1 \input texinfo @c -*- mode: texinfo; coding: utf-8 -*-
2 @comment %**start of header
3 @setfilename ../../info/eintr.info
4 @c setfilename emacs-lisp-intro.info
5 @c sethtmlfilename emacs-lisp-intro.html
6 @settitle Programming in Emacs Lisp
12 @include emacsver.texi
14 @c ================ How to Print a Book in Various Sizes ================
16 @c This book can be printed in any of three different sizes.
17 @c Set the following @-commands appropriately.
27 @c European A4 size paper:
32 @c (Note: if you edit the book so as to change the length of the
33 @c table of contents, you may have to change the value of 'pageno' below.)
35 @c <<<< For hard copy printing, this file is now
36 @c set for smallbook, which works for all sizes
37 @c of paper, and with PostScript figures >>>>
45 @c ================ Included Figures ================
47 @c If you clear this, the figures will be printed as ASCII diagrams
48 @c rather than PostScript/PDF.
49 @c (This is not relevant to Info, since Info only handles ASCII.)
50 @set print-postscript-figures
51 @c clear print-postscript-figures
53 @comment %**end of header
55 @c per rms and peterb, use 10pt fonts for the main text, mostly to
56 @c save on paper cost.
57 @c Do this inside @tex for now, so current makeinfo does not complain.
63 \global\hbadness=6666 % don't worry about not-too-underfull boxes
66 @c These refer to the printed book sold by the FSF.
67 @set edition-number 3.10
68 @set update-date 28 October 2009
70 @c For next or subsequent edition:
71 @c create function using with-output-to-temp-buffer
72 @c create a major mode, with keymaps
73 @c run an asynchronous process, like grep or diff
75 @c For 8.5 by 11 inch format: do not use such a small amount of
76 @c whitespace between paragraphs as smallbook format
79 \global\parskip 6pt plus 1pt
83 @c For all sized formats: print within-book cross
84 @c reference with ``...'' rather than [...]
86 @c This works with the texinfo.tex file, version 2003-05-04.08,
87 @c in the Texinfo version 4.6 of the 2003 Jun 13 distribution.
90 \if \xrefprintnodename
91 \global\def\xrefprintnodename#1{\unskip, ``#1''}
93 \global\def\xrefprintnodename#1{ ``#1''}
95 % \global\def\xrefprintnodename#1{, ``#1''}
98 @c ----------------------------------------------------
100 @dircategory Emacs lisp
102 * Emacs Lisp Intro: (eintr). A simple introduction to Emacs Lisp programming.
106 This is an @cite{Introduction to Programming in Emacs Lisp}, for
107 people who are not programmers.
110 Edition @value{edition-number}, @value{update-date}
113 Distributed with Emacs version @value{EMACSVER}.
116 Copyright @copyright{} 1990--1995, 1997, 2001--2018 Free Software
123 GNU Press, @hfill @uref{https://www.fsf.org/licensing/gnu-press/}@*
124 a division of the @hfill email: @email{sales@@fsf.org}@*
125 Free Software Foundation, Inc. @hfill Tel: +1 (617) 542-5942@*
126 51 Franklin Street, Fifth Floor @hfill Fax: +1 (617) 542-2652@*
127 Boston, MA 02110-1301 USA
131 Printed copies available from @uref{https://shop.fsf.org/}. Published by:
134 GNU Press, https://www.fsf.org/licensing/gnu-press/
135 a division of the email: sales@@fsf.org
136 Free Software Foundation, Inc. Tel: +1 (617) 542-5942
137 51 Franklin Street, Fifth Floor Fax: +1 (617) 542-2652
138 Boston, MA 02110-1301 USA
146 Permission is granted to copy, distribute and/or modify this document
147 under the terms of the GNU Free Documentation License, Version 1.3 or
148 any later version published by the Free Software Foundation; there
149 being no Invariant Section, with the Front-Cover Texts being ``A GNU
150 Manual'', and with the Back-Cover Texts as in (a) below. A copy of
151 the license is included in the section entitled ``GNU Free
152 Documentation License''.
154 (a) The FSF's Back-Cover Text is: ``You have the freedom to
155 copy and modify this GNU manual. Buying copies from the FSF
156 supports it in developing GNU and promoting software freedom.''
160 @c half title; two lines here, so do not use 'shorttitlepage'
163 \hbox{}\vskip 1.5in \chaprm \centerline{An Introduction to}%
165 {\begingroup\hbox{}\vskip 0.25in \chaprm%
166 \centerline{Programming in Emacs Lisp}%
167 \endgroup\page\hbox{}\page}
172 @center @titlefont{An Introduction to}
174 @center @titlefont{Programming in Emacs Lisp}
176 @center Revised Third Edition
178 @center by Robert J. Chassell
181 @vskip 0pt plus 1filll
187 @evenheading @thispage @| @| @thischapter
188 @oddheading @thissection @| @| @thispage
192 @c Keep T.O.C. short by tightening up for largebook
195 \global\parskip 2pt plus 1pt
196 \global\advance\baselineskip by -1pt
201 @c If you think this manual is too large for an introduction, please
202 @c consider this email exchange:
204 @c >> The intro is almost 300 pages in full. I had expected 60 pages.
206 @c > This is an important point in its own right. Could you
207 @c > write a simplified introduction that is only 50 pages or so?
208 @c > That would be helpful to many potential users, I'd think.
210 @c > The problem with the introduction is that it was written when
211 @c > programming was only starting to be a skill "normal" people could
212 @c > have access to. So the text is extremely verbose and is
213 @c > sometimes hard to follow because of that. The gist of the
214 @c > document could be summarized in 50 pages.
216 @c This book is intentionally addressed to people who don't know how to
217 @c program. That is its purpose. We recommend people start learning to
218 @c program using this book.
220 @c If you DO know how to program in some other language, you can probably
221 @c learn Emacs Lisp starting with the Emacs Lisp Reference Manual.
223 @c Richard Stallman <rms@gnu.org>,
224 @c https://lists.gnu.org/archive/html/emacs-devel/2018-05/msg00374.html
231 @top An Introduction to Programming in Emacs Lisp
235 <p>The homepage for GNU Emacs is at
236 <a href="/software/emacs/">https://www.gnu.org/software/emacs/</a>.<br>
237 To view this manual in other formats, click
238 <a href="/software/emacs/manual/eintr.html">here</a>.
244 This master menu first lists each chapter and index; then it lists
245 every node in every chapter.
248 @c Uncomment the 3 lines below, starting with @iftex, if you want the
249 @c pages of Preface to be numbered in roman numerals. Use -9 instead
250 @c of -11 for smallbook format.
252 @c >>>> Set pageno appropriately <<<<
254 @c The first page of the Preface is a roman numeral; it is the first
255 @c right handed page after the Table of Contents; hence the following
256 @c setting must be for an odd negative number.
259 @c global@pageno = -11
262 @set COUNT-WORDS count-words-example
263 @c Length of variable name chosen so that things still line up when expanded.
266 * Preface:: What to look for.
267 * List Processing:: What is Lisp?
268 * Practicing Evaluation:: Running several programs.
269 * Writing Defuns:: How to write function definitions.
270 * Buffer Walk Through:: Exploring a few buffer-related functions.
271 * More Complex:: A few, even more complex functions.
272 * Narrowing & Widening:: Restricting your and Emacs attention to
274 * car cdr & cons:: Fundamental functions in Lisp.
275 * Cutting & Storing Text:: Removing text and saving it.
276 * List Implementation:: How lists are implemented in the computer.
277 * Yanking:: Pasting stored text.
278 * Loops & Recursion:: How to repeat a process.
279 * Regexp Search:: Regular expression searches.
280 * Counting Words:: A review of repetition and regexps.
281 * Words in a defun:: Counting words in a @code{defun}.
282 * Readying a Graph:: A prototype graph printing function.
283 * Emacs Initialization:: How to write a @file{.emacs} file.
284 * Debugging:: How to run the Emacs Lisp debuggers.
285 * Conclusion:: Now you have the basics.
286 * the-the:: An appendix: how to find reduplicated words.
287 * Kill Ring:: An appendix: how the kill ring works.
288 * Full Graph:: How to create a graph with labeled axes.
289 * Free Software and Free Manuals::
290 * GNU Free Documentation License::
295 --- The Detailed Node Listing ---
299 * Why:: Why learn Emacs Lisp?
300 * On Reading this Text:: Read, gain familiarity, pick up habits....
301 * Who You Are:: For whom this is written.
303 * Note for Novices:: You can read this as a novice.
308 * Lisp Lists:: What are lists?
309 * Run a Program:: Any list in Lisp is a program ready to run.
310 * Making Errors:: Generating an error message.
311 * Names & Definitions:: Names of symbols and function definitions.
312 * Lisp Interpreter:: What the Lisp interpreter does.
313 * Evaluation:: Running a program.
314 * Variables:: Returning a value from a variable.
315 * Arguments:: Passing information to a function.
316 * set & setq:: Setting the value of a variable.
317 * Summary:: The major points.
318 * Error Message Exercises::
322 * Numbers Lists:: List have numbers, other lists, in them.
323 * Lisp Atoms:: Elemental entities.
324 * Whitespace in Lists:: Formatting lists to be readable.
325 * Typing Lists:: How GNU Emacs helps you type lists.
329 * Complications:: Variables, Special forms, Lists within.
330 * Byte Compiling:: Specially processing code for speed.
334 * How the Interpreter Acts:: Returns and Side Effects...
335 * Evaluating Inner Lists:: Lists within lists...
339 * fill-column Example::
340 * Void Function:: The error message for a symbol
342 * Void Variable:: The error message for a symbol without a value.
346 * Data types:: Types of data passed to a function.
347 * Args as Variable or List:: An argument can be the value
348 of a variable or list.
349 * Variable Number of Arguments:: Some functions may take a
350 variable number of arguments.
351 * Wrong Type of Argument:: Passing an argument of the wrong type
353 * message:: A useful function for sending messages.
355 Setting the Value of a Variable
357 * Using set:: Setting values.
358 * Using setq:: Setting a quoted value.
359 * Counting:: Using @code{setq} to count.
361 Practicing Evaluation
363 * How to Evaluate:: Typing editing commands or @kbd{C-x C-e}
365 * Buffer Names:: Buffers and files are different.
366 * Getting Buffers:: Getting a buffer itself, not merely its name.
367 * Switching Buffers:: How to change to another buffer.
368 * Buffer Size & Locations:: Where point is located and the size of
370 * Evaluation Exercise::
372 How To Write Function Definitions
374 * Primitive Functions::
375 * defun:: The @code{defun} macro.
376 * Install:: Install a function definition.
377 * Interactive:: Making a function interactive.
378 * Interactive Options:: Different options for @code{interactive}.
379 * Permanent Installation:: Installing code permanently.
380 * let:: Creating and initializing local variables.
382 * else:: If--then--else expressions.
383 * Truth & Falsehood:: What Lisp considers false and true.
384 * save-excursion:: Keeping track of point and buffer.
388 Install a Function Definition
390 * Effect of installation::
391 * Change a defun:: How to change a function definition.
393 Make a Function Interactive
395 * Interactive multiply-by-seven:: An overview.
396 * multiply-by-seven in detail:: The interactive version.
400 * Prevent confusion::
401 * Parts of let Expression::
402 * Sample let Expression::
403 * Uninitialized let Variables::
405 The @code{if} Special Form
407 * if in more detail::
408 * type-of-animal in detail:: An example of an @code{if} expression.
410 Truth and Falsehood in Emacs Lisp
412 * nil explained:: @code{nil} has two meanings.
414 @code{save-excursion}
416 * Point and mark:: A review of various locations.
417 * Template for save-excursion::
419 A Few Buffer-Related Functions
421 * Finding More:: How to find more information.
422 * simplified-beginning-of-buffer:: Shows @code{goto-char},
423 @code{point-min}, and @code{push-mark}.
424 * mark-whole-buffer:: Almost the same as @code{beginning-of-buffer}.
425 * append-to-buffer:: Uses @code{save-excursion} and
426 @code{insert-buffer-substring}.
427 * Buffer Related Review:: Review.
430 The Definition of @code{mark-whole-buffer}
432 * mark-whole-buffer overview::
433 * Body of mark-whole-buffer:: Only three lines of code.
435 The Definition of @code{append-to-buffer}
437 * append-to-buffer overview::
438 * append interactive:: A two part interactive expression.
439 * append-to-buffer body:: Incorporates a @code{let} expression.
440 * append save-excursion:: How the @code{save-excursion} works.
442 A Few More Complex Functions
444 * copy-to-buffer:: With @code{set-buffer}, @code{get-buffer-create}.
445 * insert-buffer:: Read-only, and with @code{or}.
446 * beginning-of-buffer:: Shows @code{goto-char},
447 @code{point-min}, and @code{push-mark}.
448 * Second Buffer Related Review::
449 * optional Exercise::
451 The Definition of @code{insert-buffer}
453 * insert-buffer code::
454 * insert-buffer interactive:: When you can read, but not write.
455 * insert-buffer body:: The body has an @code{or} and a @code{let}.
456 * if & or:: Using an @code{if} instead of an @code{or}.
457 * Insert or:: How the @code{or} expression works.
458 * Insert let:: Two @code{save-excursion} expressions.
459 * New insert-buffer::
461 The Interactive Expression in @code{insert-buffer}
463 * Read-only buffer:: When a buffer cannot be modified.
464 * b for interactive:: An existing buffer or else its name.
466 Complete Definition of @code{beginning-of-buffer}
468 * Optional Arguments::
469 * beginning-of-buffer opt arg:: Example with optional argument.
470 * beginning-of-buffer complete::
472 @code{beginning-of-buffer} with an Argument
474 * Disentangle beginning-of-buffer::
475 * Large buffer case::
476 * Small buffer case::
478 Narrowing and Widening
480 * Narrowing advantages:: The advantages of narrowing
481 * save-restriction:: The @code{save-restriction} special form.
482 * what-line:: The number of the line that point is on.
485 @code{car}, @code{cdr}, @code{cons}: Fundamental Functions
487 * Strange Names:: An historical aside: why the strange names?
488 * car & cdr:: Functions for extracting part of a list.
489 * cons:: Constructing a list.
490 * nthcdr:: Calling @code{cdr} repeatedly.
492 * setcar:: Changing the first element of a list.
493 * setcdr:: Changing the rest of a list.
499 * length:: How to find the length of a list.
501 Cutting and Storing Text
503 * Storing Text:: Text is stored in a list.
504 * zap-to-char:: Cutting out text up to a character.
505 * kill-region:: Cutting text out of a region.
506 * copy-region-as-kill:: A definition for copying text.
507 * Digression into C:: Minor note on C programming language macros.
508 * defvar:: How to give a variable an initial value.
509 * cons & search-fwd Review::
514 * Complete zap-to-char:: The complete implementation.
515 * zap-to-char interactive:: A three part interactive expression.
516 * zap-to-char body:: A short overview.
517 * search-forward:: How to search for a string.
518 * progn:: The @code{progn} special form.
519 * Summing up zap-to-char:: Using @code{point} and @code{search-forward}.
523 * Complete kill-region:: The function definition.
524 * condition-case:: Dealing with a problem.
527 @code{copy-region-as-kill}
529 * Complete copy-region-as-kill:: The complete function definition.
530 * copy-region-as-kill body:: The body of @code{copy-region-as-kill}.
532 The Body of @code{copy-region-as-kill}
534 * last-command & this-command::
535 * kill-append function::
536 * kill-new function::
538 Initializing a Variable with @code{defvar}
540 * See variable current value::
541 * defvar and asterisk::
543 How Lists are Implemented
546 * Symbols as Chest:: Exploring a powerful metaphor.
551 * Kill Ring Overview::
552 * kill-ring-yank-pointer:: The kill ring is a list.
553 * yank nthcdr Exercises:: The @code{kill-ring-yank-pointer} variable.
557 * while:: Causing a stretch of code to repeat.
559 * Recursion:: Causing a function to call itself.
564 * Looping with while:: Repeat so long as test returns true.
565 * Loop Example:: A @code{while} loop that uses a list.
566 * print-elements-of-list:: Uses @code{while}, @code{car}, @code{cdr}.
567 * Incrementing Loop:: A loop with an incrementing counter.
568 * Incrementing Loop Details::
569 * Decrementing Loop:: A loop with a decrementing counter.
571 Details of an Incrementing Loop
573 * Incrementing Example:: Counting pebbles in a triangle.
574 * Inc Example parts:: The parts of the function definition.
575 * Inc Example altogether:: Putting the function definition together.
577 Loop with a Decrementing Counter
579 * Decrementing Example:: More pebbles on the beach.
580 * Dec Example parts:: The parts of the function definition.
581 * Dec Example altogether:: Putting the function definition together.
583 Save your time: @code{dolist} and @code{dotimes}
590 * Building Robots:: Same model, different serial number ...
591 * Recursive Definition Parts:: Walk until you stop ...
592 * Recursion with list:: Using a list as the test whether to recurse.
593 * Recursive triangle function::
594 * Recursion with cond::
595 * Recursive Patterns:: Often used templates.
596 * No Deferment:: Don't store up work ...
597 * No deferment solution::
599 Recursion in Place of a Counter
601 * Recursive Example arg of 1 or 2::
602 * Recursive Example arg of 3 or 4::
610 Regular Expression Searches
612 * sentence-end:: The regular expression for @code{sentence-end}.
613 * re-search-forward:: Very similar to @code{search-forward}.
614 * forward-sentence:: A straightforward example of regexp search.
615 * forward-paragraph:: A somewhat complex example.
617 * re-search Exercises::
619 @code{forward-sentence}
621 * Complete forward-sentence::
622 * fwd-sentence while loops:: Two @code{while} loops.
623 * fwd-sentence re-search:: A regular expression search.
625 @code{forward-paragraph}: a Goldmine of Functions
627 * forward-paragraph in brief:: Key parts of the function definition.
628 * fwd-para let:: The @code{let*} expression.
629 * fwd-para while:: The forward motion @code{while} loop.
631 Counting: Repetition and Regexps
634 * @value{COUNT-WORDS}:: Use a regexp, but find a problem.
635 * recursive-count-words:: Start with case of no words in region.
636 * Counting Exercise::
638 The @code{@value{COUNT-WORDS}} Function
640 * Design @value{COUNT-WORDS}:: The definition using a @code{while} loop.
641 * Whitespace Bug:: The Whitespace Bug in @code{@value{COUNT-WORDS}}.
643 Counting Words in a @code{defun}
645 * Divide and Conquer::
646 * Words and Symbols:: What to count?
647 * Syntax:: What constitutes a word or symbol?
648 * count-words-in-defun:: Very like @code{@value{COUNT-WORDS}}.
649 * Several defuns:: Counting several defuns in a file.
650 * Find a File:: Do you want to look at a file?
651 * lengths-list-file:: A list of the lengths of many definitions.
652 * Several files:: Counting in definitions in different files.
653 * Several files recursively:: Recursively counting in different files.
654 * Prepare the data:: Prepare the data for display in a graph.
656 Count Words in @code{defuns} in Different Files
658 * lengths-list-many-files:: Return a list of the lengths of defuns.
659 * append:: Attach one list to another.
661 Prepare the Data for Display in a Graph
663 * Data for Display in Detail::
664 * Sorting:: Sorting lists.
665 * Files List:: Making a list of files.
666 * Counting function definitions::
670 * Columns of a graph::
671 * graph-body-print:: How to print the body of a graph.
672 * recursive-graph-body-print::
674 * Line Graph Exercise::
676 Your @file{.emacs} File
678 * Default Configuration::
679 * Site-wide Init:: You can write site-wide init files.
680 * defcustom:: Emacs will write code for you.
681 * Beginning init File:: How to write a @file{.emacs} init file.
682 * Text and Auto-fill:: Automatically wrap lines.
683 * Mail Aliases:: Use abbreviations for email addresses.
684 * Indent Tabs Mode:: Don't use tabs with @TeX{}
685 * Keybindings:: Create some personal keybindings.
686 * Keymaps:: More about key binding.
687 * Loading Files:: Load (i.e., evaluate) files automatically.
688 * Autoload:: Make functions available.
689 * Simple Extension:: Define a function; bind it to a key.
690 * X11 Colors:: Colors in X.
692 * Mode Line:: How to customize your mode line.
696 * debug:: How to use the built-in debugger.
697 * debug-on-entry:: Start debugging when you call a function.
698 * debug-on-quit:: Start debugging when you quit with @kbd{C-g}.
699 * edebug:: How to use Edebug, a source level debugger.
700 * Debugging Exercises::
702 Handling the Kill Ring
704 * What the Kill Ring Does::
706 * yank:: Paste a copy of a clipped element.
707 * yank-pop:: Insert element pointed to.
710 The @code{current-kill} Function
712 * Code for current-kill::
713 * Understanding current-kill::
715 @code{current-kill} in Outline
717 * Body of current-kill::
718 * Digression concerning error:: How to mislead humans, but not computers.
719 * Determining the Element::
721 A Graph with Labeled Axes
724 * print-graph Varlist:: @code{let} expression in @code{print-graph}.
725 * print-Y-axis:: Print a label for the vertical axis.
726 * print-X-axis:: Print a horizontal label.
727 * Print Whole Graph:: The function to print a complete graph.
729 The @code{print-Y-axis} Function
731 * print-Y-axis in Detail::
732 * Height of label:: What height for the Y axis?
733 * Compute a Remainder:: How to compute the remainder of a division.
734 * Y Axis Element:: Construct a line for the Y axis.
735 * Y-axis-column:: Generate a list of Y axis labels.
736 * print-Y-axis Penultimate:: A not quite final version.
738 The @code{print-X-axis} Function
740 * Similarities differences:: Much like @code{print-Y-axis}, but not exactly.
741 * X Axis Tic Marks:: Create tic marks for the horizontal axis.
743 Printing the Whole Graph
745 * The final version:: A few changes.
746 * Test print-graph:: Run a short test.
747 * Graphing words in defuns:: Executing the final code.
748 * lambda:: How to write an anonymous function.
749 * mapcar:: Apply a function to elements of a list.
750 * Another Bug:: Yet another bug @dots{} most insidious.
751 * Final printed graph:: The graph itself!
759 Most of the GNU Emacs integrated environment is written in the programming
760 language called Emacs Lisp. The code written in this programming
761 language is the software---the sets of instructions---that tell the
762 computer what to do when you give it commands. Emacs is designed so
763 that you can write new code in Emacs Lisp and easily install it as an
764 extension to the editor.
766 (GNU Emacs is sometimes called an ``extensible editor'', but it does
767 much more than provide editing capabilities. It is better to refer to
768 Emacs as an ``extensible computing environment''. However, that
769 phrase is quite a mouthful. It is easier to refer to Emacs simply as
770 an editor. Moreover, everything you do in Emacs---find the Mayan date
771 and phases of the moon, simplify polynomials, debug code, manage
772 files, read letters, write books---all these activities are kinds of
773 editing in the most general sense of the word.)
776 * Why:: Why learn Emacs Lisp?
777 * On Reading this Text:: Read, gain familiarity, pick up habits....
778 * Who You Are:: For whom this is written.
780 * Note for Novices:: You can read this as a novice.
786 @unnumberedsec Why Study Emacs Lisp?
789 Although Emacs Lisp is usually thought of in association only with Emacs,
790 it is a full computer programming language. You can use Emacs Lisp as
791 you would any other programming language.
793 Perhaps you want to understand programming; perhaps you want to extend
794 Emacs; or perhaps you want to become a programmer. This introduction to
795 Emacs Lisp is designed to get you started: to guide you in learning the
796 fundamentals of programming, and more importantly, to show you how you
797 can teach yourself to go further.
799 @node On Reading this Text
800 @unnumberedsec On Reading this Text
802 All through this document, you will see little sample programs you can
803 run inside of Emacs. If you read this document in Info inside of GNU
804 Emacs, you can run the programs as they appear. (This is easy to do and
805 is explained when the examples are presented.) Alternatively, you can
806 read this introduction as a printed book while sitting beside a computer
807 running Emacs. (This is what I like to do; I like printed books.) If
808 you don't have a running Emacs beside you, you can still read this book,
809 but in this case, it is best to treat it as a novel or as a travel guide
810 to a country not yet visited: interesting, but not the same as being
813 Much of this introduction is dedicated to walkthroughs or guided tours
814 of code used in GNU Emacs. These tours are designed for two purposes:
815 first, to give you familiarity with real, working code (code you use
816 every day); and, second, to give you familiarity with the way Emacs
817 works. It is interesting to see how a working environment is
820 hope that you will pick up the habit of browsing through source code.
821 You can learn from it and mine it for ideas. Having GNU Emacs is like
822 having a dragon's cave of treasures.
824 In addition to learning about Emacs as an editor and Emacs Lisp as a
825 programming language, the examples and guided tours will give you an
826 opportunity to get acquainted with Emacs as a Lisp programming
827 environment. GNU Emacs supports programming and provides tools that
828 you will want to become comfortable using, such as @kbd{M-.} (the key
829 which invokes the @code{xref-find-definitions} command). You will
830 also learn about buffers and other objects that are part of the
831 environment. Learning about these features of Emacs is like learning
832 new routes around your home town.
835 In addition, I have written several programs as extended examples.
836 Although these are examples, the programs are real. I use them.
837 Other people use them. You may use them. Beyond the fragments of
838 programs used for illustrations, there is very little in here that is
839 just for teaching purposes; what you see is used. This is a great
840 advantage of Emacs Lisp: it is easy to learn to use it for work.
843 Finally, I hope to convey some of the skills for using Emacs to
844 learn aspects of programming that you don't know. You can often use
845 Emacs to help you understand what puzzles you or to find out how to do
846 something new. This self-reliance is not only a pleasure, but an
850 @unnumberedsec For Whom This is Written
852 This text is written as an elementary introduction for people who are
853 not programmers. If you are a programmer, you may not be satisfied with
854 this primer. The reason is that you may have become expert at reading
855 reference manuals and be put off by the way this text is organized.
857 An expert programmer who reviewed this text said to me:
860 @i{I prefer to learn from reference manuals. I ``dive into'' each
861 paragraph, and ``come up for air'' between paragraphs.}
863 @i{When I get to the end of a paragraph, I assume that subject is
864 done, finished, that I know everything I need (with the
865 possible exception of the case when the next paragraph starts talking
866 about it in more detail). I expect that a well written reference manual
867 will not have a lot of redundancy, and that it will have excellent
868 pointers to the (one) place where the information I want is.}
871 This introduction is not written for this person!
873 Firstly, I try to say everything at least three times: first, to
874 introduce it; second, to show it in context; and third, to show it in a
875 different context, or to review it.
877 Secondly, I hardly ever put all the information about a subject in one
878 place, much less in one paragraph. To my way of thinking, that imposes
879 too heavy a burden on the reader. Instead I try to explain only what
880 you need to know at the time. (Sometimes I include a little extra
881 information so you won't be surprised later when the additional
882 information is formally introduced.)
884 When you read this text, you are not expected to learn everything the
885 first time. Frequently, you need make only a nodding
886 acquaintance with some of the items mentioned. My hope is that I have
887 structured the text and given you enough hints that you will be alert to
888 what is important, and concentrate on it.
890 You will need to dive into some paragraphs; there is no other way
891 to read them. But I have tried to keep down the number of such
892 paragraphs. This book is intended as an approachable hill, rather than
893 as a daunting mountain.
895 This introduction to @cite{Programming in Emacs Lisp} has a companion
898 @cite{The GNU Emacs Lisp Reference Manual}.
901 @ref{Top, , The GNU Emacs Lisp Reference Manual, elisp, The GNU
902 Emacs Lisp Reference Manual}.
904 The reference manual has more detail than this introduction. In the
905 reference manual, all the information about one topic is concentrated
906 in one place. You should turn to it if you are like the programmer
907 quoted above. And, of course, after you have read this
908 @cite{Introduction}, you will find the @cite{Reference Manual} useful
909 when you are writing your own programs.
912 @unnumberedsec Lisp History
915 Lisp was first developed in the late 1950s at the Massachusetts
916 Institute of Technology for research in artificial intelligence. The
917 great power of the Lisp language makes it superior for other purposes as
918 well, such as writing editor commands and integrated environments.
922 GNU Emacs Lisp is largely inspired by Maclisp, which was written at MIT
923 in the 1960s. It is somewhat inspired by Common Lisp, which became a
924 standard in the 1980s. However, Emacs Lisp is much simpler than Common
925 Lisp. (The standard Emacs distribution contains an optional extensions
926 file, @file{cl.el}, that adds many Common Lisp features to Emacs Lisp.)
928 @node Note for Novices
929 @unnumberedsec A Note for Novices
931 If you don't know GNU Emacs, you can still read this document
932 profitably. However, I recommend you learn Emacs, if only to learn to
933 move around your computer screen. You can teach yourself how to use
934 Emacs with the built-in tutorial. To use it, type @kbd{C-h t}. (This
935 means you press and release the @key{CTRL} key and the @kbd{h} at the
936 same time, and then press and release @kbd{t}.)
938 Also, I often refer to one of Emacs's standard commands by listing the
939 keys which you press to invoke the command and then giving the name of
940 the command in parentheses, like this: @kbd{M-C-\}
941 (@code{indent-region}). What this means is that the
942 @code{indent-region} command is customarily invoked by typing
943 @kbd{M-C-\}. (You can, if you wish, change the keys that are typed to
944 invoke the command; this is called @dfn{rebinding}. @xref{Keymaps, ,
945 Keymaps}.) The abbreviation @kbd{M-C-\} means that you type your
946 @key{META} key, @key{CTRL} key and @kbd{\} key all at the same time.
947 (On many modern keyboards the @key{META} key is labeled
949 Sometimes a combination like this is called a keychord, since it is
950 similar to the way you play a chord on a piano. If your keyboard does
951 not have a @key{META} key, the @key{ESC} key prefix is used in place
952 of it. In this case, @kbd{M-C-\} means that you press and release your
953 @key{ESC} key and then type the @key{CTRL} key and the @kbd{\} key at
954 the same time. But usually @kbd{M-C-\} means press the @key{CTRL} key
955 along with the key that is labeled @key{ALT} and, at the same time,
956 press the @kbd{\} key.
958 In addition to typing a lone keychord, you can prefix what you type
959 with @kbd{C-u}, which is called the @dfn{universal argument}. The
960 @kbd{C-u} keychord passes an argument to the subsequent command.
961 Thus, to indent a region of plain text by 6 spaces, mark the region,
962 and then type @w{@kbd{C-u 6 M-C-\}}. (If you do not specify a number,
963 Emacs either passes the number 4 to the command or otherwise runs the
964 command differently than it would otherwise.) @xref{Arguments, ,
965 Numeric Arguments, emacs, The GNU Emacs Manual}.
967 If you are reading this in Info using GNU Emacs, you can read through
968 this whole document just by pressing the space bar, @key{SPC}.
969 (To learn about Info, type @kbd{C-h i} and then select Info.)
971 A note on terminology: when I use the word Lisp alone, I often am
972 referring to the various dialects of Lisp in general, but when I speak
973 of Emacs Lisp, I am referring to GNU Emacs Lisp in particular.
976 @unnumberedsec Thank You
978 My thanks to all who helped me with this book. My especial thanks to
979 @r{Jim Blandy}, @r{Noah Friedman}, @w{Jim Kingdon}, @r{Roland
980 McGrath}, @w{Frank Ritter}, @w{Randy Smith}, @w{Richard M.
981 Stallman}, and @w{Melissa Weisshaus}. My thanks also go to both
982 @w{Philip Johnson} and @w{David Stampe} for their patient
983 encouragement. My mistakes are my own.
995 @c ================ Beginning of main text ================
997 @c Start main text on right-hand (verso) page
1000 \par\vfill\supereject
1003 \par\vfill\supereject
1005 \par\vfill\supereject
1007 \par\vfill\supereject
1011 @c Note: this resetting of the page number back to 1 causes TeX to gripe
1012 @c about already having seen page numbers 1-4 before (in the preface):
1013 @c pdfTeX warning (ext4): destination with the same identifier (name{1})
1014 @c has been already used, duplicate ignored
1015 @c I guess that is harmless (what happens if a later part of the text
1016 @c makes a link to something in the first 4 pages though?).
1017 @c E.g., note that the Emacs manual has a preface, but does not bother
1018 @c resetting the page numbers back to 1 after that.
1019 @c Alternatively, uncomment the 3 lines above (search for ``pageno'')
1020 @c to have the preface numbered in roman numerals.
1023 @evenheading @thispage @| @| @thischapter
1024 @oddheading @thissection @| @| @thispage
1028 @node List Processing
1029 @chapter List Processing
1031 To the untutored eye, Lisp is a strange programming language. In Lisp
1032 code there are parentheses everywhere. Some people even claim that
1033 the name stands for ``Lots of Isolated Silly Parentheses''. But the
1034 claim is unwarranted. Lisp stands for LISt Processing, and the
1035 programming language handles @emph{lists} (and lists of lists) by
1036 putting them between parentheses. The parentheses mark the boundaries
1037 of the list. Sometimes a list is preceded by an apostrophe @samp{'},
1038 called a @dfn{single-quote} in Lisp.@footnote{A single-quote is an
1039 abbreviation for the special form @code{quote}; you need not think
1040 about special forms now.
1042 @xref{Complications}.
1045 @xref{Lisp Interpreter}.
1047 } Lists are the basis
1051 * Lisp Lists:: What are lists?
1052 * Run a Program:: Any list in Lisp is a program ready to run.
1053 * Making Errors:: Generating an error message.
1054 * Names & Definitions:: Names of symbols and function definitions.
1055 * Lisp Interpreter:: What the Lisp interpreter does.
1056 * Evaluation:: Running a program.
1057 * Variables:: Returning a value from a variable.
1058 * Arguments:: Passing information to a function.
1059 * set & setq:: Setting the value of a variable.
1060 * Summary:: The major points.
1061 * Error Message Exercises::
1068 In Lisp, a list looks like this: @code{'(rose violet daisy buttercup)}.
1069 This list is preceded by a single apostrophe. It could just as well be
1070 written as follows, which looks more like the kind of list you are likely
1071 to be familiar with:
1083 The elements of this list are the names of the four different flowers,
1084 separated from each other by whitespace and surrounded by parentheses,
1085 like flowers in a field with a stone wall around them.
1086 @cindex Flowers in a field
1089 * Numbers Lists:: List have numbers, other lists, in them.
1090 * Lisp Atoms:: Elemental entities.
1091 * Whitespace in Lists:: Formatting lists to be readable.
1092 * Typing Lists:: How GNU Emacs helps you type lists.
1097 @unnumberedsubsec Numbers, Lists inside of Lists
1100 Lists can also have numbers in them, as in this list: @code{(+ 2 2)}.
1101 This list has a plus-sign, @samp{+}, followed by two @samp{2}s, each
1102 separated by whitespace.
1104 In Lisp, both data and programs are represented the same way; that is,
1105 they are both lists of words, numbers, or other lists, separated by
1106 whitespace and surrounded by parentheses. (Since a program looks like
1107 data, one program may easily serve as data for another; this is a very
1108 powerful feature of Lisp.) (Incidentally, these two parenthetical
1109 remarks are @emph{not} Lisp lists, because they contain @samp{;} and
1110 @samp{.} as punctuation marks.)
1113 Here is another list, this time with a list inside of it:
1116 '(this list has (a list inside of it))
1119 The components of this list are the words @samp{this}, @samp{list},
1120 @samp{has}, and the list @samp{(a list inside of it)}. The interior
1121 list is made up of the words @samp{a}, @samp{list}, @samp{inside},
1122 @samp{of}, @samp{it}.
1125 @subsection Lisp Atoms
1128 In Lisp, what we have been calling words are called @dfn{atoms}. This
1129 term comes from the historical meaning of the word atom, which means
1130 ``indivisible''. As far as Lisp is concerned, the words we have been
1131 using in the lists cannot be divided into any smaller parts and still
1132 mean the same thing as part of a program; likewise with numbers and
1133 single character symbols like @samp{+}. On the other hand, unlike an
1134 ancient atom, a list can be split into parts. (@xref{car cdr & cons,
1135 , @code{car} @code{cdr} & @code{cons} Fundamental Functions}.)
1137 In a list, atoms are separated from each other by whitespace. They can be
1138 right next to a parenthesis.
1140 @cindex @samp{empty list} defined
1141 Technically speaking, a list in Lisp consists of parentheses surrounding
1142 atoms separated by whitespace or surrounding other lists or surrounding
1143 both atoms and other lists. A list can have just one atom in it or
1144 have nothing in it at all. A list with nothing in it looks like this:
1145 @code{()}, and is called the @dfn{empty list}. Unlike anything else, an
1146 empty list is considered both an atom and a list at the same time.
1148 @cindex Symbolic expressions, introduced
1149 @cindex @samp{expression} defined
1150 @cindex @samp{form} defined
1151 The printed representation of both atoms and lists are called
1152 @dfn{symbolic expressions} or, more concisely, @dfn{s-expressions}.
1153 The word @dfn{expression} by itself can refer to either the printed
1154 representation, or to the atom or list as it is held internally in the
1155 computer. Often, people use the term @dfn{expression}
1156 indiscriminately. (Also, in many texts, the word @dfn{form} is used
1157 as a synonym for expression.)
1159 Incidentally, the atoms that make up our universe were named such when
1160 they were thought to be indivisible; but it has been found that physical
1161 atoms are not indivisible. Parts can split off an atom or it can
1162 fission into two parts of roughly equal size. Physical atoms were named
1163 prematurely, before their truer nature was found. In Lisp, certain
1164 kinds of atom, such as an array, can be separated into parts; but the
1165 mechanism for doing this is different from the mechanism for splitting a
1166 list. As far as list operations are concerned, the atoms of a list are
1169 As in English, the meanings of the component letters of a Lisp atom
1170 are different from the meaning the letters make as a word. For
1171 example, the word for the South American sloth, the @samp{ai}, is
1172 completely different from the two words, @samp{a}, and @samp{i}.
1174 There are many kinds of atom in nature but only a few in Lisp: for
1175 example, @dfn{numbers}, such as 37, 511, or 1729, and @dfn{symbols}, such
1176 as @samp{+}, @samp{foo}, or @samp{forward-line}. The words we have
1177 listed in the examples above are all symbols. In everyday Lisp
1178 conversation, the word ``atom'' is not often used, because programmers
1179 usually try to be more specific about what kind of atom they are dealing
1180 with. Lisp programming is mostly about symbols (and sometimes numbers)
1181 within lists. (Incidentally, the preceding three word parenthetical
1182 remark is a proper list in Lisp, since it consists of atoms, which in
1183 this case are symbols, separated by whitespace and enclosed by
1184 parentheses, without any non-Lisp punctuation.)
1187 Text between double quotation marks---even sentences or
1188 paragraphs---is also an atom. Here is an example:
1189 @cindex Text between double quotation marks
1192 '(this list includes "text between quotation marks.")
1195 @cindex @samp{string} defined
1197 In Lisp, all of the quoted text including the punctuation mark and the
1198 blank spaces is a single atom. This kind of atom is called a
1199 @dfn{string} (for ``string of characters'') and is the sort of thing that
1200 is used for messages that a computer can print for a human to read.
1201 Strings are a different kind of atom than numbers or symbols and are
1204 @node Whitespace in Lists
1205 @subsection Whitespace in Lists
1206 @cindex Whitespace in lists
1209 The amount of whitespace in a list does not matter. From the point of view
1210 of the Lisp language,
1221 is exactly the same as this:
1224 '(this list looks like this)
1227 Both examples show what to Lisp is the same list, the list made up of
1228 the symbols @samp{this}, @samp{list}, @samp{looks}, @samp{like}, and
1229 @samp{this} in that order.
1231 Extra whitespace and newlines are designed to make a list more readable
1232 by humans. When Lisp reads the expression, it gets rid of all the extra
1233 whitespace (but it needs to have at least one space between atoms in
1234 order to tell them apart.)
1236 Odd as it seems, the examples we have seen cover almost all of what Lisp
1237 lists look like! Every other list in Lisp looks more or less like one
1238 of these examples, except that the list may be longer and more complex.
1239 In brief, a list is between parentheses, a string is between quotation
1240 marks, a symbol looks like a word, and a number looks like a number.
1241 (For certain situations, square brackets, dots and a few other special
1242 characters may be used; however, we will go quite far without them.)
1245 @subsection GNU Emacs Helps You Type Lists
1246 @cindex Help typing lists
1247 @cindex Formatting help
1249 When you type a Lisp expression in GNU Emacs using either Lisp
1250 Interaction mode or Emacs Lisp mode, you have available to you several
1251 commands to format the Lisp expression so it is easy to read. For
1252 example, pressing the @key{TAB} key automatically indents the line the
1253 cursor is on by the right amount. A command to properly indent the
1254 code in a region is customarily bound to @kbd{M-C-\}. Indentation is
1255 designed so that you can see which elements of a list belong to which
1256 list---elements of a sub-list are indented more than the elements of
1259 In addition, when you type a closing parenthesis, Emacs momentarily
1260 jumps the cursor back to the matching opening parenthesis, so you can
1261 see which one it is. This is very useful, since every list you type
1262 in Lisp must have its closing parenthesis match its opening
1263 parenthesis. (@xref{Major Modes, , Major Modes, emacs, The GNU Emacs
1264 Manual}, for more information about Emacs's modes.)
1267 @section Run a Program
1268 @cindex Run a program
1269 @cindex Program, running one
1271 @cindex @samp{evaluate} defined
1272 A list in Lisp---any list---is a program ready to run. If you run it
1273 (for which the Lisp jargon is @dfn{evaluate}), the computer will do one
1274 of three things: do nothing except return to you the list itself; send
1275 you an error message; or, treat the first symbol in the list as a
1276 command to do something. (Usually, of course, it is the last of these
1277 three things that you really want!)
1279 @c use code for the single apostrophe, not samp.
1281 @cindex @code{'} for quoting
1282 @cindex quoting using apostrophe
1283 @cindex apostrophe for quoting
1284 The single apostrophe, @code{'}, that I put in front of some of the
1285 example lists in preceding sections is called a @dfn{quote}; when it
1286 precedes a list, it tells Lisp to do nothing with the list, other than
1287 take it as it is written. But if there is no quote preceding a list,
1288 the first item of the list is special: it is a command for the computer
1289 to obey. (In Lisp, these commands are called @emph{functions}.) The list
1290 @code{(+ 2 2)} shown above did not have a quote in front of it, so Lisp
1291 understands that the @code{+} is an instruction to do something with the
1292 rest of the list: add the numbers that follow.
1295 If you are reading this inside of GNU Emacs in Info, here is how you can
1296 evaluate such a list: place your cursor immediately after the right
1297 hand parenthesis of the following list and then type @kbd{C-x C-e}:
1303 @c use code for the number four, not samp.
1305 You will see the number @code{4} appear in the echo area. (What
1306 you have just done is evaluate the list. The echo area
1307 is the line at the bottom of the screen that displays or echoes
1308 text.) Now try the same thing with a quoted list: place the cursor
1309 right after the following list and type @kbd{C-x C-e}:
1312 '(this is a quoted list)
1316 You will see @code{(this is a quoted list)} appear in the echo area.
1318 @cindex Lisp interpreter, explained
1319 @cindex Interpreter, Lisp, explained
1320 In both cases, what you are doing is giving a command to the program
1321 inside of GNU Emacs called the @dfn{Lisp interpreter}---giving the
1322 interpreter a command to evaluate the expression. The name of the Lisp
1323 interpreter comes from the word for the task done by a human who comes
1324 up with the meaning of an expression---who interprets it.
1326 You can also evaluate an atom that is not part of a list---one that is
1327 not surrounded by parentheses; again, the Lisp interpreter translates
1328 from the humanly readable expression to the language of the computer.
1329 But before discussing this (@pxref{Variables}), we will discuss what the
1330 Lisp interpreter does when you make an error.
1333 @section Generate an Error Message
1334 @cindex Generate an error message
1335 @cindex Error message generation
1337 Partly so you won't worry if you do it accidentally, we will now give
1338 a command to the Lisp interpreter that generates an error message.
1339 This is a harmless activity; and indeed, we will often try to generate
1340 error messages intentionally. Once you understand the jargon, error
1341 messages can be informative. Instead of being called ``error''
1342 messages, they should be called ``help'' messages. They are like
1343 signposts to a traveler in a strange country; deciphering them can be
1344 hard, but once understood, they can point the way.
1346 The error message is generated by a built-in GNU Emacs debugger. We
1347 will enter the debugger. You get out of the debugger by typing @code{q}.
1349 What we will do is evaluate a list that is not quoted and does not
1350 have a meaningful command as its first element. Here is a list almost
1351 exactly the same as the one we just used, but without the single-quote
1352 in front of it. Position the cursor right after it and type @kbd{C-x
1356 (this is an unquoted list)
1361 What you see depends on which version of Emacs you are running. GNU
1362 Emacs version 22 provides more information than version 20 and before.
1363 First, the more recent result of generating an error; then the
1364 earlier, version 20 result.
1368 In GNU Emacs version 22, a @file{*Backtrace*} window will open up and
1369 you will see the following in it:
1372 A @file{*Backtrace*} window will open up and you should see the
1377 ---------- Buffer: *Backtrace* ----------
1378 Debugger entered--Lisp error: (void-function this)
1379 (this is an unquoted list)
1380 eval((this is an unquoted list))
1381 eval-last-sexp-1(nil)
1383 call-interactively(eval-last-sexp)
1384 ---------- Buffer: *Backtrace* ----------
1390 Your cursor will be in this window (you may have to wait a few seconds
1391 before it becomes visible). To quit the debugger and make the
1392 debugger window go away, type:
1399 Please type @kbd{q} right now, so you become confident that you can
1400 get out of the debugger. Then, type @kbd{C-x C-e} again to re-enter
1403 @cindex @samp{function} defined
1404 Based on what we already know, we can almost read this error message.
1406 You read the @file{*Backtrace*} buffer from the bottom up; it tells
1407 you what Emacs did. When you typed @kbd{C-x C-e}, you made an
1408 interactive call to the command @code{eval-last-sexp}. @code{eval} is
1409 an abbreviation for ``evaluate'' and @code{sexp} is an abbreviation for
1410 ``symbolic expression''. The command means ``evaluate last symbolic
1411 expression'', which is the expression just before your cursor.
1413 Each line above tells you what the Lisp interpreter evaluated next.
1414 The most recent action is at the top. The buffer is called the
1415 @file{*Backtrace*} buffer because it enables you to track Emacs
1419 At the top of the @file{*Backtrace*} buffer, you see the line:
1422 Debugger entered--Lisp error: (void-function this)
1426 The Lisp interpreter tried to evaluate the first atom of the list, the
1427 word @samp{this}. It is this action that generated the error message
1428 @samp{void-function this}.
1430 The message contains the words @samp{void-function} and @samp{this}.
1432 @cindex @samp{function} defined
1433 The word @samp{function} was mentioned once before. It is a very
1434 important word. For our purposes, we can define it by saying that a
1435 @dfn{function} is a set of instructions to the computer that tell the
1436 computer to do something.
1438 Now we can begin to understand the error message: @samp{void-function
1439 this}. The function (that is, the word @samp{this}) does not have a
1440 definition of any set of instructions for the computer to carry out.
1442 The slightly odd word, @samp{void-function}, is designed to cover the
1443 way Emacs Lisp is implemented, which is that when a symbol does not
1444 have a function definition attached to it, the place that should
1445 contain the instructions is void.
1447 On the other hand, since we were able to add 2 plus 2 successfully, by
1448 evaluating @code{(+ 2 2)}, we can infer that the symbol @code{+} must
1449 have a set of instructions for the computer to obey and those
1450 instructions must be to add the numbers that follow the @code{+}.
1452 It is possible to prevent Emacs entering the debugger in cases like
1453 this. We do not explain how to do that here, but we will mention what
1454 the result looks like, because you may encounter a similar situation
1455 if there is a bug in some Emacs code that you are using. In such
1456 cases, you will see only one line of error message; it will appear in
1457 the echo area and look like this:
1460 Symbol's function definition is void:@: this
1465 (Also, your terminal may beep at you---some do, some don't; and others
1466 blink. This is just a device to get your attention.)
1468 The message goes away as soon as you type a key, even just to
1471 We know the meaning of the word @samp{Symbol}. It refers to the first
1472 atom of the list, the word @samp{this}. The word @samp{function}
1473 refers to the instructions that tell the computer what to do.
1474 (Technically, the symbol tells the computer where to find the
1475 instructions, but this is a complication we can ignore for the
1478 The error message can be understood: @samp{Symbol's function
1479 definition is void:@: this}. The symbol (that is, the word
1480 @samp{this}) lacks instructions for the computer to carry out.
1482 @node Names & Definitions
1483 @section Symbol Names and Function Definitions
1484 @cindex Symbol names
1486 We can articulate another characteristic of Lisp based on what we have
1487 discussed so far---an important characteristic: a symbol, like
1488 @code{+}, is not itself the set of instructions for the computer to
1489 carry out. Instead, the symbol is used, perhaps temporarily, as a way
1490 of locating the definition or set of instructions. What we see is the
1491 name through which the instructions can be found. Names of people
1492 work the same way. I can be referred to as @samp{Bob}; however, I am
1493 not the letters @samp{B}, @samp{o}, @samp{b} but am, or was, the
1494 consciousness consistently associated with a particular life-form.
1495 The name is not me, but it can be used to refer to me.
1497 In Lisp, one set of instructions can be attached to several names.
1498 For example, the computer instructions for adding numbers can be
1499 linked to the symbol @code{plus} as well as to the symbol @code{+}
1500 (and are in some dialects of Lisp). Among humans, I can be referred
1501 to as @samp{Robert} as well as @samp{Bob} and by other words as well.
1503 On the other hand, a symbol can have only one function definition
1504 attached to it at a time. Otherwise, the computer would be confused as
1505 to which definition to use. If this were the case among people, only
1506 one person in the world could be named @samp{Bob}. However, the function
1507 definition to which the name refers can be changed readily.
1508 (@xref{Install, , Install a Function Definition}.)
1510 Since Emacs Lisp is large, it is customary to name symbols in a way
1511 that identifies the part of Emacs to which the function belongs.
1512 Thus, all the names for functions that deal with Texinfo start with
1513 @samp{texinfo-} and those for functions that deal with reading mail
1514 start with @samp{rmail-}.
1516 @node Lisp Interpreter
1517 @section The Lisp Interpreter
1518 @cindex Lisp interpreter, what it does
1519 @cindex Interpreter, what it does
1521 Based on what we have seen, we can now start to figure out what the
1522 Lisp interpreter does when we command it to evaluate a list.
1523 First, it looks to see whether there is a quote before the list; if
1524 there is, the interpreter just gives us the list. On the other
1525 hand, if there is no quote, the interpreter looks at the first element
1526 in the list and sees whether it has a function definition. If it does,
1527 the interpreter carries out the instructions in the function definition.
1528 Otherwise, the interpreter prints an error message.
1530 This is how Lisp works. Simple. There are added complications which we
1531 will get to in a minute, but these are the fundamentals. Of course, to
1532 write Lisp programs, you need to know how to write function definitions
1533 and attach them to names, and how to do this without confusing either
1534 yourself or the computer.
1537 * Complications:: Variables, Special forms, Lists within.
1538 * Byte Compiling:: Specially processing code for speed.
1543 @unnumberedsubsec Complications
1546 Now, for the first complication. In addition to lists, the Lisp
1547 interpreter can evaluate a symbol that is not quoted and does not have
1548 parentheses around it. The Lisp interpreter will attempt to determine
1549 the symbol's value as a @dfn{variable}. This situation is described
1550 in the section on variables. (@xref{Variables}.)
1552 @cindex Special form
1553 The second complication occurs because some functions are unusual and
1554 do not work in the usual manner. Those that don't are called
1555 @dfn{special forms}. They are used for special jobs, like defining a
1556 function, and there are not many of them. In the next few chapters,
1557 you will be introduced to several of the more important special forms.
1559 As well as special forms, there are also @dfn{macros}. A macro
1560 is a construct defined in Lisp, which differs from a function in that it
1561 translates a Lisp expression into another expression that is to be
1562 evaluated in place of the original expression. (@xref{Lisp macro}.)
1564 For the purposes of this introduction, you do not need to worry too much
1565 about whether something is a special form, macro, or ordinary function.
1566 For example, @code{if} is a special form (@pxref{if}), but @code{when}
1567 is a macro (@pxref{Lisp macro}). In earlier versions of Emacs,
1568 @code{defun} was a special form, but now it is a macro (@pxref{defun}).
1569 It still behaves in the same way.
1571 The final complication is this: if the function that the
1572 Lisp interpreter is looking at is not a special form, and if it is part
1573 of a list, the Lisp interpreter looks to see whether the list has a list
1574 inside of it. If there is an inner list, the Lisp interpreter first
1575 figures out what it should do with the inside list, and then it works on
1576 the outside list. If there is yet another list embedded inside the
1577 inner list, it works on that one first, and so on. It always works on
1578 the innermost list first. The interpreter works on the innermost list
1579 first, to evaluate the result of that list. The result may be
1580 used by the enclosing expression.
1582 Otherwise, the interpreter works left to right, from one expression to
1585 @node Byte Compiling
1586 @subsection Byte Compiling
1587 @cindex Byte compiling
1589 One other aspect of interpreting: the Lisp interpreter is able to
1590 interpret two kinds of entity: humanly readable code, on which we will
1591 focus exclusively, and specially processed code, called @dfn{byte
1592 compiled} code, which is not humanly readable. Byte compiled code
1593 runs faster than humanly readable code.
1595 You can transform humanly readable code into byte compiled code by
1596 running one of the compile commands such as @code{byte-compile-file}.
1597 Byte compiled code is usually stored in a file that ends with a
1598 @file{.elc} extension rather than a @file{.el} extension. You will
1599 see both kinds of file in the @file{emacs/lisp} directory; the files
1600 to read are those with @file{.el} extensions.
1602 As a practical matter, for most things you might do to customize or
1603 extend Emacs, you do not need to byte compile; and I will not discuss
1604 the topic here. @xref{Byte Compilation, , Byte Compilation, elisp,
1605 The GNU Emacs Lisp Reference Manual}, for a full description of byte
1612 When the Lisp interpreter works on an expression, the term for the
1613 activity is called @dfn{evaluation}. We say that the interpreter
1614 ``evaluates the expression''. I've used this term several times before.
1615 The word comes from its use in everyday language, ``to ascertain the
1616 value or amount of; to appraise'', according to @cite{Webster's New
1617 Collegiate Dictionary}.
1620 * How the Interpreter Acts:: Returns and Side Effects...
1621 * Evaluating Inner Lists:: Lists within lists...
1625 @node How the Interpreter Acts
1626 @unnumberedsubsec How the Lisp Interpreter Acts
1629 @cindex @samp{returned value} explained
1630 After evaluating an expression, the Lisp interpreter will most likely
1631 @dfn{return} the value that the computer produces by carrying out the
1632 instructions it found in the function definition, or perhaps it will
1633 give up on that function and produce an error message. (The interpreter
1634 may also find itself tossed, so to speak, to a different function or it
1635 may attempt to repeat continually what it is doing for ever and ever in
1636 an infinite loop. These actions are less common; and
1637 we can ignore them.) Most frequently, the interpreter returns a value.
1639 @cindex @samp{side effect} defined
1640 At the same time the interpreter returns a value, it may do something
1641 else as well, such as move a cursor or copy a file; this other kind of
1642 action is called a @dfn{side effect}. Actions that we humans think are
1643 important, such as printing results, are often side effects to the
1644 Lisp interpreter. It is fairly easy to learn to use side effects.
1646 In summary, evaluating a symbolic expression most commonly causes the
1647 Lisp interpreter to return a value and perhaps carry out a side effect;
1648 or else produce an error.
1650 @node Evaluating Inner Lists
1651 @subsection Evaluating Inner Lists
1652 @cindex Inner list evaluation
1653 @cindex Evaluating inner lists
1655 If evaluation applies to a list that is inside another list, the outer
1656 list may use the value returned by the first evaluation as information
1657 when the outer list is evaluated. This explains why inner expressions
1658 are evaluated first: the values they return are used by the outer
1662 We can investigate this process by evaluating another addition example.
1663 Place your cursor after the following expression and type @kbd{C-x C-e}:
1670 The number 8 will appear in the echo area.
1672 What happens is that the Lisp interpreter first evaluates the inner
1673 expression, @code{(+ 3 3)}, for which the value 6 is returned; then it
1674 evaluates the outer expression as if it were written @code{(+ 2 6)}, which
1675 returns the value 8. Since there are no more enclosing expressions to
1676 evaluate, the interpreter prints that value in the echo area.
1678 Now it is easy to understand the name of the command invoked by the
1679 keystrokes @kbd{C-x C-e}: the name is @code{eval-last-sexp}. The
1680 letters @code{sexp} are an abbreviation for ``symbolic expression'', and
1681 @code{eval} is an abbreviation for ``evaluate''. The command
1682 evaluates the last symbolic expression.
1684 As an experiment, you can try evaluating the expression by putting the
1685 cursor at the beginning of the next line immediately following the
1686 expression, or inside the expression.
1689 Here is another copy of the expression:
1696 If you place the cursor at the beginning of the blank line that
1697 immediately follows the expression and type @kbd{C-x C-e}, you will
1698 still get the value 8 printed in the echo area. Now try putting the
1699 cursor inside the expression. If you put it right after the next to
1700 last parenthesis (so it appears to sit on top of the last parenthesis),
1701 you will get a 6 printed in the echo area! This is because the command
1702 evaluates the expression @code{(+ 3 3)}.
1704 Now put the cursor immediately after a number. Type @kbd{C-x C-e} and
1705 you will get the number itself. In Lisp, if you evaluate a number, you
1706 get the number itself---this is how numbers differ from symbols. If you
1707 evaluate a list starting with a symbol like @code{+}, you will get a
1708 value returned that is the result of the computer carrying out the
1709 instructions in the function definition attached to that name. If a
1710 symbol by itself is evaluated, something different happens, as we will
1711 see in the next section.
1717 In Emacs Lisp, a symbol can have a value attached to it just as it can
1718 have a function definition attached to it. The two are different.
1719 The function definition is a set of instructions that a computer will
1720 obey. A value, on the other hand, is something, such as number or a
1721 name, that can vary (which is why such a symbol is called a variable).
1722 The value of a symbol can be any expression in Lisp, such as a symbol,
1723 number, list, or string. A symbol that has a value is often called a
1726 A symbol can have both a function definition and a value attached to
1727 it at the same time. Or it can have just one or the other.
1728 The two are separate. This is somewhat similar
1729 to the way the name Cambridge can refer to the city in Massachusetts
1730 and have some information attached to the name as well, such as
1731 ``great programming center''.
1734 (Incidentally, in Emacs Lisp, a symbol can have two
1735 other things attached to it, too: a property list and a documentation
1736 string; these are discussed later.)
1739 Another way to think about this is to imagine a symbol as being a chest
1740 of drawers. The function definition is put in one drawer, the value in
1741 another, and so on. What is put in the drawer holding the value can be
1742 changed without affecting the contents of the drawer holding the
1743 function definition, and vice versa.
1746 * fill-column Example::
1747 * Void Function:: The error message for a symbol
1749 * Void Variable:: The error message for a symbol without a value.
1753 @node fill-column Example
1754 @unnumberedsubsec @code{fill-column}, an Example Variable
1757 @findex fill-column@r{, an example variable}
1758 @cindex Example variable, @code{fill-column}
1759 @cindex Variable, example of, @code{fill-column}
1760 The variable @code{fill-column} illustrates a symbol with a value
1761 attached to it: in every GNU Emacs buffer, this symbol is set to some
1762 value, usually 72 or 70, but sometimes to some other value. To find the
1763 value of this symbol, evaluate it by itself. If you are reading this in
1764 Info inside of GNU Emacs, you can do this by putting the cursor after
1765 the symbol and typing @kbd{C-x C-e}:
1772 After I typed @kbd{C-x C-e}, Emacs printed the number 72 in my echo
1773 area. This is the value for which @code{fill-column} is set for me as I
1774 write this. It may be different for you in your Info buffer. Notice
1775 that the value returned as a variable is printed in exactly the same way
1776 as the value returned by a function carrying out its instructions. From
1777 the point of view of the Lisp interpreter, a value returned is a value
1778 returned. What kind of expression it came from ceases to matter once
1781 A symbol can have any value attached to it or, to use the jargon, we can
1782 @dfn{bind} the variable to a value: to a number, such as 72; to a
1783 string, @code{"such as this"}; to a list, such as @code{(spruce pine
1784 oak)}; we can even bind a variable to a function definition.
1786 A symbol can be bound to a value in several ways. @xref{set & setq, ,
1787 Setting the Value of a Variable}, for information about one way to do
1791 @subsection Error Message for a Symbol Without a Function
1792 @cindex Symbol without function error
1793 @cindex Error for symbol without function
1795 When we evaluated @code{fill-column} to find its value as a variable,
1796 we did not place parentheses around the word. This is because we did
1797 not intend to use it as a function name.
1799 If @code{fill-column} were the first or only element of a list, the
1800 Lisp interpreter would attempt to find the function definition
1801 attached to it. But @code{fill-column} has no function definition.
1802 Try evaluating this:
1810 You will create a @file{*Backtrace*} buffer that says:
1814 ---------- Buffer: *Backtrace* ----------
1815 Debugger entered--Lisp error: (void-function fill-column)
1818 eval-last-sexp-1(nil)
1820 call-interactively(eval-last-sexp)
1821 ---------- Buffer: *Backtrace* ----------
1826 (Remember, to quit the debugger and make the debugger window go away,
1827 type @kbd{q} in the @file{*Backtrace*} buffer.)
1831 In GNU Emacs 20 and before, you will produce an error message that says:
1834 Symbol's function definition is void:@: fill-column
1838 (The message will go away as soon as you move the cursor or type
1843 @subsection Error Message for a Symbol Without a Value
1844 @cindex Symbol without value error
1845 @cindex Error for symbol without value
1847 If you attempt to evaluate a symbol that does not have a value bound to
1848 it, you will receive an error message. You can see this by
1849 experimenting with our 2 plus 2 addition. In the following expression,
1850 put your cursor right after the @code{+}, before the first number 2,
1859 In GNU Emacs 22, you will create a @file{*Backtrace*} buffer that
1864 ---------- Buffer: *Backtrace* ----------
1865 Debugger entered--Lisp error: (void-variable +)
1867 eval-last-sexp-1(nil)
1869 call-interactively(eval-last-sexp)
1870 ---------- Buffer: *Backtrace* ----------
1875 (Again, you can quit the debugger by
1876 typing @kbd{q} in the @file{*Backtrace*} buffer.)
1878 This backtrace is different from the very first error message we saw,
1879 which said, @samp{Debugger entered--Lisp error: (void-function this)}.
1880 In this case, the function does not have a value as a variable; while
1881 in the other error message, the function (the word @samp{this}) did not
1884 In this experiment with the @code{+}, what we did was cause the Lisp
1885 interpreter to evaluate the @code{+} and look for the value of the
1886 variable instead of the function definition. We did this by placing the
1887 cursor right after the symbol rather than after the parenthesis of the
1888 enclosing list as we did before. As a consequence, the Lisp interpreter
1889 evaluated the preceding s-expression, which in this case was
1892 Since @code{+} does not have a value bound to it, just the function
1893 definition, the error message reported that the symbol's value as a
1898 In GNU Emacs version 20 and before, your error message will say:
1901 Symbol's value as variable is void:@: +
1905 The meaning is the same as in GNU Emacs 22.
1911 @cindex Passing information to functions
1913 To see how information is passed to functions, let's look again at
1914 our old standby, the addition of two plus two. In Lisp, this is written
1921 If you evaluate this expression, the number 4 will appear in your echo
1922 area. What the Lisp interpreter does is add the numbers that follow
1925 @cindex @samp{argument} defined
1926 The numbers added by @code{+} are called the @dfn{arguments} of the
1927 function @code{+}. These numbers are the information that is given to
1928 or @dfn{passed} to the function.
1930 The word ``argument'' comes from the way it is used in mathematics and
1931 does not refer to a disputation between two people; instead it refers to
1932 the information presented to the function, in this case, to the
1933 @code{+}. In Lisp, the arguments to a function are the atoms or lists
1934 that follow the function. The values returned by the evaluation of
1935 these atoms or lists are passed to the function. Different functions
1936 require different numbers of arguments; some functions require none at
1937 all.@footnote{It is curious to track the path by which the word ``argument''
1938 came to have two different meanings, one in mathematics and the other in
1939 everyday English. According to the @cite{Oxford English Dictionary},
1940 the word derives from the Latin for @samp{to make clear, prove}; thus it
1941 came to mean, by one thread of derivation, ``the evidence offered as
1942 proof'', which is to say, ``the information offered'', which led to its
1943 meaning in Lisp. But in the other thread of derivation, it came to mean
1944 ``to assert in a manner against which others may make counter
1945 assertions'', which led to the meaning of the word as a disputation.
1946 (Note here that the English word has two different definitions attached
1947 to it at the same time. By contrast, in Emacs Lisp, a symbol cannot
1948 have two different function definitions at the same time.)}
1951 * Data types:: Types of data passed to a function.
1952 * Args as Variable or List:: An argument can be the value
1953 of a variable or list.
1954 * Variable Number of Arguments:: Some functions may take a
1955 variable number of arguments.
1956 * Wrong Type of Argument:: Passing an argument of the wrong type
1958 * message:: A useful function for sending messages.
1962 @subsection Arguments' Data Types
1964 @cindex Types of data
1965 @cindex Arguments' data types
1967 The type of data that should be passed to a function depends on what
1968 kind of information it uses. The arguments to a function such as
1969 @code{+} must have values that are numbers, since @code{+} adds numbers.
1970 Other functions use different kinds of data for their arguments.
1974 For example, the @code{concat} function links together or unites two or
1975 more strings of text to produce a string. The arguments are strings.
1976 Concatenating the two character strings @code{abc}, @code{def} produces
1977 the single string @code{abcdef}. This can be seen by evaluating the
1981 (concat "abc" "def")
1985 The value produced by evaluating this expression is @code{"abcdef"}.
1988 A function such as @code{substring} uses both a string and numbers as
1989 arguments. The function returns a part of the string, a @dfn{substring} of
1990 the first argument. This function takes three arguments. Its first
1991 argument is the string of characters, the second and third arguments
1992 are numbers that indicate the beginning (inclusive) and end
1993 (exclusive) of the substring. The numbers are a count of the number
1994 of characters (including spaces and punctuation) from the beginning of
1995 the string. Note that the characters in a string are numbered from
1999 For example, if you evaluate the following:
2002 (substring "The quick brown fox jumped." 16 19)
2006 you will see @code{"fox"} appear in the echo area. The arguments are the
2007 string and the two numbers.
2009 Note that the string passed to @code{substring} is a single atom even
2010 though it is made up of several words separated by spaces. Lisp counts
2011 everything between the two quotation marks as part of the string,
2012 including the spaces. You can think of the @code{substring} function as
2013 a kind of atom smasher since it takes an otherwise indivisible atom
2014 and extracts a part. However, @code{substring} is only able to extract
2015 a substring from an argument that is a string, not from another type of
2016 atom such as a number or symbol.
2018 @node Args as Variable or List
2019 @subsection An Argument as the Value of a Variable or List
2021 An argument can be a symbol that returns a value when it is evaluated.
2022 For example, when the symbol @code{fill-column} by itself is evaluated,
2023 it returns a number. This number can be used in an addition.
2026 Position the cursor after the following expression and type @kbd{C-x
2034 The value will be a number two more than what you get by evaluating
2035 @code{fill-column} alone. For me, this is 74, because my value of
2036 @code{fill-column} is 72.
2038 As we have just seen, an argument can be a symbol that returns a value
2039 when evaluated. In addition, an argument can be a list that returns a
2040 value when it is evaluated. For example, in the following expression,
2041 the arguments to the function @code{concat} are the strings
2042 @w{@code{"The "}} and @w{@code{" red foxes."}} and the list
2043 @code{(number-to-string (+ 2 fill-column))}.
2045 @c For GNU Emacs 22, need number-to-string
2047 (concat "The " (number-to-string (+ 2 fill-column)) " red foxes.")
2051 If you evaluate this expression---and if, as with my Emacs,
2052 @code{fill-column} evaluates to 72---@code{"The 74 red foxes."} will
2053 appear in the echo area. (Note that you must put spaces after the
2054 word @samp{The} and before the word @samp{red} so they will appear in
2055 the final string. The function @code{number-to-string} converts the
2056 integer that the addition function returns to a string.
2057 @code{number-to-string} is also known as @code{int-to-string}.)
2059 @node Variable Number of Arguments
2060 @subsection Variable Number of Arguments
2061 @cindex Variable number of arguments
2062 @cindex Arguments, variable number of
2064 Some functions, such as @code{concat}, @code{+} or @code{*}, take any
2065 number of arguments. (The @code{*} is the symbol for multiplication.)
2066 This can be seen by evaluating each of the following expressions in
2067 the usual way. What you will see in the echo area is printed in this
2068 text after @samp{@result{}}, which you may read as ``evaluates to''.
2071 In the first set, the functions have no arguments:
2082 In this set, the functions have one argument each:
2093 In this set, the functions have three arguments each:
2097 (+ 3 4 5) @result{} 12
2099 (* 3 4 5) @result{} 60
2103 @node Wrong Type of Argument
2104 @subsection Using the Wrong Type Object as an Argument
2105 @cindex Wrong type of argument
2106 @cindex Argument, wrong type of
2108 When a function is passed an argument of the wrong type, the Lisp
2109 interpreter produces an error message. For example, the @code{+}
2110 function expects the values of its arguments to be numbers. As an
2111 experiment we can pass it the quoted symbol @code{hello} instead of a
2112 number. Position the cursor after the following expression and type
2120 When you do this you will generate an error message. What has happened
2121 is that @code{+} has tried to add the 2 to the value returned by
2122 @code{'hello}, but the value returned by @code{'hello} is the symbol
2123 @code{hello}, not a number. Only numbers can be added. So @code{+}
2124 could not carry out its addition.
