h} (@code{mark-defun}), and then call
14674 @code{@value{COUNT-WORDS}} on it, we will find that
14675 @code{@value{COUNT-WORDS}} claims the definition has eleven words, not
14676 ten! Something is wrong!
14678 The problem is twofold: @code{@value{COUNT-WORDS}} does not count the
14679 @samp{*} as a word, and it counts the single symbol,
14680 @code{multiply-by-seven}, as containing three words. The hyphens are
14681 treated as if they were interword spaces rather than intraword
14682 connectors: @samp{multiply-by-seven} is counted as if it were written
14683 @samp{multiply by seven}.
14685 The cause of this confusion is the regular expression search within
14686 the @code{@value{COUNT-WORDS}} definition that moves point forward word
14687 by word. In the canonical version of @code{@value{COUNT-WORDS}}, the
14695 This regular expression is a pattern defining one or more word
14696 constituent characters possibly followed by one or more characters
14697 that are not word constituents. What is meant by `word constituent
14698 characters' brings us to the issue of syntax, which is worth a section
14702 @section What Constitutes a Word or Symbol?
14703 @cindex Syntax categories and tables
14705 Emacs treats different characters as belonging to different
14706 @dfn{syntax categories}. For example, the regular expression,
14707 @samp{\\w+}, is a pattern specifying one or more @emph{word
14708 constituent} characters. Word constituent characters are members of
14709 one syntax category. Other syntax categories include the class of
14710 punctuation characters, such as the period and the comma, and the
14711 class of whitespace characters, such as the blank space and the tab
14712 character. (For more information, @pxref{Syntax Tables, , Syntax
14713 Tables, elisp, The GNU Emacs Lisp Reference Manual}.)
14715 Syntax tables specify which characters belong to which categories.
14716 Usually, a hyphen is not specified as a `word constituent character'.
14717 Instead, it is specified as being in the `class of characters that are
14718 part of symbol names but not words.' This means that the
14719 @code{@value{COUNT-WORDS}} function treats it in the same way it treats
14720 an interword white space, which is why @code{@value{COUNT-WORDS}}
14721 counts @samp{multiply-by-seven} as three words.
14723 There are two ways to cause Emacs to count @samp{multiply-by-seven} as
14724 one symbol: modify the syntax table or modify the regular expression.
14726 We could redefine a hyphen as a word constituent character by
14727 modifying the syntax table that Emacs keeps for each mode. This
14728 action would serve our purpose, except that a hyphen is merely the
14729 most common character within symbols that is not typically a word
14730 constituent character; there are others, too.
14732 Alternatively, we can redefine the regexp used in the
14733 @code{@value{COUNT-WORDS}} definition so as to include symbols. This
14734 procedure has the merit of clarity, but the task is a little tricky.
14737 The first part is simple enough: the pattern must match ``at least one
14738 character that is a word or symbol constituent''. Thus:
14741 "\\(\\w\\|\\s_\\)+"
14745 The @samp{\\(} is the first part of the grouping construct that
14746 includes the @samp{\\w} and the @samp{\\s_} as alternatives, separated
14747 by the @samp{\\|}. The @samp{\\w} matches any word-constituent
14748 character and the @samp{\\s_} matches any character that is part of a
14749 symbol name but not a word-constituent character. The @samp{+}
14750 following the group indicates that the word or symbol constituent
14751 characters must be matched at least once.
14753 However, the second part of the regexp is more difficult to design.
14754 What we want is to follow the first part with ``optionally one or more
14755 characters that are not constituents of a word or symbol''. At first,
14756 I thought I could define this with the following:
14759 "\\(\\W\\|\\S_\\)*"
14763 The upper case @samp{W} and @samp{S} match characters that are
14764 @emph{not} word or symbol constituents. Unfortunately, this
14765 expression matches any character that is either not a word constituent
14766 or not a symbol constituent. This matches any character!
14768 I then noticed that every word or symbol in my test region was
14769 followed by white space (blank space, tab, or newline). So I tried
14770 placing a pattern to match one or more blank spaces after the pattern
14771 for one or more word or symbol constituents. This failed, too. Words
14772 and symbols are often separated by whitespace, but in actual code
14773 parentheses may follow symbols and punctuation may follow words. So
14774 finally, I designed a pattern in which the word or symbol constituents
14775 are followed optionally by characters that are not white space and
14776 then followed optionally by white space.
14779 Here is the full regular expression:
14782 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*"
14785 @node count-words-in-defun
14786 @section The @code{count-words-in-defun} Function
14787 @cindex Counting words in a @code{defun}
14789 We have seen that there are several ways to write a
14790 @code{count-words-region} function. To write a
14791 @code{count-words-in-defun}, we need merely adapt one of these
14794 The version that uses a @code{while} loop is easy to understand, so I
14795 am going to adapt that. Because @code{count-words-in-defun} will be
14796 part of a more complex program, it need not be interactive and it need
14797 not display a message but just return the count. These considerations
14798 simplify the definition a little.
14800 On the other hand, @code{count-words-in-defun} will be used within a
14801 buffer that contains function definitions. Consequently, it is
14802 reasonable to ask that the function determine whether it is called
14803 when point is within a function definition, and if it is, to return
14804 the count for that definition. This adds complexity to the
14805 definition, but saves us from needing to pass arguments to the
14809 These considerations lead us to prepare the following template:
14813 (defun count-words-in-defun ()
14814 "@var{documentation}@dots{}"
14815 (@var{set up}@dots{}
14816 (@var{while loop}@dots{})
14817 @var{return count})
14822 As usual, our job is to fill in the slots.
14826 We are presuming that this function will be called within a buffer
14827 containing function definitions. Point will either be within a
14828 function definition or not. For @code{count-words-in-defun} to work,
14829 point must move to the beginning of the definition, a counter must
14830 start at zero, and the counting loop must stop when point reaches the
14831 end of the definition.
14833 The @code{beginning-of-defun} function searches backwards for an
14834 opening delimiter such as a @samp{(} at the beginning of a line, and
14835 moves point to that position, or else to the limit of the search. In
14836 practice, this means that @code{beginning-of-defun} moves point to the
14837 beginning of an enclosing or preceding function definition, or else to
14838 the beginning of the buffer. We can use @code{beginning-of-defun} to
14839 place point where we wish to start.
14841 The @code{while} loop requires a counter to keep track of the words or
14842 symbols being counted. A @code{let} expression can be used to create
14843 a local variable for this purpose, and bind it to an initial value of zero.
14845 The @code{end-of-defun} function works like @code{beginning-of-defun}
14846 except that it moves point to the end of the definition.
14847 @code{end-of-defun} can be used as part of an expression that
14848 determines the position of the end of the definition.
14850 The set up for @code{count-words-in-defun} takes shape rapidly: first
14851 we move point to the beginning of the definition, then we create a
14852 local variable to hold the count, and finally, we record the position
14853 of the end of the definition so the @code{while} loop will know when to stop
14857 The code looks like this:
14861 (beginning-of-defun)
14863 (end (save-excursion (end-of-defun) (point))))
14868 The code is simple. The only slight complication is likely to concern
14869 @code{end}: it is bound to the position of the end of the definition
14870 by a @code{save-excursion} expression that returns the value of point
14871 after @code{end-of-defun} temporarily moves it to the end of the
14874 The second part of the @code{count-words-in-defun}, after the set up,
14875 is the @code{while} loop.
14877 The loop must contain an expression that jumps point forward word by
14878 word and symbol by symbol, and another expression that counts the
14879 jumps. The true-or-false-test for the @code{while} loop should test
14880 true so long as point should jump forward, and false when point is at
14881 the end of the definition. We have already redefined the regular
14882 expression for this, so the loop is straightforward:
14886 (while (and (< (point) end)
14888 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*" end t))
14889 (setq count (1+ count)))
14893 The third part of the function definition returns the count of words
14894 and symbols. This part is the last expression within the body of the
14895 @code{let} expression, and can be, very simply, the local variable
14896 @code{count}, which when evaluated returns the count.
14899 Put together, the @code{count-words-in-defun} definition looks like this:
14901 @findex count-words-in-defun
14904 (defun count-words-in-defun ()
14905 "Return the number of words and symbols in a defun."
14906 (beginning-of-defun)
14908 (end (save-excursion (end-of-defun) (point))))
14912 (and (< (point) end)
14914 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*"
14916 (setq count (1+ count)))
14921 How to test this? The function is not interactive, but it is easy to
14922 put a wrapper around the function to make it interactive; we can use
14923 almost the same code as for the recursive version of
14924 @code{@value{COUNT-WORDS}}:
14928 ;;; @r{Interactive version.}
14929 (defun count-words-defun ()
14930 "Number of words and symbols in a function definition."
14933 "Counting words and symbols in function definition ... ")
14936 (let ((count (count-words-in-defun)))
14940 "The definition does NOT have any words or symbols."))
14945 "The definition has 1 word or symbol."))
14948 "The definition has %d words or symbols." count)))))
14954 Let's re-use @kbd{C-c =} as a convenient keybinding:
14957 (global-set-key "\C-c=" 'count-words-defun)
14960 Now we can try out @code{count-words-defun}: install both
14961 @code{count-words-in-defun} and @code{count-words-defun}, and set the
14962 keybinding, and then place the cursor within the following definition:
14966 (defun multiply-by-seven (number)
14967 "Multiply NUMBER by seven."
14974 Success! The definition has 10 words and symbols.
14976 The next problem is to count the numbers of words and symbols in
14977 several definitions within a single file.
14979 @node Several defuns
14980 @section Count Several @code{defuns} Within a File
14982 A file such as @file{simple.el} may have a hundred or more function
14983 definitions within it. Our long term goal is to collect statistics on
14984 many files, but as a first step, our immediate goal is to collect
14985 statistics on one file.
14987 The information will be a series of numbers, each number being the
14988 length of a function definition. We can store the numbers in a list.
14990 We know that we will want to incorporate the information regarding one
14991 file with information about many other files; this means that the
14992 function for counting definition lengths within one file need only
14993 return the list of lengths. It need not and should not display any
14996 The word count commands contain one expression to jump point forward
14997 word by word and another expression to count the jumps. The function
14998 to return the lengths of definitions can be designed to work the same
14999 way, with one expression to jump point forward definition by
15000 definition and another expression to construct the lengths' list.
15002 This statement of the problem makes it elementary to write the
15003 function definition. Clearly, we will start the count at the
15004 beginning of the file, so the first command will be @code{(goto-char
15005 (point-min))}. Next, we start the @code{while} loop; and the
15006 true-or-false test of the loop can be a regular expression search for
15007 the next function definition---so long as the search succeeds, point
15008 is moved forward and then the body of the loop is evaluated. The body
15009 needs an expression that constructs the lengths' list. @code{cons},
15010 the list construction command, can be used to create the list. That
15011 is almost all there is to it.
15014 Here is what this fragment of code looks like:
15018 (goto-char (point-min))
15019 (while (re-search-forward "^(defun" nil t)
15021 (cons (count-words-in-defun) lengths-list)))
15025 What we have left out is the mechanism for finding the file that
15026 contains the function definitions.
15028 In previous examples, we either used this, the Info file, or we
15029 switched back and forth to some other buffer, such as the
15030 @file{*scratch*} buffer.
15032 Finding a file is a new process that we have not yet discussed.
15035 @section Find a File
15036 @cindex Find a File
15038 To find a file in Emacs, you use the @kbd{C-x C-f} (@code{find-file})
15039 command. This command is almost, but not quite right for the lengths
15043 Let's look at the source for @code{find-file}:
15047 (defun find-file (filename)
15048 "Edit file FILENAME.
15049 Switch to a buffer visiting file FILENAME,
15050 creating one if none already exists."
15051 (interactive "FFind file: ")
15052 (switch-to-buffer (find-file-noselect filename)))
15057 (The most recent version of the @code{find-file} function definition
15058 permits you to specify optional wildcards to visit multiple files; that
15059 makes the definition more complex and we will not discuss it here,
15060 since it is not relevant. You can see its source using either
15061 @kbd{M-.} (@code{find-tag}) or @kbd{C-h f} (@code{describe-function}).)
15065 (defun find-file (filename &optional wildcards)
15066 "Edit file FILENAME.
15067 Switch to a buffer visiting file FILENAME,
15068 creating one if none already exists.
15069 Interactively, the default if you just type RET is the current directory,
15070 but the visited file name is available through the minibuffer history:
15071 type M-n to pull it into the minibuffer.
15073 Interactively, or if WILDCARDS is non-nil in a call from Lisp,
15074 expand wildcards (if any) and visit multiple files. You can
15075 suppress wildcard expansion by setting `find-file-wildcards' to nil.
15077 To visit a file without any kind of conversion and without
15078 automatically choosing a major mode, use \\[find-file-literally]."
15079 (interactive (find-file-read-args "Find file: " nil))
15080 (let ((value (find-file-noselect filename nil nil wildcards)))
15082 (mapcar 'switch-to-buffer (nreverse value))
15083 (switch-to-buffer value))))
15086 The definition I am showing possesses short but complete documentation
15087 and an interactive specification that prompts you for a file name when
15088 you use the command interactively. The body of the definition
15089 contains two functions, @code{find-file-noselect} and
15090 @code{switch-to-buffer}.
15092 According to its documentation as shown by @kbd{C-h f} (the
15093 @code{describe-function} command), the @code{find-file-noselect}
15094 function reads the named file into a buffer and returns the buffer.
15095 (Its most recent version includes an optional wildcards argument,
15096 too, as well as another to read a file literally and an other you
15097 suppress warning messages. These optional arguments are irrelevant.)
15099 However, the @code{find-file-noselect} function does not select the
15100 buffer in which it puts the file. Emacs does not switch its attention
15101 (or yours if you are using @code{find-file-noselect}) to the selected
15102 buffer. That is what @code{switch-to-buffer} does: it switches the
15103 buffer to which Emacs attention is directed; and it switches the
15104 buffer displayed in the window to the new buffer. We have discussed
15105 buffer switching elsewhere. (@xref{Switching Buffers}.)
15107 In this histogram project, we do not need to display each file on the
15108 screen as the program determines the length of each definition within
15109 it. Instead of employing @code{switch-to-buffer}, we can work with
15110 @code{set-buffer}, which redirects the attention of the computer
15111 program to a different buffer but does not redisplay it on the screen.
15112 So instead of calling on @code{find-file} to do the job, we must write
15113 our own expression.
15115 The task is easy: use @code{find-file-noselect} and @code{set-buffer}.
15117 @node lengths-list-file
15118 @section @code{lengths-list-file} in Detail
15120 The core of the @code{lengths-list-file} function is a @code{while}
15121 loop containing a function to move point forward `defun by defun' and
15122 a function to count the number of words and symbols in each defun.
15123 This core must be surrounded by functions that do various other tasks,
15124 including finding the file, and ensuring that point starts out at the
15125 beginning of the file. The function definition looks like this:
15126 @findex lengths-list-file
15130 (defun lengths-list-file (filename)
15131 "Return list of definitions' lengths within FILE.
15132 The returned list is a list of numbers.
15133 Each number is the number of words or
15134 symbols in one function definition."
15137 (message "Working on `%s' ... " filename)
15139 (let ((buffer (find-file-noselect filename))
15141 (set-buffer buffer)
15142 (setq buffer-read-only t)
15144 (goto-char (point-min))
15145 (while (re-search-forward "^(defun" nil t)
15147 (cons (count-words-in-defun) lengths-list)))
15148 (kill-buffer buffer)
15154 The function is passed one argument, the name of the file on which it
15155 will work. It has four lines of documentation, but no interactive
15156 specification. Since people worry that a computer is broken if they
15157 don't see anything going on, the first line of the body is a
15160 The next line contains a @code{save-excursion} that returns Emacs's
15161 attention to the current buffer when the function completes. This is
15162 useful in case you embed this function in another function that
15163 presumes point is restored to the original buffer.
15165 In the varlist of the @code{let} expression, Emacs finds the file and
15166 binds the local variable @code{buffer} to the buffer containing the
15167 file. At the same time, Emacs creates @code{lengths-list} as a local
15170 Next, Emacs switches its attention to the buffer.
15172 In the following line, Emacs makes the buffer read-only. Ideally,
15173 this line is not necessary. None of the functions for counting words
15174 and symbols in a function definition should change the buffer.
15175 Besides, the buffer is not going to be saved, even if it were changed.
15176 This line is entirely the consequence of great, perhaps excessive,
15177 caution. The reason for the caution is that this function and those
15178 it calls work on the sources for Emacs and it is inconvenient if they
15179 are inadvertently modified. It goes without saying that I did not
15180 realize a need for this line until an experiment went awry and started
15181 to modify my Emacs source files @dots{}
15183 Next comes a call to widen the buffer if it is narrowed. This
15184 function is usually not needed---Emacs creates a fresh buffer if none
15185 already exists; but if a buffer visiting the file already exists Emacs
15186 returns that one. In this case, the buffer may be narrowed and must
15187 be widened. If we wanted to be fully `user-friendly', we would
15188 arrange to save the restriction and the location of point, but we
15191 The @code{(goto-char (point-min))} expression moves point to the
15192 beginning of the buffer.
15194 Then comes a @code{while} loop in which the `work' of the function is
15195 carried out. In the loop, Emacs determines the length of each
15196 definition and constructs a lengths' list containing the information.
15198 Emacs kills the buffer after working through it. This is to save
15199 space inside of Emacs. My version of GNU Emacs 19 contained over 300
15200 source files of interest; GNU Emacs 22 contains over a thousand source
15201 files. Another function will apply @code{lengths-list-file} to each
15204 Finally, the last expression within the @code{let} expression is the
15205 @code{lengths-list} variable; its value is returned as the value of
15206 the whole function.
15208 You can try this function by installing it in the usual fashion. Then
15209 place your cursor after the following expression and type @kbd{C-x
15210 C-e} (@code{eval-last-sexp}).
15212 @c !!! 22.1.1 lisp sources location here
15215 "/usr/local/share/emacs/22.1.1/lisp/emacs-lisp/debug.el")
15219 (You may need to change the pathname of the file; the one here is for
15220 GNU Emacs version 22.1.1. To change the expression, copy it to
15221 the @file{*scratch*} buffer and edit it.
15225 (Also, to see the full length of the list, rather than a truncated
15226 version, you may have to evaluate the following:
15229 (custom-set-variables '(eval-expression-print-length nil))
15233 (@xref{defcustom, , Specifying Variables using @code{defcustom}}.
15234 Then evaluate the @code{lengths-list-file} expression.)
15237 The lengths' list for @file{debug.el} takes less than a second to
15238 produce and looks like this in GNU Emacs 22:
15241 (83 113 105 144 289 22 30 97 48 89 25 52 52 88 28 29 77 49 43 290 232 587)
15245 (Using my old machine, the version 19 lengths' list for @file{debug.el}
15246 took seven seconds to produce and looked like this:
15249 (75 41 80 62 20 45 44 68 45 12 34 235)
15252 (The newer version of @file{debug.el} contains more defuns than the
15253 earlier one; and my new machine is much faster than the old one.)
15255 Note that the length of the last definition in the file is first in
15258 @node Several files
15259 @section Count Words in @code{defuns} in Different Files
15261 In the previous section, we created a function that returns a list of
15262 the lengths of each definition in a file. Now, we want to define a
15263 function to return a master list of the lengths of the definitions in
15266 Working on each of a list of files is a repetitious act, so we can use
15267 either a @code{while} loop or recursion.
15270 * lengths-list-many-files:: Return a list of the lengths of defuns.
15271 * append:: Attach one list to another.
15275 @node lengths-list-many-files
15276 @unnumberedsubsec Determine the lengths of @code{defuns}
15279 The design using a @code{while} loop is routine. The argument passed
15280 the function is a list of files. As we saw earlier (@pxref{Loop
15281 Example}), you can write a @code{while} loop so that the body of the
15282 loop is evaluated if such a list contains elements, but to exit the
15283 loop if the list is empty. For this design to work, the body of the
15284 loop must contain an expression that shortens the list each time the
15285 body is evaluated, so that eventually the list is empty. The usual
15286 technique is to set the value of the list to the value of the @sc{cdr}
15287 of the list each time the body is evaluated.
15290 The template looks like this:
15294 (while @var{test-whether-list-is-empty}
15296 @var{set-list-to-cdr-of-list})
15300 Also, we remember that a @code{while} loop returns @code{nil} (the
15301 result of evaluating the true-or-false-test), not the result of any
15302 evaluation within its body. (The evaluations within the body of the
15303 loop are done for their side effects.) However, the expression that
15304 sets the lengths' list is part of the body---and that is the value
15305 that we want returned by the function as a whole. To do this, we
15306 enclose the @code{while} loop within a @code{let} expression, and
15307 arrange that the last element of the @code{let} expression contains
15308 the value of the lengths' list. (@xref{Incrementing Example, , Loop
15309 Example with an Incrementing Counter}.)
15311 @findex lengths-list-many-files
15313 These considerations lead us directly to the function itself:
15317 ;;; @r{Use @code{while} loop.}
15318 (defun lengths-list-many-files (list-of-files)
15319 "Return list of lengths of defuns in LIST-OF-FILES."
15322 (let (lengths-list)
15324 ;;; @r{true-or-false-test}
15325 (while list-of-files
15330 ;;; @r{Generate a lengths' list.}
15332 (expand-file-name (car list-of-files)))))
15336 ;;; @r{Make files' list shorter.}
15337 (setq list-of-files (cdr list-of-files)))
15339 ;;; @r{Return final value of lengths' list.}
15344 @code{expand-file-name} is a built-in function that converts a file
15345 name to the absolute, long, path name form. The function employs the
15346 name of the directory in which the function is called.
15348 @c !!! 22.1.1 lisp sources location here
15350 Thus, if @code{expand-file-name} is called on @code{debug.el} when
15351 Emacs is visiting the
15352 @file{/usr/local/share/emacs/22.1.1/lisp/emacs-lisp/} directory,
15362 @c !!! 22.1.1 lisp sources location here
15364 /usr/local/share/emacs/22.1.1/lisp/emacs-lisp/debug.el
15367 The only other new element of this function definition is the as yet
15368 unstudied function @code{append}, which merits a short section for
15372 @subsection The @code{append} Function
15375 The @code{append} function attaches one list to another. Thus,
15378 (append '(1 2 3 4) '(5 6 7 8))
15389 This is exactly how we want to attach two lengths' lists produced by
15390 @code{lengths-list-file} to each other. The results contrast with
15394 (cons '(1 2 3 4) '(5 6 7 8))
15399 which constructs a new list in which the first argument to @code{cons}
15400 becomes the first element of the new list:
15403 ((1 2 3 4) 5 6 7 8)
15406 @node Several files recursively
15407 @section Recursively Count Words in Different Files
15409 Besides a @code{while} loop, you can work on each of a list of files
15410 with recursion. A recursive version of @code{lengths-list-many-files}
15411 is short and simple.
15413 The recursive function has the usual parts: the `do-again-test', the
15414 `next-step-expression', and the recursive call. The `do-again-test'
15415 determines whether the function should call itself again, which it
15416 will do if the @code{list-of-files} contains any remaining elements;
15417 the `next-step-expression' resets the @code{list-of-files} to the
15418 @sc{cdr} of itself, so eventually the list will be empty; and the
15419 recursive call calls itself on the shorter list. The complete
15420 function is shorter than this description!
15421 @findex recursive-lengths-list-many-files
15425 (defun recursive-lengths-list-many-files (list-of-files)
15426 "Return list of lengths of each defun in LIST-OF-FILES."
15427 (if list-of-files ; @r{do-again-test}
15430 (expand-file-name (car list-of-files)))
15431 (recursive-lengths-list-many-files
15432 (cdr list-of-files)))))
15437 In a sentence, the function returns the lengths' list for the first of
15438 the @code{list-of-files} appended to the result of calling itself on
15439 the rest of the @code{list-of-files}.
15441 Here is a test of @code{recursive-lengths-list-many-files}, along with
15442 the results of running @code{lengths-list-file} on each of the files
15445 Install @code{recursive-lengths-list-many-files} and
15446 @code{lengths-list-file}, if necessary, and then evaluate the
15447 following expressions. You may need to change the files' pathnames;
15448 those here work when this Info file and the Emacs sources are located
15449 in their customary places. To change the expressions, copy them to
15450 the @file{*scratch*} buffer, edit them, and then evaluate them.
15452 The results are shown after the @samp{@result{}}. (These results are
15453 for files from Emacs version 22.1.1; files from other versions of
15454 Emacs may produce different results.)
15456 @c !!! 22.1.1 lisp sources location here
15459 (cd "/usr/local/share/emacs/22.1.1/")
15461 (lengths-list-file "./lisp/macros.el")
15462 @result{} (283 263 480 90)
15466 (lengths-list-file "./lisp/mail/mailalias.el")
15467 @result{} (38 32 29 95 178 180 321 218 324)
15471 (lengths-list-file "./lisp/makesum.el")
15476 (recursive-lengths-list-many-files
15477 '("./lisp/macros.el"
15478 "./lisp/mail/mailalias.el"
15479 "./lisp/makesum.el"))
15480 @result{} (283 263 480 90 38 32 29 95 178 180 321 218 324 85 181)
15484 The @code{recursive-lengths-list-many-files} function produces the
15487 The next step is to prepare the data in the list for display in a graph.
15489 @node Prepare the data
15490 @section Prepare the Data for Display in a Graph
15492 The @code{recursive-lengths-list-many-files} function returns a list
15493 of numbers. Each number records the length of a function definition.
15494 What we need to do now is transform this data into a list of numbers
15495 suitable for generating a graph. The new list will tell how many
15496 functions definitions contain less than 10 words and
15497 symbols, how many contain between 10 and 19 words and symbols, how
15498 many contain between 20 and 29 words and symbols, and so on.
15500 In brief, we need to go through the lengths' list produced by the
15501 @code{recursive-lengths-list-many-files} function and count the number
15502 of defuns within each range of lengths, and produce a list of those
15506 * Data for Display in Detail::
15507 * Sorting:: Sorting lists.
15508 * Files List:: Making a list of files.
15509 * Counting function definitions::
15513 @node Data for Display in Detail
15514 @unnumberedsubsec The Data for Display in Detail
15517 Based on what we have done before, we can readily foresee that it
15518 should not be too hard to write a function that `@sc{cdr}s' down the
15519 lengths' list, looks at each element, determines which length range it
15520 is in, and increments a counter for that range.