2127 You will create and enter a @file{*Backtrace*} buffer that says:
2132 ---------- Buffer: *Backtrace* ----------
2133 Debugger entered--Lisp error:
2134 (wrong-type-argument number-or-marker-p hello)
2136 eval((+ 2 (quote hello)))
2137 eval-last-sexp-1(nil)
2139 call-interactively(eval-last-sexp)
2140 ---------- Buffer: *Backtrace* ----------
2145 As usual, the error message tries to be helpful and makes sense after you
2146 learn how to read it.@footnote{@code{(quote hello)} is an expansion of
2147 the abbreviation @code{'hello}.}
2149 The first part of the error message is straightforward; it says
2150 @samp{wrong type argument}. Next comes the mysterious jargon word
2151 @w{@samp{number-or-marker-p}}. This word is trying to tell you what
2152 kind of argument the @code{+} expected.
2154 The symbol @code{number-or-marker-p} says that the Lisp interpreter is
2155 trying to determine whether the information presented it (the value of
2156 the argument) is a number or a marker (a special object representing a
2157 buffer position). What it does is test to see whether the @code{+} is
2158 being given numbers to add. It also tests to see whether the
2159 argument is something called a marker, which is a specific feature of
2160 Emacs Lisp. (In Emacs, locations in a buffer are recorded as markers.
2161 When the mark is set with the @kbd{C-@@} or @kbd{C-@key{SPC}} command,
2162 its position is kept as a marker. The mark can be considered a
2163 number---the number of characters the location is from the beginning
2164 of the buffer.) In Emacs Lisp, @code{+} can be used to add the
2165 numeric value of marker positions as numbers.
2167 The @samp{p} of @code{number-or-marker-p} is the embodiment of a
2168 practice started in the early days of Lisp programming. The @samp{p}
2169 stands for ``predicate''. In the jargon used by the early Lisp
2170 researchers, a predicate refers to a function to determine whether some
2171 property is true or false. So the @samp{p} tells us that
2172 @code{number-or-marker-p} is the name of a function that determines
2173 whether it is true or false that the argument supplied is a number or
2174 a marker. Other Lisp symbols that end in @samp{p} include @code{zerop},
2175 a function that tests whether its argument has the value of zero, and
2176 @code{listp}, a function that tests whether its argument is a list.
2178 Finally, the last part of the error message is the symbol @code{hello}.
2179 This is the value of the argument that was passed to @code{+}. If the
2180 addition had been passed the correct type of object, the value passed
2181 would have been a number, such as 37, rather than a symbol like
2182 @code{hello}. But then you would not have got the error message.
2186 In GNU Emacs version 20 and before, the echo area displays an error
2190 Wrong type argument:@: number-or-marker-p, hello
2193 This says, in different words, the same as the top line of the
2194 @file{*Backtrace*} buffer.
2198 @subsection The @code{message} Function
2201 Like @code{+}, the @code{message} function takes a variable number of
2202 arguments. It is used to send messages to the user and is so useful
2203 that we will describe it here.
2206 A message is printed in the echo area. For example, you can print a
2207 message in your echo area by evaluating the following list:
2210 (message "This message appears in the echo area!")
2213 The whole string between double quotation marks is a single argument
2214 and is printed @i{in toto}. (Note that in this example, the message
2215 itself will appear in the echo area within double quotes; that is
2216 because you see the value returned by the @code{message} function. In
2217 most uses of @code{message} in programs that you write, the text will
2218 be printed in the echo area as a side-effect, without the quotes.
2219 @xref{multiply-by-seven in detail, , @code{multiply-by-seven} in
2220 detail}, for an example of this.)
2222 However, if there is a @samp{%s} in the quoted string of characters, the
2223 @code{message} function does not print the @samp{%s} as such, but looks
2224 to the argument that follows the string. It evaluates the second
2225 argument and prints the value at the location in the string where the
2229 You can see this by positioning the cursor after the following
2230 expression and typing @kbd{C-x C-e}:
2233 (message "The name of this buffer is: %s." (buffer-name))
2237 In Info, @code{"The name of this buffer is: *info*."} will appear in the
2238 echo area. The function @code{buffer-name} returns the name of the
2239 buffer as a string, which the @code{message} function inserts in place
2242 To print a value as an integer, use @samp{%d} in the same way as
2243 @samp{%s}. For example, to print a message in the echo area that
2244 states the value of the @code{fill-column}, evaluate the following:
2247 (message "The value of fill-column is %d." fill-column)
2251 On my system, when I evaluate this list, @code{"The value of
2252 fill-column is 72."} appears in my echo area@footnote{Actually, you
2253 can use @code{%s} to print a number. It is non-specific. @code{%d}
2254 prints only the part of a number left of a decimal point, and not
2255 anything that is not a number.}.
2257 If there is more than one @samp{%s} in the quoted string, the value of
2258 the first argument following the quoted string is printed at the
2259 location of the first @samp{%s} and the value of the second argument is
2260 printed at the location of the second @samp{%s}, and so on.
2263 For example, if you evaluate the following,
2267 (message "There are %d %s in the office!"
2268 (- fill-column 14) "pink elephants")
2273 a rather whimsical message will appear in your echo area. On my system
2274 it says, @code{"There are 58 pink elephants in the office!"}.
2276 The expression @code{(- fill-column 14)} is evaluated and the resulting
2277 number is inserted in place of the @samp{%d}; and the string in double
2278 quotes, @code{"pink elephants"}, is treated as a single argument and
2279 inserted in place of the @samp{%s}. (That is to say, a string between
2280 double quotes evaluates to itself, like a number.)
2282 Finally, here is a somewhat complex example that not only illustrates
2283 the computation of a number, but also shows how you can use an
2284 expression within an expression to generate the text that is substituted
2289 (message "He saw %d %s"
2293 "The quick brown foxes jumped." 16 21)
2298 In this example, @code{message} has three arguments: the string,
2299 @code{"He saw %d %s"}, the expression, @code{(- fill-column 32)}, and
2300 the expression beginning with the function @code{concat}. The value
2301 resulting from the evaluation of @code{(- fill-column 32)} is inserted
2302 in place of the @samp{%d}; and the value returned by the expression
2303 beginning with @code{concat} is inserted in place of the @samp{%s}.
2305 When your fill column is 70 and you evaluate the expression, the
2306 message @code{"He saw 38 red foxes leaping."} appears in your echo
2310 @section Setting the Value of a Variable
2311 @cindex Variable, setting value
2312 @cindex Setting value of variable
2314 @cindex @samp{bind} defined
2315 There are several ways by which a variable can be given a value. One of
2316 the ways is to use either the function @code{set} or the function
2317 @code{setq}. Another way is to use @code{let} (@pxref{let}). (The
2318 jargon for this process is to @dfn{bind} a variable to a value.)
2320 The following sections not only describe how @code{set} and @code{setq}
2321 work but also illustrate how arguments are passed.
2324 * Using set:: Setting values.
2325 * Using setq:: Setting a quoted value.
2326 * Counting:: Using @code{setq} to count.
2330 @subsection Using @code{set}
2333 To set the value of the symbol @code{flowers} to the list @code{'(rose
2334 violet daisy buttercup)}, evaluate the following expression by
2335 positioning the cursor after the expression and typing @kbd{C-x C-e}.
2338 (set 'flowers '(rose violet daisy buttercup))
2342 The list @code{(rose violet daisy buttercup)} will appear in the echo
2343 area. This is what is @emph{returned} by the @code{set} function. As a
2344 side effect, the symbol @code{flowers} is bound to the list; that is,
2345 the symbol @code{flowers}, which can be viewed as a variable, is given
2346 the list as its value. (This process, by the way, illustrates how a
2347 side effect to the Lisp interpreter, setting the value, can be the
2348 primary effect that we humans are interested in. This is because every
2349 Lisp function must return a value if it does not get an error, but it
2350 will only have a side effect if it is designed to have one.)
2352 After evaluating the @code{set} expression, you can evaluate the symbol
2353 @code{flowers} and it will return the value you just set. Here is the
2354 symbol. Place your cursor after it and type @kbd{C-x C-e}.
2361 When you evaluate @code{flowers}, the list
2362 @code{(rose violet daisy buttercup)} appears in the echo area.
2364 Incidentally, if you evaluate @code{'flowers}, the variable with a quote
2365 in front of it, what you will see in the echo area is the symbol itself,
2366 @code{flowers}. Here is the quoted symbol, so you can try this:
2372 Note also, that when you use @code{set}, you need to quote both
2373 arguments to @code{set}, unless you want them evaluated. Since we do
2374 not want either argument evaluated, neither the variable
2375 @code{flowers} nor the list @code{(rose violet daisy buttercup)}, both
2376 are quoted. (When you use @code{set} without quoting its first
2377 argument, the first argument is evaluated before anything else is
2378 done. If you did this and @code{flowers} did not have a value
2379 already, you would get an error message that the @samp{Symbol's value
2380 as variable is void}; on the other hand, if @code{flowers} did return
2381 a value after it was evaluated, the @code{set} would attempt to set
2382 the value that was returned. There are situations where this is the
2383 right thing for the function to do; but such situations are rare.)
2386 @subsection Using @code{setq}
2389 As a practical matter, you almost always quote the first argument to
2390 @code{set}. The combination of @code{set} and a quoted first argument
2391 is so common that it has its own name: the special form @code{setq}.
2392 This special form is just like @code{set} except that the first argument
2393 is quoted automatically, so you don't need to type the quote mark
2394 yourself. Also, as an added convenience, @code{setq} permits you to set
2395 several different variables to different values, all in one expression.
2397 To set the value of the variable @code{carnivores} to the list
2398 @code{'(lion tiger leopard)} using @code{setq}, the following expression
2402 (setq carnivores '(lion tiger leopard))
2406 This is exactly the same as using @code{set} except the first argument
2407 is automatically quoted by @code{setq}. (The @samp{q} in @code{setq}
2408 means @code{quote}.)
2411 With @code{set}, the expression would look like this:
2414 (set 'carnivores '(lion tiger leopard))
2417 Also, @code{setq} can be used to assign different values to
2418 different variables. The first argument is bound to the value
2419 of the second argument, the third argument is bound to the value of the
2420 fourth argument, and so on. For example, you could use the following to
2421 assign a list of trees to the symbol @code{trees} and a list of herbivores
2422 to the symbol @code{herbivores}:
2426 (setq trees '(pine fir oak maple)
2427 herbivores '(gazelle antelope zebra))
2432 (The expression could just as well have been on one line, but it might
2433 not have fit on a page; and humans find it easier to read nicely
2436 Although I have been using the term ``assign'', there is another way of
2437 thinking about the workings of @code{set} and @code{setq}; and that is to
2438 say that @code{set} and @code{setq} make the symbol @emph{point} to the
2439 list. This latter way of thinking is very common and in forthcoming
2440 chapters we shall come upon at least one symbol that has ``pointer'' as
2441 part of its name. The name is chosen because the symbol has a value,
2442 specifically a list, attached to it; or, expressed another way,
2443 the symbol is set to point to the list.
2446 @subsection Counting
2449 Here is an example that shows how to use @code{setq} in a counter. You
2450 might use this to count how many times a part of your program repeats
2451 itself. First set a variable to zero; then add one to the number each
2452 time the program repeats itself. To do this, you need a variable that
2453 serves as a counter, and two expressions: an initial @code{setq}
2454 expression that sets the counter variable to zero; and a second
2455 @code{setq} expression that increments the counter each time it is
2460 (setq counter 0) ; @r{Let's call this the initializer.}
2462 (setq counter (+ counter 1)) ; @r{This is the incrementer.}
2464 counter ; @r{This is the counter.}
2469 (The text following the @samp{;} are comments. @xref{Change a
2470 defun, , Change a Function Definition}.)
2472 If you evaluate the first of these expressions, the initializer,
2473 @code{(setq counter 0)}, and then evaluate the third expression,
2474 @code{counter}, the number @code{0} will appear in the echo area. If
2475 you then evaluate the second expression, the incrementer, @code{(setq
2476 counter (+ counter 1))}, the counter will get the value 1. So if you
2477 again evaluate @code{counter}, the number @code{1} will appear in the
2478 echo area. Each time you evaluate the second expression, the value of
2479 the counter will be incremented.
2481 When you evaluate the incrementer, @code{(setq counter (+ counter 1))},
2482 the Lisp interpreter first evaluates the innermost list; this is the
2483 addition. In order to evaluate this list, it must evaluate the variable
2484 @code{counter} and the number @code{1}. When it evaluates the variable
2485 @code{counter}, it receives its current value. It passes this value and
2486 the number @code{1} to the @code{+} which adds them together. The sum
2487 is then returned as the value of the inner list and passed to the
2488 @code{setq} which sets the variable @code{counter} to this new value.
2489 Thus, the value of the variable, @code{counter}, is changed.
2494 Learning Lisp is like climbing a hill in which the first part is the
2495 steepest. You have now climbed the most difficult part; what remains
2496 becomes easier as you progress onwards.
2504 Lisp programs are made up of expressions, which are lists or single atoms.
2507 Lists are made up of zero or more atoms or inner lists, separated by whitespace and
2508 surrounded by parentheses. A list can be empty.
2511 Atoms are multi-character symbols, like @code{forward-paragraph}, single
2512 character symbols like @code{+}, strings of characters between double
2513 quotation marks, or numbers.
2516 A number evaluates to itself.
2519 A string between double quotes also evaluates to itself.
2522 When you evaluate a symbol by itself, its value is returned.
2525 When you evaluate a list, the Lisp interpreter looks at the first symbol
2526 in the list and then at the function definition bound to that symbol.
2527 Then the instructions in the function definition are carried out.
2530 A single-quote @samp{'} tells the Lisp interpreter that it should
2531 return the following expression as written, and not evaluate it as it
2532 would if the quote were not there.
2535 Arguments are the information passed to a function. The arguments to a
2536 function are computed by evaluating the rest of the elements of the list
2537 of which the function is the first element.
2540 A function always returns a value when it is evaluated (unless it gets
2541 an error); in addition, it may also carry out some action that is a
2542 side effect. In many cases, a function's primary purpose is to
2543 create a side effect.
2546 @node Error Message Exercises
2549 A few simple exercises:
2553 Generate an error message by evaluating an appropriate symbol that is
2554 not within parentheses.
2557 Generate an error message by evaluating an appropriate symbol that is
2558 between parentheses.
2561 Create a counter that increments by two rather than one.
2564 Write an expression that prints a message in the echo area when
2568 @node Practicing Evaluation
2569 @chapter Practicing Evaluation
2570 @cindex Practicing evaluation
2571 @cindex Evaluation practice
2573 Before learning how to write a function definition in Emacs Lisp, it is
2574 useful to spend a little time evaluating various expressions that have
2575 already been written. These expressions will be lists with the
2576 functions as their first (and often only) element. Since some of the
2577 functions associated with buffers are both simple and interesting, we
2578 will start with those. In this section, we will evaluate a few of
2579 these. In another section, we will study the code of several other
2580 buffer-related functions, to see how they were written.
2583 * How to Evaluate:: Typing editing commands or @kbd{C-x C-e}
2585 * Buffer Names:: Buffers and files are different.
2586 * Getting Buffers:: Getting a buffer itself, not merely its name.
2587 * Switching Buffers:: How to change to another buffer.
2588 * Buffer Size & Locations:: Where point is located and the size of
2590 * Evaluation Exercise::
2594 @node How to Evaluate
2595 @unnumberedsec How to Evaluate
2598 @i{Whenever you give an editing command} to Emacs Lisp, such as the
2599 command to move the cursor or to scroll the screen, @i{you are evaluating
2600 an expression,} the first element of which is a function. @i{This is
2603 @cindex @samp{interactive function} defined
2604 @cindex @samp{command} defined
2605 When you type keys, you cause the Lisp interpreter to evaluate an
2606 expression and that is how you get your results. Even typing plain text
2607 involves evaluating an Emacs Lisp function, in this case, one that uses
2608 @code{self-insert-command}, which simply inserts the character you
2609 typed. The functions you evaluate by typing keystrokes are called
2610 @dfn{interactive} functions, or @dfn{commands}; how you make a function
2611 interactive will be illustrated in the chapter on how to write function
2612 definitions. @xref{Interactive, , Making a Function Interactive}.
2614 In addition to typing keyboard commands, we have seen a second way to
2615 evaluate an expression: by positioning the cursor after a list and
2616 typing @kbd{C-x C-e}. This is what we will do in the rest of this
2617 section. There are other ways to evaluate an expression as well; these
2618 will be described as we come to them.
2620 Besides being used for practicing evaluation, the functions shown in the
2621 next few sections are important in their own right. A study of these
2622 functions makes clear the distinction between buffers and files, how to
2623 switch to a buffer, and how to determine a location within it.
2626 @section Buffer Names
2628 @findex buffer-file-name
2630 The two functions, @code{buffer-name} and @code{buffer-file-name}, show
2631 the difference between a file and a buffer. When you evaluate the
2632 following expression, @code{(buffer-name)}, the name of the buffer
2633 appears in the echo area. When you evaluate @code{(buffer-file-name)},
2634 the name of the file to which the buffer refers appears in the echo
2635 area. Usually, the name returned by @code{(buffer-name)} is the same as
2636 the name of the file to which it refers, and the name returned by
2637 @code{(buffer-file-name)} is the full path-name of the file.
2639 A file and a buffer are two different entities. A file is information
2640 recorded permanently in the computer (unless you delete it). A buffer,
2641 on the other hand, is information inside of Emacs that will vanish at
2642 the end of the editing session (or when you kill the buffer). Usually,
2643 a buffer contains information that you have copied from a file; we say
2644 the buffer is @dfn{visiting} that file. This copy is what you work on
2645 and modify. Changes to the buffer do not change the file, until you
2646 save the buffer. When you save the buffer, the buffer is copied to the file
2647 and is thus saved permanently.
2650 If you are reading this in Info inside of GNU Emacs, you can evaluate
2651 each of the following expressions by positioning the cursor after it and
2652 typing @kbd{C-x C-e}.
2663 When I do this in Info, the value returned by evaluating
2664 @code{(buffer-name)} is @file{"*info*"}, and the value returned by
2665 evaluating @code{(buffer-file-name)} is @file{nil}.
2667 On the other hand, while I am writing this document, the value
2668 returned by evaluating @code{(buffer-name)} is
2669 @file{"introduction.texinfo"}, and the value returned by evaluating
2670 @code{(buffer-file-name)} is
2671 @file{"/gnu/work/intro/introduction.texinfo"}.
2673 @cindex @code{nil}, history of word
2674 The former is the name of the buffer and the latter is the name of the
2675 file. In Info, the buffer name is @file{"*info*"}. Info does not
2676 point to any file, so the result of evaluating
2677 @code{(buffer-file-name)} is @file{nil}. The symbol @code{nil} is
2678 from the Latin word for ``nothing''; in this case, it means that the
2679 buffer is not associated with any file. (In Lisp, @code{nil} is also
2680 used to mean ``false'' and is a synonym for the empty list, @code{()}.)
2682 When I am writing, the name of my buffer is
2683 @file{"introduction.texinfo"}. The name of the file to which it
2684 points is @file{"/gnu/work/intro/introduction.texinfo"}.
2686 (In the expressions, the parentheses tell the Lisp interpreter to
2687 treat @w{@code{buffer-name}} and @w{@code{buffer-file-name}} as
2688 functions; without the parentheses, the interpreter would attempt to
2689 evaluate the symbols as variables. @xref{Variables}.)
2691 In spite of the distinction between files and buffers, you will often
2692 find that people refer to a file when they mean a buffer and vice versa.
2693 Indeed, most people say, ``I am editing a file,'' rather than saying,
2694 ``I am editing a buffer which I will soon save to a file.'' It is
2695 almost always clear from context what people mean. When dealing with
2696 computer programs, however, it is important to keep the distinction in mind,
2697 since the computer is not as smart as a person.
2699 @cindex Buffer, history of word
2700 The word ``buffer'', by the way, comes from the meaning of the word as a
2701 cushion that deadens the force of a collision. In early computers, a
2702 buffer cushioned the interaction between files and the computer's
2703 central processing unit. The drums or tapes that held a file and the
2704 central processing unit were pieces of equipment that were very
2705 different from each other, working at their own speeds, in spurts. The
2706 buffer made it possible for them to work together effectively.
2707 Eventually, the buffer grew from being an intermediary, a temporary
2708 holding place, to being the place where work is done. This
2709 transformation is rather like that of a small seaport that grew into a
2710 great city: once it was merely the place where cargo was warehoused
2711 temporarily before being loaded onto ships; then it became a business
2712 and cultural center in its own right.
2714 Not all buffers are associated with files. For example, a
2715 @file{*scratch*} buffer does not visit any file. Similarly, a
2716 @file{*Help*} buffer is not associated with any file.
2718 In the old days, when you lacked a @file{~/.emacs} file and started an
2719 Emacs session by typing the command @code{emacs} alone, without naming
2720 any files, Emacs started with the @file{*scratch*} buffer visible.
2721 Nowadays, you will see a splash screen. You can follow one of the
2722 commands suggested on the splash screen, visit a file, or press the
2723 spacebar to reach the @file{*scratch*} buffer.
2725 If you switch to the @file{*scratch*} buffer, type
2726 @code{(buffer-name)}, position the cursor after it, and then type
2727 @kbd{C-x C-e} to evaluate the expression. The name @code{"*scratch*"}
2728 will be returned and will appear in the echo area. @code{"*scratch*"}
2729 is the name of the buffer. When you type @code{(buffer-file-name)} in
2730 the @file{*scratch*} buffer and evaluate that, @code{nil} will appear
2731 in the echo area, just as it does when you evaluate
2732 @code{(buffer-file-name)} in Info.
2734 Incidentally, if you are in the @file{*scratch*} buffer and want the
2735 value returned by an expression to appear in the @file{*scratch*}
2736 buffer itself rather than in the echo area, type @kbd{C-u C-x C-e}
2737 instead of @kbd{C-x C-e}. This causes the value returned to appear
2738 after the expression. The buffer will look like this:
2741 (buffer-name)"*scratch*"
2745 You cannot do this in Info since Info is read-only and it will not allow
2746 you to change the contents of the buffer. But you can do this in any
2747 buffer you can edit; and when you write code or documentation (such as
2748 this book), this feature is very useful.
2750 @node Getting Buffers
2751 @section Getting Buffers
2752 @findex current-buffer
2753 @findex other-buffer
2754 @cindex Getting a buffer
2756 The @code{buffer-name} function returns the @emph{name} of the buffer;
2757 to get the buffer @emph{itself}, a different function is needed: the
2758 @code{current-buffer} function. If you use this function in code, what
2759 you get is the buffer itself.
2761 A name and the object or entity to which the name refers are different
2762 from each other. You are not your name. You are a person to whom
2763 others refer by name. If you ask to speak to George and someone hands you
2764 a card with the letters @samp{G}, @samp{e}, @samp{o}, @samp{r},
2765 @samp{g}, and @samp{e} written on it, you might be amused, but you would
2766 not be satisfied. You do not want to speak to the name, but to the
2767 person to whom the name refers. A buffer is similar: the name of the
2768 scratch buffer is @file{*scratch*}, but the name is not the buffer. To
2769 get a buffer itself, you need to use a function such as
2770 @code{current-buffer}.
2772 However, there is a slight complication: if you evaluate
2773 @code{current-buffer} in an expression on its own, as we will do here,
2774 what you see is a printed representation of the name of the buffer
2775 without the contents of the buffer. Emacs works this way for two
2776 reasons: the buffer may be thousands of lines long---too long to be
2777 conveniently displayed; and, another buffer may have the same contents
2778 but a different name, and it is important to distinguish between them.
2781 Here is an expression containing the function:
2788 If you evaluate this expression in Info in Emacs in the usual way,
2789 @file{#<buffer *info*>} will appear in the echo area. The special
2790 format indicates that the buffer itself is being returned, rather than
2793 Incidentally, while you can type a number or symbol into a program, you
2794 cannot do that with the printed representation of a buffer: the only way
2795 to get a buffer itself is with a function such as @code{current-buffer}.
2797 A related function is @code{other-buffer}. This returns the most
2798 recently selected buffer other than the one you are in currently, not
2799 a printed representation of its name. If you have recently switched
2800 back and forth from the @file{*scratch*} buffer, @code{other-buffer}
2801 will return that buffer.
2804 You can see this by evaluating the expression:
2811 You should see @file{#<buffer *scratch*>} appear in the echo area, or
2812 the name of whatever other buffer you switched back from most
2813 recently@footnote{Actually, by default, if the buffer from which you
2814 just switched is visible to you in another window, @code{other-buffer}
2815 will choose the most recent buffer that you cannot see; this is a
2816 subtlety that I often forget.}.
2818 @node Switching Buffers
2819 @section Switching Buffers
2820 @findex switch-to-buffer
2822 @cindex Switching to a buffer
2824 The @code{other-buffer} function actually provides a buffer when it is
2825 used as an argument to a function that requires one. We can see this
2826 by using @code{other-buffer} and @code{switch-to-buffer} to switch to a
2829 But first, a brief introduction to the @code{switch-to-buffer}
2830 function. When you switched back and forth from Info to the
2831 @file{*scratch*} buffer to evaluate @code{(buffer-name)}, you most
2832 likely typed @kbd{C-x b} and then typed @file{*scratch*}@footnote{Or
2833 rather, to save typing, you probably only typed @kbd{RET} if the
2834 default buffer was @file{*scratch*}, or if it was different, then you
2835 typed just part of the name, such as @code{*sc}, pressed your
2836 @kbd{TAB} key to cause it to expand to the full name, and then typed
2837 @kbd{RET}.} when prompted in the minibuffer for the name of
2838 the buffer to which you wanted to switch. The keystrokes, @kbd{C-x
2839 b}, cause the Lisp interpreter to evaluate the interactive function
2840 @code{switch-to-buffer}. As we said before, this is how Emacs works:
2841 different keystrokes call or run different functions. For example,
2842 @kbd{C-f} calls @code{forward-char}, @kbd{M-e} calls
2843 @code{forward-sentence}, and so on.
2845 By writing @code{switch-to-buffer} in an expression, and giving it a
2846 buffer to switch to, we can switch buffers just the way @kbd{C-x b}
2850 (switch-to-buffer (other-buffer))
2854 The symbol @code{switch-to-buffer} is the first element of the list,
2855 so the Lisp interpreter will treat it as a function and carry out the
2856 instructions that are attached to it. But before doing that, the
2857 interpreter will note that @code{other-buffer} is inside parentheses
2858 and work on that symbol first. @code{other-buffer} is the first (and
2859 in this case, the only) element of this list, so the Lisp interpreter
2860 calls or runs the function. It returns another buffer. Next, the
2861 interpreter runs @code{switch-to-buffer}, passing to it, as an
2862 argument, the other buffer, which is what Emacs will switch to. If
2863 you are reading this in Info, try this now. Evaluate the expression.
2864 (To get back, type @kbd{C-x b @key{RET}}.)@footnote{Remember, this
2865 expression will move you to your most recent other buffer that you
2866 cannot see. If you really want to go to your most recently selected
2867 buffer, even if you can still see it, you need to evaluate the
2868 following more complex expression:
2871 (switch-to-buffer (other-buffer (current-buffer) t))
2875 In this case, the first argument to @code{other-buffer} tells it which
2876 buffer to skip---the current one---and the second argument tells
2877 @code{other-buffer} it is OK to switch to a visible buffer. In
2878 regular use, @code{switch-to-buffer} takes you to a buffer not visible
2879 in windows since you would most likely use @kbd{C-x o}
2880 (@code{other-window}) to go to another visible buffer.}
2882 In the programming examples in later sections of this document, you will
2883 see the function @code{set-buffer} more often than
2884 @code{switch-to-buffer}. This is because of a difference between
2885 computer programs and humans: humans have eyes and expect to see the
2886 buffer on which they are working on their computer terminals. This is
2887 so obvious, it almost goes without saying. However, programs do not
2888 have eyes. When a computer program works on a buffer, that buffer does
2889 not need to be visible on the screen.
2891 @code{switch-to-buffer} is designed for humans and does two different
2892 things: it switches the buffer to which Emacs's attention is directed; and
2893 it switches the buffer displayed in the window to the new buffer.
2894 @code{set-buffer}, on the other hand, does only one thing: it switches
2895 the attention of the computer program to a different buffer. The buffer
2896 on the screen remains unchanged (of course, normally nothing happens
2897 there until the command finishes running).
2899 @cindex @samp{call} defined
2900 Also, we have just introduced another jargon term, the word @dfn{call}.
2901 When you evaluate a list in which the first symbol is a function, you
2902 are calling that function. The use of the term comes from the notion of
2903 the function as an entity that can do something for you if you call
2904 it---just as a plumber is an entity who can fix a leak if you call him
2907 @node Buffer Size & Locations
2908 @section Buffer Size and the Location of Point
2909 @cindex Size of buffer
2911 @cindex Point location
2912 @cindex Location of point
2914 Finally, let's look at several rather simple functions,
2915 @code{buffer-size}, @code{point}, @code{point-min}, and
2916 @code{point-max}. These give information about the size of a buffer and
2917 the location of point within it.
2919 The function @code{buffer-size} tells you the size of the current
2920 buffer; that is, the function returns a count of the number of
2921 characters in the buffer.
2928 You can evaluate this in the usual way, by positioning the
2929 cursor after the expression and typing @kbd{C-x C-e}.
2931 @cindex @samp{point} defined
2932 In Emacs, the current position of the cursor is called @dfn{point}.
2933 The expression @code{(point)} returns a number that tells you where the
2934 cursor is located as a count of the number of characters from the
2935 beginning of the buffer up to point.
2938 You can see the character count for point in this buffer by evaluating
2939 the following expression in the usual way:
2946 As I write this, the value of point is 65724. The @code{point}
2947 function is frequently used in some of the examples later in this
2951 The value of point depends, of course, on its location within the
2952 buffer. If you evaluate point in this spot, the number will be larger:
2959 For me, the value of point in this location is 66043, which means that
2960 there are 319 characters (including spaces) between the two
2961 expressions. (Doubtless, you will see different numbers, since I will
2962 have edited this since I first evaluated point.)
2964 @cindex @samp{narrowing} defined
2965 The function @code{point-min} is somewhat similar to @code{point}, but
2966 it returns the value of the minimum permissible value of point in the
2967 current buffer. This is the number 1 unless @dfn{narrowing} is in
2968 effect. (Narrowing is a mechanism whereby you can restrict yourself,
2969 or a program, to operations on just a part of a buffer.
2970 @xref{Narrowing & Widening, , Narrowing and Widening}.) Likewise, the
2971 function @code{point-max} returns the value of the maximum permissible
2972 value of point in the current buffer.
2974 @node Evaluation Exercise
2977 Find a file with which you are working and move towards its middle.
2978 Find its buffer name, file name, length, and your position in the file.
2980 @node Writing Defuns
2981 @chapter How To Write Function Definitions
2982 @cindex Definition writing
2983 @cindex Function definition writing
2984 @cindex Writing a function definition
2986 When the Lisp interpreter evaluates a list, it looks to see whether the
2987 first symbol on the list has a function definition attached to it; or,
2988 put another way, whether the symbol points to a function definition. If
2989 it does, the computer carries out the instructions in the definition. A
2990 symbol that has a function definition is called, simply, a function
2991 (although, properly speaking, the definition is the function and the
2992 symbol refers to it.)
2995 * Primitive Functions::
2996 * defun:: The @code{defun} macro.
2997 * Install:: Install a function definition.
2998 * Interactive:: Making a function interactive.
2999 * Interactive Options:: Different options for @code{interactive}.
3000 * Permanent Installation:: Installing code permanently.
3001 * let:: Creating and initializing local variables.
3003 * else:: If--then--else expressions.
3004 * Truth & Falsehood:: What Lisp considers false and true.
3005 * save-excursion:: Keeping track of point and buffer.
3011 @node Primitive Functions
3012 @unnumberedsec An Aside about Primitive Functions
3014 @cindex Primitive functions
3015 @cindex Functions, primitive
3017 @cindex C language primitives
3018 @cindex Primitives written in C
3019 All functions are defined in terms of other functions, except for a few
3020 @dfn{primitive} functions that are written in the C programming
3021 language. When you write functions' definitions, you will write them in
3022 Emacs Lisp and use other functions as your building blocks. Some of the
3023 functions you will use will themselves be written in Emacs Lisp (perhaps
3024 by you) and some will be primitives written in C@. The primitive
3025 functions are used exactly like those written in Emacs Lisp and behave
3026 like them. They are written in C so we can easily run GNU Emacs on any
3027 computer that has sufficient power and can run C.
3029 Let me re-emphasize this: when you write code in Emacs Lisp, you do not
3030 distinguish between the use of functions written in C and the use of
3031 functions written in Emacs Lisp. The difference is irrelevant. I
3032 mention the distinction only because it is interesting to know. Indeed,
3033 unless you investigate, you won't know whether an already-written
3034 function is written in Emacs Lisp or C.
3037 @section The @code{defun} Macro
3040 @cindex @samp{function definition} defined
3041 In Lisp, a symbol such as @code{mark-whole-buffer} has code attached to
3042 it that tells the computer what to do when the function is called.
3043 This code is called the @dfn{function definition} and is created by
3044 evaluating a Lisp expression that starts with the symbol @code{defun}
3045 (which is an abbreviation for @emph{define function}).
3047 In subsequent sections, we will look at function definitions from the
3048 Emacs source code, such as @code{mark-whole-buffer}. In this section,
3049 we will describe a simple function definition so you can see how it
3050 looks. This function definition uses arithmetic because it makes for a
3051 simple example. Some people dislike examples using arithmetic; however,
3052 if you are such a person, do not despair. Hardly any of the code we
3053 will study in the remainder of this introduction involves arithmetic or
3054 mathematics. The examples mostly involve text in one way or another.
3056 A function definition has up to five parts following the word
3061 The name of the symbol to which the function definition should be
3065 A list of the arguments that will be passed to the function. If no
3066 arguments will be passed to the function, this is an empty list,
3070 Documentation describing the function. (Technically optional, but
3071 strongly recommended.)
3074 Optionally, an expression to make the function interactive so you can
3075 use it by typing @kbd{M-x} and then the name of the function; or by
3076 typing an appropriate key or keychord.
3078 @cindex @samp{body} defined
3080 The code that instructs the computer what to do: the @dfn{body} of the
3081 function definition.
3084 It is helpful to think of the five parts of a function definition as
3085 being organized in a template, with slots for each part:
3089 (defun @var{function-name} (@var{arguments}@dots{})
3090 "@var{optional-documentation}@dots{}"
3091 (interactive @var{argument-passing-info}) ; @r{optional}
3096 As an example, here is the code for a function that multiplies its
3097 argument by 7. (This example is not interactive. @xref{Interactive,
3098 , Making a Function Interactive}, for that information.)
3102 (defun multiply-by-seven (number)
3103 "Multiply NUMBER by seven."
3108 This definition begins with a parenthesis and the symbol @code{defun},
3109 followed by the name of the function.
3111 @cindex @samp{argument list} defined
3112 The name of the function is followed by a list that contains the
3113 arguments that will be passed to the function. This list is called
3114 the @dfn{argument list}. In this example, the list has only one
3115 element, the symbol, @code{number}. When the function is used, the
3116 symbol will be bound to the value that is used as the argument to the
3119 Instead of choosing the word @code{number} for the name of the argument,
3120 I could have picked any other name. For example, I could have chosen
3121 the word @code{multiplicand}. I picked the word ``number'' because it
3122 tells what kind of value is intended for this slot; but I could just as
3123 well have chosen the word ``multiplicand'' to indicate the role that the
3124 value placed in this slot will play in the workings of the function. I
3125 could have called it @code{foogle}, but that would have been a bad
3126 choice because it would not tell humans what it means. The choice of
3127 name is up to the programmer and should be chosen to make the meaning of
3130 Indeed, you can choose any name you wish for a symbol in an argument
3131 list, even the name of a symbol used in some other function: the name
3132 you use in an argument list is private to that particular definition.
3133 In that definition, the name refers to a different entity than any use
3134 of the same name outside the function definition. Suppose you have a
3135 nick-name ``Shorty'' in your family; when your family members refer to
3136 ``Shorty'', they mean you. But outside your family, in a movie, for
3137 example, the name ``Shorty'' refers to someone else. Because a name in an
3138 argument list is private to the function definition, you can change the
3139 value of such a symbol inside the body of a function without changing
3140 its value outside the function. The effect is similar to that produced
3141 by a @code{let} expression. (@xref{let, , @code{let}}.)
3144 Note also that we discuss the word ``number'' in two different ways: as a
3145 symbol that appears in the code, and as the name of something that will
3146 be replaced by a something else during the evaluation of the function.
3147 In the first case, @code{number} is a symbol, not a number; it happens
3148 that within the function, it is a variable who value is the number in
3149 question, but our primary interest in it is as a symbol. On the other
3150 hand, when we are talking about the function, our interest is that we
3151 will substitute a number for the word @var{number}. To keep this
3152 distinction clear, we use different typography for the two
3153 circumstances. When we talk about this function, or about how it works,
3154 we refer to this number by writing @var{number}. In the function
3155 itself, we refer to it by writing @code{number}.
3158 The argument list is followed by the documentation string that
3159 describes the function. This is what you see when you type
3160 @w{@kbd{C-h f}} and the name of a function. Incidentally, when you
3161 write a documentation string like this, you should make the first line
3162 a complete sentence since some commands, such as @code{apropos}, print
3163 only the first line of a multi-line documentation string. Also, you
3164 should not indent the second line of a documentation string, if you
3165 have one, because that looks odd when you use @kbd{C-h f}
3166 (@code{describe-function}). The documentation string is optional, but
3167 it is so useful, it should be included in almost every function you
3170 @findex * @r{(multiplication)}
3171 The third line of the example consists of the body of the function
3172 definition. (Most functions' definitions, of course, are longer than
3173 this.) In this function, the body is the list, @code{(* 7 number)}, which
3174 says to multiply the value of @var{number} by 7. (In Emacs Lisp,
3175 @code{*} is the function for multiplication, just as @code{+} is the
3176 function for addition.)
3178 When you use the @code{multiply-by-seven} function, the argument
3179 @code{number} evaluates to the actual number you want used. Here is an
3180 example that shows how @code{multiply-by-seven} is used; but don't try
3181 to evaluate this yet!
3184 (multiply-by-seven 3)
3188 The symbol @code{number}, specified in the function definition in the
3189 next section, is bound to the value 3 in the actual use of
3190 the function. Note that although @code{number} was inside parentheses
3191 in the function definition, the argument passed to the
3192 @code{multiply-by-seven} function is not in parentheses. The
3193 parentheses are written in the function definition so the computer can
3194 figure out where the argument list ends and the rest of the function
3197 If you evaluate this example, you are likely to get an error message.
3198 (Go ahead, try it!) This is because we have written the function
3199 definition, but not yet told the computer about the definition---we have
3200 not yet loaded the function definition in Emacs.
3201 Installing a function is the process that tells the Lisp interpreter the
3202 definition of the function. Installation is described in the next
3206 @section Install a Function Definition
3207 @cindex Install a Function Definition
3208 @cindex Definition installation
3209 @cindex Function definition installation
3211 If you are reading this inside of Info in Emacs, you can try out the
3212 @code{multiply-by-seven} function by first evaluating the function
3213 definition and then evaluating @code{(multiply-by-seven 3)}. A copy of
3214 the function definition follows. Place the cursor after the last
3215 parenthesis of the function definition and type @kbd{C-x C-e}. When you
3216 do this, @code{multiply-by-seven} will appear in the echo area. (What
3217 this means is that when a function definition is evaluated, the value it
3218 returns is the name of the defined function.) At the same time, this
3219 action installs the function definition.
3223 (defun multiply-by-seven (number)
3224 "Multiply NUMBER by seven."
3230 By evaluating this @code{defun}, you have just installed
3231 @code{multiply-by-seven} in Emacs. The function is now just as much a
3232 part of Emacs as @code{forward-word} or any other editing function you
3233 use. (@code{multiply-by-seven} will stay installed until you quit
3234 Emacs. To reload code automatically whenever you start Emacs, see
3235 @ref{Permanent Installation, , Installing Code Permanently}.)
3238 * Effect of installation::
3239 * Change a defun:: How to change a function definition.
3243 @node Effect of installation
3244 @unnumberedsubsec The effect of installation
3247 You can see the effect of installing @code{multiply-by-seven} by
3248 evaluating the following sample. Place the cursor after the following
3249 expression and type @kbd{C-x C-e}. The number 21 will appear in the
3253 (multiply-by-seven 3)
3256 If you wish, you can read the documentation for the function by typing
3257 @kbd{C-h f} (@code{describe-function}) and then the name of the
3258 function, @code{multiply-by-seven}. When you do this, a
3259 @file{*Help*} window will appear on your screen that says:
3263 multiply-by-seven is a Lisp function.
3265 (multiply-by-seven NUMBER)
3267 Multiply NUMBER by seven.
3272 (To return to a single window on your screen, type @kbd{C-x 1}.)
3274 @node Change a defun
3275 @subsection Change a Function Definition
3276 @cindex Changing a function definition
3277 @cindex Function definition, how to change
3278 @cindex Definition, how to change
3280 If you want to change the code in @code{multiply-by-seven}, just rewrite
3281 it. To install the new version in place of the old one, evaluate the
3282 function definition again. This is how you modify code in Emacs. It is
3285 As an example, you can change the @code{multiply-by-seven} function to
3286 add the number to itself seven times instead of multiplying the number
3287 by seven. It produces the same answer, but by a different path. At
3288 the same time, we will add a comment to the code; a comment is text
3289 that the Lisp interpreter ignores, but that a human reader may find
3290 useful or enlightening. The comment is that this is the second
3295 (defun multiply-by-seven (number) ; @r{Second version.}
3296 "Multiply NUMBER by seven."
3297 (+ number number number number number number number))
3301 @cindex Comments in Lisp code
3302 The comment follows a semicolon, @samp{;}. In Lisp, everything on a
3303 line that follows a semicolon is a comment. The end of the line is the
3304 end of the comment. To stretch a comment over two or more lines, begin
3305 each line with a semicolon.
3307 @xref{Beginning init File, , Beginning a @file{.emacs}
3308 File}, and @ref{Comments, , Comments, elisp, The GNU Emacs Lisp
3309 Reference Manual}, for more about comments.
3311 You can install this version of the @code{multiply-by-seven} function by
3312 evaluating it in the same way you evaluated the first function: place
3313 the cursor after the last parenthesis and type @kbd{C-x C-e}.
3315 In summary, this is how you write code in Emacs Lisp: you write a
3316 function; install it; test it; and then make fixes or enhancements and
3320 @section Make a Function Interactive
3321 @cindex Interactive functions
3324 You make a function interactive by placing a list that begins with
3325 the special form @code{interactive} immediately after the
3326 documentation. A user can invoke an interactive function by typing
3327 @kbd{M-x} and then the name of the function; or by typing the keys to
3328 which it is bound, for example, by typing @kbd{C-n} for
3329 @code{next-line} or @kbd{C-x h} for @code{mark-whole-buffer}.
3331 Interestingly, when you call an interactive function interactively,
3332 the value returned is not automatically displayed in the echo area.
3333 This is because you often call an interactive function for its side
3334 effects, such as moving forward by a word or line, and not for the
3335 value returned. If the returned value were displayed in the echo area
3336 each time you typed a key, it would be very distracting.
3339 * Interactive multiply-by-seven:: An overview.
3340 * multiply-by-seven in detail:: The interactive version.
3344 @node Interactive multiply-by-seven
3345 @unnumberedsubsec An Interactive @code{multiply-by-seven}, An Overview
3348 Both the use of the special form @code{interactive} and one way to
3349 display a value in the echo area can be illustrated by creating an
3350 interactive version of @code{multiply-by-seven}.
3357 (defun multiply-by-seven (number) ; @r{Interactive version.}
3358 "Multiply NUMBER by seven."
3360 (message "The result is %d" (* 7 number)))
3365 You can install this code by placing your cursor after it and typing
3366 @kbd{C-x C-e}. The name of the function will appear in your echo area.
3367 Then, you can use this code by typing @kbd{C-u} and a number and then
3368 typing @kbd{M-x multiply-by-seven} and pressing @key{RET}. The phrase
3369 @samp{The result is @dots{}} followed by the product will appear in the
3372 Speaking more generally, you invoke a function like this in either of two
3377 By typing a prefix argument that contains the number to be passed, and
3378 then typing @kbd{M-x} and the name of the function, as with
3379 @kbd{C-u 3 M-x forward-sentence}; or,
3382 By typing whatever key or keychord the function is bound to, as with
3387 Both the examples just mentioned work identically to move point forward
3388 three sentences. (Since @code{multiply-by-seven} is not bound to a key,
3389 it could not be used as an example of key binding.)
3391 (@xref{Keybindings, , Some Keybindings}, to learn how to bind a command
3394 A @dfn{prefix argument} is passed to an interactive function by typing the
3395 @key{META} key followed by a number, for example, @kbd{M-3 M-e}, or by
3396 typing @kbd{C-u} and then a number, for example, @kbd{C-u 3 M-e} (if you
3397 type @kbd{C-u} without a number, it defaults to 4).
3399 @node multiply-by-seven in detail
3400 @subsection An Interactive @code{multiply-by-seven}
3402 Let's look at the use of the special form @code{interactive} and then at
3403 the function @code{message} in the interactive version of
3404 @code{multiply-by-seven}. You will recall that the function definition
3409 (defun multiply-by-seven (number) ; @r{Interactive version.}
3410 "Multiply NUMBER by seven."
3412 (message "The result is %d" (* 7 number)))
3416 In this function, the expression, @code{(interactive "p")}, is a list of
3417 two elements. The @code{"p"} tells Emacs to pass the prefix argument to
3418 the function and use its value for the argument of the function.
3421 The argument will be a number. This means that the symbol
3422 @code{number} will be bound to a number in the line:
3425 (message "The result is %d" (* 7 number))
3430 For example, if your prefix argument is 5, the Lisp interpreter will
3431 evaluate the line as if it were:
3434 (message "The result is %d" (* 7 5))
3438 (If you are reading this in GNU Emacs, you can evaluate this expression
3439 yourself.) First, the interpreter will evaluate the inner list, which
3440 is @code{(* 7 5)}. This returns a value of 35. Next, it
3441 will evaluate the outer list, passing the values of the second and
3442 subsequent elements of the list to the function @code{message}.
3444 As we have seen, @code{message} is an Emacs Lisp function especially
3445 designed for sending a one line message to a user. (@xref{message, ,
3446 The @code{message} function}.) In summary, the @code{message}
3447 function prints its first argument in the echo area as is, except for
3448 occurrences of @samp{%d} or @samp{%s} (and various other %-sequences
3449 which we have not mentioned). When it sees a control sequence, the
3450 function looks to the second or subsequent arguments and prints the
3451 value of the argument in the location in the string where the control
3452 sequence is located.
3454 In the interactive @code{multiply-by-seven} function, the control string
3455 is @samp{%d}, which requires a number, and the value returned by
3456 evaluating @code{(* 7 5)} is the number 35. Consequently, the number 35
3457 is printed in place of the @samp{%d} and the message is @samp{The result
3460 (Note that when you call the function @code{multiply-by-seven}, the
3461 message is printed without quotes, but when you call @code{message}, the
3462 text is printed in double quotes. This is because the value returned by
3463 @code{message} is what appears in the echo area when you evaluate an
3464 expression whose first element is @code{message}; but when embedded in a
3465 function, @code{message} prints the text as a side effect without
3468 @node Interactive Options
3469 @section Different Options for @code{interactive}
3470 @cindex Options for @code{interactive}
3471 @cindex Interactive options
3473 In the example, @code{multiply-by-seven} used @code{"p"} as the
3474 argument to @code{interactive}. This argument told Emacs to interpret
3475 your typing either @kbd{C-u} followed by a number or @key{META}
3476 followed by a number as a command to pass that number to the function
3477 as its argument. Emacs has more than twenty characters predefined for
3478 use with @code{interactive}. In almost every case, one of these
3479 options will enable you to pass the right information interactively to
3480 a function. (@xref{Interactive Codes, , Code Characters for
3481 @code{interactive}, elisp, The GNU Emacs Lisp Reference Manual}.)
3484 Consider the function @code{zap-to-char}. Its interactive expression
3487 @c FIXME: the interactive expression of zap-to-char has been changed
3491 (interactive "p\ncZap to char: ")
3494 The first part of the argument to @code{interactive} is @samp{p}, with
3495 which you are already familiar. This argument tells Emacs to
3496 interpret a prefix, as a number to be passed to the function. You
3497 can specify a prefix either by typing @kbd{C-u} followed by a number
3498 or by typing @key{META} followed by a number. The prefix is the
3499 number of specified characters. Thus, if your prefix is three and the
3500 specified character is @samp{x}, then you will delete all the text up
3501 to and including the third next @samp{x}. If you do not set a prefix,
3502 then you delete all the text up to and including the specified
3503 character, but no more.
3505 The @samp{c} tells the function the name of the character to which to delete.
3507 More formally, a function with two or more arguments can have
3508 information passed to each argument by adding parts to the string that
3509 follows @code{interactive}. When you do this, the information is
3510 passed to each argument in the same order it is specified in the
3511 @code{interactive} list. In the string, each part is separated from
3512 the next part by a @samp{\n}, which is a newline. For example, you
3513 can follow @samp{p} with a @samp{\n} and an @samp{cZap to char:@: }.
3514 This causes Emacs to pass the value of the prefix argument (if there
3515 is one) and the character.
3517 In this case, the function definition looks like the following, where
3518 @code{arg} and @code{char} are the symbols to which @code{interactive}
3519 binds the prefix argument and the specified character:
3523 (defun @var{name-of-function} (arg char)
3524 "@var{documentation}@dots{}"
3525 (interactive "p\ncZap to char: ")
3526 @var{body-of-function}@dots{})
3531 (The space after the colon in the prompt makes it look better when you
3532 are prompted. @xref{copy-to-buffer, , The Definition of
3533 @code{copy-to-buffer}}, for an example.)
3535 When a function does not take arguments, @code{interactive} does not
3536 require any. Such a function contains the simple expression
3537 @code{(interactive)}. The @code{mark-whole-buffer} function is like
3540 Alternatively, if the special letter-codes are not right for your
3541 application, you can pass your own arguments to @code{interactive} as
3544 @xref{append-to-buffer, , The Definition of @code{append-to-buffer}},
3545 for an example. @xref{Using Interactive, , Using @code{Interactive},
3546 elisp, The GNU Emacs Lisp Reference Manual}, for a more complete
3547 explanation about this technique.
3549 @node Permanent Installation
3550 @section Install Code Permanently
3551 @cindex Install code permanently
3552 @cindex Permanent code installation
3553 @cindex Code installation
3555 When you install a function definition by evaluating it, it will stay
3556 installed until you quit Emacs. The next time you start a new session
3557 of Emacs, the function will not be installed unless you evaluate the
3558 function definition again.
3560 At some point, you may want to have code installed automatically
3561 whenever you start a new session of Emacs. There are several ways of
3566 If you have code that is just for yourself, you can put the code for the
3567 function definition in your @file{.emacs} initialization file. When you
3568 start Emacs, your @file{.emacs} file is automatically evaluated and all
3569 the function definitions within it are installed.
3570 @xref{Emacs Initialization, , Your @file{.emacs} File}.
3573 Alternatively, you can put the function definitions that you want
3574 installed in one or more files of their own and use the @code{load}
3575 function to cause Emacs to evaluate and thereby install each of the
3576 functions in the files.
3577 @xref{Loading Files, , Loading Files}.
3580 Thirdly, if you have code that your whole site will use, it is usual
3581 to put it in a file called @file{site-init.el} that is loaded when
3582 Emacs is built. This makes the code available to everyone who uses
3583 your machine. (See the @file{INSTALL} file that is part of the Emacs
3587 Finally, if you have code that everyone who uses Emacs may want, you
3588 can post it on a computer network or send a copy to the Free Software
3589 Foundation. (When you do this, please license the code and its
3590 documentation under a license that permits other people to run, copy,
3591 study, modify, and redistribute the code and which protects you from
3592 having your work taken from you.) If you send a copy of your code to
3593 the Free Software Foundation, and properly protect yourself and
3594 others, it may be included in the next release of Emacs. In large
3595 part, this is how Emacs has grown over the past years, by donations.
3601 The @code{let} expression is a special form in Lisp that you will need
3602 to use in most function definitions.
3604 @code{let} is used to attach or bind a symbol to a value in such a way
3605 that the Lisp interpreter will not confuse the variable with a
3606 variable of the same name that is not part of the function.
3608 To understand why the @code{let} special form is necessary, consider
3609 the situation in which you own a home that you generally refer to as
3610 ``the house'', as in the sentence, ``The house needs painting.'' If you
3611 are visiting a friend and your host refers to ``the house'', he is
3612 likely to be referring to @emph{his} house, not yours, that is, to a
3615 If your friend is referring to his house and you think he is referring
3616 to your house, you may be in for some confusion. The same thing could
3617 happen in Lisp if a variable that is used inside of one function has
3618 the same name as a variable that is used inside of another function,
3619 and the two are not intended to refer to the same value. The
3620 @code{let} special form prevents this kind of confusion.
3623 * Prevent confusion::
3624 * Parts of let Expression::
3625 * Sample let Expression::
3626 * Uninitialized let Variables::
3630 @node Prevent confusion
3631 @unnumberedsubsec @code{let} Prevents Confusion
3634 @cindex @samp{local variable} defined
3635 @cindex @samp{variable, local}, defined
3636 The @code{let} special form prevents confusion. @code{let} creates a
3637 name for a @dfn{local variable} that overshadows any use of the same
3638 name outside the @code{let} expression. This is like understanding
3639 that whenever your host refers to ``the house'', he means his house, not
3640 yours. (Symbols used in argument lists work the same way.
3641 @xref{defun, , The @code{defun} Macro}.)
3643 Local variables created by a @code{let} expression retain their value
3644 @emph{only} within the @code{let} expression itself (and within
3645 expressions called within the @code{let} expression); the local
3646 variables have no effect outside the @code{let} expression.
3648 Another way to think about @code{let} is that it is like a @code{setq}
3649 that is temporary and local. The values set by @code{let} are
3650 automatically undone when the @code{let} is finished. The setting
3651 only affects expressions that are inside the bounds of the @code{let}
3652 expression. In computer science jargon, we would say the binding of
3653 a symbol is visible only in functions called in the @code{let} form;
3654 in Emacs Lisp, scoping is dynamic, not lexical.
3656 @code{let} can create more than one variable at once. Also,
3657 @code{let} gives each variable it creates an initial value, either a
3658 value specified by you, or @code{nil}. (In the jargon, this is
3659 binding the variable to the value.) After @code{let} has created
3660 and bound the variables, it executes the code in the body of the
3661 @code{let}, and returns the value of the last expression in the body,
3662 as the value of the whole @code{let} expression. (``Execute'' is a jargon
3663 term that means to evaluate a list; it comes from the use of the word
3664 meaning ``to give practical effect to'' (@cite{Oxford English
3665 Dictionary}). Since you evaluate an expression to perform an action,
3666 ``execute'' has evolved as a synonym to ``evaluate''.)
3668 @node Parts of let Expression
3669 @subsection The Parts of a @code{let} Expression
3670 @cindex @code{let} expression, parts of
3671 @cindex Parts of @code{let} expression
3673 @cindex @samp{varlist} defined
3674 A @code{let} expression is a list of three parts. The first part is
3675 the symbol @code{let}. The second part is a list, called a
3676 @dfn{varlist}, each element of which is either a symbol by itself or a
3677 two-element list, the first element of which is a symbol. The third
3678 part of the @code{let} expression is the body of the @code{let}. The
3679 body usually consists of one or more lists.
3682 A template for a @code{let} expression looks like this:
3685 (let @var{varlist} @var{body}@dots{})
3689 The symbols in the varlist are the variables that are given initial
3690 values by the @code{let} special form. Symbols by themselves are given
3691 the initial value of @code{nil}; and each symbol that is the first
3692 element of a two-element list is bound to the value that is returned
3693 when the Lisp interpreter evaluates the second element.
3695 Thus, a varlist might look like this: @code{(thread (needles 3))}. In
3696 this case, in a @code{let} expression, Emacs binds the symbol
3697 @code{thread} to an initial value of @code{nil}, and binds the symbol
3698 @code{needles} to an initial value of 3.
3700 When you write a @code{let} expression, what you do is put the
3701 appropriate expressions in the slots of the @code{let} expression
3704 If the varlist is composed of two-element lists, as is often the case,
3705 the template for the @code{let} expression looks like this:
3709 (let ((@var{variable} @var{value})
3710 (@var{variable} @var{value})
3716 @node Sample let Expression
3717 @subsection Sample @code{let} Expression
3718 @cindex Sample @code{let} expression
3719 @cindex @code{let} expression sample
3721 The following expression creates and gives initial values
3722 to the two variables @code{zebra} and @code{tiger}. The body of the
3723 @code{let} expression is a list which calls the @code{message} function.
3727 (let ((zebra "stripes")
3729 (message "One kind of animal has %s and another is %s."
3734 Here, the varlist is @code{((zebra "stripes") (tiger "fierce"))}.
3736 The two variables are @code{zebra} and @code{tiger}. Each variable is
3737 the first element of a two-element list and each value is the second
3738 element of its two-element list. In the varlist, Emacs binds the
3739 variable @code{zebra} to the value @code{"stripes"}@footnote{According
3740 to Jared Diamond in @cite{Guns, Germs, and Steel}, ``@dots{} zebras
3741 become impossibly dangerous as they grow older'' but the claim here is
3742 that they do not become fierce like a tiger. (1997, W. W. Norton and
3743 Co., ISBN 0-393-03894-2, page 171)}, and binds the
3744 variable @code{tiger} to the value @code{"fierce"}. In this example,
3745 both values are strings. The values could just as well have been
3746 another list or a symbol. The body of the @code{let}
3747 follows after the list holding the variables. In this example, the
3748 body is a list that uses the @code{message} function to print a string
3752 You may evaluate the example in the usual fashion, by placing the
3753 cursor after the last parenthesis and typing @kbd{C-x C-e}. When you do
3754 this, the following will appear in the echo area:
3757 "One kind of animal has stripes and another is fierce."
3760 As we have seen before, the @code{message} function prints its first
3761 argument, except for @samp{%s}. In this example, the value of the variable
3762 @code{zebra} is printed at the location of the first @samp{%s} and the
3763 value of the variable @code{tiger} is printed at the location of the
3766 @node Uninitialized let Variables
3767 @subsection Uninitialized Variables in a @code{let} Statement
3768 @cindex Uninitialized @code{let} variables
3769 @cindex @code{let} variables uninitialized
3771 If you do not bind the variables in a @code{let} statement to specific
3772 initial values, they will automatically be bound to an initial value of
3773 @code{nil}, as in the following expression:
3782 "Here are %d variables with %s, %s, and %s value."
3783 birch pine fir oak))
3788 Here, the varlist is @code{((birch 3) pine fir (oak 'some))}.
3791 If you evaluate this expression in the usual way, the following will
3792 appear in your echo area:
3795 "Here are 3 variables with nil, nil, and some value."
3799 In this example, Emacs binds the symbol @code{birch} to the number 3,
3800 binds the symbols @code{pine} and @code{fir} to @code{nil}, and binds
3801 the symbol @code{oak} to the value @code{some}.
3803 Note that in the first part of the @code{let}, the variables @code{pine}
3804 and @code{fir} stand alone as atoms that are not surrounded by
3805 parentheses; this is because they are being bound to @code{nil}, the
3806 empty list. But @code{oak} is bound to @code{some} and so is a part of
3807 the list @code{(oak 'some)}. Similarly, @code{birch} is bound to the
3808 number 3 and so is in a list with that number. (Since a number
3809 evaluates to itself, the number does not need to be quoted. Also, the
3810 number is printed in the message using a @samp{%d} rather than a
3811 @samp{%s}.) The four variables as a group are put into a list to
3812 delimit them from the body of the @code{let}.
3815 @section The @code{if} Special Form
3817 @cindex Conditional with @code{if}
3819 A third special form, in addition to @code{defun} and @code{let}, is the
3820 conditional @code{if}. This form is used to instruct the computer to
3821 make decisions. You can write function definitions without using
3822 @code{if}, but it is used often enough, and is important enough, to be
3823 included here. It is used, for example, in the code for the
3824 function @code{beginning-of-buffer}.
3826 The basic idea behind an @code{if}, is that @emph{if} a test is true,
3827 @emph{then} an expression is evaluated. If the test is not true, the
3828 expression is not evaluated. For example, you might make a decision
3829 such as, ``if it is warm and sunny, then go to the beach!''
3832 * if in more detail::
3833 * type-of-animal in detail:: An example of an @code{if} expression.
3837 @node if in more detail
3838 @unnumberedsubsec @code{if} in more detail
3841 @cindex @samp{if-part} defined
3842 @cindex @samp{then-part} defined
3843 An @code{if} expression written in Lisp does not use the word ``then'';
3844 the test and the action are the second and third elements of the list
3845 whose first element is @code{if}. Nonetheless, the test part of an
3846 @code{if} expression is often called the @dfn{if-part} and the second
3847 argument is often called the @dfn{then-part}.
3849 Also, when an @code{if} expression is written, the true-or-false-test
3850 is usually written on the same line as the symbol @code{if}, but the
3851 action to carry out if the test is true, the then-part, is written
3852 on the second and subsequent lines. This makes the @code{if}
3853 expression easier to read.
3857 (if @var{true-or-false-test}
3858 @var{action-to-carry-out-if-test-is-true})
3863 The true-or-false-test will be an expression that
3864 is evaluated by the Lisp interpreter.
3866 Here is an example that you can evaluate in the usual manner. The test
3867 is whether the number 5 is greater than the number 4. Since it is, the
3868 message @samp{5 is greater than 4!} will be printed.
3872 (if (> 5 4) ; @r{if-part}
3873 (message "5 is greater than 4!")) ; @r{then-part}
3878 (The function @code{>} tests whether its first argument is greater than
3879 its second argument and returns true if it is.)
3880 @findex > @r{(greater than)}
3882 Of course, in actual use, the test in an @code{if} expression will not
3883 be fixed for all time as it is by the expression @code{(> 5 4)}.
3884 Instead, at least one of the variables used in the test will be bound to
3885 a value that is not known ahead of time. (If the value were known ahead
3886 of time, we would not need to run the test!)
3888 For example, the value may be bound to an argument of a function
3889 definition. In the following function definition, the character of the
3890 animal is a value that is passed to the function. If the value bound to
3891 @code{characteristic} is @code{"fierce"}, then the message, @samp{It is a
3892 tiger!} will be printed; otherwise, @code{nil} will be returned.
3896 (defun type-of-animal (characteristic)
3897 "Print message in echo area depending on CHARACTERISTIC.
3898 If the CHARACTERISTIC is the string \"fierce\",
3899 then warn of a tiger."
3900 (if (equal characteristic "fierce")
3901 (message "It is a tiger!")))
3907 If you are reading this inside of GNU Emacs, you can evaluate the
3908 function definition in the usual way to install it in Emacs, and then you
3909 can evaluate the following two expressions to see the results:
3913 (type-of-animal "fierce")
3915 (type-of-animal "striped")
3920 @c Following sentences rewritten to prevent overfull hbox.
3922 When you evaluate @code{(type-of-animal "fierce")}, you will see the
3923 following message printed in the echo area: @code{"It is a tiger!"}; and
3924 when you evaluate @code{(type-of-animal "striped")} you will see @code{nil}
3925 printed in the echo area.
3927 @node type-of-animal in detail
3928 @subsection The @code{type-of-animal} Function in Detail
3930 Let's look at the @code{type-of-animal} function in detail.
3932 The function definition for @code{type-of-animal} was written by filling
3933 the slots of two templates, one for a function definition as a whole, and
3934 a second for an @code{if} expression.
3937 The template for every function that is not interactive is:
3941 (defun @var{name-of-function} (@var{argument-list})
3942 "@var{documentation}@dots{}"
3948 The parts of the function that match this template look like this:
3952 (defun type-of-animal (characteristic)
3953 "Print message in echo area depending on CHARACTERISTIC.
3954 If the CHARACTERISTIC is the string \"fierce\",
3955 then warn of a tiger."
3956 @var{body: the} @code{if} @var{expression})
3960 The name of function is @code{type-of-animal}; it is passed the value
3961 of one argument. The argument list is followed by a multi-line
3962 documentation string. The documentation string is included in the
3963 example because it is a good habit to write documentation string for
3964 every function definition. The body of the function definition
3965 consists of the @code{if} expression.
3968 The template for an @code{if} expression looks like this:
3972 (if @var{true-or-false-test}
3973 @var{action-to-carry-out-if-the-test-returns-true})
3978 In the @code{type-of-animal} function, the code for the @code{if}
3983 (if (equal characteristic "fierce")
3984 (message "It is a tiger!")))
3989 Here, the true-or-false-test is the expression:
3992 (equal characteristic "fierce")
3996 In Lisp, @code{equal} is a function that determines whether its first
3997 argument is equal to its second argument. The second argument is the
3998 string @code{"fierce"} and the first argument is the value of the
3999 symbol @code{characteristic}---in other words, the argument passed to
4002 In the first exercise of @code{type-of-animal}, the argument
4003 @code{"fierce"} is passed to @code{type-of-animal}. Since @code{"fierce"}
4004 is equal to @code{"fierce"}, the expression, @code{(equal characteristic
4005 "fierce")}, returns a value of true. When this happens, the @code{if}
4006 evaluates the second argument or then-part of the @code{if}:
4007 @code{(message "It is a tiger!")}.
4009 On the other hand, in the second exercise of @code{type-of-animal}, the
4010 argument @code{"striped"} is passed to @code{type-of-animal}. @code{"striped"}
4011 is not equal to @code{"fierce"}, so the then-part is not evaluated and
4012 @code{nil} is returned by the @code{if} expression.
4015 @section If--then--else Expressions
4018 An @code{if} expression may have an optional third argument, called
4019 the @dfn{else-part}, for the case when the true-or-false-test returns
4020 false. When this happens, the second argument or then-part of the
4021 overall @code{if} expression is @emph{not} evaluated, but the third or
4022 else-part @emph{is} evaluated. You might think of this as the cloudy
4023 day alternative for the decision ``if it is warm and sunny, then go to
4024 the beach, else read a book!''.
4026 The word ``else'' is not written in the Lisp code; the else-part of an
4027 @code{if} expression comes after the then-part. In the written Lisp, the
4028 else-part is usually written to start on a line of its own and is
4029 indented less than the then-part:
4033 (if @var{true-or-false-test}
4034 @var{action-to-carry-out-if-the-test-returns-true}
4035 @var{action-to-carry-out-if-the-test-returns-false})
4039 For example, the following @code{if} expression prints the message @samp{4
4040 is not greater than 5!} when you evaluate it in the usual way:
4044 (if (> 4 5) ; @r{if-part}
4045 (message "4 falsely greater than 5!") ; @r{then-part}
4046 (message "4 is not greater than 5!")) ; @r{else-part}
4051 Note that the different levels of indentation make it easy to
4052 distinguish the then-part from the else-part. (GNU Emacs has several
4053 commands that automatically indent @code{if} expressions correctly.
4054 @xref{Typing Lists, , GNU Emacs Helps You Type Lists}.)
4056 We can extend the @code{type-of-animal} function to include an
4057 else-part by simply incorporating an additional part to the @code{if}
4061 You can see the consequences of doing this if you evaluate the following
4062 version of the @code{type-of-animal} function definition to install it
4063 and then evaluate the two subsequent expressions to pass different
4064 arguments to the function.
4068 (defun type-of-animal (characteristic) ; @r{Second version.}
4069 "Print message in echo area depending on CHARACTERISTIC.
4070 If the CHARACTERISTIC is the string \"fierce\",
4071 then warn of a tiger; else say it is not fierce."
4072 (if (equal characteristic "fierce")
4073 (message "It is a tiger!")
4074 (message "It is not fierce!")))
4081 (type-of-animal "fierce")
4083 (type-of-animal "striped")
4088 @c Following sentence rewritten to prevent overfull hbox.
4090 When you evaluate @code{(type-of-animal "fierce")}, you will see the
4091 following message printed in the echo area: @code{"It is a tiger!"}; but
4092 when you evaluate @code{(type-of-animal "striped")}, you will see
4093 @code{"It is not fierce!"}.
4095 (Of course, if the @var{characteristic} were @code{"ferocious"}, the
4096 message @code{"It is not fierce!"} would be printed; and it would be
4097 misleading! When you write code, you need to take into account the
4098 possibility that some such argument will be tested by the @code{if}
4099 and write your program accordingly.)
4101 @node Truth & Falsehood
4102 @section Truth and Falsehood in Emacs Lisp
4103 @cindex Truth and falsehood in Emacs Lisp
4104 @cindex Falsehood and truth in Emacs Lisp
4107 There is an important aspect to the truth test in an @code{if}
4108 expression. So far, we have spoken of ``true'' and ``false'' as values of
4109 predicates as if they were new kinds of Emacs Lisp objects. In fact,
4110 ``false'' is just our old friend @code{nil}. Anything else---anything
4111 at all---is ``true''.
4113 The expression that tests for truth is interpreted as @dfn{true}
4114 if the result of evaluating it is a value that is not @code{nil}. In
4115 other words, the result of the test is considered true if the value
4116 returned is a number such as 47, a string such as @code{"hello"}, or a
4117 symbol (other than @code{nil}) such as @code{flowers}, or a list (so
4118 long as it is not empty), or even a buffer!
4121 * nil explained:: @code{nil} has two meanings.
4126 @unnumberedsubsec An explanation of @code{nil}
4129 Before illustrating a test for truth, we need an explanation of @code{nil}.
4131 In Emacs Lisp, the symbol @code{nil} has two meanings. First, it means the
4132 empty list. Second, it means false and is the value returned when a
4133 true-or-false-test tests false. @code{nil} can be written as an empty
4134 list, @code{()}, or as @code{nil}. As far as the Lisp interpreter is
4135 concerned, @code{()} and @code{nil} are the same. Humans, however, tend
4136 to use @code{nil} for false and @code{()} for the empty list.
4138 In Emacs Lisp, any value that is not @code{nil}---is not the empty
4139 list---is considered true. This means that if an evaluation returns
4140 something that is not an empty list, an @code{if} expression will test
4141 true. For example, if a number is put in the slot for the test, it
4142 will be evaluated and will return itself, since that is what numbers
4143 do when evaluated. In this conditional, the @code{if} expression will
4144 test true. The expression tests false only when @code{nil}, an empty
4145 list, is returned by evaluating the expression.
4147 You can see this by evaluating the two expressions in the following examples.
4149 In the first example, the number 4 is evaluated as the test in the
4150 @code{if} expression and returns itself; consequently, the then-part
4151 of the expression is evaluated and returned: @samp{true} appears in
4152 the echo area. In the second example, the @code{nil} indicates false;
4153 consequently, the else-part of the expression is evaluated and
4154 returned: @samp{false} appears in the echo area.
4171 Incidentally, if some other useful value is not available for a test that
4172 returns true, then the Lisp interpreter will return the symbol @code{t}
4173 for true. For example, the expression @code{(> 5 4)} returns @code{t}
4174 when evaluated, as you can see by evaluating it in the usual way:
4182 On the other hand, this function returns @code{nil} if the test is false.
4188 @node save-excursion
4189 @section @code{save-excursion}
4190 @findex save-excursion
4191 @cindex Region, what it is
4192 @cindex Preserving point and buffer
4193 @cindex Point and buffer preservation
4197 The @code{save-excursion} function is the third and final special form
4198 that we will discuss in this chapter.
4200 In Emacs Lisp programs used for editing, the @code{save-excursion}
4201 function is very common. It saves the location of point,
4202 executes the body of the function, and then restores point to
4203 its previous position if its location was changed. Its primary
4204 purpose is to keep the user from being surprised and disturbed by
4205 unexpected movement of point.
4208 * Point and mark:: A review of various locations.
4209 * Template for save-excursion::
4213 @node Point and mark
4214 @unnumberedsubsec Point and Mark
4217 Before discussing @code{save-excursion}, however, it may be useful
4218 first to review what point and mark are in GNU Emacs. @dfn{Point} is
4219 the current location of the cursor. Wherever the cursor
4220 is, that is point. More precisely, on terminals where the cursor
4221 appears to be on top of a character, point is immediately before the
4222 character. In Emacs Lisp, point is an integer. The first character in
4223 a buffer is number one, the second is number two, and so on. The
4224 function @code{point} returns the current position of the cursor as a
4225 number. Each buffer has its own value for point.
4227 The @dfn{mark} is another position in the buffer; its value can be set
4228 with a command such as @kbd{C-@key{SPC}} (@code{set-mark-command}). If
4229 a mark has been set, you can use the command @kbd{C-x C-x}
4230 (@code{exchange-point-and-mark}) to cause the cursor to jump to the mark
4231 and set the mark to be the previous position of point. In addition, if
4232 you set another mark, the position of the previous mark is saved in the
4233 mark ring. Many mark positions can be saved this way. You can jump the
4234 cursor to a saved mark by typing @kbd{C-u C-@key{SPC}} one or more
4237 The part of the buffer between point and mark is called @dfn{the
4238 region}. Numerous commands work on the region, including
4239 @code{center-region}, @code{count-lines-region}, @code{kill-region}, and
4240 @code{print-region}.
4242 The @code{save-excursion} special form saves the location of point and
4243 restores this position after the code within the body of the
4244 special form is evaluated by the Lisp interpreter. Thus, if point were
4245 in the beginning of a piece of text and some code moved point to the end
4246 of the buffer, the @code{save-excursion} would put point back to where
4247 it was before, after the expressions in the body of the function were
4250 In Emacs, a function frequently moves point as part of its internal
4251 workings even though a user would not expect this. For example,
4252 @code{count-lines-region} moves point. To prevent the user from being
4253 bothered by jumps that are both unexpected and (from the user's point of
4254 view) unnecessary, @code{save-excursion} is often used to keep point in
4255 the location expected by the user. The use of
4256 @code{save-excursion} is good housekeeping.
4258 To make sure the house stays clean, @code{save-excursion} restores the
4259 value of point even if something goes wrong in the code inside
4260 of it (or, to be more precise and to use the proper jargon, ``in case of
4261 abnormal exit''). This feature is very helpful.
4263 In addition to recording the value of point,
4264 @code{save-excursion} keeps track of the current buffer, and restores
4265 it, too. This means you can write code that will change the buffer and
4266 have @code{save-excursion} switch you back to the original buffer.
4267 This is how @code{save-excursion} is used in @code{append-to-buffer}.
4268 (@xref{append-to-buffer, , The Definition of @code{append-to-buffer}}.)
4270 @node Template for save-excursion
4271 @subsection Template for a @code{save-excursion} Expression
4274 The template for code using @code{save-excursion} is simple:
4284 The body of the function is one or more expressions that will be
4285 evaluated in sequence by the Lisp interpreter. If there is more than
4286 one expression in the body, the value of the last one will be returned
4287 as the value of the @code{save-excursion} function. The other
4288 expressions in the body are evaluated only for their side effects; and
4289 @code{save-excursion} itself is used only for its side effect (which
4290 is restoring the position of point).
4293 In more detail, the template for a @code{save-excursion} expression
4299 @var{first-expression-in-body}
4300 @var{second-expression-in-body}
4301 @var{third-expression-in-body}
4303 @var{last-expression-in-body})
4308 An expression, of course, may be a symbol on its own or a list.
4310 In Emacs Lisp code, a @code{save-excursion} expression often occurs
4311 within the body of a @code{let} expression. It looks like this:
4324 In the last few chapters we have introduced a macro and a fair number
4325 of functions and special forms. Here they are described in brief,
4326 along with a few similar functions that have not been mentioned yet.
4329 @item eval-last-sexp
4330 Evaluate the last symbolic expression before the current location of
4331 point. The value is printed in the echo area unless the function is
4332 invoked with an argument; in that case, the output is printed in the
4333 current buffer. This command is normally bound to @kbd{C-x C-e}.
4336 Define function. This macro has up to five parts: the name, a
4337 template for the arguments that will be passed to the function,
4338 documentation, an optional interactive declaration, and the body of
4342 For example, in Emacs the function definition of
4343 @code{dired-unmark-all-marks} is as follows.
4347 (defun dired-unmark-all-marks ()
4348 "Remove all marks from all files in the Dired buffer."
4350 (dired-unmark-all-files ?\r))
4355 Declare to the interpreter that the function can be used
4356 interactively. This special form may be followed by a string with one
4357 or more parts that pass the information to the arguments of the
4358 function, in sequence. These parts may also tell the interpreter to
4359 prompt for information. Parts of the string are separated by
4360 newlines, @samp{\n}.
4363 Common code characters are:
4367 The name of an existing buffer.
4370 The name of an existing file.
4373 The numeric prefix argument. (Note that this @code{p} is lower case.)
4376 Point and the mark, as two numeric arguments, smallest first. This
4377 is the only code letter that specifies two successive arguments
4381 @xref{Interactive Codes, , Code Characters for @samp{interactive},
4382 elisp, The GNU Emacs Lisp Reference Manual}, for a complete list of
4386 Declare that a list of variables is for use within the body of the
4387 @code{let} and give them an initial value, either @code{nil} or a
4388 specified value; then evaluate the rest of the expressions in the body
4389 of the @code{let} and return the value of the last one. Inside the
4390 body of the @code{let}, the Lisp interpreter does not see the values of
4391 the variables of the same names that are bound outside of the
4399 (let ((foo (buffer-name))
4400 (bar (buffer-size)))
4402 "This buffer is %s and has %d characters."
4407 @item save-excursion
4408 Record the values of point and the current buffer before
4409 evaluating the body of this special form. Restore the value of point and
4417 (message "We are %d characters into this buffer."
4420 (goto-char (point-min)) (point))))
4425 Evaluate the first argument to the function; if it is true, evaluate
4426 the second argument; else evaluate the third argument, if there is one.
4428 The @code{if} special form is called a @dfn{conditional}. There are
4429 other conditionals in Emacs Lisp, but @code{if} is perhaps the most
4437 (if (= 22 emacs-major-version)
4438 (message "This is version 22 Emacs")
4439 (message "This is not version 22 Emacs"))
4448 The @code{<} function tests whether its first argument is smaller than
4449 its second argument. A corresponding function, @code{>}, tests whether
4450 the first argument is greater than the second. Likewise, @code{<=}
4451 tests whether the first argument is less than or equal to the second and
4452 @code{>=} tests whether the first argument is greater than or equal to
4453 the second. In all cases, both arguments must be numbers or markers
4454 (markers indicate positions in buffers).
4458 The @code{=} function tests whether two arguments, both numbers or
4464 Test whether two objects are the same. @code{equal} uses one meaning
4465 of the word ``same'' and @code{eq} uses another: @code{equal} returns
4466 true if the two objects have a similar structure and contents, such as
4467 two copies of the same book. On the other hand, @code{eq}, returns
4468 true if both arguments are actually the same object.
4477 The @code{string-lessp} function tests whether its first argument is
4478 smaller than the second argument. A shorter, alternative name for the
4479 same function (a @code{defalias}) is @code{string<}.
4481 The arguments to @code{string-lessp} must be strings or symbols; the
4482 ordering is lexicographic, so case is significant. The print names of
4483 symbols are used instead of the symbols themselves.
4485 @cindex @samp{empty string} defined
4486 An empty string, @samp{""}, a string with no characters in it, is
4487 smaller than any string of characters.
4489 @code{string-equal} provides the corresponding test for equality. Its
4490 shorter, alternative name is @code{string=}. There are no string test
4491 functions that correspond to @var{>}, @code{>=}, or @code{<=}.
4494 Print a message in the echo area. The first argument is a string that
4495 can contain @samp{%s}, @samp{%d}, or @samp{%c} to print the value of
4496 arguments that follow the string. The argument used by @samp{%s} must
4497 be a string or a symbol; the argument used by @samp{%d} must be a
4498 number. The argument used by @samp{%c} must be an @sc{ascii} code
4499 number; it will be printed as the character with that @sc{ascii} code.
4500 (Various other %-sequences have not been mentioned.)
4504 The @code{setq} function sets the value of its first argument to the
4505 value of the second argument. The first argument is automatically
4506 quoted by @code{setq}. It does the same for succeeding pairs of
4507 arguments. Another function, @code{set}, takes only two arguments and
4508 evaluates both of them before setting the value returned by its first
4509 argument to the value returned by its second argument.
4512 Without an argument, return the name of the buffer, as a string.
4514 @item buffer-file-name
4515 Without an argument, return the name of the file the buffer is
4518 @item current-buffer
4519 Return the buffer in which Emacs is active; it may not be
4520 the buffer that is visible on the screen.
4523 Return the most recently selected buffer (other than the buffer passed
4524 to @code{other-buffer} as an argument and other than the current
4527 @item switch-to-buffer
4528 Select a buffer for Emacs to be active in and display it in the current
4529 window so users can look at it. Usually bound to @kbd{C-x b}.
4532 Switch Emacs's attention to a buffer on which programs will run. Don't
4533 alter what the window is showing.
4536 Return the number of characters in the current buffer.
4539 Return the value of the current position of the cursor, as an
4540 integer counting the number of characters from the beginning of the
4544 Return the minimum permissible value of point in
4545 the current buffer. This is 1, unless narrowing is in effect.
4548 Return the value of the maximum permissible value of point in the
4549 current buffer. This is the end of the buffer, unless narrowing is in
4554 @node defun Exercises
4559 Write a non-interactive function that doubles the value of its
4560 argument, a number. Make that function interactive.
4563 Write a function that tests whether the current value of
4564 @code{fill-column} is greater than the argument passed to the function,
4565 and if so, prints an appropriate message.
4568 @node Buffer Walk Through
4569 @chapter A Few Buffer-Related Functions
4571 In this chapter we study in detail several of the functions used in GNU
4572 Emacs. This is called a ``walk-through''. These functions are used as
4573 examples of Lisp code, but are not imaginary examples; with the
4574 exception of the first, simplified function definition, these functions
4575 show the actual code used in GNU Emacs. You can learn a great deal from
4576 these definitions. The functions described here are all related to
4577 buffers. Later, we will study other functions.
4580 * Finding More:: How to find more information.
4581 * simplified-beginning-of-buffer:: Shows @code{goto-char},
4582 @code{point-min}, and @code{push-mark}.
4583 * mark-whole-buffer:: Almost the same as @code{beginning-of-buffer}.
4584 * append-to-buffer:: Uses @code{save-excursion} and
4585 @code{insert-buffer-substring}.
4586 * Buffer Related Review:: Review.
4587 * Buffer Exercises::
4591 @section Finding More Information
4593 @findex describe-function@r{, introduced}
4594 @cindex Find function documentation
4595 In this walk-through, I will describe each new function as we come to
4596 it, sometimes in detail and sometimes briefly. If you are interested,
4597 you can get the full documentation of any Emacs Lisp function at any
4598 time by typing @kbd{C-h f} and then the name of the function (and then
4599 @key{RET}). Similarly, you can get the full documentation for a
4600 variable by typing @kbd{C-h v} and then the name of the variable (and
4603 @cindex Find source of function
4604 @c In version 22, tells location both of C and of Emacs Lisp
4605 Also, @code{describe-function} will tell you the location of the
4606 function definition.
4608 Put point into the name of the file that contains the function and
4609 press the @key{RET} key. In this case, @key{RET} means
4610 @code{push-button} rather than ``return'' or ``enter''. Emacs will take
4611 you directly to the function definition.
4616 If you move point over the file name and press
4617 the @key{RET} key, which in this case means @code{help-follow} rather
4618 than ``return'' or ``enter'', Emacs will take you directly to the function
4622 More generally, if you want to see a function in its original source
4623 file, you can use the @code{xref-find-definitions} function to jump to
4624 it. @code{xref-find-definitions} works with a wide variety of
4625 languages, not just Lisp, and C, and it works with non-programming
4626 text as well. For example, @code{xref-find-definitions} will jump to
4627 the various nodes in the Texinfo source file of this document.
4629 To use the @code{xref-find-definitions} command, type @kbd{M-.}
4630 (i.e., press the period key while holding down the @key{META} key, or
4631 else type the @key{ESC} key and then type the period key), and then,
4632 at the prompt, type in the name of the function whose source code you
4633 want to see, such as @code{mark-whole-buffer}, and then type
4634 @key{RET}. Emacs will switch buffers and display the source code for
4635 the function on your screen. To switch back to your current buffer,
4636 type @kbd{C-x b @key{RET}}. (On some keyboards, the @key{META} key is
4639 @cindex Library, as term for ``file''
4640 Incidentally, the files that contain Lisp code are conventionally
4641 called @dfn{libraries}. The metaphor is derived from that of a
4642 specialized library, such as a law library or an engineering library,
4643 rather than a general library. Each library, or file, contains
4644 functions that relate to a particular topic or activity, such as
4645 @file{abbrev.el} for handling abbreviations and other typing
4646 shortcuts, and @file{help.el} for help. (Sometimes several
4647 libraries provide code for a single activity, as the various
4648 @file{rmail@dots{}} files provide code for reading electronic mail.)
4649 In @cite{The GNU Emacs Manual}, you will see sentences such as ``The
4650 @kbd{C-h p} command lets you search the standard Emacs Lisp libraries
4651 by topic keywords.''
4653 @node simplified-beginning-of-buffer
4654 @section A Simplified @code{beginning-of-buffer} Definition
4655 @findex simplified-beginning-of-buffer
4657 The @code{beginning-of-buffer} command is a good function to start with
4658 since you are likely to be familiar with it and it is easy to
4659 understand. Used as an interactive command, @code{beginning-of-buffer}
4660 moves the cursor to the beginning of the buffer, leaving the mark at the
4661 previous position. It is generally bound to @kbd{M-<}.
4663 In this section, we will discuss a shortened version of the function
4664 that shows how it is most frequently used. This shortened function
4665 works as written, but it does not contain the code for a complex option.
4666 In another section, we will describe the entire function.
4667 (@xref{beginning-of-buffer, , Complete Definition of
4668 @code{beginning-of-buffer}}.)
4670 Before looking at the code, let's consider what the function
4671 definition has to contain: it must include an expression that makes
4672 the function interactive so it can be called by typing @kbd{M-x
4673 beginning-of-buffer} or by typing a keychord such as @kbd{M-<}; it
4674 must include code to leave a mark at the original position in the
4675 buffer; and it must include code to move the cursor to the beginning
4679 Here is the complete text of the shortened version of the function:
4683 (defun simplified-beginning-of-buffer ()
4684 "Move point to the beginning of the buffer;
4685 leave mark at previous position."
4688 (goto-char (point-min)))
4692 Like all function definitions, this definition has five parts following
4693 the macro @code{defun}:
4697 The name: in this example, @code{simplified-beginning-of-buffer}.
4700 A list of the arguments: in this example, an empty list, @code{()},
4703 The documentation string.
4706 The interactive expression.
4713 In this function definition, the argument list is empty; this means that
4714 this function does not require any arguments. (When we look at the
4715 definition for the complete function, we will see that it may be passed
4716 an optional argument.)
4718 The interactive expression tells Emacs that the function is intended to
4719 be used interactively. In this example, @code{interactive} does not have
4720 an argument because @code{simplified-beginning-of-buffer} does not
4724 The body of the function consists of the two lines:
4729 (goto-char (point-min))
4733 The first of these lines is the expression, @code{(push-mark)}. When
4734 this expression is evaluated by the Lisp interpreter, it sets a mark at
4735 the current position of the cursor, wherever that may be. The position
4736 of this mark is saved in the mark ring.
4738 The next line is @code{(goto-char (point-min))}. This expression
4739 jumps the cursor to the minimum point in the buffer, that is, to the
4740 beginning of the buffer (or to the beginning of the accessible portion
4741 of the buffer if it is narrowed. @xref{Narrowing & Widening, ,
4742 Narrowing and Widening}.)
4744 The @code{push-mark} command sets a mark at the place where the cursor
4745 was located before it was moved to the beginning of the buffer by the
4746 @code{(goto-char (point-min))} expression. Consequently, you can, if
4747 you wish, go back to where you were originally by typing @kbd{C-x C-x}.
4749 That is all there is to the function definition!
4751 @findex describe-function
4752 When you are reading code such as this and come upon an unfamiliar
4753 function, such as @code{goto-char}, you can find out what it does by
4754 using the @code{describe-function} command. To use this command, type
4755 @kbd{C-h f} and then type in the name of the function and press
4756 @key{RET}. The @code{describe-function} command will print the
4757 function's documentation string in a @file{*Help*} window. For
4758 example, the documentation for @code{goto-char} is:
4762 Set point to POSITION, a number or marker.
4763 Beginning of buffer is position (point-min), end is (point-max).
4768 The function's one argument is the desired position.
4771 (The prompt for @code{describe-function} will offer you the symbol
4772 under or preceding the cursor, so you can save typing by positioning
4773 the cursor right over or after the function and then typing @kbd{C-h f
4776 The @code{end-of-buffer} function definition is written in the same way as
4777 the @code{beginning-of-buffer} definition except that the body of the
4778 function contains the expression @code{(goto-char (point-max))} in place
4779 of @code{(goto-char (point-min))}.
4781 @node mark-whole-buffer
4782 @section The Definition of @code{mark-whole-buffer}
4783 @findex mark-whole-buffer
4785 The @code{mark-whole-buffer} function is no harder to understand than the
4786 @code{simplified-beginning-of-buffer} function. In this case, however,
4787 we will look at the complete function, not a shortened version.
4789 The @code{mark-whole-buffer} function is not as commonly used as the
4790 @code{beginning-of-buffer} function, but is useful nonetheless: it
4791 marks a whole buffer as a region by putting point at the beginning and
4792 a mark at the end of the buffer. It is generally bound to @kbd{C-x
4796 * mark-whole-buffer overview::
4797 * Body of mark-whole-buffer:: Only three lines of code.
4801 @node mark-whole-buffer overview
4802 @unnumberedsubsec An overview of @code{mark-whole-buffer}
4806 In GNU Emacs 22, the code for the complete function looks like this:
4810 (defun mark-whole-buffer ()
4811 "Put point at beginning and mark at end of buffer.
4812 You probably should not use this function in Lisp programs;
4813 it is usually a mistake for a Lisp function to use any subroutine
4814 that uses or sets the mark."
4817 (push-mark (point-max) nil t)
4818 (goto-char (point-min)))
4823 Like all other functions, the @code{mark-whole-buffer} function fits
4824 into the template for a function definition. The template looks like
4829 (defun @var{name-of-function} (@var{argument-list})
4830 "@var{documentation}@dots{}"
4831 (@var{interactive-expression}@dots{})
4836 Here is how the function works: the name of the function is
4837 @code{mark-whole-buffer}; it is followed by an empty argument list,
4838 @samp{()}, which means that the function does not require arguments.
4839 The documentation comes next.
4841 The next line is an @code{(interactive)} expression that tells Emacs
4842 that the function will be used interactively. These details are similar
4843 to the @code{simplified-beginning-of-buffer} function described in the
4847 @node Body of mark-whole-buffer
4848 @subsection Body of @code{mark-whole-buffer}
4850 The body of the @code{mark-whole-buffer} function consists of three
4857 (push-mark (point-max) nil t)
4858 (goto-char (point-min))
4862 The first of these lines is the expression, @code{(push-mark (point))}.
4864 This line does exactly the same job as the first line of the body of
4865 the @code{simplified-beginning-of-buffer} function, which is written
4866 @code{(push-mark)}. In both cases, the Lisp interpreter sets a mark
4867 at the current position of the cursor.
4869 I don't know why the expression in @code{mark-whole-buffer} is written
4870 @code{(push-mark (point))} and the expression in
4871 @code{beginning-of-buffer} is written @code{(push-mark)}. Perhaps
4872 whoever wrote the code did not know that the arguments for
4873 @code{push-mark} are optional and that if @code{push-mark} is not
4874 passed an argument, the function automatically sets mark at the
4875 location of point by default. Or perhaps the expression was written
4876 so as to parallel the structure of the next line. In any case, the
4877 line causes Emacs to determine the position of point and set a mark
4880 In earlier versions of GNU Emacs, the next line of
4881 @code{mark-whole-buffer} was @code{(push-mark (point-max))}. This
4882 expression sets a mark at the point in the buffer that has the highest
4883 number. This will be the end of the buffer (or, if the buffer is
4884 narrowed, the end of the accessible portion of the buffer.
4885 @xref{Narrowing & Widening, , Narrowing and Widening}, for more about
4886 narrowing.) After this mark has been set, the previous mark, the one
4887 set at point, is no longer set, but Emacs remembers its position, just
4888 as all other recent marks are always remembered. This means that you
4889 can, if you wish, go back to that position by typing @kbd{C-u
4893 In GNU Emacs 22, the @code{(point-max)} is slightly more complicated.
4897 (push-mark (point-max) nil t)
4901 The expression works nearly the same as before. It sets a mark at the
4902 highest numbered place in the buffer that it can. However, in this
4903 version, @code{push-mark} has two additional arguments. The second
4904 argument to @code{push-mark} is @code{nil}. This tells the function
4905 it @emph{should} display a message that says ``Mark set'' when it pushes
4906 the mark. The third argument is @code{t}. This tells
4907 @code{push-mark} to activate the mark when Transient Mark mode is
4908 turned on. Transient Mark mode highlights the currently active
4909 region. It is often turned off.
4911 Finally, the last line of the function is @code{(goto-char
4912 (point-min)))}. This is written exactly the same way as it is written
4913 in @code{beginning-of-buffer}. The expression moves the cursor to
4914 the minimum point in the buffer, that is, to the beginning of the buffer
4915 (or to the beginning of the accessible portion of the buffer). As a
4916 result of this, point is placed at the beginning of the buffer and mark
4917 is set at the end of the buffer. The whole buffer is, therefore, the
4920 @c FIXME: the definition of append-to-buffer has been changed (in
4922 @node append-to-buffer
4923 @section The Definition of @code{append-to-buffer}
4924 @findex append-to-buffer
4926 The @code{append-to-buffer} command is more complex than the
4927 @code{mark-whole-buffer} command. What it does is copy the region
4928 (that is, the part of the buffer between point and mark) from the
4929 current buffer to a specified buffer.
4932 * append-to-buffer overview::
4933 * append interactive:: A two part interactive expression.
4934 * append-to-buffer body:: Incorporates a @code{let} expression.
4935 * append save-excursion:: How the @code{save-excursion} works.
4939 @node append-to-buffer overview
4940 @unnumberedsubsec An Overview of @code{append-to-buffer}
4943 @findex insert-buffer-substring
4944 The @code{append-to-buffer} command uses the
4945 @code{insert-buffer-substring} function to copy the region.
4946 @code{insert-buffer-substring} is described by its name: it takes a
4947 substring from a buffer, and inserts it into another buffer.
4949 Most of @code{append-to-buffer} is
4950 concerned with setting up the conditions for
4951 @code{insert-buffer-substring} to work: the code must specify both the
4952 buffer to which the text will go, the window it comes from and goes
4953 to, and the region that will be copied.
4956 Here is the complete text of the function:
4960 (defun append-to-buffer (buffer start end)
4961 "Append to specified buffer the text of the region.
4962 It is inserted into that buffer before its point.
4966 When calling from a program, give three arguments:
4967 BUFFER (or buffer name), START and END.
4968 START and END specify the portion of the current buffer to be copied."
4970 (list (read-buffer "Append to buffer: " (other-buffer
4971 (current-buffer) t))
4972 (region-beginning) (region-end)))
4975 (let ((oldbuf (current-buffer)))
4977 (let* ((append-to (get-buffer-create buffer))
4978 (windows (get-buffer-window-list append-to t t))
4980 (set-buffer append-to)
4981 (setq point (point))
4982 (barf-if-buffer-read-only)
4983 (insert-buffer-substring oldbuf start end)
4984 (dolist (window windows)
4985 (when (= (window-point window) point)
4986 (set-window-point window (point))))))))
4990 The function can be understood by looking at it as a series of
4991 filled-in templates.
4993 The outermost template is for the function definition. In this
4994 function, it looks like this (with several slots filled in):
4998 (defun append-to-buffer (buffer start end)
4999 "@var{documentation}@dots{}"
5000 (interactive @dots{})
5005 The first line of the function includes its name and three arguments.
5006 The arguments are the @code{buffer} to which the text will be copied, and
5007 the @code{start} and @code{end} of the region in the current buffer that
5010 The next part of the function is the documentation, which is clear and
5011 complete. As is conventional, the three arguments are written in
5012 upper case so you will notice them easily. Even better, they are
5013 described in the same order as in the argument list.
5015 Note that the documentation distinguishes between a buffer and its
5016 name. (The function can handle either.)
5018 @node append interactive
5019 @subsection The @code{append-to-buffer} Interactive Expression
5021 Since the @code{append-to-buffer} function will be used interactively,
5022 the function must have an @code{interactive} expression. (For a
5023 review of @code{interactive}, see @ref{Interactive, , Making a
5024 Function Interactive}.) The expression reads as follows:
5030 "Append to buffer: "
5031 (other-buffer (current-buffer) t))
5038 This expression is not one with letters standing for parts, as
5039 described earlier. Instead, it starts a list with these parts:
5041 The first part of the list is an expression to read the name of a
5042 buffer and return it as a string. That is @code{read-buffer}. The
5043 function requires a prompt as its first argument, @samp{"Append to
5044 buffer: "}. Its second argument tells the command what value to
5045 provide if you don't specify anything.
5047 In this case that second argument is an expression containing the
5048 function @code{other-buffer}, an exception, and a @samp{t}, standing
5051 The first argument to @code{other-buffer}, the exception, is yet
5052 another function, @code{current-buffer}. That is not going to be
5053 returned. The second argument is the symbol for true, @code{t}. that
5054 tells @code{other-buffer} that it may show visible buffers (except in
5055 this case, it will not show the current buffer, which makes sense).
5058 The expression looks like this:
5061 (other-buffer (current-buffer) t)
5064 The second and third arguments to the @code{list} expression are
5065 @code{(region-beginning)} and @code{(region-end)}. These two
5066 functions specify the beginning and end of the text to be appended.
5069 Originally, the command used the letters @samp{B} and @samp{r}.
5070 The whole @code{interactive} expression looked like this:
5073 (interactive "BAppend to buffer:@: \nr")
5077 But when that was done, the default value of the buffer switched to
5078 was invisible. That was not wanted.
5080 (The prompt was separated from the second argument with a newline,
5081 @samp{\n}. It was followed by an @samp{r} that told Emacs to bind the
5082 two arguments that follow the symbol @code{buffer} in the function's
5083 argument list (that is, @code{start} and @code{end}) to the values of
5084 point and mark. That argument worked fine.)
5086 @node append-to-buffer body
5087 @subsection The Body of @code{append-to-buffer}
5090 in GNU Emacs 22 in /usr/local/src/emacs/lisp/simple.el
5092 (defun append-to-buffer (buffer start end)
5093 "Append to specified buffer the text of the region.
5094 It is inserted into that buffer before its point.
5096 When calling from a program, give three arguments:
5097 BUFFER (or buffer name), START and END.
5098 START and END specify the portion of the current buffer to be copied."
5100 (list (read-buffer "Append to buffer: " (other-buffer (current-buffer) t))
5101 (region-beginning) (region-end)))
5102 (let ((oldbuf (current-buffer)))
5104 (let* ((append-to (get-buffer-create buffer))
5105 (windows (get-buffer-window-list append-to t t))
5107 (set-buffer append-to)
5108 (setq point (point))
5109 (barf-if-buffer-read-only)
5110 (insert-buffer-substring oldbuf start end)
5111 (dolist (window windows)
5112 (when (= (window-point window) point)
5113 (set-window-point window (point))))))))
5116 The body of the @code{append-to-buffer} function begins with @code{let}.
5118 As we have seen before (@pxref{let, , @code{let}}), the purpose of a
5119 @code{let} expression is to create and give initial values to one or
5120 more variables that will only be used within the body of the
5121 @code{let}. This means that such a variable will not be confused with
5122 any variable of the same name outside the @code{let} expression.
5124 We can see how the @code{let} expression fits into the function as a
5125 whole by showing a template for @code{append-to-buffer} with the
5126 @code{let} expression in outline:
5130 (defun append-to-buffer (buffer start end)
5131 "@var{documentation}@dots{}"
5132 (interactive @dots{})
5133 (let ((@var{variable} @var{value}))
5138 The @code{let} expression has three elements:
5142 The symbol @code{let};
5145 A varlist containing, in this case, a single two-element list,
5146 @code{(@var{variable} @var{value})};
5149 The body of the @code{let} expression.
5153 In the @code{append-to-buffer} function, the varlist looks like this:
5156 (oldbuf (current-buffer))
5160 In this part of the @code{let} expression, the one variable,
5161 @code{oldbuf}, is bound to the value returned by the
5162 @code{(current-buffer)} expression. The variable, @code{oldbuf}, is
5163 used to keep track of the buffer in which you are working and from
5164 which you will copy.
5166 The element or elements of a varlist are surrounded by a set of
5167 parentheses so the Lisp interpreter can distinguish the varlist from
5168 the body of the @code{let}. As a consequence, the two-element list
5169 within the varlist is surrounded by a circumscribing set of parentheses.
5170 The line looks like this:
5174 (let ((oldbuf (current-buffer)))
5180 The two parentheses before @code{oldbuf} might surprise you if you did
5181 not realize that the first parenthesis before @code{oldbuf} marks the
5182 boundary of the varlist and the second parenthesis marks the beginning
5183 of the two-element list, @code{(oldbuf (current-buffer))}.
5185 @node append save-excursion
5186 @subsection @code{save-excursion} in @code{append-to-buffer}
5188 The body of the @code{let} expression in @code{append-to-buffer}
5189 consists of a @code{save-excursion} expression.
5191 The @code{save-excursion} function saves the location of point, and restores it
5192 to that position after the expressions in the
5193 body of the @code{save-excursion} complete execution. In addition,
5194 @code{save-excursion} keeps track of the original buffer, and
5195 restores it. This is how @code{save-excursion} is used in
5196 @code{append-to-buffer}.
5199 @cindex Indentation for formatting
5200 @cindex Formatting convention
5201 Incidentally, it is worth noting here that a Lisp function is normally
5202 formatted so that everything that is enclosed in a multi-line spread is
5203 indented more to the right than the first symbol. In this function
5204 definition, the @code{let} is indented more than the @code{defun}, and
5205 the @code{save-excursion} is indented more than the @code{let}, like
5221 This formatting convention makes it easy to see that the lines in
5222 the body of the @code{save-excursion} are enclosed by the parentheses
5223 associated with @code{save-excursion}, just as the
5224 @code{save-excursion} itself is enclosed by the parentheses associated
5225 with the @code{let}:
5229 (let ((oldbuf (current-buffer)))
5232 (set-buffer @dots{})
5233 (insert-buffer-substring oldbuf start end)
5239 The use of the @code{save-excursion} function can be viewed as a process
5240 of filling in the slots of a template:
5245 @var{first-expression-in-body}
5246 @var{second-expression-in-body}
5248 @var{last-expression-in-body})
5254 In this function, the body of the @code{save-excursion} contains only
5255 one expression, the @code{let*} expression. You know about a
5256 @code{let} function. The @code{let*} function is different. It has a
5257 @samp{*} in its name. It enables Emacs to set each variable in its
5258 varlist in sequence, one after another.
5260 Its critical feature is that variables later in the varlist can make
5261 use of the values to which Emacs set variables earlier in the varlist.
5262 @xref{fwd-para let, , The @code{let*} expression}.
5264 We will skip functions like @code{let*} and focus on two: the
5265 @code{set-buffer} function and the @code{insert-buffer-substring}
5269 In the old days, the @code{set-buffer} expression was simply
5272 (set-buffer (get-buffer-create buffer))
5280 (set-buffer append-to)
5284 @code{append-to} is bound to @code{(get-buffer-create buffer)} earlier
5285 on in the @code{let*} expression. That extra binding would not be
5286 necessary except for that @code{append-to} is used later in the
5287 varlist as an argument to @code{get-buffer-window-list}.
5292 (let ((oldbuf (current-buffer)))
5294 (let* ((append-to (get-buffer-create buffer))
5295 (windows (get-buffer-window-list append-to t t))
5297 (set-buffer append-to)
5298 (setq point (point))
5299 (barf-if-buffer-read-only)
5300 (insert-buffer-substring oldbuf start end)
5301 (dolist (window windows)
5302 (when (= (window-point window) point)
5303 (set-window-point window (point))))))))
5306 The @code{append-to-buffer} function definition inserts text from the
5307 buffer in which you are currently to a named buffer. It happens that
5308 @code{insert-buffer-substring} copies text from another buffer to the
5309 current buffer, just the reverse---that is why the
5310 @code{append-to-buffer} definition starts out with a @code{let} that
5311 binds the local symbol @code{oldbuf} to the value returned by
5312 @code{current-buffer}.
5315 The @code{insert-buffer-substring} expression looks like this:
5318 (insert-buffer-substring oldbuf start end)
5322 The @code{insert-buffer-substring} function copies a string
5323 @emph{from} the buffer specified as its first argument and inserts the
5324 string into the present buffer. In this case, the argument to
5325 @code{insert-buffer-substring} is the value of the variable created
5326 and bound by the @code{let}, namely the value of @code{oldbuf}, which
5327 was the current buffer when you gave the @code{append-to-buffer}
5330 After @code{insert-buffer-substring} has done its work,
5331 @code{save-excursion} will restore the action to the original buffer
5332 and @code{append-to-buffer} will have done its job.
5335 Written in skeletal form, the workings of the body look like this:
5339 (let (@var{bind-}@code{oldbuf}@var{-to-value-of-}@code{current-buffer})
5340 (save-excursion ; @r{Keep track of buffer.}
5342 @var{insert-substring-from-}@code{oldbuf}@var{-into-buffer})
5344 @var{change-back-to-original-buffer-when-finished}
5345 @var{let-the-local-meaning-of-}@code{oldbuf}@var{-disappear-when-finished}
5349 In summary, @code{append-to-buffer} works as follows: it saves the
5350 value of the current buffer in the variable called @code{oldbuf}. It
5351 gets the new buffer (creating one if need be) and switches Emacs's
5352 attention to it. Using the value of @code{oldbuf}, it inserts the
5353 region of text from the old buffer into the new buffer; and then using
5354 @code{save-excursion}, it brings you back to your original buffer.
5356 In looking at @code{append-to-buffer}, you have explored a fairly
5357 complex function. It shows how to use @code{let} and
5358 @code{save-excursion}, and how to change to and come back from another
5359 buffer. Many function definitions use @code{let},
5360 @code{save-excursion}, and @code{set-buffer} this way.
5362 @node Buffer Related Review
5365 Here is a brief summary of the various functions discussed in this chapter.
5368 @item describe-function
5369 @itemx describe-variable
5370 Print the documentation for a function or variable.
5371 Conventionally bound to @kbd{C-h f} and @kbd{C-h v}.
5373 @item xref-find-definitions
5374 Find the file containing the source for a function or variable and
5375 switch buffers to it, positioning point at the beginning of the item.
5376 Conventionally bound to @kbd{M-.} (that's a period following the
5379 @item save-excursion
5380 Save the location of point and restore its value after the
5381 arguments to @code{save-excursion} have been evaluated. Also, remember
5382 the current buffer and return to it.
5385 Set mark at a location and record the value of the previous mark on the
5386 mark ring. The mark is a location in the buffer that will keep its
5387 relative position even if text is added to or removed from the buffer.
5390 Set point to the location specified by the value of the argument, which
5391 can be a number, a marker, or an expression that returns the number of
5392 a position, such as @code{(point-min)}.
5394 @item insert-buffer-substring
5395 Copy a region of text from a buffer that is passed to the function as
5396 an argument and insert the region into the current buffer.
5398 @item mark-whole-buffer
5399 Mark the whole buffer as a region. Normally bound to @kbd{C-x h}.
5402 Switch the attention of Emacs to another buffer, but do not change the
5403 window being displayed. Used when the program rather than a human is
5404 to work on a different buffer.
5406 @item get-buffer-create
5408 Find a named buffer or create one if a buffer of that name does not
5409 exist. The @code{get-buffer} function returns @code{nil} if the named
5410 buffer does not exist.
5414 @node Buffer Exercises
5419 Write your own @code{simplified-end-of-buffer} function definition;
5420 then test it to see whether it works.
5423 Use @code{if} and @code{get-buffer} to write a function that prints a
5424 message telling you whether a buffer exists.
5427 Using @code{xref-find-definitions}, find the source for the
5428 @code{copy-to-buffer} function.
5432 @chapter A Few More Complex Functions
5434 In this chapter, we build on what we have learned in previous chapters
5435 by looking at more complex functions. The @code{copy-to-buffer}
5436 function illustrates use of two @code{save-excursion} expressions in
5437 one definition, while the @code{insert-buffer} function illustrates
5438 use of an asterisk in an @code{interactive} expression, use of
5439 @code{or}, and the important distinction between a name and the object
5440 to which the name refers.
5443 * copy-to-buffer:: With @code{set-buffer}, @code{get-buffer-create}.
5444 * insert-buffer:: Read-only, and with @code{or}.
5445 * beginning-of-buffer:: Shows @code{goto-char},
5446 @code{point-min}, and @code{push-mark}.
5447 * Second Buffer Related Review::
5448 * optional Exercise::
5451 @node copy-to-buffer
5452 @section The Definition of @code{copy-to-buffer}
5453 @findex copy-to-buffer
5455 After understanding how @code{append-to-buffer} works, it is easy to
5456 understand @code{copy-to-buffer}. This function copies text into a
5457 buffer, but instead of adding to the second buffer, it replaces all the
5458 previous text in the second buffer.
5461 The body of @code{copy-to-buffer} looks like this,
5466 (interactive "BCopy to buffer: \nr")
5467 (let ((oldbuf (current-buffer)))
5468 (with-current-buffer (get-buffer-create buffer)
5469 (barf-if-buffer-read-only)
5472 (insert-buffer-substring oldbuf start end)))))
5476 The @code{copy-to-buffer} function has a simpler @code{interactive}
5477 expression than @code{append-to-buffer}.
5480 The definition then says
5483 (with-current-buffer (get-buffer-create buffer) @dots{}
5486 First, look at the earliest inner expression; that is evaluated first.
5487 That expression starts with @code{get-buffer-create buffer}. The
5488 function tells the computer to use the buffer with the name specified
5489 as the one to which you are copying, or if such a buffer does not
5490 exist, to create it. Then, the @code{with-current-buffer} function
5491 evaluates its body with that buffer temporarily current.
5493 (This demonstrates another way to shift the computer's attention but
5494 not the user's. The @code{append-to-buffer} function showed how to do
5495 the same with @code{save-excursion} and @code{set-buffer}.
5496 @code{with-current-buffer} is a newer, and arguably easier,
5499 The @code{barf-if-buffer-read-only} function sends you an error
5500 message saying the buffer is read-only if you cannot modify it.
5502 The next line has the @code{erase-buffer} function as its sole
5503 contents. That function erases the buffer.
5505 Finally, the last two lines contain the @code{save-excursion}
5506 expression with @code{insert-buffer-substring} as its body.
5507 The @code{insert-buffer-substring} expression copies the text from
5508 the buffer you are in (and you have not seen the computer shift its
5509 attention, so you don't know that that buffer is now called
5512 Incidentally, this is what is meant by ``replacement''. To replace text,
5513 Emacs erases the previous text and then inserts new text.
5516 In outline, the body of @code{copy-to-buffer} looks like this:
5520 (let (@var{bind-}@code{oldbuf}@var{-to-value-of-}@code{current-buffer})
5521 (@var{with-the-buffer-you-are-copying-to}
5522 (@var{but-do-not-erase-or-copy-to-a-read-only-buffer})
5525 @var{insert-substring-from-}@code{oldbuf}@var{-into-buffer})))
5530 @section The Definition of @code{insert-buffer}
5531 @findex insert-buffer
5533 @code{insert-buffer} is yet another buffer-related function. This
5534 command copies another buffer @emph{into} the current buffer. It is the
5535 reverse of @code{append-to-buffer} or @code{copy-to-buffer}, since they
5536 copy a region of text @emph{from} the current buffer to another buffer.
5538 Here is a discussion based on the original code. The code was
5539 simplified in 2003 and is harder to understand.
5541 (@xref{New insert-buffer, , New Body for @code{insert-buffer}}, to see
5542 a discussion of the new body.)
5544 In addition, this code illustrates the use of @code{interactive} with a
5545 buffer that might be @dfn{read-only} and the important distinction
5546 between the name of an object and the object actually referred to.
5549 * insert-buffer code::
5550 * insert-buffer interactive:: When you can read, but not write.
5551 * insert-buffer body:: The body has an @code{or} and a @code{let}.
5552 * if & or:: Using an @code{if} instead of an @code{or}.
5553 * Insert or:: How the @code{or} expression works.
5554 * Insert let:: Two @code{save-excursion} expressions.
5555 * New insert-buffer::
5559 @node insert-buffer code
5560 @unnumberedsubsec The Code for @code{insert-buffer}
5564 Here is the earlier code:
5568 (defun insert-buffer (buffer)
5569 "Insert after point the contents of BUFFER.
5570 Puts mark after the inserted text.
5571 BUFFER may be a buffer or a buffer name."
5572 (interactive "*bInsert buffer:@: ")
5575 (or (bufferp buffer)
5576 (setq buffer (get-buffer buffer)))
5577 (let (start end newmark)
5581 (setq start (point-min) end (point-max)))
5584 (insert-buffer-substring buffer start end)
5585 (setq newmark (point)))
5586 (push-mark newmark)))
5591 As with other function definitions, you can use a template to see an
5592 outline of the function:
5596 (defun insert-buffer (buffer)
5597 "@var{documentation}@dots{}"
5598 (interactive "*bInsert buffer:@: ")
5603 @node insert-buffer interactive
5604 @subsection The Interactive Expression in @code{insert-buffer}
5605 @findex interactive@r{, example use of}
5607 In @code{insert-buffer}, the argument to the @code{interactive}
5608 declaration has two parts, an asterisk, @samp{*}, and @samp{bInsert
5612 * Read-only buffer:: When a buffer cannot be modified.
5613 * b for interactive:: An existing buffer or else its name.
5616 @node Read-only buffer
5617 @unnumberedsubsubsec A Read-only Buffer
5618 @cindex Read-only buffer
5619 @cindex Asterisk for read-only buffer
5620 @findex * @r{for read-only buffer}
5622 The asterisk is for the situation when the current buffer is a
5623 read-only buffer---a buffer that cannot be modified. If
5624 @code{insert-buffer} is called when the current buffer is read-only, a
5625 message to this effect is printed in the echo area and the terminal
5626 may beep or blink at you; you will not be permitted to insert anything
5627 into current buffer. The asterisk does not need to be followed by a
5628 newline to separate it from the next argument.
5630 @node b for interactive
5631 @unnumberedsubsubsec @samp{b} in an Interactive Expression
5633 The next argument in the interactive expression starts with a lower
5634 case @samp{b}. (This is different from the code for
5635 @code{append-to-buffer}, which uses an upper-case @samp{B}.
5636 @xref{append-to-buffer, , The Definition of @code{append-to-buffer}}.)
5637 The lower-case @samp{b} tells the Lisp interpreter that the argument
5638 for @code{insert-buffer} should be an existing buffer or else its
5639 name. (The upper-case @samp{B} option provides for the possibility
5640 that the buffer does not exist.) Emacs will prompt you for the name
5641 of the buffer, offering you a default buffer, with name completion
5642 enabled. If the buffer does not exist, you receive a message that
5643 says ``No match''; your terminal may beep at you as well.
5645 The new and simplified code generates a list for @code{interactive}.
5646 It uses the @code{barf-if-buffer-read-only} and @code{read-buffer}
5647 functions with which we are already familiar and the @code{progn}
5648 special form with which we are not. (It will be described later.)
5650 @node insert-buffer body
5651 @subsection The Body of the @code{insert-buffer} Function
5653 The body of the @code{insert-buffer} function has two major parts: an
5654 @code{or} expression and a @code{let} expression. The purpose of the
5655 @code{or} expression is to ensure that the argument @code{buffer} is
5656 bound to a buffer and not just the name of a buffer. The body of the
5657 @code{let} expression contains the code which copies the other buffer
5658 into the current buffer.
5661 In outline, the two expressions fit into the @code{insert-buffer}
5666 (defun insert-buffer (buffer)
5667 "@var{documentation}@dots{}"
5668 (interactive "*bInsert buffer:@: ")
5673 (let (@var{varlist})
5674 @var{body-of-}@code{let}@dots{} )
5678 To understand how the @code{or} expression ensures that the argument
5679 @code{buffer} is bound to a buffer and not to the name of a buffer, it
5680 is first necessary to understand the @code{or} function.
5682 Before doing this, let me rewrite this part of the function using
5683 @code{if} so that you can see what is done in a manner that will be familiar.
5686 @subsection @code{insert-buffer} With an @code{if} Instead of an @code{or}
5688 The job to be done is to make sure the value of @code{buffer} is a
5689 buffer itself and not the name of a buffer. If the value is the name,
5690 then the buffer itself must be got.
5692 You can imagine yourself at a conference where an usher is wandering
5693 around holding a list with your name on it and looking for you: the
5694 usher is bound to your name, not to you; but when the usher finds
5695 you and takes your arm, the usher becomes bound to you.
5698 In Lisp, you might describe this situation like this:
5702 (if (not (holding-on-to-guest))
5703 (find-and-take-arm-of-guest))
5707 We want to do the same thing with a buffer---if we do not have the
5708 buffer itself, we want to get it.
5711 Using a predicate called @code{bufferp} that tells us whether we have a
5712 buffer (rather than its name), we can write the code like this:
5716 (if (not (bufferp buffer)) ; @r{if-part}
5717 (setq buffer (get-buffer buffer))) ; @r{then-part}
5722 Here, the true-or-false-test of the @code{if} expression is
5723 @w{@code{(not (bufferp buffer))}}; and the then-part is the expression
5724 @w{@code{(setq buffer (get-buffer buffer))}}.
5726 In the test, the function @code{bufferp} returns true if its argument is
5727 a buffer---but false if its argument is the name of the buffer. (The
5728 last character of the function name @code{bufferp} is the character
5729 @samp{p}; as we saw earlier, such use of @samp{p} is a convention that
5730 indicates that the function is a predicate, which is a term that means
5731 that the function will determine whether some property is true or false.
5732 @xref{Wrong Type of Argument, , Using the Wrong Type Object as an
5736 The function @code{not} precedes the expression @code{(bufferp buffer)},
5737 so the true-or-false-test looks like this:
5740 (not (bufferp buffer))
5744 @code{not} is a function that returns true if its argument is false
5745 and false if its argument is true. So if @code{(bufferp buffer)}
5746 returns true, the @code{not} expression returns false and vice versa.
5748 Using this test, the @code{if} expression works as follows: when the
5749 value of the variable @code{buffer} is actually a buffer rather than
5750 its name, the true-or-false-test returns false and the @code{if}
5751 expression does not evaluate the then-part. This is fine, since we do
5752 not need to do anything to the variable @code{buffer} if it really is
5755 On the other hand, when the value of @code{buffer} is not a buffer
5756 itself, but the name of a buffer, the true-or-false-test returns true
5757 and the then-part of the expression is evaluated. In this case, the
5758 then-part is @code{(setq buffer (get-buffer buffer))}. This
5759 expression uses the @code{get-buffer} function to return an actual
5760 buffer itself, given its name. The @code{setq} then sets the variable
5761 @code{buffer} to the value of the buffer itself, replacing its previous
5762 value (which was the name of the buffer).
5765 @subsection The @code{or} in the Body
5767 The purpose of the @code{or} expression in the @code{insert-buffer}
5768 function is to ensure that the argument @code{buffer} is bound to a
5769 buffer and not just to the name of a buffer. The previous section shows
5770 how the job could have been done using an @code{if} expression.
5771 However, the @code{insert-buffer} function actually uses @code{or}.
5772 To understand this, it is necessary to understand how @code{or} works.
5775 An @code{or} function can have any number of arguments. It evaluates
5776 each argument in turn and returns the value of the first of its
5777 arguments that is not @code{nil}. Also, and this is a crucial feature
5778 of @code{or}, it does not evaluate any subsequent arguments after
5779 returning the first non-@code{nil} value.
5782 The @code{or} expression looks like this:
5786 (or (bufferp buffer)
5787 (setq buffer (get-buffer buffer)))
5792 The first argument to @code{or} is the expression @code{(bufferp buffer)}.
5793 This expression returns true (a non-@code{nil} value) if the buffer is
5794 actually a buffer, and not just the name of a buffer. In the @code{or}
5795 expression, if this is the case, the @code{or} expression returns this
5796 true value and does not evaluate the next expression---and this is fine
5797 with us, since we do not want to do anything to the value of
5798 @code{buffer} if it really is a buffer.
5800 On the other hand, if the value of @code{(bufferp buffer)} is @code{nil},
5801 which it will be if the value of @code{buffer} is the name of a buffer,
5802 the Lisp interpreter evaluates the next element of the @code{or}
5803 expression. This is the expression @code{(setq buffer (get-buffer
5804 buffer))}. This expression returns a non-@code{nil} value, which
5805 is the value to which it sets the variable @code{buffer}---and this
5806 value is a buffer itself, not the name of a buffer.
5808 The result of all this is that the symbol @code{buffer} is always
5809 bound to a buffer itself rather than to the name of a buffer. All
5810 this is necessary because the @code{set-buffer} function in a
5811 following line only works with a buffer itself, not with the name to a
5815 Incidentally, using @code{or}, the situation with the usher would be
5819 (or (holding-on-to-guest) (find-and-take-arm-of-guest))
5823 @subsection The @code{let} Expression in @code{insert-buffer}
5825 After ensuring that the variable @code{buffer} refers to a buffer itself
5826 and not just to the name of a buffer, the @code{insert-buffer function}
5827 continues with a @code{let} expression. This specifies three local
5828 variables, @code{start}, @code{end}, and @code{newmark} and binds them
5829 to the initial value @code{nil}. These variables are used inside the
5830 remainder of the @code{let} and temporarily hide any other occurrence of
5831 variables of the same name in Emacs until the end of the @code{let}.
5834 The body of the @code{let} contains two @code{save-excursion}
5835 expressions. First, we will look at the inner @code{save-excursion}
5836 expression in detail. The expression looks like this:
5842 (setq start (point-min) end (point-max)))
5847 The expression @code{(set-buffer buffer)} changes Emacs's attention
5848 from the current buffer to the one from which the text will copied.
5849 In that buffer, the variables @code{start} and @code{end} are set to
5850 the beginning and end of the buffer, using the commands
5851 @code{point-min} and @code{point-max}. Note that we have here an
5852 illustration of how @code{setq} is able to set two variables in the
5853 same expression. The first argument of @code{setq} is set to the
5854 value of its second, and its third argument is set to the value of its
5857 After the body of the inner @code{save-excursion} is evaluated, the
5858 @code{save-excursion} restores the original buffer, but @code{start} and
5859 @code{end} remain set to the values of the beginning and end of the
5860 buffer from which the text will be copied.
5863 The outer @code{save-excursion} expression looks like this:
5868 (@var{inner-}@code{save-excursion}@var{-expression}
5869 (@var{go-to-new-buffer-and-set-}@code{start}@var{-and-}@code{end})
5870 (insert-buffer-substring buffer start end)
5871 (setq newmark (point)))
5876 The @code{insert-buffer-substring} function copies the text
5877 @emph{into} the current buffer @emph{from} the region indicated by
5878 @code{start} and @code{end} in @code{buffer}. Since the whole of the
5879 second buffer lies between @code{start} and @code{end}, the whole of
5880 the second buffer is copied into the buffer you are editing. Next,
5881 the value of point, which will be at the end of the inserted text, is
5882 recorded in the variable @code{newmark}.
5884 After the body of the outer @code{save-excursion} is evaluated, point
5885 is relocated to its original place.
5887 However, it is convenient to locate a mark at the end of the newly
5888 inserted text and locate point at its beginning. The @code{newmark}
5889 variable records the end of the inserted text. In the last line of
5890 the @code{let} expression, the @code{(push-mark newmark)} expression
5891 function sets a mark to this location. (The previous location of the
5892 mark is still accessible; it is recorded on the mark ring and you can
5893 go back to it with @kbd{C-u C-@key{SPC}}.) Meanwhile, point is
5894 located at the beginning of the inserted text, which is where it was
5895 before you called the insert function, the position of which was saved
5896 by the first @code{save-excursion}.
5899 The whole @code{let} expression looks like this:
5903 (let (start end newmark)
5907 (setq start (point-min) end (point-max)))
5908 (insert-buffer-substring buffer start end)
5909 (setq newmark (point)))
5910 (push-mark newmark))
5914 Like the @code{append-to-buffer} function, the @code{insert-buffer}
5915 function uses @code{let}, @code{save-excursion}, and
5916 @code{set-buffer}. In addition, the function illustrates one way to
5917 use @code{or}. All these functions are building blocks that we will
5918 find and use again and again.
5920 @node New insert-buffer
5921 @subsection New Body for @code{insert-buffer}
5922 @findex insert-buffer@r{, new version body}
5923 @cindex new version body for @code{insert-buffer}
5925 The body in the GNU Emacs 22 version is more confusing than the original.
5928 It consists of two expressions,
5934 (insert-buffer-substring (get-buffer buffer))
5942 except, and this is what confuses novices, very important work is done
5943 inside the @code{push-mark} expression.
5945 The @code{get-buffer} function returns a buffer with the name
5946 provided. You will note that the function is @emph{not} called
5947 @code{get-buffer-create}; it does not create a buffer if one does not
5948 already exist. The buffer returned by @code{get-buffer}, an existing
5949 buffer, is passed to @code{insert-buffer-substring}, which inserts the
5950 whole of the buffer (since you did not specify anything else).
5952 The location into which the buffer is inserted is recorded by
5953 @code{push-mark}. Then the function returns @code{nil}, the value of
5954 its last command. Put another way, the @code{insert-buffer} function
5955 exists only to produce a side effect, inserting another buffer, not to
5958 @node beginning-of-buffer
5959 @section Complete Definition of @code{beginning-of-buffer}
5960 @findex beginning-of-buffer
5962 The basic structure of the @code{beginning-of-buffer} function has
5963 already been discussed. (@xref{simplified-beginning-of-buffer, , A
5964 Simplified @code{beginning-of-buffer} Definition}.)
5965 This section describes the complex part of the definition.
5967 As previously described, when invoked without an argument,
5968 @code{beginning-of-buffer} moves the cursor to the beginning of the
5969 buffer (in truth, the beginning of the accessible portion of the
5970 buffer), leaving the mark at the previous position. However, when the
5971 command is invoked with a number between one and ten, the function
5972 considers that number to be a fraction of the length of the buffer,
5973 measured in tenths, and Emacs moves the cursor that fraction of the
5974 way from the beginning of the buffer. Thus, you can either call this
5975 function with the key command @kbd{M-<}, which will move the cursor to
5976 the beginning of the buffer, or with a key command such as @kbd{C-u 7
5977 M-<} which will move the cursor to a point 70% of the way through the
5978 buffer. If a number bigger than ten is used for the argument, it
5979 moves to the end of the buffer.
5981 The @code{beginning-of-buffer} function can be called with or without an
5982 argument. The use of the argument is optional.
5985 * Optional Arguments::
5986 * beginning-of-buffer opt arg:: Example with optional argument.
5987 * beginning-of-buffer complete::
5990 @node Optional Arguments
5991 @subsection Optional Arguments
5993 Unless told otherwise, Lisp expects that a function with an argument in
5994 its function definition will be called with a value for that argument.
5995 If that does not happen, you get an error and a message that says
5996 @samp{Wrong number of arguments}.
5998 @cindex Optional arguments
6001 However, optional arguments are a feature of Lisp: a particular
6002 @dfn{keyword} is used to tell the Lisp interpreter that an argument is
6003 optional. The keyword is @code{&optional}. (The @samp{&} in front of
6004 @samp{optional} is part of the keyword.) In a function definition, if
6005 an argument follows the keyword @code{&optional}, no value need be
6006 passed to that argument when the function is called.
6009 The first line of the function definition of @code{beginning-of-buffer}
6010 therefore looks like this:
6013 (defun beginning-of-buffer (&optional arg)
6017 In outline, the whole function looks like this:
6021 (defun beginning-of-buffer (&optional arg)
6022 "@var{documentation}@dots{}"
6024 (or (@var{is-the-argument-a-cons-cell} arg)
6025 (and @var{are-both-transient-mark-mode-and-mark-active-true})
6027 (let (@var{determine-size-and-set-it})
6029 (@var{if-there-is-an-argument}
6030 @var{figure-out-where-to-go}
6037 The function is similar to the @code{simplified-beginning-of-buffer}
6038 function except that the @code{interactive} expression has @code{"P"}
6039 as an argument and the @code{goto-char} function is followed by an
6040 if-then-else expression that figures out where to put the cursor if
6041 there is an argument that is not a cons cell.
6043 (Since I do not explain a cons cell for many more chapters, please
6044 consider ignoring the function @code{consp}. @xref{List
6045 Implementation, , How Lists are Implemented}, and @ref{Cons Cell Type,
6046 , Cons Cell and List Types, elisp, The GNU Emacs Lisp Reference
6049 The @code{"P"} in the @code{interactive} expression tells Emacs to
6050 pass a prefix argument, if there is one, to the function in raw form.
6051 A prefix argument is made by typing the @key{META} key followed by a
6052 number, or by typing @kbd{C-u} and then a number. (If you don't type
6053 a number, @kbd{C-u} defaults to a cons cell with a 4. A lowercase
6054 @code{"p"} in the @code{interactive} expression causes the function to
6055 convert a prefix arg to a number.)
6057 The true-or-false-test of the @code{if} expression looks complex, but
6058 it is not: it checks whether @code{arg} has a value that is not
6059 @code{nil} and whether it is a cons cell. (That is what @code{consp}
6060 does; it checks whether its argument is a cons cell.) If @code{arg}
6061 has a value that is not @code{nil} (and is not a cons cell), which
6062 will be the case if @code{beginning-of-buffer} is called with a
6063 numeric argument, then this true-or-false-test will return true and
6064 the then-part of the @code{if} expression will be evaluated. On the
6065 other hand, if @code{beginning-of-buffer} is not called with an
6066 argument, the value of @code{arg} will be @code{nil} and the else-part
6067 of the @code{if} expression will be evaluated. The else-part is
6068 simply @code{point-min}, and when this is the outcome, the whole
6069 @code{goto-char} expression is @code{(goto-char (point-min))}, which
6070 is how we saw the @code{beginning-of-buffer} function in its
6073 @node beginning-of-buffer opt arg
6074 @subsection @code{beginning-of-buffer} with an Argument
6076 When @code{beginning-of-buffer} is called with an argument, an
6077 expression is evaluated which calculates what value to pass to
6078 @code{goto-char}. This expression is rather complicated at first sight.
6079 It includes an inner @code{if} expression and much arithmetic. It looks
6084 (if (> (buffer-size) 10000)
6085 ;; @r{Avoid overflow for large buffer sizes!}
6086 (* (prefix-numeric-value arg)
6091 size (prefix-numeric-value arg))) 10)))
6096 * Disentangle beginning-of-buffer::
6097 * Large buffer case::
6098 * Small buffer case::
6102 @node Disentangle beginning-of-buffer
6103 @unnumberedsubsubsec Disentangle @code{beginning-of-buffer}
6106 Like other complex-looking expressions, the conditional expression
6107 within @code{beginning-of-buffer} can be disentangled by looking at it
6108 as parts of a template, in this case, the template for an if-then-else
6109 expression. In skeletal form, the expression looks like this:
6113 (if (@var{buffer-is-large}
6114 @var{divide-buffer-size-by-10-and-multiply-by-arg}
6115 @var{else-use-alternate-calculation}
6119 The true-or-false-test of this inner @code{if} expression checks the
6120 size of the buffer. The reason for this is that the old version 18
6121 Emacs used numbers that are no bigger than eight million or so and in
6122 the computation that followed, the programmer feared that Emacs might
6123 try to use over-large numbers if the buffer were large. The term
6124 ``overflow'', mentioned in the comment, means numbers that are over
6125 large. More recent versions of Emacs use larger numbers, but this
6126 code has not been touched, if only because people now look at buffers
6127 that are far, far larger than ever before.
6129 There are two cases: if the buffer is large and if it is not.
6131 @node Large buffer case
6132 @unnumberedsubsubsec What happens in a large buffer
6134 In @code{beginning-of-buffer}, the inner @code{if} expression tests
6135 whether the size of the buffer is greater than 10,000 characters. To do
6136 this, it uses the @code{>} function and the computation of @code{size}
6137 that comes from the let expression.
6139 In the old days, the function @code{buffer-size} was used. Not only
6140 was that function called several times, it gave the size of the whole
6141 buffer, not the accessible part. The computation makes much more
6142 sense when it handles just the accessible part. (@xref{Narrowing &
6143 Widening, , Narrowing and Widening}, for more information on focusing
6144 attention to an accessible part.)
6147 The line looks like this:
6155 When the buffer is large, the then-part of the @code{if} expression is
6156 evaluated. It reads like this (after formatting for easy reading):
6161 (prefix-numeric-value arg)
6167 This expression is a multiplication, with two arguments to the function
6170 The first argument is @code{(prefix-numeric-value arg)}. When
6171 @code{"P"} is used as the argument for @code{interactive}, the value
6172 passed to the function as its argument is passed a @dfn{raw prefix
6173 argument}, and not a number. (It is a number in a list.) To perform
6174 the arithmetic, a conversion is necessary, and
6175 @code{prefix-numeric-value} does the job.
6177 @findex / @r{(division)}
6179 The second argument is @code{(/ size 10)}. This expression divides
6180 the numeric value by ten---the numeric value of the size of the
6181 accessible portion of the buffer. This produces a number that tells
6182 how many characters make up one tenth of the buffer size. (In Lisp,
6183 @code{/} is used for division, just as @code{*} is used for
6187 In the multiplication expression as a whole, this amount is multiplied
6188 by the value of the prefix argument---the multiplication looks like this:
6192 (* @var{numeric-value-of-prefix-arg}
6193 @var{number-of-characters-in-one-tenth-of-the-accessible-buffer})
6198 If, for example, the prefix argument is @samp{7}, the one-tenth value
6199 will be multiplied by 7 to give a position 70% of the way through.
6202 The result of all this is that if the accessible portion of the buffer
6203 is large, the @code{goto-char} expression reads like this:
6207 (goto-char (* (prefix-numeric-value arg)
6212 This puts the cursor where we want it.
6214 @node Small buffer case
6215 @unnumberedsubsubsec What happens in a small buffer
6217 If the buffer contains fewer than 10,000 characters, a slightly
6218 different computation is performed. You might think this is not
6219 necessary, since the first computation could do the job. However, in
6220 a small buffer, the first method may not put the cursor on exactly the
6221 desired line; the second method does a better job.
6224 The code looks like this:
6226 @c Keep this on one line.
6228 (/ (+ 10 (* size (prefix-numeric-value arg))) 10))
6233 This is code in which you figure out what happens by discovering how the
6234 functions are embedded in parentheses. It is easier to read if you
6235 reformat it with each expression indented more deeply than its
6236 enclosing expression:
6244 (prefix-numeric-value arg)))
6251 Looking at parentheses, we see that the innermost operation is
6252 @code{(prefix-numeric-value arg)}, which converts the raw argument to
6253 a number. In the following expression, this number is multiplied by
6254 the size of the accessible portion of the buffer:
6257 (* size (prefix-numeric-value arg))
6261 This multiplication creates a number that may be larger than the size of
6262 the buffer---seven times larger if the argument is 7, for example. Ten
6263 is then added to this number and finally the large number is divided by
6264 ten to provide a value that is one character larger than the percentage
6265 position in the buffer.
6267 The number that results from all this is passed to @code{goto-char} and
6268 the cursor is moved to that point.
6271 @node beginning-of-buffer complete
6272 @subsection The Complete @code{beginning-of-buffer}
6275 Here is the complete text of the @code{beginning-of-buffer} function:
6281 (defun beginning-of-buffer (&optional arg)
6282 "Move point to the beginning of the buffer;
6283 leave mark at previous position.
6284 With \\[universal-argument] prefix,
6285 do not set mark at previous position.
6287 put point N/10 of the way from the beginning.
6289 If the buffer is narrowed,
6290 this command uses the beginning and size
6291 of the accessible part of the buffer.
6295 Don't use this command in Lisp programs!
6296 \(goto-char (point-min)) is faster
6297 and avoids clobbering the mark."
6300 (and transient-mark-mode mark-active)
6304 (let ((size (- (point-max) (point-min))))
6305 (goto-char (if (and arg (not (consp arg)))
6308 ;; Avoid overflow for large buffer sizes!
6309 (* (prefix-numeric-value arg)
6311 (/ (+ 10 (* size (prefix-numeric-value arg)))
6314 (if (and arg (not (consp arg))) (forward-line 1)))
6319 From before GNU Emacs 22
6322 (defun beginning-of-buffer (&optional arg)
6323 "Move point to the beginning of the buffer;
6324 leave mark at previous position.
6325 With arg N, put point N/10 of the way
6326 from the true beginning.
6329 Don't use this in Lisp programs!
6330 \(goto-char (point-min)) is faster
6331 and does not set the mark."
6338 (if (> (buffer-size) 10000)
6339 ;; @r{Avoid overflow for large buffer sizes!}
6340 (* (prefix-numeric-value arg)
6341 (/ (buffer-size) 10))
6344 (/ (+ 10 (* (buffer-size)
6345 (prefix-numeric-value arg)))
6348 (if arg (forward-line 1)))
6354 Except for two small points, the previous discussion shows how this
6355 function works. The first point deals with a detail in the
6356 documentation string, and the second point concerns the last line of
6360 In the documentation string, there is reference to an expression:
6363 \\[universal-argument]
6367 A @samp{\\} is used before the first square bracket of this
6368 expression. This @samp{\\} tells the Lisp interpreter to substitute
6369 whatever key is currently bound to the @samp{[@dots{}]}. In the case
6370 of @code{universal-argument}, that is usually @kbd{C-u}, but it might
6371 be different. (@xref{Documentation Tips, , Tips for Documentation
6372 Strings, elisp, The GNU Emacs Lisp Reference Manual}, for more
6376 Finally, the last line of the @code{beginning-of-buffer} command says
6377 to move point to the beginning of the next line if the command is
6378 invoked with an argument:
6381 (if (and arg (not (consp arg))) (forward-line 1))
6385 This puts the cursor at the beginning of the first line after the
6386 appropriate tenths position in the buffer. This is a flourish that
6387 means that the cursor is always located @emph{at least} the requested
6388 tenths of the way through the buffer, which is a nicety that is,
6389 perhaps, not necessary, but which, if it did not occur, would be sure
6390 to draw complaints. (The @code{(not (consp arg))} portion is so that
6391 if you specify the command with a @kbd{C-u}, but without a number,
6392 that is to say, if the raw prefix argument is simply a cons cell,
6393 the command does not put you at the beginning of the second line.)
6395 @node Second Buffer Related Review
6398 Here is a brief summary of some of the topics covered in this chapter.
6402 Evaluate each argument in sequence, and return the value of the first
6403 argument that is not @code{nil}; if none return a value that is not
6404 @code{nil}, return @code{nil}. In brief, return the first true value
6405 of the arguments; return a true value if one @emph{or} any of the
6409 Evaluate each argument in sequence, and if any are @code{nil}, return
6410 @code{nil}; if none are @code{nil}, return the value of the last
6411 argument. In brief, return a true value only if all the arguments are
6412 true; return a true value if one @emph{and} each of the others is
6416 A keyword used to indicate that an argument to a function definition
6417 is optional; this means that the function can be evaluated without the
6418 argument, if desired.