15522 However, before beginning to write such a function, we should consider
15523 the advantages of sorting the lengths' list first, so the numbers are
15524 ordered from smallest to largest. First, sorting will make it easier
15525 to count the numbers in each range, since two adjacent numbers will
15526 either be in the same length range or in adjacent ranges. Second, by
15527 inspecting a sorted list, we can discover the highest and lowest
15528 number, and thereby determine the largest and smallest length range
15532 @subsection Sorting Lists
15535 Emacs contains a function to sort lists, called (as you might guess)
15536 @code{sort}. The @code{sort} function takes two arguments, the list
15537 to be sorted, and a predicate that determines whether the first of
15538 two list elements is ``less'' than the second.
15540 As we saw earlier (@pxref{Wrong Type of Argument, , Using the Wrong
15541 Type Object as an Argument}), a predicate is a function that
15542 determines whether some property is true or false. The @code{sort}
15543 function will reorder a list according to whatever property the
15544 predicate uses; this means that @code{sort} can be used to sort
15545 non-numeric lists by non-numeric criteria---it can, for example,
15546 alphabetize a list.
15549 The @code{<} function is used when sorting a numeric list. For example,
15552 (sort '(4 8 21 17 33 7 21 7) '<)
15560 (4 7 7 8 17 21 21 33)
15564 (Note that in this example, both the arguments are quoted so that the
15565 symbols are not evaluated before being passed to @code{sort} as
15568 Sorting the list returned by the
15569 @code{recursive-lengths-list-many-files} function is straightforward;
15570 it uses the @code{<} function:
15574 In GNU Emacs 22, eval
15576 (cd "/usr/local/share/emacs/22.0.50/")
15578 (recursive-lengths-list-many-files
15579 '("./lisp/macros.el"
15580 "./lisp/mail/mailalias.el"
15581 "./lisp/makesum.el"))
15589 (recursive-lengths-list-many-files
15590 '("./lisp/macros.el"
15591 "./lisp/mailalias.el"
15592 "./lisp/makesum.el"))
15602 (29 32 38 85 90 95 178 180 181 218 263 283 321 324 480)
15606 (Note that in this example, the first argument to @code{sort} is not
15607 quoted, since the expression must be evaluated so as to produce the
15608 list that is passed to @code{sort}.)
15611 @subsection Making a List of Files
15613 The @code{recursive-lengths-list-many-files} function requires a list
15614 of files as its argument. For our test examples, we constructed such
15615 a list by hand; but the Emacs Lisp source directory is too large for
15616 us to do for that. Instead, we will write a function to do the job
15617 for us. In this function, we will use both a @code{while} loop and a
15620 @findex directory-files
15621 We did not have to write a function like this for older versions of
15622 GNU Emacs, since they placed all the @samp{.el} files in one
15623 directory. Instead, we were able to use the @code{directory-files}
15624 function, which lists the names of files that match a specified
15625 pattern within a single directory.
15627 However, recent versions of Emacs place Emacs Lisp files in
15628 sub-directories of the top level @file{lisp} directory. This
15629 re-arrangement eases navigation. For example, all the mail related
15630 files are in a @file{lisp} sub-directory called @file{mail}. But at
15631 the same time, this arrangement forces us to create a file listing
15632 function that descends into the sub-directories.
15634 @findex files-in-below-directory
15635 We can create this function, called @code{files-in-below-directory},
15636 using familiar functions such as @code{car}, @code{nthcdr}, and
15637 @code{substring} in conjunction with an existing function called
15638 @code{directory-files-and-attributes}. This latter function not only
15639 lists all the filenames in a directory, including the names
15640 of sub-directories, but also their attributes.
15642 To restate our goal: to create a function that will enable us
15643 to feed filenames to @code{recursive-lengths-list-many-files}
15644 as a list that looks like this (but with more elements):
15648 ("./lisp/macros.el"
15649 "./lisp/mail/rmail.el"
15650 "./lisp/makesum.el")
15654 The @code{directory-files-and-attributes} function returns a list of
15655 lists. Each of the lists within the main list consists of 13
15656 elements. The first element is a string that contains the name of the
15657 file---which, in GNU/Linux, may be a `directory file', that is to
15658 say, a file with the special attributes of a directory. The second
15659 element of the list is @code{t} for a directory, a string
15660 for symbolic link (the string is the name linked to), or @code{nil}.
15662 For example, the first @samp{.el} file in the @file{lisp/} directory
15663 is @file{abbrev.el}. Its name is
15664 @file{/usr/local/share/emacs/22.1.1/lisp/abbrev.el} and it is not a
15665 directory or a symbolic link.
15668 This is how @code{directory-files-and-attributes} lists that file and
15680 (20615 27034 579989 697000)
15682 (20615 26327 734791 805000)
15694 On the other hand, @file{mail/} is a directory within the @file{lisp/}
15695 directory. The beginning of its listing looks like this:
15706 (To learn about the different attributes, look at the documentation of
15707 @code{file-attributes}. Bear in mind that the @code{file-attributes}
15708 function does not list the filename, so its first element is
15709 @code{directory-files-and-attributes}'s second element.)
15711 We will want our new function, @code{files-in-below-directory}, to
15712 list the @samp{.el} files in the directory it is told to check, and in
15713 any directories below that directory.
15715 This gives us a hint on how to construct
15716 @code{files-in-below-directory}: within a directory, the function
15717 should add @samp{.el} filenames to a list; and if, within a directory,
15718 the function comes upon a sub-directory, it should go into that
15719 sub-directory and repeat its actions.
15721 However, we should note that every directory contains a name that
15722 refers to itself, called @file{.}, (``dot'') and a name that refers to
15723 its parent directory, called @file{..} (``double dot''). (In
15724 @file{/}, the root directory, @file{..} refers to itself, since
15725 @file{/} has no parent.) Clearly, we do not want our
15726 @code{files-in-below-directory} function to enter those directories,
15727 since they always lead us, directly or indirectly, to the current
15730 Consequently, our @code{files-in-below-directory} function must do
15735 Check to see whether it is looking at a filename that ends in
15736 @samp{.el}; and if so, add its name to a list.
15739 Check to see whether it is looking at a filename that is the name of a
15740 directory; and if so,
15744 Check to see whether it is looking at @file{.} or @file{..}; and if
15748 Or else, go into that directory and repeat the process.
15752 Let's write a function definition to do these tasks. We will use a
15753 @code{while} loop to move from one filename to another within a
15754 directory, checking what needs to be done; and we will use a recursive
15755 call to repeat the actions on each sub-directory. The recursive
15756 pattern is `accumulate'
15757 (@pxref{Accumulate, , Recursive Pattern: @emph{accumulate}}),
15758 using @code{append} as the combiner.
15761 (directory-files "/usr/local/src/emacs/lisp/" t "\\.el$")
15762 (shell-command "find /usr/local/src/emacs/lisp/ -name '*.el'")
15764 (directory-files "/usr/local/share/emacs/22.1.1/lisp/" t "\\.el$")
15765 (shell-command "find /usr/local/share/emacs/22.1.1/lisp/ -name '*.el'")
15768 @c /usr/local/share/emacs/22.1.1/lisp/
15771 Here is the function:
15775 (defun files-in-below-directory (directory)
15776 "List the .el files in DIRECTORY and in its sub-directories."
15777 ;; Although the function will be used non-interactively,
15778 ;; it will be easier to test if we make it interactive.
15779 ;; The directory will have a name such as
15780 ;; "/usr/local/share/emacs/22.1.1/lisp/"
15781 (interactive "DDirectory name: ")
15784 (let (el-files-list
15785 (current-directory-list
15786 (directory-files-and-attributes directory t)))
15787 ;; while we are in the current directory
15788 (while current-directory-list
15792 ;; check to see whether filename ends in `.el'
15793 ;; and if so, append its name to a list.
15794 ((equal ".el" (substring (car (car current-directory-list)) -3))
15795 (setq el-files-list
15796 (cons (car (car current-directory-list)) el-files-list)))
15799 ;; check whether filename is that of a directory
15800 ((eq t (car (cdr (car current-directory-list))))
15801 ;; decide whether to skip or recurse
15804 (substring (car (car current-directory-list)) -1))
15805 ;; then do nothing since filename is that of
15806 ;; current directory or parent, "." or ".."
15810 ;; else descend into the directory and repeat the process
15811 (setq el-files-list
15813 (files-in-below-directory
15814 (car (car current-directory-list)))
15816 ;; move to the next filename in the list; this also
15817 ;; shortens the list so the while loop eventually comes to an end
15818 (setq current-directory-list (cdr current-directory-list)))
15819 ;; return the filenames
15824 @c (files-in-below-directory "/usr/local/src/emacs/lisp/")
15825 @c (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/")
15827 The @code{files-in-below-directory} @code{directory-files} function
15828 takes one argument, the name of a directory.
15831 Thus, on my system,
15833 @c (length (files-in-below-directory "/usr/local/src/emacs/lisp/"))
15835 @c !!! 22.1.1 lisp sources location here
15839 (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/"))
15844 tells me that in and below my Lisp sources directory are 1031
15847 @code{files-in-below-directory} returns a list in reverse alphabetical
15848 order. An expression to sort the list in alphabetical order looks
15854 (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/")
15861 "Test how long it takes to find lengths of all sorted elisp defuns."
15862 (insert "\n" (current-time-string) "\n")
15865 (recursive-lengths-list-many-files
15866 (files-in-below-directory "/usr/local/src/emacs/lisp/"))
15868 (insert (format "%s" (current-time-string))))
15871 @node Counting function definitions
15872 @subsection Counting function definitions
15874 Our immediate goal is to generate a list that tells us how many
15875 function definitions contain fewer than 10 words and symbols, how many
15876 contain between 10 and 19 words and symbols, how many contain between
15877 20 and 29 words and symbols, and so on.
15879 With a sorted list of numbers, this is easy: count how many elements
15880 of the list are smaller than 10, then, after moving past the numbers
15881 just counted, count how many are smaller than 20, then, after moving
15882 past the numbers just counted, count how many are smaller than 30, and
15883 so on. Each of the numbers, 10, 20, 30, 40, and the like, is one
15884 larger than the top of that range. We can call the list of such
15885 numbers the @code{top-of-ranges} list.
15888 If we wished, we could generate this list automatically, but it is
15889 simpler to write a list manually. Here it is:
15890 @vindex top-of-ranges
15894 (defvar top-of-ranges
15897 110 120 130 140 150
15898 160 170 180 190 200
15899 210 220 230 240 250
15900 260 270 280 290 300)
15901 "List specifying ranges for `defuns-per-range'.")
15905 To change the ranges, we edit this list.
15907 Next, we need to write the function that creates the list of the
15908 number of definitions within each range. Clearly, this function must
15909 take the @code{sorted-lengths} and the @code{top-of-ranges} lists
15912 The @code{defuns-per-range} function must do two things again and
15913 again: it must count the number of definitions within a range
15914 specified by the current top-of-range value; and it must shift to the
15915 next higher value in the @code{top-of-ranges} list after counting the
15916 number of definitions in the current range. Since each of these
15917 actions is repetitive, we can use @code{while} loops for the job.
15918 One loop counts the number of definitions in the range defined by the
15919 current top-of-range value, and the other loop selects each of the
15920 top-of-range values in turn.
15922 Several entries of the @code{sorted-lengths} list are counted for each
15923 range; this means that the loop for the @code{sorted-lengths} list
15924 will be inside the loop for the @code{top-of-ranges} list, like a
15925 small gear inside a big gear.
15927 The inner loop counts the number of definitions within the range. It
15928 is a simple counting loop of the type we have seen before.
15929 (@xref{Incrementing Loop, , A loop with an incrementing counter}.)
15930 The true-or-false test of the loop tests whether the value from the
15931 @code{sorted-lengths} list is smaller than the current value of the
15932 top of the range. If it is, the function increments the counter and
15933 tests the next value from the @code{sorted-lengths} list.
15936 The inner loop looks like this:
15940 (while @var{length-element-smaller-than-top-of-range}
15941 (setq number-within-range (1+ number-within-range))
15942 (setq sorted-lengths (cdr sorted-lengths)))
15946 The outer loop must start with the lowest value of the
15947 @code{top-of-ranges} list, and then be set to each of the succeeding
15948 higher values in turn. This can be done with a loop like this:
15952 (while top-of-ranges
15953 @var{body-of-loop}@dots{}
15954 (setq top-of-ranges (cdr top-of-ranges)))
15959 Put together, the two loops look like this:
15963 (while top-of-ranges
15965 ;; @r{Count the number of elements within the current range.}
15966 (while @var{length-element-smaller-than-top-of-range}
15967 (setq number-within-range (1+ number-within-range))
15968 (setq sorted-lengths (cdr sorted-lengths)))
15970 ;; @r{Move to next range.}
15971 (setq top-of-ranges (cdr top-of-ranges)))
15975 In addition, in each circuit of the outer loop, Emacs should record
15976 the number of definitions within that range (the value of
15977 @code{number-within-range}) in a list. We can use @code{cons} for
15978 this purpose. (@xref{cons, , @code{cons}}.)
15980 The @code{cons} function works fine, except that the list it
15981 constructs will contain the number of definitions for the highest
15982 range at its beginning and the number of definitions for the lowest
15983 range at its end. This is because @code{cons} attaches new elements
15984 of the list to the beginning of the list, and since the two loops are
15985 working their way through the lengths' list from the lower end first,
15986 the @code{defuns-per-range-list} will end up largest number first.
15987 But we will want to print our graph with smallest values first and the
15988 larger later. The solution is to reverse the order of the
15989 @code{defuns-per-range-list}. We can do this using the
15990 @code{nreverse} function, which reverses the order of a list.
15997 (nreverse '(1 2 3 4))
16008 Note that the @code{nreverse} function is ``destructive''---that is,
16009 it changes the list to which it is applied; this contrasts with the
16010 @code{car} and @code{cdr} functions, which are non-destructive. In
16011 this case, we do not want the original @code{defuns-per-range-list},
16012 so it does not matter that it is destroyed. (The @code{reverse}
16013 function provides a reversed copy of a list, leaving the original list
16018 Put all together, the @code{defuns-per-range} looks like this:
16022 (defun defuns-per-range (sorted-lengths top-of-ranges)
16023 "SORTED-LENGTHS defuns in each TOP-OF-RANGES range."
16024 (let ((top-of-range (car top-of-ranges))
16025 (number-within-range 0)
16026 defuns-per-range-list)
16031 (while top-of-ranges
16037 ;; @r{Need number for numeric test.}
16038 (car sorted-lengths)
16039 (< (car sorted-lengths) top-of-range))
16043 ;; @r{Count number of definitions within current range.}
16044 (setq number-within-range (1+ number-within-range))
16045 (setq sorted-lengths (cdr sorted-lengths)))
16047 ;; @r{Exit inner loop but remain within outer loop.}
16051 (setq defuns-per-range-list
16052 (cons number-within-range defuns-per-range-list))
16053 (setq number-within-range 0) ; @r{Reset count to zero.}
16057 ;; @r{Move to next range.}
16058 (setq top-of-ranges (cdr top-of-ranges))
16059 ;; @r{Specify next top of range value.}
16060 (setq top-of-range (car top-of-ranges)))
16064 ;; @r{Exit outer loop and count the number of defuns larger than}
16065 ;; @r{ the largest top-of-range value.}
16066 (setq defuns-per-range-list
16068 (length sorted-lengths)
16069 defuns-per-range-list))
16073 ;; @r{Return a list of the number of definitions within each range,}
16074 ;; @r{ smallest to largest.}
16075 (nreverse defuns-per-range-list)))
16081 The function is straightforward except for one subtle feature. The
16082 true-or-false test of the inner loop looks like this:
16086 (and (car sorted-lengths)
16087 (< (car sorted-lengths) top-of-range))
16093 instead of like this:
16096 (< (car sorted-lengths) top-of-range)
16099 The purpose of the test is to determine whether the first item in the
16100 @code{sorted-lengths} list is less than the value of the top of the
16103 The simple version of the test works fine unless the
16104 @code{sorted-lengths} list has a @code{nil} value. In that case, the
16105 @code{(car sorted-lengths)} expression function returns
16106 @code{nil}. The @code{<} function cannot compare a number to
16107 @code{nil}, which is an empty list, so Emacs signals an error and
16108 stops the function from attempting to continue to execute.
16110 The @code{sorted-lengths} list always becomes @code{nil} when the
16111 counter reaches the end of the list. This means that any attempt to
16112 use the @code{defuns-per-range} function with the simple version of
16113 the test will fail.
16115 We solve the problem by using the @code{(car sorted-lengths)}
16116 expression in conjunction with the @code{and} expression. The
16117 @code{(car sorted-lengths)} expression returns a non-@code{nil}
16118 value so long as the list has at least one number within it, but
16119 returns @code{nil} if the list is empty. The @code{and} expression
16120 first evaluates the @code{(car sorted-lengths)} expression, and
16121 if it is @code{nil}, returns false @emph{without} evaluating the
16122 @code{<} expression. But if the @code{(car sorted-lengths)}
16123 expression returns a non-@code{nil} value, the @code{and} expression
16124 evaluates the @code{<} expression, and returns that value as the value
16125 of the @code{and} expression.
16127 @c colon in printed section title causes problem in Info cross reference
16128 This way, we avoid an error.
16131 (For information about @code{and}, see
16132 @ref{kill-new function, , The @code{kill-new} function}.)
16136 (@xref{kill-new function, , The @code{kill-new} function}, for
16137 information about @code{and}.)
16140 Here is a short test of the @code{defuns-per-range} function. First,
16141 evaluate the expression that binds (a shortened)
16142 @code{top-of-ranges} list to the list of values, then evaluate the
16143 expression for binding the @code{sorted-lengths} list, and then
16144 evaluate the @code{defuns-per-range} function.
16148 ;; @r{(Shorter list than we will use later.)}
16149 (setq top-of-ranges
16150 '(110 120 130 140 150
16151 160 170 180 190 200))
16153 (setq sorted-lengths
16154 '(85 86 110 116 122 129 154 176 179 200 265 300 300))
16156 (defuns-per-range sorted-lengths top-of-ranges)
16162 The list returned looks like this:
16165 (2 2 2 0 0 1 0 2 0 0 4)
16169 Indeed, there are two elements of the @code{sorted-lengths} list
16170 smaller than 110, two elements between 110 and 119, two elements
16171 between 120 and 129, and so on. There are four elements with a value
16174 @c The next step is to turn this numbers' list into a graph.
16175 @node Readying a Graph
16176 @chapter Readying a Graph
16177 @cindex Readying a graph
16178 @cindex Graph prototype
16179 @cindex Prototype graph
16180 @cindex Body of graph
16182 Our goal is to construct a graph showing the numbers of function
16183 definitions of various lengths in the Emacs lisp sources.
16185 As a practical matter, if you were creating a graph, you would
16186 probably use a program such as @code{gnuplot} to do the job.
16187 (@code{gnuplot} is nicely integrated into GNU Emacs.) In this case,
16188 however, we create one from scratch, and in the process we will
16189 re-acquaint ourselves with some of what we learned before and learn
16192 In this chapter, we will first write a simple graph printing function.
16193 This first definition will be a @dfn{prototype}, a rapidly written
16194 function that enables us to reconnoiter this unknown graph-making
16195 territory. We will discover dragons, or find that they are myth.
16196 After scouting the terrain, we will feel more confident and enhance
16197 the function to label the axes automatically.
16200 * Columns of a graph::
16201 * graph-body-print:: How to print the body of a graph.
16202 * recursive-graph-body-print::
16204 * Line Graph Exercise::
16208 @node Columns of a graph
16209 @unnumberedsec Printing the Columns of a Graph
16212 Since Emacs is designed to be flexible and work with all kinds of
16213 terminals, including character-only terminals, the graph will need to
16214 be made from one of the `typewriter' symbols. An asterisk will do; as
16215 we enhance the graph-printing function, we can make the choice of
16216 symbol a user option.
16218 We can call this function @code{graph-body-print}; it will take a
16219 @code{numbers-list} as its only argument. At this stage, we will not
16220 label the graph, but only print its body.
16222 The @code{graph-body-print} function inserts a vertical column of
16223 asterisks for each element in the @code{numbers-list}. The height of
16224 each line is determined by the value of that element of the
16225 @code{numbers-list}.
16227 Inserting columns is a repetitive act; that means that this function can
16228 be written either with a @code{while} loop or recursively.
16230 Our first challenge is to discover how to print a column of asterisks.
16231 Usually, in Emacs, we print characters onto a screen horizontally,
16232 line by line, by typing. We have two routes we can follow: write our
16233 own column-insertion function or discover whether one exists in Emacs.
16235 To see whether there is one in Emacs, we can use the @kbd{M-x apropos}
16236 command. This command is like the @kbd{C-h a} (@code{command-apropos})
16237 command, except that the latter finds only those functions that are
16238 commands. The @kbd{M-x apropos} command lists all symbols that match
16239 a regular expression, including functions that are not interactive.
16242 What we want to look for is some command that prints or inserts
16243 columns. Very likely, the name of the function will contain either
16244 the word `print' or the word `insert' or the word `column'.
16245 Therefore, we can simply type @kbd{M-x apropos RET
16246 print\|insert\|column RET} and look at the result. On my system, this
16247 command once too takes quite some time, and then produced a list of 79
16248 functions and variables. Now it does not take much time at all and
16249 produces a list of 211 functions and variables. Scanning down the
16250 list, the only function that looks as if it might do the job is
16251 @code{insert-rectangle}.
16254 Indeed, this is the function we want; its documentation says:
16259 Insert text of RECTANGLE with upper left corner at point.
16260 RECTANGLE's first line is inserted at point,
16261 its second line is inserted at a point vertically under point, etc.
16262 RECTANGLE should be a list of strings.
16263 After this command, the mark is at the upper left corner
16264 and point is at the lower right corner.
16268 We can run a quick test, to make sure it does what we expect of it.
16270 Here is the result of placing the cursor after the
16271 @code{insert-rectangle} expression and typing @kbd{C-u C-x C-e}
16272 (@code{eval-last-sexp}). The function inserts the strings
16273 @samp{"first"}, @samp{"second"}, and @samp{"third"} at and below
16274 point. Also the function returns @code{nil}.
16278 (insert-rectangle '("first" "second" "third"))first
16285 Of course, we won't be inserting the text of the
16286 @code{insert-rectangle} expression itself into the buffer in which we
16287 are making the graph, but will call the function from our program. We
16288 shall, however, have to make sure that point is in the buffer at the
16289 place where the @code{insert-rectangle} function will insert its
16292 If you are reading this in Info, you can see how this works by
16293 switching to another buffer, such as the @file{*scratch*} buffer,
16294 placing point somewhere in the buffer, typing @kbd{M-:}, typing the
16295 @code{insert-rectangle} expression into the minibuffer at the prompt,
16296 and then typing @key{RET}. This causes Emacs to evaluate the
16297 expression in the minibuffer, but to use as the value of point the
16298 position of point in the @file{*scratch*} buffer. (@kbd{M-:} is the
16299 keybinding for @code{eval-expression}. Also, @code{nil} does not
16300 appear in the @file{*scratch*} buffer since the expression is
16301 evaluated in the minibuffer.)
16303 We find when we do this that point ends up at the end of the last
16304 inserted line---that is to say, this function moves point as a
16305 side-effect. If we were to repeat the command, with point at this
16306 position, the next insertion would be below and to the right of the
16307 previous insertion. We don't want this! If we are going to make a
16308 bar graph, the columns need to be beside each other.
16310 So we discover that each cycle of the column-inserting @code{while}
16311 loop must reposition point to the place we want it, and that place
16312 will be at the top, not the bottom, of the column. Moreover, we
16313 remember that when we print a graph, we do not expect all the columns
16314 to be the same height. This means that the top of each column may be
16315 at a different height from the previous one. We cannot simply
16316 reposition point to the same line each time, but moved over to the
16317 right---or perhaps we can@dots{}
16319 We are planning to make the columns of the bar graph out of asterisks.
16320 The number of asterisks in the column is the number specified by the
16321 current element of the @code{numbers-list}. We need to construct a
16322 list of asterisks of the right length for each call to
16323 @code{insert-rectangle}. If this list consists solely of the requisite
16324 number of asterisks, then we will have position point the right number
16325 of lines above the base for the graph to print correctly. This could
16328 Alternatively, if we can figure out some way to pass
16329 @code{insert-rectangle} a list of the same length each time, then we
16330 can place point on the same line each time, but move it over one
16331 column to the right for each new column. If we do this, however, some
16332 of the entries in the list passed to @code{insert-rectangle} must be
16333 blanks rather than asterisks. For example, if the maximum height of
16334 the graph is 5, but the height of the column is 3, then
16335 @code{insert-rectangle} requires an argument that looks like this:
16338 (" " " " "*" "*" "*")
16341 This last proposal is not so difficult, so long as we can determine
16342 the column height. There are two ways for us to specify the column
16343 height: we can arbitrarily state what it will be, which would work
16344 fine for graphs of that height; or we can search through the list of
16345 numbers and use the maximum height of the list as the maximum height
16346 of the graph. If the latter operation were difficult, then the former
16347 procedure would be easiest, but there is a function built into Emacs
16348 that determines the maximum of its arguments. We can use that
16349 function. The function is called @code{max} and it returns the
16350 largest of all its arguments, which must be numbers. Thus, for
16358 returns 7. (A corresponding function called @code{min} returns the
16359 smallest of all its arguments.)
16363 However, we cannot simply call @code{max} on the @code{numbers-list};
16364 the @code{max} function expects numbers as its argument, not a list of
16365 numbers. Thus, the following expression,
16368 (max '(3 4 6 5 7 3))
16373 produces the following error message;
16376 Wrong type of argument: number-or-marker-p, (3 4 6 5 7 3)
16380 We need a function that passes a list of arguments to a function.
16381 This function is @code{apply}. This function `applies' its first
16382 argument (a function) to its remaining arguments, the last of which
16389 (apply 'max 3 4 7 3 '(4 8 5))
16395 (Incidentally, I don't know how you would learn of this function
16396 without a book such as this. It is possible to discover other
16397 functions, like @code{search-forward} or @code{insert-rectangle}, by
16398 guessing at a part of their names and then using @code{apropos}. Even
16399 though its base in metaphor is clear---`apply' its first argument to
16400 the rest---I doubt a novice would come up with that particular word
16401 when using @code{apropos} or other aid. Of course, I could be wrong;
16402 after all, the function was first named by someone who had to invent
16405 The second and subsequent arguments to @code{apply} are optional, so
16406 we can use @code{apply} to call a function and pass the elements of a
16407 list to it, like this, which also returns 8:
16410 (apply 'max '(4 8 5))
16413 This latter way is how we will use @code{apply}. The
16414 @code{recursive-lengths-list-many-files} function returns a numbers'
16415 list to which we can apply @code{max} (we could also apply @code{max} to
16416 the sorted numbers' list; it does not matter whether the list is
16420 Hence, the operation for finding the maximum height of the graph is this:
16423 (setq max-graph-height (apply 'max numbers-list))
16426 Now we can return to the question of how to create a list of strings
16427 for a column of the graph. Told the maximum height of the graph
16428 and the number of asterisks that should appear in the column, the
16429 function should return a list of strings for the
16430 @code{insert-rectangle} command to insert.
16432 Each column is made up of asterisks or blanks. Since the function is
16433 passed the value of the height of the column and the number of
16434 asterisks in the column, the number of blanks can be found by
16435 subtracting the number of asterisks from the height of the column.
16436 Given the number of blanks and the number of asterisks, two
16437 @code{while} loops can be used to construct the list:
16441 ;;; @r{First version.}
16442 (defun column-of-graph (max-graph-height actual-height)
16443 "Return list of strings that is one column of a graph."
16444 (let ((insert-list nil)
16445 (number-of-top-blanks
16446 (- max-graph-height actual-height)))
16450 ;; @r{Fill in asterisks.}
16451 (while (> actual-height 0)
16452 (setq insert-list (cons "*" insert-list))
16453 (setq actual-height (1- actual-height)))
16457 ;; @r{Fill in blanks.}
16458 (while (> number-of-top-blanks 0)
16459 (setq insert-list (cons " " insert-list))
16460 (setq number-of-top-blanks
16461 (1- number-of-top-blanks)))
16465 ;; @r{Return whole list.}
16470 If you install this function and then evaluate the following
16471 expression you will see that it returns the list as desired:
16474 (column-of-graph 5 3)
16482 (" " " " "*" "*" "*")
16485 As written, @code{column-of-graph} contains a major flaw: the symbols
16486 used for the blank and for the marked entries in the column are
16487 `hard-coded' as a space and asterisk. This is fine for a prototype,
16488 but you, or another user, may wish to use other symbols. For example,
16489 in testing the graph function, you many want to use a period in place
16490 of the space, to make sure the point is being repositioned properly
16491 each time the @code{insert-rectangle} function is called; or you might
16492 want to substitute a @samp{+} sign or other symbol for the asterisk.
16493 You might even want to make a graph-column that is more than one
16494 display column wide. The program should be more flexible. The way to
16495 do that is to replace the blank and the asterisk with two variables
16496 that we can call @code{graph-blank} and @code{graph-symbol} and define
16497 those variables separately.
16499 Also, the documentation is not well written. These considerations
16500 lead us to the second version of the function:
16504 (defvar graph-symbol "*"
16505 "String used as symbol in graph, usually an asterisk.")
16509 (defvar graph-blank " "
16510 "String used as blank in graph, usually a blank space.
16511 graph-blank must be the same number of columns wide
16517 (For an explanation of @code{defvar}, see
16518 @ref{defvar, , Initializing a Variable with @code{defvar}}.)
16522 ;;; @r{Second version.}
16523 (defun column-of-graph (max-graph-height actual-height)
16524 "Return MAX-GRAPH-HEIGHT strings; ACTUAL-HEIGHT are graph-symbols.
16528 The graph-symbols are contiguous entries at the end
16530 The list will be inserted as one column of a graph.
16531 The strings are either graph-blank or graph-symbol."
16535 (let ((insert-list nil)
16536 (number-of-top-blanks
16537 (- max-graph-height actual-height)))
16541 ;; @r{Fill in @code{graph-symbols}.}
16542 (while (> actual-height 0)
16543 (setq insert-list (cons graph-symbol insert-list))
16544 (setq actual-height (1- actual-height)))
16548 ;; @r{Fill in @code{graph-blanks}.}
16549 (while (> number-of-top-blanks 0)
16550 (setq insert-list (cons graph-blank insert-list))
16551 (setq number-of-top-blanks
16552 (1- number-of-top-blanks)))
16554 ;; @r{Return whole list.}
16559 If we wished, we could rewrite @code{column-of-graph} a third time to
16560 provide optionally for a line graph as well as for a bar graph. This
16561 would not be hard to do. One way to think of a line graph is that it
16562 is no more than a bar graph in which the part of each bar that is
16563 below the top is blank. To construct a column for a line graph, the
16564 function first constructs a list of blanks that is one shorter than
16565 the value, then it uses @code{cons} to attach a graph symbol to the
16566 list; then it uses @code{cons} again to attach the `top blanks' to
16569 It is easy to see how to write such a function, but since we don't
16570 need it, we will not do it. But the job could be done, and if it were
16571 done, it would be done with @code{column-of-graph}. Even more
16572 important, it is worth noting that few changes would have to be made
16573 anywhere else. The enhancement, if we ever wish to make it, is
16576 Now, finally, we come to our first actual graph printing function.
16577 This prints the body of a graph, not the labels for the vertical and
16578 horizontal axes, so we can call this @code{graph-body-print}.
16580 @node graph-body-print
16581 @section The @code{graph-body-print} Function
16582 @findex graph-body-print
16584 After our preparation in the preceding section, the
16585 @code{graph-body-print} function is straightforward. The function
16586 will print column after column of asterisks and blanks, using the
16587 elements of a numbers' list to specify the number of asterisks in each
16588 column. This is a repetitive act, which means we can use a
16589 decrementing @code{while} loop or recursive function for the job. In
16590 this section, we will write the definition using a @code{while} loop.
16592 The @code{column-of-graph} function requires the height of the graph
16593 as an argument, so we should determine and record that as a local variable.
16595 This leads us to the following template for the @code{while} loop
16596 version of this function:
16600 (defun graph-body-print (numbers-list)
16601 "@var{documentation}@dots{}"
16602 (let ((height @dots{}
16607 (while numbers-list
16608 @var{insert-columns-and-reposition-point}
16609 (setq numbers-list (cdr numbers-list)))))
16614 We need to fill in the slots of the template.
16616 Clearly, we can use the @code{(apply 'max numbers-list)} expression to
16617 determine the height of the graph.
16619 The @code{while} loop will cycle through the @code{numbers-list} one
16620 element at a time. As it is shortened by the @code{(setq numbers-list
16621 (cdr numbers-list))} expression, the @sc{car} of each instance of the
16622 list is the value of the argument for @code{column-of-graph}.
16624 At each cycle of the @code{while} loop, the @code{insert-rectangle}
16625 function inserts the list returned by @code{column-of-graph}. Since
16626 the @code{insert-rectangle} function moves point to the lower right of
16627 the inserted rectangle, we need to save the location of point at the
16628 time the rectangle is inserted, move back to that position after the
16629 rectangle is inserted, and then move horizontally to the next place
16630 from which @code{insert-rectangle} is called.
16632 If the inserted columns are one character wide, as they will be if
16633 single blanks and asterisks are used, the repositioning command is
16634 simply @code{(forward-char 1)}; however, the width of a column may be
16635 greater than one. This means that the repositioning command should be
16636 written @code{(forward-char symbol-width)}. The @code{symbol-width}
16637 itself is the length of a @code{graph-blank} and can be found using
16638 the expression @code{(length graph-blank)}. The best place to bind
16639 the @code{symbol-width} variable to the value of the width of graph
16640 column is in the varlist of the @code{let} expression.
16643 These considerations lead to the following function definition:
16647 (defun graph-body-print (numbers-list)
16648 "Print a bar graph of the NUMBERS-LIST.
16649 The numbers-list consists of the Y-axis values."
16651 (let ((height (apply 'max numbers-list))
16652 (symbol-width (length graph-blank))
16657 (while numbers-list
16658 (setq from-position (point))
16660 (column-of-graph height (car numbers-list)))
16661 (goto-char from-position)
16662 (forward-char symbol-width)
16665 ;; @r{Draw graph column by column.}
16667 (setq numbers-list (cdr numbers-list)))
16670 ;; @r{Place point for X axis labels.}
16671 (forward-line height)
16678 The one unexpected expression in this function is the
16679 @w{@code{(sit-for 0)}} expression in the @code{while} loop. This
16680 expression makes the graph printing operation more interesting to
16681 watch than it would be otherwise. The expression causes Emacs to
16682 `sit' or do nothing for a zero length of time and then redraw the
16683 screen. Placed here, it causes Emacs to redraw the screen column by
16684 column. Without it, Emacs would not redraw the screen until the
16687 We can test @code{graph-body-print} with a short list of numbers.
16691 Install @code{graph-symbol}, @code{graph-blank},
16692 @code{column-of-graph}, which are in
16694 @ref{Readying a Graph, , Readying a Graph},
16697 @ref{Columns of a graph},
16699 and @code{graph-body-print}.
16703 Copy the following expression:
16706 (graph-body-print '(1 2 3 4 6 4 3 5 7 6 5 2 3))
16710 Switch to the @file{*scratch*} buffer and place the cursor where you
16711 want the graph to start.
16714 Type @kbd{M-:} (@code{eval-expression}).
16717 Yank the @code{graph-body-print} expression into the minibuffer
16718 with @kbd{C-y} (@code{yank)}.
16721 Press @key{RET} to evaluate the @code{graph-body-print} expression.
16725 Emacs will print a graph like this:
16739 @node recursive-graph-body-print
16740 @section The @code{recursive-graph-body-print} Function
16741 @findex recursive-graph-body-print
16743 The @code{graph-body-print} function may also be written recursively.
16744 The recursive solution is divided into two parts: an outside `wrapper'
16745 that uses a @code{let} expression to determine the values of several
16746 variables that need only be found once, such as the maximum height of
16747 the graph, and an inside function that is called recursively to print
16751 The `wrapper' is uncomplicated:
16755 (defun recursive-graph-body-print (numbers-list)
16756 "Print a bar graph of the NUMBERS-LIST.
16757 The numbers-list consists of the Y-axis values."
16758 (let ((height (apply 'max numbers-list))
16759 (symbol-width (length graph-blank))
16761 (recursive-graph-body-print-internal
16768 The recursive function is a little more difficult. It has four parts:
16769 the `do-again-test', the printing code, the recursive call, and the
16770 `next-step-expression'. The `do-again-test' is a @code{when}
16771 expression that determines whether the @code{numbers-list} contains
16772 any remaining elements; if it does, the function prints one column of
16773 the graph using the printing code and calls itself again. The
16774 function calls itself again according to the value produced by the
16775 `next-step-expression' which causes the call to act on a shorter
16776 version of the @code{numbers-list}.
16780 (defun recursive-graph-body-print-internal
16781 (numbers-list height symbol-width)
16782 "Print a bar graph.
16783 Used within recursive-graph-body-print function."
16788 (setq from-position (point))
16790 (column-of-graph height (car numbers-list)))
16793 (goto-char from-position)
16794 (forward-char symbol-width)
16795 (sit-for 0) ; @r{Draw graph column by column.}
16796 (recursive-graph-body-print-internal
16797 (cdr numbers-list) height symbol-width)))
16802 After installation, this expression can be tested; here is a sample:
16805 (recursive-graph-body-print '(3 2 5 6 7 5 3 4 6 4 3 2 1))
16809 Here is what @code{recursive-graph-body-print} produces:
16823 Either of these two functions, @code{graph-body-print} or
16824 @code{recursive-graph-body-print}, create the body of a graph.
16827 @section Need for Printed Axes
16829 A graph needs printed axes, so you can orient yourself. For a do-once
16830 project, it may be reasonable to draw the axes by hand using Emacs's
16831 Picture mode; but a graph drawing function may be used more than once.
16833 For this reason, I have written enhancements to the basic
16834 @code{print-graph-body} function that automatically print labels for
16835 the horizontal and vertical axes. Since the label printing functions
16836 do not contain much new material, I have placed their description in
16837 an appendix. @xref{Full Graph, , A Graph with Labeled Axes}.
16839 @node Line Graph Exercise
16842 Write a line graph version of the graph printing functions.
16844 @node Emacs Initialization
16845 @chapter Your @file{.emacs} File
16846 @cindex @file{.emacs} file
16847 @cindex Customizing your @file{.emacs} file
16848 @cindex Initialization file
16850 ``You don't have to like Emacs to like it''---this seemingly
16851 paradoxical statement is the secret of GNU Emacs. The plain, `out of
16852 the box' Emacs is a generic tool. Most people who use it, customize
16853 it to suit themselves.
16855 GNU Emacs is mostly written in Emacs Lisp; this means that by writing
16856 expressions in Emacs Lisp you can change or extend Emacs.
16859 * Default Configuration::
16860 * Site-wide Init:: You can write site-wide init files.
16861 * defcustom:: Emacs will write code for you.
16862 * Beginning a .emacs File:: How to write a @code{.emacs file}.
16863 * Text and Auto-fill:: Automatically wrap lines.
16864 * Mail Aliases:: Use abbreviations for email addresses.
16865 * Indent Tabs Mode:: Don't use tabs with @TeX{}
16866 * Keybindings:: Create some personal keybindings.
16867 * Keymaps:: More about key binding.
16868 * Loading Files:: Load (i.e., evaluate) files automatically.
16869 * Autoload:: Make functions available.
16870 * Simple Extension:: Define a function; bind it to a key.
16871 * X11 Colors:: Colors in X.
16873 * Mode Line:: How to customize your mode line.
16877 @node Default Configuration
16878 @unnumberedsec Emacs's Default Configuration
16881 There are those who appreciate Emacs's default configuration. After
16882 all, Emacs starts you in C mode when you edit a C file, starts you in
16883 Fortran mode when you edit a Fortran file, and starts you in
16884 Fundamental mode when you edit an unadorned file. This all makes
16885 sense, if you do not know who is going to use Emacs. Who knows what a
16886 person hopes to do with an unadorned file? Fundamental mode is the
16887 right default for such a file, just as C mode is the right default for
16888 editing C code. (Enough programming languages have syntaxes
16889 that enable them to share or nearly share features, so C mode is
16890 now provided by CC mode, the `C Collection'.)
16892 But when you do know who is going to use Emacs---you,
16893 yourself---then it makes sense to customize Emacs.
16895 For example, I seldom want Fundamental mode when I edit an
16896 otherwise undistinguished file; I want Text mode. This is why I
16897 customize Emacs: so it suits me.
16899 You can customize and extend Emacs by writing or adapting a
16900 @file{~/.emacs} file. This is your personal initialization file; its
16901 contents, written in Emacs Lisp, tell Emacs what to do.@footnote{You
16902 may also add @file{.el} to @file{~/.emacs} and call it a
16903 @file{~/.emacs.el} file. In the past, you were forbidden to type the
16904 extra keystrokes that the name @file{~/.emacs.el} requires, but now
16905 you may. The new format is consistent with the Emacs Lisp file
16906 naming conventions; the old format saves typing.}
16908 A @file{~/.emacs} file contains Emacs Lisp code. You can write this
16909 code yourself; or you can use Emacs's @code{customize} feature to write
16910 the code for you. You can combine your own expressions and
16911 auto-written Customize expressions in your @file{.emacs} file.
16913 (I myself prefer to write my own expressions, except for those,
16914 particularly fonts, that I find easier to manipulate using the
16915 @code{customize} command. I combine the two methods.)
16917 Most of this chapter is about writing expressions yourself. It
16918 describes a simple @file{.emacs} file; for more information, see
16919 @ref{Init File, , The Init File, emacs, The GNU Emacs Manual}, and
16920 @ref{Init File, , The Init File, elisp, The GNU Emacs Lisp Reference
16923 @node Site-wide Init
16924 @section Site-wide Initialization Files
16926 @cindex @file{default.el} init file
16927 @cindex @file{site-init.el} init file
16928 @cindex @file{site-load.el} init file
16929 In addition to your personal initialization file, Emacs automatically
16930 loads various site-wide initialization files, if they exist. These
16931 have the same form as your @file{.emacs} file, but are loaded by
16934 Two site-wide initialization files, @file{site-load.el} and
16935 @file{site-init.el}, are loaded into Emacs and then `dumped' if a
16936 `dumped' version of Emacs is created, as is most common. (Dumped
16937 copies of Emacs load more quickly. However, once a file is loaded and
16938 dumped, a change to it does not lead to a change in Emacs unless you
16939 load it yourself or re-dump Emacs. @xref{Building Emacs, , Building
16940 Emacs, elisp, The GNU Emacs Lisp Reference Manual}, and the
16941 @file{INSTALL} file.)
16943 Three other site-wide initialization files are loaded automatically
16944 each time you start Emacs, if they exist. These are
16945 @file{site-start.el}, which is loaded @emph{before} your @file{.emacs}
16946 file, and @file{default.el}, and the terminal type file, which are both
16947 loaded @emph{after} your @file{.emacs} file.
16949 Settings and definitions in your @file{.emacs} file will overwrite
16950 conflicting settings and definitions in a @file{site-start.el} file,
16951 if it exists; but the settings and definitions in a @file{default.el}
16952 or terminal type file will overwrite those in your @file{.emacs} file.
16953 (You can prevent interference from a terminal type file by setting
16954 @code{term-file-prefix} to @code{nil}. @xref{Simple Extension, , A
16955 Simple Extension}.)
16957 @c Rewritten to avoid overfull hbox.
16958 The @file{INSTALL} file that comes in the distribution contains
16959 descriptions of the @file{site-init.el} and @file{site-load.el} files.
16961 The @file{loadup.el}, @file{startup.el}, and @file{loaddefs.el} files
16962 control loading. These files are in the @file{lisp} directory of the
16963 Emacs distribution and are worth perusing.
16965 The @file{loaddefs.el} file contains a good many suggestions as to
16966 what to put into your own @file{.emacs} file, or into a site-wide
16967 initialization file.
16970 @section Specifying Variables using @code{defcustom}
16973 You can specify variables using @code{defcustom} so that you and
16974 others can then use Emacs's @code{customize} feature to set their
16975 values. (You cannot use @code{customize} to write function
16976 definitions; but you can write @code{defuns} in your @file{.emacs}
16977 file. Indeed, you can write any Lisp expression in your @file{.emacs}
16980 The @code{customize} feature depends on the @code{defcustom} special
16981 form. Although you can use @code{defvar} or @code{setq} for variables
16982 that users set, the @code{defcustom} special form is designed for the
16985 You can use your knowledge of @code{defvar} for writing the
16986 first three arguments for @code{defcustom}. The first argument to
16987 @code{defcustom} is the name of the variable. The second argument is
16988 the variable's initial value, if any; and this value is set only if
16989 the value has not already been set. The third argument is the
16992 The fourth and subsequent arguments to @code{defcustom} specify types
16993 and options; these are not featured in @code{defvar}. (These
16994 arguments are optional.)
16996 Each of these arguments consists of a keyword followed by a value.
16997 Each keyword starts with the colon character @samp{:}.
17000 For example, the customizable user option variable
17001 @code{text-mode-hook} looks like this:
17005 (defcustom text-mode-hook nil
17006 "Normal hook run when entering Text mode and many related modes."
17008 :options '(turn-on-auto-fill flyspell-mode)
17014 The name of the variable is @code{text-mode-hook}; it has no default
17015 value; and its documentation string tells you what it does.
17017 The @code{:type} keyword tells Emacs the kind of data to which
17018 @code{text-mode-hook} should be set and how to display the value in a
17019 Customization buffer.
17021 The @code{:options} keyword specifies a suggested list of values for
17022 the variable. Usually, @code{:options} applies to a hook.
17023 The list is only a suggestion; it is not exclusive; a person who sets
17024 the variable may set it to other values; the list shown following the
17025 @code{:options} keyword is intended to offer convenient choices to a
17028 Finally, the @code{:group} keyword tells the Emacs Customization
17029 command in which group the variable is located. This tells where to
17032 The @code{defcustom} function recognizes more than a dozen keywords.
17033 For more information, see @ref{Customization, , Writing Customization
17034 Definitions, elisp, The GNU Emacs Lisp Reference Manual}.
17036 Consider @code{text-mode-hook} as an example.
17038 There are two ways to customize this variable. You can use the
17039 customization command or write the appropriate expressions yourself.
17042 Using the customization command, you can type:
17049 and find that the group for editing files of data is called `data'.
17050 Enter that group. Text Mode Hook is the first member. You can click
17051 on its various options, such as @code{turn-on-auto-fill}, to set the
17052 values. After you click on the button to
17055 Save for Future Sessions
17059 Emacs will write an expression into your @file{.emacs} file.
17060 It will look like this:
17064 (custom-set-variables
17065 ;; custom-set-variables was added by Custom.
17066 ;; If you edit it by hand, you could mess it up, so be careful.
17067 ;; Your init file should contain only one such instance.
17068 ;; If there is more than one, they won't work right.
17069 '(text-mode-hook (quote (turn-on-auto-fill text-mode-hook-identify))))
17074 (The @code{text-mode-hook-identify} function tells
17075 @code{toggle-text-mode-auto-fill} which buffers are in Text mode.
17076 It comes on automatically.)
17078 The @code{custom-set-variables} function works somewhat differently
17079 than a @code{setq}. While I have never learned the differences, I
17080 modify the @code{custom-set-variables} expressions in my @file{.emacs}
17081 file by hand: I make the changes in what appears to me to be a
17082 reasonable manner and have not had any problems. Others prefer to use
17083 the Customization command and let Emacs do the work for them.
17085 Another @code{custom-set-@dots{}} function is @code{custom-set-faces}.
17086 This function sets the various font faces. Over time, I have set a
17087 considerable number of faces. Some of the time, I re-set them using
17088 @code{customize}; other times, I simply edit the
17089 @code{custom-set-faces} expression in my @file{.emacs} file itself.
17091 The second way to customize your @code{text-mode-hook} is to set it
17092 yourself in your @file{.emacs} file using code that has nothing to do
17093 with the @code{custom-set-@dots{}} functions.
17096 When you do this, and later use @code{customize}, you will see a
17100 CHANGED outside Customize; operating on it here may be unreliable.
17104 This message is only a warning. If you click on the button to
17107 Save for Future Sessions
17111 Emacs will write a @code{custom-set-@dots{}} expression near the end
17112 of your @file{.emacs} file that will be evaluated after your
17113 hand-written expression. It will, therefore, overrule your
17114 hand-written expression. No harm will be done. When you do this,
17115 however, be careful to remember which expression is active; if you
17116 forget, you may confuse yourself.
17118 So long as you remember where the values are set, you will have no
17119 trouble. In any event, the values are always set in your
17120 initialization file, which is usually called @file{.emacs}.
17122 I myself use @code{customize} for hardly anything. Mostly, I write
17123 expressions myself.
17127 Incidentally, to be more complete concerning defines: @code{defsubst}
17128 defines an inline function. The syntax is just like that of
17129 @code{defun}. @code{defconst} defines a symbol as a constant. The
17130 intent is that neither programs nor users should ever change a value
17131 set by @code{defconst}. (You can change it; the value set is a
17132 variable; but please do not.)
17134 @node Beginning a .emacs File
17135 @section Beginning a @file{.emacs} File
17136 @cindex @file{.emacs} file, beginning of
17138 When you start Emacs, it loads your @file{.emacs} file unless you tell
17139 it not to by specifying @samp{-q} on the command line. (The
17140 @code{emacs -q} command gives you a plain, out-of-the-box Emacs.)
17142 A @file{.emacs} file contains Lisp expressions. Often, these are no
17143 more than expressions to set values; sometimes they are function
17146 @xref{Init File, , The Init File @file{~/.emacs}, emacs, The GNU Emacs
17147 Manual}, for a short description of initialization files.
17149 This chapter goes over some of the same ground, but is a walk among
17150 extracts from a complete, long-used @file{.emacs} file---my own.
17152 The first part of the file consists of comments: reminders to myself.
17153 By now, of course, I remember these things, but when I started, I did
17159 ;;;; Bob's .emacs file
17160 ; Robert J. Chassell
17161 ; 26 September 1985
17166 Look at that date! I started this file a long time ago. I have been
17167 adding to it ever since.
17171 ; Each section in this file is introduced by a
17172 ; line beginning with four semicolons; and each
17173 ; entry is introduced by a line beginning with
17174 ; three semicolons.
17179 This describes the usual conventions for comments in Emacs Lisp.
17180 Everything on a line that follows a semicolon is a comment. Two,
17181 three, and four semicolons are used as subsection and section markers.
17182 (@xref{Comments, ,, elisp, The GNU Emacs Lisp Reference Manual}, for
17183 more about comments.)
17188 ; Control-h is the help key;
17189 ; after typing control-h, type a letter to
17190 ; indicate the subject about which you want help.
17191 ; For an explanation of the help facility,
17192 ; type control-h two times in a row.
17197 Just remember: type @kbd{C-h} two times for help.
17201 ; To find out about any mode, type control-h m
17202 ; while in that mode. For example, to find out
17203 ; about mail mode, enter mail mode and then type
17209 `Mode help', as I call this, is very helpful. Usually, it tells you
17210 all you need to know.
17212 Of course, you don't need to include comments like these in your
17213 @file{.emacs} file. I included them in mine because I kept forgetting
17214 about Mode help or the conventions for comments---but I was able to
17215 remember to look here to remind myself.
17217 @node Text and Auto-fill
17218 @section Text and Auto Fill Mode
17220 Now we come to the part that `turns on' Text mode and
17225 ;;; Text mode and Auto Fill mode
17226 ;; The next two lines put Emacs into Text mode
17227 ;; and Auto Fill mode, and are for writers who
17228 ;; want to start writing prose rather than code.
17229 (setq-default major-mode 'text-mode)
17230 (add-hook 'text-mode-hook 'turn-on-auto-fill)
17234 Here is the first part of this @file{.emacs} file that does something
17235 besides remind a forgetful human!
17237 The first of the two lines in parentheses tells Emacs to turn on Text
17238 mode when you find a file, @emph{unless} that file should go into some
17239 other mode, such as C mode.
17241 @cindex Per-buffer, local variables list
17242 @cindex Local variables list, per-buffer,
17243 @cindex Automatic mode selection
17244 @cindex Mode selection, automatic
17245 When Emacs reads a file, it looks at the extension to the file name,
17246 if any. (The extension is the part that comes after a @samp{.}.) If
17247 the file ends with a @samp{.c} or @samp{.h} extension then Emacs turns
17248 on C mode. Also, Emacs looks at first nonblank line of the file; if
17249 the line says @w{@samp{-*- C -*-}}, Emacs turns on C mode. Emacs
17250 possesses a list of extensions and specifications that it uses
17251 automatically. In addition, Emacs looks near the last page for a
17252 per-buffer, ``local variables list'', if any.