6420 @item prefix-numeric-value
6421 Convert the raw prefix argument produced by @code{(interactive
6422 "P")} to a numeric value.
6425 Move point forward to the beginning of the next line, or if the argument
6426 is greater than one, forward that many lines. If it can't move as far
6427 forward as it is supposed to, @code{forward-line} goes forward as far as
6428 it can and then returns a count of the number of additional lines it was
6429 supposed to move but couldn't.
6432 Delete the entire contents of the current buffer.
6435 Return @code{t} if its argument is a buffer; otherwise return @code{nil}.
6438 @node optional Exercise
6439 @section @code{optional} Argument Exercise
6441 Write an interactive function with an optional argument that tests
6442 whether its argument, a number, is greater than or equal to, or else,
6443 less than the value of @code{fill-column}, and tells you which, in a
6444 message. However, if you do not pass an argument to the function, use
6445 56 as a default value.
6447 @node Narrowing & Widening
6448 @chapter Narrowing and Widening
6449 @cindex Focusing attention (narrowing)
6453 Narrowing is a feature of Emacs that makes it possible for you to focus
6454 on a specific part of a buffer, and work without accidentally changing
6455 other parts. Narrowing is normally disabled since it can confuse
6459 * Narrowing advantages:: The advantages of narrowing
6460 * save-restriction:: The @code{save-restriction} special form.
6461 * what-line:: The number of the line that point is on.
6466 @node Narrowing advantages
6467 @unnumberedsec The Advantages of Narrowing
6470 With narrowing, the rest of a buffer is made invisible, as if it weren't
6471 there. This is an advantage if, for example, you want to replace a word
6472 in one part of a buffer but not in another: you narrow to the part you want
6473 and the replacement is carried out only in that section, not in the rest
6474 of the buffer. Searches will only work within a narrowed region, not
6475 outside of one, so if you are fixing a part of a document, you can keep
6476 yourself from accidentally finding parts you do not need to fix by
6477 narrowing just to the region you want.
6478 (The key binding for @code{narrow-to-region} is @kbd{C-x n n}.)
6480 However, narrowing does make the rest of the buffer invisible, which
6481 can scare people who inadvertently invoke narrowing and think they
6482 have deleted a part of their file. Moreover, the @code{undo} command
6483 (which is usually bound to @kbd{C-x u}) does not turn off narrowing
6484 (nor should it), so people can become quite desperate if they do not
6485 know that they can return the rest of a buffer to visibility with the
6486 @code{widen} command.
6487 (The key binding for @code{widen} is @kbd{C-x n w}.)
6489 Narrowing is just as useful to the Lisp interpreter as to a human.
6490 Often, an Emacs Lisp function is designed to work on just part of a
6491 buffer; or conversely, an Emacs Lisp function needs to work on all of a
6492 buffer that has been narrowed. The @code{what-line} function, for
6493 example, removes the narrowing from a buffer, if it has any narrowing
6494 and when it has finished its job, restores the narrowing to what it was.
6495 On the other hand, the @code{count-lines} function
6496 uses narrowing to restrict itself to just that portion
6497 of the buffer in which it is interested and then restores the previous
6500 @node save-restriction
6501 @section The @code{save-restriction} Special Form
6502 @findex save-restriction
6504 In Emacs Lisp, you can use the @code{save-restriction} special form to
6505 keep track of whatever narrowing is in effect, if any. When the Lisp
6506 interpreter meets with @code{save-restriction}, it executes the code
6507 in the body of the @code{save-restriction} expression, and then undoes
6508 any changes to narrowing that the code caused. If, for example, the
6509 buffer is narrowed and the code that follows @code{save-restriction}
6510 gets rid of the narrowing, @code{save-restriction} returns the buffer
6511 to its narrowed region afterwards. In the @code{what-line} command,
6512 any narrowing the buffer may have is undone by the @code{widen}
6513 command that immediately follows the @code{save-restriction} command.
6514 Any original narrowing is restored just before the completion of the
6518 The template for a @code{save-restriction} expression is simple:
6528 The body of the @code{save-restriction} is one or more expressions that
6529 will be evaluated in sequence by the Lisp interpreter.
6531 Finally, a point to note: when you use both @code{save-excursion} and
6532 @code{save-restriction}, one right after the other, you should use
6533 @code{save-excursion} outermost. If you write them in reverse order,
6534 you may fail to record narrowing in the buffer to which Emacs switches
6535 after calling @code{save-excursion}. Thus, when written together,
6536 @code{save-excursion} and @code{save-restriction} should be written
6547 In other circumstances, when not written together, the
6548 @code{save-excursion} and @code{save-restriction} special forms must
6549 be written in the order appropriate to the function.
6565 /usr/local/src/emacs/lisp/simple.el
6568 "Print the current buffer line number and narrowed line number of point."
6570 (let ((start (point-min))
6571 (n (line-number-at-pos)))
6573 (message "Line %d" n)
6577 (message "line %d (narrowed line %d)"
6578 (+ n (line-number-at-pos start) -1) n))))))
6580 (defun line-number-at-pos (&optional pos)
6581 "Return (narrowed) buffer line number at position POS.
6582 If POS is nil, use current buffer location.
6583 Counting starts at (point-min), so the value refers
6584 to the contents of the accessible portion of the buffer."
6585 (let ((opoint (or pos (point))) start)
6587 (goto-char (point-min))
6588 (setq start (point))
6591 (1+ (count-lines start (point))))))
6593 (defun count-lines (start end)
6594 "Return number of lines between START and END.
6595 This is usually the number of newlines between them,
6596 but can be one more if START is not equal to END
6597 and the greater of them is not at the start of a line."
6600 (narrow-to-region start end)
6601 (goto-char (point-min))
6602 (if (eq selective-display t)
6605 (while (re-search-forward "[\n\C-m]" nil t 40)
6606 (setq done (+ 40 done)))
6607 (while (re-search-forward "[\n\C-m]" nil t 1)
6608 (setq done (+ 1 done)))
6609 (goto-char (point-max))
6610 (if (and (/= start end)
6614 (- (buffer-size) (forward-line (buffer-size)))))))
6618 @section @code{what-line}
6620 @cindex Widening, example of
6622 The @code{what-line} command tells you the number of the line in which
6623 the cursor is located. The function illustrates the use of the
6624 @code{save-restriction} and @code{save-excursion} commands. Here is the
6625 original text of the function:
6630 "Print the current line number (in the buffer) of point."
6637 (1+ (count-lines 1 (point)))))))
6641 (In recent versions of GNU Emacs, the @code{what-line} function has
6642 been expanded to tell you your line number in a narrowed buffer as
6643 well as your line number in a widened buffer. The recent version is
6644 more complex than the version shown here. If you feel adventurous,
6645 you might want to look at it after figuring out how this version
6646 works. You will probably need to use @kbd{C-h f}
6647 (@code{describe-function}). The newer version uses a conditional to
6648 determine whether the buffer has been narrowed.
6650 (Also, it uses @code{line-number-at-pos}, which among other simple
6651 expressions, such as @code{(goto-char (point-min))}, moves point to
6652 the beginning of the current line with @code{(forward-line 0)} rather
6653 than @code{beginning-of-line}.)
6655 The @code{what-line} function as shown here has a documentation line
6656 and is interactive, as you would expect. The next two lines use the
6657 functions @code{save-restriction} and @code{widen}.
6659 The @code{save-restriction} special form notes whatever narrowing is in
6660 effect, if any, in the current buffer and restores that narrowing after
6661 the code in the body of the @code{save-restriction} has been evaluated.
6663 The @code{save-restriction} special form is followed by @code{widen}.
6664 This function undoes any narrowing the current buffer may have had
6665 when @code{what-line} was called. (The narrowing that was there is
6666 the narrowing that @code{save-restriction} remembers.) This widening
6667 makes it possible for the line counting commands to count from the
6668 beginning of the buffer. Otherwise, they would have been limited to
6669 counting within the accessible region. Any original narrowing is
6670 restored just before the completion of the function by the
6671 @code{save-restriction} special form.
6673 The call to @code{widen} is followed by @code{save-excursion}, which
6674 saves the location of the cursor (i.e., of point), and
6675 restores it after the code in the body of the @code{save-excursion}
6676 uses the @code{beginning-of-line} function to move point.
6678 (Note that the @code{(widen)} expression comes between the
6679 @code{save-restriction} and @code{save-excursion} special forms. When
6680 you write the two @code{save- @dots{}} expressions in sequence, write
6681 @code{save-excursion} outermost.)
6684 The last two lines of the @code{what-line} function are functions to
6685 count the number of lines in the buffer and then print the number in the
6691 (1+ (count-lines 1 (point)))))))
6695 The @code{message} function prints a one-line message at the bottom of
6696 the Emacs screen. The first argument is inside of quotation marks and
6697 is printed as a string of characters. However, it may contain a
6698 @samp{%d} expression to print a following argument. @samp{%d} prints
6699 the argument as a decimal, so the message will say something such as
6703 The number that is printed in place of the @samp{%d} is computed by the
6704 last line of the function:
6707 (1+ (count-lines 1 (point)))
6713 (defun count-lines (start end)
6714 "Return number of lines between START and END.
6715 This is usually the number of newlines between them,
6716 but can be one more if START is not equal to END
6717 and the greater of them is not at the start of a line."
6720 (narrow-to-region start end)
6721 (goto-char (point-min))
6722 (if (eq selective-display t)
6725 (while (re-search-forward "[\n\C-m]" nil t 40)
6726 (setq done (+ 40 done)))
6727 (while (re-search-forward "[\n\C-m]" nil t 1)
6728 (setq done (+ 1 done)))
6729 (goto-char (point-max))
6730 (if (and (/= start end)
6734 (- (buffer-size) (forward-line (buffer-size)))))))
6738 What this does is count the lines from the first position of the
6739 buffer, indicated by the @code{1}, up to @code{(point)}, and then add
6740 one to that number. (The @code{1+} function adds one to its
6741 argument.) We add one to it because line 2 has only one line before
6742 it, and @code{count-lines} counts only the lines @emph{before} the
6745 After @code{count-lines} has done its job, and the message has been
6746 printed in the echo area, the @code{save-excursion} restores point to
6747 its original position; and @code{save-restriction} restores
6748 the original narrowing, if any.
6750 @node narrow Exercise
6751 @section Exercise with Narrowing
6753 Write a function that will display the first 60 characters of the
6754 current buffer, even if you have narrowed the buffer to its latter
6755 half so that the first line is inaccessible. Restore point, mark, and
6756 narrowing. For this exercise, you need to use a whole potpourri of
6757 functions, including @code{save-restriction}, @code{widen},
6758 @code{goto-char}, @code{point-min}, @code{message}, and
6759 @code{buffer-substring}.
6761 @cindex Properties, mention of @code{buffer-substring-no-properties}
6762 (@code{buffer-substring} is a previously unmentioned function you will
6763 have to investigate yourself; or perhaps you will have to use
6764 @code{buffer-substring-no-properties} or
6765 @code{filter-buffer-substring} @dots{}, yet other functions. Text
6766 properties are a feature otherwise not discussed here. @xref{Text
6767 Properties, , Text Properties, elisp, The GNU Emacs Lisp Reference
6770 Additionally, do you really need @code{goto-char} or @code{point-min}?
6771 Or can you write the function without them?
6773 @node car cdr & cons
6774 @chapter @code{car}, @code{cdr}, @code{cons}: Fundamental Functions
6775 @findex car@r{, introduced}
6776 @findex cdr@r{, introduced}
6778 In Lisp, @code{car}, @code{cdr}, and @code{cons} are fundamental
6779 functions. The @code{cons} function is used to construct lists, and
6780 the @code{car} and @code{cdr} functions are used to take them apart.
6782 In the walk through of the @code{copy-region-as-kill} function, we
6783 will see @code{cons} as well as two variants on @code{cdr},
6784 namely, @code{setcdr} and @code{nthcdr}. (@xref{copy-region-as-kill}.)
6787 * Strange Names:: An historical aside: why the strange names?
6788 * car & cdr:: Functions for extracting part of a list.
6789 * cons:: Constructing a list.
6790 * nthcdr:: Calling @code{cdr} repeatedly.
6792 * setcar:: Changing the first element of a list.
6793 * setcdr:: Changing the rest of a list.
6799 @unnumberedsec Strange Names
6802 The name of the @code{cons} function is not unreasonable: it is an
6803 abbreviation of the word ``construct''. The origins of the names for
6804 @code{car} and @code{cdr}, on the other hand, are esoteric: @code{car}
6805 is an acronym from the phrase ``Contents of the Address part of the
6806 Register''; and @code{cdr} (pronounced ``could-er'') is an acronym from
6807 the phrase ``Contents of the Decrement part of the Register''. These
6808 phrases refer to specific pieces of hardware on the very early
6809 computer on which the original Lisp was developed. Besides being
6810 obsolete, the phrases have been completely irrelevant for more than 25
6811 years to anyone thinking about Lisp. Nonetheless, although a few
6812 brave scholars have begun to use more reasonable names for these
6813 functions, the old terms are still in use. In particular, since the
6814 terms are used in the Emacs Lisp source code, we will use them in this
6818 @section @code{car} and @code{cdr}
6820 The @sc{car} of a list is, quite simply, the first item in the list.
6821 Thus the @sc{car} of the list @code{(rose violet daisy buttercup)} is
6825 If you are reading this in Info in GNU Emacs, you can see this by
6826 evaluating the following:
6829 (car '(rose violet daisy buttercup))
6833 After evaluating the expression, @code{rose} will appear in the echo
6836 Clearly, a more reasonable name for the @code{car} function would be
6837 @code{first} and this is often suggested.
6839 @code{car} does not remove the first item from the list; it only reports
6840 what it is. After @code{car} has been applied to a list, the list is
6841 still the same as it was. In the jargon, @code{car} is
6842 ``non-destructive''. This feature turns out to be important.
6844 The @sc{cdr} of a list is the rest of the list, that is, the
6845 @code{cdr} function returns the part of the list that follows the
6846 first item. Thus, while the @sc{car} of the list @code{'(rose violet
6847 daisy buttercup)} is @code{rose}, the rest of the list, the value
6848 returned by the @code{cdr} function, is @code{(violet daisy
6852 You can see this by evaluating the following in the usual way:
6855 (cdr '(rose violet daisy buttercup))
6859 When you evaluate this, @code{(violet daisy buttercup)} will appear in
6862 Like @code{car}, @code{cdr} does not remove any elements from the
6863 list---it just returns a report of what the second and subsequent
6866 Incidentally, in the example, the list of flowers is quoted. If it were
6867 not, the Lisp interpreter would try to evaluate the list by calling
6868 @code{rose} as a function. In this example, we do not want to do that.
6870 Clearly, a more reasonable name for @code{cdr} would be @code{rest}.
6872 (There is a lesson here: when you name new functions, consider very
6873 carefully what you are doing, since you may be stuck with the names
6874 for far longer than you expect. The reason this document perpetuates
6875 these names is that the Emacs Lisp source code uses them, and if I did
6876 not use them, you would have a hard time reading the code; but do,
6877 please, try to avoid using these terms yourself. The people who come
6878 after you will be grateful to you.)
6880 When @code{car} and @code{cdr} are applied to a list made up of symbols,
6881 such as the list @code{(pine fir oak maple)}, the element of the list
6882 returned by the function @code{car} is the symbol @code{pine} without
6883 any parentheses around it. @code{pine} is the first element in the
6884 list. However, the @sc{cdr} of the list is a list itself, @code{(fir
6885 oak maple)}, as you can see by evaluating the following expressions in
6890 (car '(pine fir oak maple))
6892 (cdr '(pine fir oak maple))
6896 On the other hand, in a list of lists, the first element is itself a
6897 list. @code{car} returns this first element as a list. For example,
6898 the following list contains three sub-lists, a list of carnivores, a
6899 list of herbivores and a list of sea mammals:
6903 (car '((lion tiger cheetah)
6904 (gazelle antelope zebra)
6905 (whale dolphin seal)))
6910 In this example, the first element or @sc{car} of the list is the list of
6911 carnivores, @code{(lion tiger cheetah)}, and the rest of the list is
6912 @code{((gazelle antelope zebra) (whale dolphin seal))}.
6916 (cdr '((lion tiger cheetah)
6917 (gazelle antelope zebra)
6918 (whale dolphin seal)))
6922 It is worth saying again that @code{car} and @code{cdr} are
6923 non-destructive---that is, they do not modify or change lists to which
6924 they are applied. This is very important for how they are used.
6926 Also, in the first chapter, in the discussion about atoms, I said that
6927 in Lisp, certain kinds of atom, such as an array, can be separated
6928 into parts; but the mechanism for doing this is different from the
6929 mechanism for splitting a list. As far as Lisp is concerned, the
6930 atoms of a list are unsplittable. (@xref{Lisp Atoms}.) The
6931 @code{car} and @code{cdr} functions are used for splitting lists and
6932 are considered fundamental to Lisp. Since they cannot split or gain
6933 access to the parts of an array, an array is considered an atom.
6934 Conversely, the other fundamental function, @code{cons}, can put
6935 together or construct a list, but not an array. (Arrays are handled
6936 by array-specific functions. @xref{Arrays, , Arrays, elisp, The GNU
6937 Emacs Lisp Reference Manual}.)
6940 @section @code{cons}
6941 @findex cons@r{, introduced}
6943 The @code{cons} function constructs lists; it is the inverse of
6944 @code{car} and @code{cdr}. For example, @code{cons} can be used to make
6945 a four element list from the three element list, @code{(fir oak maple)}:
6948 (cons 'pine '(fir oak maple))
6953 After evaluating this list, you will see
6956 (pine fir oak maple)
6960 appear in the echo area. @code{cons} causes the creation of a new
6961 list in which the element is followed by the elements of the original
6964 We often say that @code{cons} puts a new element at the beginning of
6965 a list, or that it attaches or pushes elements onto the list, but this
6966 phrasing can be misleading, since @code{cons} does not change an
6967 existing list, but creates a new one.
6969 Like @code{car} and @code{cdr}, @code{cons} is non-destructive.
6973 * length:: How to find the length of a list.
6978 @unnumberedsubsec Build a list
6981 @code{cons} must have a list to attach to.@footnote{Actually, you can
6982 @code{cons} an element to an atom to produce a dotted pair. Dotted
6983 pairs are not discussed here; see @ref{Dotted Pair Notation, , Dotted
6984 Pair Notation, elisp, The GNU Emacs Lisp Reference Manual}.} You
6985 cannot start from absolutely nothing. If you are building a list, you
6986 need to provide at least an empty list at the beginning. Here is a
6987 series of @code{cons} expressions that build up a list of flowers. If
6988 you are reading this in Info in GNU Emacs, you can evaluate each of
6989 the expressions in the usual way; the value is printed in this text
6990 after @samp{@result{}}, which you may read as ``evaluates to''.
6994 (cons 'buttercup ())
6995 @result{} (buttercup)
6999 (cons 'daisy '(buttercup))
7000 @result{} (daisy buttercup)
7004 (cons 'violet '(daisy buttercup))
7005 @result{} (violet daisy buttercup)
7009 (cons 'rose '(violet daisy buttercup))
7010 @result{} (rose violet daisy buttercup)
7015 In the first example, the empty list is shown as @code{()} and a list
7016 made up of @code{buttercup} followed by the empty list is constructed.
7017 As you can see, the empty list is not shown in the list that was
7018 constructed. All that you see is @code{(buttercup)}. The empty list is
7019 not counted as an element of a list because there is nothing in an empty
7020 list. Generally speaking, an empty list is invisible.
7022 The second example, @code{(cons 'daisy '(buttercup))} constructs a new,
7023 two element list by putting @code{daisy} in front of @code{buttercup};
7024 and the third example constructs a three element list by putting
7025 @code{violet} in front of @code{daisy} and @code{buttercup}.
7028 @subsection Find the Length of a List: @code{length}
7031 You can find out how many elements there are in a list by using the Lisp
7032 function @code{length}, as in the following examples:
7036 (length '(buttercup))
7041 (length '(daisy buttercup))
7046 (length (cons 'violet '(daisy buttercup)))
7052 In the third example, the @code{cons} function is used to construct a
7053 three element list which is then passed to the @code{length} function as
7057 We can also use @code{length} to count the number of elements in an
7068 As you would expect, the number of elements in an empty list is zero.
7070 An interesting experiment is to find out what happens if you try to find
7071 the length of no list at all; that is, if you try to call @code{length}
7072 without giving it an argument, not even an empty list:
7080 What you see, if you evaluate this, is the error message
7083 Lisp error: (wrong-number-of-arguments length 0)
7087 This means that the function receives the wrong number of
7088 arguments, zero, when it expects some other number of arguments. In
7089 this case, one argument is expected, the argument being a list whose
7090 length the function is measuring. (Note that @emph{one} list is
7091 @emph{one} argument, even if the list has many elements inside it.)
7093 The part of the error message that says @samp{length} is the name of
7097 @code{length} is still a subroutine, but you need C-h f to discover that.
7099 In an earlier version:
7100 This is written with a special notation, @samp{#<subr},
7101 that indicates that the function @code{length} is one of the primitive
7102 functions written in C rather than in Emacs Lisp. (@samp{subr} is an
7103 abbreviation for ``subroutine''.) @xref{What Is a Function, , What Is a
7104 Function?, elisp , The GNU Emacs Lisp Reference Manual}, for more
7109 @section @code{nthcdr}
7112 The @code{nthcdr} function is associated with the @code{cdr} function.
7113 What it does is take the @sc{cdr} of a list repeatedly.
7115 If you take the @sc{cdr} of the list @code{(pine fir
7116 oak maple)}, you will be returned the list @code{(fir oak maple)}. If you
7117 repeat this on what was returned, you will be returned the list
7118 @code{(oak maple)}. (Of course, repeated @sc{cdr}ing on the original
7119 list will just give you the original @sc{cdr} since the function does
7120 not change the list. You need to evaluate the @sc{cdr} of the
7121 @sc{cdr} and so on.) If you continue this, eventually you will be
7122 returned an empty list, which in this case, instead of being shown as
7123 @code{()} is shown as @code{nil}.
7126 For review, here is a series of repeated @sc{cdr}s, the text following
7127 the @samp{@result{}} shows what is returned.
7131 (cdr '(pine fir oak maple))
7132 @result{}(fir oak maple)
7136 (cdr '(fir oak maple))
7137 @result{} (oak maple)
7162 You can also do several @sc{cdr}s without printing the values in
7167 (cdr (cdr '(pine fir oak maple)))
7168 @result{} (oak maple)
7173 In this example, the Lisp interpreter evaluates the innermost list first.
7174 The innermost list is quoted, so it just passes the list as it is to the
7175 innermost @code{cdr}. This @code{cdr} passes a list made up of the
7176 second and subsequent elements of the list to the outermost @code{cdr},
7177 which produces a list composed of the third and subsequent elements of
7178 the original list. In this example, the @code{cdr} function is repeated
7179 and returns a list that consists of the original list without its
7182 The @code{nthcdr} function does the same as repeating the call to
7183 @code{cdr}. In the following example, the argument 2 is passed to the
7184 function @code{nthcdr}, along with the list, and the value returned is
7185 the list without its first two items, which is exactly the same
7186 as repeating @code{cdr} twice on the list:
7190 (nthcdr 2 '(pine fir oak maple))
7191 @result{} (oak maple)
7196 Using the original four element list, we can see what happens when
7197 various numeric arguments are passed to @code{nthcdr}, including 0, 1,
7202 ;; @r{Leave the list as it was.}
7203 (nthcdr 0 '(pine fir oak maple))
7204 @result{} (pine fir oak maple)
7208 ;; @r{Return a copy without the first element.}
7209 (nthcdr 1 '(pine fir oak maple))
7210 @result{} (fir oak maple)
7214 ;; @r{Return a copy of the list without three elements.}
7215 (nthcdr 3 '(pine fir oak maple))
7220 ;; @r{Return a copy lacking all four elements.}
7221 (nthcdr 4 '(pine fir oak maple))
7226 ;; @r{Return a copy lacking all elements.}
7227 (nthcdr 5 '(pine fir oak maple))
7236 The @code{nthcdr} function takes the @sc{cdr} of a list repeatedly.
7237 The @code{nth} function takes the @sc{car} of the result returned by
7238 @code{nthcdr}. It returns the Nth element of the list.
7241 Thus, if it were not defined in C for speed, the definition of
7242 @code{nth} would be:
7247 "Returns the Nth element of LIST.
7248 N counts from zero. If LIST is not that long, nil is returned."
7249 (car (nthcdr n list)))
7254 (Originally, @code{nth} was defined in Emacs Lisp in @file{subr.el},
7255 but its definition was redone in C in the 1980s.)
7257 The @code{nth} function returns a single element of a list.
7258 This can be very convenient.
7260 Note that the elements are numbered from zero, not one. That is to
7261 say, the first element of a list, its @sc{car} is the zeroth element.
7262 This zero-based counting often bothers people who
7263 are accustomed to the first element in a list being number one, which
7271 (nth 0 '("one" "two" "three"))
7274 (nth 1 '("one" "two" "three"))
7279 It is worth mentioning that @code{nth}, like @code{nthcdr} and
7280 @code{cdr}, does not change the original list---the function is
7281 non-destructive. This is in sharp contrast to the @code{setcar} and
7282 @code{setcdr} functions.
7285 @section @code{setcar}
7288 As you might guess from their names, the @code{setcar} and @code{setcdr}
7289 functions set the @sc{car} or the @sc{cdr} of a list to a new value.
7290 They actually change the original list, unlike @code{car} and @code{cdr}
7291 which leave the original list as it was. One way to find out how this
7292 works is to experiment. We will start with the @code{setcar} function.
7295 First, we can make a list and then set the value of a variable to the
7296 list, using the @code{setq} function. Here is a list of animals:
7299 (setq animals '(antelope giraffe lion tiger))
7303 If you are reading this in Info inside of GNU Emacs, you can evaluate
7304 this expression in the usual fashion, by positioning the cursor after
7305 the expression and typing @kbd{C-x C-e}. (I'm doing this right here
7306 as I write this. This is one of the advantages of having the
7307 interpreter built into the computing environment. Incidentally, when
7308 there is nothing on the line after the final parentheses, such as a
7309 comment, point can be on the next line. Thus, if your cursor is in
7310 the first column of the next line, you do not need to move it.
7311 Indeed, Emacs permits any amount of white space after the final
7315 When we evaluate the variable @code{animals}, we see that it is bound to
7316 the list @code{(antelope giraffe lion tiger)}:
7321 @result{} (antelope giraffe lion tiger)
7326 Put another way, the variable @code{animals} points to the list
7327 @code{(antelope giraffe lion tiger)}.
7329 Next, evaluate the function @code{setcar} while passing it two
7330 arguments, the variable @code{animals} and the quoted symbol
7331 @code{hippopotamus}; this is done by writing the three element list
7332 @code{(setcar animals 'hippopotamus)} and then evaluating it in the
7336 (setcar animals 'hippopotamus)
7341 After evaluating this expression, evaluate the variable @code{animals}
7342 again. You will see that the list of animals has changed:
7347 @result{} (hippopotamus giraffe lion tiger)
7352 The first element on the list, @code{antelope} is replaced by
7353 @code{hippopotamus}.
7355 So we can see that @code{setcar} did not add a new element to the list
7356 as @code{cons} would have; it replaced @code{antelope} with
7357 @code{hippopotamus}; it @emph{changed} the list.
7360 @section @code{setcdr}
7363 The @code{setcdr} function is similar to the @code{setcar} function,
7364 except that the function replaces the second and subsequent elements of
7365 a list rather than the first element.
7367 (To see how to change the last element of a list, look ahead to
7368 @ref{kill-new function, , The @code{kill-new} function}, which uses
7369 the @code{nthcdr} and @code{setcdr} functions.)
7372 To see how this works, set the value of the variable to a list of
7373 domesticated animals by evaluating the following expression:
7376 (setq domesticated-animals '(horse cow sheep goat))
7381 If you now evaluate the list, you will be returned the list
7382 @code{(horse cow sheep goat)}:
7386 domesticated-animals
7387 @result{} (horse cow sheep goat)
7392 Next, evaluate @code{setcdr} with two arguments, the name of the
7393 variable which has a list as its value, and the list to which the
7394 @sc{cdr} of the first list will be set;
7397 (setcdr domesticated-animals '(cat dog))
7401 If you evaluate this expression, the list @code{(cat dog)} will appear
7402 in the echo area. This is the value returned by the function. The
7403 result we are interested in is the side effect, which we can see by
7404 evaluating the variable @code{domesticated-animals}:
7408 domesticated-animals
7409 @result{} (horse cat dog)
7414 Indeed, the list is changed from @code{(horse cow sheep goat)} to
7415 @code{(horse cat dog)}. The @sc{cdr} of the list is changed from
7416 @code{(cow sheep goat)} to @code{(cat dog)}.
7421 Construct a list of four birds by evaluating several expressions with
7422 @code{cons}. Find out what happens when you @code{cons} a list onto
7423 itself. Replace the first element of the list of four birds with a
7424 fish. Replace the rest of that list with a list of other fish.
7426 @node Cutting & Storing Text
7427 @chapter Cutting and Storing Text
7428 @cindex Cutting and storing text
7429 @cindex Storing and cutting text
7430 @cindex Killing text
7431 @cindex Clipping text
7432 @cindex Erasing text
7433 @cindex Deleting text
7435 Whenever you cut or clip text out of a buffer with a @dfn{kill} command in
7436 GNU Emacs, it is stored in a list and you can bring it back with a
7439 (The use of the word ``kill'' in Emacs for processes which specifically
7440 @emph{do not} destroy the values of the entities is an unfortunate
7441 historical accident. A much more appropriate word would be ``clip'' since
7442 that is what the kill commands do; they clip text out of a buffer and
7443 put it into storage from which it can be brought back. I have often
7444 been tempted to replace globally all occurrences of ``kill'' in the Emacs
7445 sources with ``clip'' and all occurrences of ``killed'' with ``clipped''.)
7448 * Storing Text:: Text is stored in a list.
7449 * zap-to-char:: Cutting out text up to a character.
7450 * kill-region:: Cutting text out of a region.
7451 * copy-region-as-kill:: A definition for copying text.
7452 * Digression into C:: Minor note on C programming language macros.
7453 * defvar:: How to give a variable an initial value.
7454 * cons & search-fwd Review::
7455 * search Exercises::
7460 @unnumberedsec Storing Text in a List
7463 When text is cut out of a buffer, it is stored on a list. Successive
7464 pieces of text are stored on the list successively, so the list might
7468 ("a piece of text" "previous piece")
7473 The function @code{cons} can be used to create a new list from a piece
7474 of text (an ``atom'', to use the jargon) and an existing list, like
7479 (cons "another piece"
7480 '("a piece of text" "previous piece"))
7486 If you evaluate this expression, a list of three elements will appear in
7490 ("another piece" "a piece of text" "previous piece")
7493 With the @code{car} and @code{nthcdr} functions, you can retrieve
7494 whichever piece of text you want. For example, in the following code,
7495 @code{nthcdr 1 @dots{}} returns the list with the first item removed;
7496 and the @code{car} returns the first element of that remainder---the
7497 second element of the original list:
7501 (car (nthcdr 1 '("another piece"
7504 @result{} "a piece of text"
7508 The actual functions in Emacs are more complex than this, of course.
7509 The code for cutting and retrieving text has to be written so that
7510 Emacs can figure out which element in the list you want---the first,
7511 second, third, or whatever. In addition, when you get to the end of
7512 the list, Emacs should give you the first element of the list, rather
7513 than nothing at all.
7515 The list that holds the pieces of text is called the @dfn{kill ring}.
7516 This chapter leads up to a description of the kill ring and how it is
7517 used by first tracing how the @code{zap-to-char} function works. This
7518 function calls a function that invokes a function that
7519 manipulates the kill ring. Thus, before reaching the mountains, we
7520 climb the foothills.
7522 A subsequent chapter describes how text that is cut from the buffer is
7523 retrieved. @xref{Yanking, , Yanking Text Back}.
7526 @section @code{zap-to-char}
7529 Let us look at the interactive @code{zap-to-char} function.
7532 * Complete zap-to-char:: The complete implementation.
7533 * zap-to-char interactive:: A three part interactive expression.
7534 * zap-to-char body:: A short overview.
7535 * search-forward:: How to search for a string.
7536 * progn:: The @code{progn} special form.
7537 * Summing up zap-to-char:: Using @code{point} and @code{search-forward}.
7541 @node Complete zap-to-char
7542 @unnumberedsubsec The Complete @code{zap-to-char} Implementation
7545 The @code{zap-to-char} function removes the text in the region between
7546 the location of the cursor (i.e., of point) up to and including the
7547 next occurrence of a specified character. The text that
7548 @code{zap-to-char} removes is put in the kill ring; and it can be
7549 retrieved from the kill ring by typing @kbd{C-y} (@code{yank}). If
7550 the command is given an argument, it removes text through that number
7551 of occurrences. Thus, if the cursor were at the beginning of this
7552 sentence and the character were @samp{s}, @samp{Thus} would be
7553 removed. If the argument were two, @samp{Thus, if the curs} would be
7554 removed, up to and including the @samp{s} in @samp{cursor}.
7556 If the specified character is not found, @code{zap-to-char} will say
7557 ``Search failed'', tell you the character you typed, and not remove
7560 In order to determine how much text to remove, @code{zap-to-char} uses
7561 a search function. Searches are used extensively in code that
7562 manipulates text, and we will focus attention on them as well as on the
7566 @c GNU Emacs version 19
7567 (defun zap-to-char (arg char) ; version 19 implementation
7568 "Kill up to and including ARG'th occurrence of CHAR.
7569 Goes backward if ARG is negative; error if CHAR not found."
7570 (interactive "*p\ncZap to char: ")
7571 (kill-region (point)
7574 (char-to-string char) nil nil arg)
7579 Here is the complete text of the version 22 implementation of the function:
7584 (defun zap-to-char (arg char)
7585 "Kill up to and including ARG'th occurrence of CHAR.
7586 Case is ignored if `case-fold-search' is non-nil in the current buffer.
7587 Goes backward if ARG is negative; error if CHAR not found."
7588 (interactive "p\ncZap to char: ")
7589 (if (char-table-p translation-table-for-input)
7590 (setq char (or (aref translation-table-for-input char) char)))
7591 (kill-region (point) (progn
7592 (search-forward (char-to-string char)
7598 The documentation is thorough. You do need to know the jargon meaning
7599 of the word ``kill''.
7601 @cindex curved quotes
7602 @cindex curly quotes
7603 The version 22 documentation string for @code{zap-to-char} uses ASCII
7604 grave accent and apostrophe to quote a symbol, so it appears as
7605 @t{`case-fold-search'}. This quoting style was inspired by 1970s-era
7606 displays in which grave accent and apostrophe were often mirror images
7607 suitable for use as quotes. On most modern displays this is no longer
7608 true, and when these two ASCII characters appear in documentation
7609 strings or diagnostic message formats, Emacs typically transliterates
7610 them to @dfn{curved quotes} (left and right single quotation marks),
7611 so that the abovequoted symbol appears
7612 as @t{‘case-fold-search’}. Source-code strings can also simply use
7613 curved quotes directly.
7615 @node zap-to-char interactive
7616 @subsection The @code{interactive} Expression
7619 The interactive expression in the @code{zap-to-char} command looks like
7623 (interactive "p\ncZap to char: ")
7626 The part within quotation marks, @code{"p\ncZap to char:@: "}, specifies
7627 two different things. First, and most simply, is the @samp{p}.
7628 This part is separated from the next part by a newline, @samp{\n}.
7629 The @samp{p} means that the first argument to the function will be
7630 passed the value of a @dfn{processed prefix}. The prefix argument is
7631 passed by typing @kbd{C-u} and a number, or @kbd{M-} and a number. If
7632 the function is called interactively without a prefix, 1 is passed to
7635 The second part of @code{"p\ncZap to char:@: "} is
7636 @samp{cZap to char:@: }. In this part, the lower case @samp{c}
7637 indicates that @code{interactive} expects a prompt and that the
7638 argument will be a character. The prompt follows the @samp{c} and is
7639 the string @samp{Zap to char:@: } (with a space after the colon to
7642 What all this does is prepare the arguments to @code{zap-to-char} so they
7643 are of the right type, and give the user a prompt.
7645 In a read-only buffer, the @code{zap-to-char} function copies the text
7646 to the kill ring, but does not remove it. The echo area displays a
7647 message saying that the buffer is read-only. Also, the terminal may
7648 beep or blink at you.
7650 @node zap-to-char body
7651 @subsection The Body of @code{zap-to-char}
7653 The body of the @code{zap-to-char} function contains the code that
7654 kills (that is, removes) the text in the region from the current
7655 position of the cursor up to and including the specified character.
7657 The first part of the code looks like this:
7660 (if (char-table-p translation-table-for-input)
7661 (setq char (or (aref translation-table-for-input char) char)))
7662 (kill-region (point) (progn
7663 (search-forward (char-to-string char) nil nil arg)
7668 @code{char-table-p} is an hitherto unseen function. It determines
7669 whether its argument is a character table. When it is, it sets the
7670 character passed to @code{zap-to-char} to one of them, if that
7671 character exists, or to the character itself. (This becomes important
7672 for certain characters in non-European languages. The @code{aref}
7673 function extracts an element from an array. It is an array-specific
7674 function that is not described in this document. @xref{Arrays, ,
7675 Arrays, elisp, The GNU Emacs Lisp Reference Manual}.)
7678 @code{(point)} is the current position of the cursor.
7680 The next part of the code is an expression using @code{progn}. The body
7681 of the @code{progn} consists of calls to @code{search-forward} and
7684 It is easier to understand how @code{progn} works after learning about
7685 @code{search-forward}, so we will look at @code{search-forward} and
7686 then at @code{progn}.
7688 @node search-forward
7689 @subsection The @code{search-forward} Function
7690 @findex search-forward
7692 The @code{search-forward} function is used to locate the
7693 zapped-for-character in @code{zap-to-char}. If the search is
7694 successful, @code{search-forward} leaves point immediately after the
7695 last character in the target string. (In @code{zap-to-char}, the
7696 target string is just one character long. @code{zap-to-char} uses the
7697 function @code{char-to-string} to ensure that the computer treats that
7698 character as a string.) If the search is backwards,
7699 @code{search-forward} leaves point just before the first character in
7700 the target. Also, @code{search-forward} returns @code{t} for true.
7701 (Moving point is therefore a side effect.)
7704 In @code{zap-to-char}, the @code{search-forward} function looks like this:
7707 (search-forward (char-to-string char) nil nil arg)
7710 The @code{search-forward} function takes four arguments:
7714 The first argument is the target, what is searched for. This must be a
7715 string, such as @samp{"z"}.
7717 As it happens, the argument passed to @code{zap-to-char} is a single
7718 character. Because of the way computers are built, the Lisp
7719 interpreter may treat a single character as being different from a
7720 string of characters. Inside the computer, a single character has a
7721 different electronic format than a string of one character. (A single
7722 character can often be recorded in the computer using exactly one
7723 byte; but a string may be longer, and the computer needs to be ready
7724 for this.) Since the @code{search-forward} function searches for a
7725 string, the character that the @code{zap-to-char} function receives as
7726 its argument must be converted inside the computer from one format to
7727 the other; otherwise the @code{search-forward} function will fail.
7728 The @code{char-to-string} function is used to make this conversion.
7731 The second argument bounds the search; it is specified as a position in
7732 the buffer. In this case, the search can go to the end of the buffer,
7733 so no bound is set and the second argument is @code{nil}.
7736 The third argument tells the function what it should do if the search
7737 fails---it can signal an error (and print a message) or it can return
7738 @code{nil}. A @code{nil} as the third argument causes the function to
7739 signal an error when the search fails.
7742 The fourth argument to @code{search-forward} is the repeat count---how
7743 many occurrences of the string to look for. This argument is optional
7744 and if the function is called without a repeat count, this argument is
7745 passed the value 1. If this argument is negative, the search goes
7750 In template form, a @code{search-forward} expression looks like this:
7754 (search-forward "@var{target-string}"
7755 @var{limit-of-search}
7756 @var{what-to-do-if-search-fails}
7761 We will look at @code{progn} next.
7764 @subsection The @code{progn} Special Form
7767 @code{progn} is a special form that causes each of its arguments to be
7768 evaluated in sequence and then returns the value of the last one. The
7769 preceding expressions are evaluated only for the side effects they
7770 perform. The values produced by them are discarded.
7773 The template for a @code{progn} expression is very simple:
7782 In @code{zap-to-char}, the @code{progn} expression has to do two things:
7783 put point in exactly the right position; and return the location of
7784 point so that @code{kill-region} will know how far to kill to.
7786 The first argument to the @code{progn} is @code{search-forward}. When
7787 @code{search-forward} finds the string, the function leaves point
7788 immediately after the last character in the target string. (In this
7789 case the target string is just one character long.) If the search is
7790 backwards, @code{search-forward} leaves point just before the first
7791 character in the target. The movement of point is a side effect.
7793 The second and last argument to @code{progn} is the expression
7794 @code{(point)}. This expression returns the value of point, which in
7795 this case will be the location to which it has been moved by
7796 @code{search-forward}. (In the source, a line that tells the function
7797 to go to the previous character, if it is going forward, was commented
7798 out in 1999; I don't remember whether that feature or mis-feature was
7799 ever a part of the distributed source.) The value of @code{point} is
7800 returned by the @code{progn} expression and is passed to
7801 @code{kill-region} as @code{kill-region}'s second argument.
7803 @node Summing up zap-to-char
7804 @subsection Summing up @code{zap-to-char}
7806 Now that we have seen how @code{search-forward} and @code{progn} work,
7807 we can see how the @code{zap-to-char} function works as a whole.
7809 The first argument to @code{kill-region} is the position of the cursor
7810 when the @code{zap-to-char} command is given---the value of point at
7811 that time. Within the @code{progn}, the search function then moves
7812 point to just after the zapped-to-character and @code{point} returns the
7813 value of this location. The @code{kill-region} function puts together
7814 these two values of point, the first one as the beginning of the region
7815 and the second one as the end of the region, and removes the region.
7817 The @code{progn} special form is necessary because the
7818 @code{kill-region} command takes two arguments; and it would fail if
7819 @code{search-forward} and @code{point} expressions were written in
7820 sequence as two additional arguments. The @code{progn} expression is
7821 a single argument to @code{kill-region} and returns the one value that
7822 @code{kill-region} needs for its second argument.
7825 @section @code{kill-region}
7828 The @code{zap-to-char} function uses the @code{kill-region} function.
7829 This function clips text from a region and copies that text to
7830 the kill ring, from which it may be retrieved.
7835 (defun kill-region (beg end &optional yank-handler)
7836 "Kill (\"cut\") text between point and mark.
7837 This deletes the text from the buffer and saves it in the kill ring.
7838 The command \\[yank] can retrieve it from there.
7839 \(If you want to kill and then yank immediately, use \\[kill-ring-save].)
7841 If you want to append the killed region to the last killed text,
7842 use \\[append-next-kill] before \\[kill-region].
7844 If the buffer is read-only, Emacs will beep and refrain from deleting
7845 the text, but put the text in the kill ring anyway. This means that
7846 you can use the killing commands to copy text from a read-only buffer.
7848 This is the primitive for programs to kill text (as opposed to deleting it).
7849 Supply two arguments, character positions indicating the stretch of text
7851 Any command that calls this function is a \"kill command\".
7852 If the previous command was also a kill command,
7853 the text killed this time appends to the text killed last time
7854 to make one entry in the kill ring.
7856 In Lisp code, optional third arg YANK-HANDLER, if non-nil,
7857 specifies the yank-handler text property to be set on the killed
7858 text. See `insert-for-yank'."
7859 ;; Pass point first, then mark, because the order matters
7860 ;; when calling kill-append.
7861 (interactive (list (point) (mark)))
7862 (unless (and beg end)
7863 (error "The mark is not set now, so there is no region"))
7865 (let ((string (filter-buffer-substring beg end t)))
7866 (when string ;STRING is nil if BEG = END
7867 ;; Add that string to the kill ring, one way or another.
7868 (if (eq last-command 'kill-region)
7869 (kill-append string (< end beg) yank-handler)
7870 (kill-new string nil yank-handler)))
7871 (when (or string (eq last-command 'kill-region))
7872 (setq this-command 'kill-region))
7874 ((buffer-read-only text-read-only)
7875 ;; The code above failed because the buffer, or some of the characters
7876 ;; in the region, are read-only.
7877 ;; We should beep, in case the user just isn't aware of this.
7878 ;; However, there's no harm in putting
7879 ;; the region's text in the kill ring, anyway.
7880 (copy-region-as-kill beg end)
7881 ;; Set this-command now, so it will be set even if we get an error.
7882 (setq this-command 'kill-region)
7883 ;; This should barf, if appropriate, and give us the correct error.
7884 (if kill-read-only-ok
7885 (progn (message "Read only text copied to kill ring") nil)
7886 ;; Signal an error if the buffer is read-only.
7887 (barf-if-buffer-read-only)
7888 ;; If the buffer isn't read-only, the text is.
7889 (signal 'text-read-only (list (current-buffer)))))))
7892 The Emacs 22 version of that function uses @code{condition-case} and
7893 @code{copy-region-as-kill}, both of which we will explain.
7894 @code{condition-case} is an important special form.
7896 In essence, the @code{kill-region} function calls
7897 @code{condition-case}, which takes three arguments. In this function,
7898 the first argument does nothing. The second argument contains the
7899 code that does the work when all goes well. The third argument
7900 contains the code that is called in the event of an error.
7903 * Complete kill-region:: The function definition.
7904 * condition-case:: Dealing with a problem.
7909 @node Complete kill-region
7910 @unnumberedsubsec The Complete @code{kill-region} Definition
7914 We will go through the @code{condition-case} code in a moment. First,
7915 let us look at the definition of @code{kill-region}, with comments
7921 (defun kill-region (beg end)
7922 "Kill (\"cut\") text between point and mark.
7923 This deletes the text from the buffer and saves it in the kill ring.
7924 The command \\[yank] can retrieve it from there. @dots{} "
7928 ;; @bullet{} Since order matters, pass point first.
7929 (interactive (list (point) (mark)))
7930 ;; @bullet{} And tell us if we cannot cut the text.
7931 ;; 'unless' is an 'if' without a then-part.
7932 (unless (and beg end)
7933 (error "The mark is not set now, so there is no region"))
7937 ;; @bullet{} 'condition-case' takes three arguments.
7938 ;; If the first argument is nil, as it is here,
7939 ;; information about the error signal is not
7940 ;; stored for use by another function.
7945 ;; @bullet{} The second argument to 'condition-case' tells the
7946 ;; Lisp interpreter what to do when all goes well.
7950 ;; It starts with a 'let' function that extracts the string
7951 ;; and tests whether it exists. If so (that is what the
7952 ;; 'when' checks), it calls an 'if' function that determines
7953 ;; whether the previous command was another call to
7954 ;; 'kill-region'; if it was, then the new text is appended to
7955 ;; the previous text; if not, then a different function,
7956 ;; 'kill-new', is called.
7960 ;; The 'kill-append' function concatenates the new string and
7961 ;; the old. The 'kill-new' function inserts text into a new
7962 ;; item in the kill ring.
7966 ;; 'when' is an 'if' without an else-part. The second 'when'
7967 ;; again checks whether the current string exists; in
7968 ;; addition, it checks whether the previous command was
7969 ;; another call to 'kill-region'. If one or the other
7970 ;; condition is true, then it sets the current command to
7971 ;; be 'kill-region'.
7974 (let ((string (filter-buffer-substring beg end t)))
7975 (when string ;STRING is nil if BEG = END
7976 ;; Add that string to the kill ring, one way or another.
7977 (if (eq last-command 'kill-region)
7980 ;; @minus{} 'yank-handler' is an optional argument to
7981 ;; 'kill-region' that tells the 'kill-append' and
7982 ;; 'kill-new' functions how deal with properties
7983 ;; added to the text, such as 'bold' or 'italics'.
7984 (kill-append string (< end beg) yank-handler)
7985 (kill-new string nil yank-handler)))
7986 (when (or string (eq last-command 'kill-region))
7987 (setq this-command 'kill-region))
7992 ;; @bullet{} The third argument to 'condition-case' tells the interpreter
7993 ;; what to do with an error.
7996 ;; The third argument has a conditions part and a body part.
7997 ;; If the conditions are met (in this case,
7998 ;; if text or buffer are read-only)
7999 ;; then the body is executed.
8002 ;; The first part of the third argument is the following:
8003 ((buffer-read-only text-read-only) ;; the if-part
8004 ;; @dots{} the then-part
8005 (copy-region-as-kill beg end)
8008 ;; Next, also as part of the then-part, set this-command, so
8009 ;; it will be set in an error
8010 (setq this-command 'kill-region)
8011 ;; Finally, in the then-part, send a message if you may copy
8012 ;; the text to the kill ring without signaling an error, but
8013 ;; don't if you may not.
8016 (if kill-read-only-ok
8017 (progn (message "Read only text copied to kill ring") nil)
8018 (barf-if-buffer-read-only)
8019 ;; If the buffer isn't read-only, the text is.
8020 (signal 'text-read-only (list (current-buffer)))))
8028 (defun kill-region (beg end)
8029 "Kill between point and mark.
8030 The text is deleted but saved in the kill ring."
8035 ;; 1. 'condition-case' takes three arguments.
8036 ;; If the first argument is nil, as it is here,
8037 ;; information about the error signal is not
8038 ;; stored for use by another function.
8043 ;; 2. The second argument to 'condition-case'
8044 ;; tells the Lisp interpreter what to do when all goes well.
8048 ;; The 'delete-and-extract-region' function usually does the
8049 ;; work. If the beginning and ending of the region are both
8050 ;; the same, then the variable 'string' will be empty, or nil
8051 (let ((string (delete-and-extract-region beg end)))
8055 ;; 'when' is an 'if' clause that cannot take an 'else-part'.
8056 ;; Emacs normally sets the value of 'last-command' to the
8057 ;; previous command.
8060 ;; 'kill-append' concatenates the new string and the old.
8061 ;; 'kill-new' inserts text into a new item in the kill ring.
8063 (if (eq last-command 'kill-region)
8064 ;; if true, prepend string
8065 (kill-append string (< end beg))
8067 (setq this-command 'kill-region))
8071 ;; 3. The third argument to 'condition-case' tells the interpreter
8072 ;; what to do with an error.
8075 ;; The third argument has a conditions part and a body part.
8076 ;; If the conditions are met (in this case,
8077 ;; if text or buffer are read-only)
8078 ;; then the body is executed.
8081 ((buffer-read-only text-read-only) ;; this is the if-part
8083 (copy-region-as-kill beg end)
8086 (if kill-read-only-ok ;; usually this variable is nil
8087 (message "Read only text copied to kill ring")
8088 ;; or else, signal an error if the buffer is read-only;
8089 (barf-if-buffer-read-only)
8090 ;; and, in any case, signal that the text is read-only.
8091 (signal 'text-read-only (list (current-buffer)))))))
8096 @node condition-case
8097 @subsection @code{condition-case}
8098 @findex condition-case
8100 As we have seen earlier (@pxref{Making Errors, , Generate an Error
8101 Message}), when the Emacs Lisp interpreter has trouble evaluating an
8102 expression, it provides you with help; in the jargon, this is called
8103 ``signaling an error''. Usually, the computer stops the program and
8104 shows you a message.
8106 However, some programs undertake complicated actions. They should not
8107 simply stop on an error. In the @code{kill-region} function, the most
8108 likely error is that you will try to kill text that is read-only and
8109 cannot be removed. So the @code{kill-region} function contains code
8110 to handle this circumstance. This code, which makes up the body of
8111 the @code{kill-region} function, is inside of a @code{condition-case}
8115 The template for @code{condition-case} looks like this:
8122 @var{error-handler}@dots{})
8126 The second argument, @var{bodyform}, is straightforward. The
8127 @code{condition-case} special form causes the Lisp interpreter to
8128 evaluate the code in @var{bodyform}. If no error occurs, the special
8129 form returns the code's value and produces the side-effects, if any.
8131 In short, the @var{bodyform} part of a @code{condition-case}
8132 expression determines what should happen when everything works
8135 However, if an error occurs, among its other actions, the function
8136 generating the error signal will define one or more error condition
8139 An error handler is the third argument to @code{condition-case}.
8140 An error handler has two parts, a @var{condition-name} and a
8141 @var{body}. If the @var{condition-name} part of an error handler
8142 matches a condition name generated by an error, then the @var{body}
8143 part of the error handler is run.
8145 As you will expect, the @var{condition-name} part of an error handler
8146 may be either a single condition name or a list of condition names.
8148 Also, a complete @code{condition-case} expression may contain more
8149 than one error handler. When an error occurs, the first applicable
8152 Lastly, the first argument to the @code{condition-case} expression,
8153 the @var{var} argument, is sometimes bound to a variable that
8154 contains information about the error. However, if that argument is
8155 nil, as is the case in @code{kill-region}, that information is
8159 In brief, in the @code{kill-region} function, the code
8160 @code{condition-case} works like this:
8164 @var{If no errors}, @var{run only this code}
8165 @var{but}, @var{if errors}, @var{run this other code}.
8172 copy-region-as-kill is short, 12 lines, and uses
8173 filter-buffer-substring, which is longer, 39 lines
8174 and has delete-and-extract-region in it.
8175 delete-and-extract-region is written in C.
8177 see Initializing a Variable with @code{defvar}
8179 Initializing a Variable with @code{defvar} includes line 8350
8183 @subsection Lisp macro
8187 The part of the @code{condition-case} expression that is evaluated in
8188 the expectation that all goes well has a @code{when}. The code uses
8189 @code{when} to determine whether the @code{string} variable points to
8192 A @code{when} expression is simply a programmers' convenience. It is
8193 an @code{if} without the possibility of an else clause. In your mind,
8194 you can replace @code{when} with @code{if} and understand what goes
8195 on. That is what the Lisp interpreter does.
8197 Technically speaking, @code{when} is a Lisp macro. A Lisp macro
8198 enables you to define new control constructs and other language
8199 features. It tells the interpreter how to compute another Lisp
8200 expression which will in turn compute the value. In this case, the
8201 other expression is an @code{if} expression.
8203 The @code{kill-region} function definition also has an @code{unless}
8204 macro; it is the converse of @code{when}. The @code{unless} macro is
8205 an @code{if} without a then clause
8207 For more about Lisp macros, see @ref{Macros, , Macros, elisp, The GNU
8208 Emacs Lisp Reference Manual}. The C programming language also
8209 provides macros. These are different, but also useful.
8212 We will briefly look at C macros in
8213 @ref{Digression into C}.
8217 Regarding the @code{when} macro, in the @code{condition-case}
8218 expression, when the string has content, then another conditional
8219 expression is executed. This is an @code{if} with both a then-part
8224 (if (eq last-command 'kill-region)
8225 (kill-append string (< end beg) yank-handler)
8226 (kill-new string nil yank-handler))
8230 The then-part is evaluated if the previous command was another call to
8231 @code{kill-region}; if not, the else-part is evaluated.
8233 @code{yank-handler} is an optional argument to @code{kill-region} that
8234 tells the @code{kill-append} and @code{kill-new} functions how deal
8235 with properties added to the text, such as bold or italics.
8237 @code{last-command} is a variable that comes with Emacs that we have
8238 not seen before. Normally, whenever a function is executed, Emacs
8239 sets the value of @code{last-command} to the previous command.
8242 In this segment of the definition, the @code{if} expression checks
8243 whether the previous command was @code{kill-region}. If it was,
8246 (kill-append string (< end beg) yank-handler)
8250 concatenates a copy of the newly clipped text to the just previously
8251 clipped text in the kill ring.
8253 @node copy-region-as-kill
8254 @section @code{copy-region-as-kill}
8255 @findex copy-region-as-kill
8258 The @code{copy-region-as-kill} function copies a region of text from a
8259 buffer and (via either @code{kill-append} or @code{kill-new}) saves it
8260 in the @code{kill-ring}.
8262 If you call @code{copy-region-as-kill} immediately after a
8263 @code{kill-region} command, Emacs appends the newly copied text to the
8264 previously copied text. This means that if you yank back the text, you
8265 get it all, from both this and the previous operation. On the other
8266 hand, if some other command precedes the @code{copy-region-as-kill},
8267 the function copies the text into a separate entry in the kill ring.
8270 * Complete copy-region-as-kill:: The complete function definition.
8271 * copy-region-as-kill body:: The body of @code{copy-region-as-kill}.
8275 @node Complete copy-region-as-kill
8276 @unnumberedsubsec The complete @code{copy-region-as-kill} function definition
8280 Here is the complete text of the version 22 @code{copy-region-as-kill}
8285 (defun copy-region-as-kill (beg end)
8286 "Save the region as if killed, but don't kill it.
8287 In Transient Mark mode, deactivate the mark.
8288 If `interprogram-cut-function' is non-nil, also save the text for a window
8289 system cut and paste."
8293 (if (eq last-command 'kill-region)
8294 (kill-append (filter-buffer-substring beg end) (< end beg))
8295 (kill-new (filter-buffer-substring beg end)))
8298 (if transient-mark-mode
8299 (setq deactivate-mark t))
8305 As usual, this function can be divided into its component parts:
8309 (defun copy-region-as-kill (@var{argument-list})
8310 "@var{documentation}@dots{}"
8316 The arguments are @code{beg} and @code{end} and the function is
8317 interactive with @code{"r"}, so the two arguments must refer to the
8318 beginning and end of the region. If you have been reading through this
8319 document from the beginning, understanding these parts of a function is
8320 almost becoming routine.
8322 The documentation is somewhat confusing unless you remember that the
8323 word ``kill'' has a meaning different from usual. The Transient Mark
8324 and @code{interprogram-cut-function} comments explain certain
8327 After you once set a mark, a buffer always contains a region. If you
8328 wish, you can use Transient Mark mode to highlight the region
8329 temporarily. (No one wants to highlight the region all the time, so
8330 Transient Mark mode highlights it only at appropriate times. Many
8331 people turn off Transient Mark mode, so the region is never
8334 Also, a windowing system allows you to copy, cut, and paste among
8335 different programs. In the X windowing system, for example, the
8336 @code{interprogram-cut-function} function is @code{x-select-text},
8337 which works with the windowing system's equivalent of the Emacs kill
8340 The body of the @code{copy-region-as-kill} function starts with an
8341 @code{if} clause. What this clause does is distinguish between two
8342 different situations: whether or not this command is executed
8343 immediately after a previous @code{kill-region} command. In the first
8344 case, the new region is appended to the previously copied text.
8345 Otherwise, it is inserted into the beginning of the kill ring as a
8346 separate piece of text from the previous piece.
8348 The last two lines of the function prevent the region from lighting up
8349 if Transient Mark mode is turned on.
8351 The body of @code{copy-region-as-kill} merits discussion in detail.
8353 @node copy-region-as-kill body
8354 @subsection The Body of @code{copy-region-as-kill}
8356 The @code{copy-region-as-kill} function works in much the same way as
8357 the @code{kill-region} function. Both are written so that two or more
8358 kills in a row combine their text into a single entry. If you yank
8359 back the text from the kill ring, you get it all in one piece.
8360 Moreover, kills that kill forward from the current position of the
8361 cursor are added to the end of the previously copied text and commands
8362 that copy text backwards add it to the beginning of the previously
8363 copied text. This way, the words in the text stay in the proper
8366 Like @code{kill-region}, the @code{copy-region-as-kill} function makes
8367 use of the @code{last-command} variable that keeps track of the
8368 previous Emacs command.
8371 * last-command & this-command::
8372 * kill-append function::
8373 * kill-new function::
8377 @node last-command & this-command
8378 @unnumberedsubsubsec @code{last-command} and @code{this-command}
8381 Normally, whenever a function is executed, Emacs sets the value of
8382 @code{this-command} to the function being executed (which in this case
8383 would be @code{copy-region-as-kill}). At the same time, Emacs sets
8384 the value of @code{last-command} to the previous value of
8385 @code{this-command}.
8387 In the first part of the body of the @code{copy-region-as-kill}
8388 function, an @code{if} expression determines whether the value of
8389 @code{last-command} is @code{kill-region}. If so, the then-part of
8390 the @code{if} expression is evaluated; it uses the @code{kill-append}
8391 function to concatenate the text copied at this call to the function
8392 with the text already in the first element (the @sc{car}) of the kill
8393 ring. On the other hand, if the value of @code{last-command} is not
8394 @code{kill-region}, then the @code{copy-region-as-kill} function
8395 attaches a new element to the kill ring using the @code{kill-new}
8399 The @code{if} expression reads as follows; it uses @code{eq}:
8403 (if (eq last-command 'kill-region)
8405 (kill-append (filter-buffer-substring beg end) (< end beg))
8407 (kill-new (filter-buffer-substring beg end)))
8411 @findex filter-buffer-substring
8412 (The @code{filter-buffer-substring} function returns a filtered
8413 substring of the buffer, if any. Optionally---the arguments are not
8414 here, so neither is done---the function may delete the initial text or
8415 return the text without its properties; this function is a replacement
8416 for the older @code{buffer-substring} function, which came before text
8417 properties were implemented.)
8419 @findex eq @r{(example of use)}
8421 The @code{eq} function tests whether its first argument is the same Lisp
8422 object as its second argument. The @code{eq} function is similar to the
8423 @code{equal} function in that it is used to test for equality, but
8424 differs in that it determines whether two representations are actually
8425 the same object inside the computer, but with different names.
8426 @code{equal} determines whether the structure and contents of two
8427 expressions are the same.
8429 If the previous command was @code{kill-region}, then the Emacs Lisp
8430 interpreter calls the @code{kill-append} function
8432 @node kill-append function
8433 @unnumberedsubsubsec The @code{kill-append} function
8437 The @code{kill-append} function looks like this:
8442 (defun kill-append (string before-p &optional yank-handler)
8443 "Append STRING to the end of the latest kill in the kill ring.
8444 If BEFORE-P is non-nil, prepend STRING to the kill.
8446 (let* ((cur (car kill-ring)))
8447 (kill-new (if before-p (concat string cur) (concat cur string))
8448 (or (= (length cur) 0)
8450 (get-text-property 0 'yank-handler cur)))
8457 (defun kill-append (string before-p)
8458 "Append STRING to the end of the latest kill in the kill ring.
8459 If BEFORE-P is non-nil, prepend STRING to the kill.
8460 If `interprogram-cut-function' is set, pass the resulting kill to
8462 (kill-new (if before-p
8463 (concat string (car kill-ring))
8464 (concat (car kill-ring) string))
8469 The @code{kill-append} function is fairly straightforward. It uses
8470 the @code{kill-new} function, which we will discuss in more detail in
8473 (Also, the function provides an optional argument called
8474 @code{yank-handler}; when invoked, this argument tells the function
8475 how to deal with properties added to the text, such as bold or
8478 @c !!! bug in GNU Emacs 22 version of kill-append ?
8479 It has a @code{let*} function to set the value of the first element of
8480 the kill ring to @code{cur}. (I do not know why the function does not
8481 use @code{let} instead; only one value is set in the expression.
8482 Perhaps this is a bug that produces no problems?)
8484 Consider the conditional that is one of the two arguments to
8485 @code{kill-new}. It uses @code{concat} to concatenate the new text to
8486 the @sc{car} of the kill ring. Whether it prepends or appends the
8487 text depends on the results of an @code{if} expression:
8491 (if before-p ; @r{if-part}
8492 (concat string cur) ; @r{then-part}
8493 (concat cur string)) ; @r{else-part}
8498 If the region being killed is before the region that was killed in the
8499 last command, then it should be prepended before the material that was
8500 saved in the previous kill; and conversely, if the killed text follows
8501 what was just killed, it should be appended after the previous text.
8502 The @code{if} expression depends on the predicate @code{before-p} to
8503 decide whether the newly saved text should be put before or after the
8504 previously saved text.
8506 The symbol @code{before-p} is the name of one of the arguments to
8507 @code{kill-append}. When the @code{kill-append} function is
8508 evaluated, it is bound to the value returned by evaluating the actual
8509 argument. In this case, this is the expression @code{(< end beg)}.
8510 This expression does not directly determine whether the killed text in
8511 this command is located before or after the kill text of the last
8512 command; what it does is determine whether the value of the variable
8513 @code{end} is less than the value of the variable @code{beg}. If it
8514 is, it means that the user is most likely heading towards the
8515 beginning of the buffer. Also, the result of evaluating the predicate
8516 expression, @code{(< end beg)}, will be true and the text will be
8517 prepended before the previous text. On the other hand, if the value of
8518 the variable @code{end} is greater than the value of the variable
8519 @code{beg}, the text will be appended after the previous text.
8522 When the newly saved text will be prepended, then the string with the new
8523 text will be concatenated before the old text:
8531 But if the text will be appended, it will be concatenated
8535 (concat cur string))
8538 To understand how this works, we first need to review the
8539 @code{concat} function. The @code{concat} function links together or
8540 unites two strings of text. The result is a string. For example:
8544 (concat "abc" "def")
8550 (car '("first element" "second element")))
8551 @result{} "new first element"
8554 '("first element" "second element")) " modified")
8555 @result{} "first element modified"
8559 We can now make sense of @code{kill-append}: it modifies the contents
8560 of the kill ring. The kill ring is a list, each element of which is
8561 saved text. The @code{kill-append} function uses the @code{kill-new}
8562 function which in turn uses the @code{setcar} function.
8564 @node kill-new function
8565 @unnumberedsubsubsec The @code{kill-new} function
8569 In version 22 the @code{kill-new} function looks like this:
8573 (defun kill-new (string &optional replace yank-handler)
8574 "Make STRING the latest kill in the kill ring.
8575 Set `kill-ring-yank-pointer' to point to it.
8577 If `interprogram-cut-function' is non-nil, apply it to STRING.
8578 Optional second argument REPLACE non-nil means that STRING will replace
8579 the front of the kill ring, rather than being added to the list.
8583 (if (> (length string) 0)
8585 (put-text-property 0 (length string)
8586 'yank-handler yank-handler string))
8588 (signal 'args-out-of-range
8589 (list string "yank-handler specified for empty string"))))
8592 (if (fboundp 'menu-bar-update-yank-menu)
8593 (menu-bar-update-yank-menu string (and replace (car kill-ring))))
8596 (if (and replace kill-ring)
8597 (setcar kill-ring string)
8598 (push string kill-ring)
8599 (if (> (length kill-ring) kill-ring-max)
8600 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)))
8603 (setq kill-ring-yank-pointer kill-ring)
8604 (if interprogram-cut-function
8605 (funcall interprogram-cut-function string (not replace))))
8610 (defun kill-new (string &optional replace)
8611 "Make STRING the latest kill in the kill ring.
8612 Set the kill-ring-yank pointer to point to it.
8613 If `interprogram-cut-function' is non-nil, apply it to STRING.
8614 Optional second argument REPLACE non-nil means that STRING will replace
8615 the front of the kill ring, rather than being added to the list."
8616 (and (fboundp 'menu-bar-update-yank-menu)
8617 (menu-bar-update-yank-menu string (and replace (car kill-ring))))
8618 (if (and replace kill-ring)
8619 (setcar kill-ring string)
8620 (setq kill-ring (cons string kill-ring))
8621 (if (> (length kill-ring) kill-ring-max)
8622 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)))
8623 (setq kill-ring-yank-pointer kill-ring)
8624 (if interprogram-cut-function
8625 (funcall interprogram-cut-function string (not replace))))
8628 (Notice that the function is not interactive.)
8630 As usual, we can look at this function in parts.
8632 The function definition has an optional @code{yank-handler} argument,
8633 which when invoked tells the function how to deal with properties
8634 added to the text, such as bold or italics. We will skip that.
8637 The first line of the documentation makes sense:
8640 Make STRING the latest kill in the kill ring.
8644 Let's skip over the rest of the documentation for the moment.
8647 Also, let's skip over the initial @code{if} expression and those lines
8648 of code involving @code{menu-bar-update-yank-menu}. We will explain
8652 The critical lines are these:
8656 (if (and replace kill-ring)
8658 (setcar kill-ring string)
8662 (push string kill-ring)
8665 (if (> (length kill-ring) kill-ring-max)
8666 ;; @r{avoid overly long kill ring}
8667 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)))
8670 (setq kill-ring-yank-pointer kill-ring)
8671 (if interprogram-cut-function
8672 (funcall interprogram-cut-function string (not replace))))
8676 The conditional test is @w{@code{(and replace kill-ring)}}.
8677 This will be true when two conditions are met: the kill ring has
8678 something in it, and the @code{replace} variable is true.
8681 When the @code{kill-append} function sets @code{replace} to be true
8682 and when the kill ring has at least one item in it, the @code{setcar}
8683 expression is executed:
8686 (setcar kill-ring string)
8689 The @code{setcar} function actually changes the first element of the
8690 @code{kill-ring} list to the value of @code{string}. It replaces the
8694 On the other hand, if the kill ring is empty, or replace is false, the
8695 else-part of the condition is executed:
8698 (push string kill-ring)
8703 @code{push} puts its first argument onto the second. It is similar to
8707 (setq kill-ring (cons string kill-ring))
8715 (add-to-list kill-ring string)
8719 When it is false, the expression first constructs a new version of the
8720 kill ring by prepending @code{string} to the existing kill ring as a
8721 new element (that is what the @code{push} does). Then it executes a
8722 second @code{if} clause. This second @code{if} clause keeps the kill
8723 ring from growing too long.
8725 Let's look at these two expressions in order.
8727 The @code{push} line of the else-part sets the new value of the kill
8728 ring to what results from adding the string being killed to the old
8731 We can see how this works with an example.
8737 (setq example-list '("here is a clause" "another clause"))
8742 After evaluating this expression with @kbd{C-x C-e}, you can evaluate
8743 @code{example-list} and see what it returns:
8748 @result{} ("here is a clause" "another clause")
8754 Now, we can add a new element on to this list by evaluating the
8755 following expression:
8756 @findex push@r{, example}
8759 (push "a third clause" example-list)
8764 When we evaluate @code{example-list}, we find its value is:
8769 @result{} ("a third clause" "here is a clause" "another clause")
8774 Thus, the third clause is added to the list by @code{push}.
8777 Now for the second part of the @code{if} clause. This expression
8778 keeps the kill ring from growing too long. It looks like this:
8782 (if (> (length kill-ring) kill-ring-max)
8783 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil))
8787 The code checks whether the length of the kill ring is greater than
8788 the maximum permitted length. This is the value of
8789 @code{kill-ring-max} (which is 60, by default). If the length of the
8790 kill ring is too long, then this code sets the last element of the
8791 kill ring to @code{nil}. It does this by using two functions,
8792 @code{nthcdr} and @code{setcdr}.