17255 @xref{Choosing Modes, , How Major Modes are Chosen, emacs, The GNU
17258 @xref{File Variables, , Local Variables in Files, emacs, The GNU Emacs
17262 See sections ``How Major Modes are Chosen'' and ``Local Variables in
17263 Files'' in @cite{The GNU Emacs Manual}.
17266 Now, back to the @file{.emacs} file.
17269 Here is the line again; how does it work?
17271 @cindex Text Mode turned on
17273 (setq major-mode 'text-mode)
17277 This line is a short, but complete Emacs Lisp expression.
17279 We are already familiar with @code{setq}. It sets the following variable,
17280 @code{major-mode}, to the subsequent value, which is @code{text-mode}.
17281 The single quote mark before @code{text-mode} tells Emacs to deal directly
17282 with the @code{text-mode} symbol, not with whatever it might stand for.
17283 @xref{set & setq, , Setting the Value of a Variable},
17284 for a reminder of how @code{setq} works.
17285 The main point is that there is no difference between the procedure you
17286 use to set a value in your @file{.emacs} file and the procedure you use
17287 anywhere else in Emacs.
17290 Here is the next line:
17292 @cindex Auto Fill mode turned on
17295 (add-hook 'text-mode-hook 'turn-on-auto-fill)
17299 In this line, the @code{add-hook} command adds
17300 @code{turn-on-auto-fill} to the variable.
17302 @code{turn-on-auto-fill} is the name of a program, that, you guessed
17303 it!, turns on Auto Fill mode.
17305 Every time Emacs turns on Text mode, Emacs runs the commands `hooked'
17306 onto Text mode. So every time Emacs turns on Text mode, Emacs also
17307 turns on Auto Fill mode.
17309 In brief, the first line causes Emacs to enter Text mode when you edit a
17310 file, unless the file name extension, a first non-blank line, or local
17311 variables to tell Emacs otherwise.
17313 Text mode among other actions, sets the syntax table to work
17314 conveniently for writers. In Text mode, Emacs considers an apostrophe
17315 as part of a word like a letter; but Emacs does not consider a period
17316 or a space as part of a word. Thus, @kbd{M-f} moves you over
17317 @samp{it's}. On the other hand, in C mode, @kbd{M-f} stops just after
17318 the @samp{t} of @samp{it's}.
17320 The second line causes Emacs to turn on Auto Fill mode when it turns
17321 on Text mode. In Auto Fill mode, Emacs automatically breaks a line
17322 that is too wide and brings the excessively wide part of the line down
17323 to the next line. Emacs breaks lines between words, not within them.
17325 When Auto Fill mode is turned off, lines continue to the right as you
17326 type them. Depending on how you set the value of
17327 @code{truncate-lines}, the words you type either disappear off the
17328 right side of the screen, or else are shown, in a rather ugly and
17329 unreadable manner, as a continuation line on the screen.
17332 In addition, in this part of my @file{.emacs} file, I tell the Emacs
17333 fill commands to insert two spaces after a colon:
17336 (setq colon-double-space t)
17340 @section Mail Aliases
17342 Here is a @code{setq} that `turns on' mail aliases, along with more
17348 ; To enter mail mode, type `C-x m'
17349 ; To enter RMAIL (for reading mail),
17351 (setq mail-aliases t)
17355 @cindex Mail aliases
17357 This @code{setq} command sets the value of the variable
17358 @code{mail-aliases} to @code{t}. Since @code{t} means true, the line
17359 says, in effect, ``Yes, use mail aliases.''
17361 Mail aliases are convenient short names for long email addresses or
17362 for lists of email addresses. The file where you keep your `aliases'
17363 is @file{~/.mailrc}. You write an alias like this:
17366 alias geo george@@foobar.wiz.edu
17370 When you write a message to George, address it to @samp{geo}; the
17371 mailer will automatically expand @samp{geo} to the full address.
17373 @node Indent Tabs Mode
17374 @section Indent Tabs Mode
17375 @cindex Tabs, preventing
17376 @findex indent-tabs-mode
17378 By default, Emacs inserts tabs in place of multiple spaces when it
17379 formats a region. (For example, you might indent many lines of text
17380 all at once with the @code{indent-region} command.) Tabs look fine on
17381 a terminal or with ordinary printing, but they produce badly indented
17382 output when you use @TeX{} or Texinfo since @TeX{} ignores tabs.
17385 The following turns off Indent Tabs mode:
17389 ;;; Prevent Extraneous Tabs
17390 (setq-default indent-tabs-mode nil)
17394 Note that this line uses @code{setq-default} rather than the
17395 @code{setq} command that we have seen before. The @code{setq-default}
17396 command sets values only in buffers that do not have their own local
17397 values for the variable.
17400 @xref{Just Spaces, , Tabs vs. Spaces, emacs, The GNU Emacs Manual}.
17402 @xref{File Variables, , Local Variables in Files, emacs, The GNU Emacs
17406 See sections ``Tabs vs.@: Spaces'' and ``Local Variables in
17407 Files'' in @cite{The GNU Emacs Manual}.
17412 @section Some Keybindings
17414 Now for some personal keybindings:
17418 ;;; Compare windows
17419 (global-set-key "\C-cw" 'compare-windows)
17423 @findex compare-windows
17424 @code{compare-windows} is a nifty command that compares the text in
17425 your current window with text in the next window. It makes the
17426 comparison by starting at point in each window, moving over text in
17427 each window as far as they match. I use this command all the time.
17429 This also shows how to set a key globally, for all modes.
17431 @cindex Setting a key globally
17432 @cindex Global set key
17433 @cindex Key setting globally
17434 @findex global-set-key
17435 The command is @code{global-set-key}. It is followed by the
17436 keybinding. In a @file{.emacs} file, the keybinding is written as
17437 shown: @code{\C-c} stands for `control-c', which means `press the
17438 control key and the @key{c} key at the same time'. The @code{w} means
17439 `press the @key{w} key'. The keybinding is surrounded by double
17440 quotation marks. In documentation, you would write this as
17441 @w{@kbd{C-c w}}. (If you were binding a @key{META} key, such as
17442 @kbd{M-c}, rather than a @key{CTRL} key, you would write
17443 @w{@code{\M-c}} in your @file{.emacs} file. @xref{Init Rebinding, ,
17444 Rebinding Keys in Your Init File, emacs, The GNU Emacs Manual}, for
17447 The command invoked by the keys is @code{compare-windows}. Note that
17448 @code{compare-windows} is preceded by a single quote; otherwise, Emacs
17449 would first try to evaluate the symbol to determine its value.
17451 These three things, the double quotation marks, the backslash before
17452 the @samp{C}, and the single quote mark are necessary parts of
17453 keybinding that I tend to forget. Fortunately, I have come to
17454 remember that I should look at my existing @file{.emacs} file, and
17455 adapt what is there.
17457 As for the keybinding itself: @kbd{C-c w}. This combines the prefix
17458 key, @kbd{C-c}, with a single character, in this case, @kbd{w}. This
17459 set of keys, @kbd{C-c} followed by a single character, is strictly
17460 reserved for individuals' own use. (I call these `own' keys, since
17461 these are for my own use.) You should always be able to create such a
17462 keybinding for your own use without stomping on someone else's
17463 keybinding. If you ever write an extension to Emacs, please avoid
17464 taking any of these keys for public use. Create a key like @kbd{C-c
17465 C-w} instead. Otherwise, we will run out of `own' keys.
17468 Here is another keybinding, with a comment:
17472 ;;; Keybinding for `occur'
17473 ; I use occur a lot, so let's bind it to a key:
17474 (global-set-key "\C-co" 'occur)
17479 The @code{occur} command shows all the lines in the current buffer
17480 that contain a match for a regular expression. Matching lines are
17481 shown in a buffer called @file{*Occur*}. That buffer serves as a menu
17482 to jump to occurrences.
17484 @findex global-unset-key
17485 @cindex Unbinding key
17486 @cindex Key unbinding
17488 Here is how to unbind a key, so it does not
17494 (global-unset-key "\C-xf")
17498 There is a reason for this unbinding: I found I inadvertently typed
17499 @w{@kbd{C-x f}} when I meant to type @kbd{C-x C-f}. Rather than find a
17500 file, as I intended, I accidentally set the width for filled text,
17501 almost always to a width I did not want. Since I hardly ever reset my
17502 default width, I simply unbound the key.
17504 @findex list-buffers, @r{rebound}
17505 @findex buffer-menu, @r{bound to key}
17507 The following rebinds an existing key:
17511 ;;; Rebind `C-x C-b' for `buffer-menu'
17512 (global-set-key "\C-x\C-b" 'buffer-menu)
17516 By default, @kbd{C-x C-b} runs the
17517 @code{list-buffers} command. This command lists
17518 your buffers in @emph{another} window. Since I
17519 almost always want to do something in that
17520 window, I prefer the @code{buffer-menu}
17521 command, which not only lists the buffers,
17522 but moves point into that window.
17527 @cindex Rebinding keys
17529 Emacs uses @dfn{keymaps} to record which keys call which commands.
17530 When you use @code{global-set-key} to set the keybinding for a single
17531 command in all parts of Emacs, you are specifying the keybinding in
17532 @code{current-global-map}.
17534 Specific modes, such as C mode or Text mode, have their own keymaps;
17535 the mode-specific keymaps override the global map that is shared by
17538 The @code{global-set-key} function binds, or rebinds, the global
17539 keymap. For example, the following binds the key @kbd{C-x C-b} to the
17540 function @code{buffer-menu}:
17543 (global-set-key "\C-x\C-b" 'buffer-menu)
17546 Mode-specific keymaps are bound using the @code{define-key} function,
17547 which takes a specific keymap as an argument, as well as the key and
17548 the command. For example, my @file{.emacs} file contains the
17549 following expression to bind the @code{texinfo-insert-@@group} command
17550 to @kbd{C-c C-c g}:
17554 (define-key texinfo-mode-map "\C-c\C-cg" 'texinfo-insert-@@group)
17559 The @code{texinfo-insert-@@group} function itself is a little extension
17560 to Texinfo mode that inserts @samp{@@group} into a Texinfo file. I
17561 use this command all the time and prefer to type the three strokes
17562 @kbd{C-c C-c g} rather than the six strokes @kbd{@@ g r o u p}.
17563 (@samp{@@group} and its matching @samp{@@end group} are commands that
17564 keep all enclosed text together on one page; many multi-line examples
17565 in this book are surrounded by @samp{@@group @dots{} @@end group}.)
17568 Here is the @code{texinfo-insert-@@group} function definition:
17572 (defun texinfo-insert-@@group ()
17573 "Insert the string @@group in a Texinfo buffer."
17575 (beginning-of-line)
17576 (insert "@@group\n"))
17580 (Of course, I could have used Abbrev mode to save typing, rather than
17581 write a function to insert a word; but I prefer key strokes consistent
17582 with other Texinfo mode key bindings.)
17584 You will see numerous @code{define-key} expressions in
17585 @file{loaddefs.el} as well as in the various mode libraries, such as
17586 @file{cc-mode.el} and @file{lisp-mode.el}.
17588 @xref{Key Bindings, , Customizing Key Bindings, emacs, The GNU Emacs
17589 Manual}, and @ref{Keymaps, , Keymaps, elisp, The GNU Emacs Lisp
17590 Reference Manual}, for more information about keymaps.
17592 @node Loading Files
17593 @section Loading Files
17594 @cindex Loading files
17597 Many people in the GNU Emacs community have written extensions to
17598 Emacs. As time goes by, these extensions are often included in new
17599 releases. For example, the Calendar and Diary packages are now part
17600 of the standard GNU Emacs, as is Calc.
17602 You can use a @code{load} command to evaluate a complete file and
17603 thereby install all the functions and variables in the file into Emacs.
17606 @c (auto-compression-mode t)
17609 (load "~/emacs/slowsplit")
17612 This evaluates, i.e., loads, the @file{slowsplit.el} file or if it
17613 exists, the faster, byte compiled @file{slowsplit.elc} file from the
17614 @file{emacs} sub-directory of your home directory. The file contains
17615 the function @code{split-window-quietly}, which John Robinson wrote in
17618 The @code{split-window-quietly} function splits a window with the
17619 minimum of redisplay. I installed it in 1989 because it worked well
17620 with the slow 1200 baud terminals I was then using. Nowadays, I only
17621 occasionally come across such a slow connection, but I continue to use
17622 the function because I like the way it leaves the bottom half of a
17623 buffer in the lower of the new windows and the top half in the upper
17627 To replace the key binding for the default
17628 @code{split-window-vertically}, you must also unset that key and bind
17629 the keys to @code{split-window-quietly}, like this:
17633 (global-unset-key "\C-x2")
17634 (global-set-key "\C-x2" 'split-window-quietly)
17639 If you load many extensions, as I do, then instead of specifying the
17640 exact location of the extension file, as shown above, you can specify
17641 that directory as part of Emacs's @code{load-path}. Then, when Emacs
17642 loads a file, it will search that directory as well as its default
17643 list of directories. (The default list is specified in @file{paths.h}
17644 when Emacs is built.)
17647 The following command adds your @file{~/emacs} directory to the
17648 existing load path:
17652 ;;; Emacs Load Path
17653 (setq load-path (cons "~/emacs" load-path))
17657 Incidentally, @code{load-library} is an interactive interface to the
17658 @code{load} function. The complete function looks like this:
17660 @findex load-library
17663 (defun load-library (library)
17664 "Load the library named LIBRARY.
17665 This is an interface to the function `load'."
17667 (list (completing-read "Load library: "
17668 (apply-partially 'locate-file-completion-table
17670 (get-load-suffixes)))))
17675 The name of the function, @code{load-library}, comes from the use of
17676 `library' as a conventional synonym for `file'. The source for the
17677 @code{load-library} command is in the @file{files.el} library.
17679 Another interactive command that does a slightly different job is
17680 @code{load-file}. @xref{Lisp Libraries, , Libraries of Lisp Code for
17681 Emacs, emacs, The GNU Emacs Manual}, for information on the
17682 distinction between @code{load-library} and this command.
17685 @section Autoloading
17688 Instead of installing a function by loading the file that contains it,
17689 or by evaluating the function definition, you can make the function
17690 available but not actually install it until it is first called. This
17691 is called @dfn{autoloading}.
17693 When you execute an autoloaded function, Emacs automatically evaluates
17694 the file that contains the definition, and then calls the function.
17696 Emacs starts quicker with autoloaded functions, since their libraries
17697 are not loaded right away; but you need to wait a moment when you
17698 first use such a function, while its containing file is evaluated.
17700 Rarely used functions are frequently autoloaded. The
17701 @file{loaddefs.el} library contains hundreds of autoloaded functions,
17702 from @code{bookmark-set} to @code{wordstar-mode}. Of course, you may
17703 come to use a `rare' function frequently. When you do, you should
17704 load that function's file with a @code{load} expression in your
17705 @file{.emacs} file.
17707 In my @file{.emacs} file, I load 14 libraries that contain functions
17708 that would otherwise be autoloaded. (Actually, it would have been
17709 better to include these files in my `dumped' Emacs, but I forgot.
17710 @xref{Building Emacs, , Building Emacs, elisp, The GNU Emacs Lisp
17711 Reference Manual}, and the @file{INSTALL} file for more about
17714 You may also want to include autoloaded expressions in your @file{.emacs}
17715 file. @code{autoload} is a built-in function that takes up to five
17716 arguments, the final three of which are optional. The first argument
17717 is the name of the function to be autoloaded; the second is the name
17718 of the file to be loaded. The third argument is documentation for the
17719 function, and the fourth tells whether the function can be called
17720 interactively. The fifth argument tells what type of
17721 object---@code{autoload} can handle a keymap or macro as well as a
17722 function (the default is a function).
17725 Here is a typical example:
17729 (autoload 'html-helper-mode
17730 "html-helper-mode" "Edit HTML documents" t)
17735 (@code{html-helper-mode} is an older alternative to @code{html-mode},
17736 which is a standard part of the distribution.)
17739 This expression autoloads the @code{html-helper-mode} function. It
17740 takes it from the @file{html-helper-mode.el} file (or from the byte
17741 compiled version @file{html-helper-mode.elc}, if that exists.) The
17742 file must be located in a directory specified by @code{load-path}.
17743 The documentation says that this is a mode to help you edit documents
17744 written in the HyperText Markup Language. You can call this mode
17745 interactively by typing @kbd{M-x html-helper-mode}. (You need to
17746 duplicate the function's regular documentation in the autoload
17747 expression because the regular function is not yet loaded, so its
17748 documentation is not available.)
17750 @xref{Autoload, , Autoload, elisp, The GNU Emacs Lisp Reference
17751 Manual}, for more information.
17753 @node Simple Extension
17754 @section A Simple Extension: @code{line-to-top-of-window}
17755 @findex line-to-top-of-window
17756 @cindex Simple extension in @file{.emacs} file
17758 Here is a simple extension to Emacs that moves the line point is on to
17759 the top of the window. I use this all the time, to make text easier
17762 You can put the following code into a separate file and then load it
17763 from your @file{.emacs} file, or you can include it within your
17764 @file{.emacs} file.
17767 Here is the definition:
17771 ;;; Line to top of window;
17772 ;;; replace three keystroke sequence C-u 0 C-l
17773 (defun line-to-top-of-window ()
17774 "Move the line point is on to top of window."
17781 Now for the keybinding.
17783 Nowadays, function keys as well as mouse button events and
17784 non-@sc{ascii} characters are written within square brackets, without
17785 quotation marks. (In Emacs version 18 and before, you had to write
17786 different function key bindings for each different make of terminal.)
17788 I bind @code{line-to-top-of-window} to my @key{F6} function key like
17792 (global-set-key [f6] 'line-to-top-of-window)
17795 For more information, see @ref{Init Rebinding, , Rebinding Keys in
17796 Your Init File, emacs, The GNU Emacs Manual}.
17798 @cindex Conditional 'twixt two versions of Emacs
17799 @cindex Version of Emacs, choosing
17800 @cindex Emacs version, choosing
17801 If you run two versions of GNU Emacs, such as versions 22 and 23, and
17802 use one @file{.emacs} file, you can select which code to evaluate with
17803 the following conditional:
17808 ((= 22 emacs-major-version)
17809 ;; evaluate version 22 code
17811 ((= 23 emacs-major-version)
17812 ;; evaluate version 23 code
17817 For example, recent versions blink
17818 their cursors by default. I hate such blinking, as well as other
17819 features, so I placed the following in my @file{.emacs}
17820 file@footnote{When I start instances of Emacs that do not load my
17821 @file{.emacs} file or any site file, I also turn off blinking:
17824 emacs -q --no-site-file -eval '(blink-cursor-mode nil)'
17826 @exdent Or nowadays, using an even more sophisticated set of options,
17834 (when (>= emacs-major-version 21)
17835 (blink-cursor-mode 0)
17836 ;; Insert newline when you press `C-n' (next-line)
17837 ;; at the end of the buffer
17838 (setq next-line-add-newlines t)
17841 ;; Turn on image viewing
17842 (auto-image-file-mode t)
17845 ;; Turn on menu bar (this bar has text)
17846 ;; (Use numeric argument to turn on)
17850 ;; Turn off tool bar (this bar has icons)
17851 ;; (Use numeric argument to turn on)
17852 (tool-bar-mode nil)
17855 ;; Turn off tooltip mode for tool bar
17856 ;; (This mode causes icon explanations to pop up)
17857 ;; (Use numeric argument to turn on)
17859 ;; If tooltips turned on, make tips appear promptly
17860 (setq tooltip-delay 0.1) ; default is 0.7 second
17866 @section X11 Colors
17868 You can specify colors when you use Emacs with the MIT X Windowing
17871 I dislike the default colors and specify my own.
17874 Here are the expressions in my @file{.emacs}
17875 file that set values:
17879 ;; Set cursor color
17880 (set-cursor-color "white")
17883 (set-mouse-color "white")
17885 ;; Set foreground and background
17886 (set-foreground-color "white")
17887 (set-background-color "darkblue")
17891 ;;; Set highlighting colors for isearch and drag
17892 (set-face-foreground 'highlight "white")
17893 (set-face-background 'highlight "blue")
17897 (set-face-foreground 'region "cyan")
17898 (set-face-background 'region "blue")
17902 (set-face-foreground 'secondary-selection "skyblue")
17903 (set-face-background 'secondary-selection "darkblue")
17907 ;; Set calendar highlighting colors
17908 (setq calendar-load-hook
17910 (set-face-foreground 'diary-face "skyblue")
17911 (set-face-background 'holiday-face "slate blue")
17912 (set-face-foreground 'holiday-face "white")))
17916 The various shades of blue soothe my eye and prevent me from seeing
17917 the screen flicker.
17919 Alternatively, I could have set my specifications in various X
17920 initialization files. For example, I could set the foreground,
17921 background, cursor, and pointer (i.e., mouse) colors in my
17922 @file{~/.Xresources} file like this:
17926 Emacs*foreground: white
17927 Emacs*background: darkblue
17928 Emacs*cursorColor: white
17929 Emacs*pointerColor: white
17933 In any event, since it is not part of Emacs, I set the root color of
17934 my X window in my @file{~/.xinitrc} file, like this@footnote{I also
17935 run more modern window managers, such as Enlightenment, Gnome, or KDE;
17936 in those cases, I often specify an image rather than a plain color.}:
17939 xsetroot -solid Navy -fg white &
17943 @node Miscellaneous
17944 @section Miscellaneous Settings for a @file{.emacs} File
17947 Here are a few miscellaneous settings:
17952 Set the shape and color of the mouse cursor:
17956 ; Cursor shapes are defined in
17957 ; `/usr/include/X11/cursorfont.h';
17958 ; for example, the `target' cursor is number 128;
17959 ; the `top_left_arrow' cursor is number 132.
17963 (let ((mpointer (x-get-resource "*mpointer"
17964 "*emacs*mpointer")))
17965 ;; If you have not set your mouse pointer
17966 ;; then set it, otherwise leave as is:
17967 (if (eq mpointer nil)
17968 (setq mpointer "132")) ; top_left_arrow
17971 (setq x-pointer-shape (string-to-int mpointer))
17972 (set-mouse-color "white"))
17977 Or you can set the values of a variety of features in an alist, like
17983 default-frame-alist
17984 '((cursor-color . "white")
17985 (mouse-color . "white")
17986 (foreground-color . "white")
17987 (background-color . "DodgerBlue4")
17988 ;; (cursor-type . bar)
17989 (cursor-type . box)
17992 (tool-bar-lines . 0)
17993 (menu-bar-lines . 1)
17997 "-Misc-Fixed-Medium-R-Normal--20-200-75-75-C-100-ISO8859-1")
18003 Convert @kbd{@key{CTRL}-h} into @key{DEL} and @key{DEL}
18004 into @kbd{@key{CTRL}-h}.@*
18005 (Some older keyboards needed this, although I have not seen the
18010 ;; Translate `C-h' to <DEL>.
18011 ; (keyboard-translate ?\C-h ?\C-?)
18013 ;; Translate <DEL> to `C-h'.
18014 (keyboard-translate ?\C-? ?\C-h)
18018 @item Turn off a blinking cursor!
18022 (if (fboundp 'blink-cursor-mode)
18023 (blink-cursor-mode -1))
18028 or start GNU Emacs with the command @code{emacs -nbc}.
18031 @item When using `grep'@*
18032 @samp{-i}@w{ } Ignore case distinctions@*
18033 @samp{-n}@w{ } Prefix each line of output with line number@*
18034 @samp{-H}@w{ } Print the filename for each match.@*
18035 @samp{-e}@w{ } Protect patterns beginning with a hyphen character, @samp{-}
18038 (setq grep-command "grep -i -nH -e ")
18042 @c Evidently, no longer needed in GNU Emacs 22
18044 item Automatically uncompress compressed files when visiting them
18047 (load "uncompress")
18052 @item Find an existing buffer, even if it has a different name@*
18053 This avoids problems with symbolic links.
18056 (setq find-file-existing-other-name t)
18059 @item Set your language environment and default input method
18063 (set-language-environment "latin-1")
18064 ;; Remember you can enable or disable multilingual text input
18065 ;; with the @code{toggle-input-method'} (@kbd{C-\}) command
18066 (setq default-input-method "latin-1-prefix")
18070 If you want to write with Chinese `GB' characters, set this instead:
18074 (set-language-environment "Chinese-GB")
18075 (setq default-input-method "chinese-tonepy")
18080 @subsubheading Fixing Unpleasant Key Bindings
18081 @cindex Key bindings, fixing
18082 @cindex Bindings, key, fixing unpleasant
18084 Some systems bind keys unpleasantly. Sometimes, for example, the
18085 @key{CTRL} key appears in an awkward spot rather than at the far left
18088 Usually, when people fix these sorts of keybindings, they do not
18089 change their @file{~/.emacs} file. Instead, they bind the proper keys
18090 on their consoles with the @code{loadkeys} or @code{install-keymap}
18091 commands in their boot script and then include @code{xmodmap} commands
18092 in their @file{.xinitrc} or @file{.Xsession} file for X Windows.
18100 loadkeys /usr/share/keymaps/i386/qwerty/emacs2.kmap.gz
18102 install-keymap emacs2
18108 For a @file{.xinitrc} or @file{.Xsession} file when the @key{Caps
18109 Lock} key is at the far left of the home row:
18113 # Bind the key labeled `Caps Lock' to `Control'
18114 # (Such a broken user interface suggests that keyboard manufacturers
18115 # think that computers are typewriters from 1885.)
18117 xmodmap -e "clear Lock"
18118 xmodmap -e "add Control = Caps_Lock"
18124 In a @file{.xinitrc} or @file{.Xsession} file, to convert an @key{ALT}
18125 key to a @key{META} key:
18129 # Some ill designed keyboards have a key labeled ALT and no Meta
18130 xmodmap -e "keysym Alt_L = Meta_L Alt_L"
18136 @section A Modified Mode Line
18137 @vindex mode-line-format
18138 @cindex Mode line format
18140 Finally, a feature I really like: a modified mode line.
18142 When I work over a network, I forget which machine I am using. Also,
18143 I tend to I lose track of where I am, and which line point is on.
18145 So I reset my mode line to look like this:
18148 -:-- foo.texi rattlesnake:/home/bob/ Line 1 (Texinfo Fill) Top
18151 I am visiting a file called @file{foo.texi}, on my machine
18152 @file{rattlesnake} in my @file{/home/bob} buffer. I am on line 1, in
18153 Texinfo mode, and am at the top of the buffer.
18156 My @file{.emacs} file has a section that looks like this:
18160 ;; Set a Mode Line that tells me which machine, which directory,
18161 ;; and which line I am on, plus the other customary information.
18162 (setq-default mode-line-format
18166 "mouse-1: select window, mouse-2: delete others ..."))
18167 mode-line-mule-info
18169 mode-line-frame-identification
18173 mode-line-buffer-identification
18176 (system-name) 0 (string-match "\\..+" (system-name))))
18181 "mouse-1: select window, mouse-2: delete others ..."))
18182 (line-number-mode " Line %l ")
18188 "mouse-1: select window, mouse-2: delete others ..."))
18189 (:eval (mode-line-mode-name))
18192 #("%n" 0 2 (help-echo "mouse-2: widen" local-map (keymap ...)))
18201 Here, I redefine the default mode line. Most of the parts are from
18202 the original; but I make a few changes. I set the @emph{default} mode
18203 line format so as to permit various modes, such as Info, to override
18206 Many elements in the list are self-explanatory:
18207 @code{mode-line-modified} is a variable that tells whether the buffer
18208 has been modified, @code{mode-name} tells the name of the mode, and so
18209 on. However, the format looks complicated because of two features we
18210 have not discussed.
18212 @cindex Properties, in mode line example
18213 The first string in the mode line is a dash or hyphen, @samp{-}. In
18214 the old days, it would have been specified simply as @code{"-"}. But
18215 nowadays, Emacs can add properties to a string, such as highlighting
18216 or, as in this case, a help feature. If you place your mouse cursor
18217 over the hyphen, some help information appears (By default, you must
18218 wait seven-tenths of a second before the information appears. You can
18219 change that timing by changing the value of @code{tooltip-delay}.)
18222 The new string format has a special syntax:
18225 #("-" 0 1 (help-echo "mouse-1: select window, ..."))
18229 The @code{#(} begins a list. The first element of the list is the
18230 string itself, just one @samp{-}. The second and third
18231 elements specify the range over which the fourth element applies. A
18232 range starts @emph{after} a character, so a zero means the range
18233 starts just before the first character; a 1 means that the range ends
18234 just after the first character. The third element is the property for
18235 the range. It consists of a property list, a
18236 property name, in this case, @samp{help-echo}, followed by a value, in this
18237 case, a string. The second, third, and fourth elements of this new
18238 string format can be repeated.
18240 @xref{Text Properties, , Text Properties, elisp, The GNU Emacs Lisp
18241 Reference Manual}, and see @ref{Mode Line Format, , Mode Line Format,
18242 elisp, The GNU Emacs Lisp Reference Manual}, for more information.
18244 @code{mode-line-buffer-identification}
18245 displays the current buffer name. It is a list
18246 beginning @code{(#("%12b" 0 4 @dots{}}.
18247 The @code{#(} begins the list.
18249 The @samp{"%12b"} displays the current buffer name, using the
18250 @code{buffer-name} function with which we are familiar; the `12'
18251 specifies the maximum number of characters that will be displayed.
18252 When a name has fewer characters, whitespace is added to fill out to
18253 this number. (Buffer names can and often should be longer than 12
18254 characters; this length works well in a typical 80 column wide
18257 @code{:eval} says to evaluate the following form and use the result as
18258 a string to display. In this case, the expression displays the first
18259 component of the full system name. The end of the first component is
18260 a @samp{.} (`period'), so I use the @code{string-match} function to
18261 tell me the length of the first component. The substring from the
18262 zeroth character to that length is the name of the machine.
18265 This is the expression:
18270 (system-name) 0 (string-match "\\..+" (system-name))))
18274 @samp{%[} and @samp{%]} cause a pair of square brackets
18275 to appear for each recursive editing level. @samp{%n} says `Narrow'
18276 when narrowing is in effect. @samp{%P} tells you the percentage of
18277 the buffer that is above the bottom of the window, or `Top', `Bottom',
18278 or `All'. (A lower case @samp{p} tell you the percentage above the
18279 @emph{top} of the window.) @samp{%-} inserts enough dashes to fill
18282 Remember, ``You don't have to like Emacs to like it''---your own
18283 Emacs can have different colors, different commands, and different
18284 keys than a default Emacs.
18286 On the other hand, if you want to bring up a plain `out of the box'
18287 Emacs, with no customization, type:
18294 This will start an Emacs that does @emph{not} load your
18295 @file{~/.emacs} initialization file. A plain, default Emacs. Nothing
18302 GNU Emacs has two debuggers, @code{debug} and @code{edebug}. The
18303 first is built into the internals of Emacs and is always with you;
18304 the second requires that you instrument a function before you can use it.
18306 Both debuggers are described extensively in @ref{Debugging, ,
18307 Debugging Lisp Programs, elisp, The GNU Emacs Lisp Reference Manual}.
18308 In this chapter, I will walk through a short example of each.
18311 * debug:: How to use the built-in debugger.
18312 * debug-on-entry:: Start debugging when you call a function.
18313 * debug-on-quit:: Start debugging when you quit with @kbd{C-g}.
18314 * edebug:: How to use Edebug, a source level debugger.
18315 * Debugging Exercises::
18319 @section @code{debug}
18322 Suppose you have written a function definition that is intended to
18323 return the sum of the numbers 1 through a given number. (This is the
18324 @code{triangle} function discussed earlier. @xref{Decrementing
18325 Example, , Example with Decrementing Counter}, for a discussion.)
18326 @c xref{Decrementing Loop,, Loop with a Decrementing Counter}, for a discussion.)
18328 However, your function definition has a bug. You have mistyped
18329 @samp{1=} for @samp{1-}. Here is the broken definition:
18331 @findex triangle-bugged
18334 (defun triangle-bugged (number)
18335 "Return sum of numbers 1 through NUMBER inclusive."
18337 (while (> number 0)
18338 (setq total (+ total number))
18339 (setq number (1= number))) ; @r{Error here.}
18344 If you are reading this in Info, you can evaluate this definition in
18345 the normal fashion. You will see @code{triangle-bugged} appear in the
18349 Now evaluate the @code{triangle-bugged} function with an
18353 (triangle-bugged 4)
18357 In a recent GNU Emacs, you will create and enter a @file{*Backtrace*}
18363 ---------- Buffer: *Backtrace* ----------
18364 Debugger entered--Lisp error: (void-function 1=)
18366 (setq number (1= number))
18367 (while (> number 0) (setq total (+ total number))
18368 (setq number (1= number)))
18369 (let ((total 0)) (while (> number 0) (setq total ...)
18370 (setq number ...)) total)
18374 eval((triangle-bugged 4))
18375 eval-last-sexp-1(nil)
18376 eval-last-sexp(nil)
18377 call-interactively(eval-last-sexp)
18378 ---------- Buffer: *Backtrace* ----------
18383 (I have reformatted this example slightly; the debugger does not fold
18384 long lines. As usual, you can quit the debugger by typing @kbd{q} in
18385 the @file{*Backtrace*} buffer.)
18387 In practice, for a bug as simple as this, the `Lisp error' line will
18388 tell you what you need to know to correct the definition. The
18389 function @code{1=} is `void'.
18393 In GNU Emacs 20 and before, you will see:
18396 Symbol's function definition is void:@: 1=
18400 which has the same meaning as the @file{*Backtrace*} buffer line in
18404 However, suppose you are not quite certain what is going on?
18405 You can read the complete backtrace.
18407 In this case, you need to run a recent GNU Emacs, which automatically
18408 starts the debugger that puts you in the @file{*Backtrace*} buffer; or
18409 else, you need to start the debugger manually as described below.
18411 Read the @file{*Backtrace*} buffer from the bottom up; it tells you
18412 what Emacs did that led to the error. Emacs made an interactive call
18413 to @kbd{C-x C-e} (@code{eval-last-sexp}), which led to the evaluation
18414 of the @code{triangle-bugged} expression. Each line above tells you
18415 what the Lisp interpreter evaluated next.
18418 The third line from the top of the buffer is
18421 (setq number (1= number))
18425 Emacs tried to evaluate this expression; in order to do so, it tried
18426 to evaluate the inner expression shown on the second line from the
18435 This is where the error occurred; as the top line says:
18438 Debugger entered--Lisp error: (void-function 1=)
18442 You can correct the mistake, re-evaluate the function definition, and
18443 then run your test again.
18445 @node debug-on-entry
18446 @section @code{debug-on-entry}
18447 @findex debug-on-entry
18449 A recent GNU Emacs starts the debugger automatically when your
18450 function has an error.
18453 GNU Emacs version 20 and before did not; it simply
18454 presented you with an error message. You had to start the debugger
18458 Incidentally, you can start the debugger manually for all versions of
18459 Emacs; the advantage is that the debugger runs even if you do not have
18460 a bug in your code. Sometimes your code will be free of bugs!
18462 You can enter the debugger when you call the function by calling
18463 @code{debug-on-entry}.
18470 M-x debug-on-entry RET triangle-bugged RET
18475 Now, evaluate the following:
18478 (triangle-bugged 5)
18482 All versions of Emacs will create a @file{*Backtrace*} buffer and tell
18483 you that it is beginning to evaluate the @code{triangle-bugged}
18488 ---------- Buffer: *Backtrace* ----------
18489 Debugger entered--entering a function:
18490 * triangle-bugged(5)
18491 eval((triangle-bugged 5))
18494 eval-last-sexp-1(nil)
18495 eval-last-sexp(nil)
18496 call-interactively(eval-last-sexp)
18497 ---------- Buffer: *Backtrace* ----------
18501 In the @file{*Backtrace*} buffer, type @kbd{d}. Emacs will evaluate
18502 the first expression in @code{triangle-bugged}; the buffer will look
18507 ---------- Buffer: *Backtrace* ----------
18508 Debugger entered--beginning evaluation of function call form:
18509 * (let ((total 0)) (while (> number 0) (setq total ...)
18510 (setq number ...)) total)
18511 * triangle-bugged(5)
18512 eval((triangle-bugged 5))
18515 eval-last-sexp-1(nil)
18516 eval-last-sexp(nil)
18517 call-interactively(eval-last-sexp)
18518 ---------- Buffer: *Backtrace* ----------
18523 Now, type @kbd{d} again, eight times, slowly. Each time you type
18524 @kbd{d}, Emacs will evaluate another expression in the function
18528 Eventually, the buffer will look like this:
18532 ---------- Buffer: *Backtrace* ----------
18533 Debugger entered--beginning evaluation of function call form:
18534 * (setq number (1= number))
18535 * (while (> number 0) (setq total (+ total number))
18536 (setq number (1= number)))
18539 * (let ((total 0)) (while (> number 0) (setq total ...)
18540 (setq number ...)) total)
18541 * triangle-bugged(5)
18542 eval((triangle-bugged 5))
18545 eval-last-sexp-1(nil)
18546 eval-last-sexp(nil)
18547 call-interactively(eval-last-sexp)
18548 ---------- Buffer: *Backtrace* ----------
18554 Finally, after you type @kbd{d} two more times, Emacs will reach the
18555 error, and the top two lines of the @file{*Backtrace*} buffer will look
18560 ---------- Buffer: *Backtrace* ----------
18561 Debugger entered--Lisp error: (void-function 1=)
18564 ---------- Buffer: *Backtrace* ----------
18568 By typing @kbd{d}, you were able to step through the function.
18570 You can quit a @file{*Backtrace*} buffer by typing @kbd{q} in it; this
18571 quits the trace, but does not cancel @code{debug-on-entry}.
18573 @findex cancel-debug-on-entry
18574 To cancel the effect of @code{debug-on-entry}, call
18575 @code{cancel-debug-on-entry} and the name of the function, like this:
18578 M-x cancel-debug-on-entry RET triangle-bugged RET
18582 (If you are reading this in Info, cancel @code{debug-on-entry} now.)
18584 @node debug-on-quit
18585 @section @code{debug-on-quit} and @code{(debug)}
18587 In addition to setting @code{debug-on-error} or calling @code{debug-on-entry},
18588 there are two other ways to start @code{debug}.
18590 @findex debug-on-quit
18591 You can start @code{debug} whenever you type @kbd{C-g}
18592 (@code{keyboard-quit}) by setting the variable @code{debug-on-quit} to
18593 @code{t}. This is useful for debugging infinite loops.
18596 @cindex @code{(debug)} in code
18597 Or, you can insert a line that says @code{(debug)} into your code
18598 where you want the debugger to start, like this:
18602 (defun triangle-bugged (number)
18603 "Return sum of numbers 1 through NUMBER inclusive."
18605 (while (> number 0)
18606 (setq total (+ total number))
18607 (debug) ; @r{Start debugger.}
18608 (setq number (1= number))) ; @r{Error here.}
18613 The @code{debug} function is described in detail in @ref{Debugger, ,
18614 The Lisp Debugger, elisp, The GNU Emacs Lisp Reference Manual}.
18617 @section The @code{edebug} Source Level Debugger
18618 @cindex Source level debugger
18621 Edebug is a source level debugger. Edebug normally displays the
18622 source of the code you are debugging, with an arrow at the left that
18623 shows which line you are currently executing.
18625 You can walk through the execution of a function, line by line, or run
18626 quickly until reaching a @dfn{breakpoint} where execution stops.
18628 Edebug is described in @ref{Edebug, , , elisp, The GNU Emacs
18629 Lisp Reference Manual}.
18632 Here is a bugged function definition for @code{triangle-recursively}.
18633 @xref{Recursive triangle function, , Recursion in place of a counter},
18634 for a review of it.
18638 (defun triangle-recursively-bugged (number)
18639 "Return sum of numbers 1 through NUMBER inclusive.
18644 (triangle-recursively-bugged
18645 (1= number))))) ; @r{Error here.}
18650 Normally, you would install this definition by positioning your cursor
18651 after the function's closing parenthesis and typing @kbd{C-x C-e}
18652 (@code{eval-last-sexp}) or else by positioning your cursor within the
18653 definition and typing @kbd{C-M-x} (@code{eval-defun}). (By default,
18654 the @code{eval-defun} command works only in Emacs Lisp mode or in Lisp
18658 However, to prepare this function definition for Edebug, you must
18659 first @dfn{instrument} the code using a different command. You can do
18660 this by positioning your cursor within or just after the definition
18664 M-x edebug-defun RET
18668 This will cause Emacs to load Edebug automatically if it is not
18669 already loaded, and properly instrument the function.
18671 After instrumenting the function, place your cursor after the
18672 following expression and type @kbd{C-x C-e} (@code{eval-last-sexp}):
18675 (triangle-recursively-bugged 3)
18679 You will be jumped back to the source for
18680 @code{triangle-recursively-bugged} and the cursor positioned at the
18681 beginning of the @code{if} line of the function. Also, you will see
18682 an arrowhead at the left hand side of that line. The arrowhead marks
18683 the line where the function is executing. (In the following examples,
18684 we show the arrowhead with @samp{=>}; in a windowing system, you may
18685 see the arrowhead as a solid triangle in the window `fringe'.)
18688 =>@point{}(if (= number 1)
18693 In the example, the location of point is displayed with a star,
18694 @samp{@point{}} (in Info, it is displayed as @samp{-!-}).
18697 In the example, the location of point is displayed as @samp{@point{}}
18698 (in a printed book, it is displayed with a five pointed star).
18701 If you now press @key{SPC}, point will move to the next expression to
18702 be executed; the line will look like this:
18705 =>(if @point{}(= number 1)
18709 As you continue to press @key{SPC}, point will move from expression to
18710 expression. At the same time, whenever an expression returns a value,
18711 that value will be displayed in the echo area. For example, after you
18712 move point past @code{number}, you will see the following:
18715 Result: 3 (#o3, #x3, ?\C-c)
18719 This means the value of @code{number} is 3, which is octal three,
18720 hexadecimal three, and @sc{ascii} `control-c' (the third letter of the
18721 alphabet, in case you need to know this information).
18723 You can continue moving through the code until you reach the line with
18724 the error. Before evaluation, that line looks like this:
18727 => @point{}(1= number))))) ; @r{Error here.}
18732 When you press @key{SPC} once again, you will produce an error message
18736 Symbol's function definition is void:@: 1=
18742 Press @kbd{q} to quit Edebug.
18744 To remove instrumentation from a function definition, simply
18745 re-evaluate it with a command that does not instrument it.
18746 For example, you could place your cursor after the definition's
18747 closing parenthesis and type @kbd{C-x C-e}.
18749 Edebug does a great deal more than walk with you through a function.
18750 You can set it so it races through on its own, stopping only at an
18751 error or at specified stopping points; you can cause it to display the
18752 changing values of various expressions; you can find out how many
18753 times a function is called, and more.
18755 Edebug is described in @ref{Edebug, , , elisp, The GNU Emacs
18756 Lisp Reference Manual}.
18759 @node Debugging Exercises
18760 @section Debugging Exercises
18764 Install the @code{@value{COUNT-WORDS}} function and then cause it to
18765 enter the built-in debugger when you call it. Run the command on a
18766 region containing two words. You will need to press @kbd{d} a
18767 remarkable number of times. On your system, is a `hook' called after
18768 the command finishes? (For information on hooks, see @ref{Command
18769 Overview, , Command Loop Overview, elisp, The GNU Emacs Lisp Reference
18773 Copy @code{@value{COUNT-WORDS}} into the @file{*scratch*} buffer,
18774 instrument the function for Edebug, and walk through its execution.
18775 The function does not need to have a bug, although you can introduce
18776 one if you wish. If the function lacks a bug, the walk-through
18777 completes without problems.
18780 While running Edebug, type @kbd{?} to see a list of all the Edebug commands.
18781 (The @code{global-edebug-prefix} is usually @kbd{C-x X}, i.e.,
18782 @kbd{@key{CTRL}-x} followed by an upper case @kbd{X}; use this prefix
18783 for commands made outside of the Edebug debugging buffer.)
18786 In the Edebug debugging buffer, use the @kbd{p}
18787 (@code{edebug-bounce-point}) command to see where in the region the
18788 @code{@value{COUNT-WORDS}} is working.
18791 Move point to some spot further down the function and then type the
18792 @kbd{h} (@code{edebug-goto-here}) command to jump to that location.
18795 Use the @kbd{t} (@code{edebug-trace-mode}) command to cause Edebug to
18796 walk through the function on its own; use an upper case @kbd{T} for
18797 @code{edebug-Trace-fast-mode}.
18800 Set a breakpoint, then run Edebug in Trace mode until it reaches the
18805 @chapter Conclusion
18807 We have now reached the end of this Introduction. You have now
18808 learned enough about programming in Emacs Lisp to set values, to write
18809 simple @file{.emacs} files for yourself and your friends, and write
18810 simple customizations and extensions to Emacs.
18812 This is a place to stop. Or, if you wish, you can now go onward, and
18815 You have learned some of the basic nuts and bolts of programming. But
18816 only some. There are a great many more brackets and hinges that are
18817 easy to use that we have not touched.
18819 A path you can follow right now lies among the sources to GNU Emacs
18822 @cite{The GNU Emacs Lisp Reference Manual}.
18825 @ref{Top, , The GNU Emacs Lisp Reference Manual, elisp, The GNU
18826 Emacs Lisp Reference Manual}.
18829 The Emacs Lisp sources are an adventure. When you read the sources and
18830 come across a function or expression that is unfamiliar, you need to
18831 figure out or find out what it does.
18833 Go to the Reference Manual. It is a thorough, complete, and fairly
18834 easy-to-read description of Emacs Lisp. It is written not only for
18835 experts, but for people who know what you know. (The @cite{Reference
18836 Manual} comes with the standard GNU Emacs distribution. Like this
18837 introduction, it comes as a Texinfo source file, so you can read it
18838 on-line and as a typeset, printed book.)
18840 Go to the other on-line help that is part of GNU Emacs: the on-line
18841 documentation for all functions and variables, and @code{find-tag},
18842 the program that takes you to sources.
18844 Here is an example of how I explore the sources. Because of its name,
18845 @file{simple.el} is the file I looked at first, a long time ago. As
18846 it happens some of the functions in @file{simple.el} are complicated,
18847 or at least look complicated at first sight. The @code{open-line}
18848 function, for example, looks complicated.
18850 You may want to walk through this function slowly, as we did with the
18851 @code{forward-sentence} function. (@xref{forward-sentence, The
18852 @code{forward-sentence} function}.) Or you may want to skip that
18853 function and look at another, such as @code{split-line}. You don't
18854 need to read all the functions. According to
18855 @code{count-words-in-defun}, the @code{split-line} function contains
18856 102 words and symbols.
18858 Even though it is short, @code{split-line} contains expressions
18859 we have not studied: @code{skip-chars-forward}, @code{indent-to},
18860 @code{current-column} and @code{insert-and-inherit}.
18862 Consider the @code{skip-chars-forward} function. (It is part of the
18863 function definition for @code{back-to-indentation}, which is shown in
18864 @ref{Review, , Review}.)
18866 In GNU Emacs, you can find out more about @code{skip-chars-forward} by
18867 typing @kbd{C-h f} (@code{describe-function}) and the name of the
18868 function. This gives you the function documentation.
18870 You may be able to guess what is done by a well named function such as
18871 @code{indent-to}; or you can look it up, too. Incidentally, the
18872 @code{describe-function} function itself is in @file{help.el}; it is
18873 one of those long, but decipherable functions. You can look up
18874 @code{describe-function} using the @kbd{C-h f} command!
18876 In this instance, since the code is Lisp, the @file{*Help*} buffer
18877 contains the name of the library containing the function's source.
18878 You can put point over the name of the library and press the RET key,
18879 which in this situation is bound to @code{help-follow}, and be taken
18880 directly to the source, in the same way as @kbd{M-.}
18883 The definition for @code{describe-function} illustrates how to
18884 customize the @code{interactive} expression without using the standard
18885 character codes; and it shows how to create a temporary buffer.
18887 (The @code{indent-to} function is written in C rather than Emacs Lisp;
18888 it is a `built-in' function. @code{help-follow} takes you to its
18889 source as does @code{find-tag}, when properly set up.)
18891 You can look at a function's source using @code{find-tag}, which is
18892 bound to @kbd{M-.} Finally, you can find out what the Reference
18893 Manual has to say by visiting the manual in Info, and typing @kbd{i}
18894 (@code{Info-index}) and the name of the function, or by looking up the
18895 function in the index to a printed copy of the manual.
18897 Similarly, you can find out what is meant by
18898 @code{insert-and-inherit}.
18900 Other interesting source files include @file{paragraphs.el},
18901 @file{loaddefs.el}, and @file{loadup.el}. The @file{paragraphs.el}
18902 file includes short, easily understood functions as well as longer
18903 ones. The @file{loaddefs.el} file contains the many standard
18904 autoloads and many keymaps. I have never looked at it all; only at
18905 parts. @file{loadup.el} is the file that loads the standard parts of
18906 Emacs; it tells you a great deal about how Emacs is built.
18907 (@xref{Building Emacs, , Building Emacs, elisp, The GNU Emacs Lisp
18908 Reference Manual}, for more about building.)
18910 As I said, you have learned some nuts and bolts; however, and very
18911 importantly, we have hardly touched major aspects of programming; I
18912 have said nothing about how to sort information, except to use the
18913 predefined @code{sort} function; I have said nothing about how to store
18914 information, except to use variables and lists; I have said nothing
18915 about how to write programs that write programs. These are topics for
18916 another, and different kind of book, a different kind of learning.
18918 What you have done is learn enough for much practical work with GNU
18919 Emacs. What you have done is get started. This is the end of a
18922 @c ================ Appendix ================
18925 @appendix The @code{the-the} Function
18927 @cindex Duplicated words function
18928 @cindex Words, duplicated
18930 Sometimes when you you write text, you duplicate words---as with ``you
18931 you'' near the beginning of this sentence. I find that most
18932 frequently, I duplicate ``the''; hence, I call the function for
18933 detecting duplicated words, @code{the-the}.
18936 As a first step, you could use the following regular expression to
18937 search for duplicates:
18940 \\(\\w+[ \t\n]+\\)\\1
18944 This regexp matches one or more word-constituent characters followed
18945 by one or more spaces, tabs, or newlines. However, it does not detect
18946 duplicated words on different lines, since the ending of the first
18947 word, the end of the line, is different from the ending of the second
18948 word, a space. (For more information about regular expressions, see
18949 @ref{Regexp Search, , Regular Expression Searches}, as well as
18950 @ref{Regexps, , Syntax of Regular Expressions, emacs, The GNU Emacs
18951 Manual}, and @ref{Regular Expressions, , Regular Expressions, elisp,
18952 The GNU Emacs Lisp Reference Manual}.)
18954 You might try searching just for duplicated word-constituent
18955 characters but that does not work since the pattern detects doubles
18956 such as the two occurrences of `th' in `with the'.
18958 Another possible regexp searches for word-constituent characters
18959 followed by non-word-constituent characters, reduplicated. Here,
18960 @w{@samp{\\w+}} matches one or more word-constituent characters and
18961 @w{@samp{\\W*}} matches zero or more non-word-constituent characters.
18964 \\(\\(\\w+\\)\\W*\\)\\1
18970 Here is the pattern that I use. It is not perfect, but good enough.
18971 @w{@samp{\\b}} matches the empty string, provided it is at the beginning
18972 or end of a word; @w{@samp{[^@@ \n\t]+}} matches one or more occurrences of
18973 any characters that are @emph{not} an @@-sign, space, newline, or tab.
18976 \\b\\([^@@ \n\t]+\\)[ \n\t]+\\1\\b
18979 One can write more complicated expressions, but I found that this
18980 expression is good enough, so I use it.
18982 Here is the @code{the-the} function, as I include it in my
18983 @file{.emacs} file, along with a handy global key binding:
18988 "Search forward for for a duplicated word."
18990 (message "Searching for for duplicated words ...")
18994 ;; This regexp is not perfect
18995 ;; but is fairly good over all:
18996 (if (re-search-forward
18997 "\\b\\([^@@ \n\t]+\\)[ \n\t]+\\1\\b" nil 'move)
18998 (message "Found duplicated word.")