8794 We looked at @code{setcdr} earlier (@pxref{setcdr, , @code{setcdr}}).
8795 It sets the @sc{cdr} of a list, just as @code{setcar} sets the
8796 @sc{car} of a list. In this case, however, @code{setcdr} will not be
8797 setting the @sc{cdr} of the whole kill ring; the @code{nthcdr}
8798 function is used to cause it to set the @sc{cdr} of the next to last
8799 element of the kill ring---this means that since the @sc{cdr} of the
8800 next to last element is the last element of the kill ring, it will set
8801 the last element of the kill ring.
8803 @findex nthcdr@r{, example}
8804 The @code{nthcdr} function works by repeatedly taking the @sc{cdr} of a
8805 list---it takes the @sc{cdr} of the @sc{cdr} of the @sc{cdr}
8806 @dots{} It does this @var{N} times and returns the results.
8807 (@xref{nthcdr, , @code{nthcdr}}.)
8809 @findex setcdr@r{, example}
8810 Thus, if we had a four element list that was supposed to be three
8811 elements long, we could set the @sc{cdr} of the next to last element
8812 to @code{nil}, and thereby shorten the list. (If you set the last
8813 element to some other value than @code{nil}, which you could do, then
8814 you would not have shortened the list. @xref{setcdr, ,
8817 You can see shortening by evaluating the following three expressions
8818 in turn. First set the value of @code{trees} to @code{(maple oak pine
8819 birch)}, then set the @sc{cdr} of its second @sc{cdr} to @code{nil}
8820 and then find the value of @code{trees}:
8824 (setq trees '(maple oak pine birch))
8825 @result{} (maple oak pine birch)
8829 (setcdr (nthcdr 2 trees) nil)
8833 @result{} (maple oak pine)
8838 (The value returned by the @code{setcdr} expression is @code{nil} since
8839 that is what the @sc{cdr} is set to.)
8841 To repeat, in @code{kill-new}, the @code{nthcdr} function takes the
8842 @sc{cdr} a number of times that is one less than the maximum permitted
8843 size of the kill ring and @code{setcdr} sets the @sc{cdr} of that
8844 element (which will be the rest of the elements in the kill ring) to
8845 @code{nil}. This prevents the kill ring from growing too long.
8848 The next to last expression in the @code{kill-new} function is
8851 (setq kill-ring-yank-pointer kill-ring)
8854 The @code{kill-ring-yank-pointer} is a global variable that is set to be
8855 the @code{kill-ring}.
8857 Even though the @code{kill-ring-yank-pointer} is called a
8858 @samp{pointer}, it is a variable just like the kill ring. However, the
8859 name has been chosen to help humans understand how the variable is used.
8862 Now, to return to an early expression in the body of the function:
8866 (if (fboundp 'menu-bar-update-yank-menu)
8867 (menu-bar-update-yank-menu string (and replace (car kill-ring))))
8872 It starts with an @code{if} expression
8874 In this case, the expression tests first to see whether
8875 @code{menu-bar-update-yank-menu} exists as a function, and if so,
8876 calls it. The @code{fboundp} function returns true if the symbol it
8877 is testing has a function definition that is not void. If the
8878 symbol's function definition were void, we would receive an error
8879 message, as we did when we created errors intentionally (@pxref{Making
8880 Errors, , Generate an Error Message}).
8883 The then-part contains an expression whose first element is the
8884 function @code{and}.
8887 The @code{and} special form evaluates each of its arguments until one
8888 of the arguments returns a value of @code{nil}, in which case the
8889 @code{and} expression returns @code{nil}; however, if none of the
8890 arguments returns a value of @code{nil}, the value resulting from
8891 evaluating the last argument is returned. (Since such a value is not
8892 @code{nil}, it is considered true in Emacs Lisp.) In other words, an
8893 @code{and} expression returns a true value only if all its arguments
8894 are true. (@xref{Second Buffer Related Review}.)
8896 The expression determines whether the second argument to
8897 @code{menu-bar-update-yank-menu} is true or not.
8899 ;; If we're supposed to be extending an existing string, and that
8900 ;; string really is at the front of the menu, then update it in place.
8903 @code{menu-bar-update-yank-menu} is one of the functions that make it
8904 possible to use the ``Select and Paste'' menu in the Edit item of a menu
8905 bar; using a mouse, you can look at the various pieces of text you
8906 have saved and select one piece to paste.
8908 The last expression in the @code{kill-new} function adds the newly
8909 copied string to whatever facility exists for copying and pasting
8910 among different programs running in a windowing system. In the X
8911 Windowing system, for example, the @code{x-select-text} function takes
8912 the string and stores it in memory operated by X@. You can paste the
8913 string in another program, such as an Xterm.
8916 The expression looks like this:
8920 (if interprogram-cut-function
8921 (funcall interprogram-cut-function string (not replace))))
8925 If an @code{interprogram-cut-function} exists, then Emacs executes
8926 @code{funcall}, which in turn calls its first argument as a function
8927 and passes the remaining arguments to it. (Incidentally, as far as I
8928 can see, this @code{if} expression could be replaced by an @code{and}
8929 expression similar to the one in the first part of the function.)
8931 We are not going to discuss windowing systems and other programs
8932 further, but merely note that this is a mechanism that enables GNU
8933 Emacs to work easily and well with other programs.
8935 This code for placing text in the kill ring, either concatenated with
8936 an existing element or as a new element, leads us to the code for
8937 bringing back text that has been cut out of the buffer---the yank
8938 commands. However, before discussing the yank commands, it is better
8939 to learn how lists are implemented in a computer. This will make
8940 clear such mysteries as the use of the term ``pointer''. But before
8941 that, we will digress into C.
8944 @c is this true in Emacs 22? Does not seems to be
8946 (If the @w{@code{(< end beg))}}
8947 expression is true, @code{kill-append} prepends the string to the just
8948 previously clipped text. For a detailed discussion, see
8949 @ref{kill-append function, , The @code{kill-append} function}.)
8951 If you then yank back the text, i.e., paste it, you get both
8952 pieces of text at once. That way, if you delete two words in a row,
8953 and then yank them back, you get both words, in their proper order,
8954 with one yank. (The @w{@code{(< end beg))}} expression makes sure the
8957 On the other hand, if the previous command is not @code{kill-region},
8958 then the @code{kill-new} function is called, which adds the text to
8959 the kill ring as the latest item, and sets the
8960 @code{kill-ring-yank-pointer} variable to point to it.
8964 @c Evidently, changed for Emacs 22. The zap-to-char command does not
8965 @c use the delete-and-extract-region function
8967 2006 Oct 26, the Digression into C is now OK but should come after
8968 copy-region-as-kill and filter-buffer-substring
8972 copy-region-as-kill is short, 12 lines, and uses
8973 filter-buffer-substring, which is longer, 39 lines
8974 and has delete-and-extract-region in it.
8975 delete-and-extract-region is written in C.
8977 see Initializing a Variable with @code{defvar}
8980 @node Digression into C
8981 @section Digression into C
8982 @findex delete-and-extract-region
8983 @cindex C, a digression into
8984 @cindex Digression into C
8986 The @code{copy-region-as-kill} function (@pxref{copy-region-as-kill, ,
8987 @code{copy-region-as-kill}}) uses the @code{filter-buffer-substring}
8988 function, which in turn uses the @code{delete-and-extract-region}
8989 function. It removes the contents of a region and you cannot get them
8992 Unlike the other code discussed here, the
8993 @code{delete-and-extract-region} function is not written in Emacs
8994 Lisp; it is written in C and is one of the primitives of the GNU Emacs
8995 system. Since it is very simple, I will digress briefly from Lisp and
8998 @c GNU Emacs 24 in src/editfns.c
8999 @c the DEFUN for delete-and-extract-region
9002 Like many of the other Emacs primitives,
9003 @code{delete-and-extract-region} is written as an instance of a C
9004 macro, a macro being a template for code. The complete macro looks
9009 DEFUN ("delete-and-extract-region", Fdelete_and_extract_region,
9010 Sdelete_and_extract_region, 2, 2, 0,
9011 doc: /* Delete the text between START and END and return it. */)
9012 (Lisp_Object start, Lisp_Object end)
9014 validate_region (&start, &end);
9015 if (XINT (start) == XINT (end))
9016 return empty_unibyte_string;
9017 return del_range_1 (XINT (start), XINT (end), 1, 1);
9022 Without going into the details of the macro writing process, let me
9023 point out that this macro starts with the word @code{DEFUN}. The word
9024 @code{DEFUN} was chosen since the code serves the same purpose as
9025 @code{defun} does in Lisp. (The @code{DEFUN} C macro is defined in
9026 @file{emacs/src/lisp.h}.)
9028 The word @code{DEFUN} is followed by seven parts inside of
9033 The first part is the name given to the function in Lisp,
9034 @code{delete-and-extract-region}.
9037 The second part is the name of the function in C,
9038 @code{Fdelete_and_extract_region}. By convention, it starts with
9039 @samp{F}. Since C does not use hyphens in names, underscores are used
9043 The third part is the name for the C constant structure that records
9044 information on this function for internal use. It is the name of the
9045 function in C but begins with an @samp{S} instead of an @samp{F}.
9048 The fourth and fifth parts specify the minimum and maximum number of
9049 arguments the function can have. This function demands exactly 2
9053 The sixth part is nearly like the argument that follows the
9054 @code{interactive} declaration in a function written in Lisp: a letter
9055 followed, perhaps, by a prompt. The only difference from Lisp is
9056 when the macro is called with no arguments. Then you write a @code{0}
9057 (which is a null string), as in this macro.
9059 If you were to specify arguments, you would place them between
9060 quotation marks. The C macro for @code{goto-char} includes
9061 @code{"NGoto char: "} in this position to indicate that the function
9062 expects a raw prefix, in this case, a numerical location in a buffer,
9063 and provides a prompt.
9066 The seventh part is a documentation string, just like the one for a
9067 function written in Emacs Lisp. This is written as a C comment. (When
9068 you build Emacs, the program @command{lib-src/make-docfile} extracts
9069 these comments and uses them to make the documentation.)
9073 In a C macro, the formal parameters come next, with a statement of
9074 what kind of object they are, followed by the body
9075 of the macro. For @code{delete-and-extract-region} the body
9076 consists of the following four lines:
9080 validate_region (&start, &end);
9081 if (XINT (start) == XINT (end))
9082 return empty_unibyte_string;
9083 return del_range_1 (XINT (start), XINT (end), 1, 1);
9087 The @code{validate_region} function checks whether the values
9088 passed as the beginning and end of the region are the proper type and
9089 are within range. If the beginning and end positions are the same,
9090 then return an empty string.
9092 The @code{del_range_1} function actually deletes the text. It is a
9093 complex function we will not look into. It updates the buffer and
9094 does other things. However, it is worth looking at the two arguments
9095 passed to @code{del_range_1}. These are @w{@code{XINT (start)}} and
9096 @w{@code{XINT (end)}}.
9098 As far as the C language is concerned, @code{start} and @code{end} are
9099 two integers that mark the beginning and end of the region to be
9100 deleted@footnote{More precisely, and requiring more expert knowledge
9101 to understand, the two integers are of type @code{Lisp_Object}, which can
9102 also be a C union instead of an integer type.}.
9104 Integer widths depend on the machine, and are typically 32 or 64 bits.
9105 A few of the bits are used to specify the type of information; the
9106 remaining bits are used as content.
9108 @samp{XINT} is a C macro that extracts the relevant number from the
9109 longer collection of bits; the type bits are discarded.
9112 The command in @code{delete-and-extract-region} looks like this:
9115 del_range_1 (XINT (start), XINT (end), 1, 1);
9119 It deletes the region between the beginning position, @code{start},
9120 and the ending position, @code{end}.
9122 From the point of view of the person writing Lisp, Emacs is all very
9123 simple; but hidden underneath is a great deal of complexity to make it
9127 @section Initializing a Variable with @code{defvar}
9129 @cindex Initializing a variable
9130 @cindex Variable initialization
9135 copy-region-as-kill is short, 12 lines, and uses
9136 filter-buffer-substring, which is longer, 39 lines
9137 and has delete-and-extract-region in it.
9138 delete-and-extract-region is written in C.
9140 see Initializing a Variable with @code{defvar}
9144 The @code{copy-region-as-kill} function is written in Emacs Lisp. Two
9145 functions within it, @code{kill-append} and @code{kill-new}, copy a
9146 region in a buffer and save it in a variable called the
9147 @code{kill-ring}. This section describes how the @code{kill-ring}
9148 variable is created and initialized using the @code{defvar} special
9151 (Again we note that the term @code{kill-ring} is a misnomer. The text
9152 that is clipped out of the buffer can be brought back; it is not a ring
9153 of corpses, but a ring of resurrectable text.)
9155 In Emacs Lisp, a variable such as the @code{kill-ring} is created and
9156 given an initial value by using the @code{defvar} special form. The
9157 name comes from ``define variable''.
9159 The @code{defvar} special form is similar to @code{setq} in that it sets
9160 the value of a variable. It is unlike @code{setq} in two ways: first,
9161 it only sets the value of the variable if the variable does not already
9162 have a value. If the variable already has a value, @code{defvar} does
9163 not override the existing value. Second, @code{defvar} has a
9164 documentation string.
9166 (There is a related macro, @code{defcustom}, designed for variables
9167 that people customize. It has more features than @code{defvar}.
9168 (@xref{defcustom, , Setting Variables with @code{defcustom}}.)
9171 * See variable current value::
9172 * defvar and asterisk::
9176 @node See variable current value
9177 @unnumberedsubsec Seeing the Current Value of a Variable
9180 You can see the current value of a variable, any variable, by using
9181 the @code{describe-variable} function, which is usually invoked by
9182 typing @kbd{C-h v}. If you type @kbd{C-h v} and then @code{kill-ring}
9183 (followed by @key{RET}) when prompted, you will see what is in your
9184 current kill ring---this may be quite a lot! Conversely, if you have
9185 been doing nothing this Emacs session except read this document, you
9186 may have nothing in it. Also, you will see the documentation for
9192 List of killed text sequences.
9193 Since the kill ring is supposed to interact nicely with cut-and-paste
9194 facilities offered by window systems, use of this variable should
9197 interact nicely with `interprogram-cut-function' and
9198 `interprogram-paste-function'. The functions `kill-new',
9199 `kill-append', and `current-kill' are supposed to implement this
9200 interaction; you may want to use them instead of manipulating the kill
9206 The kill ring is defined by a @code{defvar} in the following way:
9210 (defvar kill-ring nil
9211 "List of killed text sequences.
9217 In this variable definition, the variable is given an initial value of
9218 @code{nil}, which makes sense, since if you have saved nothing, you want
9219 nothing back if you give a @code{yank} command. The documentation
9220 string is written just like the documentation string of a @code{defun}.
9221 As with the documentation string of the @code{defun}, the first line of
9222 the documentation should be a complete sentence, since some commands,
9223 like @code{apropos}, print only the first line of documentation.
9224 Succeeding lines should not be indented; otherwise they look odd when
9225 you use @kbd{C-h v} (@code{describe-variable}).
9227 @node defvar and asterisk
9228 @subsection @code{defvar} and an asterisk
9229 @findex defvar @r{for a user customizable variable}
9230 @findex defvar @r{with an asterisk}
9232 In the past, Emacs used the @code{defvar} special form both for
9233 internal variables that you would not expect a user to change and for
9234 variables that you do expect a user to change. Although you can still
9235 use @code{defvar} for user customizable variables, please use
9236 @code{defcustom} instead, since it provides a path into
9237 the Customization commands. (@xref{defcustom, , Specifying Variables
9238 using @code{defcustom}}.)
9240 When you specified a variable using the @code{defvar} special form,
9241 you could distinguish a variable that a user might want to change from
9242 others by typing an asterisk, @samp{*}, in the first column of its
9243 documentation string. For example:
9247 (defvar shell-command-default-error-buffer nil
9248 "*Buffer name for `shell-command' @dots{} error output.
9253 @findex set-variable
9255 You could (and still can) use the @code{set-variable} command to
9256 change the value of @code{shell-command-default-error-buffer}
9257 temporarily. However, options set using @code{set-variable} are set
9258 only for the duration of your editing session. The new values are not
9259 saved between sessions. Each time Emacs starts, it reads the original
9260 value, unless you change the value within your @file{.emacs} file,
9261 either by setting it manually or by using @code{customize}.
9262 @xref{Emacs Initialization, , Your @file{.emacs} File}.
9264 For me, the major use of the @code{set-variable} command is to suggest
9265 variables that I might want to set in my @file{.emacs} file. There
9266 are now more than 700 such variables, far too many to remember
9267 readily. Fortunately, you can press @key{TAB} after calling the
9268 @code{M-x set-variable} command to see the list of variables.
9269 (@xref{Examining, , Examining and Setting Variables, emacs,
9270 The GNU Emacs Manual}.)
9273 @node cons & search-fwd Review
9276 Here is a brief summary of some recently introduced functions.
9281 @code{car} returns the first element of a list; @code{cdr} returns the
9282 second and subsequent elements of a list.
9289 (car '(1 2 3 4 5 6 7))
9291 (cdr '(1 2 3 4 5 6 7))
9292 @result{} (2 3 4 5 6 7)
9297 @code{cons} constructs a list by prepending its first argument to its
9311 @code{funcall} evaluates its first argument as a function. It passes
9312 its remaining arguments to its first argument.
9315 Return the result of taking @sc{cdr} @var{n} times on a list.
9323 The ``rest of the rest'', as it were.
9330 (nthcdr 3 '(1 2 3 4 5 6 7))
9337 @code{setcar} changes the first element of a list; @code{setcdr}
9338 changes the second and subsequent elements of a list.
9345 (setq triple '(1 2 3))
9352 (setcdr triple '("foo" "bar"))
9355 @result{} (37 "foo" "bar")
9360 Evaluate each argument in sequence and then return the value of the
9373 @item save-restriction
9374 Record whatever narrowing is in effect in the current buffer, if any,
9375 and restore that narrowing after evaluating the arguments.
9377 @item search-forward
9378 Search for a string, and if the string is found, move point. With a
9379 regular expression, use the similar @code{re-search-forward}.
9380 (@xref{Regexp Search, , Regular Expression Searches}, for an
9381 explanation of regular expression patterns and searches.)
9385 @code{search-forward} and @code{re-search-forward} take four
9390 The string or regular expression to search for.
9393 Optionally, the limit of the search.
9396 Optionally, what to do if the search fails, return @code{nil} or an
9400 Optionally, how many times to repeat the search; if negative, the
9401 search goes backwards.
9405 @itemx delete-and-extract-region
9406 @itemx copy-region-as-kill
9408 @code{kill-region} cuts the text between point and mark from the
9409 buffer and stores that text in the kill ring, so you can get it back
9412 @code{copy-region-as-kill} copies the text between point and mark into
9413 the kill ring, from which you can get it by yanking. The function
9414 does not cut or remove the text from the buffer.
9417 @code{delete-and-extract-region} removes the text between point and
9418 mark from the buffer and throws it away. You cannot get it back.
9419 (This is not an interactive command.)
9422 @node search Exercises
9423 @section Searching Exercises
9427 Write an interactive function that searches for a string. If the
9428 search finds the string, leave point after it and display a message
9429 that says ``Found!''. (Do not use @code{search-forward} for the name
9430 of this function; if you do, you will overwrite the existing version of
9431 @code{search-forward} that comes with Emacs. Use a name such as
9432 @code{test-search} instead.)
9435 Write a function that prints the third element of the kill ring in the
9436 echo area, if any; if the kill ring does not contain a third element,
9437 print an appropriate message.
9440 @node List Implementation
9441 @chapter How Lists are Implemented
9442 @cindex Lists in a computer
9444 In Lisp, atoms are recorded in a straightforward fashion; if the
9445 implementation is not straightforward in practice, it is, nonetheless,
9446 straightforward in theory. The atom @samp{rose}, for example, is
9447 recorded as the four contiguous letters @samp{r}, @samp{o}, @samp{s},
9448 @samp{e}. A list, on the other hand, is kept differently. The mechanism
9449 is equally simple, but it takes a moment to get used to the idea. A
9450 list is kept using a series of pairs of pointers. In the series, the
9451 first pointer in each pair points to an atom or to another list, and the
9452 second pointer in each pair points to the next pair, or to the symbol
9453 @code{nil}, which marks the end of the list.
9455 A pointer itself is quite simply the electronic address of what is
9456 pointed to. Hence, a list is kept as a series of electronic addresses.
9459 * Lists diagrammed::
9460 * Symbols as Chest:: Exploring a powerful metaphor.
9465 @node Lists diagrammed
9466 @unnumberedsec Lists diagrammed
9469 For example, the list @code{(rose violet buttercup)} has three elements,
9470 @samp{rose}, @samp{violet}, and @samp{buttercup}. In the computer, the
9471 electronic address of @samp{rose} is recorded in a segment of computer
9472 memory along with the address that gives the electronic address of where
9473 the atom @samp{violet} is located; and that address (the one that tells
9474 where @samp{violet} is located) is kept along with an address that tells
9475 where the address for the atom @samp{buttercup} is located.
9478 This sounds more complicated than it is and is easier seen in a diagram:
9480 @c clear print-postscript-figures
9481 @c !!! cons-cell-diagram #1
9485 ___ ___ ___ ___ ___ ___
9486 |___|___|--> |___|___|--> |___|___|--> nil
9489 --> rose --> violet --> buttercup
9493 @ifset print-postscript-figures
9496 @center @image{cons-1}
9500 @ifclear print-postscript-figures
9504 ___ ___ ___ ___ ___ ___
9505 |___|___|--> |___|___|--> |___|___|--> nil
9508 --> rose --> violet --> buttercup
9515 In the diagram, each box represents a word of computer memory that
9516 holds a Lisp object, usually in the form of a memory address. The boxes,
9517 i.e., the addresses, are in pairs. Each arrow points to what the address
9518 is the address of, either an atom or another pair of addresses. The
9519 first box is the electronic address of @samp{rose} and the arrow points
9520 to @samp{rose}; the second box is the address of the next pair of boxes,
9521 the first part of which is the address of @samp{violet} and the second
9522 part of which is the address of the next pair. The very last box
9523 points to the symbol @code{nil}, which marks the end of the list.
9526 When a variable is set to a list with a function such as @code{setq},
9527 it stores the address of the first box in the variable. Thus,
9528 evaluation of the expression
9531 (setq bouquet '(rose violet buttercup))
9536 creates a situation like this:
9538 @c cons-cell-diagram #2
9544 | ___ ___ ___ ___ ___ ___
9545 --> |___|___|--> |___|___|--> |___|___|--> nil
9548 --> rose --> violet --> buttercup
9552 @ifset print-postscript-figures
9555 @center @image{cons-2}
9559 @ifclear print-postscript-figures
9565 | ___ ___ ___ ___ ___ ___
9566 --> |___|___|--> |___|___|--> |___|___|--> nil
9569 --> rose --> violet --> buttercup
9576 In this example, the symbol @code{bouquet} holds the address of the first
9580 This same list can be illustrated in a different sort of box notation
9583 @c cons-cell-diagram #2a
9589 | -------------- --------------- ----------------
9590 | | car | cdr | | car | cdr | | car | cdr |
9591 -->| rose | o------->| violet | o------->| butter- | nil |
9592 | | | | | | | cup | |
9593 -------------- --------------- ----------------
9597 @ifset print-postscript-figures
9600 @center @image{cons-2a}
9604 @ifclear print-postscript-figures
9610 | -------------- --------------- ----------------
9611 | | car | cdr | | car | cdr | | car | cdr |
9612 -->| rose | o------->| violet | o------->| butter- | nil |
9613 | | | | | | | cup | |
9614 -------------- --------------- ----------------
9620 (Symbols consist of more than pairs of addresses, but the structure of
9621 a symbol is made up of addresses. Indeed, the symbol @code{bouquet}
9622 consists of a group of address-boxes, one of which is the address of
9623 the printed word @samp{bouquet}, a second of which is the address of a
9624 function definition attached to the symbol, if any, a third of which
9625 is the address of the first pair of address-boxes for the list
9626 @code{(rose violet buttercup)}, and so on. Here we are showing that
9627 the symbol's third address-box points to the first pair of
9628 address-boxes for the list.)
9630 If a symbol is set to the @sc{cdr} of a list, the list itself is not
9631 changed; the symbol simply has an address further down the list. (In
9632 the jargon, @sc{car} and @sc{cdr} are ``non-destructive''.) Thus,
9633 evaluation of the following expression
9636 (setq flowers (cdr bouquet))
9643 @c cons-cell-diagram #3
9650 | ___ ___ | ___ ___ ___ ___
9651 --> | | | --> | | | | | |
9652 |___|___|----> |___|___|--> |___|___|--> nil
9655 --> rose --> violet --> buttercup
9660 @ifset print-postscript-figures
9663 @center @image{cons-3}
9667 @ifclear print-postscript-figures
9674 | ___ ___ | ___ ___ ___ ___
9675 --> | | | --> | | | | | |
9676 |___|___|----> |___|___|--> |___|___|--> nil
9679 --> rose --> violet --> buttercup
9687 The value of @code{flowers} is @code{(violet buttercup)}, which is
9688 to say, the symbol @code{flowers} holds the address of the pair of
9689 address-boxes, the first of which holds the address of @code{violet},
9690 and the second of which holds the address of @code{buttercup}.
9692 A pair of address-boxes is called a @dfn{cons cell} or @dfn{dotted
9693 pair}. @xref{Cons Cell Type, , Cons Cell and List Types, elisp, The GNU Emacs Lisp
9694 Reference Manual}, and @ref{Dotted Pair Notation, , Dotted Pair
9695 Notation, elisp, The GNU Emacs Lisp Reference Manual}, for more
9696 information about cons cells and dotted pairs.
9699 The function @code{cons} adds a new pair of addresses to the front of
9700 a series of addresses like that shown above. For example, evaluating
9704 (setq bouquet (cons 'lily bouquet))
9711 @c cons-cell-diagram #4
9718 | ___ ___ ___ ___ | ___ ___ ___ ___
9719 --> | | | | | | --> | | | | | |
9720 |___|___|----> |___|___|----> |___|___|---->|___|___|--> nil
9723 --> lily --> rose --> violet --> buttercup
9728 @ifset print-postscript-figures
9731 @center @image{cons-4}
9735 @ifclear print-postscript-figures
9742 | ___ ___ ___ ___ | ___ ___ ___ ___
9743 --> | | | | | | --> | | | | | |
9744 |___|___|----> |___|___|----> |___|___|---->|___|___|--> nil
9747 --> lily --> rose --> violet --> buttercup
9756 However, this does not change the value of the symbol
9757 @code{flowers}, as you can see by evaluating the following,
9760 (eq (cdr (cdr bouquet)) flowers)
9764 which returns @code{t} for true.
9766 Until it is reset, @code{flowers} still has the value
9767 @code{(violet buttercup)}; that is, it has the address of the cons
9768 cell whose first address is of @code{violet}. Also, this does not
9769 alter any of the pre-existing cons cells; they are all still there.
9771 Thus, in Lisp, to get the @sc{cdr} of a list, you just get the address
9772 of the next cons cell in the series; to get the @sc{car} of a list,
9773 you get the address of the first element of the list; to @code{cons} a
9774 new element on a list, you add a new cons cell to the front of the list.
9775 That is all there is to it! The underlying structure of Lisp is
9778 And what does the last address in a series of cons cells refer to? It
9779 is the address of the empty list, of @code{nil}.
9781 In summary, when a Lisp variable is set to a value, it is provided with
9782 the address of the list to which the variable refers.
9784 @node Symbols as Chest
9785 @section Symbols as a Chest of Drawers
9786 @cindex Symbols as a Chest of Drawers
9787 @cindex Chest of Drawers, metaphor for a symbol
9788 @cindex Drawers, Chest of, metaphor for a symbol
9790 In an earlier section, I suggested that you might imagine a symbol as
9791 being a chest of drawers. The function definition is put in one
9792 drawer, the value in another, and so on. What is put in the drawer
9793 holding the value can be changed without affecting the contents of the
9794 drawer holding the function definition, and vice versa.
9796 Actually, what is put in each drawer is the address of the value or
9797 function definition. It is as if you found an old chest in the attic,
9798 and in one of its drawers you found a map giving you directions to
9799 where the buried treasure lies.
9801 (In addition to its name, symbol definition, and variable value, a
9802 symbol has a drawer for a @dfn{property list} which can be used to
9803 record other information. Property lists are not discussed here; see
9804 @ref{Property Lists, , Property Lists, elisp, The GNU Emacs Lisp
9808 Here is a fanciful representation:
9810 @c chest-of-drawers diagram
9815 Chest of Drawers Contents of Drawers
9819 ---------------------
9820 | directions to | [map to]
9821 | symbol name | bouquet
9823 +---------------------+
9825 | symbol definition | [none]
9827 +---------------------+
9828 | directions to | [map to]
9829 | variable value | (rose violet buttercup)
9831 +---------------------+
9833 | property list | [not described here]
9835 +---------------------+
9841 @ifset print-postscript-figures
9844 @center @image{drawers}
9848 @ifclear print-postscript-figures
9853 Chest of Drawers Contents of Drawers
9857 ---------------------
9858 | directions to | [map to]
9859 | symbol name | bouquet
9861 +---------------------+
9863 | symbol definition | [none]
9865 +---------------------+
9866 | directions to | [map to]
9867 | variable value | (rose violet buttercup)
9869 +---------------------+
9871 | property list | [not described here]
9873 +---------------------+
9884 Set @code{flowers} to @code{violet} and @code{buttercup}. Cons two
9885 more flowers on to this list and set this new list to
9886 @code{more-flowers}. Set the @sc{car} of @code{flowers} to a fish.
9887 What does the @code{more-flowers} list now contain?
9890 @chapter Yanking Text Back
9892 @cindex Text retrieval
9893 @cindex Retrieving text
9894 @cindex Pasting text
9896 Whenever you cut text out of a buffer with a kill command in GNU Emacs,
9897 you can bring it back with a yank command. The text that is cut out of
9898 the buffer is put in the kill ring and the yank commands insert the
9899 appropriate contents of the kill ring back into a buffer (not necessarily
9900 the original buffer).
9902 A simple @kbd{C-y} (@code{yank}) command inserts the first item from
9903 the kill ring into the current buffer. If the @kbd{C-y} command is
9904 followed immediately by @kbd{M-y}, the first element is replaced by
9905 the second element. Successive @kbd{M-y} commands replace the second
9906 element with the third, fourth, or fifth element, and so on. When the
9907 last element in the kill ring is reached, it is replaced by the first
9908 element and the cycle is repeated. (Thus the kill ring is called a
9909 ``ring'' rather than just a ``list''. However, the actual data structure
9910 that holds the text is a list.
9911 @xref{Kill Ring, , Handling the Kill Ring}, for the details of how the
9912 list is handled as a ring.)
9915 * Kill Ring Overview::
9916 * kill-ring-yank-pointer:: The kill ring is a list.
9917 * yank nthcdr Exercises:: The @code{kill-ring-yank-pointer} variable.
9920 @node Kill Ring Overview
9921 @section Kill Ring Overview
9922 @cindex Kill ring overview
9924 The kill ring is a list of textual strings. This is what it looks like:
9927 ("some text" "a different piece of text" "yet more text")
9930 If this were the contents of my kill ring and I pressed @kbd{C-y}, the
9931 string of characters saying @samp{some text} would be inserted in this
9932 buffer where my cursor is located.
9934 The @code{yank} command is also used for duplicating text by copying it.
9935 The copied text is not cut from the buffer, but a copy of it is put on the
9936 kill ring and is inserted by yanking it back.
9938 Three functions are used for bringing text back from the kill ring:
9939 @code{yank}, which is usually bound to @kbd{C-y}; @code{yank-pop},
9940 which is usually bound to @kbd{M-y}; and @code{rotate-yank-pointer},
9941 which is used by the two other functions.
9943 These functions refer to the kill ring through a variable called the
9944 @code{kill-ring-yank-pointer}. Indeed, the insertion code for both the
9945 @code{yank} and @code{yank-pop} functions is:
9948 (insert (car kill-ring-yank-pointer))
9952 (Well, no more. In GNU Emacs 22, the function has been replaced by
9953 @code{insert-for-yank} which calls @code{insert-for-yank-1}
9954 repetitively for each @code{yank-handler} segment. In turn,
9955 @code{insert-for-yank-1} strips text properties from the inserted text
9956 according to @code{yank-excluded-properties}. Otherwise, it is just
9957 like @code{insert}. We will stick with plain @code{insert} since it
9958 is easier to understand.)
9960 To begin to understand how @code{yank} and @code{yank-pop} work, it is
9961 first necessary to look at the @code{kill-ring-yank-pointer} variable.
9963 @node kill-ring-yank-pointer
9964 @section The @code{kill-ring-yank-pointer} Variable
9966 @code{kill-ring-yank-pointer} is a variable, just as @code{kill-ring} is
9967 a variable. It points to something by being bound to the value of what
9968 it points to, like any other Lisp variable.
9971 Thus, if the value of the kill ring is:
9974 ("some text" "a different piece of text" "yet more text")
9979 and the @code{kill-ring-yank-pointer} points to the second clause, the
9980 value of @code{kill-ring-yank-pointer} is:
9983 ("a different piece of text" "yet more text")
9986 As explained in the previous chapter (@pxref{List Implementation}), the
9987 computer does not keep two different copies of the text being pointed to
9988 by both the @code{kill-ring} and the @code{kill-ring-yank-pointer}. The
9989 words ``a different piece of text'' and ``yet more text'' are not
9990 duplicated. Instead, the two Lisp variables point to the same pieces of
9991 text. Here is a diagram:
9993 @c cons-cell-diagram #5
9997 kill-ring kill-ring-yank-pointer
9999 | ___ ___ | ___ ___ ___ ___
10000 ---> | | | --> | | | | | |
10001 |___|___|----> |___|___|--> |___|___|--> nil
10004 | | --> "yet more text"
10006 | --> "a different piece of text"
10013 @ifset print-postscript-figures
10016 @center @image{cons-5}
10020 @ifclear print-postscript-figures
10024 kill-ring kill-ring-yank-pointer
10026 | ___ ___ | ___ ___ ___ ___
10027 ---> | | | --> | | | | | |
10028 |___|___|----> |___|___|--> |___|___|--> nil
10031 | | --> "yet more text"
10033 | --> "a different piece of text
10042 Both the variable @code{kill-ring} and the variable
10043 @code{kill-ring-yank-pointer} are pointers. But the kill ring itself is
10044 usually described as if it were actually what it is composed of. The
10045 @code{kill-ring} is spoken of as if it were the list rather than that it
10046 points to the list. Conversely, the @code{kill-ring-yank-pointer} is
10047 spoken of as pointing to a list.
10049 These two ways of talking about the same thing sound confusing at first but
10050 make sense on reflection. The kill ring is generally thought of as the
10051 complete structure of data that holds the information of what has recently
10052 been cut out of the Emacs buffers. The @code{kill-ring-yank-pointer}
10053 on the other hand, serves to indicate---that is, to point to---that part
10054 of the kill ring of which the first element (the @sc{car}) will be
10058 In GNU Emacs 22, the @code{kill-new} function calls
10060 @code{(setq kill-ring-yank-pointer kill-ring)}
10062 (defun rotate-yank-pointer (arg)
10063 "Rotate the yanking point in the kill ring.
10064 With argument, rotate that many kills forward (or backward, if negative)."
10066 (current-kill arg))
10068 (defun current-kill (n &optional do-not-move)
10069 "Rotate the yanking point by N places, and then return that kill.
10070 If N is zero, `interprogram-paste-function' is set, and calling it
10071 returns a string, then that string is added to the front of the
10072 kill ring and returned as the latest kill.
10073 If optional arg DO-NOT-MOVE is non-nil, then don't actually move the
10074 yanking point; just return the Nth kill forward."
10075 (let ((interprogram-paste (and (= n 0)
10076 interprogram-paste-function
10077 (funcall interprogram-paste-function))))
10078 (if interprogram-paste
10080 ;; Disable the interprogram cut function when we add the new
10081 ;; text to the kill ring, so Emacs doesn't try to own the
10082 ;; selection, with identical text.
10083 (let ((interprogram-cut-function nil))
10084 (kill-new interprogram-paste))
10085 interprogram-paste)
10086 (or kill-ring (error "Kill ring is empty"))
10087 (let ((ARGth-kill-element
10088 (nthcdr (mod (- n (length kill-ring-yank-pointer))
10089 (length kill-ring))
10092 (setq kill-ring-yank-pointer ARGth-kill-element))
10093 (car ARGth-kill-element)))))
10098 @node yank nthcdr Exercises
10099 @section Exercises with @code{yank} and @code{nthcdr}
10103 Using @kbd{C-h v} (@code{describe-variable}), look at the value of
10104 your kill ring. Add several items to your kill ring; look at its
10105 value again. Using @kbd{M-y} (@code{yank-pop)}, move all the way
10106 around the kill ring. How many items were in your kill ring? Find
10107 the value of @code{kill-ring-max}. Was your kill ring full, or could
10108 you have kept more blocks of text within it?
10111 Using @code{nthcdr} and @code{car}, construct a series of expressions
10112 to return the first, second, third, and fourth elements of a list.
10115 @node Loops & Recursion
10116 @chapter Loops and Recursion
10117 @cindex Loops and recursion
10118 @cindex Recursion and loops
10119 @cindex Repetition (loops)
10121 Emacs Lisp has two primary ways to cause an expression, or a series of
10122 expressions, to be evaluated repeatedly: one uses a @code{while}
10123 loop, and the other uses @dfn{recursion}.
10125 Repetition can be very valuable. For example, to move forward four
10126 sentences, you need only write a program that will move forward one
10127 sentence and then repeat the process four times. Since a computer does
10128 not get bored or tired, such repetitive action does not have the
10129 deleterious effects that excessive or the wrong kinds of repetition can
10132 People mostly write Emacs Lisp functions using @code{while} loops and
10133 their kin; but you can use recursion, which provides a very powerful
10134 way to think about and then to solve problems@footnote{You can write
10135 recursive functions to be frugal or wasteful of mental or computer
10136 resources; as it happens, methods that people find easy---that are
10137 frugal of mental resources---sometimes use considerable computer
10138 resources. Emacs was designed to run on machines that we now consider
10139 limited and its default settings are conservative. You may want to
10140 increase the values of @code{max-specpdl-size} and
10141 @code{max-lisp-eval-depth}. In my @file{.emacs} file, I set them to
10142 15 and 30 times their default value.}.
10145 * while:: Causing a stretch of code to repeat.
10147 * Recursion:: Causing a function to call itself.
10148 * Looping exercise::
10152 @section @code{while}
10156 The @code{while} special form tests whether the value returned by
10157 evaluating its first argument is true or false. This is similar to what
10158 the Lisp interpreter does with an @code{if}; what the interpreter does
10159 next, however, is different.
10161 In a @code{while} expression, if the value returned by evaluating the
10162 first argument is false, the Lisp interpreter skips the rest of the
10163 expression (the @dfn{body} of the expression) and does not evaluate it.
10164 However, if the value is true, the Lisp interpreter evaluates the body
10165 of the expression and then again tests whether the first argument to
10166 @code{while} is true or false. If the value returned by evaluating the
10167 first argument is again true, the Lisp interpreter again evaluates the
10168 body of the expression.
10171 The template for a @code{while} expression looks like this:
10175 (while @var{true-or-false-test}
10181 * Looping with while:: Repeat so long as test returns true.
10182 * Loop Example:: A @code{while} loop that uses a list.
10183 * print-elements-of-list:: Uses @code{while}, @code{car}, @code{cdr}.
10184 * Incrementing Loop:: A loop with an incrementing counter.
10185 * Incrementing Loop Details::
10186 * Decrementing Loop:: A loop with a decrementing counter.
10190 @node Looping with while
10191 @unnumberedsubsec Looping with @code{while}
10194 So long as the true-or-false-test of the @code{while} expression
10195 returns a true value when it is evaluated, the body is repeatedly
10196 evaluated. This process is called a loop since the Lisp interpreter
10197 repeats the same thing again and again, like an airplane doing a loop.
10198 When the result of evaluating the true-or-false-test is false, the
10199 Lisp interpreter does not evaluate the rest of the @code{while}
10200 expression and exits the loop.
10202 Clearly, if the value returned by evaluating the first argument to
10203 @code{while} is always true, the body following will be evaluated
10204 again and again @dots{} and again @dots{} forever. Conversely, if the
10205 value returned is never true, the expressions in the body will never
10206 be evaluated. The craft of writing a @code{while} loop consists of
10207 choosing a mechanism such that the true-or-false-test returns true
10208 just the number of times that you want the subsequent expressions to
10209 be evaluated, and then have the test return false.
10211 The value returned by evaluating a @code{while} is the value of the
10212 true-or-false-test. An interesting consequence of this is that a
10213 @code{while} loop that evaluates without error will return @code{nil}
10214 or false regardless of whether it has looped 1 or 100 times or none at
10215 all. A @code{while} expression that evaluates successfully never
10216 returns a true value! What this means is that @code{while} is always
10217 evaluated for its side effects, which is to say, the consequences of
10218 evaluating the expressions within the body of the @code{while} loop.
10219 This makes sense. It is not the mere act of looping that is desired,
10220 but the consequences of what happens when the expressions in the loop
10221 are repeatedly evaluated.
10224 @subsection A @code{while} Loop and a List
10226 A common way to control a @code{while} loop is to test whether a list
10227 has any elements. If it does, the loop is repeated; but if it does not,
10228 the repetition is ended. Since this is an important technique, we will
10229 create a short example to illustrate it.
10231 A simple way to test whether a list has elements is to evaluate the
10232 list: if it has no elements, it is an empty list and will return the
10233 empty list, @code{()}, which is a synonym for @code{nil} or false. On
10234 the other hand, a list with elements will return those elements when it
10235 is evaluated. Since Emacs Lisp considers as true any value that is not
10236 @code{nil}, a list that returns elements will test true in a
10240 For example, you can set the variable @code{empty-list} to @code{nil} by
10241 evaluating the following @code{setq} expression:
10244 (setq empty-list ())
10248 After evaluating the @code{setq} expression, you can evaluate the
10249 variable @code{empty-list} in the usual way, by placing the cursor after
10250 the symbol and typing @kbd{C-x C-e}; @code{nil} will appear in your
10257 On the other hand, if you set a variable to be a list with elements, the
10258 list will appear when you evaluate the variable, as you can see by
10259 evaluating the following two expressions:
10263 (setq animals '(gazelle giraffe lion tiger))
10269 Thus, to create a @code{while} loop that tests whether there are any
10270 items in the list @code{animals}, the first part of the loop will be
10281 When the @code{while} tests its first argument, the variable
10282 @code{animals} is evaluated. It returns a list. So long as the list
10283 has elements, the @code{while} considers the results of the test to be
10284 true; but when the list is empty, it considers the results of the test
10287 To prevent the @code{while} loop from running forever, some mechanism
10288 needs to be provided to empty the list eventually. An oft-used
10289 technique is to have one of the subsequent forms in the @code{while}
10290 expression set the value of the list to be the @sc{cdr} of the list.
10291 Each time the @code{cdr} function is evaluated, the list will be made
10292 shorter, until eventually only the empty list will be left. At this
10293 point, the test of the @code{while} loop will return false, and the
10294 arguments to the @code{while} will no longer be evaluated.
10296 For example, the list of animals bound to the variable @code{animals}
10297 can be set to be the @sc{cdr} of the original list with the
10298 following expression:
10301 (setq animals (cdr animals))
10305 If you have evaluated the previous expressions and then evaluate this
10306 expression, you will see @code{(giraffe lion tiger)} appear in the echo
10307 area. If you evaluate the expression again, @code{(lion tiger)} will
10308 appear in the echo area. If you evaluate it again and yet again,
10309 @code{(tiger)} appears and then the empty list, shown by @code{nil}.
10311 A template for a @code{while} loop that uses the @code{cdr} function
10312 repeatedly to cause the true-or-false-test eventually to test false
10317 (while @var{test-whether-list-is-empty}
10319 @var{set-list-to-cdr-of-list})
10323 This test and use of @code{cdr} can be put together in a function that
10324 goes through a list and prints each element of the list on a line of its
10327 @node print-elements-of-list
10328 @subsection An Example: @code{print-elements-of-list}
10329 @findex print-elements-of-list
10331 The @code{print-elements-of-list} function illustrates a @code{while}
10334 @cindex @file{*scratch*} buffer
10335 The function requires several lines for its output. If you are
10336 reading this in a recent instance of GNU Emacs,
10337 @c GNU Emacs 21, GNU Emacs 22, or a later version,
10338 you can evaluate the following expression inside of Info, as usual.
10340 If you are using an earlier version of Emacs, you need to copy the
10341 necessary expressions to your @file{*scratch*} buffer and evaluate
10342 them there. This is because the echo area had only one line in the
10345 You can copy the expressions by marking the beginning of the region
10346 with @kbd{C-@key{SPC}} (@code{set-mark-command}), moving the cursor to
10347 the end of the region and then copying the region using @kbd{M-w}
10348 (@code{kill-ring-save}, which calls @code{copy-region-as-kill} and
10349 then provides visual feedback). In the @file{*scratch*}
10350 buffer, you can yank the expressions back by typing @kbd{C-y}
10353 After you have copied the expressions to the @file{*scratch*} buffer,
10354 evaluate each expression in turn. Be sure to evaluate the last
10355 expression, @code{(print-elements-of-list animals)}, by typing
10356 @kbd{C-u C-x C-e}, that is, by giving an argument to
10357 @code{eval-last-sexp}. This will cause the result of the evaluation
10358 to be printed in the @file{*scratch*} buffer instead of being printed
10359 in the echo area. (Otherwise you will see something like this in your
10360 echo area: @code{^Jgazelle^J^Jgiraffe^J^Jlion^J^Jtiger^Jnil}, in which
10361 each @samp{^J} stands for a newline.)
10364 In a recent instance of GNU Emacs, you can evaluate these expressions
10365 directly in the Info buffer, and the echo area will grow to show the
10370 (setq animals '(gazelle giraffe lion tiger))
10372 (defun print-elements-of-list (list)
10373 "Print each element of LIST on a line of its own."
10376 (setq list (cdr list))))
10378 (print-elements-of-list animals)
10384 When you evaluate the three expressions in sequence, you will see
10400 Each element of the list is printed on a line of its own (that is what
10401 the function @code{print} does) and then the value returned by the
10402 function is printed. Since the last expression in the function is the
10403 @code{while} loop, and since @code{while} loops always return
10404 @code{nil}, a @code{nil} is printed after the last element of the list.
10406 @node Incrementing Loop
10407 @subsection A Loop with an Incrementing Counter
10409 A loop is not useful unless it stops when it ought. Besides
10410 controlling a loop with a list, a common way of stopping a loop is to
10411 write the first argument as a test that returns false when the correct
10412 number of repetitions are complete. This means that the loop must
10413 have a counter---an expression that counts how many times the loop
10417 @node Incrementing Loop Details
10418 @unnumberedsubsec Details of an Incrementing Loop
10421 The test for a loop with an incrementing counter can be an expression
10422 such as @code{(< count desired-number)} which returns @code{t} for
10423 true if the value of @code{count} is less than the
10424 @code{desired-number} of repetitions and @code{nil} for false if the
10425 value of @code{count} is equal to or is greater than the
10426 @code{desired-number}. The expression that increments the count can
10427 be a simple @code{setq} such as @code{(setq count (1+ count))}, where
10428 @code{1+} is a built-in function in Emacs Lisp that adds 1 to its
10429 argument. (The expression @w{@code{(1+ count)}} has the same result
10430 as @w{@code{(+ count 1)}}, but is easier for a human to read.)
10433 The template for a @code{while} loop controlled by an incrementing
10434 counter looks like this:
10438 @var{set-count-to-initial-value}
10439 (while (< count desired-number) ; @r{true-or-false-test}
10441 (setq count (1+ count))) ; @r{incrementer}
10446 Note that you need to set the initial value of @code{count}; usually it
10450 * Incrementing Example:: Counting pebbles in a triangle.
10451 * Inc Example parts:: The parts of the function definition.
10452 * Inc Example altogether:: Putting the function definition together.
10455 @node Incrementing Example
10456 @unnumberedsubsubsec Example with incrementing counter
10458 Suppose you are playing on the beach and decide to make a triangle of
10459 pebbles, putting one pebble in the first row, two in the second row,
10460 three in the third row and so on, like this:
10478 @bullet{} @bullet{}
10479 @bullet{} @bullet{} @bullet{}
10480 @bullet{} @bullet{} @bullet{} @bullet{}
10487 (About 2500 years ago, Pythagoras and others developed the beginnings of
10488 number theory by considering questions such as this.)
10490 Suppose you want to know how many pebbles you will need to make a
10491 triangle with 7 rows?
10493 Clearly, what you need to do is add up the numbers from 1 to 7. There
10494 are two ways to do this; start with the smallest number, one, and add up
10495 the list in sequence, 1, 2, 3, 4 and so on; or start with the largest
10496 number and add the list going down: 7, 6, 5, 4 and so on. Because both
10497 mechanisms illustrate common ways of writing @code{while} loops, we will
10498 create two examples, one counting up and the other counting down. In
10499 this first example, we will start with 1 and add 2, 3, 4 and so on.
10501 If you are just adding up a short list of numbers, the easiest way to do
10502 it is to add up all the numbers at once. However, if you do not know
10503 ahead of time how many numbers your list will have, or if you want to be
10504 prepared for a very long list, then you need to design your addition so
10505 that what you do is repeat a simple process many times instead of doing
10506 a more complex process once.
10508 For example, instead of adding up all the pebbles all at once, what you
10509 can do is add the number of pebbles in the first row, 1, to the number
10510 in the second row, 2, and then add the total of those two rows to the
10511 third row, 3. Then you can add the number in the fourth row, 4, to the
10512 total of the first three rows; and so on.
10514 The critical characteristic of the process is that each repetitive
10515 action is simple. In this case, at each step we add only two numbers,
10516 the number of pebbles in the row and the total already found. This
10517 process of adding two numbers is repeated again and again until the last
10518 row has been added to the total of all the preceding rows. In a more
10519 complex loop the repetitive action might not be so simple, but it will
10520 be simpler than doing everything all at once.
10522 @node Inc Example parts
10523 @unnumberedsubsubsec The parts of the function definition
10525 The preceding analysis gives us the bones of our function definition:
10526 first, we will need a variable that we can call @code{total} that will
10527 be the total number of pebbles. This will be the value returned by
10530 Second, we know that the function will require an argument: this
10531 argument will be the total number of rows in the triangle. It can be
10532 called @code{number-of-rows}.
10534 Finally, we need a variable to use as a counter. We could call this
10535 variable @code{counter}, but a better name is @code{row-number}. That
10536 is because what the counter does in this function is count rows, and a
10537 program should be written to be as understandable as possible.
10539 When the Lisp interpreter first starts evaluating the expressions in the
10540 function, the value of @code{total} should be set to zero, since we have
10541 not added anything to it. Then the function should add the number of
10542 pebbles in the first row to the total, and then add the number of
10543 pebbles in the second to the total, and then add the number of
10544 pebbles in the third row to the total, and so on, until there are no
10545 more rows left to add.
10547 Both @code{total} and @code{row-number} are used only inside the
10548 function, so they can be declared as local variables with @code{let}
10549 and given initial values. Clearly, the initial value for @code{total}
10550 should be 0. The initial value of @code{row-number} should be 1,
10551 since we start with the first row. This means that the @code{let}
10552 statement will look like this:
10562 After the internal variables are declared and bound to their initial
10563 values, we can begin the @code{while} loop. The expression that serves
10564 as the test should return a value of @code{t} for true so long as the
10565 @code{row-number} is less than or equal to the @code{number-of-rows}.
10566 (If the expression tests true only so long as the row number is less
10567 than the number of rows in the triangle, the last row will never be
10568 added to the total; hence the row number has to be either less than or
10569 equal to the number of rows.)
10572 @findex <= @r{(less than or equal)}
10573 Lisp provides the @code{<=} function that returns true if the value of
10574 its first argument is less than or equal to the value of its second
10575 argument and false otherwise. So the expression that the @code{while}
10576 will evaluate as its test should look like this:
10579 (<= row-number number-of-rows)
10582 The total number of pebbles can be found by repeatedly adding the number
10583 of pebbles in a row to the total already found. Since the number of
10584 pebbles in the row is equal to the row number, the total can be found by
10585 adding the row number to the total. (Clearly, in a more complex
10586 situation, the number of pebbles in the row might be related to the row
10587 number in a more complicated way; if this were the case, the row number
10588 would be replaced by the appropriate expression.)
10591 (setq total (+ total row-number))
10595 What this does is set the new value of @code{total} to be equal to the
10596 sum of adding the number of pebbles in the row to the previous total.
10598 After setting the value of @code{total}, the conditions need to be
10599 established for the next repetition of the loop, if there is one. This
10600 is done by incrementing the value of the @code{row-number} variable,
10601 which serves as a counter. After the @code{row-number} variable has
10602 been incremented, the true-or-false-test at the beginning of the
10603 @code{while} loop tests whether its value is still less than or equal to
10604 the value of the @code{number-of-rows} and if it is, adds the new value
10605 of the @code{row-number} variable to the @code{total} of the previous
10606 repetition of the loop.
10609 The built-in Emacs Lisp function @code{1+} adds 1 to a number, so the
10610 @code{row-number} variable can be incremented with this expression:
10613 (setq row-number (1+ row-number))
10616 @node Inc Example altogether
10617 @unnumberedsubsubsec Putting the function definition together
10619 We have created the parts for the function definition; now we need to
10623 First, the contents of the @code{while} expression:
10627 (while (<= row-number number-of-rows) ; @r{true-or-false-test}
10628 (setq total (+ total row-number))
10629 (setq row-number (1+ row-number))) ; @r{incrementer}
10633 Along with the @code{let} expression varlist, this very nearly
10634 completes the body of the function definition. However, it requires
10635 one final element, the need for which is somewhat subtle.
10637 The final touch is to place the variable @code{total} on a line by
10638 itself after the @code{while} expression. Otherwise, the value returned
10639 by the whole function is the value of the last expression that is
10640 evaluated in the body of the @code{let}, and this is the value
10641 returned by the @code{while}, which is always @code{nil}.
10643 This may not be evident at first sight. It almost looks as if the
10644 incrementing expression is the last expression of the whole function.
10645 But that expression is part of the body of the @code{while}; it is the
10646 last element of the list that starts with the symbol @code{while}.
10647 Moreover, the whole of the @code{while} loop is a list within the body
10651 In outline, the function will look like this:
10655 (defun @var{name-of-function} (@var{argument-list})
10656 "@var{documentation}@dots{}"
10657 (let (@var{varlist})
10658 (while (@var{true-or-false-test})
10659 @var{body-of-while}@dots{} )
10660 @dots{} )) ; @r{Need final expression here.}
10664 The result of evaluating the @code{let} is what is going to be returned
10665 by the @code{defun} since the @code{let} is not embedded within any
10666 containing list, except for the @code{defun} as a whole. However, if
10667 the @code{while} is the last element of the @code{let} expression, the
10668 function will always return @code{nil}. This is not what we want!
10669 Instead, what we want is the value of the variable @code{total}. This
10670 is returned by simply placing the symbol as the last element of the list
10671 starting with @code{let}. It gets evaluated after the preceding
10672 elements of the list are evaluated, which means it gets evaluated after
10673 it has been assigned the correct value for the total.
10675 It may be easier to see this by printing the list starting with
10676 @code{let} all on one line. This format makes it evident that the
10677 @var{varlist} and @code{while} expressions are the second and third
10678 elements of the list starting with @code{let}, and the @code{total} is
10683 (let (@var{varlist}) (while (@var{true-or-false-test}) @var{body-of-while}@dots{} ) total)
10688 Putting everything together, the @code{triangle} function definition
10693 (defun triangle (number-of-rows) ; @r{Version with}
10694 ; @r{ incrementing counter.}
10695 "Add up the number of pebbles in a triangle.
10696 The first row has one pebble, the second row two pebbles,
10697 the third row three pebbles, and so on.
10698 The argument is NUMBER-OF-ROWS."
10703 (while (<= row-number number-of-rows)
10704 (setq total (+ total row-number))
10705 (setq row-number (1+ row-number)))
10711 After you have installed @code{triangle} by evaluating the function, you
10712 can try it out. Here are two examples:
10723 The sum of the first four numbers is 10 and the sum of the first seven
10726 @node Decrementing Loop
10727 @subsection Loop with a Decrementing Counter
10729 Another common way to write a @code{while} loop is to write the test
10730 so that it determines whether a counter is greater than zero. So long
10731 as the counter is greater than zero, the loop is repeated. But when
10732 the counter is equal to or less than zero, the loop is stopped. For
10733 this to work, the counter has to start out greater than zero and then
10734 be made smaller and smaller by a form that is evaluated
10737 The test will be an expression such as @code{(> counter 0)} which
10738 returns @code{t} for true if the value of @code{counter} is greater
10739 than zero, and @code{nil} for false if the value of @code{counter} is
10740 equal to or less than zero. The expression that makes the number
10741 smaller and smaller can be a simple @code{setq} such as @code{(setq
10742 counter (1- counter))}, where @code{1-} is a built-in function in
10743 Emacs Lisp that subtracts 1 from its argument.
10746 The template for a decrementing @code{while} loop looks like this:
10750 (while (> counter 0) ; @r{true-or-false-test}
10752 (setq counter (1- counter))) ; @r{decrementer}
10757 * Decrementing Example:: More pebbles on the beach.
10758 * Dec Example parts:: The parts of the function definition.
10759 * Dec Example altogether:: Putting the function definition together.
10762 @node Decrementing Example
10763 @unnumberedsubsubsec Example with decrementing counter
10765 To illustrate a loop with a decrementing counter, we will rewrite the
10766 @code{triangle} function so the counter decreases to zero.
10768 This is the reverse of the earlier version of the function. In this
10769 case, to find out how many pebbles are needed to make a triangle with
10770 3 rows, add the number of pebbles in the third row, 3, to the number
10771 in the preceding row, 2, and then add the total of those two rows to
10772 the row that precedes them, which is 1.
10774 Likewise, to find the number of pebbles in a triangle with 7 rows, add
10775 the number of pebbles in the seventh row, 7, to the number in the
10776 preceding row, which is 6, and then add the total of those two rows to
10777 the row that precedes them, which is 5, and so on. As in the previous
10778 example, each addition only involves adding two numbers, the total of
10779 the rows already added up and the number of pebbles in the row that is
10780 being added to the total. This process of adding two numbers is
10781 repeated again and again until there are no more pebbles to add.
10783 We know how many pebbles to start with: the number of pebbles in the
10784 last row is equal to the number of rows. If the triangle has seven
10785 rows, the number of pebbles in the last row is 7. Likewise, we know how
10786 many pebbles are in the preceding row: it is one less than the number in
10789 @node Dec Example parts
10790 @unnumberedsubsubsec The parts of the function definition
10792 We start with three variables: the total number of rows in the
10793 triangle; the number of pebbles in a row; and the total number of
10794 pebbles, which is what we want to calculate. These variables can be
10795 named @code{number-of-rows}, @code{number-of-pebbles-in-row}, and
10796 @code{total}, respectively.
10798 Both @code{total} and @code{number-of-pebbles-in-row} are used only
10799 inside the function and are declared with @code{let}. The initial
10800 value of @code{total} should, of course, be zero. However, the
10801 initial value of @code{number-of-pebbles-in-row} should be equal to
10802 the number of rows in the triangle, since the addition will start with
10806 This means that the beginning of the @code{let} expression will look
10812 (number-of-pebbles-in-row number-of-rows))
10817 The total number of pebbles can be found by repeatedly adding the number
10818 of pebbles in a row to the total already found, that is, by repeatedly
10819 evaluating the following expression:
10822 (setq total (+ total number-of-pebbles-in-row))
10826 After the @code{number-of-pebbles-in-row} is added to the @code{total},
10827 the @code{number-of-pebbles-in-row} should be decremented by one, since
10828 the next time the loop repeats, the preceding row will be
10829 added to the total.
10831 The number of pebbles in a preceding row is one less than the number of
10832 pebbles in a row, so the built-in Emacs Lisp function @code{1-} can be
10833 used to compute the number of pebbles in the preceding row. This can be
10834 done with the following expression:
10838 (setq number-of-pebbles-in-row
10839 (1- number-of-pebbles-in-row))
10843 Finally, we know that the @code{while} loop should stop making repeated
10844 additions when there are no pebbles in a row. So the test for
10845 the @code{while} loop is simply:
10848 (while (> number-of-pebbles-in-row 0)
10851 @node Dec Example altogether
10852 @unnumberedsubsubsec Putting the function definition together
10854 We can put these expressions together to create a function definition
10855 that works. However, on examination, we find that one of the local
10856 variables is unneeded!
10859 The function definition looks like this:
10863 ;;; @r{First subtractive version.}
10864 (defun triangle (number-of-rows)
10865 "Add up the number of pebbles in a triangle."
10867 (number-of-pebbles-in-row number-of-rows))
10868 (while (> number-of-pebbles-in-row 0)
10869 (setq total (+ total number-of-pebbles-in-row))
10870 (setq number-of-pebbles-in-row
10871 (1- number-of-pebbles-in-row)))
10876 As written, this function works.
10878 However, we do not need @code{number-of-pebbles-in-row}.
10880 @cindex Argument as local variable
10881 When the @code{triangle} function is evaluated, the symbol
10882 @code{number-of-rows} will be bound to a number, giving it an initial
10883 value. That number can be changed in the body of the function as if
10884 it were a local variable, without any fear that such a change will
10885 effect the value of the variable outside of the function. This is a
10886 very useful characteristic of Lisp; it means that the variable
10887 @code{number-of-rows} can be used anywhere in the function where
10888 @code{number-of-pebbles-in-row} is used.
10891 Here is a second version of the function written a bit more cleanly:
10895 (defun triangle (number) ; @r{Second version.}
10896 "Return sum of numbers 1 through NUMBER inclusive."
10898 (while (> number 0)
10899 (setq total (+ total number))
10900 (setq number (1- number)))
10905 In brief, a properly written @code{while} loop will consist of three parts:
10909 A test that will return false after the loop has repeated itself the
10910 correct number of times.
10913 An expression the evaluation of which will return the value desired
10914 after being repeatedly evaluated.
10917 An expression to change the value passed to the true-or-false-test so
10918 that the test returns false after the loop has repeated itself the right
10922 @node dolist dotimes
10923 @section Save your time: @code{dolist} and @code{dotimes}
10925 In addition to @code{while}, both @code{dolist} and @code{dotimes}
10926 provide for looping. Sometimes these are quicker to write than the
10927 equivalent @code{while} loop. Both are Lisp macros. (@xref{Macros, ,
10928 Macros, elisp, The GNU Emacs Lisp Reference Manual}. )
10930 @code{dolist} works like a @code{while} loop that @sc{cdr}s down a
10931 list: @code{dolist} automatically shortens the list each time it
10932 loops---takes the @sc{cdr} of the list---and binds the @sc{car} of
10933 each shorter version of the list to the first of its arguments.
10935 @code{dotimes} loops a specific number of times: you specify the number.
10943 @unnumberedsubsec The @code{dolist} Macro
10946 Suppose, for example, you want to reverse a list, so that
10947 ``first'' ``second'' ``third'' becomes ``third'' ``second'' ``first''.
10950 In practice, you would use the @code{reverse} function, like this:
10954 (setq animals '(gazelle giraffe lion tiger))
10962 Here is how you could reverse the list using a @code{while} loop:
10966 (setq animals '(gazelle giraffe lion tiger))
10968 (defun reverse-list-with-while (list)
10969 "Using while, reverse the order of LIST."
10970 (let (value) ; make sure list starts empty
10972 (setq value (cons (car list) value))
10973 (setq list (cdr list)))
10976 (reverse-list-with-while animals)
10982 And here is how you could use the @code{dolist} macro:
10986 (setq animals '(gazelle giraffe lion tiger))
10988 (defun reverse-list-with-dolist (list)
10989 "Using dolist, reverse the order of LIST."
10990 (let (value) ; make sure list starts empty
10991 (dolist (element list value)
10992 (setq value (cons element value)))))
10994 (reverse-list-with-dolist animals)
11000 In Info, you can place your cursor after the closing parenthesis of
11001 each expression and type @kbd{C-x C-e}; in each case, you should see
11004 (tiger lion giraffe gazelle)
11010 For this example, the existing @code{reverse} function is obviously best.
11011 The @code{while} loop is just like our first example (@pxref{Loop
11012 Example, , A @code{while} Loop and a List}). The @code{while} first
11013 checks whether the list has elements; if so, it constructs a new list
11014 by adding the first element of the list to the existing list (which in
11015 the first iteration of the loop is @code{nil}). Since the second
11016 element is prepended in front of the first element, and the third
11017 element is prepended in front of the second element, the list is reversed.
11019 In the expression using a @code{while} loop,
11020 the @w{@code{(setq list (cdr list))}}
11021 expression shortens the list, so the @code{while} loop eventually
11022 stops. In addition, it provides the @code{cons} expression with a new
11023 first element by creating a new and shorter list at each repetition of
11026 The @code{dolist} expression does very much the same as the
11027 @code{while} expression, except that the @code{dolist} macro does some
11028 of the work you have to do when writing a @code{while} expression.
11030 Like a @code{while} loop, a @code{dolist} loops. What is different is
11031 that it automatically shortens the list each time it loops---it
11032 @sc{cdr}s down the list on its own---and it automatically binds
11033 the @sc{car} of each shorter version of the list to the first of its
11036 In the example, the @sc{car} of each shorter version of the list is
11037 referred to using the symbol @samp{element}, the list itself is called
11038 @samp{list}, and the value returned is called @samp{value}. The
11039 remainder of the @code{dolist} expression is the body.
11041 The @code{dolist} expression binds the @sc{car} of each shorter
11042 version of the list to @code{element} and then evaluates the body of
11043 the expression; and repeats the loop. The result is returned in
11047 @unnumberedsubsec The @code{dotimes} Macro
11050 The @code{dotimes} macro is similar to @code{dolist}, except that it
11051 loops a specific number of times.
11053 The first argument to @code{dotimes} is assigned the numbers 0, 1, 2
11054 and so forth each time around the loop, and the value of the third
11055 argument is returned. You need to provide the value of the second
11056 argument, which is how many times the macro loops.
11059 For example, the following binds the numbers from 0 up to, but not
11060 including, the number 3 to the first argument, @var{number}, and then
11061 constructs a list of the three numbers. (The first number is 0, the
11062 second number is 1, and the third number is 2; this makes a total of
11063 three numbers in all, starting with zero as the first number.)
11067 (let (value) ; otherwise a value is a void variable
11068 (dotimes (number 3 value)
11069 (setq value (cons number value))))
11076 @code{dotimes} returns @code{value}, so the way to use
11077 @code{dotimes} is to operate on some expression @var{number} number of
11078 times and then return the result, either as a list or an atom.
11081 Here is an example of a @code{defun} that uses @code{dotimes} to add
11082 up the number of pebbles in a triangle.
11086 (defun triangle-using-dotimes (number-of-rows)
11087 "Using `dotimes', add up the number of pebbles in a triangle."
11088 (let ((total 0)) ; otherwise a total is a void variable
11089 (dotimes (number number-of-rows total)
11090 (setq total (+ total (1+ number))))))
11092 (triangle-using-dotimes 4)
11100 A recursive function contains code that tells the Lisp interpreter to
11101 call a program that runs exactly like itself, but with slightly
11102 different arguments. The code runs exactly the same because it has
11103 the same name. However, even though the program has the same name, it
11104 is not the same entity. It is different. In the jargon, it is a
11105 different ``instance''.
11107 Eventually, if the program is written correctly, the slightly
11108 different arguments will become sufficiently different from the first
11109 arguments that the final instance will stop.
11112 * Building Robots:: Same model, different serial number ...
11113 * Recursive Definition Parts:: Walk until you stop ...
11114 * Recursion with list:: Using a list as the test whether to recurse.
11115 * Recursive triangle function::
11116 * Recursion with cond::
11117 * Recursive Patterns:: Often used templates.
11118 * No Deferment:: Don't store up work ...
11119 * No deferment solution::
11122 @node Building Robots
11123 @subsection Building Robots: Extending the Metaphor
11124 @cindex Building robots
11125 @cindex Robots, building
11127 It is sometimes helpful to think of a running program as a robot that
11128 does a job. In doing its job, a recursive function calls on a second
11129 robot to help it. The second robot is identical to the first in every
11130 way, except that the second robot helps the first and has been
11131 passed different arguments than the first.
11133 In a recursive function, the second robot may call a third; and the
11134 third may call a fourth, and so on. Each of these is a different
11135 entity; but all are clones.
11137 Since each robot has slightly different instructions---the arguments
11138 will differ from one robot to the next---the last robot should know
11141 Let's expand on the metaphor in which a computer program is a robot.
11143 A function definition provides the blueprints for a robot. When you
11144 install a function definition, that is, when you evaluate a
11145 @code{defun} macro, you install the necessary equipment to build
11146 robots. It is as if you were in a factory, setting up an assembly
11147 line. Robots with the same name are built according to the same
11148 blueprints. So they have the same model number, but a
11149 different serial number.
11151 We often say that a recursive function ``calls itself''. What we mean
11152 is that the instructions in a recursive function cause the Lisp
11153 interpreter to run a different function that has the same name and
11154 does the same job as the first, but with different arguments.