18999 (message "End of buffer")))
19003 ;; Bind `the-the' to C-c \
19004 (global-set-key "\C-c\\" 'the-the)
19013 one two two three four five
19018 You can substitute the other regular expressions shown above in the
19019 function definition and try each of them on this list.
19022 @appendix Handling the Kill Ring
19023 @cindex Kill ring handling
19024 @cindex Handling the kill ring
19025 @cindex Ring, making a list like a
19027 The kill ring is a list that is transformed into a ring by the
19028 workings of the @code{current-kill} function. The @code{yank} and
19029 @code{yank-pop} commands use the @code{current-kill} function.
19031 This appendix describes the @code{current-kill} function as well as
19032 both the @code{yank} and the @code{yank-pop} commands, but first,
19033 consider the workings of the kill ring.
19036 * What the Kill Ring Does::
19038 * yank:: Paste a copy of a clipped element.
19039 * yank-pop:: Insert element pointed to.
19044 @node What the Kill Ring Does
19045 @unnumberedsec What the Kill Ring Does
19049 The kill ring has a default maximum length of sixty items; this number
19050 is too large for an explanation. Instead, set it to four. Please
19051 evaluate the following:
19055 (setq old-kill-ring-max kill-ring-max)
19056 (setq kill-ring-max 4)
19061 Then, please copy each line of the following indented example into the
19062 kill ring. You may kill each line with @kbd{C-k} or mark it and copy
19066 (In a read-only buffer, such as the @file{*info*} buffer, the kill
19067 command, @kbd{C-k} (@code{kill-line}), will not remove the text,
19068 merely copy it to the kill ring. However, your machine may beep at
19069 you. Alternatively, for silence, you may copy the region of each line
19070 with the @kbd{M-w} (@code{kill-ring-save}) command. You must mark
19071 each line for this command to succeed, but it does not matter at which
19072 end you put point or mark.)
19076 Please invoke the calls in order, so that five elements attempt to
19077 fill the kill ring:
19082 second piece of text
19084 fourth line of text
19091 Then find the value of @code{kill-ring} by evaluating
19103 ("fifth bit of text" "fourth line of text"
19104 "third line" "second piece of text")
19109 The first element, @samp{first some text}, was dropped.
19112 To return to the old value for the length of the kill ring, evaluate:
19115 (setq kill-ring-max old-kill-ring-max)
19119 @appendixsec The @code{current-kill} Function
19120 @findex current-kill
19122 The @code{current-kill} function changes the element in the kill ring
19123 to which @code{kill-ring-yank-pointer} points. (Also, the
19124 @code{kill-new} function sets @code{kill-ring-yank-pointer} to point
19125 to the latest element of the kill ring. The @code{kill-new}
19126 function is used directly or indirectly by @code{kill-append},
19127 @code{copy-region-as-kill}, @code{kill-ring-save}, @code{kill-line},
19128 and @code{kill-region}.)
19131 * Code for current-kill::
19132 * Understanding current-kill::
19136 @node Code for current-kill
19137 @unnumberedsubsec The code for @code{current-kill}
19142 The @code{current-kill} function is used by @code{yank} and by
19143 @code{yank-pop}. Here is the code for @code{current-kill}:
19147 (defun current-kill (n &optional do-not-move)
19148 "Rotate the yanking point by N places, and then return that kill.
19149 If N is zero, `interprogram-paste-function' is set, and calling it
19150 returns a string, then that string is added to the front of the
19151 kill ring and returned as the latest kill.
19154 If optional arg DO-NOT-MOVE is non-nil, then don't actually move the
19155 yanking point; just return the Nth kill forward."
19156 (let ((interprogram-paste (and (= n 0)
19157 interprogram-paste-function
19158 (funcall interprogram-paste-function))))
19161 (if interprogram-paste
19163 ;; Disable the interprogram cut function when we add the new
19164 ;; text to the kill ring, so Emacs doesn't try to own the
19165 ;; selection, with identical text.
19166 (let ((interprogram-cut-function nil))
19167 (kill-new interprogram-paste))
19168 interprogram-paste)
19171 (or kill-ring (error "Kill ring is empty"))
19172 (let ((ARGth-kill-element
19173 (nthcdr (mod (- n (length kill-ring-yank-pointer))
19174 (length kill-ring))
19177 (setq kill-ring-yank-pointer ARGth-kill-element))
19178 (car ARGth-kill-element)))))
19182 Remember also that the @code{kill-new} function sets
19183 @code{kill-ring-yank-pointer} to the latest element of the kill
19184 ring, which means that all the functions that call it set the value
19185 indirectly: @code{kill-append}, @code{copy-region-as-kill},
19186 @code{kill-ring-save}, @code{kill-line}, and @code{kill-region}.
19189 Here is the line in @code{kill-new}, which is explained in
19190 @ref{kill-new function, , The @code{kill-new} function}.
19193 (setq kill-ring-yank-pointer kill-ring)
19197 @node Understanding current-kill
19198 @unnumberedsubsec @code{current-kill} in Outline
19201 The @code{current-kill} function looks complex, but as usual, it can
19202 be understood by taking it apart piece by piece. First look at it in
19207 (defun current-kill (n &optional do-not-move)
19208 "Rotate the yanking point by N places, and then return that kill."
19214 This function takes two arguments, one of which is optional. It has a
19215 documentation string. It is @emph{not} interactive.
19218 * Body of current-kill::
19219 * Digression concerning error:: How to mislead humans, but not computers.
19220 * Determining the Element::
19224 @node Body of current-kill
19225 @unnumberedsubsubsec The Body of @code{current-kill}
19228 The body of the function definition is a @code{let} expression, which
19229 itself has a body as well as a @var{varlist}.
19231 The @code{let} expression declares a variable that will be only usable
19232 within the bounds of this function. This variable is called
19233 @code{interprogram-paste} and is for copying to another program. It
19234 is not for copying within this instance of GNU Emacs. Most window
19235 systems provide a facility for interprogram pasting. Sadly, that
19236 facility usually provides only for the last element. Most windowing
19237 systems have not adopted a ring of many possibilities, even though
19238 Emacs has provided it for decades.
19240 The @code{if} expression has two parts, one if there exists
19241 @code{interprogram-paste} and one if not.
19244 Let us consider the `if not' or else-part of the @code{current-kill}
19245 function. (The then-part uses the @code{kill-new} function, which
19246 we have already described. @xref{kill-new function, , The
19247 @code{kill-new} function}.)
19251 (or kill-ring (error "Kill ring is empty"))
19252 (let ((ARGth-kill-element
19253 (nthcdr (mod (- n (length kill-ring-yank-pointer))
19254 (length kill-ring))
19257 (setq kill-ring-yank-pointer ARGth-kill-element))
19258 (car ARGth-kill-element))
19263 The code first checks whether the kill ring has content; otherwise it
19267 Note that the @code{or} expression is very similar to testing length
19274 (if (zerop (length kill-ring)) ; @r{if-part}
19275 (error "Kill ring is empty")) ; @r{then-part}
19281 If there is not anything in the kill ring, its length must be zero and
19282 an error message sent to the user: @samp{Kill ring is empty}. The
19283 @code{current-kill} function uses an @code{or} expression which is
19284 simpler. But an @code{if} expression reminds us what goes on.
19286 This @code{if} expression uses the function @code{zerop} which returns
19287 true if the value it is testing is zero. When @code{zerop} tests
19288 true, the then-part of the @code{if} is evaluated. The then-part is a
19289 list starting with the function @code{error}, which is a function that
19290 is similar to the @code{message} function
19291 (@pxref{message, , The @code{message} Function}) in that
19292 it prints a one-line message in the echo area. However, in addition
19293 to printing a message, @code{error} also stops evaluation of the
19294 function within which it is embedded. This means that the rest of the
19295 function will not be evaluated if the length of the kill ring is zero.
19297 Then the @code{current-kill} function selects the element to return.
19298 The selection depends on the number of places that @code{current-kill}
19299 rotates and on where @code{kill-ring-yank-pointer} points.
19301 Next, either the optional @code{do-not-move} argument is true or the
19302 current value of @code{kill-ring-yank-pointer} is set to point to the
19303 list. Finally, another expression returns the first element of the
19304 list even if the @code{do-not-move} argument is true.
19307 @node Digression concerning error
19308 @unnumberedsubsubsec Digression about the word `error'
19311 In my opinion, it is slightly misleading, at least to humans, to use
19312 the term `error' as the name of the @code{error} function. A better
19313 term would be `cancel'. Strictly speaking, of course, you cannot
19314 point to, much less rotate a pointer to a list that has no length, so
19315 from the point of view of the computer, the word `error' is correct.
19316 But a human expects to attempt this sort of thing, if only to find out
19317 whether the kill ring is full or empty. This is an act of
19320 From the human point of view, the act of exploration and discovery is
19321 not necessarily an error, and therefore should not be labeled as one,
19322 even in the bowels of a computer. As it is, the code in Emacs implies
19323 that a human who is acting virtuously, by exploring his or her
19324 environment, is making an error. This is bad. Even though the computer
19325 takes the same steps as it does when there is an `error', a term such as
19326 `cancel' would have a clearer connotation.
19329 @node Determining the Element
19330 @unnumberedsubsubsec Determining the Element
19333 Among other actions, the else-part of the @code{if} expression sets
19334 the value of @code{kill-ring-yank-pointer} to
19335 @code{ARGth-kill-element} when the kill ring has something in it and
19336 the value of @code{do-not-move} is @code{nil}.
19339 The code looks like this:
19343 (nthcdr (mod (- n (length kill-ring-yank-pointer))
19344 (length kill-ring))
19349 This needs some examination. Unless it is not supposed to move the
19350 pointer, the @code{current-kill} function changes where
19351 @code{kill-ring-yank-pointer} points.
19353 @w{@code{(setq kill-ring-yank-pointer ARGth-kill-element))}}
19354 expression does. Also, clearly, @code{ARGth-kill-element} is being
19355 set to be equal to some @sc{cdr} of the kill ring, using the
19356 @code{nthcdr} function that is described in an earlier section.
19357 (@xref{copy-region-as-kill}.) How does it do this?
19359 As we have seen before (@pxref{nthcdr}), the @code{nthcdr} function
19360 works by repeatedly taking the @sc{cdr} of a list---it takes the
19361 @sc{cdr} of the @sc{cdr} of the @sc{cdr} @dots{}
19364 The two following expressions produce the same result:
19368 (setq kill-ring-yank-pointer (cdr kill-ring))
19370 (setq kill-ring-yank-pointer (nthcdr 1 kill-ring))
19374 However, the @code{nthcdr} expression is more complicated. It uses
19375 the @code{mod} function to determine which @sc{cdr} to select.
19377 (You will remember to look at inner functions first; indeed, we will
19378 have to go inside the @code{mod}.)
19380 The @code{mod} function returns the value of its first argument modulo
19381 the second; that is to say, it returns the remainder after dividing
19382 the first argument by the second. The value returned has the same
19383 sign as the second argument.
19391 @result{} 0 ;; @r{because there is no remainder}
19398 In this case, the first argument is often smaller than the second.
19410 We can guess what the @code{-} function does. It is like @code{+} but
19411 subtracts instead of adds; the @code{-} function subtracts its second
19412 argument from its first. Also, we already know what the @code{length}
19413 function does (@pxref{length}). It returns the length of a list.
19415 And @code{n} is the name of the required argument to the
19416 @code{current-kill} function.
19419 So when the first argument to @code{nthcdr} is zero, the @code{nthcdr}
19420 expression returns the whole list, as you can see by evaluating the
19425 ;; kill-ring-yank-pointer @r{and} kill-ring @r{have a length of four}
19426 ;; @r{and} (mod (- 0 4) 4) @result{} 0
19427 (nthcdr (mod (- 0 4) 4)
19428 '("fourth line of text"
19430 "second piece of text"
19431 "first some text"))
19436 When the first argument to the @code{current-kill} function is one,
19437 the @code{nthcdr} expression returns the list without its first
19442 (nthcdr (mod (- 1 4) 4)
19443 '("fourth line of text"
19445 "second piece of text"
19446 "first some text"))
19450 @cindex @samp{global variable} defined
19451 @cindex @samp{variable, global}, defined
19452 Incidentally, both @code{kill-ring} and @code{kill-ring-yank-pointer}
19453 are @dfn{global variables}. That means that any expression in Emacs
19454 Lisp can access them. They are not like the local variables set by
19455 @code{let} or like the symbols in an argument list.
19456 Local variables can only be accessed
19457 within the @code{let} that defines them or the function that specifies
19458 them in an argument list (and within expressions called by them).
19461 @c texi2dvi fails when the name of the section is within ifnottex ...
19462 (@xref{Prevent confusion, , @code{let} Prevents Confusion}, and
19463 @ref{defun, , The @code{defun} Special Form}.)
19467 @appendixsec @code{yank}
19470 After learning about @code{current-kill}, the code for the
19471 @code{yank} function is almost easy.
19473 The @code{yank} function does not use the
19474 @code{kill-ring-yank-pointer} variable directly. It calls
19475 @code{insert-for-yank} which calls @code{current-kill} which sets the
19476 @code{kill-ring-yank-pointer} variable.
19479 The code looks like this:
19484 (defun yank (&optional arg)
19485 "Reinsert (\"paste\") the last stretch of killed text.
19486 More precisely, reinsert the stretch of killed text most recently
19487 killed OR yanked. Put point at end, and set mark at beginning.
19488 With just \\[universal-argument] as argument, same but put point at
19489 beginning (and mark at end). With argument N, reinsert the Nth most
19490 recently killed stretch of killed text.
19492 When this command inserts killed text into the buffer, it honors
19493 `yank-excluded-properties' and `yank-handler' as described in the
19494 doc string for `insert-for-yank-1', which see.
19496 See also the command \\[yank-pop]."
19500 (setq yank-window-start (window-start))
19501 ;; If we don't get all the way thru, make last-command indicate that
19502 ;; for the following command.
19503 (setq this-command t)
19504 (push-mark (point))
19507 (insert-for-yank (current-kill (cond
19512 ;; This is like exchange-point-and-mark,
19513 ;; but doesn't activate the mark.
19514 ;; It is cleaner to avoid activation, even though the command
19515 ;; loop would deactivate the mark because we inserted text.
19516 (goto-char (prog1 (mark t)
19517 (set-marker (mark-marker) (point) (current-buffer)))))
19520 ;; If we do get all the way thru, make this-command indicate that.
19521 (if (eq this-command t)
19522 (setq this-command 'yank))
19527 The key expression is @code{insert-for-yank}, which inserts the string
19528 returned by @code{current-kill}, but removes some text properties from
19531 However, before getting to that expression, the function sets the value
19532 of @code{yank-window-start} to the position returned by the
19533 @code{(window-start)} expression, the position at which the display
19534 currently starts. The @code{yank} function also sets
19535 @code{this-command} and pushes the mark.
19537 After it yanks the appropriate element, if the optional argument is a
19538 @sc{cons} rather than a number or nothing, it puts point at beginning
19539 of the yanked text and mark at its end.
19541 (The @code{prog1} function is like @code{progn} but returns the value
19542 of its first argument rather than the value of its last argument. Its
19543 first argument is forced to return the buffer's mark as an integer.
19544 You can see the documentation for these functions by placing point
19545 over them in this buffer and then typing @kbd{C-h f}
19546 (@code{describe-function}) followed by a @kbd{RET}; the default is the
19549 The last part of the function tells what to do when it succeeds.
19552 @appendixsec @code{yank-pop}
19555 After understanding @code{yank} and @code{current-kill}, you know how
19556 to approach the @code{yank-pop} function. Leaving out the
19557 documentation to save space, it looks like this:
19562 (defun yank-pop (&optional arg)
19565 (if (not (eq last-command 'yank))
19566 (error "Previous command was not a yank"))
19569 (setq this-command 'yank)
19570 (unless arg (setq arg 1))
19571 (let ((inhibit-read-only t)
19572 (before (< (point) (mark t))))
19576 (funcall (or yank-undo-function 'delete-region) (point) (mark t))
19577 (funcall (or yank-undo-function 'delete-region) (mark t) (point)))
19578 (setq yank-undo-function nil)
19581 (set-marker (mark-marker) (point) (current-buffer))
19582 (insert-for-yank (current-kill arg))
19583 ;; Set the window start back where it was in the yank command,
19585 (set-window-start (selected-window) yank-window-start t)
19589 ;; This is like exchange-point-and-mark,
19590 ;; but doesn't activate the mark.
19591 ;; It is cleaner to avoid activation, even though the command
19592 ;; loop would deactivate the mark because we inserted text.
19593 (goto-char (prog1 (mark t)
19594 (set-marker (mark-marker)
19596 (current-buffer))))))
19601 The function is interactive with a small @samp{p} so the prefix
19602 argument is processed and passed to the function. The command can
19603 only be used after a previous yank; otherwise an error message is
19604 sent. This check uses the variable @code{last-command} which is set
19605 by @code{yank} and is discussed elsewhere.
19606 (@xref{copy-region-as-kill}.)
19608 The @code{let} clause sets the variable @code{before} to true or false
19609 depending whether point is before or after mark and then the region
19610 between point and mark is deleted. This is the region that was just
19611 inserted by the previous yank and it is this text that will be
19614 @code{funcall} calls its first argument as a function, passing
19615 remaining arguments to it. The first argument is whatever the
19616 @code{or} expression returns. The two remaining arguments are the
19617 positions of point and mark set by the preceding @code{yank} command.
19619 There is more, but that is the hardest part.
19622 @appendixsec The @file{ring.el} File
19623 @cindex @file{ring.el} file
19625 Interestingly, GNU Emacs posses a file called @file{ring.el} that
19626 provides many of the features we just discussed. But functions such
19627 as @code{kill-ring-yank-pointer} do not use this library, possibly
19628 because they were written earlier.
19631 @appendix A Graph with Labeled Axes
19633 Printed axes help you understand a graph. They convey scale. In an
19634 earlier chapter (@pxref{Readying a Graph, , Readying a Graph}), we
19635 wrote the code to print the body of a graph. Here we write the code
19636 for printing and labeling vertical and horizontal axes, along with the
19640 * Labeled Example::
19641 * print-graph Varlist:: @code{let} expression in @code{print-graph}.
19642 * print-Y-axis:: Print a label for the vertical axis.
19643 * print-X-axis:: Print a horizontal label.
19644 * Print Whole Graph:: The function to print a complete graph.
19648 @node Labeled Example
19649 @unnumberedsec Labeled Example Graph
19652 Since insertions fill a buffer to the right and below point, the new
19653 graph printing function should first print the Y or vertical axis,
19654 then the body of the graph, and finally the X or horizontal axis.
19655 This sequence lays out for us the contents of the function:
19665 Print body of graph.
19672 Here is an example of how a finished graph should look:
19685 1 - ****************
19692 In this graph, both the vertical and the horizontal axes are labeled
19693 with numbers. However, in some graphs, the horizontal axis is time
19694 and would be better labeled with months, like this:
19708 Indeed, with a little thought, we can easily come up with a variety of
19709 vertical and horizontal labeling schemes. Our task could become
19710 complicated. But complications breed confusion. Rather than permit
19711 this, it is better choose a simple labeling scheme for our first
19712 effort, and to modify or replace it later.
19715 These considerations suggest the following outline for the
19716 @code{print-graph} function:
19720 (defun print-graph (numbers-list)
19721 "@var{documentation}@dots{}"
19722 (let ((height @dots{}
19726 (print-Y-axis height @dots{} )
19727 (graph-body-print numbers-list)
19728 (print-X-axis @dots{} )))
19732 We can work on each part of the @code{print-graph} function definition
19735 @node print-graph Varlist
19736 @appendixsec The @code{print-graph} Varlist
19737 @cindex @code{print-graph} varlist
19739 In writing the @code{print-graph} function, the first task is to write
19740 the varlist in the @code{let} expression. (We will leave aside for the
19741 moment any thoughts about making the function interactive or about the
19742 contents of its documentation string.)
19744 The varlist should set several values. Clearly, the top of the label
19745 for the vertical axis must be at least the height of the graph, which
19746 means that we must obtain this information here. Note that the
19747 @code{print-graph-body} function also requires this information. There
19748 is no reason to calculate the height of the graph in two different
19749 places, so we should change @code{print-graph-body} from the way we
19750 defined it earlier to take advantage of the calculation.
19752 Similarly, both the function for printing the X axis labels and the
19753 @code{print-graph-body} function need to learn the value of the width of
19754 each symbol. We can perform the calculation here and change the
19755 definition for @code{print-graph-body} from the way we defined it in the
19758 The length of the label for the horizontal axis must be at least as long
19759 as the graph. However, this information is used only in the function
19760 that prints the horizontal axis, so it does not need to be calculated here.
19762 These thoughts lead us directly to the following form for the varlist
19763 in the @code{let} for @code{print-graph}:
19767 (let ((height (apply 'max numbers-list)) ; @r{First version.}
19768 (symbol-width (length graph-blank)))
19773 As we shall see, this expression is not quite right.
19777 @appendixsec The @code{print-Y-axis} Function
19778 @cindex Axis, print vertical
19779 @cindex Y axis printing
19780 @cindex Vertical axis printing
19781 @cindex Print vertical axis
19783 The job of the @code{print-Y-axis} function is to print a label for
19784 the vertical axis that looks like this:
19802 The function should be passed the height of the graph, and then should
19803 construct and insert the appropriate numbers and marks.
19806 * print-Y-axis in Detail::
19807 * Height of label:: What height for the Y axis?
19808 * Compute a Remainder:: How to compute the remainder of a division.
19809 * Y Axis Element:: Construct a line for the Y axis.
19810 * Y-axis-column:: Generate a list of Y axis labels.
19811 * print-Y-axis Penultimate:: A not quite final version.
19815 @node print-Y-axis in Detail
19816 @unnumberedsubsec The @code{print-Y-axis} Function in Detail
19819 It is easy enough to see in the figure what the Y axis label should
19820 look like; but to say in words, and then to write a function
19821 definition to do the job is another matter. It is not quite true to
19822 say that we want a number and a tic every five lines: there are only
19823 three lines between the @samp{1} and the @samp{5} (lines 2, 3, and 4),
19824 but four lines between the @samp{5} and the @samp{10} (lines 6, 7, 8,
19825 and 9). It is better to say that we want a number and a tic mark on
19826 the base line (number 1) and then that we want a number and a tic on
19827 the fifth line from the bottom and on every line that is a multiple of
19831 @node Height of label
19832 @unnumberedsubsec What height should the label be?
19835 The next issue is what height the label should be? Suppose the maximum
19836 height of tallest column of the graph is seven. Should the highest
19837 label on the Y axis be @samp{5 -}, and should the graph stick up above
19838 the label? Or should the highest label be @samp{7 -}, and mark the peak
19839 of the graph? Or should the highest label be @code{10 -}, which is a
19840 multiple of five, and be higher than the topmost value of the graph?
19842 The latter form is preferred. Most graphs are drawn within rectangles
19843 whose sides are an integral number of steps long---5, 10, 15, and so
19844 on for a step distance of five. But as soon as we decide to use a
19845 step height for the vertical axis, we discover that the simple
19846 expression in the varlist for computing the height is wrong. The
19847 expression is @code{(apply 'max numbers-list)}. This returns the
19848 precise height, not the maximum height plus whatever is necessary to
19849 round up to the nearest multiple of five. A more complex expression
19852 As usual in cases like this, a complex problem becomes simpler if it is
19853 divided into several smaller problems.
19855 First, consider the case when the highest value of the graph is an
19856 integral multiple of five---when it is 5, 10, 15, or some higher
19857 multiple of five. We can use this value as the Y axis height.
19859 A fairly simply way to determine whether a number is a multiple of
19860 five is to divide it by five and see if the division results in a
19861 remainder. If there is no remainder, the number is a multiple of
19862 five. Thus, seven divided by five has a remainder of two, and seven
19863 is not an integral multiple of five. Put in slightly different
19864 language, more reminiscent of the classroom, five goes into seven
19865 once, with a remainder of two. However, five goes into ten twice,
19866 with no remainder: ten is an integral multiple of five.
19868 @node Compute a Remainder
19869 @appendixsubsec Side Trip: Compute a Remainder
19871 @findex % @r{(remainder function)}
19872 @cindex Remainder function, @code{%}
19873 In Lisp, the function for computing a remainder is @code{%}. The
19874 function returns the remainder of its first argument divided by its
19875 second argument. As it happens, @code{%} is a function in Emacs Lisp
19876 that you cannot discover using @code{apropos}: you find nothing if you
19877 type @kbd{M-x apropos @key{RET} remainder @key{RET}}. The only way to
19878 learn of the existence of @code{%} is to read about it in a book such
19879 as this or in the Emacs Lisp sources.
19881 You can try the @code{%} function by evaluating the following two
19893 The first expression returns 2 and the second expression returns 0.
19895 To test whether the returned value is zero or some other number, we
19896 can use the @code{zerop} function. This function returns @code{t} if
19897 its argument, which must be a number, is zero.
19909 Thus, the following expression will return @code{t} if the height
19910 of the graph is evenly divisible by five:
19913 (zerop (% height 5))
19917 (The value of @code{height}, of course, can be found from @code{(apply
19918 'max numbers-list)}.)
19920 On the other hand, if the value of @code{height} is not a multiple of
19921 five, we want to reset the value to the next higher multiple of five.
19922 This is straightforward arithmetic using functions with which we are
19923 already familiar. First, we divide the value of @code{height} by five
19924 to determine how many times five goes into the number. Thus, five
19925 goes into twelve twice. If we add one to this quotient and multiply by
19926 five, we will obtain the value of the next multiple of five that is
19927 larger than the height. Five goes into twelve twice. Add one to two,
19928 and multiply by five; the result is fifteen, which is the next multiple
19929 of five that is higher than twelve. The Lisp expression for this is:
19932 (* (1+ (/ height 5)) 5)
19936 For example, if you evaluate the following, the result is 15:
19939 (* (1+ (/ 12 5)) 5)
19942 All through this discussion, we have been using `five' as the value
19943 for spacing labels on the Y axis; but we may want to use some other
19944 value. For generality, we should replace `five' with a variable to
19945 which we can assign a value. The best name I can think of for this
19946 variable is @code{Y-axis-label-spacing}.