11156 It is important that the arguments differ from one instance to the
11157 next; otherwise, the process will never stop.
11159 @node Recursive Definition Parts
11160 @subsection The Parts of a Recursive Definition
11161 @cindex Parts of a Recursive Definition
11162 @cindex Recursive Definition Parts
11164 A recursive function typically contains a conditional expression which
11169 A true-or-false-test that determines whether the function is called
11170 again, here called the @dfn{do-again-test}.
11173 The name of the function. When this name is called, a new instance of
11174 the function---a new robot, as it were---is created and told what to do.
11177 An expression that returns a different value each time the function is
11178 called, here called the @dfn{next-step-expression}. Consequently, the
11179 argument (or arguments) passed to the new instance of the function
11180 will be different from that passed to the previous instance. This
11181 causes the conditional expression, the @dfn{do-again-test}, to test
11182 false after the correct number of repetitions.
11185 Recursive functions can be much simpler than any other kind of
11186 function. Indeed, when people first start to use them, they often look
11187 so mysteriously simple as to be incomprehensible. Like riding a
11188 bicycle, reading a recursive function definition takes a certain knack
11189 which is hard at first but then seems simple.
11192 There are several different common recursive patterns. A very simple
11193 pattern looks like this:
11197 (defun @var{name-of-recursive-function} (@var{argument-list})
11198 "@var{documentation}@dots{}"
11199 (if @var{do-again-test}
11201 (@var{name-of-recursive-function}
11202 @var{next-step-expression})))
11206 Each time a recursive function is evaluated, a new instance of it is
11207 created and told what to do. The arguments tell the instance what to do.
11209 An argument is bound to the value of the next-step-expression. Each
11210 instance runs with a different value of the next-step-expression.
11212 The value in the next-step-expression is used in the do-again-test.
11214 The value returned by the next-step-expression is passed to the new
11215 instance of the function, which evaluates it (or some
11216 transmogrification of it) to determine whether to continue or stop.
11217 The next-step-expression is designed so that the do-again-test returns
11218 false when the function should no longer be repeated.
11220 The do-again-test is sometimes called the @dfn{stop condition},
11221 since it stops the repetitions when it tests false.
11223 @node Recursion with list
11224 @subsection Recursion with a List
11226 The example of a @code{while} loop that printed the elements of a list
11227 of numbers can be written recursively. Here is the code, including
11228 an expression to set the value of the variable @code{animals} to a list.
11230 If you are reading this in Info in Emacs, you can evaluate this
11231 expression directly in Info. Otherwise, you must copy the example
11232 to the @file{*scratch*} buffer and evaluate each expression there.
11233 Use @kbd{C-u C-x C-e} to evaluate the
11234 @code{(print-elements-recursively animals)} expression so that the
11235 results are printed in the buffer; otherwise the Lisp interpreter will
11236 try to squeeze the results into the one line of the echo area.
11238 Also, place your cursor immediately after the last closing parenthesis
11239 of the @code{print-elements-recursively} function, before the comment.
11240 Otherwise, the Lisp interpreter will try to evaluate the comment.
11242 @findex print-elements-recursively
11245 (setq animals '(gazelle giraffe lion tiger))
11247 (defun print-elements-recursively (list)
11248 "Print each element of LIST on a line of its own.
11250 (when list ; @r{do-again-test}
11251 (print (car list)) ; @r{body}
11252 (print-elements-recursively ; @r{recursive call}
11253 (cdr list)))) ; @r{next-step-expression}
11255 (print-elements-recursively animals)
11259 The @code{print-elements-recursively} function first tests whether
11260 there is any content in the list; if there is, the function prints the
11261 first element of the list, the @sc{car} of the list. Then the
11262 function invokes itself, but gives itself as its argument, not the
11263 whole list, but the second and subsequent elements of the list, the
11264 @sc{cdr} of the list.
11266 Put another way, if the list is not empty, the function invokes
11267 another instance of code that is similar to the initial code, but is a
11268 different thread of execution, with different arguments than the first
11271 Put in yet another way, if the list is not empty, the first robot
11272 assembles a second robot and tells it what to do; the second robot is
11273 a different individual from the first, but is the same model.
11275 When the second evaluation occurs, the @code{when} expression is
11276 evaluated and if true, prints the first element of the list it
11277 receives as its argument (which is the second element of the original
11278 list). Then the function calls itself with the @sc{cdr} of the list
11279 it is invoked with, which (the second time around) is the @sc{cdr} of
11280 the @sc{cdr} of the original list.
11282 Note that although we say that the function ``calls itself'', what we
11283 mean is that the Lisp interpreter assembles and instructs a new
11284 instance of the program. The new instance is a clone of the first,
11285 but is a separate individual.
11287 Each time the function invokes itself, it does so on a
11288 shorter version of the original list. It creates a new instance that
11289 works on a shorter list.
11291 Eventually, the function invokes itself on an empty list. It creates
11292 a new instance whose argument is @code{nil}. The conditional expression
11293 tests the value of @code{list}. Since the value of @code{list} is
11294 @code{nil}, the @code{when} expression tests false so the then-part is
11295 not evaluated. The function as a whole then returns @code{nil}.
11298 When you evaluate the expression @code{(print-elements-recursively
11299 animals)} in the @file{*scratch*} buffer, you see this result:
11315 @node Recursive triangle function
11316 @subsection Recursion in Place of a Counter
11317 @findex triangle-recursively
11320 The @code{triangle} function described in a previous section can also
11321 be written recursively. It looks like this:
11325 (defun triangle-recursively (number)
11326 "Return the sum of the numbers 1 through NUMBER inclusive.
11328 (if (= number 1) ; @r{do-again-test}
11330 (+ number ; @r{else-part}
11331 (triangle-recursively ; @r{recursive call}
11332 (1- number))))) ; @r{next-step-expression}
11334 (triangle-recursively 7)
11339 You can install this function by evaluating it and then try it by
11340 evaluating @code{(triangle-recursively 7)}. (Remember to put your
11341 cursor immediately after the last parenthesis of the function
11342 definition, before the comment.) The function evaluates to 28.
11344 To understand how this function works, let's consider what happens in the
11345 various cases when the function is passed 1, 2, 3, or 4 as the value of
11349 * Recursive Example arg of 1 or 2::
11350 * Recursive Example arg of 3 or 4::
11354 @node Recursive Example arg of 1 or 2
11355 @unnumberedsubsubsec An argument of 1 or 2
11358 First, what happens if the value of the argument is 1?
11360 The function has an @code{if} expression after the documentation
11361 string. It tests whether the value of @code{number} is equal to 1; if
11362 so, Emacs evaluates the then-part of the @code{if} expression, which
11363 returns the number 1 as the value of the function. (A triangle with
11364 one row has one pebble in it.)
11366 Suppose, however, that the value of the argument is 2. In this case,
11367 Emacs evaluates the else-part of the @code{if} expression.
11370 The else-part consists of an addition, the recursive call to
11371 @code{triangle-recursively} and a decrementing action; and it looks like
11375 (+ number (triangle-recursively (1- number)))
11378 When Emacs evaluates this expression, the innermost expression is
11379 evaluated first; then the other parts in sequence. Here are the steps
11383 @item Step 1 @w{ } Evaluate the innermost expression.
11385 The innermost expression is @code{(1- number)} so Emacs decrements the
11386 value of @code{number} from 2 to 1.
11388 @item Step 2 @w{ } Evaluate the @code{triangle-recursively} function.
11390 The Lisp interpreter creates an individual instance of
11391 @code{triangle-recursively}. It does not matter that this function is
11392 contained within itself. Emacs passes the result Step 1 as the
11393 argument used by this instance of the @code{triangle-recursively}
11396 In this case, Emacs evaluates @code{triangle-recursively} with an
11397 argument of 1. This means that this evaluation of
11398 @code{triangle-recursively} returns 1.
11400 @item Step 3 @w{ } Evaluate the value of @code{number}.
11402 The variable @code{number} is the second element of the list that
11403 starts with @code{+}; its value is 2.
11405 @item Step 4 @w{ } Evaluate the @code{+} expression.
11407 The @code{+} expression receives two arguments, the first
11408 from the evaluation of @code{number} (Step 3) and the second from the
11409 evaluation of @code{triangle-recursively} (Step 2).
11411 The result of the addition is the sum of 2 plus 1, and the number 3 is
11412 returned, which is correct. A triangle with two rows has three
11416 @node Recursive Example arg of 3 or 4
11417 @unnumberedsubsubsec An argument of 3 or 4
11419 Suppose that @code{triangle-recursively} is called with an argument of
11423 @item Step 1 @w{ } Evaluate the do-again-test.
11425 The @code{if} expression is evaluated first. This is the do-again
11426 test and returns false, so the else-part of the @code{if} expression
11427 is evaluated. (Note that in this example, the do-again-test causes
11428 the function to call itself when it tests false, not when it tests
11431 @item Step 2 @w{ } Evaluate the innermost expression of the else-part.
11433 The innermost expression of the else-part is evaluated, which decrements
11434 3 to 2. This is the next-step-expression.
11436 @item Step 3 @w{ } Evaluate the @code{triangle-recursively} function.
11438 The number 2 is passed to the @code{triangle-recursively} function.
11440 We already know what happens when Emacs evaluates @code{triangle-recursively} with
11441 an argument of 2. After going through the sequence of actions described
11442 earlier, it returns a value of 3. So that is what will happen here.
11444 @item Step 4 @w{ } Evaluate the addition.
11446 3 will be passed as an argument to the addition and will be added to the
11447 number with which the function was called, which is 3.
11451 The value returned by the function as a whole will be 6.
11453 Now that we know what will happen when @code{triangle-recursively} is
11454 called with an argument of 3, it is evident what will happen if it is
11455 called with an argument of 4:
11459 In the recursive call, the evaluation of
11462 (triangle-recursively (1- 4))
11467 will return the value of evaluating
11470 (triangle-recursively 3)
11474 which is 6 and this value will be added to 4 by the addition in the
11479 The value returned by the function as a whole will be 10.
11481 Each time @code{triangle-recursively} is evaluated, it evaluates a
11482 version of itself---a different instance of itself---with a smaller
11483 argument, until the argument is small enough so that it does not
11486 Note that this particular design for a recursive function
11487 requires that operations be deferred.
11489 Before @code{(triangle-recursively 7)} can calculate its answer, it
11490 must call @code{(triangle-recursively 6)}; and before
11491 @code{(triangle-recursively 6)} can calculate its answer, it must call
11492 @code{(triangle-recursively 5)}; and so on. That is to say, the
11493 calculation that @code{(triangle-recursively 7)} makes must be
11494 deferred until @code{(triangle-recursively 6)} makes its calculation;
11495 and @code{(triangle-recursively 6)} must defer until
11496 @code{(triangle-recursively 5)} completes; and so on.
11498 If each of these instances of @code{triangle-recursively} are thought
11499 of as different robots, the first robot must wait for the second to
11500 complete its job, which must wait until the third completes, and so
11503 There is a way around this kind of waiting, which we will discuss in
11504 @ref{No Deferment, , Recursion without Deferments}.
11506 @node Recursion with cond
11507 @subsection Recursion Example Using @code{cond}
11510 The version of @code{triangle-recursively} described earlier is written
11511 with the @code{if} special form. It can also be written using another
11512 special form called @code{cond}. The name of the special form
11513 @code{cond} is an abbreviation of the word @samp{conditional}.
11515 Although the @code{cond} special form is not used as often in the
11516 Emacs Lisp sources as @code{if}, it is used often enough to justify
11520 The template for a @code{cond} expression looks like this:
11530 where the @var{body} is a series of lists.
11533 Written out more fully, the template looks like this:
11538 (@var{first-true-or-false-test} @var{first-consequent})
11539 (@var{second-true-or-false-test} @var{second-consequent})
11540 (@var{third-true-or-false-test} @var{third-consequent})
11545 When the Lisp interpreter evaluates the @code{cond} expression, it
11546 evaluates the first element (the @sc{car} or true-or-false-test) of
11547 the first expression in a series of expressions within the body of the
11550 If the true-or-false-test returns @code{nil} the rest of that
11551 expression, the consequent, is skipped and the true-or-false-test of the
11552 next expression is evaluated. When an expression is found whose
11553 true-or-false-test returns a value that is not @code{nil}, the
11554 consequent of that expression is evaluated. The consequent can be one
11555 or more expressions. If the consequent consists of more than one
11556 expression, the expressions are evaluated in sequence and the value of
11557 the last one is returned. If the expression does not have a consequent,
11558 the value of the true-or-false-test is returned.
11560 If none of the true-or-false-tests test true, the @code{cond} expression
11561 returns @code{nil}.
11564 Written using @code{cond}, the @code{triangle} function looks like this:
11568 (defun triangle-using-cond (number)
11569 (cond ((<= number 0) 0)
11572 (+ number (triangle-using-cond (1- number))))))
11577 In this example, the @code{cond} returns 0 if the number is less than or
11578 equal to 0, it returns 1 if the number is 1 and it evaluates @code{(+
11579 number (triangle-using-cond (1- number)))} if the number is greater than
11582 @node Recursive Patterns
11583 @subsection Recursive Patterns
11584 @cindex Recursive Patterns
11586 Here are three common recursive patterns. Each involves a list.
11587 Recursion does not need to involve lists, but Lisp is designed for lists
11588 and this provides a sense of its primal capabilities.
11597 @unnumberedsubsubsec Recursive Pattern: @emph{every}
11598 @cindex Every, type of recursive pattern
11599 @cindex Recursive pattern - every
11601 In the @code{every} recursive pattern, an action is performed on every
11605 The basic pattern is:
11609 If a list be empty, return @code{nil}.
11611 Else, act on the beginning of the list (the @sc{car} of the list)
11614 through a recursive call by the function on the rest (the
11615 @sc{cdr}) of the list,
11617 and, optionally, combine the acted-on element, using @code{cons},
11618 with the results of acting on the rest.
11623 Here is an example:
11627 (defun square-each (numbers-list)
11628 "Square each of a NUMBERS LIST, recursively."
11629 (if (not numbers-list) ; do-again-test
11632 (* (car numbers-list) (car numbers-list))
11633 (square-each (cdr numbers-list))))) ; next-step-expression
11637 (square-each '(1 2 3))
11644 If @code{numbers-list} is empty, do nothing. But if it has content,
11645 construct a list combining the square of the first number in the list
11646 with the result of the recursive call.
11648 (The example follows the pattern exactly: @code{nil} is returned if
11649 the numbers' list is empty. In practice, you would write the
11650 conditional so it carries out the action when the numbers' list is not
11653 The @code{print-elements-recursively} function (@pxref{Recursion with
11654 list, , Recursion with a List}) is another example of an @code{every}
11655 pattern, except in this case, rather than bring the results together
11656 using @code{cons}, we print each element of output.
11659 The @code{print-elements-recursively} function looks like this:
11663 (setq animals '(gazelle giraffe lion tiger))
11667 (defun print-elements-recursively (list)
11668 "Print each element of LIST on a line of its own.
11670 (when list ; @r{do-again-test}
11671 (print (car list)) ; @r{body}
11672 (print-elements-recursively ; @r{recursive call}
11673 (cdr list)))) ; @r{next-step-expression}
11675 (print-elements-recursively animals)
11680 The pattern for @code{print-elements-recursively} is:
11684 When the list is empty, do nothing.
11686 But when the list has at least one element,
11689 act on the beginning of the list (the @sc{car} of the list),
11691 and make a recursive call on the rest (the @sc{cdr}) of the list.
11696 @unnumberedsubsubsec Recursive Pattern: @emph{accumulate}
11697 @cindex Accumulate, type of recursive pattern
11698 @cindex Recursive pattern - accumulate
11700 Another recursive pattern is called the @code{accumulate} pattern. In
11701 the @code{accumulate} recursive pattern, an action is performed on
11702 every element of a list and the result of that action is accumulated
11703 with the results of performing the action on the other elements.
11705 This is very like the @code{every} pattern using @code{cons}, except that
11706 @code{cons} is not used, but some other combiner.
11713 If a list be empty, return zero or some other constant.
11715 Else, act on the beginning of the list (the @sc{car} of the list),
11718 and combine that acted-on element, using @code{+} or
11719 some other combining function, with
11721 a recursive call by the function on the rest (the @sc{cdr}) of the list.
11726 Here is an example:
11730 (defun add-elements (numbers-list)
11731 "Add the elements of NUMBERS-LIST together."
11732 (if (not numbers-list)
11734 (+ (car numbers-list) (add-elements (cdr numbers-list)))))
11738 (add-elements '(1 2 3 4))
11743 @xref{Files List, , Making a List of Files}, for an example of the
11744 accumulate pattern.
11747 @unnumberedsubsubsec Recursive Pattern: @emph{keep}
11748 @cindex Keep, type of recursive pattern
11749 @cindex Recursive pattern - keep
11751 A third recursive pattern is called the @code{keep} pattern.
11752 In the @code{keep} recursive pattern, each element of a list is tested;
11753 the element is acted on and the results are kept only if the element
11756 Again, this is very like the @code{every} pattern, except the element is
11757 skipped unless it meets a criterion.
11760 The pattern has three parts:
11764 If a list be empty, return @code{nil}.
11766 Else, if the beginning of the list (the @sc{car} of the list) passes
11770 act on that element and combine it, using @code{cons} with
11772 a recursive call by the function on the rest (the @sc{cdr}) of the list.
11775 Otherwise, if the beginning of the list (the @sc{car} of the list) fails
11779 skip on that element,
11781 and, recursively call the function on the rest (the @sc{cdr}) of the list.
11786 Here is an example that uses @code{cond}:
11790 (defun keep-three-letter-words (word-list)
11791 "Keep three letter words in WORD-LIST."
11793 ;; First do-again-test: stop-condition
11794 ((not word-list) nil)
11796 ;; Second do-again-test: when to act
11797 ((eq 3 (length (symbol-name (car word-list))))
11798 ;; combine acted-on element with recursive call on shorter list
11799 (cons (car word-list) (keep-three-letter-words (cdr word-list))))
11801 ;; Third do-again-test: when to skip element;
11802 ;; recursively call shorter list with next-step expression
11803 (t (keep-three-letter-words (cdr word-list)))))
11807 (keep-three-letter-words '(one two three four five six))
11808 @result{} (one two six)
11812 It goes without saying that you need not use @code{nil} as the test for
11813 when to stop; and you can, of course, combine these patterns.
11816 @subsection Recursion without Deferments
11817 @cindex Deferment in recursion
11818 @cindex Recursion without Deferments
11820 Let's consider again what happens with the @code{triangle-recursively}
11821 function. We will find that the intermediate calculations are
11822 deferred until all can be done.
11825 Here is the function definition:
11829 (defun triangle-recursively (number)
11830 "Return the sum of the numbers 1 through NUMBER inclusive.
11832 (if (= number 1) ; @r{do-again-test}
11834 (+ number ; @r{else-part}
11835 (triangle-recursively ; @r{recursive call}
11836 (1- number))))) ; @r{next-step-expression}
11840 What happens when we call this function with an argument of 7?
11842 The first instance of the @code{triangle-recursively} function adds
11843 the number 7 to the value returned by a second instance of
11844 @code{triangle-recursively}, an instance that has been passed an
11845 argument of 6. That is to say, the first calculation is:
11848 (+ 7 (triangle-recursively 6))
11852 The first instance of @code{triangle-recursively}---you may want to
11853 think of it as a little robot---cannot complete its job. It must hand
11854 off the calculation for @code{(triangle-recursively 6)} to a second
11855 instance of the program, to a second robot. This second individual is
11856 completely different from the first one; it is, in the jargon, a
11857 ``different instantiation''. Or, put another way, it is a different
11858 robot. It is the same model as the first; it calculates triangle
11859 numbers recursively; but it has a different serial number.
11861 And what does @code{(triangle-recursively 6)} return? It returns the
11862 number 6 added to the value returned by evaluating
11863 @code{triangle-recursively} with an argument of 5. Using the robot
11864 metaphor, it asks yet another robot to help it.
11870 (+ 7 6 (triangle-recursively 5))
11874 And what happens next?
11877 (+ 7 6 5 (triangle-recursively 4))
11880 Each time @code{triangle-recursively} is called, except for the last
11881 time, it creates another instance of the program---another robot---and
11882 asks it to make a calculation.
11885 Eventually, the full addition is set up and performed:
11891 This design for the function defers the calculation of the first step
11892 until the second can be done, and defers that until the third can be
11893 done, and so on. Each deferment means the computer must remember what
11894 is being waited on. This is not a problem when there are only a few
11895 steps, as in this example. But it can be a problem when there are
11898 @node No deferment solution
11899 @subsection No Deferment Solution
11900 @cindex No deferment solution
11901 @cindex Solution without deferment
11903 The solution to the problem of deferred operations is to write in a
11904 manner that does not defer operations@footnote{The phrase @dfn{tail
11905 recursive} is used to describe such a process, one that uses
11906 constant space.}. This requires
11907 writing to a different pattern, often one that involves writing two
11908 function definitions, an initialization function and a helper
11911 The initialization function sets up the job; the helper function
11915 Here are the two function definitions for adding up numbers. They are
11916 so simple, I find them hard to understand.
11920 (defun triangle-initialization (number)
11921 "Return the sum of the numbers 1 through NUMBER inclusive.
11922 This is the initialization component of a two function
11923 duo that uses recursion."
11924 (triangle-recursive-helper 0 0 number))
11930 (defun triangle-recursive-helper (sum counter number)
11931 "Return SUM, using COUNTER, through NUMBER inclusive.
11932 This is the helper component of a two function duo
11933 that uses recursion."
11934 (if (> counter number)
11936 (triangle-recursive-helper (+ sum counter) ; @r{sum}
11937 (1+ counter) ; @r{counter}
11938 number))) ; @r{number}
11943 Install both function definitions by evaluating them, then call
11944 @code{triangle-initialization} with 2 rows:
11948 (triangle-initialization 2)
11953 The initialization function calls the first instance of the helper
11954 function with three arguments: zero, zero, and a number which is the
11955 number of rows in the triangle.
11957 The first two arguments passed to the helper function are
11958 initialization values. These values are changed when
11959 @code{triangle-recursive-helper} invokes new instances.@footnote{The
11960 jargon is mildly confusing: @code{triangle-recursive-helper} uses a
11961 process that is iterative in a procedure that is recursive. The
11962 process is called iterative because the computer need only record the
11963 three values, @code{sum}, @code{counter}, and @code{number}; the
11964 procedure is recursive because the function calls itself. On the
11965 other hand, both the process and the procedure used by
11966 @code{triangle-recursively} are called recursive. The word
11967 ``recursive'' has different meanings in the two contexts.}
11969 Let's see what happens when we have a triangle that has one row. (This
11970 triangle will have one pebble in it!)
11973 @code{triangle-initialization} will call its helper with
11974 the arguments @w{@code{0 0 1}}. That function will run the conditional
11975 test whether @code{(> counter number)}:
11983 and find that the result is false, so it will invoke
11984 the else-part of the @code{if} clause:
11988 (triangle-recursive-helper
11989 (+ sum counter) ; @r{sum plus counter} @result{} @r{sum}
11990 (1+ counter) ; @r{increment counter} @result{} @r{counter}
11991 number) ; @r{number stays the same}
11997 which will first compute:
12001 (triangle-recursive-helper (+ 0 0) ; @r{sum}
12002 (1+ 0) ; @r{counter}
12006 (triangle-recursive-helper 0 1 1)
12010 Again, @code{(> counter number)} will be false, so again, the Lisp
12011 interpreter will evaluate @code{triangle-recursive-helper}, creating a
12012 new instance with new arguments.
12015 This new instance will be;
12019 (triangle-recursive-helper
12020 (+ sum counter) ; @r{sum plus counter} @result{} @r{sum}
12021 (1+ counter) ; @r{increment counter} @result{} @r{counter}
12022 number) ; @r{number stays the same}
12026 (triangle-recursive-helper 1 2 1)
12030 In this case, the @code{(> counter number)} test will be true! So the
12031 instance will return the value of the sum, which will be 1, as
12034 Now, let's pass @code{triangle-initialization} an argument
12035 of 2, to find out how many pebbles there are in a triangle with two rows.
12037 That function calls @code{(triangle-recursive-helper 0 0 2)}.
12040 In stages, the instances called will be:
12044 @r{sum counter number}
12045 (triangle-recursive-helper 0 1 2)
12047 (triangle-recursive-helper 1 2 2)
12049 (triangle-recursive-helper 3 3 2)
12053 When the last instance is called, the @code{(> counter number)} test
12054 will be true, so the instance will return the value of @code{sum},
12057 This kind of pattern helps when you are writing functions that can use
12058 many resources in a computer.
12061 @node Looping exercise
12062 @section Looping Exercise
12066 Write a function similar to @code{triangle} in which each row has a
12067 value which is the square of the row number. Use a @code{while} loop.
12070 Write a function similar to @code{triangle} that multiplies instead of
12074 Rewrite these two functions recursively. Rewrite these functions
12077 @c comma in printed title causes problem in Info cross reference
12079 Write a function for Texinfo mode that creates an index entry at the
12080 beginning of a paragraph for every @samp{@@dfn} within the paragraph.
12081 (In a Texinfo file, @samp{@@dfn} marks a definition. This book is
12082 written in Texinfo.)
12084 Many of the functions you will need are described in two of the
12085 previous chapters, @ref{Cutting & Storing Text, , Cutting and Storing
12086 Text}, and @ref{Yanking, , Yanking Text Back}. If you use
12087 @code{forward-paragraph} to put the index entry at the beginning of
12088 the paragraph, you will have to use @w{@kbd{C-h f}}
12089 (@code{describe-function}) to find out how to make the command go
12092 For more information, see
12094 @ref{Indicating, , Indicating Definitions, texinfo}.
12097 @ref{Indicating, , Indicating, texinfo, Texinfo Manual}, which goes to
12098 a Texinfo manual in the current directory. Or, if you are on the
12100 @uref{https://www.gnu.org/software/texinfo/manual/texinfo/}
12103 ``Indicating Definitions, Commands, etc.''@: in @cite{Texinfo, The GNU
12104 Documentation Format}.
12108 @node Regexp Search
12109 @chapter Regular Expression Searches
12110 @cindex Searches, illustrating
12111 @cindex Regular expression searches
12112 @cindex Patterns, searching for
12113 @cindex Motion by sentence and paragraph
12114 @cindex Sentences, movement by
12115 @cindex Paragraphs, movement by
12117 Regular expression searches are used extensively in GNU Emacs. The
12118 two functions, @code{forward-sentence} and @code{forward-paragraph},
12119 illustrate these searches well. They use regular expressions to find
12120 where to move point. The phrase ``regular expression'' is often written
12123 Regular expression searches are described in @ref{Regexp Search, ,
12124 Regular Expression Search, emacs, The GNU Emacs Manual}, as well as in
12125 @ref{Regular Expressions, , , elisp, The GNU Emacs Lisp Reference
12126 Manual}. In writing this chapter, I am presuming that you have at
12127 least a mild acquaintance with them. The major point to remember is
12128 that regular expressions permit you to search for patterns as well as
12129 for literal strings of characters. For example, the code in
12130 @code{forward-sentence} searches for the pattern of possible
12131 characters that could mark the end of a sentence, and moves point to
12134 Before looking at the code for the @code{forward-sentence} function, it
12135 is worth considering what the pattern that marks the end of a sentence
12136 must be. The pattern is discussed in the next section; following that
12137 is a description of the regular expression search function,
12138 @code{re-search-forward}. The @code{forward-sentence} function
12139 is described in the section following. Finally, the
12140 @code{forward-paragraph} function is described in the last section of
12141 this chapter. @code{forward-paragraph} is a complex function that
12142 introduces several new features.
12145 * sentence-end:: The regular expression for @code{sentence-end}.
12146 * re-search-forward:: Very similar to @code{search-forward}.
12147 * forward-sentence:: A straightforward example of regexp search.
12148 * forward-paragraph:: A somewhat complex example.
12150 * re-search Exercises::
12154 @section The Regular Expression for @code{sentence-end}
12155 @findex sentence-end
12157 The symbol @code{sentence-end} is bound to the pattern that marks the
12158 end of a sentence. What should this regular expression be?
12160 Clearly, a sentence may be ended by a period, a question mark, or an
12161 exclamation mark. Indeed, in English, only clauses that end with one
12162 of those three characters should be considered the end of a sentence.
12163 This means that the pattern should include the character set:
12169 However, we do not want @code{forward-sentence} merely to jump to a
12170 period, a question mark, or an exclamation mark, because such a character
12171 might be used in the middle of a sentence. A period, for example, is
12172 used after abbreviations. So other information is needed.
12174 According to convention, you type two spaces after every sentence, but
12175 only one space after a period, a question mark, or an exclamation mark in
12176 the body of a sentence. So a period, a question mark, or an exclamation
12177 mark followed by two spaces is a good indicator of an end of sentence.
12178 However, in a file, the two spaces may instead be a tab or the end of a
12179 line. This means that the regular expression should include these three
12180 items as alternatives.
12183 This group of alternatives will look like this:
12194 Here, @samp{$} indicates the end of the line, and I have pointed out
12195 where the tab and two spaces are inserted in the expression. Both are
12196 inserted by putting the actual characters into the expression.
12198 Two backslashes, @samp{\\}, are required before the parentheses and
12199 vertical bars: the first backslash quotes the following backslash in
12200 Emacs; and the second indicates that the following character, the
12201 parenthesis or the vertical bar, is special.
12204 Also, a sentence may be followed by one or more carriage returns, like
12215 Like tabs and spaces, a carriage return is inserted into a regular
12216 expression by inserting it literally. The asterisk indicates that the
12217 @key{RET} is repeated zero or more times.
12219 But a sentence end does not consist only of a period, a question mark or
12220 an exclamation mark followed by appropriate space: a closing quotation
12221 mark or a closing brace of some kind may precede the space. Indeed more
12222 than one such mark or brace may precede the space. These require a
12223 expression that looks like this:
12229 In this expression, the first @samp{]} is the first character in the
12230 expression; the second character is @samp{"}, which is preceded by a
12231 @samp{\} to tell Emacs the @samp{"} is @emph{not} special. The last
12232 three characters are @samp{'}, @samp{)}, and @samp{@}}.
12234 All this suggests what the regular expression pattern for matching the
12235 end of a sentence should be; and, indeed, if we evaluate
12236 @code{sentence-end} we find that it returns the following value:
12241 @result{} "[.?!][]\"')@}]*\\($\\| \\| \\)[
12247 (Well, not in GNU Emacs 22; that is because of an effort to make the
12248 process simpler and to handle more glyphs and languages. When the
12249 value of @code{sentence-end} is @code{nil}, then use the value defined
12250 by the function @code{sentence-end}. (Here is a use of the difference
12251 between a value and a function in Emacs Lisp.) The function returns a
12252 value constructed from the variables @code{sentence-end-base},
12253 @code{sentence-end-double-space}, @code{sentence-end-without-period},
12254 and @code{sentence-end-without-space}. The critical variable is
12255 @code{sentence-end-base}; its global value is similar to the one
12256 described above but it also contains two additional quotation marks.
12257 These have differing degrees of curliness. The
12258 @code{sentence-end-without-period} variable, when true, tells Emacs
12259 that a sentence may end without a period, such as text in Thai.)
12263 (Note that here the @key{TAB}, two spaces, and @key{RET} are shown
12264 literally in the pattern.)
12266 This regular expression can be deciphered as follows:
12270 The first part of the pattern is the three characters, a period, a question
12271 mark and an exclamation mark, within square brackets. The pattern must
12272 begin with one or other of these characters.
12275 The second part of the pattern is the group of closing braces and
12276 quotation marks, which can appear zero or more times. These may follow
12277 the period, question mark or exclamation mark. In a regular expression,
12278 the backslash, @samp{\}, followed by the double quotation mark,
12279 @samp{"}, indicates the class of string-quote characters. Usually, the
12280 double quotation mark is the only character in this class. The
12281 asterisk, @samp{*}, indicates that the items in the previous group (the
12282 group surrounded by square brackets, @samp{[]}) may be repeated zero or
12285 @item \\($\\| \\| \\)
12286 The third part of the pattern is one or other of: either the end of a
12287 line, or two blank spaces, or a tab. The double back-slashes are used
12288 to prevent Emacs from reading the parentheses and vertical bars as part
12289 of the search pattern; the parentheses are used to mark the group and
12290 the vertical bars are used to indicated that the patterns to either side
12291 of them are alternatives. The dollar sign is used to indicate the end
12292 of a line and both the two spaces and the tab are each inserted as is to
12293 indicate what they are.
12296 Finally, the last part of the pattern indicates that the end of the line
12297 or the whitespace following the period, question mark or exclamation
12298 mark may, but need not, be followed by one or more carriage returns. In
12299 the pattern, the carriage return is inserted as an actual carriage
12300 return between square brackets but here it is shown as @key{RET}.
12304 @node re-search-forward
12305 @section The @code{re-search-forward} Function
12306 @findex re-search-forward
12308 The @code{re-search-forward} function is very like the
12309 @code{search-forward} function. (@xref{search-forward, , The
12310 @code{search-forward} Function}.)
12312 @code{re-search-forward} searches for a regular expression. If the
12313 search is successful, it leaves point immediately after the last
12314 character in the target. If the search is backwards, it leaves point
12315 just before the first character in the target. You may tell
12316 @code{re-search-forward} to return @code{t} for true. (Moving point
12317 is therefore a side effect.)
12319 Like @code{search-forward}, the @code{re-search-forward} function takes
12324 The first argument is the regular expression that the function searches
12325 for. The regular expression will be a string between quotation marks.
12328 The optional second argument limits how far the function will search; it is a
12329 bound, which is specified as a position in the buffer.
12332 The optional third argument specifies how the function responds to
12333 failure: @code{nil} as the third argument causes the function to
12334 signal an error (and print a message) when the search fails; any other
12335 value causes it to return @code{nil} if the search fails and @code{t}
12336 if the search succeeds.
12339 The optional fourth argument is the repeat count. A negative repeat
12340 count causes @code{re-search-forward} to search backwards.
12344 The template for @code{re-search-forward} looks like this:
12348 (re-search-forward "@var{regular-expression}"
12349 @var{limit-of-search}
12350 @var{what-to-do-if-search-fails}
12351 @var{repeat-count})
12355 The second, third, and fourth arguments are optional. However, if you
12356 want to pass a value to either or both of the last two arguments, you
12357 must also pass a value to all the preceding arguments. Otherwise, the
12358 Lisp interpreter will mistake which argument you are passing the value
12362 In the @code{forward-sentence} function, the regular expression will be
12363 the value of the variable @code{sentence-end}. In simple form, that is:
12367 "[.?!][]\"')@}]*\\($\\| \\| \\)[
12373 The limit of the search will be the end of the paragraph (since a
12374 sentence cannot go beyond a paragraph). If the search fails, the
12375 function will return @code{nil}; and the repeat count will be provided
12376 by the argument to the @code{forward-sentence} function.
12378 @node forward-sentence
12379 @section @code{forward-sentence}
12380 @findex forward-sentence
12382 The command to move the cursor forward a sentence is a straightforward
12383 illustration of how to use regular expression searches in Emacs Lisp.
12384 Indeed, the function looks longer and more complicated than it is; this
12385 is because the function is designed to go backwards as well as forwards;
12386 and, optionally, over more than one sentence. The function is usually
12387 bound to the key command @kbd{M-e}.
12390 * Complete forward-sentence::
12391 * fwd-sentence while loops:: Two @code{while} loops.
12392 * fwd-sentence re-search:: A regular expression search.
12396 @node Complete forward-sentence
12397 @unnumberedsubsec Complete @code{forward-sentence} function definition
12401 Here is the code for @code{forward-sentence}:
12406 (defun forward-sentence (&optional arg)
12407 "Move forward to next end of sentence. With argument, repeat.
12408 With negative argument, move backward repeatedly to start of sentence.
12410 The variable `sentence-end' is a regular expression that matches ends of
12411 sentences. Also, every paragraph boundary terminates sentences as well."
12415 (or arg (setq arg 1))
12416 (let ((opoint (point))
12417 (sentence-end (sentence-end)))
12419 (let ((pos (point))
12420 (par-beg (save-excursion (start-of-paragraph-text) (point))))
12421 (if (and (re-search-backward sentence-end par-beg t)
12422 (or (< (match-end 0) pos)
12423 (re-search-backward sentence-end par-beg t)))
12424 (goto-char (match-end 0))
12425 (goto-char par-beg)))
12426 (setq arg (1+ arg)))
12430 (let ((par-end (save-excursion (end-of-paragraph-text) (point))))
12431 (if (re-search-forward sentence-end par-end t)
12432 (skip-chars-backward " \t\n")
12433 (goto-char par-end)))
12434 (setq arg (1- arg)))
12435 (constrain-to-field nil opoint t)))
12443 (defun forward-sentence (&optional arg)
12444 "Move forward to next sentence-end. With argument, repeat.
12445 With negative argument, move backward repeatedly to sentence-beginning.
12446 Sentence ends are identified by the value of sentence-end
12447 treated as a regular expression. Also, every paragraph boundary
12448 terminates sentences as well."
12452 (or arg (setq arg 1))
12455 (save-excursion (start-of-paragraph-text) (point))))
12456 (if (re-search-backward
12457 (concat sentence-end "[^ \t\n]") par-beg t)
12458 (goto-char (1- (match-end 0)))
12459 (goto-char par-beg)))
12460 (setq arg (1+ arg)))
12463 (save-excursion (end-of-paragraph-text) (point))))
12464 (if (re-search-forward sentence-end par-end t)
12465 (skip-chars-backward " \t\n")
12466 (goto-char par-end)))
12467 (setq arg (1- arg))))
12472 The function looks long at first sight and it is best to look at its
12473 skeleton first, and then its muscle. The way to see the skeleton is to
12474 look at the expressions that start in the left-most columns:
12478 (defun forward-sentence (&optional arg)
12479 "@var{documentation}@dots{}"
12481 (or arg (setq arg 1))
12482 (let ((opoint (point)) (sentence-end (sentence-end)))
12484 (let ((pos (point))
12485 (par-beg (save-excursion (start-of-paragraph-text) (point))))
12486 @var{rest-of-body-of-while-loop-when-going-backwards}
12488 (let ((par-end (save-excursion (end-of-paragraph-text) (point))))
12489 @var{rest-of-body-of-while-loop-when-going-forwards}
12490 @var{handle-forms-and-equivalent}
12494 This looks much simpler! The function definition consists of
12495 documentation, an @code{interactive} expression, an @code{or}
12496 expression, a @code{let} expression, and @code{while} loops.
12498 Let's look at each of these parts in turn.
12500 We note that the documentation is thorough and understandable.
12502 The function has an @code{interactive "p"} declaration. This means
12503 that the processed prefix argument, if any, is passed to the
12504 function as its argument. (This will be a number.) If the function
12505 is not passed an argument (it is optional) then the argument
12506 @code{arg} will be bound to 1.
12508 When @code{forward-sentence} is called non-interactively without an
12509 argument, @code{arg} is bound to @code{nil}. The @code{or} expression
12510 handles this. What it does is either leave the value of @code{arg} as
12511 it is, but only if @code{arg} is bound to a value; or it sets the
12512 value of @code{arg} to 1, in the case when @code{arg} is bound to
12515 Next is a @code{let}. That specifies the values of two local
12516 variables, @code{opoint} and @code{sentence-end}. The local value of
12517 point, from before the search, is used in the
12518 @code{constrain-to-field} function which handles forms and
12519 equivalents. The @code{sentence-end} variable is set by the
12520 @code{sentence-end} function.
12522 @node fwd-sentence while loops
12523 @unnumberedsubsec The @code{while} loops
12525 Two @code{while} loops follow. The first @code{while} has a
12526 true-or-false-test that tests true if the prefix argument for
12527 @code{forward-sentence} is a negative number. This is for going
12528 backwards. The body of this loop is similar to the body of the second
12529 @code{while} clause, but it is not exactly the same. We will skip
12530 this @code{while} loop and concentrate on the second @code{while}
12534 The second @code{while} loop is for moving point forward. Its skeleton
12539 (while (> arg 0) ; @r{true-or-false-test}
12541 (if (@var{true-or-false-test})
12544 (setq arg (1- arg)))) ; @code{while} @r{loop decrementer}
12548 The @code{while} loop is of the decrementing kind.
12549 (@xref{Decrementing Loop, , A Loop with a Decrementing Counter}.) It
12550 has a true-or-false-test that tests true so long as the counter (in
12551 this case, the variable @code{arg}) is greater than zero; and it has a
12552 decrementer that subtracts 1 from the value of the counter every time
12555 If no prefix argument is given to @code{forward-sentence}, which is
12556 the most common way the command is used, this @code{while} loop will
12557 run once, since the value of @code{arg} will be 1.
12559 The body of the @code{while} loop consists of a @code{let} expression,
12560 which creates and binds a local variable, and has, as its body, an
12561 @code{if} expression.
12564 The body of the @code{while} loop looks like this:
12569 (save-excursion (end-of-paragraph-text) (point))))
12570 (if (re-search-forward sentence-end par-end t)
12571 (skip-chars-backward " \t\n")
12572 (goto-char par-end)))
12576 The @code{let} expression creates and binds the local variable
12577 @code{par-end}. As we shall see, this local variable is designed to
12578 provide a bound or limit to the regular expression search. If the
12579 search fails to find a proper sentence ending in the paragraph, it will
12580 stop on reaching the end of the paragraph.
12582 But first, let us examine how @code{par-end} is bound to the value of
12583 the end of the paragraph. What happens is that the @code{let} sets the
12584 value of @code{par-end} to the value returned when the Lisp interpreter
12585 evaluates the expression
12589 (save-excursion (end-of-paragraph-text) (point))
12594 In this expression, @code{(end-of-paragraph-text)} moves point to the
12595 end of the paragraph, @code{(point)} returns the value of point, and then
12596 @code{save-excursion} restores point to its original position. Thus,
12597 the @code{let} binds @code{par-end} to the value returned by the
12598 @code{save-excursion} expression, which is the position of the end of
12599 the paragraph. (The @code{end-of-paragraph-text} function uses
12600 @code{forward-paragraph}, which we will discuss shortly.)
12603 Emacs next evaluates the body of the @code{let}, which is an @code{if}
12604 expression that looks like this:
12608 (if (re-search-forward sentence-end par-end t) ; @r{if-part}
12609 (skip-chars-backward " \t\n") ; @r{then-part}
12610 (goto-char par-end))) ; @r{else-part}
12614 The @code{if} tests whether its first argument is true and if so,
12615 evaluates its then-part; otherwise, the Emacs Lisp interpreter
12616 evaluates the else-part. The true-or-false-test of the @code{if}
12617 expression is the regular expression search.
12619 It may seem odd to have what looks like the real work of
12620 the @code{forward-sentence} function buried here, but this is a common
12621 way this kind of operation is carried out in Lisp.
12623 @node fwd-sentence re-search
12624 @unnumberedsubsec The regular expression search
12626 The @code{re-search-forward} function searches for the end of the
12627 sentence, that is, for the pattern defined by the @code{sentence-end}
12628 regular expression. If the pattern is found---if the end of the sentence is
12629 found---then the @code{re-search-forward} function does two things:
12633 The @code{re-search-forward} function carries out a side effect, which
12634 is to move point to the end of the occurrence found.
12637 The @code{re-search-forward} function returns a value of true. This is
12638 the value received by the @code{if}, and means that the search was
12643 The side effect, the movement of point, is completed before the
12644 @code{if} function is handed the value returned by the successful
12645 conclusion of the search.
12647 When the @code{if} function receives the value of true from a successful
12648 call to @code{re-search-forward}, the @code{if} evaluates the then-part,
12649 which is the expression @code{(skip-chars-backward " \t\n")}. This
12650 expression moves backwards over any blank spaces, tabs or carriage
12651 returns until a printed character is found and then leaves point after
12652 the character. Since point has already been moved to the end of the
12653 pattern that marks the end of the sentence, this action leaves point
12654 right after the closing printed character of the sentence, which is
12657 On the other hand, if the @code{re-search-forward} function fails to
12658 find a pattern marking the end of the sentence, the function returns
12659 false. The false then causes the @code{if} to evaluate its third
12660 argument, which is @code{(goto-char par-end)}: it moves point to the
12661 end of the paragraph.
12663 (And if the text is in a form or equivalent, and point may not move
12664 fully, then the @code{constrain-to-field} function comes into play.)
12666 Regular expression searches are exceptionally useful and the pattern
12667 illustrated by @code{re-search-forward}, in which the search is the
12668 test of an @code{if} expression, is handy. You will see or write code
12669 incorporating this pattern often.
12671 @node forward-paragraph
12672 @section @code{forward-paragraph}: a Goldmine of Functions
12673 @findex forward-paragraph
12677 (defun forward-paragraph (&optional arg)
12678 "Move forward to end of paragraph.
12679 With argument ARG, do it ARG times;
12680 a negative argument ARG = -N means move backward N paragraphs.
12682 A line which `paragraph-start' matches either separates paragraphs
12683 \(if `paragraph-separate' matches it also) or is the first line of a paragraph.
12684 A paragraph end is the beginning of a line which is not part of the paragraph
12685 to which the end of the previous line belongs, or the end of the buffer.
12686 Returns the count of paragraphs left to move."
12688 (or arg (setq arg 1))
12689 (let* ((opoint (point))
12690 (fill-prefix-regexp
12691 (and fill-prefix (not (equal fill-prefix ""))
12692 (not paragraph-ignore-fill-prefix)
12693 (regexp-quote fill-prefix)))
12694 ;; Remove ^ from paragraph-start and paragraph-sep if they are there.
12695 ;; These regexps shouldn't be anchored, because we look for them
12696 ;; starting at the left-margin. This allows paragraph commands to
12697 ;; work normally with indented text.
12698 ;; This hack will not find problem cases like "whatever\\|^something".
12699 (parstart (if (and (not (equal "" paragraph-start))
12700 (equal ?^ (aref paragraph-start 0)))
12701 (substring paragraph-start 1)
12703 (parsep (if (and (not (equal "" paragraph-separate))
12704 (equal ?^ (aref paragraph-separate 0)))
12705 (substring paragraph-separate 1)
12706 paragraph-separate))
12708 (if fill-prefix-regexp
12709 (concat parsep "\\|"
12710 fill-prefix-regexp "[ \t]*$")
12712 ;; This is used for searching.
12713 (sp-parstart (concat "^[ \t]*\\(?:" parstart "\\|" parsep "\\)"))
12715 (while (and (< arg 0) (not (bobp)))
12716 (if (and (not (looking-at parsep))
12717 (re-search-backward "^\n" (max (1- (point)) (point-min)) t)
12718 (looking-at parsep))
12719 (setq arg (1+ arg))
12720 (setq start (point))
12721 ;; Move back over paragraph-separating lines.
12722 (forward-char -1) (beginning-of-line)
12723 (while (and (not (bobp))
12724 (progn (move-to-left-margin)
12725 (looking-at parsep)))
12729 (setq arg (1+ arg))
12730 ;; Go to end of the previous (non-separating) line.
12732 ;; Search back for line that starts or separates paragraphs.
12733 (if (if fill-prefix-regexp
12734 ;; There is a fill prefix; it overrides parstart.
12735 (let (multiple-lines)
12736 (while (and (progn (beginning-of-line) (not (bobp)))
12737 (progn (move-to-left-margin)
12738 (not (looking-at parsep)))
12739 (looking-at fill-prefix-regexp))
12740 (unless (= (point) start)
12741 (setq multiple-lines t))
12743 (move-to-left-margin)
12744 ;; This deleted code caused a long hanging-indent line
12745 ;; not to be filled together with the following lines.
12746 ;; ;; Don't move back over a line before the paragraph
12747 ;; ;; which doesn't start with fill-prefix
12748 ;; ;; unless that is the only line we've moved over.
12749 ;; (and (not (looking-at fill-prefix-regexp))
12751 ;; (forward-line 1))
12753 (while (and (re-search-backward sp-parstart nil 1)
12754 (setq found-start t)
12755 ;; Found a candidate, but need to check if it is a
12757 (progn (setq start (point))
12758 (move-to-left-margin)
12759 (not (looking-at parsep)))
12760 (not (and (looking-at parstart)
12761 (or (not use-hard-newlines)
12764 (1- start) 'hard)))))
12765 (setq found-start nil)
12770 ;; Move forward over paragraph separators.
12771 ;; We know this cannot reach the place we started
12772 ;; because we know we moved back over a non-separator.
12773 (while (and (not (eobp))
12774 (progn (move-to-left-margin)
12775 (looking-at parsep)))
12777 ;; If line before paragraph is just margin, back up to there.
12779 (if (> (current-column) (current-left-margin))
12781 (skip-chars-backward " \t")
12783 (forward-line 1))))
12784 ;; No starter or separator line => use buffer beg.
12785 (goto-char (point-min))))))
12787 (while (and (> arg 0) (not (eobp)))
12788 ;; Move forward over separator lines...
12789 (while (and (not (eobp))
12790 (progn (move-to-left-margin) (not (eobp)))
12791 (looking-at parsep))
12793 (unless (eobp) (setq arg (1- arg)))
12794 ;; ... and one more line.
12796 (if fill-prefix-regexp
12797 ;; There is a fill prefix; it overrides parstart.
12798 (while (and (not (eobp))
12799 (progn (move-to-left-margin) (not (eobp)))
12800 (not (looking-at parsep))
12801 (looking-at fill-prefix-regexp))
12803 (while (and (re-search-forward sp-parstart nil 1)
12804 (progn (setq start (match-beginning 0))
12807 (progn (move-to-left-margin)
12808 (not (looking-at parsep)))
12809 (or (not (looking-at parstart))
12810 (and use-hard-newlines
12811 (not (get-text-property (1- start) 'hard)))))
12813 (if (< (point) (point-max))
12814 (goto-char start))))
12815 (constrain-to-field nil opoint t)
12816 ;; Return the number of steps that could not be done.
12820 The @code{forward-paragraph} function moves point forward to the end
12821 of the paragraph. It is usually bound to @kbd{M-@}} and makes use of a
12822 number of functions that are important in themselves, including
12823 @code{let*}, @code{match-beginning}, and @code{looking-at}.
12825 The function definition for @code{forward-paragraph} is considerably
12826 longer than the function definition for @code{forward-sentence}
12827 because it works with a paragraph, each line of which may begin with a
12830 A fill prefix consists of a string of characters that are repeated at
12831 the beginning of each line. For example, in Lisp code, it is a
12832 convention to start each line of a paragraph-long comment with
12833 @samp{;;; }. In Text mode, four blank spaces make up another common
12834 fill prefix, creating an indented paragraph. (@xref{Fill Prefix, , ,
12835 emacs, The GNU Emacs Manual}, for more information about fill
12838 The existence of a fill prefix means that in addition to being able to
12839 find the end of a paragraph whose lines begin on the left-most
12840 column, the @code{forward-paragraph} function must be able to find the
12841 end of a paragraph when all or many of the lines in the buffer begin
12842 with the fill prefix.
12844 Moreover, it is sometimes practical to ignore a fill prefix that
12845 exists, especially when blank lines separate paragraphs.
12846 This is an added complication.
12849 * forward-paragraph in brief:: Key parts of the function definition.
12850 * fwd-para let:: The @code{let*} expression.
12851 * fwd-para while:: The forward motion @code{while} loop.
12855 @node forward-paragraph in brief
12856 @unnumberedsubsec Shortened @code{forward-paragraph} function definition
12859 Rather than print all of the @code{forward-paragraph} function, we
12860 will only print parts of it. Read without preparation, the function
12864 In outline, the function looks like this:
12868 (defun forward-paragraph (&optional arg)
12869 "@var{documentation}@dots{}"
12871 (or arg (setq arg 1))
12874 (while (and (< arg 0) (not (bobp))) ; @r{backward-moving-code}
12876 (while (and (> arg 0) (not (eobp))) ; @r{forward-moving-code}
12881 The first parts of the function are routine: the function's argument
12882 list consists of one optional argument. Documentation follows.
12884 The lower case @samp{p} in the @code{interactive} declaration means
12885 that the processed prefix argument, if any, is passed to the function.
12886 This will be a number, and is the repeat count of how many paragraphs
12887 point will move. The @code{or} expression in the next line handles
12888 the common case when no argument is passed to the function, which occurs
12889 if the function is called from other code rather than interactively.
12890 This case was described earlier. (@xref{forward-sentence, The
12891 @code{forward-sentence} function}.) Now we reach the end of the
12892 familiar part of this function.
12895 @unnumberedsubsec The @code{let*} expression
12897 The next line of the @code{forward-paragraph} function begins a
12898 @code{let*} expression. This is a different than @code{let}. The
12899 symbol is @code{let*} not @code{let}.
12902 The @code{let*} special form is like @code{let} except that Emacs sets
12903 each variable in sequence, one after another, and variables in the
12904 latter part of the varlist can make use of the values to which Emacs
12905 set variables in the earlier part of the varlist.
12908 ( refappend save-excursion, , code save-excursion in code append-to-buffer .)
12911 (@ref{append save-excursion, , @code{save-excursion} in @code{append-to-buffer}}.)
12913 In the @code{let*} expression in this function, Emacs binds a total of
12914 seven variables: @code{opoint}, @code{fill-prefix-regexp},
12915 @code{parstart}, @code{parsep}, @code{sp-parstart}, @code{start}, and
12916 @code{found-start}.
12918 The variable @code{parsep} appears twice, first, to remove instances
12919 of @samp{^}, and second, to handle fill prefixes.
12921 The variable @code{opoint} is just the value of @code{point}. As you
12922 can guess, it is used in a @code{constrain-to-field} expression, just
12923 as in @code{forward-sentence}.
12925 The variable @code{fill-prefix-regexp} is set to the value returned by
12926 evaluating the following list:
12931 (not (equal fill-prefix ""))
12932 (not paragraph-ignore-fill-prefix)
12933 (regexp-quote fill-prefix))
12938 This is an expression whose first element is the @code{and} special form.
12940 As we learned earlier (@pxref{kill-new function, , The @code{kill-new}
12941 function}), the @code{and} special form evaluates each of its
12942 arguments until one of the arguments returns a value of @code{nil}, in
12943 which case the @code{and} expression returns @code{nil}; however, if
12944 none of the arguments returns a value of @code{nil}, the value
12945 resulting from evaluating the last argument is returned. (Since such
12946 a value is not @code{nil}, it is considered true in Lisp.) In other
12947 words, an @code{and} expression returns a true value only if all its
12948 arguments are true.
12951 In this case, the variable @code{fill-prefix-regexp} is bound to a
12952 non-@code{nil} value only if the following four expressions produce a
12953 true (i.e., a non-@code{nil}) value when they are evaluated; otherwise,
12954 @code{fill-prefix-regexp} is bound to @code{nil}.
12958 When this variable is evaluated, the value of the fill prefix, if any,
12959 is returned. If there is no fill prefix, this variable returns
12962 @item (not (equal fill-prefix "")
12963 This expression checks whether an existing fill prefix is an empty
12964 string, that is, a string with no characters in it. An empty string is
12965 not a useful fill prefix.
12967 @item (not paragraph-ignore-fill-prefix)
12968 This expression returns @code{nil} if the variable
12969 @code{paragraph-ignore-fill-prefix} has been turned on by being set to a
12970 true value such as @code{t}.
12972 @item (regexp-quote fill-prefix)
12973 This is the last argument to the @code{and} special form. If all the
12974 arguments to the @code{and} are true, the value resulting from
12975 evaluating this expression will be returned by the @code{and} expression
12976 and bound to the variable @code{fill-prefix-regexp},
12979 @findex regexp-quote
12981 The result of evaluating this @code{and} expression successfully is that
12982 @code{fill-prefix-regexp} will be bound to the value of
12983 @code{fill-prefix} as modified by the @code{regexp-quote} function.
12984 What @code{regexp-quote} does is read a string and return a regular
12985 expression that will exactly match the string and match nothing else.
12986 This means that @code{fill-prefix-regexp} will be set to a value that
12987 will exactly match the fill prefix if the fill prefix exists.
12988 Otherwise, the variable will be set to @code{nil}.
12990 The next two local variables in the @code{let*} expression are
12991 designed to remove instances of @samp{^} from @code{parstart} and
12992 @code{parsep}, the local variables which indicate the paragraph start
12993 and the paragraph separator. The next expression sets @code{parsep}
12994 again. That is to handle fill prefixes.
12996 This is the setting that requires the definition call @code{let*}
12997 rather than @code{let}. The true-or-false-test for the @code{if}
12998 depends on whether the variable @code{fill-prefix-regexp} evaluates to
12999 @code{nil} or some other value.
13001 If @code{fill-prefix-regexp} does not have a value, Emacs evaluates
13002 the else-part of the @code{if} expression and binds @code{parsep} to
13003 its local value. (@code{parsep} is a regular expression that matches
13004 what separates paragraphs.)
13006 But if @code{fill-prefix-regexp} does have a value, Emacs evaluates
13007 the then-part of the @code{if} expression and binds @code{parsep} to a
13008 regular expression that includes the @code{fill-prefix-regexp} as part
13011 Specifically, @code{parsep} is set to the original value of the
13012 paragraph separate regular expression concatenated with an alternative
13013 expression that consists of the @code{fill-prefix-regexp} followed by
13014 optional whitespace to the end of the line. The whitespace is defined
13015 by @w{@code{"[ \t]*$"}}.) The @samp{\\|} defines this portion of the
13016 regexp as an alternative to @code{parsep}.
13018 According to a comment in the code, the next local variable,
13019 @code{sp-parstart}, is used for searching, and then the final two,
13020 @code{start} and @code{found-start}, are set to @code{nil}.
13022 Now we get into the body of the @code{let*}. The first part of the body
13023 of the @code{let*} deals with the case when the function is given a
13024 negative argument and is therefore moving backwards. We will skip this
13027 @node fwd-para while
13028 @unnumberedsubsec The forward motion @code{while} loop
13030 The second part of the body of the @code{let*} deals with forward
13031 motion. It is a @code{while} loop that repeats itself so long as the
13032 value of @code{arg} is greater than zero. In the most common use of
13033 the function, the value of the argument is 1, so the body of the
13034 @code{while} loop is evaluated exactly once, and the cursor moves
13035 forward one paragraph.
13038 (while (and (> arg 0) (not (eobp)))
13040 ;; Move forward over separator lines...
13041 (while (and (not (eobp))
13042 (progn (move-to-left-margin) (not (eobp)))
13043 (looking-at parsep))
13045 (unless (eobp) (setq arg (1- arg)))
13046 ;; ... and one more line.
13049 (if fill-prefix-regexp
13050 ;; There is a fill prefix; it overrides parstart.
13051 (while (and (not (eobp))
13052 (progn (move-to-left-margin) (not (eobp)))
13053 (not (looking-at parsep))
13054 (looking-at fill-prefix-regexp))
13057 (while (and (re-search-forward sp-parstart nil 1)
13058 (progn (setq start (match-beginning 0))
13061 (progn (move-to-left-margin)
13062 (not (looking-at parsep)))
13063 (or (not (looking-at parstart))
13064 (and use-hard-newlines
13065 (not (get-text-property (1- start) 'hard)))))
13068 (if (< (point) (point-max))
13069 (goto-char start))))
13072 This part handles three situations: when point is between paragraphs,
13073 when there is a fill prefix and when there is no fill prefix.
13076 The @code{while} loop looks like this:
13080 ;; @r{going forwards and not at the end of the buffer}
13081 (while (and (> arg 0) (not (eobp)))
13083 ;; @r{between paragraphs}
13084 ;; Move forward over separator lines...
13085 (while (and (not (eobp))
13086 (progn (move-to-left-margin) (not (eobp)))
13087 (looking-at parsep))
13089 ;; @r{This decrements the loop}
13090 (unless (eobp) (setq arg (1- arg)))
13091 ;; ... and one more line.
13096 (if fill-prefix-regexp
13097 ;; There is a fill prefix; it overrides parstart;
13098 ;; we go forward line by line
13099 (while (and (not (eobp))
13100 (progn (move-to-left-margin) (not (eobp)))
13101 (not (looking-at parsep))
13102 (looking-at fill-prefix-regexp))
13107 ;; There is no fill prefix;
13108 ;; we go forward character by character
13109 (while (and (re-search-forward sp-parstart nil 1)
13110 (progn (setq start (match-beginning 0))
13113 (progn (move-to-left-margin)
13114 (not (looking-at parsep)))
13115 (or (not (looking-at parstart))
13116 (and use-hard-newlines
13117 (not (get-text-property (1- start) 'hard)))))
13122 ;; and if there is no fill prefix and if we are not at the end,
13123 ;; go to whatever was found in the regular expression search
13125 (if (< (point) (point-max))
13126 (goto-char start))))
13131 We can see that this is a decrementing counter @code{while} loop,
13132 using the expression @code{(setq arg (1- arg))} as the decrementer.
13133 That expression is not far from the @code{while}, but is hidden in
13134 another Lisp macro, an @code{unless} macro. Unless we are at the end
13135 of the buffer---that is what the @code{eobp} function determines; it
13136 is an abbreviation of @samp{End Of Buffer P}---we decrease the value
13137 of @code{arg} by one.
13139 (If we are at the end of the buffer, we cannot go forward any more and
13140 the next loop of the @code{while} expression will test false since the
13141 test is an @code{and} with @code{(not (eobp))}. The @code{not}
13142 function means exactly as you expect; it is another name for
13143 @code{null}, a function that returns true when its argument is false.)
13145 Interestingly, the loop count is not decremented until we leave the
13146 space between paragraphs, unless we come to the end of buffer or stop
13147 seeing the local value of the paragraph separator.
13149 That second @code{while} also has a @code{(move-to-left-margin)}
13150 expression. The function is self-explanatory. It is inside a
13151 @code{progn} expression and not the last element of its body, so it is
13152 only invoked for its side effect, which is to move point to the left
13153 margin of the current line.
13156 The @code{looking-at} function is also self-explanatory; it returns
13157 true if the text after point matches the regular expression given as
13160 The rest of the body of the loop looks difficult at first, but makes
13161 sense as you come to understand it.
13164 First consider what happens if there is a fill prefix:
13168 (if fill-prefix-regexp
13169 ;; There is a fill prefix; it overrides parstart;
13170 ;; we go forward line by line
13171 (while (and (not (eobp))
13172 (progn (move-to-left-margin) (not (eobp)))
13173 (not (looking-at parsep))
13174 (looking-at fill-prefix-regexp))
13180 This expression moves point forward line by line so long
13181 as four conditions are true:
13185 Point is not at the end of the buffer.
13188 We can move to the left margin of the text and are
13189 not at the end of the buffer.
13192 The text following point does not separate paragraphs.
13195 The pattern following point is the fill prefix regular expression.
13198 The last condition may be puzzling, until you remember that point was
13199 moved to the beginning of the line early in the @code{forward-paragraph}
13200 function. This means that if the text has a fill prefix, the
13201 @code{looking-at} function will see it.
13204 Consider what happens when there is no fill prefix.
13208 (while (and (re-search-forward sp-parstart nil 1)
13209 (progn (setq start (match-beginning 0))
13212 (progn (move-to-left-margin)
13213 (not (looking-at parsep)))
13214 (or (not (looking-at parstart))
13215 (and use-hard-newlines
13216 (not (get-text-property (1- start) 'hard)))))
13222 This @code{while} loop has us searching forward for
13223 @code{sp-parstart}, which is the combination of possible whitespace
13224 with the local value of the start of a paragraph or of a paragraph
13225 separator. (The latter two are within an expression starting
13226 @code{\(?:} so that they are not referenced by the
13227 @code{match-beginning} function.)
13230 The two expressions,
13234 (setq start (match-beginning 0))
13240 mean go to the start of the text matched by the regular expression
13243 The @code{(match-beginning 0)} expression is new. It returns a number
13244 specifying the location of the start of the text that was matched by
13247 The @code{match-beginning} function is used here because of a
13248 characteristic of a forward search: a successful forward search,
13249 regardless of whether it is a plain search or a regular expression
13250 search, moves point to the end of the text that is found. In this
13251 case, a successful search moves point to the end of the pattern for
13252 @code{sp-parstart}.
13254 However, we want to put point at the end of the current paragraph, not
13255 somewhere else. Indeed, since the search possibly includes the
13256 paragraph separator, point may end up at the beginning of the next one
13257 unless we use an expression that includes @code{match-beginning}.
13259 @findex match-beginning
13260 When given an argument of 0, @code{match-beginning} returns the
13261 position that is the start of the text matched by the most recent
13262 search. In this case, the most recent search looks for
13263 @code{sp-parstart}. The @code{(match-beginning 0)} expression returns
13264 the beginning position of that pattern, rather than the end position
13267 (Incidentally, when passed a positive number as an argument, the
13268 @code{match-beginning} function returns the location of point at that
13269 parenthesized expression in the last search unless that parenthesized
13270 expression begins with @code{\(?:}. I don't know why @code{\(?:}
13271 appears here since the argument is 0.)
13274 The last expression when there is no fill prefix is
13278 (if (< (point) (point-max))
13279 (goto-char start))))
13284 This says that if there is no fill prefix and if we are not at the
13285 end, point should move to the beginning of whatever was found by the
13286 regular expression search for @code{sp-parstart}.
13288 The full definition for the @code{forward-paragraph} function not only
13289 includes code for going forwards, but also code for going backwards.
13291 If you are reading this inside of GNU Emacs and you want to see the
13292 whole function, you can type @kbd{C-h f} (@code{describe-function})
13293 and the name of the function. This gives you the function
13294 documentation and the name of the library containing the function's
13295 source. Place point over the name of the library and press the @key{RET}
13296 key; you will be taken directly to the source. (Be sure to install
13297 your sources! Without them, you are like a person who tries to drive
13298 a car with his eyes shut!)
13300 @node Regexp Review
13303 Here is a brief summary of some recently introduced functions.
13307 Repeatedly evaluate the body of the expression so long as the first
13308 element of the body tests true. Then return @code{nil}. (The
13309 expression is evaluated only for its side effects.)
13318 (insert (format "foo is %d.\n" foo))
13319 (setq foo (1- foo))))
13321 @result{} foo is 2.
13328 (The @code{insert} function inserts its arguments at point; the
13329 @code{format} function returns a string formatted from its arguments
13330 the way @code{message} formats its arguments; @code{\n} produces a new
13333 @item re-search-forward
13334 Search for a pattern, and if the pattern is found, move point to rest
13338 Takes four arguments, like @code{search-forward}:
13342 A regular expression that specifies the pattern to search for.
13343 (Remember to put quotation marks around this argument!)
13346 Optionally, the limit of the search.
13349 Optionally, what to do if the search fails, return @code{nil} or an
13353 Optionally, how many times to repeat the search; if negative, the
13354 search goes backwards.
13358 Bind some variables locally to particular values,
13359 and then evaluate the remaining arguments, returning the value of the
13360 last one. While binding the local variables, use the local values of
13361 variables bound earlier, if any.
13370 (message "`bar' is %d." bar))
13371 @result{} ‘bar’ is 21.
13375 @item match-beginning
13376 Return the position of the start of the text found by the last regular
13380 Return @code{t} for true if the text after point matches the argument,
13381 which should be a regular expression.
13384 Return @code{t} for true if point is at the end of the accessible part
13385 of a buffer. The end of the accessible part is the end of the buffer
13386 if the buffer is not narrowed; it is the end of the narrowed part if
13387 the buffer is narrowed.
13391 @node re-search Exercises
13392 @section Exercises with @code{re-search-forward}
13396 Write a function to search for a regular expression that matches two
13397 or more blank lines in sequence.
13400 Write a function to search for duplicated words, such as ``the the''.
13401 @xref{Regexps, , Syntax of Regular Expressions, emacs, The GNU Emacs
13402 Manual}, for information on how to write a regexp (a regular
13403 expression) to match a string that is composed of two identical
13404 halves. You can devise several regexps; some are better than others.
13405 The function I use is described in an appendix, along with several
13406 regexps. @xref{the-the, , @code{the-the} Duplicated Words Function}.
13409 @node Counting Words
13410 @chapter Counting via Repetition and Regexps
13411 @cindex Repetition for word counting
13412 @cindex Regular expressions for word counting
13414 Repetition and regular expression searches are powerful tools that you
13415 often use when you write code in Emacs Lisp. This chapter illustrates
13416 the use of regular expression searches through the construction of
13417 word count commands using @code{while} loops and recursion.
13420 * Why Count Words::
13421 * @value{COUNT-WORDS}:: Use a regexp, but find a problem.
13422 * recursive-count-words:: Start with case of no words in region.
13423 * Counting Exercise::
13427 @node Why Count Words
13428 @unnumberedsec Counting words
13431 The standard Emacs distribution contains functions for counting the
13432 number of lines and words within a region.
13434 Certain types of writing ask you to count words. Thus, if you write
13435 an essay, you may be limited to 800 words; if you write a novel, you
13436 may discipline yourself to write 1000 words a day. It seems odd, but
13437 for a long time, Emacs lacked a word count command. Perhaps people used
13438 Emacs mostly for code or types of documentation that did not require
13439 word counts; or perhaps they restricted themselves to the operating
13440 system word count command, @code{wc}. Alternatively, people may have
13441 followed the publishers' convention and computed a word count by
13442 dividing the number of characters in a document by five.
13444 There are many ways to implement a command to count words. Here are
13445 some examples, which you may wish to compare with the standard Emacs
13446 command, @code{count-words-region}.
13448 @node @value{COUNT-WORDS}
13449 @section The @code{@value{COUNT-WORDS}} Function
13450 @findex @value{COUNT-WORDS}
13452 A word count command could count words in a line, paragraph, region,
13453 or buffer. What should the command cover? You could design the
13454 command to count the number of words in a complete buffer. However,
13455 the Emacs tradition encourages flexibility---you may want to count
13456 words in just a section, rather than all of a buffer. So it makes
13457 more sense to design the command to count the number of words in a
13458 region. Once you have a command to count words in a region, you can,
13459 if you wish, count words in a whole buffer by marking it with
13460 @w{@kbd{C-x h}} (@code{mark-whole-buffer}).