19949 Using this term, and an @code{if} expression, we produce the
19954 (if (zerop (% height Y-axis-label-spacing))
19957 (* (1+ (/ height Y-axis-label-spacing))
19958 Y-axis-label-spacing))
19963 This expression returns the value of @code{height} itself if the height
19964 is an even multiple of the value of the @code{Y-axis-label-spacing} or
19965 else it computes and returns a value of @code{height} that is equal to
19966 the next higher multiple of the value of the @code{Y-axis-label-spacing}.
19968 We can now include this expression in the @code{let} expression of the
19969 @code{print-graph} function (after first setting the value of
19970 @code{Y-axis-label-spacing}):
19971 @vindex Y-axis-label-spacing
19975 (defvar Y-axis-label-spacing 5
19976 "Number of lines from one Y axis label to next.")
19981 (let* ((height (apply 'max numbers-list))
19982 (height-of-top-line
19983 (if (zerop (% height Y-axis-label-spacing))
19988 (* (1+ (/ height Y-axis-label-spacing))
19989 Y-axis-label-spacing)))
19990 (symbol-width (length graph-blank))))
19996 (Note use of the @code{let*} function: the initial value of height is
19997 computed once by the @code{(apply 'max numbers-list)} expression and
19998 then the resulting value of @code{height} is used to compute its
19999 final value. @xref{fwd-para let, , The @code{let*} expression}, for
20000 more about @code{let*}.)
20002 @node Y Axis Element
20003 @appendixsubsec Construct a Y Axis Element
20005 When we print the vertical axis, we want to insert strings such as
20006 @w{@samp{5 -}} and @w{@samp{10 - }} every five lines.
20007 Moreover, we want the numbers and dashes to line up, so shorter
20008 numbers must be padded with leading spaces. If some of the strings
20009 use two digit numbers, the strings with single digit numbers must
20010 include a leading blank space before the number.
20012 @findex number-to-string
20013 To figure out the length of the number, the @code{length} function is
20014 used. But the @code{length} function works only with a string, not with
20015 a number. So the number has to be converted from being a number to
20016 being a string. This is done with the @code{number-to-string} function.
20021 (length (number-to-string 35))
20024 (length (number-to-string 100))
20030 (@code{number-to-string} is also called @code{int-to-string}; you will
20031 see this alternative name in various sources.)
20033 In addition, in each label, each number is followed by a string such
20034 as @w{@samp{ - }}, which we will call the @code{Y-axis-tic} marker.
20035 This variable is defined with @code{defvar}:
20040 (defvar Y-axis-tic " - "
20041 "String that follows number in a Y axis label.")
20045 The length of the Y label is the sum of the length of the Y axis tic
20046 mark and the length of the number of the top of the graph.
20049 (length (concat (number-to-string height) Y-axis-tic)))
20052 This value will be calculated by the @code{print-graph} function in
20053 its varlist as @code{full-Y-label-width} and passed on. (Note that we
20054 did not think to include this in the varlist when we first proposed it.)
20056 To make a complete vertical axis label, a tic mark is concatenated
20057 with a number; and the two together may be preceded by one or more
20058 spaces depending on how long the number is. The label consists of
20059 three parts: the (optional) leading spaces, the number, and the tic
20060 mark. The function is passed the value of the number for the specific
20061 row, and the value of the width of the top line, which is calculated
20062 (just once) by @code{print-graph}.
20066 (defun Y-axis-element (number full-Y-label-width)
20067 "Construct a NUMBERed label element.
20068 A numbered element looks like this ` 5 - ',
20069 and is padded as needed so all line up with
20070 the element for the largest number."
20073 (let* ((leading-spaces
20074 (- full-Y-label-width
20076 (concat (number-to-string number)
20081 (make-string leading-spaces ? )
20082 (number-to-string number)
20087 The @code{Y-axis-element} function concatenates together the leading
20088 spaces, if any; the number, as a string; and the tic mark.
20090 To figure out how many leading spaces the label will need, the
20091 function subtracts the actual length of the label---the length of the
20092 number plus the length of the tic mark---from the desired label width.
20094 @findex make-string
20095 Blank spaces are inserted using the @code{make-string} function. This
20096 function takes two arguments: the first tells it how long the string
20097 will be and the second is a symbol for the character to insert, in a
20098 special format. The format is a question mark followed by a blank
20099 space, like this, @samp{? }. @xref{Character Type, , Character Type,
20100 elisp, The GNU Emacs Lisp Reference Manual}, for a description of the
20101 syntax for characters. (Of course, you might want to replace the
20102 blank space by some other character @dots{} You know what to do.)
20104 The @code{number-to-string} function is used in the concatenation
20105 expression, to convert the number to a string that is concatenated
20106 with the leading spaces and the tic mark.
20108 @node Y-axis-column
20109 @appendixsubsec Create a Y Axis Column
20111 The preceding functions provide all the tools needed to construct a
20112 function that generates a list of numbered and blank strings to insert
20113 as the label for the vertical axis:
20115 @findex Y-axis-column
20118 (defun Y-axis-column (height width-of-label)
20119 "Construct list of Y axis labels and blank strings.
20120 For HEIGHT of line above base and WIDTH-OF-LABEL."
20124 (while (> height 1)
20125 (if (zerop (% height Y-axis-label-spacing))
20126 ;; @r{Insert label.}
20129 (Y-axis-element height width-of-label)
20133 ;; @r{Else, insert blanks.}
20136 (make-string width-of-label ? )
20138 (setq height (1- height)))
20139 ;; @r{Insert base line.}
20141 (cons (Y-axis-element 1 width-of-label) Y-axis))
20142 (nreverse Y-axis)))
20146 In this function, we start with the value of @code{height} and
20147 repetitively subtract one from its value. After each subtraction, we
20148 test to see whether the value is an integral multiple of the
20149 @code{Y-axis-label-spacing}. If it is, we construct a numbered label
20150 using the @code{Y-axis-element} function; if not, we construct a
20151 blank label using the @code{make-string} function. The base line
20152 consists of the number one followed by a tic mark.
20155 @node print-Y-axis Penultimate
20156 @appendixsubsec The Not Quite Final Version of @code{print-Y-axis}
20158 The list constructed by the @code{Y-axis-column} function is passed to
20159 the @code{print-Y-axis} function, which inserts the list as a column.
20161 @findex print-Y-axis
20164 (defun print-Y-axis (height full-Y-label-width)
20165 "Insert Y axis using HEIGHT and FULL-Y-LABEL-WIDTH.
20166 Height must be the maximum height of the graph.
20167 Full width is the width of the highest label element."
20168 ;; Value of height and full-Y-label-width
20169 ;; are passed by `print-graph'.
20172 (let ((start (point)))
20174 (Y-axis-column height full-Y-label-width))
20175 ;; @r{Place point ready for inserting graph.}
20177 ;; @r{Move point forward by value of} full-Y-label-width
20178 (forward-char full-Y-label-width)))
20182 The @code{print-Y-axis} uses the @code{insert-rectangle} function to
20183 insert the Y axis labels created by the @code{Y-axis-column} function.
20184 In addition, it places point at the correct position for printing the body of
20187 You can test @code{print-Y-axis}:
20195 Y-axis-label-spacing
20204 Copy the following expression:
20207 (print-Y-axis 12 5)
20211 Switch to the @file{*scratch*} buffer and place the cursor where you
20212 want the axis labels to start.
20215 Type @kbd{M-:} (@code{eval-expression}).
20218 Yank the @code{graph-body-print} expression into the minibuffer
20219 with @kbd{C-y} (@code{yank)}.
20222 Press @key{RET} to evaluate the expression.
20225 Emacs will print labels vertically, the top one being @w{@samp{10 -@w{
20226 }}}. (The @code{print-graph} function will pass the value of
20227 @code{height-of-top-line}, which in this case will end up as 15,
20228 thereby getting rid of what might appear as a bug.)
20232 @appendixsec The @code{print-X-axis} Function
20233 @cindex Axis, print horizontal
20234 @cindex X axis printing
20235 @cindex Print horizontal axis
20236 @cindex Horizontal axis printing
20238 X axis labels are much like Y axis labels, except that the ticks are on a
20239 line above the numbers. Labels should look like this:
20248 The first tic is under the first column of the graph and is preceded by
20249 several blank spaces. These spaces provide room in rows above for the Y
20250 axis labels. The second, third, fourth, and subsequent ticks are all
20251 spaced equally, according to the value of @code{X-axis-label-spacing}.
20253 The second row of the X axis consists of numbers, preceded by several
20254 blank spaces and also separated according to the value of the variable
20255 @code{X-axis-label-spacing}.
20257 The value of the variable @code{X-axis-label-spacing} should itself be
20258 measured in units of @code{symbol-width}, since you may want to change
20259 the width of the symbols that you are using to print the body of the
20260 graph without changing the ways the graph is labeled.
20263 * Similarities differences:: Much like @code{print-Y-axis}, but not exactly.
20264 * X Axis Tic Marks:: Create tic marks for the horizontal axis.
20268 @node Similarities differences
20269 @unnumberedsubsec Similarities and differences
20272 The @code{print-X-axis} function is constructed in more or less the
20273 same fashion as the @code{print-Y-axis} function except that it has
20274 two lines: the line of tic marks and the numbers. We will write a
20275 separate function to print each line and then combine them within the
20276 @code{print-X-axis} function.
20278 This is a three step process:
20282 Write a function to print the X axis tic marks, @code{print-X-axis-tic-line}.
20285 Write a function to print the X numbers, @code{print-X-axis-numbered-line}.
20288 Write a function to print both lines, the @code{print-X-axis} function,
20289 using @code{print-X-axis-tic-line} and
20290 @code{print-X-axis-numbered-line}.
20293 @node X Axis Tic Marks
20294 @appendixsubsec X Axis Tic Marks
20296 The first function should print the X axis tic marks. We must specify
20297 the tic marks themselves and their spacing:
20301 (defvar X-axis-label-spacing
20302 (if (boundp 'graph-blank)
20303 (* 5 (length graph-blank)) 5)
20304 "Number of units from one X axis label to next.")
20309 (Note that the value of @code{graph-blank} is set by another
20310 @code{defvar}. The @code{boundp} predicate checks whether it has
20311 already been set; @code{boundp} returns @code{nil} if it has not. If
20312 @code{graph-blank} were unbound and we did not use this conditional
20313 construction, in a recent GNU Emacs, we would enter the debugger and
20314 see an error message saying @samp{@w{Debugger entered--Lisp error:}
20315 @w{(void-variable graph-blank)}}.)
20318 Here is the @code{defvar} for @code{X-axis-tic-symbol}:
20322 (defvar X-axis-tic-symbol "|"
20323 "String to insert to point to a column in X axis.")
20328 The goal is to make a line that looks like this:
20334 The first tic is indented so that it is under the first column, which is
20335 indented to provide space for the Y axis labels.
20337 A tic element consists of the blank spaces that stretch from one tic to
20338 the next plus a tic symbol. The number of blanks is determined by the
20339 width of the tic symbol and the @code{X-axis-label-spacing}.
20342 The code looks like this:
20346 ;;; X-axis-tic-element
20350 ;; @r{Make a string of blanks.}
20351 (- (* symbol-width X-axis-label-spacing)
20352 (length X-axis-tic-symbol))
20354 ;; @r{Concatenate blanks with tic symbol.}
20360 Next, we determine how many blanks are needed to indent the first tic
20361 mark to the first column of the graph. This uses the value of
20362 @code{full-Y-label-width} passed it by the @code{print-graph} function.
20365 The code to make @code{X-axis-leading-spaces}
20370 ;; X-axis-leading-spaces
20372 (make-string full-Y-label-width ? )
20377 We also need to determine the length of the horizontal axis, which is
20378 the length of the numbers list, and the number of ticks in the horizontal
20385 (length numbers-list)
20391 (* symbol-width X-axis-label-spacing)
20395 ;; number-of-X-ticks
20396 (if (zerop (% (X-length tic-width)))
20397 (/ (X-length tic-width))
20398 (1+ (/ (X-length tic-width))))
20403 All this leads us directly to the function for printing the X axis tic line:
20405 @findex print-X-axis-tic-line
20408 (defun print-X-axis-tic-line
20409 (number-of-X-tics X-axis-leading-spaces X-axis-tic-element)
20410 "Print ticks for X axis."
20411 (insert X-axis-leading-spaces)
20412 (insert X-axis-tic-symbol) ; @r{Under first column.}
20415 ;; @r{Insert second tic in the right spot.}
20418 (- (* symbol-width X-axis-label-spacing)
20419 ;; @r{Insert white space up to second tic symbol.}
20420 (* 2 (length X-axis-tic-symbol)))
20422 X-axis-tic-symbol))
20425 ;; @r{Insert remaining ticks.}
20426 (while (> number-of-X-tics 1)
20427 (insert X-axis-tic-element)
20428 (setq number-of-X-tics (1- number-of-X-tics))))
20432 The line of numbers is equally straightforward:
20435 First, we create a numbered element with blank spaces before each number:
20437 @findex X-axis-element
20440 (defun X-axis-element (number)
20441 "Construct a numbered X axis element."
20442 (let ((leading-spaces
20443 (- (* symbol-width X-axis-label-spacing)
20444 (length (number-to-string number)))))
20445 (concat (make-string leading-spaces ? )
20446 (number-to-string number))))
20450 Next, we create the function to print the numbered line, starting with
20451 the number ``1'' under the first column:
20453 @findex print-X-axis-numbered-line
20456 (defun print-X-axis-numbered-line
20457 (number-of-X-tics X-axis-leading-spaces)
20458 "Print line of X-axis numbers"
20459 (let ((number X-axis-label-spacing))
20460 (insert X-axis-leading-spaces)
20466 ;; @r{Insert white space up to next number.}
20467 (- (* symbol-width X-axis-label-spacing) 2)
20469 (number-to-string number)))
20472 ;; @r{Insert remaining numbers.}
20473 (setq number (+ number X-axis-label-spacing))
20474 (while (> number-of-X-tics 1)
20475 (insert (X-axis-element number))
20476 (setq number (+ number X-axis-label-spacing))
20477 (setq number-of-X-tics (1- number-of-X-tics)))))
20481 Finally, we need to write the @code{print-X-axis} that uses
20482 @code{print-X-axis-tic-line} and
20483 @code{print-X-axis-numbered-line}.
20485 The function must determine the local values of the variables used by both
20486 @code{print-X-axis-tic-line} and @code{print-X-axis-numbered-line}, and
20487 then it must call them. Also, it must print the carriage return that
20488 separates the two lines.
20490 The function consists of a varlist that specifies five local variables,
20491 and calls to each of the two line printing functions:
20493 @findex print-X-axis
20496 (defun print-X-axis (numbers-list)
20497 "Print X axis labels to length of NUMBERS-LIST."
20498 (let* ((leading-spaces
20499 (make-string full-Y-label-width ? ))
20502 ;; symbol-width @r{is provided by} graph-body-print
20503 (tic-width (* symbol-width X-axis-label-spacing))
20504 (X-length (length numbers-list))
20512 ;; @r{Make a string of blanks.}
20513 (- (* symbol-width X-axis-label-spacing)
20514 (length X-axis-tic-symbol))
20518 ;; @r{Concatenate blanks with tic symbol.}
20519 X-axis-tic-symbol))
20523 (if (zerop (% X-length tic-width))
20524 (/ X-length tic-width)
20525 (1+ (/ X-length tic-width)))))
20528 (print-X-axis-tic-line tic-number leading-spaces X-tic)
20530 (print-X-axis-numbered-line tic-number leading-spaces)))
20535 You can test @code{print-X-axis}:
20539 Install @code{X-axis-tic-symbol}, @code{X-axis-label-spacing},
20540 @code{print-X-axis-tic-line}, as well as @code{X-axis-element},
20541 @code{print-X-axis-numbered-line}, and @code{print-X-axis}.
20544 Copy the following expression:
20549 (let ((full-Y-label-width 5)
20552 '(1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16))))
20557 Switch to the @file{*scratch*} buffer and place the cursor where you
20558 want the axis labels to start.
20561 Type @kbd{M-:} (@code{eval-expression}).
20564 Yank the test expression into the minibuffer
20565 with @kbd{C-y} (@code{yank)}.
20568 Press @key{RET} to evaluate the expression.
20572 Emacs will print the horizontal axis like this:
20582 @node Print Whole Graph
20583 @appendixsec Printing the Whole Graph
20584 @cindex Printing the whole graph
20585 @cindex Whole graph printing
20586 @cindex Graph, printing all
20588 Now we are nearly ready to print the whole graph.
20590 The function to print the graph with the proper labels follows the
20591 outline we created earlier (@pxref{Full Graph, , A Graph with Labeled
20592 Axes}), but with additions.
20595 Here is the outline:
20599 (defun print-graph (numbers-list)
20600 "@var{documentation}@dots{}"
20601 (let ((height @dots{}
20605 (print-Y-axis height @dots{} )
20606 (graph-body-print numbers-list)
20607 (print-X-axis @dots{} )))
20612 * The final version:: A few changes.
20613 * Test print-graph:: Run a short test.
20614 * Graphing words in defuns:: Executing the final code.
20615 * lambda:: How to write an anonymous function.
20616 * mapcar:: Apply a function to elements of a list.
20617 * Another Bug:: Yet another bug @dots{} most insidious.
20618 * Final printed graph:: The graph itself!
20622 @node The final version
20623 @unnumberedsubsec Changes for the Final Version
20626 The final version is different from what we planned in two ways:
20627 first, it contains additional values calculated once in the varlist;
20628 second, it carries an option to specify the labels' increment per row.
20629 This latter feature turns out to be essential; otherwise, a graph may
20630 have more rows than fit on a display or on a sheet of paper.
20633 This new feature requires a change to the @code{Y-axis-column}
20634 function, to add @code{vertical-step} to it. The function looks like
20637 @findex Y-axis-column @r{Final version.}
20640 ;;; @r{Final version.}
20641 (defun Y-axis-column
20642 (height width-of-label &optional vertical-step)
20643 "Construct list of labels for Y axis.
20644 HEIGHT is maximum height of graph.
20645 WIDTH-OF-LABEL is maximum width of label.
20646 VERTICAL-STEP, an option, is a positive integer
20647 that specifies how much a Y axis label increments
20648 for each line. For example, a step of 5 means
20649 that each line is five units of the graph."
20653 (number-per-line (or vertical-step 1)))
20654 (while (> height 1)
20655 (if (zerop (% height Y-axis-label-spacing))
20658 ;; @r{Insert label.}
20662 (* height number-per-line)
20667 ;; @r{Else, insert blanks.}
20670 (make-string width-of-label ? )
20672 (setq height (1- height)))
20675 ;; @r{Insert base line.}
20676 (setq Y-axis (cons (Y-axis-element
20677 (or vertical-step 1)
20680 (nreverse Y-axis)))
20684 The values for the maximum height of graph and the width of a symbol
20685 are computed by @code{print-graph} in its @code{let} expression; so
20686 @code{graph-body-print} must be changed to accept them.
20688 @findex graph-body-print @r{Final version.}
20691 ;;; @r{Final version.}
20692 (defun graph-body-print (numbers-list height symbol-width)
20693 "Print a bar graph of the NUMBERS-LIST.
20694 The numbers-list consists of the Y-axis values.
20695 HEIGHT is maximum height of graph.
20696 SYMBOL-WIDTH is number of each column."
20699 (let (from-position)
20700 (while numbers-list
20701 (setq from-position (point))
20703 (column-of-graph height (car numbers-list)))
20704 (goto-char from-position)
20705 (forward-char symbol-width)
20708 ;; @r{Draw graph column by column.}
20710 (setq numbers-list (cdr numbers-list)))
20711 ;; @r{Place point for X axis labels.}
20712 (forward-line height)
20718 Finally, the code for the @code{print-graph} function:
20720 @findex print-graph @r{Final version.}
20723 ;;; @r{Final version.}
20725 (numbers-list &optional vertical-step)
20726 "Print labeled bar graph of the NUMBERS-LIST.
20727 The numbers-list consists of the Y-axis values.
20731 Optionally, VERTICAL-STEP, a positive integer,
20732 specifies how much a Y axis label increments for
20733 each line. For example, a step of 5 means that
20734 each row is five units."
20737 (let* ((symbol-width (length graph-blank))
20738 ;; @code{height} @r{is both the largest number}
20739 ;; @r{and the number with the most digits.}
20740 (height (apply 'max numbers-list))
20743 (height-of-top-line
20744 (if (zerop (% height Y-axis-label-spacing))
20747 (* (1+ (/ height Y-axis-label-spacing))
20748 Y-axis-label-spacing)))
20751 (vertical-step (or vertical-step 1))
20752 (full-Y-label-width
20758 (* height-of-top-line vertical-step))
20764 height-of-top-line full-Y-label-width vertical-step)
20768 numbers-list height-of-top-line symbol-width)
20769 (print-X-axis numbers-list)))
20773 @node Test print-graph
20774 @appendixsubsec Testing @code{print-graph}
20777 We can test the @code{print-graph} function with a short list of numbers:
20781 Install the final versions of @code{Y-axis-column},
20782 @code{graph-body-print}, and @code{print-graph} (in addition to the
20786 Copy the following expression:
20789 (print-graph '(3 2 5 6 7 5 3 4 6 4 3 2 1))
20793 Switch to the @file{*scratch*} buffer and place the cursor where you
20794 want the axis labels to start.
20797 Type @kbd{M-:} (@code{eval-expression}).
20800 Yank the test expression into the minibuffer
20801 with @kbd{C-y} (@code{yank)}.
20804 Press @key{RET} to evaluate the expression.
20808 Emacs will print a graph that looks like this:
20829 On the other hand, if you pass @code{print-graph} a
20830 @code{vertical-step} value of 2, by evaluating this expression:
20833 (print-graph '(3 2 5 6 7 5 3 4 6 4 3 2 1) 2)
20838 The graph looks like this:
20859 (A question: is the `2' on the bottom of the vertical axis a bug or a
20860 feature? If you think it is a bug, and should be a `1' instead, (or
20861 even a `0'), you can modify the sources.)
20863 @node Graphing words in defuns
20864 @appendixsubsec Graphing Numbers of Words and Symbols
20866 Now for the graph for which all this code was written: a graph that
20867 shows how many function definitions contain fewer than 10 words and
20868 symbols, how many contain between 10 and 19 words and symbols, how
20869 many contain between 20 and 29 words and symbols, and so on.
20871 This is a multi-step process. First make sure you have loaded all the
20875 It is a good idea to reset the value of @code{top-of-ranges} in case
20876 you have set it to some different value. You can evaluate the
20881 (setq top-of-ranges
20884 110 120 130 140 150
20885 160 170 180 190 200
20886 210 220 230 240 250
20887 260 270 280 290 300)
20892 Next create a list of the number of words and symbols in each range.
20896 Evaluate the following:
20900 (setq list-for-graph
20903 (recursive-lengths-list-many-files
20904 (directory-files "/usr/local/emacs/lisp"
20912 On my old machine, this took about an hour. It looked though 303 Lisp
20913 files in my copy of Emacs version 19.23. After all that computing,
20914 the @code{list-for-graph} had this value:
20918 (537 1027 955 785 594 483 349 292 224 199 166 120 116 99
20919 90 80 67 48 52 45 41 33 28 26 25 20 12 28 11 13 220)
20924 This means that my copy of Emacs had 537 function definitions with
20925 fewer than 10 words or symbols in them, 1,027 function definitions
20926 with 10 to 19 words or symbols in them, 955 function definitions with
20927 20 to 29 words or symbols in them, and so on.
20929 Clearly, just by looking at this list we can see that most function
20930 definitions contain ten to thirty words and symbols.
20932 Now for printing. We do @emph{not} want to print a graph that is
20933 1,030 lines high @dots{} Instead, we should print a graph that is
20934 fewer than twenty-five lines high. A graph that height can be
20935 displayed on almost any monitor, and easily printed on a sheet of paper.
20937 This means that each value in @code{list-for-graph} must be reduced to
20938 one-fiftieth its present value.
20940 Here is a short function to do just that, using two functions we have
20941 not yet seen, @code{mapcar} and @code{lambda}.
20945 (defun one-fiftieth (full-range)
20946 "Return list, each number one-fiftieth of previous."
20947 (mapcar (lambda (arg) (/ arg 50)) full-range))
20952 @appendixsubsec A @code{lambda} Expression: Useful Anonymity
20953 @cindex Anonymous function
20956 @code{lambda} is the symbol for an anonymous function, a function
20957 without a name. Every time you use an anonymous function, you need to
20958 include its whole body.
20965 (lambda (arg) (/ arg 50))
20969 is a function definition that says `return the value resulting from
20970 dividing whatever is passed to me as @code{arg} by 50'.
20973 Earlier, for example, we had a function @code{multiply-by-seven}; it
20974 multiplied its argument by 7. This function is similar, except it
20975 divides its argument by 50; and, it has no name. The anonymous
20976 equivalent of @code{multiply-by-seven} is:
20979 (lambda (number) (* 7 number))
20983 (@xref{defun, , The @code{defun} Special Form}.)
20987 If we want to multiply 3 by 7, we can write:
20989 @c !!! Clear print-postscript-figures if the computer formatting this
20990 @c document is too small and cannot handle all the diagrams and figures.
20991 @c clear print-postscript-figures
20992 @c set print-postscript-figures
20993 @c lambda example diagram #1
20997 (multiply-by-seven 3)
20998 \_______________/ ^
21004 @ifset print-postscript-figures
21007 @center @image{lambda-1}
21008 %%%% old method of including an image
21009 % \input /usr/local/lib/tex/inputs/psfig.tex
21010 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/lambda-1.eps}}
21015 @ifclear print-postscript-figures
21019 (multiply-by-seven 3)
21020 \_______________/ ^
21029 This expression returns 21.
21033 Similarly, we can write:
21035 @c lambda example diagram #2
21039 ((lambda (number) (* 7 number)) 3)
21040 \____________________________/ ^
21042 anonymous function argument
21046 @ifset print-postscript-figures
21049 @center @image{lambda-2}
21050 %%%% old method of including an image
21051 % \input /usr/local/lib/tex/inputs/psfig.tex
21052 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/lambda-2.eps}}
21057 @ifclear print-postscript-figures
21061 ((lambda (number) (* 7 number)) 3)
21062 \____________________________/ ^
21064 anonymous function argument
21072 If we want to divide 100 by 50, we can write:
21074 @c lambda example diagram #3
21078 ((lambda (arg) (/ arg 50)) 100)
21079 \______________________/ \_/
21081 anonymous function argument
21085 @ifset print-postscript-figures
21088 @center @image{lambda-3}
21089 %%%% old method of including an image
21090 % \input /usr/local/lib/tex/inputs/psfig.tex
21091 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/lambda-3.eps}}
21096 @ifclear print-postscript-figures
21100 ((lambda (arg) (/ arg 50)) 100)
21101 \______________________/ \_/
21103 anonymous function argument
21110 This expression returns 2. The 100 is passed to the function, which
21111 divides that number by 50.