13462 Clearly, counting words is a repetitive act: starting from the
13463 beginning of the region, you count the first word, then the second
13464 word, then the third word, and so on, until you reach the end of the
13465 region. This means that word counting is ideally suited to recursion
13466 or to a @code{while} loop.
13469 * Design @value{COUNT-WORDS}:: The definition using a @code{while} loop.
13470 * Whitespace Bug:: The Whitespace Bug in @code{@value{COUNT-WORDS}}.
13474 @node Design @value{COUNT-WORDS}
13475 @unnumberedsubsec Designing @code{@value{COUNT-WORDS}}
13478 First, we will implement the word count command with a @code{while}
13479 loop, then with recursion. The command will, of course, be
13483 The template for an interactive function definition is, as always:
13487 (defun @var{name-of-function} (@var{argument-list})
13488 "@var{documentation}@dots{}"
13489 (@var{interactive-expression}@dots{})
13494 What we need to do is fill in the slots.
13496 The name of the function should be self-explanatory and similar to the
13497 existing @code{count-lines-region} name. This makes the name easier
13498 to remember. @code{count-words-region} is the obvious choice. Since
13499 that name is now used for the standard Emacs command to count words, we
13500 will name our implementation @code{@value{COUNT-WORDS}}.
13502 The function counts words within a region. This means that the
13503 argument list must contain symbols that are bound to the two
13504 positions, the beginning and end of the region. These two positions
13505 can be called @samp{beginning} and @samp{end} respectively. The first
13506 line of the documentation should be a single sentence, since that is
13507 all that is printed as documentation by a command such as
13508 @code{apropos}. The interactive expression will be of the form
13509 @samp{(interactive "r")}, since that will cause Emacs to pass the
13510 beginning and end of the region to the function's argument list. All
13513 The body of the function needs to be written to do three tasks:
13514 first, to set up conditions under which the @code{while} loop can
13515 count words, second, to run the @code{while} loop, and third, to send
13516 a message to the user.
13518 When a user calls @code{@value{COUNT-WORDS}}, point may be at the
13519 beginning or the end of the region. However, the counting process
13520 must start at the beginning of the region. This means we will want
13521 to put point there if it is not already there. Executing
13522 @code{(goto-char beginning)} ensures this. Of course, we will want to
13523 return point to its expected position when the function finishes its
13524 work. For this reason, the body must be enclosed in a
13525 @code{save-excursion} expression.
13527 The central part of the body of the function consists of a
13528 @code{while} loop in which one expression jumps point forward word by
13529 word, and another expression counts those jumps. The true-or-false-test
13530 of the @code{while} loop should test true so long as point should jump
13531 forward, and false when point is at the end of the region.
13533 We could use @code{(forward-word 1)} as the expression for moving point
13534 forward word by word, but it is easier to see what Emacs identifies as a
13535 ``word'' if we use a regular expression search.
13537 A regular expression search that finds the pattern for which it is
13538 searching leaves point after the last character matched. This means
13539 that a succession of successful word searches will move point forward
13542 As a practical matter, we want the regular expression search to jump
13543 over whitespace and punctuation between words as well as over the
13544 words themselves. A regexp that refuses to jump over interword
13545 whitespace would never jump more than one word! This means that
13546 the regexp should include the whitespace and punctuation that follows
13547 a word, if any, as well as the word itself. (A word may end a buffer
13548 and not have any following whitespace or punctuation, so that part of
13549 the regexp must be optional.)
13551 Thus, what we want for the regexp is a pattern defining one or more
13552 word constituent characters followed, optionally, by one or more
13553 characters that are not word constituents. The regular expression for
13561 The buffer's syntax table determines which characters are and are not
13562 word constituents. For more information about syntax,
13563 @pxref{Syntax Tables, , Syntax Tables, elisp, The GNU Emacs Lisp
13567 The search expression looks like this:
13570 (re-search-forward "\\w+\\W*")
13574 (Note that paired backslashes precede the @samp{w} and @samp{W}. A
13575 single backslash has special meaning to the Emacs Lisp interpreter.
13576 It indicates that the following character is interpreted differently
13577 than usual. For example, the two characters, @samp{\n}, stand for
13578 @samp{newline}, rather than for a backslash followed by @samp{n}. Two
13579 backslashes in a row stand for an ordinary, unspecial backslash, so
13580 Emacs Lisp interpreter ends of seeing a single backslash followed by a
13581 letter. So it discovers the letter is special.)
13583 We need a counter to count how many words there are; this variable
13584 must first be set to 0 and then incremented each time Emacs goes
13585 around the @code{while} loop. The incrementing expression is simply:
13588 (setq count (1+ count))
13591 Finally, we want to tell the user how many words there are in the
13592 region. The @code{message} function is intended for presenting this
13593 kind of information to the user. The message has to be phrased so
13594 that it reads properly regardless of how many words there are in the
13595 region: we don't want to say that ``there are 1 words in the region''.
13596 The conflict between singular and plural is ungrammatical. We can
13597 solve this problem by using a conditional expression that evaluates
13598 different messages depending on the number of words in the region.
13599 There are three possibilities: no words in the region, one word in the
13600 region, and more than one word. This means that the @code{cond}
13601 special form is appropriate.
13604 All this leads to the following function definition:
13608 ;;; @r{First version; has bugs!}
13609 (defun @value{COUNT-WORDS} (beginning end)
13610 "Print number of words in the region.
13611 Words are defined as at least one word-constituent
13612 character followed by at least one character that
13613 is not a word-constituent. The buffer's syntax
13614 table determines which characters these are."
13616 (message "Counting words in region ... ")
13620 ;;; @r{1. Set up appropriate conditions.}
13622 (goto-char beginning)
13627 ;;; @r{2. Run the} while @r{loop.}
13628 (while (< (point) end)
13629 (re-search-forward "\\w+\\W*")
13630 (setq count (1+ count)))
13634 ;;; @r{3. Send a message to the user.}
13635 (cond ((zerop count)
13637 "The region does NOT have any words."))
13640 "The region has 1 word."))
13643 "The region has %d words." count))))))
13648 As written, the function works, but not in all circumstances.
13650 @node Whitespace Bug
13651 @subsection The Whitespace Bug in @code{@value{COUNT-WORDS}}
13653 The @code{@value{COUNT-WORDS}} command described in the preceding
13654 section has two bugs, or rather, one bug with two manifestations.
13655 First, if you mark a region containing only whitespace in the middle
13656 of some text, the @code{@value{COUNT-WORDS}} command tells you that the
13657 region contains one word! Second, if you mark a region containing
13658 only whitespace at the end of the buffer or the accessible portion of
13659 a narrowed buffer, the command displays an error message that looks
13663 Search failed: "\\w+\\W*"
13666 If you are reading this in Info in GNU Emacs, you can test for these
13669 First, evaluate the function in the usual manner to install it.
13671 Here is a copy of the definition. Place your cursor after the closing
13672 parenthesis and type @kbd{C-x C-e} to install it.
13676 ;; @r{First version; has bugs!}
13677 (defun @value{COUNT-WORDS} (beginning end)
13678 "Print number of words in the region.
13679 Words are defined as at least one word-constituent character followed
13680 by at least one character that is not a word-constituent. The buffer's
13681 syntax table determines which characters these are."
13685 (message "Counting words in region ... ")
13689 ;;; @r{1. Set up appropriate conditions.}
13691 (goto-char beginning)
13696 ;;; @r{2. Run the} while @r{loop.}
13697 (while (< (point) end)
13698 (re-search-forward "\\w+\\W*")
13699 (setq count (1+ count)))
13703 ;;; @r{3. Send a message to the user.}
13704 (cond ((zerop count)
13705 (message "The region does NOT have any words."))
13706 ((= 1 count) (message "The region has 1 word."))
13707 (t (message "The region has %d words." count))))))
13713 If you wish, you can also install this keybinding by evaluating it:
13716 (global-set-key "\C-c=" '@value{COUNT-WORDS})
13719 To conduct the first test, set mark and point to the beginning and end
13720 of the following line and then type @kbd{C-c =} (or @kbd{M-x
13721 @value{COUNT-WORDS}} if you have not bound @kbd{C-c =}):
13728 Emacs will tell you, correctly, that the region has three words.
13730 Repeat the test, but place mark at the beginning of the line and place
13731 point just @emph{before} the word @samp{one}. Again type the command
13732 @kbd{C-c =} (or @kbd{M-x @value{COUNT-WORDS}}). Emacs should tell you
13733 that the region has no words, since it is composed only of the
13734 whitespace at the beginning of the line. But instead Emacs tells you
13735 that the region has one word!
13737 For the third test, copy the sample line to the end of the
13738 @file{*scratch*} buffer and then type several spaces at the end of the
13739 line. Place mark right after the word @samp{three} and point at the
13740 end of line. (The end of the line will be the end of the buffer.)
13741 Type @kbd{C-c =} (or @kbd{M-x @value{COUNT-WORDS}}) as you did before.
13742 Again, Emacs should tell you that the region has no words, since it is
13743 composed only of the whitespace at the end of the line. Instead,
13744 Emacs displays an error message saying @samp{Search failed}.
13746 The two bugs stem from the same problem.
13748 Consider the first manifestation of the bug, in which the command
13749 tells you that the whitespace at the beginning of the line contains
13750 one word. What happens is this: The @code{M-x @value{COUNT-WORDS}}
13751 command moves point to the beginning of the region. The @code{while}
13752 tests whether the value of point is smaller than the value of
13753 @code{end}, which it is. Consequently, the regular expression search
13754 looks for and finds the first word. It leaves point after the word.
13755 @code{count} is set to one. The @code{while} loop repeats; but this
13756 time the value of point is larger than the value of @code{end}, the
13757 loop is exited; and the function displays a message saying the number
13758 of words in the region is one. In brief, the regular expression
13759 search looks for and finds the word even though it is outside
13762 In the second manifestation of the bug, the region is whitespace at
13763 the end of the buffer. Emacs says @samp{Search failed}. What happens
13764 is that the true-or-false-test in the @code{while} loop tests true, so
13765 the search expression is executed. But since there are no more words
13766 in the buffer, the search fails.
13768 In both manifestations of the bug, the search extends or attempts to
13769 extend outside of the region.
13771 The solution is to limit the search to the region---this is a fairly
13772 simple action, but as you may have come to expect, it is not quite as
13773 simple as you might think.
13775 As we have seen, the @code{re-search-forward} function takes a search
13776 pattern as its first argument. But in addition to this first,
13777 mandatory argument, it accepts three optional arguments. The optional
13778 second argument bounds the search. The optional third argument, if
13779 @code{t}, causes the function to return @code{nil} rather than signal
13780 an error if the search fails. The optional fourth argument is a
13781 repeat count. (In Emacs, you can see a function's documentation by
13782 typing @kbd{C-h f}, the name of the function, and then @key{RET}.)
13784 In the @code{@value{COUNT-WORDS}} definition, the value of the end of
13785 the region is held by the variable @code{end} which is passed as an
13786 argument to the function. Thus, we can add @code{end} as an argument
13787 to the regular expression search expression:
13790 (re-search-forward "\\w+\\W*" end)
13793 However, if you make only this change to the @code{@value{COUNT-WORDS}}
13794 definition and then test the new version of the definition on a
13795 stretch of whitespace, you will receive an error message saying
13796 @samp{Search failed}.
13798 What happens is this: the search is limited to the region, and fails
13799 as you expect because there are no word-constituent characters in the
13800 region. Since it fails, we receive an error message. But we do not
13801 want to receive an error message in this case; we want to receive the
13802 message ``The region does NOT have any words.''
13804 The solution to this problem is to provide @code{re-search-forward}
13805 with a third argument of @code{t}, which causes the function to return
13806 @code{nil} rather than signal an error if the search fails.
13808 However, if you make this change and try it, you will see the message
13809 ``Counting words in region ... '' and @dots{} you will keep on seeing
13810 that message @dots{}, until you type @kbd{C-g} (@code{keyboard-quit}).
13812 Here is what happens: the search is limited to the region, as before,
13813 and it fails because there are no word-constituent characters in the
13814 region, as expected. Consequently, the @code{re-search-forward}
13815 expression returns @code{nil}. It does nothing else. In particular,
13816 it does not move point, which it does as a side effect if it finds the
13817 search target. After the @code{re-search-forward} expression returns
13818 @code{nil}, the next expression in the @code{while} loop is evaluated.
13819 This expression increments the count. Then the loop repeats. The
13820 true-or-false-test tests true because the value of point is still less
13821 than the value of end, since the @code{re-search-forward} expression
13822 did not move point. @dots{} and the cycle repeats @dots{}
13824 The @code{@value{COUNT-WORDS}} definition requires yet another
13825 modification, to cause the true-or-false-test of the @code{while} loop
13826 to test false if the search fails. Put another way, there are two
13827 conditions that must be satisfied in the true-or-false-test before the
13828 word count variable is incremented: point must still be within the
13829 region and the search expression must have found a word to count.
13831 Since both the first condition and the second condition must be true
13832 together, the two expressions, the region test and the search
13833 expression, can be joined with an @code{and} special form and embedded in
13834 the @code{while} loop as the true-or-false-test, like this:
13837 (and (< (point) end) (re-search-forward "\\w+\\W*" end t))
13840 @c colon in printed section title causes problem in Info cross reference
13841 @c also trouble with an overfull hbox
13844 (For information about @code{and}, see
13845 @ref{kill-new function, , The @code{kill-new} function}.)
13849 (@xref{kill-new function, , The @code{kill-new} function}, for
13850 information about @code{and}.)
13853 The @code{re-search-forward} expression returns @code{t} if the search
13854 succeeds and as a side effect moves point. Consequently, as words are
13855 found, point is moved through the region. When the search expression
13856 fails to find another word, or when point reaches the end of the
13857 region, the true-or-false-test tests false, the @code{while} loop
13858 exits, and the @code{@value{COUNT-WORDS}} function displays one or
13859 other of its messages.
13861 After incorporating these final changes, the @code{@value{COUNT-WORDS}}
13862 works without bugs (or at least, without bugs that I have found!).
13863 Here is what it looks like:
13867 ;;; @r{Final version:} @code{while}
13868 (defun @value{COUNT-WORDS} (beginning end)
13869 "Print number of words in the region."
13871 (message "Counting words in region ... ")
13875 ;;; @r{1. Set up appropriate conditions.}
13878 (goto-char beginning)
13882 ;;; @r{2. Run the} while @r{loop.}
13883 (while (and (< (point) end)
13884 (re-search-forward "\\w+\\W*" end t))
13885 (setq count (1+ count)))
13889 ;;; @r{3. Send a message to the user.}
13890 (cond ((zerop count)
13892 "The region does NOT have any words."))
13895 "The region has 1 word."))
13898 "The region has %d words." count))))))
13902 @node recursive-count-words
13903 @section Count Words Recursively
13904 @cindex Count words recursively
13905 @cindex Recursively counting words
13906 @cindex Words, counted recursively
13908 You can write the function for counting words recursively as well as
13909 with a @code{while} loop. Let's see how this is done.
13911 First, we need to recognize that the @code{@value{COUNT-WORDS}}
13912 function has three jobs: it sets up the appropriate conditions for
13913 counting to occur; it counts the words in the region; and it sends a
13914 message to the user telling how many words there are.
13916 If we write a single recursive function to do everything, we will
13917 receive a message for every recursive call. If the region contains 13
13918 words, we will receive thirteen messages, one right after the other.
13919 We don't want this! Instead, we must write two functions to do the
13920 job, one of which (the recursive function) will be used inside of the
13921 other. One function will set up the conditions and display the
13922 message; the other will return the word count.
13924 Let us start with the function that causes the message to be displayed.
13925 We can continue to call this @code{@value{COUNT-WORDS}}.
13927 This is the function that the user will call. It will be interactive.
13928 Indeed, it will be similar to our previous versions of this
13929 function, except that it will call @code{recursive-count-words} to
13930 determine how many words are in the region.
13933 We can readily construct a template for this function, based on our
13938 ;; @r{Recursive version; uses regular expression search}
13939 (defun @value{COUNT-WORDS} (beginning end)
13940 "@var{documentation}@dots{}"
13941 (@var{interactive-expression}@dots{})
13945 ;;; @r{1. Set up appropriate conditions.}
13946 (@var{explanatory message})
13947 (@var{set-up functions}@dots{}
13951 ;;; @r{2. Count the words.}
13952 @var{recursive call}
13956 ;;; @r{3. Send a message to the user.}
13957 @var{message providing word count}))
13961 The definition looks straightforward, except that somehow the count
13962 returned by the recursive call must be passed to the message
13963 displaying the word count. A little thought suggests that this can be
13964 done by making use of a @code{let} expression: we can bind a variable
13965 in the varlist of a @code{let} expression to the number of words in
13966 the region, as returned by the recursive call; and then the
13967 @code{cond} expression, using binding, can display the value to the
13970 Often, one thinks of the binding within a @code{let} expression as
13971 somehow secondary to the primary work of a function. But in this
13972 case, what you might consider the primary job of the function,
13973 counting words, is done within the @code{let} expression.
13976 Using @code{let}, the function definition looks like this:
13980 (defun @value{COUNT-WORDS} (beginning end)
13981 "Print number of words in the region."
13986 ;;; @r{1. Set up appropriate conditions.}
13987 (message "Counting words in region ... ")
13989 (goto-char beginning)
13993 ;;; @r{2. Count the words.}
13994 (let ((count (recursive-count-words end)))
13998 ;;; @r{3. Send a message to the user.}
13999 (cond ((zerop count)
14001 "The region does NOT have any words."))
14004 "The region has 1 word."))
14007 "The region has %d words." count))))))
14011 Next, we need to write the recursive counting function.
14013 A recursive function has at least three parts: the do-again-test, the
14014 next-step-expression, and the recursive call.
14016 The do-again-test determines whether the function will or will not be
14017 called again. Since we are counting words in a region and can use a
14018 function that moves point forward for every word, the do-again-test
14019 can check whether point is still within the region. The do-again-test
14020 should find the value of point and determine whether point is before,
14021 at, or after the value of the end of the region. We can use the
14022 @code{point} function to locate point. Clearly, we must pass the
14023 value of the end of the region to the recursive counting function as an
14026 In addition, the do-again-test should also test whether the search finds a
14027 word. If it does not, the function should not call itself again.
14029 The next-step-expression changes a value so that when the recursive
14030 function is supposed to stop calling itself, it stops. More
14031 precisely, the next-step-expression changes a value so that at the
14032 right time, the do-again-test stops the recursive function from
14033 calling itself again. In this case, the next-step-expression can be
14034 the expression that moves point forward, word by word.
14036 The third part of a recursive function is the recursive call.
14038 Somewhere, we also need a part that does the work of the
14039 function, a part that does the counting. A vital part!
14042 But already, we have an outline of the recursive counting function:
14046 (defun recursive-count-words (region-end)
14047 "@var{documentation}@dots{}"
14048 @var{do-again-test}
14049 @var{next-step-expression}
14050 @var{recursive call})
14054 Now we need to fill in the slots. Let's start with the simplest cases
14055 first: if point is at or beyond the end of the region, there cannot
14056 be any words in the region, so the function should return zero.
14057 Likewise, if the search fails, there are no words to count, so the
14058 function should return zero.
14060 On the other hand, if point is within the region and the search
14061 succeeds, the function should call itself again.
14064 Thus, the do-again-test should look like this:
14068 (and (< (point) region-end)
14069 (re-search-forward "\\w+\\W*" region-end t))
14073 Note that the search expression is part of the do-again-test---the
14074 function returns @code{t} if its search succeeds and @code{nil} if it
14075 fails. (@xref{Whitespace Bug, , The Whitespace Bug in
14076 @code{@value{COUNT-WORDS}}}, for an explanation of how
14077 @code{re-search-forward} works.)
14079 The do-again-test is the true-or-false test of an @code{if} clause.
14080 Clearly, if the do-again-test succeeds, the then-part of the @code{if}
14081 clause should call the function again; but if it fails, the else-part
14082 should return zero since either point is outside the region or the
14083 search failed because there were no words to find.
14085 But before considering the recursive call, we need to consider the
14086 next-step-expression. What is it? Interestingly, it is the search
14087 part of the do-again-test.
14089 In addition to returning @code{t} or @code{nil} for the
14090 do-again-test, @code{re-search-forward} moves point forward as a side
14091 effect of a successful search. This is the action that changes the
14092 value of point so that the recursive function stops calling itself
14093 when point completes its movement through the region. Consequently,
14094 the @code{re-search-forward} expression is the next-step-expression.
14097 In outline, then, the body of the @code{recursive-count-words}
14098 function looks like this:
14102 (if @var{do-again-test-and-next-step-combined}
14104 @var{recursive-call-returning-count}
14110 How to incorporate the mechanism that counts?
14112 If you are not used to writing recursive functions, a question like
14113 this can be troublesome. But it can and should be approached
14116 We know that the counting mechanism should be associated in some way
14117 with the recursive call. Indeed, since the next-step-expression moves
14118 point forward by one word, and since a recursive call is made for
14119 each word, the counting mechanism must be an expression that adds one
14120 to the value returned by a call to @code{recursive-count-words}.
14123 Consider several cases:
14127 If there are two words in the region, the function should return
14128 a value resulting from adding one to the value returned when it counts
14129 the first word, plus the number returned when it counts the remaining
14130 words in the region, which in this case is one.
14133 If there is one word in the region, the function should return
14134 a value resulting from adding one to the value returned when it counts
14135 that word, plus the number returned when it counts the remaining
14136 words in the region, which in this case is zero.
14139 If there are no words in the region, the function should return zero.
14142 From the sketch we can see that the else-part of the @code{if} returns
14143 zero for the case of no words. This means that the then-part of the
14144 @code{if} must return a value resulting from adding one to the value
14145 returned from a count of the remaining words.
14148 The expression will look like this, where @code{1+} is a function that
14149 adds one to its argument.
14152 (1+ (recursive-count-words region-end))
14156 The whole @code{recursive-count-words} function will then look like
14161 (defun recursive-count-words (region-end)
14162 "@var{documentation}@dots{}"
14164 ;;; @r{1. do-again-test}
14165 (if (and (< (point) region-end)
14166 (re-search-forward "\\w+\\W*" region-end t))
14170 ;;; @r{2. then-part: the recursive call}
14171 (1+ (recursive-count-words region-end))
14173 ;;; @r{3. else-part}
14179 Let's examine how this works:
14181 If there are no words in the region, the else part of the @code{if}
14182 expression is evaluated and consequently the function returns zero.
14184 If there is one word in the region, the value of point is less than
14185 the value of @code{region-end} and the search succeeds. In this case,
14186 the true-or-false-test of the @code{if} expression tests true, and the
14187 then-part of the @code{if} expression is evaluated. The counting
14188 expression is evaluated. This expression returns a value (which will
14189 be the value returned by the whole function) that is the sum of one
14190 added to the value returned by a recursive call.
14192 Meanwhile, the next-step-expression has caused point to jump over the
14193 first (and in this case only) word in the region. This means that
14194 when @code{(recursive-count-words region-end)} is evaluated a second
14195 time, as a result of the recursive call, the value of point will be
14196 equal to or greater than the value of region end. So this time,
14197 @code{recursive-count-words} will return zero. The zero will be added
14198 to one, and the original evaluation of @code{recursive-count-words}
14199 will return one plus zero, which is one, which is the correct amount.
14201 Clearly, if there are two words in the region, the first call to
14202 @code{recursive-count-words} returns one added to the value returned
14203 by calling @code{recursive-count-words} on a region containing the
14204 remaining word---that is, it adds one to one, producing two, which is
14205 the correct amount.
14207 Similarly, if there are three words in the region, the first call to
14208 @code{recursive-count-words} returns one added to the value returned
14209 by calling @code{recursive-count-words} on a region containing the
14210 remaining two words---and so on and so on.
14214 With full documentation the two functions look like this:
14218 The recursive function:
14220 @findex recursive-count-words
14223 (defun recursive-count-words (region-end)
14224 "Number of words between point and REGION-END."
14228 ;;; @r{1. do-again-test}
14229 (if (and (< (point) region-end)
14230 (re-search-forward "\\w+\\W*" region-end t))
14234 ;;; @r{2. then-part: the recursive call}
14235 (1+ (recursive-count-words region-end))
14237 ;;; @r{3. else-part}
14248 ;;; @r{Recursive version}
14249 (defun @value{COUNT-WORDS} (beginning end)
14250 "Print number of words in the region.
14254 Words are defined as at least one word-constituent
14255 character followed by at least one character that is
14256 not a word-constituent. The buffer's syntax table
14257 determines which characters these are."
14261 (message "Counting words in region ... ")
14263 (goto-char beginning)
14264 (let ((count (recursive-count-words end)))
14267 (cond ((zerop count)
14269 "The region does NOT have any words."))
14273 (message "The region has 1 word."))
14276 "The region has %d words." count))))))
14280 @node Counting Exercise
14281 @section Exercise: Counting Punctuation
14283 Using a @code{while} loop, write a function to count the number of
14284 punctuation marks in a region---period, comma, semicolon, colon,
14285 exclamation mark, and question mark. Do the same using recursion.
14287 @node Words in a defun
14288 @chapter Counting Words in a @code{defun}
14289 @cindex Counting words in a @code{defun}
14290 @cindex Word counting in a @code{defun}
14292 Our next project is to count the number of words in a function
14293 definition. Clearly, this can be done using some variant of
14294 @code{@value{COUNT-WORDS}}. @xref{Counting Words, , Counting via
14295 Repetition and Regexps}. If we are just going to count the words in
14296 one definition, it is easy enough to mark the definition with the
14297 @kbd{C-M-h} (@code{mark-defun}) command, and then call
14298 @code{@value{COUNT-WORDS}}.
14300 However, I am more ambitious: I want to count the words and symbols in
14301 every definition in the Emacs sources and then print a graph that
14302 shows how many functions there are of each length: how many contain 40
14303 to 49 words or symbols, how many contain 50 to 59 words or symbols,
14304 and so on. I have often been curious how long a typical function is,
14305 and this will tell.
14308 * Divide and Conquer::
14309 * Words and Symbols:: What to count?
14310 * Syntax:: What constitutes a word or symbol?
14311 * count-words-in-defun:: Very like @code{@value{COUNT-WORDS}}.
14312 * Several defuns:: Counting several defuns in a file.
14313 * Find a File:: Do you want to look at a file?
14314 * lengths-list-file:: A list of the lengths of many definitions.
14315 * Several files:: Counting in definitions in different files.
14316 * Several files recursively:: Recursively counting in different files.
14317 * Prepare the data:: Prepare the data for display in a graph.
14321 @node Divide and Conquer
14322 @unnumberedsec Divide and Conquer
14325 Described in one phrase, the histogram project is daunting; but
14326 divided into numerous small steps, each of which we can take one at a
14327 time, the project becomes less fearsome. Let us consider what the
14332 First, write a function to count the words in one definition. This
14333 includes the problem of handling symbols as well as words.
14336 Second, write a function to list the number of words in each function
14337 in a file. This function can use the @code{count-words-in-defun}
14341 Third, write a function to list the number of words in each function
14342 in each of several files. This entails automatically finding the
14343 various files, switching to them, and counting the words in the
14344 definitions within them.
14347 Fourth, write a function to convert the list of numbers that we
14348 created in step three to a form that will be suitable for printing as
14352 Fifth, write a function to print the results as a graph.
14355 This is quite a project! But if we take each step slowly, it will not
14358 @node Words and Symbols
14359 @section What to Count?
14360 @cindex Words and symbols in defun
14362 When we first start thinking about how to count the words in a
14363 function definition, the first question is (or ought to be) what are
14364 we going to count? When we speak of ``words'' with respect to a Lisp
14365 function definition, we are actually speaking, in large part, of
14366 symbols. For example, the following @code{multiply-by-seven}
14367 function contains the five symbols @code{defun},
14368 @code{multiply-by-seven}, @code{number}, @code{*}, and @code{7}. In
14369 addition, in the documentation string, it contains the four words
14370 @samp{Multiply}, @samp{NUMBER}, @samp{by}, and @samp{seven}. The
14371 symbol @samp{number} is repeated, so the definition contains a total
14372 of ten words and symbols.
14376 (defun multiply-by-seven (number)
14377 "Multiply NUMBER by seven."
14383 However, if we mark the @code{multiply-by-seven} definition with
14384 @kbd{C-M-h} (@code{mark-defun}), and then call
14385 @code{@value{COUNT-WORDS}} on it, we will find that
14386 @code{@value{COUNT-WORDS}} claims the definition has eleven words, not
14387 ten! Something is wrong!
14389 The problem is twofold: @code{@value{COUNT-WORDS}} does not count the
14390 @samp{*} as a word, and it counts the single symbol,
14391 @code{multiply-by-seven}, as containing three words. The hyphens are
14392 treated as if they were interword spaces rather than intraword
14393 connectors: @samp{multiply-by-seven} is counted as if it were written
14394 @samp{multiply by seven}.
14396 The cause of this confusion is the regular expression search within
14397 the @code{@value{COUNT-WORDS}} definition that moves point forward word
14398 by word. In the canonical version of @code{@value{COUNT-WORDS}}, the
14406 This regular expression is a pattern defining one or more word
14407 constituent characters possibly followed by one or more characters
14408 that are not word constituents. What is meant by ``word constituent
14409 characters'' brings us to the issue of syntax, which is worth a section
14413 @section What Constitutes a Word or Symbol?
14414 @cindex Syntax categories and tables
14416 Emacs treats different characters as belonging to different
14417 @dfn{syntax categories}. For example, the regular expression,
14418 @samp{\\w+}, is a pattern specifying one or more @emph{word
14419 constituent} characters. Word constituent characters are members of
14420 one syntax category. Other syntax categories include the class of
14421 punctuation characters, such as the period and the comma, and the
14422 class of whitespace characters, such as the blank space and the tab
14423 character. (For more information, @pxref{Syntax Tables, , Syntax
14424 Tables, elisp, The GNU Emacs Lisp Reference Manual}.)
14426 Syntax tables specify which characters belong to which categories.
14427 Usually, a hyphen is not specified as a word constituent character.
14428 Instead, it is specified as being in the class of characters that are
14429 part of symbol names but not words. This means that the
14430 @code{@value{COUNT-WORDS}} function treats it in the same way it treats
14431 an interword white space, which is why @code{@value{COUNT-WORDS}}
14432 counts @samp{multiply-by-seven} as three words.
14434 There are two ways to cause Emacs to count @samp{multiply-by-seven} as
14435 one symbol: modify the syntax table or modify the regular expression.
14437 We could redefine a hyphen as a word constituent character by
14438 modifying the syntax table that Emacs keeps for each mode. This
14439 action would serve our purpose, except that a hyphen is merely the
14440 most common character within symbols that is not typically a word
14441 constituent character; there are others, too.
14443 Alternatively, we can redefine the regexp used in the
14444 @code{@value{COUNT-WORDS}} definition so as to include symbols. This
14445 procedure has the merit of clarity, but the task is a little tricky.
14448 The first part is simple enough: the pattern must match at least one
14449 character that is a word or symbol constituent. Thus:
14452 "\\(\\w\\|\\s_\\)+"
14456 The @samp{\\(} is the first part of the grouping construct that
14457 includes the @samp{\\w} and the @samp{\\s_} as alternatives, separated
14458 by the @samp{\\|}. The @samp{\\w} matches any word-constituent
14459 character and the @samp{\\s_} matches any character that is part of a
14460 symbol name but not a word-constituent character. The @samp{+}
14461 following the group indicates that the word or symbol constituent
14462 characters must be matched at least once.
14464 However, the second part of the regexp is more difficult to design.
14465 What we want is to follow the first part with optionally one or more
14466 characters that are not constituents of a word or symbol. At first,
14467 I thought I could define this with the following:
14470 "\\(\\W\\|\\S_\\)*"
14474 The upper case @samp{W} and @samp{S} match characters that are
14475 @emph{not} word or symbol constituents. Unfortunately, this
14476 expression matches any character that is either not a word constituent
14477 or not a symbol constituent. This matches any character!
14479 I then noticed that every word or symbol in my test region was
14480 followed by white space (blank space, tab, or newline). So I tried
14481 placing a pattern to match one or more blank spaces after the pattern
14482 for one or more word or symbol constituents. This failed, too. Words
14483 and symbols are often separated by whitespace, but in actual code
14484 parentheses may follow symbols and punctuation may follow words. So
14485 finally, I designed a pattern in which the word or symbol constituents
14486 are followed optionally by characters that are not white space and
14487 then followed optionally by white space.
14490 Here is the full regular expression:
14493 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*"
14496 @node count-words-in-defun
14497 @section The @code{count-words-in-defun} Function
14498 @cindex Counting words in a @code{defun}
14500 We have seen that there are several ways to write a
14501 @code{count-words-region} function. To write a
14502 @code{count-words-in-defun}, we need merely adapt one of these
14505 The version that uses a @code{while} loop is easy to understand, so I
14506 am going to adapt that. Because @code{count-words-in-defun} will be
14507 part of a more complex program, it need not be interactive and it need
14508 not display a message but just return the count. These considerations
14509 simplify the definition a little.
14511 On the other hand, @code{count-words-in-defun} will be used within a
14512 buffer that contains function definitions. Consequently, it is
14513 reasonable to ask that the function determine whether it is called
14514 when point is within a function definition, and if it is, to return
14515 the count for that definition. This adds complexity to the
14516 definition, but saves us from needing to pass arguments to the
14520 These considerations lead us to prepare the following template:
14524 (defun count-words-in-defun ()
14525 "@var{documentation}@dots{}"
14526 (@var{set up}@dots{}
14527 (@var{while loop}@dots{})
14528 @var{return count})
14533 As usual, our job is to fill in the slots.
14537 We are presuming that this function will be called within a buffer
14538 containing function definitions. Point will either be within a
14539 function definition or not. For @code{count-words-in-defun} to work,
14540 point must move to the beginning of the definition, a counter must
14541 start at zero, and the counting loop must stop when point reaches the
14542 end of the definition.
14544 The @code{beginning-of-defun} function searches backwards for an
14545 opening delimiter such as a @samp{(} at the beginning of a line, and
14546 moves point to that position, or else to the limit of the search. In
14547 practice, this means that @code{beginning-of-defun} moves point to the
14548 beginning of an enclosing or preceding function definition, or else to
14549 the beginning of the buffer. We can use @code{beginning-of-defun} to
14550 place point where we wish to start.
14552 The @code{while} loop requires a counter to keep track of the words or
14553 symbols being counted. A @code{let} expression can be used to create
14554 a local variable for this purpose, and bind it to an initial value of zero.
14556 The @code{end-of-defun} function works like @code{beginning-of-defun}
14557 except that it moves point to the end of the definition.
14558 @code{end-of-defun} can be used as part of an expression that
14559 determines the position of the end of the definition.
14561 The set up for @code{count-words-in-defun} takes shape rapidly: first
14562 we move point to the beginning of the definition, then we create a
14563 local variable to hold the count, and finally, we record the position
14564 of the end of the definition so the @code{while} loop will know when to stop
14568 The code looks like this:
14572 (beginning-of-defun)
14574 (end (save-excursion (end-of-defun) (point))))
14579 The code is simple. The only slight complication is likely to concern
14580 @code{end}: it is bound to the position of the end of the definition
14581 by a @code{save-excursion} expression that returns the value of point
14582 after @code{end-of-defun} temporarily moves it to the end of the
14585 The second part of the @code{count-words-in-defun}, after the set up,
14586 is the @code{while} loop.
14588 The loop must contain an expression that jumps point forward word by
14589 word and symbol by symbol, and another expression that counts the
14590 jumps. The true-or-false-test for the @code{while} loop should test
14591 true so long as point should jump forward, and false when point is at
14592 the end of the definition. We have already redefined the regular
14593 expression for this, so the loop is straightforward:
14597 (while (and (< (point) end)
14599 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*" end t))
14600 (setq count (1+ count)))
14604 The third part of the function definition returns the count of words
14605 and symbols. This part is the last expression within the body of the
14606 @code{let} expression, and can be, very simply, the local variable
14607 @code{count}, which when evaluated returns the count.
14610 Put together, the @code{count-words-in-defun} definition looks like this:
14612 @findex count-words-in-defun
14615 (defun count-words-in-defun ()
14616 "Return the number of words and symbols in a defun."
14617 (beginning-of-defun)
14619 (end (save-excursion (end-of-defun) (point))))
14623 (and (< (point) end)
14625 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*"
14627 (setq count (1+ count)))
14632 How to test this? The function is not interactive, but it is easy to
14633 put a wrapper around the function to make it interactive; we can use
14634 almost the same code as for the recursive version of
14635 @code{@value{COUNT-WORDS}}:
14639 ;;; @r{Interactive version.}
14640 (defun count-words-defun ()
14641 "Number of words and symbols in a function definition."
14644 "Counting words and symbols in function definition ... ")
14647 (let ((count (count-words-in-defun)))
14651 "The definition does NOT have any words or symbols."))
14656 "The definition has 1 word or symbol."))
14659 "The definition has %d words or symbols." count)))))
14665 Let's re-use @kbd{C-c =} as a convenient keybinding:
14668 (global-set-key "\C-c=" 'count-words-defun)
14671 Now we can try out @code{count-words-defun}: install both
14672 @code{count-words-in-defun} and @code{count-words-defun}, and set the
14673 keybinding, and then place the cursor within the following definition:
14677 (defun multiply-by-seven (number)
14678 "Multiply NUMBER by seven."
14685 Success! The definition has 10 words and symbols.
14687 The next problem is to count the numbers of words and symbols in
14688 several definitions within a single file.
14690 @node Several defuns
14691 @section Count Several @code{defuns} Within a File
14693 A file such as @file{simple.el} may have a hundred or more function
14694 definitions within it. Our long term goal is to collect statistics on
14695 many files, but as a first step, our immediate goal is to collect
14696 statistics on one file.
14698 The information will be a series of numbers, each number being the
14699 length of a function definition. We can store the numbers in a list.
14701 We know that we will want to incorporate the information regarding one
14702 file with information about many other files; this means that the
14703 function for counting definition lengths within one file need only
14704 return the list of lengths. It need not and should not display any
14707 The word count commands contain one expression to jump point forward
14708 word by word and another expression to count the jumps. The function
14709 to return the lengths of definitions can be designed to work the same
14710 way, with one expression to jump point forward definition by
14711 definition and another expression to construct the lengths' list.
14713 This statement of the problem makes it elementary to write the
14714 function definition. Clearly, we will start the count at the
14715 beginning of the file, so the first command will be @code{(goto-char
14716 (point-min))}. Next, we start the @code{while} loop; and the
14717 true-or-false test of the loop can be a regular expression search for
14718 the next function definition---so long as the search succeeds, point
14719 is moved forward and then the body of the loop is evaluated. The body
14720 needs an expression that constructs the lengths' list. @code{cons},
14721 the list construction command, can be used to create the list. That
14722 is almost all there is to it.
14725 Here is what this fragment of code looks like:
14729 (goto-char (point-min))
14730 (while (re-search-forward "^(defun" nil t)
14732 (cons (count-words-in-defun) lengths-list)))
14736 What we have left out is the mechanism for finding the file that
14737 contains the function definitions.
14739 In previous examples, we either used this, the Info file, or we
14740 switched back and forth to some other buffer, such as the
14741 @file{*scratch*} buffer.
14743 Finding a file is a new process that we have not yet discussed.
14746 @section Find a File
14747 @cindex Find a File
14749 To find a file in Emacs, you use the @kbd{C-x C-f} (@code{find-file})
14750 command. This command is almost, but not quite right for the lengths
14754 Let's look at the source for @code{find-file}:
14758 (defun find-file (filename)
14759 "Edit file FILENAME.
14760 Switch to a buffer visiting file FILENAME,
14761 creating one if none already exists."
14762 (interactive "FFind file: ")
14763 (switch-to-buffer (find-file-noselect filename)))
14768 (The most recent version of the @code{find-file} function definition
14769 permits you to specify optional wildcards to visit multiple files;
14770 that makes the definition more complex and we will not discuss it
14771 here, since it is not relevant. You can see its source using either
14772 @kbd{M-.} (@code{xref-find-definitions}) or @kbd{C-h f}
14773 (@code{describe-function}).)
14777 (defun find-file (filename &optional wildcards)
14778 "Edit file FILENAME.
14779 Switch to a buffer visiting file FILENAME,
14780 creating one if none already exists.
14781 Interactively, the default if you just type @key{RET} is the current directory,
14782 but the visited file name is available through the minibuffer history:
14783 type M-n to pull it into the minibuffer.
14785 Interactively, or if WILDCARDS is non-nil in a call from Lisp,
14786 expand wildcards (if any) and visit multiple files. You can
14787 suppress wildcard expansion by setting `find-file-wildcards' to nil.
14789 To visit a file without any kind of conversion and without
14790 automatically choosing a major mode, use \\[find-file-literally]."
14791 (interactive (find-file-read-args "Find file: " nil))
14792 (let ((value (find-file-noselect filename nil nil wildcards)))
14794 (mapcar 'switch-to-buffer (nreverse value))
14795 (switch-to-buffer value))))
14798 The definition I am showing possesses short but complete documentation
14799 and an interactive specification that prompts you for a file name when
14800 you use the command interactively. The body of the definition
14801 contains two functions, @code{find-file-noselect} and
14802 @code{switch-to-buffer}.
14804 According to its documentation as shown by @kbd{C-h f} (the
14805 @code{describe-function} command), the @code{find-file-noselect}
14806 function reads the named file into a buffer and returns the buffer.
14807 (Its most recent version includes an optional @var{wildcards} argument,
14808 too, as well as another to read a file literally and an other you
14809 suppress warning messages. These optional arguments are irrelevant.)
14811 However, the @code{find-file-noselect} function does not select the
14812 buffer in which it puts the file. Emacs does not switch its attention
14813 (or yours if you are using @code{find-file-noselect}) to the selected
14814 buffer. That is what @code{switch-to-buffer} does: it switches the
14815 buffer to which Emacs attention is directed; and it switches the
14816 buffer displayed in the window to the new buffer. We have discussed
14817 buffer switching elsewhere. (@xref{Switching Buffers}.)
14819 In this histogram project, we do not need to display each file on the
14820 screen as the program determines the length of each definition within
14821 it. Instead of employing @code{switch-to-buffer}, we can work with
14822 @code{set-buffer}, which redirects the attention of the computer
14823 program to a different buffer but does not redisplay it on the screen.
14824 So instead of calling on @code{find-file} to do the job, we must write
14825 our own expression.
14827 The task is easy: use @code{find-file-noselect} and @code{set-buffer}.
14829 @node lengths-list-file
14830 @section @code{lengths-list-file} in Detail
14832 The core of the @code{lengths-list-file} function is a @code{while}
14833 loop containing a function to move point forward defun by defun, and
14834 a function to count the number of words and symbols in each defun.
14835 This core must be surrounded by functions that do various other tasks,
14836 including finding the file, and ensuring that point starts out at the
14837 beginning of the file. The function definition looks like this:
14838 @findex lengths-list-file
14842 (defun lengths-list-file (filename)
14843 "Return list of definitions' lengths within FILE.
14844 The returned list is a list of numbers.
14845 Each number is the number of words or
14846 symbols in one function definition."
14849 (message "Working on `%s' ... " filename)
14851 (let ((buffer (find-file-noselect filename))
14853 (set-buffer buffer)
14854 (setq buffer-read-only t)
14856 (goto-char (point-min))
14857 (while (re-search-forward "^(defun" nil t)
14859 (cons (count-words-in-defun) lengths-list)))
14860 (kill-buffer buffer)
14866 The function is passed one argument, the name of the file on which it
14867 will work. It has four lines of documentation, but no interactive
14868 specification. Since people worry that a computer is broken if they
14869 don't see anything going on, the first line of the body is a
14872 The next line contains a @code{save-excursion} that returns Emacs's
14873 attention to the current buffer when the function completes. This is
14874 useful in case you embed this function in another function that
14875 presumes point is restored to the original buffer.
14877 In the varlist of the @code{let} expression, Emacs finds the file and
14878 binds the local variable @code{buffer} to the buffer containing the
14879 file. At the same time, Emacs creates @code{lengths-list} as a local
14882 Next, Emacs switches its attention to the buffer.
14884 In the following line, Emacs makes the buffer read-only. Ideally,
14885 this line is not necessary. None of the functions for counting words
14886 and symbols in a function definition should change the buffer.
14887 Besides, the buffer is not going to be saved, even if it were changed.
14888 This line is entirely the consequence of great, perhaps excessive,
14889 caution. The reason for the caution is that this function and those
14890 it calls work on the sources for Emacs and it is inconvenient if they
14891 are inadvertently modified. It goes without saying that I did not
14892 realize a need for this line until an experiment went awry and started
14893 to modify my Emacs source files @dots{}
14895 Next comes a call to widen the buffer if it is narrowed. This
14896 function is usually not needed---Emacs creates a fresh buffer if none
14897 already exists; but if a buffer visiting the file already exists Emacs
14898 returns that one. In this case, the buffer may be narrowed and must
14899 be widened. If we wanted to be fully user-friendly, we would
14900 arrange to save the restriction and the location of point, but we
14903 The @code{(goto-char (point-min))} expression moves point to the
14904 beginning of the buffer.
14906 Then comes a @code{while} loop in which the work of the function is
14907 carried out. In the loop, Emacs determines the length of each
14908 definition and constructs a lengths' list containing the information.
14910 Emacs kills the buffer after working through it. This is to save
14911 space inside of Emacs. My version of GNU Emacs 19 contained over 300
14912 source files of interest; GNU Emacs 22 contains over a thousand source
14913 files. Another function will apply @code{lengths-list-file} to each
14916 Finally, the last expression within the @code{let} expression is the
14917 @code{lengths-list} variable; its value is returned as the value of
14918 the whole function.
14920 You can try this function by installing it in the usual fashion. Then
14921 place your cursor after the following expression and type @kbd{C-x
14922 C-e} (@code{eval-last-sexp}).
14924 @c !!! 22.1.1 lisp sources location here
14927 "/usr/local/share/emacs/22.1/lisp/emacs-lisp/debug.el")
14931 You may need to change the pathname of the file; the one here is for
14932 GNU Emacs version 22.1. To change the expression, copy it to
14933 the @file{*scratch*} buffer and edit it.
14937 Also, to see the full length of the list, rather than a truncated
14938 version, you may have to evaluate the following:
14939 @c We do not want to insert, so do not mention the zero prefix argument.
14942 (custom-set-variables '(eval-expression-print-length nil))
14946 (@xref{defcustom, , Specifying Variables using @code{defcustom}}.
14947 Then evaluate the @code{lengths-list-file} expression.)
14950 The lengths' list for @file{debug.el} takes less than a second to
14951 produce and looks like this in GNU Emacs 22:
14954 (83 113 105 144 289 22 30 97 48 89 25 52 52 88 28 29 77 49 43 290 232 587)
14958 (Using my old machine, the version 19 lengths' list for @file{debug.el}
14959 took seven seconds to produce and looked like this:
14962 (75 41 80 62 20 45 44 68 45 12 34 235)
14966 The newer version of @file{debug.el} contains more defuns than the
14967 earlier one; and my new machine is much faster than the old one.)
14969 Note that the length of the last definition in the file is first in
14972 @node Several files
14973 @section Count Words in @code{defuns} in Different Files
14975 In the previous section, we created a function that returns a list of
14976 the lengths of each definition in a file. Now, we want to define a
14977 function to return a master list of the lengths of the definitions in
14980 Working on each of a list of files is a repetitious act, so we can use
14981 either a @code{while} loop or recursion.
14984 * lengths-list-many-files:: Return a list of the lengths of defuns.
14985 * append:: Attach one list to another.
14989 @node lengths-list-many-files
14990 @unnumberedsubsec Determine the lengths of @code{defuns}
14993 The design using a @code{while} loop is routine. The argument passed
14994 to the function is a list of files. As we saw earlier (@pxref{Loop
14995 Example}), you can write a @code{while} loop so that the body of the
14996 loop is evaluated if such a list contains elements, but to exit the
14997 loop if the list is empty. For this design to work, the body of the
14998 loop must contain an expression that shortens the list each time the
14999 body is evaluated, so that eventually the list is empty. The usual
15000 technique is to set the value of the list to the value of the @sc{cdr}
15001 of the list each time the body is evaluated.
15004 The template looks like this:
15008 (while @var{test-whether-list-is-empty}
15010 @var{set-list-to-cdr-of-list})
15014 Also, we remember that a @code{while} loop returns @code{nil} (the
15015 result of evaluating the true-or-false-test), not the result of any
15016 evaluation within its body. (The evaluations within the body of the
15017 loop are done for their side effects.) However, the expression that
15018 sets the lengths' list is part of the body---and that is the value
15019 that we want returned by the function as a whole. To do this, we
15020 enclose the @code{while} loop within a @code{let} expression, and
15021 arrange that the last element of the @code{let} expression contains
15022 the value of the lengths' list. (@xref{Incrementing Example, , Loop
15023 Example with an Incrementing Counter}.)
15025 @findex lengths-list-many-files
15027 These considerations lead us directly to the function itself:
15031 ;;; @r{Use @code{while} loop.}
15032 (defun lengths-list-many-files (list-of-files)
15033 "Return list of lengths of defuns in LIST-OF-FILES."
15036 (let (lengths-list)
15038 ;;; @r{true-or-false-test}
15039 (while list-of-files
15044 ;;; @r{Generate a lengths' list.}
15046 (expand-file-name (car list-of-files)))))
15050 ;;; @r{Make files' list shorter.}
15051 (setq list-of-files (cdr list-of-files)))
15053 ;;; @r{Return final value of lengths' list.}
15058 @code{expand-file-name} is a built-in function that converts a file
15059 name to the absolute, long, path name form. The function employs the
15060 name of the directory in which the function is called.
15062 @c !!! 22.1.1 lisp sources location here
15064 Thus, if @code{expand-file-name} is called on @code{debug.el} when
15065 Emacs is visiting the
15066 @file{/usr/local/share/emacs/22.1.1/lisp/emacs-lisp/} directory,
15076 @c !!! 22.1.1 lisp sources location here
15078 /usr/local/share/emacs/22.1.1/lisp/emacs-lisp/debug.el
15081 The only other new element of this function definition is the as yet
15082 unstudied function @code{append}, which merits a short section for
15086 @subsection The @code{append} Function
15089 The @code{append} function attaches one list to another. Thus,
15092 (append '(1 2 3 4) '(5 6 7 8))
15103 This is exactly how we want to attach two lengths' lists produced by
15104 @code{lengths-list-file} to each other. The results contrast with
15108 (cons '(1 2 3 4) '(5 6 7 8))
15113 which constructs a new list in which the first argument to @code{cons}
15114 becomes the first element of the new list:
15117 ((1 2 3 4) 5 6 7 8)
15120 @node Several files recursively
15121 @section Recursively Count Words in Different Files
15123 Besides a @code{while} loop, you can work on each of a list of files
15124 with recursion. A recursive version of @code{lengths-list-many-files}
15125 is short and simple.
15127 The recursive function has the usual parts: the do-again-test, the
15128 next-step-expression, and the recursive call. The do-again-test
15129 determines whether the function should call itself again, which it
15130 will do if the @code{list-of-files} contains any remaining elements;
15131 the next-step-expression resets the @code{list-of-files} to the
15132 @sc{cdr} of itself, so eventually the list will be empty; and the
15133 recursive call calls itself on the shorter list. The complete
15134 function is shorter than this description!
15135 @findex recursive-lengths-list-many-files
15139 (defun recursive-lengths-list-many-files (list-of-files)
15140 "Return list of lengths of each defun in LIST-OF-FILES."
15141 (if list-of-files ; @r{do-again-test}
15144 (expand-file-name (car list-of-files)))
15145 (recursive-lengths-list-many-files
15146 (cdr list-of-files)))))
15151 In a sentence, the function returns the lengths' list for the first of
15152 the @code{list-of-files} appended to the result of calling itself on
15153 the rest of the @code{list-of-files}.
15155 Here is a test of @code{recursive-lengths-list-many-files}, along with
15156 the results of running @code{lengths-list-file} on each of the files
15159 Install @code{recursive-lengths-list-many-files} and
15160 @code{lengths-list-file}, if necessary, and then evaluate the
15161 following expressions. You may need to change the files' pathnames;
15162 those here work when this Info file and the Emacs sources are located
15163 in their customary places. To change the expressions, copy them to
15164 the @file{*scratch*} buffer, edit them, and then evaluate them.
15166 The results are shown after the @samp{@result{}}. (These results are
15167 for files from Emacs version 22.1.1; files from other versions of
15168 Emacs may produce different results.)
15170 @c !!! 22.1.1 lisp sources location here
15173 (cd "/usr/local/share/emacs/22.1.1/")
15175 (lengths-list-file "./lisp/macros.el")
15176 @result{} (283 263 480 90)
15180 (lengths-list-file "./lisp/mail/mailalias.el")
15181 @result{} (38 32 29 95 178 180 321 218 324)
15185 (lengths-list-file "./lisp/makesum.el")
15190 (recursive-lengths-list-many-files
15191 '("./lisp/macros.el"
15192 "./lisp/mail/mailalias.el"
15193 "./lisp/makesum.el"))
15194 @result{} (283 263 480 90 38 32 29 95 178 180 321 218 324 85 181)
15198 The @code{recursive-lengths-list-many-files} function produces the
15201 The next step is to prepare the data in the list for display in a graph.
15203 @node Prepare the data
15204 @section Prepare the Data for Display in a Graph
15206 The @code{recursive-lengths-list-many-files} function returns a list
15207 of numbers. Each number records the length of a function definition.
15208 What we need to do now is transform this data into a list of numbers
15209 suitable for generating a graph. The new list will tell how many
15210 functions definitions contain less than 10 words and
15211 symbols, how many contain between 10 and 19 words and symbols, how
15212 many contain between 20 and 29 words and symbols, and so on.
15214 In brief, we need to go through the lengths' list produced by the
15215 @code{recursive-lengths-list-many-files} function and count the number
15216 of defuns within each range of lengths, and produce a list of those
15220 * Data for Display in Detail::
15221 * Sorting:: Sorting lists.
15222 * Files List:: Making a list of files.
15223 * Counting function definitions::
15227 @node Data for Display in Detail
15228 @unnumberedsubsec The Data for Display in Detail
15231 Based on what we have done before, we can readily foresee that it
15232 should not be too hard to write a function that @sc{cdr}s down the
15233 lengths' list, looks at each element, determines which length range it
15234 is in, and increments a counter for that range.
15236 However, before beginning to write such a function, we should consider
15237 the advantages of sorting the lengths' list first, so the numbers are
15238 ordered from smallest to largest. First, sorting will make it easier
15239 to count the numbers in each range, since two adjacent numbers will
15240 either be in the same length range or in adjacent ranges. Second, by
15241 inspecting a sorted list, we can discover the highest and lowest
15242 number, and thereby determine the largest and smallest length range
15246 @subsection Sorting Lists
15249 Emacs contains a function to sort lists, called (as you might guess)
15250 @code{sort}. The @code{sort} function takes two arguments, the list
15251 to be sorted, and a predicate that determines whether the first of
15252 two list elements is less than the second.
15254 As we saw earlier (@pxref{Wrong Type of Argument, , Using the Wrong
15255 Type Object as an Argument}), a predicate is a function that
15256 determines whether some property is true or false. The @code{sort}
15257 function will reorder a list according to whatever property the
15258 predicate uses; this means that @code{sort} can be used to sort
15259 non-numeric lists by non-numeric criteria---it can, for example,
15260 alphabetize a list.
15263 The @code{<} function is used when sorting a numeric list. For example,
15266 (sort '(4 8 21 17 33 7 21 7) '<)
15274 (4 7 7 8 17 21 21 33)
15278 (Note that in this example, both the arguments are quoted so that the
15279 symbols are not evaluated before being passed to @code{sort} as
15282 Sorting the list returned by the
15283 @code{recursive-lengths-list-many-files} function is straightforward;
15284 it uses the @code{<} function:
15288 In GNU Emacs 22, eval
15290 (cd "/usr/local/share/emacs/22.0.50/")
15292 (recursive-lengths-list-many-files
15293 '("./lisp/macros.el"
15294 "./lisp/mail/mailalias.el"
15295 "./lisp/makesum.el"))
15303 (recursive-lengths-list-many-files
15304 '("./lisp/macros.el"
15305 "./lisp/mailalias.el"
15306 "./lisp/makesum.el"))
15316 (29 32 38 85 90 95 178 180 181 218 263 283 321 324 480)
15320 (Note that in this example, the first argument to @code{sort} is not
15321 quoted, since the expression must be evaluated so as to produce the
15322 list that is passed to @code{sort}.)
15325 @subsection Making a List of Files
15327 The @code{recursive-lengths-list-many-files} function requires a list
15328 of files as its argument. For our test examples, we constructed such
15329 a list by hand; but the Emacs Lisp source directory is too large for
15330 us to do for that. Instead, we will write a function to do the job
15331 for us. In this function, we will use both a @code{while} loop and a
15334 @findex directory-files
15335 We did not have to write a function like this for older versions of
15336 GNU Emacs, since they placed all the @samp{.el} files in one
15337 directory. Instead, we were able to use the @code{directory-files}
15338 function, which lists the names of files that match a specified
15339 pattern within a single directory.
15341 However, recent versions of Emacs place Emacs Lisp files in
15342 sub-directories of the top level @file{lisp} directory. This
15343 re-arrangement eases navigation. For example, all the mail related
15344 files are in a @file{lisp} sub-directory called @file{mail}. But at
15345 the same time, this arrangement forces us to create a file listing
15346 function that descends into the sub-directories.
15348 @findex files-in-below-directory
15349 We can create this function, called @code{files-in-below-directory},
15350 using familiar functions such as @code{car}, @code{nthcdr}, and
15351 @code{substring} in conjunction with an existing function called
15352 @code{directory-files-and-attributes}. This latter function not only
15353 lists all the filenames in a directory, including the names
15354 of sub-directories, but also their attributes.
15356 To restate our goal: to create a function that will enable us
15357 to feed filenames to @code{recursive-lengths-list-many-files}
15358 as a list that looks like this (but with more elements):
15362 ("./lisp/macros.el"
15363 "./lisp/mail/rmail.el"
15364 "./lisp/makesum.el")
15368 The @code{directory-files-and-attributes} function returns a list of
15369 lists. Each of the lists within the main list consists of 13
15370 elements. The first element is a string that contains the name of the
15371 file---which, in GNU/Linux, may be a @dfn{directory file}, that is to
15372 say, a file with the special attributes of a directory. The second
15373 element of the list is @code{t} for a directory, a string
15374 for symbolic link (the string is the name linked to), or @code{nil}.
15376 For example, the first @samp{.el} file in the @file{lisp/} directory
15377 is @file{abbrev.el}. Its name is
15378 @file{/usr/local/share/emacs/22.1.1/lisp/abbrev.el} and it is not a
15379 directory or a symbolic link.
15382 This is how @code{directory-files-and-attributes} lists that file and
15394 (20615 27034 579989 697000)
15396 (20615 26327 734791 805000)
15408 On the other hand, @file{mail/} is a directory within the @file{lisp/}
15409 directory. The beginning of its listing looks like this:
15420 (To learn about the different attributes, look at the documentation of
15421 @code{file-attributes}. Bear in mind that the @code{file-attributes}
15422 function does not list the filename, so its first element is
15423 @code{directory-files-and-attributes}'s second element.)
15425 We will want our new function, @code{files-in-below-directory}, to
15426 list the @samp{.el} files in the directory it is told to check, and in
15427 any directories below that directory.
15429 This gives us a hint on how to construct
15430 @code{files-in-below-directory}: within a directory, the function
15431 should add @samp{.el} filenames to a list; and if, within a directory,
15432 the function comes upon a sub-directory, it should go into that
15433 sub-directory and repeat its actions.
15435 However, we should note that every directory contains a name that
15436 refers to itself, called @file{.} (``dot''), and a name that refers to
15437 its parent directory, called @file{..} (``dot dot''). (In
15438 @file{/}, the root directory, @file{..} refers to itself, since
15439 @file{/} has no parent.) Clearly, we do not want our
15440 @code{files-in-below-directory} function to enter those directories,
15441 since they always lead us, directly or indirectly, to the current
15444 Consequently, our @code{files-in-below-directory} function must do
15449 Check to see whether it is looking at a filename that ends in
15450 @samp{.el}; and if so, add its name to a list.
15453 Check to see whether it is looking at a filename that is the name of a
15454 directory; and if so,
15458 Check to see whether it is looking at @file{.} or @file{..}; and if
15462 Or else, go into that directory and repeat the process.
15466 Let's write a function definition to do these tasks. We will use a
15467 @code{while} loop to move from one filename to another within a
15468 directory, checking what needs to be done; and we will use a recursive
15469 call to repeat the actions on each sub-directory. The recursive
15470 pattern is Accumulate
15471 (@pxref{Accumulate}),
15472 using @code{append} as the combiner.
15475 (directory-files "/usr/local/src/emacs/lisp/" t "\\.el$")
15476 (shell-command "find /usr/local/src/emacs/lisp/ -name '*.el'")
15478 (directory-files "/usr/local/share/emacs/22.1.1/lisp/" t "\\.el$")
15479 (shell-command "find /usr/local/share/emacs/22.1.1/lisp/ -name '*.el'")
15482 @c /usr/local/share/emacs/22.1.1/lisp/
15485 Here is the function:
15489 (defun files-in-below-directory (directory)
15490 "List the .el files in DIRECTORY and in its sub-directories."
15491 ;; Although the function will be used non-interactively,
15492 ;; it will be easier to test if we make it interactive.
15493 ;; The directory will have a name such as
15494 ;; "/usr/local/share/emacs/22.1.1/lisp/"
15495 (interactive "DDirectory name: ")
15498 (let (el-files-list
15499 (current-directory-list
15500 (directory-files-and-attributes directory t)))
15501 ;; while we are in the current directory
15502 (while current-directory-list
15506 ;; check to see whether filename ends in '.el'
15507 ;; and if so, add its name to a list.
15508 ((equal ".el" (substring (car (car current-directory-list)) -3))
15509 (setq el-files-list
15510 (cons (car (car current-directory-list)) el-files-list)))
15513 ;; check whether filename is that of a directory
15514 ((eq t (car (cdr (car current-directory-list))))
15515 ;; decide whether to skip or recurse
15518 (substring (car (car current-directory-list)) -1))
15519 ;; then do nothing since filename is that of
15520 ;; current directory or parent, "." or ".."
15524 ;; else descend into the directory and repeat the process
15525 (setq el-files-list
15527 (files-in-below-directory
15528 (car (car current-directory-list)))
15530 ;; move to the next filename in the list; this also
15531 ;; shortens the list so the while loop eventually comes to an end
15532 (setq current-directory-list (cdr current-directory-list)))
15533 ;; return the filenames
15538 @c (files-in-below-directory "/usr/local/src/emacs/lisp/")
15539 @c (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/")
15541 The @code{files-in-below-directory} @code{directory-files} function
15542 takes one argument, the name of a directory.
15545 Thus, on my system,
15547 @c (length (files-in-below-directory "/usr/local/src/emacs/lisp/"))
15549 @c !!! 22.1.1 lisp sources location here
15553 (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/"))
15558 tells me that in and below my Lisp sources directory are 1031
15561 @code{files-in-below-directory} returns a list in reverse alphabetical
15562 order. An expression to sort the list in alphabetical order looks
15568 (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/")
15575 "Test how long it takes to find lengths of all sorted elisp defuns."
15576 (insert "\n" (current-time-string) "\n")
15579 (recursive-lengths-list-many-files
15580 (files-in-below-directory "/usr/local/src/emacs/lisp/"))
15582 (insert (format "%s" (current-time-string))))
15585 @node Counting function definitions
15586 @subsection Counting function definitions
15588 Our immediate goal is to generate a list that tells us how many
15589 function definitions contain fewer than 10 words and symbols, how many
15590 contain between 10 and 19 words and symbols, how many contain between
15591 20 and 29 words and symbols, and so on.
15593 With a sorted list of numbers, this is easy: count how many elements
15594 of the list are smaller than 10, then, after moving past the numbers
15595 just counted, count how many are smaller than 20, then, after moving
15596 past the numbers just counted, count how many are smaller than 30, and
15597 so on. Each of the numbers, 10, 20, 30, 40, and the like, is one
15598 larger than the top of that range. We can call the list of such
15599 numbers the @code{top-of-ranges} list.
15602 If we wished, we could generate this list automatically, but it is
15603 simpler to write a list manually. Here it is:
15604 @vindex top-of-ranges
15608 (defvar top-of-ranges
15611 110 120 130 140 150
15612 160 170 180 190 200
15613 210 220 230 240 250
15614 260 270 280 290 300)
15615 "List specifying ranges for `defuns-per-range'.")
15619 To change the ranges, we edit this list.
15621 Next, we need to write the function that creates the list of the
15622 number of definitions within each range. Clearly, this function must
15623 take the @code{sorted-lengths} and the @code{top-of-ranges} lists
15626 The @code{defuns-per-range} function must do two things again and
15627 again: it must count the number of definitions within a range
15628 specified by the current top-of-range value; and it must shift to the
15629 next higher value in the @code{top-of-ranges} list after counting the
15630 number of definitions in the current range. Since each of these
15631 actions is repetitive, we can use @code{while} loops for the job.
15632 One loop counts the number of definitions in the range defined by the
15633 current top-of-range value, and the other loop selects each of the
15634 top-of-range values in turn.
15636 Several entries of the @code{sorted-lengths} list are counted for each
15637 range; this means that the loop for the @code{sorted-lengths} list
15638 will be inside the loop for the @code{top-of-ranges} list, like a
15639 small gear inside a big gear.
15641 The inner loop counts the number of definitions within the range. It
15642 is a simple counting loop of the type we have seen before.
15643 (@xref{Incrementing Loop, , A loop with an incrementing counter}.)
15644 The true-or-false test of the loop tests whether the value from the
15645 @code{sorted-lengths} list is smaller than the current value of the
15646 top of the range. If it is, the function increments the counter and
15647 tests the next value from the @code{sorted-lengths} list.
15650 The inner loop looks like this:
15654 (while @var{length-element-smaller-than-top-of-range}
15655 (setq number-within-range (1+ number-within-range))
15656 (setq sorted-lengths (cdr sorted-lengths)))
15660 The outer loop must start with the lowest value of the
15661 @code{top-of-ranges} list, and then be set to each of the succeeding
15662 higher values in turn. This can be done with a loop like this:
15666 (while top-of-ranges
15667 @var{body-of-loop}@dots{}
15668 (setq top-of-ranges (cdr top-of-ranges)))
15673 Put together, the two loops look like this:
15677 (while top-of-ranges
15679 ;; @r{Count the number of elements within the current range.}
15680 (while @var{length-element-smaller-than-top-of-range}
15681 (setq number-within-range (1+ number-within-range))
15682 (setq sorted-lengths (cdr sorted-lengths)))
15684 ;; @r{Move to next range.}
15685 (setq top-of-ranges (cdr top-of-ranges)))
15689 In addition, in each circuit of the outer loop, Emacs should record
15690 the number of definitions within that range (the value of
15691 @code{number-within-range}) in a list. We can use @code{cons} for
15692 this purpose. (@xref{cons, , @code{cons}}.)
15694 The @code{cons} function works fine, except that the list it
15695 constructs will contain the number of definitions for the highest
15696 range at its beginning and the number of definitions for the lowest
15697 range at its end. This is because @code{cons} attaches new elements
15698 of the list to the beginning of the list, and since the two loops are
15699 working their way through the lengths' list from the lower end first,
15700 the @code{defuns-per-range-list} will end up largest number first.
15701 But we will want to print our graph with smallest values first and the
15702 larger later. The solution is to reverse the order of the
15703 @code{defuns-per-range-list}. We can do this using the
15704 @code{nreverse} function, which reverses the order of a list.
15711 (nreverse '(1 2 3 4))
15722 Note that the @code{nreverse} function is destructive---that is,
15723 it changes the list to which it is applied; this contrasts with the
15724 @code{car} and @code{cdr} functions, which are non-destructive. In
15725 this case, we do not want the original @code{defuns-per-range-list},
15726 so it does not matter that it is destroyed. (The @code{reverse}
15727 function provides a reversed copy of a list, leaving the original list
15732 Put all together, the @code{defuns-per-range} looks like this:
15736 (defun defuns-per-range (sorted-lengths top-of-ranges)
15737 "SORTED-LENGTHS defuns in each TOP-OF-RANGES range."
15738 (let ((top-of-range (car top-of-ranges))
15739 (number-within-range 0)
15740 defuns-per-range-list)
15745 (while top-of-ranges
15751 ;; @r{Need number for numeric test.}
15752 (car sorted-lengths)
15753 (< (car sorted-lengths) top-of-range))
15757 ;; @r{Count number of definitions within current range.}
15758 (setq number-within-range (1+ number-within-range))
15759 (setq sorted-lengths (cdr sorted-lengths)))
15761 ;; @r{Exit inner loop but remain within outer loop.}
15765 (setq defuns-per-range-list
15766 (cons number-within-range defuns-per-range-list))
15767 (setq number-within-range 0) ; @r{Reset count to zero.}
15771 ;; @r{Move to next range.}
15772 (setq top-of-ranges (cdr top-of-ranges))
15773 ;; @r{Specify next top of range value.}
15774 (setq top-of-range (car top-of-ranges)))
15778 ;; @r{Exit outer loop and count the number of defuns larger than}
15779 ;; @r{ the largest top-of-range value.}
15780 (setq defuns-per-range-list
15782 (length sorted-lengths)
15783 defuns-per-range-list))
15787 ;; @r{Return a list of the number of definitions within each range,}
15788 ;; @r{ smallest to largest.}
15789 (nreverse defuns-per-range-list)))
15795 The function is straightforward except for one subtle feature. The
15796 true-or-false test of the inner loop looks like this:
15800 (and (car sorted-lengths)
15801 (< (car sorted-lengths) top-of-range))
15807 instead of like this:
15810 (< (car sorted-lengths) top-of-range)
15813 The purpose of the test is to determine whether the first item in the
15814 @code{sorted-lengths} list is less than the value of the top of the
15817 The simple version of the test works fine unless the
15818 @code{sorted-lengths} list has a @code{nil} value. In that case, the
15819 @code{(car sorted-lengths)} expression function returns
15820 @code{nil}. The @code{<} function cannot compare a number to
15821 @code{nil}, which is an empty list, so Emacs signals an error and
15822 stops the function from attempting to continue to execute.