21113 @xref{Lambda Expressions, , Lambda Expressions, elisp, The GNU Emacs
21114 Lisp Reference Manual}, for more about @code{lambda}. Lisp and lambda
21115 expressions derive from the Lambda Calculus.
21118 @appendixsubsec The @code{mapcar} Function
21121 @code{mapcar} is a function that calls its first argument with each
21122 element of its second argument, in turn. The second argument must be
21125 The @samp{map} part of the name comes from the mathematical phrase,
21126 `mapping over a domain', meaning to apply a function to each of the
21127 elements in a domain. The mathematical phrase is based on the
21128 metaphor of a surveyor walking, one step at a time, over an area he is
21129 mapping. And @samp{car}, of course, comes from the Lisp notion of the
21138 (mapcar '1+ '(2 4 6))
21144 The function @code{1+} which adds one to its argument, is executed on
21145 @emph{each} element of the list, and a new list is returned.
21147 Contrast this with @code{apply}, which applies its first argument to
21149 (@xref{Readying a Graph, , Readying a Graph}, for a explanation of
21153 In the definition of @code{one-fiftieth}, the first argument is the
21154 anonymous function:
21157 (lambda (arg) (/ arg 50))
21161 and the second argument is @code{full-range}, which will be bound to
21162 @code{list-for-graph}.
21165 The whole expression looks like this:
21168 (mapcar (lambda (arg) (/ arg 50)) full-range))
21171 @xref{Mapping Functions, , Mapping Functions, elisp, The GNU Emacs
21172 Lisp Reference Manual}, for more about @code{mapcar}.
21174 Using the @code{one-fiftieth} function, we can generate a list in
21175 which each element is one-fiftieth the size of the corresponding
21176 element in @code{list-for-graph}.
21180 (setq fiftieth-list-for-graph
21181 (one-fiftieth list-for-graph))
21186 The resulting list looks like this:
21190 (10 20 19 15 11 9 6 5 4 3 3 2 2
21191 1 1 1 1 0 1 0 0 0 0 0 0 0 0 0 0 0 4)
21196 This, we are almost ready to print! (We also notice the loss of
21197 information: many of the higher ranges are 0, meaning that fewer than
21198 50 defuns had that many words or symbols---but not necessarily meaning
21199 that none had that many words or symbols.)
21202 @appendixsubsec Another Bug @dots{} Most Insidious
21203 @cindex Bug, most insidious type
21204 @cindex Insidious type of bug
21206 I said `almost ready to print'! Of course, there is a bug in the
21207 @code{print-graph} function @dots{} It has a @code{vertical-step}
21208 option, but not a @code{horizontal-step} option. The
21209 @code{top-of-range} scale goes from 10 to 300 by tens. But the
21210 @code{print-graph} function will print only by ones.
21212 This is a classic example of what some consider the most insidious
21213 type of bug, the bug of omission. This is not the kind of bug you can
21214 find by studying the code, for it is not in the code; it is an omitted
21215 feature. Your best actions are to try your program early and often;
21216 and try to arrange, as much as you can, to write code that is easy to
21217 understand and easy to change. Try to be aware, whenever you can,
21218 that whatever you have written, @emph{will} be rewritten, if not soon,
21219 eventually. A hard maxim to follow.
21221 It is the @code{print-X-axis-numbered-line} function that needs the
21222 work; and then the @code{print-X-axis} and the @code{print-graph}
21223 functions need to be adapted. Not much needs to be done; there is one
21224 nicety: the numbers ought to line up under the tic marks. This takes
21228 Here is the corrected @code{print-X-axis-numbered-line}:
21232 (defun print-X-axis-numbered-line
21233 (number-of-X-tics X-axis-leading-spaces
21234 &optional horizontal-step)
21235 "Print line of X-axis numbers"
21236 (let ((number X-axis-label-spacing)
21237 (horizontal-step (or horizontal-step 1)))
21240 (insert X-axis-leading-spaces)
21241 ;; @r{Delete extra leading spaces.}
21244 (length (number-to-string horizontal-step)))))
21249 ;; @r{Insert white space.}
21251 X-axis-label-spacing)
21254 (number-to-string horizontal-step)))
21258 (* number horizontal-step))))
21261 ;; @r{Insert remaining numbers.}
21262 (setq number (+ number X-axis-label-spacing))
21263 (while (> number-of-X-tics 1)
21264 (insert (X-axis-element
21265 (* number horizontal-step)))
21266 (setq number (+ number X-axis-label-spacing))
21267 (setq number-of-X-tics (1- number-of-X-tics)))))
21272 If you are reading this in Info, you can see the new versions of
21273 @code{print-X-axis} @code{print-graph} and evaluate them. If you are
21274 reading this in a printed book, you can see the changed lines here
21275 (the full text is too much to print).
21280 (defun print-X-axis (numbers-list horizontal-step)
21282 (print-X-axis-numbered-line
21283 tic-number leading-spaces horizontal-step))
21291 &optional vertical-step horizontal-step)
21293 (print-X-axis numbers-list horizontal-step))
21301 (defun print-X-axis (numbers-list horizontal-step)
21302 "Print X axis labels to length of NUMBERS-LIST.
21303 Optionally, HORIZONTAL-STEP, a positive integer,
21304 specifies how much an X axis label increments for
21308 ;; Value of symbol-width and full-Y-label-width
21309 ;; are passed by `print-graph'.
21310 (let* ((leading-spaces
21311 (make-string full-Y-label-width ? ))
21312 ;; symbol-width @r{is provided by} graph-body-print
21313 (tic-width (* symbol-width X-axis-label-spacing))
21314 (X-length (length numbers-list))
21320 ;; @r{Make a string of blanks.}
21321 (- (* symbol-width X-axis-label-spacing)
21322 (length X-axis-tic-symbol))
21326 ;; @r{Concatenate blanks with tic symbol.}
21327 X-axis-tic-symbol))
21329 (if (zerop (% X-length tic-width))
21330 (/ X-length tic-width)
21331 (1+ (/ X-length tic-width)))))
21335 (print-X-axis-tic-line
21336 tic-number leading-spaces X-tic)
21338 (print-X-axis-numbered-line
21339 tic-number leading-spaces horizontal-step)))
21346 (numbers-list &optional vertical-step horizontal-step)
21347 "Print labeled bar graph of the NUMBERS-LIST.
21348 The numbers-list consists of the Y-axis values.
21352 Optionally, VERTICAL-STEP, a positive integer,
21353 specifies how much a Y axis label increments for
21354 each line. For example, a step of 5 means that
21355 each row is five units.
21359 Optionally, HORIZONTAL-STEP, a positive integer,
21360 specifies how much an X axis label increments for
21362 (let* ((symbol-width (length graph-blank))
21363 ;; @code{height} @r{is both the largest number}
21364 ;; @r{and the number with the most digits.}
21365 (height (apply 'max numbers-list))
21368 (height-of-top-line
21369 (if (zerop (% height Y-axis-label-spacing))
21372 (* (1+ (/ height Y-axis-label-spacing))
21373 Y-axis-label-spacing)))
21376 (vertical-step (or vertical-step 1))
21377 (full-Y-label-width
21381 (* height-of-top-line vertical-step))
21386 height-of-top-line full-Y-label-width vertical-step)
21388 numbers-list height-of-top-line symbol-width)
21389 (print-X-axis numbers-list horizontal-step)))
21396 Graphing Definitions Re-listed
21399 Here are all the graphing definitions in their final form:
21403 (defvar top-of-ranges
21406 110 120 130 140 150
21407 160 170 180 190 200
21408 210 220 230 240 250)
21409 "List specifying ranges for `defuns-per-range'.")
21413 (defvar graph-symbol "*"
21414 "String used as symbol in graph, usually an asterisk.")
21418 (defvar graph-blank " "
21419 "String used as blank in graph, usually a blank space.
21420 graph-blank must be the same number of columns wide
21425 (defvar Y-axis-tic " - "
21426 "String that follows number in a Y axis label.")
21430 (defvar Y-axis-label-spacing 5
21431 "Number of lines from one Y axis label to next.")
21435 (defvar X-axis-tic-symbol "|"
21436 "String to insert to point to a column in X axis.")
21440 (defvar X-axis-label-spacing
21441 (if (boundp 'graph-blank)
21442 (* 5 (length graph-blank)) 5)
21443 "Number of units from one X axis label to next.")
21449 (defun count-words-in-defun ()
21450 "Return the number of words and symbols in a defun."
21451 (beginning-of-defun)
21453 (end (save-excursion (end-of-defun) (point))))
21458 (and (< (point) end)
21460 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*"
21462 (setq count (1+ count)))
21469 (defun lengths-list-file (filename)
21470 "Return list of definitions' lengths within FILE.
21471 The returned list is a list of numbers.
21472 Each number is the number of words or
21473 symbols in one function definition."
21477 (message "Working on `%s' ... " filename)
21479 (let ((buffer (find-file-noselect filename))
21481 (set-buffer buffer)
21482 (setq buffer-read-only t)
21484 (goto-char (point-min))
21488 (while (re-search-forward "^(defun" nil t)
21490 (cons (count-words-in-defun) lengths-list)))
21491 (kill-buffer buffer)
21498 (defun lengths-list-many-files (list-of-files)
21499 "Return list of lengths of defuns in LIST-OF-FILES."
21500 (let (lengths-list)
21501 ;;; @r{true-or-false-test}
21502 (while list-of-files
21508 ;;; @r{Generate a lengths' list.}
21510 (expand-file-name (car list-of-files)))))
21511 ;;; @r{Make files' list shorter.}
21512 (setq list-of-files (cdr list-of-files)))
21513 ;;; @r{Return final value of lengths' list.}
21520 (defun defuns-per-range (sorted-lengths top-of-ranges)
21521 "SORTED-LENGTHS defuns in each TOP-OF-RANGES range."
21522 (let ((top-of-range (car top-of-ranges))
21523 (number-within-range 0)
21524 defuns-per-range-list)
21529 (while top-of-ranges
21533 ;; @r{Need number for numeric test.}
21534 (car sorted-lengths)
21535 (< (car sorted-lengths) top-of-range))
21537 ;; @r{Count number of definitions within current range.}
21538 (setq number-within-range (1+ number-within-range))
21539 (setq sorted-lengths (cdr sorted-lengths)))
21543 ;; @r{Exit inner loop but remain within outer loop.}
21545 (setq defuns-per-range-list
21546 (cons number-within-range defuns-per-range-list))
21547 (setq number-within-range 0) ; @r{Reset count to zero.}
21549 ;; @r{Move to next range.}
21550 (setq top-of-ranges (cdr top-of-ranges))
21551 ;; @r{Specify next top of range value.}
21552 (setq top-of-range (car top-of-ranges)))
21556 ;; @r{Exit outer loop and count the number of defuns larger than}
21557 ;; @r{ the largest top-of-range value.}
21558 (setq defuns-per-range-list
21560 (length sorted-lengths)
21561 defuns-per-range-list))
21563 ;; @r{Return a list of the number of definitions within each range,}
21564 ;; @r{ smallest to largest.}
21565 (nreverse defuns-per-range-list)))
21571 (defun column-of-graph (max-graph-height actual-height)
21572 "Return list of MAX-GRAPH-HEIGHT strings;
21573 ACTUAL-HEIGHT are graph-symbols.
21574 The graph-symbols are contiguous entries at the end
21576 The list will be inserted as one column of a graph.
21577 The strings are either graph-blank or graph-symbol."
21581 (let ((insert-list nil)
21582 (number-of-top-blanks
21583 (- max-graph-height actual-height)))
21585 ;; @r{Fill in @code{graph-symbols}.}
21586 (while (> actual-height 0)
21587 (setq insert-list (cons graph-symbol insert-list))
21588 (setq actual-height (1- actual-height)))
21592 ;; @r{Fill in @code{graph-blanks}.}
21593 (while (> number-of-top-blanks 0)
21594 (setq insert-list (cons graph-blank insert-list))
21595 (setq number-of-top-blanks
21596 (1- number-of-top-blanks)))
21598 ;; @r{Return whole list.}
21605 (defun Y-axis-element (number full-Y-label-width)
21606 "Construct a NUMBERed label element.
21607 A numbered element looks like this ` 5 - ',
21608 and is padded as needed so all line up with
21609 the element for the largest number."
21612 (let* ((leading-spaces
21613 (- full-Y-label-width
21615 (concat (number-to-string number)
21620 (make-string leading-spaces ? )
21621 (number-to-string number)
21628 (defun print-Y-axis
21629 (height full-Y-label-width &optional vertical-step)
21630 "Insert Y axis by HEIGHT and FULL-Y-LABEL-WIDTH.
21631 Height must be the maximum height of the graph.
21632 Full width is the width of the highest label element.
21633 Optionally, print according to VERTICAL-STEP."
21636 ;; Value of height and full-Y-label-width
21637 ;; are passed by `print-graph'.
21638 (let ((start (point)))
21640 (Y-axis-column height full-Y-label-width vertical-step))
21643 ;; @r{Place point ready for inserting graph.}
21645 ;; @r{Move point forward by value of} full-Y-label-width
21646 (forward-char full-Y-label-width)))
21652 (defun print-X-axis-tic-line
21653 (number-of-X-tics X-axis-leading-spaces X-axis-tic-element)
21654 "Print ticks for X axis."
21655 (insert X-axis-leading-spaces)
21656 (insert X-axis-tic-symbol) ; @r{Under first column.}
21659 ;; @r{Insert second tic in the right spot.}
21662 (- (* symbol-width X-axis-label-spacing)
21663 ;; @r{Insert white space up to second tic symbol.}
21664 (* 2 (length X-axis-tic-symbol)))
21666 X-axis-tic-symbol))
21669 ;; @r{Insert remaining ticks.}
21670 (while (> number-of-X-tics 1)
21671 (insert X-axis-tic-element)
21672 (setq number-of-X-tics (1- number-of-X-tics))))
21678 (defun X-axis-element (number)
21679 "Construct a numbered X axis element."
21680 (let ((leading-spaces
21681 (- (* symbol-width X-axis-label-spacing)
21682 (length (number-to-string number)))))
21683 (concat (make-string leading-spaces ? )
21684 (number-to-string number))))
21690 (defun graph-body-print (numbers-list height symbol-width)
21691 "Print a bar graph of the NUMBERS-LIST.
21692 The numbers-list consists of the Y-axis values.
21693 HEIGHT is maximum height of graph.
21694 SYMBOL-WIDTH is number of each column."
21697 (let (from-position)
21698 (while numbers-list
21699 (setq from-position (point))
21701 (column-of-graph height (car numbers-list)))
21702 (goto-char from-position)
21703 (forward-char symbol-width)
21706 ;; @r{Draw graph column by column.}
21708 (setq numbers-list (cdr numbers-list)))
21709 ;; @r{Place point for X axis labels.}
21710 (forward-line height)
21717 (defun Y-axis-column
21718 (height width-of-label &optional vertical-step)
21719 "Construct list of labels for Y axis.
21720 HEIGHT is maximum height of graph.
21721 WIDTH-OF-LABEL is maximum width of label.
21724 VERTICAL-STEP, an option, is a positive integer
21725 that specifies how much a Y axis label increments
21726 for each line. For example, a step of 5 means
21727 that each line is five units of the graph."
21729 (number-per-line (or vertical-step 1)))
21732 (while (> height 1)
21733 (if (zerop (% height Y-axis-label-spacing))
21734 ;; @r{Insert label.}
21738 (* height number-per-line)
21743 ;; @r{Else, insert blanks.}
21746 (make-string width-of-label ? )
21748 (setq height (1- height)))
21751 ;; @r{Insert base line.}
21752 (setq Y-axis (cons (Y-axis-element
21753 (or vertical-step 1)
21756 (nreverse Y-axis)))
21762 (defun print-X-axis-numbered-line
21763 (number-of-X-tics X-axis-leading-spaces
21764 &optional horizontal-step)
21765 "Print line of X-axis numbers"
21766 (let ((number X-axis-label-spacing)
21767 (horizontal-step (or horizontal-step 1)))
21770 (insert X-axis-leading-spaces)
21772 (delete-char (- (1- (length (number-to-string horizontal-step)))))
21775 ;; @r{Insert white space up to next number.}
21776 (- (* symbol-width X-axis-label-spacing)
21777 (1- (length (number-to-string horizontal-step)))
21780 (number-to-string (* number horizontal-step))))
21783 ;; @r{Insert remaining numbers.}
21784 (setq number (+ number X-axis-label-spacing))
21785 (while (> number-of-X-tics 1)
21786 (insert (X-axis-element (* number horizontal-step)))
21787 (setq number (+ number X-axis-label-spacing))
21788 (setq number-of-X-tics (1- number-of-X-tics)))))
21794 (defun print-X-axis (numbers-list horizontal-step)
21795 "Print X axis labels to length of NUMBERS-LIST.
21796 Optionally, HORIZONTAL-STEP, a positive integer,
21797 specifies how much an X axis label increments for
21801 ;; Value of symbol-width and full-Y-label-width
21802 ;; are passed by `print-graph'.
21803 (let* ((leading-spaces
21804 (make-string full-Y-label-width ? ))
21805 ;; symbol-width @r{is provided by} graph-body-print
21806 (tic-width (* symbol-width X-axis-label-spacing))
21807 (X-length (length numbers-list))
21813 ;; @r{Make a string of blanks.}
21814 (- (* symbol-width X-axis-label-spacing)
21815 (length X-axis-tic-symbol))
21819 ;; @r{Concatenate blanks with tic symbol.}
21820 X-axis-tic-symbol))
21822 (if (zerop (% X-length tic-width))
21823 (/ X-length tic-width)
21824 (1+ (/ X-length tic-width)))))
21828 (print-X-axis-tic-line
21829 tic-number leading-spaces X-tic)
21831 (print-X-axis-numbered-line
21832 tic-number leading-spaces horizontal-step)))
21838 (defun one-fiftieth (full-range)
21839 "Return list, each number of which is 1/50th previous."
21840 (mapcar (lambda (arg) (/ arg 50)) full-range))
21847 (numbers-list &optional vertical-step horizontal-step)
21848 "Print labeled bar graph of the NUMBERS-LIST.
21849 The numbers-list consists of the Y-axis values.
21853 Optionally, VERTICAL-STEP, a positive integer,
21854 specifies how much a Y axis label increments for
21855 each line. For example, a step of 5 means that
21856 each row is five units.
21860 Optionally, HORIZONTAL-STEP, a positive integer,
21861 specifies how much an X axis label increments for
21863 (let* ((symbol-width (length graph-blank))
21864 ;; @code{height} @r{is both the largest number}
21865 ;; @r{and the number with the most digits.}
21866 (height (apply 'max numbers-list))
21869 (height-of-top-line
21870 (if (zerop (% height Y-axis-label-spacing))
21873 (* (1+ (/ height Y-axis-label-spacing))
21874 Y-axis-label-spacing)))
21877 (vertical-step (or vertical-step 1))
21878 (full-Y-label-width
21882 (* height-of-top-line vertical-step))
21888 height-of-top-line full-Y-label-width vertical-step)
21890 numbers-list height-of-top-line symbol-width)
21891 (print-X-axis numbers-list horizontal-step)))
21898 @node Final printed graph
21899 @appendixsubsec The Printed Graph
21901 When made and installed, you can call the @code{print-graph} command
21907 (print-graph fiftieth-list-for-graph 50 10)
21937 50 - ***************** * *
21939 10 50 100 150 200 250 300 350
21946 The largest group of functions contain 10--19 words and symbols each.
21948 @node Free Software and Free Manuals
21949 @appendix Free Software and Free Manuals
21951 @strong{by Richard M. Stallman}
21954 The biggest deficiency in free operating systems is not in the
21955 software---it is the lack of good free manuals that we can include in
21956 these systems. Many of our most important programs do not come with
21957 full manuals. Documentation is an essential part of any software
21958 package; when an important free software package does not come with a
21959 free manual, that is a major gap. We have many such gaps today.
21961 Once upon a time, many years ago, I thought I would learn Perl. I got
21962 a copy of a free manual, but I found it hard to read. When I asked
21963 Perl users about alternatives, they told me that there were better
21964 introductory manuals---but those were not free.
21966 Why was this? The authors of the good manuals had written them for
21967 O'Reilly Associates, which published them with restrictive terms---no
21968 copying, no modification, source files not available---which exclude
21969 them from the free software community.
21971 That wasn't the first time this sort of thing has happened, and (to
21972 our community's great loss) it was far from the last. Proprietary
21973 manual publishers have enticed a great many authors to restrict their
21974 manuals since then. Many times I have heard a GNU user eagerly tell me
21975 about a manual that he is writing, with which he expects to help the
21976 GNU project---and then had my hopes dashed, as he proceeded to explain
21977 that he had signed a contract with a publisher that would restrict it
21978 so that we cannot use it.
21980 Given that writing good English is a rare skill among programmers, we
21981 can ill afford to lose manuals this way.
21983 Free documentation, like free software, is a matter of freedom, not
21984 price. The problem with these manuals was not that O'Reilly Associates
21985 charged a price for printed copies---that in itself is fine. The Free
21986 Software Foundation @uref{http://shop.fsf.org, sells printed copies} of
21987 free @uref{http://www.gnu.org/doc/doc.html, GNU manuals}, too.
21988 But GNU manuals are available in source code form, while these manuals
21989 are available only on paper. GNU manuals come with permission to copy
21990 and modify; the Perl manuals do not. These restrictions are the
21993 The criterion for a free manual is pretty much the same as for free
21994 software: it is a matter of giving all users certain
21995 freedoms. Redistribution (including commercial redistribution) must be
21996 permitted, so that the manual can accompany every copy of the program,
21997 on-line or on paper. Permission for modification is crucial too.
21999 As a general rule, I don't believe that it is essential for people to
22000 have permission to modify all sorts of articles and books. The issues
22001 for writings are not necessarily the same as those for software. For
22002 example, I don't think you or I are obliged to give permission to
22003 modify articles like this one, which describe our actions and our
22006 But there is a particular reason why the freedom to modify is crucial
22007 for documentation for free software. When people exercise their right
22008 to modify the software, and add or change its features, if they are
22009 conscientious they will change the manual too---so they can provide
22010 accurate and usable documentation with the modified program. A manual
22011 which forbids programmers to be conscientious and finish the job, or
22012 more precisely requires them to write a new manual from scratch if
22013 they change the program, does not fill our community's needs.
22015 While a blanket prohibition on modification is unacceptable, some
22016 kinds of limits on the method of modification pose no problem. For
22017 example, requirements to preserve the original author's copyright
22018 notice, the distribution terms, or the list of authors, are ok. It is
22019 also no problem to require modified versions to include notice that
22020 they were modified, even to have entire sections that may not be
22021 deleted or changed, as long as these sections deal with nontechnical
22022 topics. (Some GNU manuals have them.)
22024 These kinds of restrictions are not a problem because, as a practical
22025 matter, they don't stop the conscientious programmer from adapting the
22026 manual to fit the modified program. In other words, they don't block
22027 the free software community from making full use of the manual.
22029 However, it must be possible to modify all the technical content of
22030 the manual, and then distribute the result in all the usual media,
22031 through all the usual channels; otherwise, the restrictions do block
22032 the community, the manual is not free, and so we need another manual.
22034 Unfortunately, it is often hard to find someone to write another
22035 manual when a proprietary manual exists. The obstacle is that many
22036 users think that a proprietary manual is good enough---so they don't
22037 see the need to write a free manual. They do not see that the free
22038 operating system has a gap that needs filling.
22040 Why do users think that proprietary manuals are good enough? Some have
22041 not considered the issue. I hope this article will do something to
22044 Other users consider proprietary manuals acceptable for the same
22045 reason so many people consider proprietary software acceptable: they
22046 judge in purely practical terms, not using freedom as a
22047 criterion. These people are entitled to their opinions, but since
22048 those opinions spring from values which do not include freedom, they
22049 are no guide for those of us who do value freedom.
22051 Please spread the word about this issue. We continue to lose manuals
22052 to proprietary publishing. If we spread the word that proprietary
22053 manuals are not sufficient, perhaps the next person who wants to help
22054 GNU by writing documentation will realize, before it is too late, that
22055 he must above all make it free.
22057 We can also encourage commercial publishers to sell free, copylefted
22058 manuals instead of proprietary ones. One way you can help this is to
22059 check the distribution terms of a manual before you buy it, and prefer
22060 copylefted manuals to non-copylefted ones.
22064 Note: The Free Software Foundation maintains a page on its Web site
22065 that lists free books available from other publishers:@*
22066 @uref{http://www.gnu.org/doc/other-free-books.html}
22068 @node GNU Free Documentation License
22069 @appendix GNU Free Documentation License
22071 @cindex FDL, GNU Free Documentation License
22072 @include doclicense.texi
22078 MENU ENTRY: NODE NAME.
22084 @c Place biographical information on right-hand (verso) page
22087 \par\vfill\supereject
22089 \global\evenheadline={\hfil} \global\evenfootline={\hfil}
22090 \global\oddheadline={\hfil} \global\oddfootline={\hfil}
22093 % \par\vfill\supereject
22094 \global\evenheadline={\hfil} \global\evenfootline={\hfil}
22095 \global\oddheadline={\hfil} \global\oddfootline={\hfil}
22096 %\page\hbox{}%\page
22097 %\page\hbox{}%\page
22104 @c ================ Biographical information ================
22108 @center About the Author
22113 @node About the Author
22114 @unnumbered About the Author
22118 Robert J. Chassell has worked with GNU Emacs since 1985. He writes
22119 and edits, teaches Emacs and Emacs Lisp, and speaks throughout the
22120 world on software freedom. Chassell was a founding Director and
22121 Treasurer of the Free Software Foundation, Inc. He is co-author of
22122 the @cite{Texinfo} manual, and has edited more than a dozen other
22123 books. He graduated from Cambridge University, in England. He has an
22124 abiding interest in social and economic history and flies his own
22131 @c @c Prevent page number on blank verso, so eject it first.
22133 @c \par\vfill\supereject
22138 @c @evenheading @thispage @| @| @thistitle
22139 @c @oddheading @| @| @thispage