15824 The @code{sorted-lengths} list always becomes @code{nil} when the
15825 counter reaches the end of the list. This means that any attempt to
15826 use the @code{defuns-per-range} function with the simple version of
15827 the test will fail.
15829 We solve the problem by using the @code{(car sorted-lengths)}
15830 expression in conjunction with the @code{and} expression. The
15831 @code{(car sorted-lengths)} expression returns a non-@code{nil}
15832 value so long as the list has at least one number within it, but
15833 returns @code{nil} if the list is empty. The @code{and} expression
15834 first evaluates the @code{(car sorted-lengths)} expression, and
15835 if it is @code{nil}, returns false @emph{without} evaluating the
15836 @code{<} expression. But if the @code{(car sorted-lengths)}
15837 expression returns a non-@code{nil} value, the @code{and} expression
15838 evaluates the @code{<} expression, and returns that value as the value
15839 of the @code{and} expression.
15841 @c colon in printed section title causes problem in Info cross reference
15842 This way, we avoid an error.
15845 (For information about @code{and}, see
15846 @ref{kill-new function, , The @code{kill-new} function}.)
15850 (@xref{kill-new function, , The @code{kill-new} function}, for
15851 information about @code{and}.)
15854 Here is a short test of the @code{defuns-per-range} function. First,
15855 evaluate the expression that binds (a shortened)
15856 @code{top-of-ranges} list to the list of values, then evaluate the
15857 expression for binding the @code{sorted-lengths} list, and then
15858 evaluate the @code{defuns-per-range} function.
15862 ;; @r{(Shorter list than we will use later.)}
15863 (setq top-of-ranges
15864 '(110 120 130 140 150
15865 160 170 180 190 200))
15867 (setq sorted-lengths
15868 '(85 86 110 116 122 129 154 176 179 200 265 300 300))
15870 (defuns-per-range sorted-lengths top-of-ranges)
15876 The list returned looks like this:
15879 (2 2 2 0 0 1 0 2 0 0 4)
15883 Indeed, there are two elements of the @code{sorted-lengths} list
15884 smaller than 110, two elements between 110 and 119, two elements
15885 between 120 and 129, and so on. There are four elements with a value
15888 @c The next step is to turn this numbers' list into a graph.
15889 @node Readying a Graph
15890 @chapter Readying a Graph
15891 @cindex Readying a graph
15892 @cindex Graph prototype
15893 @cindex Prototype graph
15894 @cindex Body of graph
15896 Our goal is to construct a graph showing the numbers of function
15897 definitions of various lengths in the Emacs lisp sources.
15899 As a practical matter, if you were creating a graph, you would
15900 probably use a program such as @code{gnuplot} to do the job.
15901 (@code{gnuplot} is nicely integrated into GNU Emacs.) In this case,
15902 however, we create one from scratch, and in the process we will
15903 re-acquaint ourselves with some of what we learned before and learn
15906 In this chapter, we will first write a simple graph printing function.
15907 This first definition will be a @dfn{prototype}, a rapidly written
15908 function that enables us to reconnoiter this unknown graph-making
15909 territory. We will discover dragons, or find that they are myth.
15910 After scouting the terrain, we will feel more confident and enhance
15911 the function to label the axes automatically.
15914 * Columns of a graph::
15915 * graph-body-print:: How to print the body of a graph.
15916 * recursive-graph-body-print::
15918 * Line Graph Exercise::
15922 @node Columns of a graph
15923 @unnumberedsec Printing the Columns of a Graph
15926 Since Emacs is designed to be flexible and work with all kinds of
15927 terminals, including character-only terminals, the graph will need to
15928 be made from one of the typewriter symbols. An asterisk will do; as
15929 we enhance the graph-printing function, we can make the choice of
15930 symbol a user option.
15932 We can call this function @code{graph-body-print}; it will take a
15933 @code{numbers-list} as its only argument. At this stage, we will not
15934 label the graph, but only print its body.
15936 The @code{graph-body-print} function inserts a vertical column of
15937 asterisks for each element in the @code{numbers-list}. The height of
15938 each line is determined by the value of that element of the
15939 @code{numbers-list}.
15941 Inserting columns is a repetitive act; that means that this function can
15942 be written either with a @code{while} loop or recursively.
15944 Our first challenge is to discover how to print a column of asterisks.
15945 Usually, in Emacs, we print characters onto a screen horizontally,
15946 line by line, by typing. We have two routes we can follow: write our
15947 own column-insertion function or discover whether one exists in Emacs.
15949 To see whether there is one in Emacs, we can use the @kbd{M-x apropos}
15950 command. This command is like the @kbd{C-h a} (@code{command-apropos})
15951 command, except that the latter finds only those functions that are
15952 commands. The @kbd{M-x apropos} command lists all symbols that match
15953 a regular expression, including functions that are not interactive.
15956 What we want to look for is some command that prints or inserts
15957 columns. Very likely, the name of the function will contain either
15958 the word ``print'' or the word ``insert'' or the word ``column''.
15959 Therefore, we can simply type @kbd{M-x apropos @key{RET}
15960 print\|insert\|column @key{RET}} and look at the result. On my system, this
15961 command once took quite some time, and then produced a list of 79
15962 functions and variables. Now it does not take much time at all and
15963 produces a list of 211 functions and variables. Scanning down the
15964 list, the only function that looks as if it might do the job is
15965 @code{insert-rectangle}.
15968 Indeed, this is the function we want; its documentation says:
15973 Insert text of RECTANGLE with upper left corner at point.
15974 RECTANGLE's first line is inserted at point,
15975 its second line is inserted at a point vertically under point, etc.
15976 RECTANGLE should be a list of strings.
15977 After this command, the mark is at the upper left corner
15978 and point is at the lower right corner.
15982 We can run a quick test, to make sure it does what we expect of it.
15984 Here is the result of placing the cursor after the
15985 @code{insert-rectangle} expression and typing @kbd{C-u C-x C-e}
15986 (@code{eval-last-sexp}). The function inserts the strings
15987 @samp{"first"}, @samp{"second"}, and @samp{"third"} at and below
15988 point. Also the function returns @code{nil}.
15992 (insert-rectangle '("first" "second" "third"))first
15999 Of course, we won't be inserting the text of the
16000 @code{insert-rectangle} expression itself into the buffer in which we
16001 are making the graph, but will call the function from our program. We
16002 shall, however, have to make sure that point is in the buffer at the
16003 place where the @code{insert-rectangle} function will insert its
16006 If you are reading this in Info, you can see how this works by
16007 switching to another buffer, such as the @file{*scratch*} buffer,
16008 placing point somewhere in the buffer, typing @kbd{M-:}, typing the
16009 @code{insert-rectangle} expression into the minibuffer at the prompt,
16010 and then typing @key{RET}. This causes Emacs to evaluate the
16011 expression in the minibuffer, but to use as the value of point the
16012 position of point in the @file{*scratch*} buffer. (@kbd{M-:} is the
16013 keybinding for @code{eval-expression}. Also, @code{nil} does not
16014 appear in the @file{*scratch*} buffer since the expression is
16015 evaluated in the minibuffer.)
16017 We find when we do this that point ends up at the end of the last
16018 inserted line---that is to say, this function moves point as a
16019 side-effect. If we were to repeat the command, with point at this
16020 position, the next insertion would be below and to the right of the
16021 previous insertion. We don't want this! If we are going to make a
16022 bar graph, the columns need to be beside each other.
16024 So we discover that each cycle of the column-inserting @code{while}
16025 loop must reposition point to the place we want it, and that place
16026 will be at the top, not the bottom, of the column. Moreover, we
16027 remember that when we print a graph, we do not expect all the columns
16028 to be the same height. This means that the top of each column may be
16029 at a different height from the previous one. We cannot simply
16030 reposition point to the same line each time, but moved over to the
16031 right---or perhaps we can@dots{}
16033 We are planning to make the columns of the bar graph out of asterisks.
16034 The number of asterisks in the column is the number specified by the
16035 current element of the @code{numbers-list}. We need to construct a
16036 list of asterisks of the right length for each call to
16037 @code{insert-rectangle}. If this list consists solely of the requisite
16038 number of asterisks, then we will have to position point the right number
16039 of lines above the base for the graph to print correctly. This could
16042 Alternatively, if we can figure out some way to pass
16043 @code{insert-rectangle} a list of the same length each time, then we
16044 can place point on the same line each time, but move it over one
16045 column to the right for each new column. If we do this, however, some
16046 of the entries in the list passed to @code{insert-rectangle} must be
16047 blanks rather than asterisks. For example, if the maximum height of
16048 the graph is 5, but the height of the column is 3, then
16049 @code{insert-rectangle} requires an argument that looks like this:
16052 (" " " " "*" "*" "*")
16055 This last proposal is not so difficult, so long as we can determine
16056 the column height. There are two ways for us to specify the column
16057 height: we can arbitrarily state what it will be, which would work
16058 fine for graphs of that height; or we can search through the list of
16059 numbers and use the maximum height of the list as the maximum height
16060 of the graph. If the latter operation were difficult, then the former
16061 procedure would be easiest, but there is a function built into Emacs
16062 that determines the maximum of its arguments. We can use that
16063 function. The function is called @code{max} and it returns the
16064 largest of all its arguments, which must be numbers. Thus, for
16072 returns 7. (A corresponding function called @code{min} returns the
16073 smallest of all its arguments.)
16077 However, we cannot simply call @code{max} on the @code{numbers-list};
16078 the @code{max} function expects numbers as its argument, not a list of
16079 numbers. Thus, the following expression,
16082 (max '(3 4 6 5 7 3))
16087 produces the following error message;
16090 Wrong type of argument: number-or-marker-p, (3 4 6 5 7 3)
16094 We need a function that passes a list of arguments to a function.
16095 This function is @code{apply}. This function applies its first
16096 argument (a function) to its remaining arguments, the last of which
16103 (apply 'max 3 4 7 3 '(4 8 5))
16109 (Incidentally, I don't know how you would learn of this function
16110 without a book such as this. It is possible to discover other
16111 functions, like @code{search-forward} or @code{insert-rectangle}, by
16112 guessing at a part of their names and then using @code{apropos}. Even
16113 though its base in metaphor is clear---apply its first argument to
16114 the rest---I doubt a novice would come up with that particular word
16115 when using @code{apropos} or other aid. Of course, I could be wrong;
16116 after all, the function was first named by someone who had to invent
16119 The second and subsequent arguments to @code{apply} are optional, so
16120 we can use @code{apply} to call a function and pass the elements of a
16121 list to it, like this, which also returns 8:
16124 (apply 'max '(4 8 5))
16127 This latter way is how we will use @code{apply}. The
16128 @code{recursive-lengths-list-many-files} function returns a numbers'
16129 list to which we can apply @code{max} (we could also apply @code{max} to
16130 the sorted numbers' list; it does not matter whether the list is
16134 Hence, the operation for finding the maximum height of the graph is this:
16137 (setq max-graph-height (apply 'max numbers-list))
16140 Now we can return to the question of how to create a list of strings
16141 for a column of the graph. Told the maximum height of the graph
16142 and the number of asterisks that should appear in the column, the
16143 function should return a list of strings for the
16144 @code{insert-rectangle} command to insert.
16146 Each column is made up of asterisks or blanks. Since the function is
16147 passed the value of the height of the column and the number of
16148 asterisks in the column, the number of blanks can be found by
16149 subtracting the number of asterisks from the height of the column.
16150 Given the number of blanks and the number of asterisks, two
16151 @code{while} loops can be used to construct the list:
16155 ;;; @r{First version.}
16156 (defun column-of-graph (max-graph-height actual-height)
16157 "Return list of strings that is one column of a graph."
16158 (let ((insert-list nil)
16159 (number-of-top-blanks
16160 (- max-graph-height actual-height)))
16164 ;; @r{Fill in asterisks.}
16165 (while (> actual-height 0)
16166 (setq insert-list (cons "*" insert-list))
16167 (setq actual-height (1- actual-height)))
16171 ;; @r{Fill in blanks.}
16172 (while (> number-of-top-blanks 0)
16173 (setq insert-list (cons " " insert-list))
16174 (setq number-of-top-blanks
16175 (1- number-of-top-blanks)))
16179 ;; @r{Return whole list.}
16184 If you install this function and then evaluate the following
16185 expression you will see that it returns the list as desired:
16188 (column-of-graph 5 3)
16196 (" " " " "*" "*" "*")
16199 As written, @code{column-of-graph} contains a major flaw: the symbols
16200 used for the blank and for the marked entries in the column are
16201 hard-coded as a space and asterisk. This is fine for a prototype,
16202 but you, or another user, may wish to use other symbols. For example,
16203 in testing the graph function, you may want to use a period in place
16204 of the space, to make sure the point is being repositioned properly
16205 each time the @code{insert-rectangle} function is called; or you might
16206 want to substitute a @samp{+} sign or other symbol for the asterisk.
16207 You might even want to make a graph-column that is more than one
16208 display column wide. The program should be more flexible. The way to
16209 do that is to replace the blank and the asterisk with two variables
16210 that we can call @code{graph-blank} and @code{graph-symbol} and define
16211 those variables separately.
16213 Also, the documentation is not well written. These considerations
16214 lead us to the second version of the function:
16218 (defvar graph-symbol "*"
16219 "String used as symbol in graph, usually an asterisk.")
16223 (defvar graph-blank " "
16224 "String used as blank in graph, usually a blank space.
16225 graph-blank must be the same number of columns wide
16231 (For an explanation of @code{defvar}, see
16232 @ref{defvar, , Initializing a Variable with @code{defvar}}.)
16236 ;;; @r{Second version.}
16237 (defun column-of-graph (max-graph-height actual-height)
16238 "Return MAX-GRAPH-HEIGHT strings; ACTUAL-HEIGHT are graph-symbols.
16242 The graph-symbols are contiguous entries at the end
16244 The list will be inserted as one column of a graph.
16245 The strings are either graph-blank or graph-symbol."
16249 (let ((insert-list nil)
16250 (number-of-top-blanks
16251 (- max-graph-height actual-height)))
16255 ;; @r{Fill in @code{graph-symbols}.}
16256 (while (> actual-height 0)
16257 (setq insert-list (cons graph-symbol insert-list))
16258 (setq actual-height (1- actual-height)))
16262 ;; @r{Fill in @code{graph-blanks}.}
16263 (while (> number-of-top-blanks 0)
16264 (setq insert-list (cons graph-blank insert-list))
16265 (setq number-of-top-blanks
16266 (1- number-of-top-blanks)))
16268 ;; @r{Return whole list.}
16273 If we wished, we could rewrite @code{column-of-graph} a third time to
16274 provide optionally for a line graph as well as for a bar graph. This
16275 would not be hard to do. One way to think of a line graph is that it
16276 is no more than a bar graph in which the part of each bar that is
16277 below the top is blank. To construct a column for a line graph, the
16278 function first constructs a list of blanks that is one shorter than
16279 the value, then it uses @code{cons} to attach a graph symbol to the
16280 list; then it uses @code{cons} again to attach the top blanks to
16283 It is easy to see how to write such a function, but since we don't
16284 need it, we will not do it. But the job could be done, and if it were
16285 done, it would be done with @code{column-of-graph}. Even more
16286 important, it is worth noting that few changes would have to be made
16287 anywhere else. The enhancement, if we ever wish to make it, is
16290 Now, finally, we come to our first actual graph printing function.
16291 This prints the body of a graph, not the labels for the vertical and
16292 horizontal axes, so we can call this @code{graph-body-print}.
16294 @node graph-body-print
16295 @section The @code{graph-body-print} Function
16296 @findex graph-body-print
16298 After our preparation in the preceding section, the
16299 @code{graph-body-print} function is straightforward. The function
16300 will print column after column of asterisks and blanks, using the
16301 elements of a numbers' list to specify the number of asterisks in each
16302 column. This is a repetitive act, which means we can use a
16303 decrementing @code{while} loop or recursive function for the job. In
16304 this section, we will write the definition using a @code{while} loop.
16306 The @code{column-of-graph} function requires the height of the graph
16307 as an argument, so we should determine and record that as a local variable.
16309 This leads us to the following template for the @code{while} loop
16310 version of this function:
16314 (defun graph-body-print (numbers-list)
16315 "@var{documentation}@dots{}"
16316 (let ((height @dots{}
16321 (while numbers-list
16322 @var{insert-columns-and-reposition-point}
16323 (setq numbers-list (cdr numbers-list)))))
16328 We need to fill in the slots of the template.
16330 Clearly, we can use the @code{(apply 'max numbers-list)} expression to
16331 determine the height of the graph.
16333 The @code{while} loop will cycle through the @code{numbers-list} one
16334 element at a time. As it is shortened by the @code{(setq numbers-list
16335 (cdr numbers-list))} expression, the @sc{car} of each instance of the
16336 list is the value of the argument for @code{column-of-graph}.
16338 At each cycle of the @code{while} loop, the @code{insert-rectangle}
16339 function inserts the list returned by @code{column-of-graph}. Since
16340 the @code{insert-rectangle} function moves point to the lower right of
16341 the inserted rectangle, we need to save the location of point at the
16342 time the rectangle is inserted, move back to that position after the
16343 rectangle is inserted, and then move horizontally to the next place
16344 from which @code{insert-rectangle} is called.
16346 If the inserted columns are one character wide, as they will be if
16347 single blanks and asterisks are used, the repositioning command is
16348 simply @code{(forward-char 1)}; however, the width of a column may be
16349 greater than one. This means that the repositioning command should be
16350 written @code{(forward-char symbol-width)}. The @code{symbol-width}
16351 itself is the length of a @code{graph-blank} and can be found using
16352 the expression @code{(length graph-blank)}. The best place to bind
16353 the @code{symbol-width} variable to the value of the width of graph
16354 column is in the varlist of the @code{let} expression.
16357 These considerations lead to the following function definition:
16361 (defun graph-body-print (numbers-list)
16362 "Print a bar graph of the NUMBERS-LIST.
16363 The numbers-list consists of the Y-axis values."
16365 (let ((height (apply 'max numbers-list))
16366 (symbol-width (length graph-blank))
16371 (while numbers-list
16372 (setq from-position (point))
16374 (column-of-graph height (car numbers-list)))
16375 (goto-char from-position)
16376 (forward-char symbol-width)
16379 ;; @r{Draw graph column by column.}
16381 (setq numbers-list (cdr numbers-list)))
16384 ;; @r{Place point for X axis labels.}
16385 (forward-line height)
16392 The one unexpected expression in this function is the
16393 @w{@code{(sit-for 0)}} expression in the @code{while} loop. This
16394 expression makes the graph printing operation more interesting to
16395 watch than it would be otherwise. The expression causes Emacs to
16396 @dfn{sit} or do nothing for a zero length of time and then redraw the
16397 screen. Placed here, it causes Emacs to redraw the screen column by
16398 column. Without it, Emacs would not redraw the screen until the
16401 We can test @code{graph-body-print} with a short list of numbers.
16405 Install @code{graph-symbol}, @code{graph-blank},
16406 @code{column-of-graph}, which are in
16408 @ref{Readying a Graph, , Readying a Graph},
16411 @ref{Columns of a graph},
16413 and @code{graph-body-print}.
16417 Copy the following expression:
16420 (graph-body-print '(1 2 3 4 6 4 3 5 7 6 5 2 3))
16424 Switch to the @file{*scratch*} buffer and place the cursor where you
16425 want the graph to start.
16428 Type @kbd{M-:} (@code{eval-expression}).
16431 Yank the @code{graph-body-print} expression into the minibuffer
16432 with @kbd{C-y} (@code{yank)}.
16435 Press @key{RET} to evaluate the @code{graph-body-print} expression.
16439 Emacs will print a graph like this:
16453 @node recursive-graph-body-print
16454 @section The @code{recursive-graph-body-print} Function
16455 @findex recursive-graph-body-print
16457 The @code{graph-body-print} function may also be written recursively.
16458 The recursive solution is divided into two parts: an outside wrapper
16459 that uses a @code{let} expression to determine the values of several
16460 variables that need only be found once, such as the maximum height of
16461 the graph, and an inside function that is called recursively to print
16465 The wrapper is uncomplicated:
16469 (defun recursive-graph-body-print (numbers-list)
16470 "Print a bar graph of the NUMBERS-LIST.
16471 The numbers-list consists of the Y-axis values."
16472 (let ((height (apply 'max numbers-list))
16473 (symbol-width (length graph-blank))
16475 (recursive-graph-body-print-internal
16482 The recursive function is a little more difficult. It has four parts:
16483 the do-again-test, the printing code, the recursive call, and the
16484 next-step-expression. The do-again-test is a @code{when}
16485 expression that determines whether the @code{numbers-list} contains
16486 any remaining elements; if it does, the function prints one column of
16487 the graph using the printing code and calls itself again. The
16488 function calls itself again according to the value produced by the
16489 next-step-expression which causes the call to act on a shorter
16490 version of the @code{numbers-list}.
16494 (defun recursive-graph-body-print-internal
16495 (numbers-list height symbol-width)
16496 "Print a bar graph.
16497 Used within recursive-graph-body-print function."
16502 (setq from-position (point))
16504 (column-of-graph height (car numbers-list)))
16507 (goto-char from-position)
16508 (forward-char symbol-width)
16509 (sit-for 0) ; @r{Draw graph column by column.}
16510 (recursive-graph-body-print-internal
16511 (cdr numbers-list) height symbol-width)))
16516 After installation, this expression can be tested; here is a sample:
16519 (recursive-graph-body-print '(3 2 5 6 7 5 3 4 6 4 3 2 1))
16523 Here is what @code{recursive-graph-body-print} produces:
16537 Either of these two functions, @code{graph-body-print} or
16538 @code{recursive-graph-body-print}, create the body of a graph.
16541 @section Need for Printed Axes
16543 A graph needs printed axes, so you can orient yourself. For a do-once
16544 project, it may be reasonable to draw the axes by hand using Emacs's
16545 Picture mode; but a graph drawing function may be used more than once.
16547 For this reason, I have written enhancements to the basic
16548 @code{print-graph-body} function that automatically print labels for
16549 the horizontal and vertical axes. Since the label printing functions
16550 do not contain much new material, I have placed their description in
16551 an appendix. @xref{Full Graph, , A Graph with Labeled Axes}.
16553 @node Line Graph Exercise
16556 Write a line graph version of the graph printing functions.
16558 @node Emacs Initialization
16559 @chapter Your @file{.emacs} File
16560 @cindex @file{.emacs} file
16561 @cindex Customizing your @file{.emacs} file
16562 @cindex Initialization file
16564 ``You don't have to like Emacs to like it''---this seemingly
16565 paradoxical statement is the secret of GNU Emacs. The plain, out-of-the-box
16566 Emacs is a generic tool. Most people who use it customize
16567 it to suit themselves.
16569 GNU Emacs is mostly written in Emacs Lisp; this means that by writing
16570 expressions in Emacs Lisp you can change or extend Emacs.
16573 * Default Configuration::
16574 * Site-wide Init:: You can write site-wide init files.
16575 * defcustom:: Emacs will write code for you.
16576 * Beginning init File:: How to write a @file{.emacs} init file.
16577 * Text and Auto-fill:: Automatically wrap lines.
16578 * Mail Aliases:: Use abbreviations for email addresses.
16579 * Indent Tabs Mode:: Don't use tabs with @TeX{}
16580 * Keybindings:: Create some personal keybindings.
16581 * Keymaps:: More about key binding.
16582 * Loading Files:: Load (i.e., evaluate) files automatically.
16583 * Autoload:: Make functions available.
16584 * Simple Extension:: Define a function; bind it to a key.
16585 * X11 Colors:: Colors in X.
16587 * Mode Line:: How to customize your mode line.
16591 @node Default Configuration
16592 @unnumberedsec Emacs's Default Configuration
16595 There are those who appreciate Emacs's default configuration. After
16596 all, Emacs starts you in C mode when you edit a C file, starts you in
16597 Fortran mode when you edit a Fortran file, and starts you in
16598 Fundamental mode when you edit an unadorned file. This all makes
16599 sense, if you do not know who is going to use Emacs. Who knows what a
16600 person hopes to do with an unadorned file? Fundamental mode is the
16601 right default for such a file, just as C mode is the right default for
16602 editing C code. (Enough programming languages have syntaxes
16603 that enable them to share or nearly share features, so C mode is
16604 now provided by CC mode, the C Collection.)
16606 But when you do know who is going to use Emacs---you,
16607 yourself---then it makes sense to customize Emacs.
16609 For example, I seldom want Fundamental mode when I edit an
16610 otherwise undistinguished file; I want Text mode. This is why I
16611 customize Emacs: so it suits me.
16613 You can customize and extend Emacs by writing or adapting a
16614 @file{~/.emacs} file. This is your personal initialization file; its
16615 contents, written in Emacs Lisp, tell Emacs what to do.@footnote{You
16616 may also add @file{.el} to @file{~/.emacs} and call it a
16617 @file{~/.emacs.el} file. In the past, you were forbidden to type the
16618 extra keystrokes that the name @file{~/.emacs.el} requires, but now
16619 you may. The new format is consistent with the Emacs Lisp file
16620 naming conventions; the old format saves typing.}
16622 A @file{~/.emacs} file contains Emacs Lisp code. You can write this
16623 code yourself; or you can use Emacs's @code{customize} feature to write
16624 the code for you. You can combine your own expressions and
16625 auto-written Customize expressions in your @file{.emacs} file.
16627 (I myself prefer to write my own expressions, except for those,
16628 particularly fonts, that I find easier to manipulate using the
16629 @code{customize} command. I combine the two methods.)
16631 Most of this chapter is about writing expressions yourself. It
16632 describes a simple @file{.emacs} file; for more information, see
16633 @ref{Init File, , The Init File, emacs, The GNU Emacs Manual}, and
16634 @ref{Init File, , The Init File, elisp, The GNU Emacs Lisp Reference
16637 @node Site-wide Init
16638 @section Site-wide Initialization Files
16640 @cindex @file{default.el} init file
16641 @cindex @file{site-init.el} init file
16642 @cindex @file{site-load.el} init file
16643 In addition to your personal initialization file, Emacs automatically
16644 loads various site-wide initialization files, if they exist. These
16645 have the same form as your @file{.emacs} file, but are loaded by
16648 Two site-wide initialization files, @file{site-load.el} and
16649 @file{site-init.el}, are loaded into Emacs and then dumped if a
16650 dumped version of Emacs is created, as is most common. (Dumped
16651 copies of Emacs load more quickly. However, once a file is loaded and
16652 dumped, a change to it does not lead to a change in Emacs unless you
16653 load it yourself or re-dump Emacs. @xref{Building Emacs, , Building
16654 Emacs, elisp, The GNU Emacs Lisp Reference Manual}, and the
16655 @file{INSTALL} file.)
16657 Three other site-wide initialization files are loaded automatically
16658 each time you start Emacs, if they exist. These are
16659 @file{site-start.el}, which is loaded @emph{before} your @file{.emacs}
16660 file, and @file{default.el}, and the terminal type file, which are both
16661 loaded @emph{after} your @file{.emacs} file.
16663 Settings and definitions in your @file{.emacs} file will overwrite
16664 conflicting settings and definitions in a @file{site-start.el} file,
16665 if it exists; but the settings and definitions in a @file{default.el}
16666 or terminal type file will overwrite those in your @file{.emacs} file.
16667 (You can prevent interference from a terminal type file by setting
16668 @code{term-file-prefix} to @code{nil}. @xref{Simple Extension, , A
16669 Simple Extension}.)
16671 @c Rewritten to avoid overfull hbox.
16672 The @file{INSTALL} file that comes in the distribution contains
16673 descriptions of the @file{site-init.el} and @file{site-load.el} files.
16675 The @file{loadup.el}, @file{startup.el}, and @file{loaddefs.el} files
16676 control loading. These files are in the @file{lisp} directory of the
16677 Emacs distribution and are worth perusing.
16679 The @file{loaddefs.el} file contains a good many suggestions as to
16680 what to put into your own @file{.emacs} file, or into a site-wide
16681 initialization file.
16684 @section Specifying Variables using @code{defcustom}
16687 You can specify variables using @code{defcustom} so that you and
16688 others can then use Emacs's @code{customize} feature to set their
16689 values. (You cannot use @code{customize} to write function
16690 definitions; but you can write @code{defuns} in your @file{.emacs}
16691 file. Indeed, you can write any Lisp expression in your @file{.emacs}
16694 The @code{customize} feature depends on the @code{defcustom} macro.
16695 Although you can use @code{defvar} or @code{setq} for variables that
16696 users set, the @code{defcustom} macro is designed for the job.
16698 You can use your knowledge of @code{defvar} for writing the
16699 first three arguments for @code{defcustom}. The first argument to
16700 @code{defcustom} is the name of the variable. The second argument is
16701 the variable's initial value, if any; and this value is set only if
16702 the value has not already been set. The third argument is the
16705 The fourth and subsequent arguments to @code{defcustom} specify types
16706 and options; these are not featured in @code{defvar}. (These
16707 arguments are optional.)
16709 Each of these arguments consists of a keyword followed by a value.
16710 Each keyword starts with the colon character @samp{:}.
16713 For example, the customizable user option variable
16714 @code{text-mode-hook} looks like this:
16718 (defcustom text-mode-hook nil
16719 "Normal hook run when entering Text mode and many related modes."
16721 :options '(turn-on-auto-fill flyspell-mode)
16727 The name of the variable is @code{text-mode-hook}; it has no default
16728 value; and its documentation string tells you what it does.
16730 The @code{:type} keyword tells Emacs the kind of data to which
16731 @code{text-mode-hook} should be set and how to display the value in a
16732 Customization buffer.
16734 The @code{:options} keyword specifies a suggested list of values for
16735 the variable. Usually, @code{:options} applies to a hook.
16736 The list is only a suggestion; it is not exclusive; a person who sets
16737 the variable may set it to other values; the list shown following the
16738 @code{:options} keyword is intended to offer convenient choices to a
16741 Finally, the @code{:group} keyword tells the Emacs Customization
16742 command in which group the variable is located. This tells where to
16745 The @code{defcustom} macro recognizes more than a dozen keywords.
16746 For more information, see @ref{Customization, , Writing Customization
16747 Definitions, elisp, The GNU Emacs Lisp Reference Manual}.
16749 Consider @code{text-mode-hook} as an example.
16751 There are two ways to customize this variable. You can use the
16752 customization command or write the appropriate expressions yourself.
16755 Using the customization command, you can type:
16762 and find that the group for editing files of text is called ``Text''.
16763 Enter that group. Text Mode Hook is the first member. You can click
16764 on its various options, such as @code{turn-on-auto-fill}, to set the
16765 values. After you click on the button to
16768 Save for Future Sessions
16772 Emacs will write an expression into your @file{.emacs} file.
16773 It will look like this:
16777 (custom-set-variables
16778 ;; custom-set-variables was added by Custom.
16779 ;; If you edit it by hand, you could mess it up, so be careful.
16780 ;; Your init file should contain only one such instance.
16781 ;; If there is more than one, they won't work right.
16782 '(text-mode-hook (quote (turn-on-auto-fill text-mode-hook-identify))))
16787 (The @code{text-mode-hook-identify} function tells
16788 @code{toggle-text-mode-auto-fill} which buffers are in Text mode.
16789 It comes on automatically.)
16791 The @code{custom-set-variables} function works somewhat differently
16792 than a @code{setq}. While I have never learned the differences, I
16793 modify the @code{custom-set-variables} expressions in my @file{.emacs}
16794 file by hand: I make the changes in what appears to me to be a
16795 reasonable manner and have not had any problems. Others prefer to use
16796 the Customization command and let Emacs do the work for them.
16798 Another @code{custom-set-@dots{}} function is @code{custom-set-faces}.
16799 This function sets the various font faces. Over time, I have set a
16800 considerable number of faces. Some of the time, I re-set them using
16801 @code{customize}; other times, I simply edit the
16802 @code{custom-set-faces} expression in my @file{.emacs} file itself.
16804 The second way to customize your @code{text-mode-hook} is to set it
16805 yourself in your @file{.emacs} file using code that has nothing to do
16806 with the @code{custom-set-@dots{}} functions.
16809 When you do this, and later use @code{customize}, you will see a
16813 CHANGED outside Customize; operating on it here may be unreliable.
16817 This message is only a warning. If you click on the button to
16820 Save for Future Sessions
16824 Emacs will write a @code{custom-set-@dots{}} expression near the end
16825 of your @file{.emacs} file that will be evaluated after your
16826 hand-written expression. It will, therefore, overrule your
16827 hand-written expression. No harm will be done. When you do this,
16828 however, be careful to remember which expression is active; if you
16829 forget, you may confuse yourself.
16831 So long as you remember where the values are set, you will have no
16832 trouble. In any event, the values are always set in your
16833 initialization file, which is usually called @file{.emacs}.
16835 I myself use @code{customize} for hardly anything. Mostly, I write
16836 expressions myself.
16840 Incidentally, to be more complete concerning defines: @code{defsubst}
16841 defines an inline function. The syntax is just like that of
16842 @code{defun}. @code{defconst} defines a symbol as a constant. The
16843 intent is that neither programs nor users should ever change a value
16844 set by @code{defconst}. (You can change it; the value set is a
16845 variable; but please do not.)
16847 @node Beginning init File
16848 @section Beginning a @file{.emacs} File
16849 @cindex @file{.emacs} file, beginning of
16851 When you start Emacs, it loads your @file{.emacs} file unless you tell
16852 it not to by specifying @samp{-q} on the command line. (The
16853 @code{emacs -q} command gives you a plain, out-of-the-box Emacs.)
16855 A @file{.emacs} file contains Lisp expressions. Often, these are no
16856 more than expressions to set values; sometimes they are function
16859 @xref{Init File, , The Init File @file{~/.emacs}, emacs, The GNU Emacs
16860 Manual}, for a short description of initialization files.
16862 This chapter goes over some of the same ground, but is a walk among
16863 extracts from a complete, long-used @file{.emacs} file---my own.
16865 The first part of the file consists of comments: reminders to myself.
16866 By now, of course, I remember these things, but when I started, I did
16872 ;;;; Bob's .emacs file
16873 ; Robert J. Chassell
16874 ; 26 September 1985
16879 Look at that date! I started this file a long time ago. I have been
16880 adding to it ever since.
16884 ; Each section in this file is introduced by a
16885 ; line beginning with four semicolons; and each
16886 ; entry is introduced by a line beginning with
16887 ; three semicolons.
16892 This describes the usual conventions for comments in Emacs Lisp.
16893 Everything on a line that follows a semicolon is a comment. Two,
16894 three, and four semicolons are used as subsection and section markers.
16895 (@xref{Comments, ,, elisp, The GNU Emacs Lisp Reference Manual}, for
16896 more about comments.)
16901 ; Control-h is the help key;
16902 ; after typing control-h, type a letter to
16903 ; indicate the subject about which you want help.
16904 ; For an explanation of the help facility,
16905 ; type control-h two times in a row.
16910 Just remember: type @kbd{C-h} two times for help.
16914 ; To find out about any mode, type control-h m
16915 ; while in that mode. For example, to find out
16916 ; about mail mode, enter mail mode and then type
16922 ``Mode help'', as I call this, is very helpful. Usually, it tells you
16923 all you need to know.
16925 Of course, you don't need to include comments like these in your
16926 @file{.emacs} file. I included them in mine because I kept forgetting
16927 about Mode help or the conventions for comments---but I was able to
16928 remember to look here to remind myself.
16930 @node Text and Auto-fill
16931 @section Text and Auto Fill Mode
16933 Now we come to the part that turns on Text mode and
16938 ;;; Text mode and Auto Fill mode
16939 ;; The next two lines put Emacs into Text mode
16940 ;; and Auto Fill mode, and are for writers who
16941 ;; want to start writing prose rather than code.
16942 (setq-default major-mode 'text-mode)
16943 (add-hook 'text-mode-hook 'turn-on-auto-fill)
16947 Here is the first part of this @file{.emacs} file that does something
16948 besides remind a forgetful human!
16950 The first of the two lines in parentheses tells Emacs to turn on Text
16951 mode when you find a file, @emph{unless} that file should go into some
16952 other mode, such as C mode.
16954 @cindex Per-buffer, local variables list
16955 @cindex Local variables list, per-buffer,
16956 @cindex Automatic mode selection
16957 @cindex Mode selection, automatic
16958 When Emacs reads a file, it looks at the extension to the file name,
16959 if any. (The extension is the part that comes after a @samp{.}.) If
16960 the file ends with a @samp{.c} or @samp{.h} extension then Emacs turns
16961 on C mode. Also, Emacs looks at first nonblank line of the file; if
16962 the line says @w{@samp{-*- C -*-}}, Emacs turns on C mode. Emacs
16963 possesses a list of extensions and specifications that it uses
16964 automatically. In addition, Emacs looks near the last page for a
16965 per-buffer, local variables list, if any.
16968 @xref{Choosing Modes, , How Major Modes are Chosen, emacs, The GNU
16971 @xref{File Variables, , Local Variables in Files, emacs, The GNU Emacs
16975 See sections ``How Major Modes are Chosen'' and ``Local Variables in
16976 Files'' in @cite{The GNU Emacs Manual}.
16979 Now, back to the @file{.emacs} file.
16982 Here is the line again; how does it work?
16984 @cindex Text Mode turned on
16986 (setq major-mode 'text-mode)
16990 This line is a short, but complete Emacs Lisp expression.
16992 We are already familiar with @code{setq}. It sets the following variable,
16993 @code{major-mode}, to the subsequent value, which is @code{text-mode}.
16994 The single-quote before @code{text-mode} tells Emacs to deal directly
16995 with the @code{text-mode} symbol, not with whatever it might stand for.
16996 @xref{set & setq, , Setting the Value of a Variable},
16997 for a reminder of how @code{setq} works.
16998 The main point is that there is no difference between the procedure you
16999 use to set a value in your @file{.emacs} file and the procedure you use
17000 anywhere else in Emacs.
17003 Here is the next line:
17005 @cindex Auto Fill mode turned on
17008 (add-hook 'text-mode-hook 'turn-on-auto-fill)
17012 In this line, the @code{add-hook} command adds
17013 @code{turn-on-auto-fill} to the variable.
17015 @code{turn-on-auto-fill} is the name of a program, that, you guessed
17016 it!, turns on Auto Fill mode.
17018 Every time Emacs turns on Text mode, Emacs runs the commands hooked
17019 onto Text mode. So every time Emacs turns on Text mode, Emacs also
17020 turns on Auto Fill mode.
17022 In brief, the first line causes Emacs to enter Text mode when you edit a
17023 file, unless the file name extension, a first non-blank line, or local
17024 variables to tell Emacs otherwise.
17026 Text mode among other actions, sets the syntax table to work
17027 conveniently for writers. In Text mode, Emacs considers an apostrophe
17028 as part of a word like a letter; but Emacs does not consider a period
17029 or a space as part of a word. Thus, @kbd{M-f} moves you over
17030 @samp{it's}. On the other hand, in C mode, @kbd{M-f} stops just after
17031 the @samp{t} of @samp{it's}.
17033 The second line causes Emacs to turn on Auto Fill mode when it turns
17034 on Text mode. In Auto Fill mode, Emacs automatically breaks a line
17035 that is too wide and brings the excessively wide part of the line down
17036 to the next line. Emacs breaks lines between words, not within them.
17038 When Auto Fill mode is turned off, lines continue to the right as you
17039 type them. Depending on how you set the value of
17040 @code{truncate-lines}, the words you type either disappear off the
17041 right side of the screen, or else are shown, in a rather ugly and
17042 unreadable manner, as a continuation line on the screen.
17045 In addition, in this part of my @file{.emacs} file, I tell the Emacs
17046 fill commands to insert two spaces after a colon:
17049 (setq colon-double-space t)
17053 @section Mail Aliases
17055 Here is a @code{setq} that turns on mail aliases, along with more
17061 ; To enter mail mode, type 'C-x m'
17062 ; To enter RMAIL (for reading mail),
17064 (setq mail-aliases t)
17068 @cindex Mail aliases
17070 This @code{setq} command sets the value of the variable
17071 @code{mail-aliases} to @code{t}. Since @code{t} means true, the line
17072 says, in effect, ``Yes, use mail aliases.''
17074 Mail aliases are convenient short names for long email addresses or
17075 for lists of email addresses. The file where you keep your aliases
17076 is @file{~/.mailrc}. You write an alias like this:
17079 alias geo george@@foobar.wiz.edu
17083 When you write a message to George, address it to @samp{geo}; the
17084 mailer will automatically expand @samp{geo} to the full address.
17086 @node Indent Tabs Mode
17087 @section Indent Tabs Mode
17088 @cindex Tabs, preventing
17089 @findex indent-tabs-mode
17091 By default, Emacs inserts tabs in place of multiple spaces when it
17092 formats a region. (For example, you might indent many lines of text
17093 all at once with the @code{indent-region} command.) Tabs look fine on
17094 a terminal or with ordinary printing, but they produce badly indented
17095 output when you use @TeX{} or Texinfo since @TeX{} ignores tabs.
17098 The following turns off Indent Tabs mode:
17102 ;;; Prevent Extraneous Tabs
17103 (setq-default indent-tabs-mode nil)
17107 Note that this line uses @code{setq-default} rather than the
17108 @code{setq} command that we have seen before. The @code{setq-default}
17109 command sets values only in buffers that do not have their own local
17110 values for the variable.
17113 @xref{Just Spaces, , Tabs vs.@: Spaces, emacs, The GNU Emacs Manual}.
17115 @xref{File Variables, , Local Variables in Files, emacs, The GNU Emacs
17119 See sections ``Tabs vs.@: Spaces'' and ``Local Variables in
17120 Files'' in @cite{The GNU Emacs Manual}.
17125 @section Some Keybindings
17127 Now for some personal keybindings:
17131 ;;; Compare windows
17132 (global-set-key "\C-cw" 'compare-windows)
17136 @findex compare-windows
17137 @code{compare-windows} is a nifty command that compares the text in
17138 your current window with text in the next window. It makes the
17139 comparison by starting at point in each window, moving over text in
17140 each window as far as they match. I use this command all the time.
17142 This also shows how to set a key globally, for all modes.
17144 @cindex Setting a key globally
17145 @cindex Global set key
17146 @cindex Key setting globally
17147 @findex global-set-key
17148 The command is @code{global-set-key}. It is followed by the
17149 keybinding. In a @file{.emacs} file, the keybinding is written as
17150 shown: @code{\C-c} stands for Control-C, which means to press the
17151 control key and the @kbd{c} key at the same time. The @code{w} means
17152 to press the @kbd{w} key. The keybinding is surrounded by double
17153 quotation marks. In documentation, you would write this as
17154 @w{@kbd{C-c w}}. (If you were binding a @key{META} key, such as
17155 @kbd{M-c}, rather than a @key{CTRL} key, you would write
17156 @w{@code{\M-c}} in your @file{.emacs} file. @xref{Init Rebinding, ,
17157 Rebinding Keys in Your Init File, emacs, The GNU Emacs Manual}, for
17160 The command invoked by the keys is @code{compare-windows}. Note that
17161 @code{compare-windows} is preceded by a single-quote; otherwise, Emacs
17162 would first try to evaluate the symbol to determine its value.
17164 These three things, the double quotation marks, the backslash before
17165 the @samp{C}, and the single-quote are necessary parts of
17166 keybinding that I tend to forget. Fortunately, I have come to
17167 remember that I should look at my existing @file{.emacs} file, and
17168 adapt what is there.
17170 As for the keybinding itself: @kbd{C-c w}. This combines the prefix
17171 key, @kbd{C-c}, with a single character, in this case, @kbd{w}. This
17172 set of keys, @kbd{C-c} followed by a single character, is strictly
17173 reserved for individuals' own use. (I call these @dfn{own} keys, since
17174 these are for my own use.) You should always be able to create such a
17175 keybinding for your own use without stomping on someone else's
17176 keybinding. If you ever write an extension to Emacs, please avoid
17177 taking any of these keys for public use. Create a key like @kbd{C-c
17178 C-w} instead. Otherwise, we will run out of own keys.
17181 Here is another keybinding, with a comment:
17185 ;;; Keybinding for 'occur'
17186 ; I use occur a lot, so let's bind it to a key:
17187 (global-set-key "\C-co" 'occur)
17192 The @code{occur} command shows all the lines in the current buffer
17193 that contain a match for a regular expression. When the region is
17194 active, @code{occur} restricts matches to such region. Otherwise it
17195 uses the entire buffer.
17196 Matching lines are shown in a buffer called @file{*Occur*}.
17197 That buffer serves as a menu to jump to occurrences.
17199 @findex global-unset-key
17200 @cindex Unbinding key
17201 @cindex Key unbinding
17203 Here is how to unbind a key, so it does not
17209 (global-unset-key "\C-xf")
17213 There is a reason for this unbinding: I found I inadvertently typed
17214 @w{@kbd{C-x f}} when I meant to type @kbd{C-x C-f}. Rather than find a
17215 file, as I intended, I accidentally set the width for filled text,
17216 almost always to a width I did not want. Since I hardly ever reset my
17217 default width, I simply unbound the key.
17219 @findex list-buffers@r{, rebound}
17220 @findex buffer-menu@r{, bound to key}
17222 The following rebinds an existing key:
17226 ;;; Rebind 'C-x C-b' for 'buffer-menu'
17227 (global-set-key "\C-x\C-b" 'buffer-menu)
17231 By default, @kbd{C-x C-b} runs the
17232 @code{list-buffers} command. This command lists
17233 your buffers in @emph{another} window. Since I
17234 almost always want to do something in that
17235 window, I prefer the @code{buffer-menu}
17236 command, which not only lists the buffers,
17237 but moves point into that window.
17242 @cindex Rebinding keys
17244 Emacs uses @dfn{keymaps} to record which keys call which commands.
17245 When you use @code{global-set-key} to set the keybinding for a single
17246 command in all parts of Emacs, you are specifying the keybinding in
17247 @code{current-global-map}.
17249 Specific modes, such as C mode or Text mode, have their own keymaps;
17250 the mode-specific keymaps override the global map that is shared by
17253 The @code{global-set-key} function binds, or rebinds, the global
17254 keymap. For example, the following binds the key @kbd{C-x C-b} to the
17255 function @code{buffer-menu}:
17258 (global-set-key "\C-x\C-b" 'buffer-menu)
17261 Mode-specific keymaps are bound using the @code{define-key} function,
17262 which takes a specific keymap as an argument, as well as the key and
17263 the command. For example, my @file{.emacs} file contains the
17264 following expression to bind the @code{texinfo-insert-@@group} command
17265 to @kbd{C-c C-c g}:
17269 (define-key texinfo-mode-map "\C-c\C-cg" 'texinfo-insert-@@group)
17274 The @code{texinfo-insert-@@group} function itself is a little extension
17275 to Texinfo mode that inserts @samp{@@group} into a Texinfo file. I
17276 use this command all the time and prefer to type the three strokes
17277 @kbd{C-c C-c g} rather than the six strokes @kbd{@@ g r o u p}.
17278 (@samp{@@group} and its matching @samp{@@end group} are commands that
17279 keep all enclosed text together on one page; many multi-line examples
17280 in this book are surrounded by @samp{@@group @dots{} @@end group}.)
17283 Here is the @code{texinfo-insert-@@group} function definition:
17287 (defun texinfo-insert-@@group ()
17288 "Insert the string @@group in a Texinfo buffer."
17290 (beginning-of-line)
17291 (insert "@@group\n"))
17295 (Of course, I could have used Abbrev mode to save typing, rather than
17296 write a function to insert a word; but I prefer key strokes consistent
17297 with other Texinfo mode key bindings.)
17299 You will see numerous @code{define-key} expressions in
17300 @file{loaddefs.el} as well as in the various mode libraries, such as
17301 @file{cc-mode.el} and @file{lisp-mode.el}.
17303 @xref{Key Bindings, , Customizing Key Bindings, emacs, The GNU Emacs
17304 Manual}, and @ref{Keymaps, , Keymaps, elisp, The GNU Emacs Lisp
17305 Reference Manual}, for more information about keymaps.
17307 @node Loading Files
17308 @section Loading Files
17309 @cindex Loading files
17312 Many people in the GNU Emacs community have written extensions to
17313 Emacs. As time goes by, these extensions are often included in new
17314 releases. For example, the Calendar and Diary packages are now part
17315 of the standard GNU Emacs, as is Calc.
17317 You can use a @code{load} command to evaluate a complete file and
17318 thereby install all the functions and variables in the file into Emacs.
17321 @c (auto-compression-mode t)
17324 (load "~/emacs/slowsplit")
17327 This evaluates, i.e., loads, the @file{slowsplit.el} file or if it
17328 exists, the faster, byte compiled @file{slowsplit.elc} file from the
17329 @file{emacs} sub-directory of your home directory. The file contains
17330 the function @code{split-window-quietly}, which John Robinson wrote in
17333 The @code{split-window-quietly} function splits a window with the
17334 minimum of redisplay. I installed it in 1989 because it worked well
17335 with the slow 1200 baud terminals I was then using. Nowadays, I only
17336 occasionally come across such a slow connection, but I continue to use
17337 the function because I like the way it leaves the bottom half of a
17338 buffer in the lower of the new windows and the top half in the upper
17342 To replace the key binding for the default
17343 @code{split-window-vertically}, you must also unset that key and bind
17344 the keys to @code{split-window-quietly}, like this:
17348 (global-unset-key "\C-x2")
17349 (global-set-key "\C-x2" 'split-window-quietly)
17354 If you load many extensions, as I do, then instead of specifying the
17355 exact location of the extension file, as shown above, you can specify
17356 that directory as part of Emacs's @code{load-path}. Then, when Emacs
17357 loads a file, it will search that directory as well as its default
17358 list of directories. (The default list is specified in @file{paths.h}
17359 when Emacs is built.)
17362 The following command adds your @file{~/emacs} directory to the
17363 existing load path:
17367 ;;; Emacs Load Path
17368 (setq load-path (cons "~/emacs" load-path))
17372 Incidentally, @code{load-library} is an interactive interface to the
17373 @code{load} function. The complete function looks like this:
17375 @findex load-library
17378 (defun load-library (library)
17379 "Load the Emacs Lisp library named LIBRARY.
17380 This is an interface to the function `load'. LIBRARY is searched
17381 for in `load-path', both with and without `load-suffixes' (as
17382 well as `load-file-rep-suffixes').
17384 See Info node `(emacs)Lisp Libraries' for more details.
17385 See `load-file' for a different interface to `load'."
17387 (list (completing-read "Load library: "
17388 (apply-partially 'locate-file-completion-table
17390 (get-load-suffixes)))))
17395 The name of the function, @code{load-library}, comes from the use of
17396 ``library'' as a conventional synonym for ``file''. The source for the
17397 @code{load-library} command is in the @file{files.el} library.
17399 Another interactive command that does a slightly different job is
17400 @code{load-file}. @xref{Lisp Libraries, , Libraries of Lisp Code for
17401 Emacs, emacs, The GNU Emacs Manual}, for information on the
17402 distinction between @code{load-library} and this command.
17405 @section Autoloading
17408 Instead of installing a function by loading the file that contains it,
17409 or by evaluating the function definition, you can make the function
17410 available but not actually install it until it is first called. This
17411 is called @dfn{autoloading}.
17413 When you execute an autoloaded function, Emacs automatically evaluates
17414 the file that contains the definition, and then calls the function.
17416 Emacs starts quicker with autoloaded functions, since their libraries
17417 are not loaded right away; but you need to wait a moment when you
17418 first use such a function, while its containing file is evaluated.
17420 Rarely used functions are frequently autoloaded. The
17421 @file{loaddefs.el} library contains thousands of autoloaded functions,
17422 from @code{5x5} to @code{zone}. Of course, you may
17423 come to use a rare function frequently. When you do, you should
17424 load that function's file with a @code{load} expression in your
17425 @file{.emacs} file.
17427 In my @file{.emacs} file, I load 14 libraries that contain functions
17428 that would otherwise be autoloaded. (Actually, it would have been
17429 better to include these files in my dumped Emacs, but I forgot.
17430 @xref{Building Emacs, , Building Emacs, elisp, The GNU Emacs Lisp
17431 Reference Manual}, and the @file{INSTALL} file for more about
17434 You may also want to include autoloaded expressions in your @file{.emacs}
17435 file. @code{autoload} is a built-in function that takes up to five
17436 arguments, the final three of which are optional. The first argument
17437 is the name of the function to be autoloaded; the second is the name
17438 of the file to be loaded. The third argument is documentation for the
17439 function, and the fourth tells whether the function can be called
17440 interactively. The fifth argument tells what type of
17441 object---@code{autoload} can handle a keymap or macro as well as a
17442 function (the default is a function).
17445 Here is a typical example:
17449 (autoload 'html-helper-mode
17450 "html-helper-mode" "Edit HTML documents" t)
17455 (@code{html-helper-mode} is an older alternative to @code{html-mode},
17456 which is a standard part of the distribution.)
17459 This expression autoloads the @code{html-helper-mode} function. It
17460 takes it from the @file{html-helper-mode.el} file (or from the byte
17461 compiled version @file{html-helper-mode.elc}, if that exists.) The
17462 file must be located in a directory specified by @code{load-path}.
17463 The documentation says that this is a mode to help you edit documents
17464 written in the HyperText Markup Language. You can call this mode
17465 interactively by typing @kbd{M-x html-helper-mode}. (You need to
17466 duplicate the function's regular documentation in the autoload
17467 expression because the regular function is not yet loaded, so its
17468 documentation is not available.)
17470 @xref{Autoload, , Autoload, elisp, The GNU Emacs Lisp Reference
17471 Manual}, for more information.
17473 @node Simple Extension
17474 @section A Simple Extension: @code{line-to-top-of-window}
17475 @findex line-to-top-of-window
17476 @cindex Simple extension in @file{.emacs} file
17478 Here is a simple extension to Emacs that moves the line point is on to
17479 the top of the window. I use this all the time, to make text easier
17482 You can put the following code into a separate file and then load it
17483 from your @file{.emacs} file, or you can include it within your
17484 @file{.emacs} file.
17487 Here is the definition:
17491 ;;; Line to top of window;
17492 ;;; replace three keystroke sequence C-u 0 C-l
17493 (defun line-to-top-of-window ()
17494 "Move the line point is on to top of window."
17501 Now for the keybinding.
17503 Nowadays, function keys as well as mouse button events and
17504 non-@sc{ascii} characters are written within square brackets, without
17505 quotation marks. (In Emacs version 18 and before, you had to write
17506 different function key bindings for each different make of terminal.)
17508 I bind @code{line-to-top-of-window} to my @key{F6} function key like
17512 (global-set-key [f6] 'line-to-top-of-window)
17515 For more information, see @ref{Init Rebinding, , Rebinding Keys in
17516 Your Init File, emacs, The GNU Emacs Manual}.
17518 @cindex Conditional 'twixt two versions of Emacs
17519 @cindex Version of Emacs, choosing
17520 @cindex Emacs version, choosing
17521 If you run two versions of GNU Emacs, such as versions 22 and 23, and
17522 use one @file{.emacs} file, you can select which code to evaluate with
17523 the following conditional:
17528 ((= 22 emacs-major-version)
17529 ;; evaluate version 22 code
17531 ((= 23 emacs-major-version)
17532 ;; evaluate version 23 code
17537 For example, recent versions blink
17538 their cursors by default. I hate such blinking, as well as other
17539 features, so I placed the following in my @file{.emacs}
17540 file@footnote{When I start instances of Emacs that do not load my
17541 @file{.emacs} file or any site file, I also turn off blinking:
17544 emacs -q --no-site-file -eval '(blink-cursor-mode nil)'
17546 @exdent Or nowadays, using an even more sophisticated set of options,
17554 (when (>= emacs-major-version 21)
17555 (blink-cursor-mode 0)
17556 ;; Insert newline when you press 'C-n' (next-line)
17557 ;; at the end of the buffer
17558 (setq next-line-add-newlines t)
17561 ;; Turn on image viewing
17562 (auto-image-file-mode t)
17565 ;; Turn on menu bar (this bar has text)
17566 ;; (Use numeric argument to turn on)
17570 ;; Turn off tool bar (this bar has icons)
17571 ;; (Use numeric argument to turn on)
17572 (tool-bar-mode nil)
17575 ;; Turn off tooltip mode for tool bar
17576 ;; (This mode causes icon explanations to pop up)
17577 ;; (Use numeric argument to turn on)
17579 ;; If tooltips turned on, make tips appear promptly
17580 (setq tooltip-delay 0.1) ; default is 0.7 second
17586 @section X11 Colors
17588 You can specify colors when you use Emacs with the MIT X Windowing
17591 I dislike the default colors and specify my own.
17594 Here are the expressions in my @file{.emacs}
17595 file that set values:
17599 ;; Set cursor color
17600 (set-cursor-color "white")
17603 (set-mouse-color "white")
17605 ;; Set foreground and background
17606 (set-foreground-color "white")
17607 (set-background-color "darkblue")
17611 ;;; Set highlighting colors for isearch and drag
17612 (set-face-foreground 'highlight "white")
17613 (set-face-background 'highlight "blue")
17617 (set-face-foreground 'region "cyan")
17618 (set-face-background 'region "blue")
17622 (set-face-foreground 'secondary-selection "skyblue")
17623 (set-face-background 'secondary-selection "darkblue")
17627 ;; Set calendar highlighting colors
17628 (with-eval-after-load 'calendar
17629 (set-face-foreground 'diary "skyblue")
17630 (set-face-background 'holiday "slate blue")
17631 (set-face-foreground 'holiday "white"))
17635 The various shades of blue soothe my eye and prevent me from seeing
17636 the screen flicker.
17638 Alternatively, I could have set my specifications in various X
17639 initialization files. For example, I could set the foreground,
17640 background, cursor, and pointer (i.e., mouse) colors in my
17641 @file{~/.Xresources} file like this:
17645 Emacs*foreground: white
17646 Emacs*background: darkblue
17647 Emacs*cursorColor: white
17648 Emacs*pointerColor: white
17652 In any event, since it is not part of Emacs, I set the root color of
17653 my X window in my @file{~/.xinitrc} file, like this@footnote{I also
17654 run more modern window managers, such as Enlightenment, Gnome, or KDE;
17655 in those cases, I often specify an image rather than a plain color.}:
17658 xsetroot -solid Navy -fg white &
17662 @node Miscellaneous
17663 @section Miscellaneous Settings for a @file{.emacs} File
17666 Here are a few miscellaneous settings:
17671 Set the shape and color of the mouse cursor:
17675 ; Cursor shapes are defined in
17676 ; '/usr/include/X11/cursorfont.h';
17677 ; for example, the 'target' cursor is number 128;
17678 ; the 'top_left_arrow' cursor is number 132.
17682 (let ((mpointer (x-get-resource "*mpointer"
17683 "*emacs*mpointer")))
17684 ;; If you have not set your mouse pointer
17685 ;; then set it, otherwise leave as is:
17686 (if (eq mpointer nil)
17687 (setq mpointer "132")) ; top_left_arrow
17690 (setq x-pointer-shape (string-to-number mpointer))
17691 (set-mouse-color "white"))
17696 Or you can set the values of a variety of features in an alist, like
17702 default-frame-alist
17703 '((cursor-color . "white")
17704 (mouse-color . "white")
17705 (foreground-color . "white")
17706 (background-color . "DodgerBlue4")
17707 ;; (cursor-type . bar)
17708 (cursor-type . box)
17711 (tool-bar-lines . 0)
17712 (menu-bar-lines . 1)
17716 "-Misc-Fixed-Medium-R-Normal--20-200-75-75-C-100-ISO8859-1")
17722 Convert @kbd{@key{CTRL}-h} into @key{DEL} and @key{DEL}
17723 into @kbd{@key{CTRL}-h}.@*
17724 (Some older keyboards needed this, although I have not seen the
17729 ;; Translate 'C-h' to <DEL>.
17730 ; (keyboard-translate ?\C-h ?\C-?)
17732 ;; Translate <DEL> to 'C-h'.
17733 (keyboard-translate ?\C-? ?\C-h)
17737 @item Turn off a blinking cursor!
17741 (if (fboundp 'blink-cursor-mode)
17742 (blink-cursor-mode -1))
17747 or start GNU Emacs with the command @code{emacs -nbc}.
17750 @item When using @command{grep}@*
17751 @samp{-i}@w{ } Ignore case distinctions@*
17752 @samp{-n}@w{ } Prefix each line of output with line number@*
17753 @samp{-H}@w{ } Print the filename for each match.@*
17754 @samp{-e}@w{ } Protect patterns beginning with a hyphen character, @samp{-}
17757 (setq grep-command "grep -i -nH -e ")
17761 @c Evidently, no longer needed in GNU Emacs 22
17763 item Automatically uncompress compressed files when visiting them
17766 (load "uncompress")
17771 @item Find an existing buffer, even if it has a different name@*
17772 This avoids problems with symbolic links.
17775 (setq find-file-existing-other-name t)
17778 @item Set your language environment and default input method
17782 (set-language-environment "latin-1")
17783 ;; Remember you can enable or disable multilingual text input
17784 ;; with the @code{toggle-input-method'} (@kbd{C-\}) command
17785 (setq default-input-method "latin-1-prefix")
17789 If you want to write with Chinese GB characters, set this instead:
17793 (set-language-environment "Chinese-GB")
17794 (setq default-input-method "chinese-tonepy")
17799 @subsubheading Fixing Unpleasant Key Bindings
17800 @cindex Key bindings, fixing
17801 @cindex Bindings, key, fixing unpleasant
17803 Some systems bind keys unpleasantly. Sometimes, for example, the
17804 @key{CTRL} key appears in an awkward spot rather than at the far left
17807 Usually, when people fix these sorts of keybindings, they do not
17808 change their @file{~/.emacs} file. Instead, they bind the proper keys
17809 on their consoles with the @code{loadkeys} or @code{install-keymap}
17810 commands in their boot script and then include @code{xmodmap} commands
17811 in their @file{.xinitrc} or @file{.Xsession} file for X Windows.
17819 loadkeys /usr/share/keymaps/i386/qwerty/emacs2.kmap.gz
17821 install-keymap emacs2
17827 For a @file{.xinitrc} or @file{.Xsession} file when the @key{Caps
17828 Lock} key is at the far left of the home row:
17832 # Bind the key labeled 'Caps Lock' to 'Control'
17833 # (Such a broken user interface suggests that keyboard manufacturers
17834 # think that computers are typewriters from 1885.)
17836 xmodmap -e "clear Lock"
17837 xmodmap -e "add Control = Caps_Lock"
17843 In a @file{.xinitrc} or @file{.Xsession} file, to convert an @key{ALT}
17844 key to a @key{META} key:
17848 # Some ill designed keyboards have a key labeled ALT and no Meta
17849 xmodmap -e "keysym Alt_L = Meta_L Alt_L"
17855 @section A Modified Mode Line
17856 @vindex mode-line-format
17857 @cindex Mode line format
17859 Finally, a feature I really like: a modified mode line.
17861 When I work over a network, I forget which machine I am using. Also,
17862 I tend to I lose track of where I am, and which line point is on.
17864 So I reset my mode line to look like this:
17867 -:-- foo.texi rattlesnake:/home/bob/ Line 1 (Texinfo Fill) Top
17870 I am visiting a file called @file{foo.texi}, on my machine
17871 @file{rattlesnake} in my @file{/home/bob} buffer. I am on line 1, in
17872 Texinfo mode, and am at the top of the buffer.
17875 My @file{.emacs} file has a section that looks like this:
17879 ;; Set a Mode Line that tells me which machine, which directory,
17880 ;; and which line I am on, plus the other customary information.
17881 (setq-default mode-line-format
17885 "mouse-1: select window, mouse-2: delete others ..."))
17886 mode-line-mule-info
17888 mode-line-frame-identification
17892 mode-line-buffer-identification
17895 (system-name) 0 (string-match "\\..+" (system-name))))
17900 "mouse-1: select window, mouse-2: delete others ..."))
17901 (line-number-mode " Line %l ")
17907 "mouse-1: select window, mouse-2: delete others ..."))
17908 (:eval (mode-line-mode-name))
17911 #("%n" 0 2 (help-echo "mouse-2: widen" local-map (keymap ...)))
17920 Here, I redefine the default mode line. Most of the parts are from
17921 the original; but I make a few changes. I set the @emph{default} mode
17922 line format so as to permit various modes, such as Info, to override
17925 Many elements in the list are self-explanatory:
17926 @code{mode-line-modified} is a variable that tells whether the buffer
17927 has been modified, @code{mode-name} tells the name of the mode, and so
17928 on. However, the format looks complicated because of two features we
17929 have not discussed.
17931 @cindex Properties, in mode line example
17932 The first string in the mode line is a dash or hyphen, @samp{-}. In
17933 the old days, it would have been specified simply as @code{"-"}. But
17934 nowadays, Emacs can add properties to a string, such as highlighting
17935 or, as in this case, a help feature. If you place your mouse cursor
17936 over the hyphen, some help information appears (By default, you must
17937 wait seven-tenths of a second before the information appears. You can
17938 change that timing by changing the value of @code{tooltip-delay}.)
17941 The new string format has a special syntax:
17944 #("-" 0 1 (help-echo "mouse-1: select window, ..."))
17948 The @code{#(} begins a list. The first element of the list is the
17949 string itself, just one @samp{-}. The second and third
17950 elements specify the range over which the fourth element applies. A
17951 range starts @emph{after} a character, so a zero means the range
17952 starts just before the first character; a 1 means that the range ends
17953 just after the first character. The third element is the property for
17954 the range. It consists of a property list, a
17955 property name, in this case, @samp{help-echo}, followed by a value, in this
17956 case, a string. The second, third, and fourth elements of this new
17957 string format can be repeated.
17959 @xref{Text Properties, , Text Properties, elisp, The GNU Emacs Lisp
17960 Reference Manual}, and see @ref{Mode Line Format, , Mode Line Format,
17961 elisp, The GNU Emacs Lisp Reference Manual}, for more information.
17963 @code{mode-line-buffer-identification}
17964 displays the current buffer name. It is a list
17965 beginning @code{(#("%12b" 0 4 @dots{}}.
17966 The @code{#(} begins the list.
17968 The @samp{"%12b"} displays the current buffer name, using the
17969 @code{buffer-name} function with which we are familiar; the @samp{12}
17970 specifies the maximum number of characters that will be displayed.
17971 When a name has fewer characters, whitespace is added to fill out to
17972 this number. (Buffer names can and often should be longer than 12
17973 characters; this length works well in a typical 80 column wide
17976 @code{:eval} says to evaluate the following form and use the result as
17977 a string to display. In this case, the expression displays the first
17978 component of the full system name. The end of the first component is
17979 a @samp{.} (period), so I use the @code{string-match} function to
17980 tell me the length of the first component. The substring from the
17981 zeroth character to that length is the name of the machine.
17984 This is the expression:
17989 (system-name) 0 (string-match "\\..+" (system-name))))
17993 @samp{%[} and @samp{%]} cause a pair of square brackets
17994 to appear for each recursive editing level. @samp{%n} says ``Narrow''
17995 when narrowing is in effect. @samp{%P} tells you the percentage of
17996 the buffer that is above the bottom of the window, or ``Top'', ``Bottom'',
17997 or ``All''. (A lower case @samp{p} tell you the percentage above the
17998 @emph{top} of the window.) @samp{%-} inserts enough dashes to fill
18001 Remember, you don't have to like Emacs to like it---your own
18002 Emacs can have different colors, different commands, and different
18003 keys than a default Emacs.
18005 On the other hand, if you want to bring up a plain out-of-the-box
18006 Emacs, with no customization, type:
18013 This will start an Emacs that does @emph{not} load your
18014 @file{~/.emacs} initialization file. A plain, default Emacs. Nothing
18021 GNU Emacs has two debuggers, @code{debug} and @code{edebug}. The
18022 first is built into the internals of Emacs and is always with you;
18023 the second requires that you instrument a function before you can use it.
18025 Both debuggers are described extensively in @ref{Debugging, ,
18026 Debugging Lisp Programs, elisp, The GNU Emacs Lisp Reference Manual}.
18027 In this chapter, I will walk through a short example of each.
18030 * debug:: How to use the built-in debugger.
18031 * debug-on-entry:: Start debugging when you call a function.
18032 * debug-on-quit:: Start debugging when you quit with @kbd{C-g}.
18033 * edebug:: How to use Edebug, a source level debugger.
18034 * Debugging Exercises::
18038 @section @code{debug}
18041 Suppose you have written a function definition that is intended to
18042 return the sum of the numbers 1 through a given number. (This is the
18043 @code{triangle} function discussed earlier. @xref{Decrementing
18044 Example, , Example with Decrementing Counter}, for a discussion.)
18045 @c xref{Decrementing Loop,, Loop with a Decrementing Counter}, for a discussion.)
18047 However, your function definition has a bug. You have mistyped
18048 @samp{1=} for @samp{1-}. Here is the broken definition:
18050 @findex triangle-bugged
18053 (defun triangle-bugged (number)
18054 "Return sum of numbers 1 through NUMBER inclusive."
18056 (while (> number 0)
18057 (setq total (+ total number))
18058 (setq number (1= number))) ; @r{Error here.}
18063 If you are reading this in Info, you can evaluate this definition in
18064 the normal fashion. You will see @code{triangle-bugged} appear in the
18068 Now evaluate the @code{triangle-bugged} function with an
18072 (triangle-bugged 4)
18076 In a recent GNU Emacs, you will create and enter a @file{*Backtrace*}
18082 ---------- Buffer: *Backtrace* ----------
18083 Debugger entered--Lisp error: (void-function 1=)
18085 (setq number (1= number))
18086 (while (> number 0) (setq total (+ total number))
18087 (setq number (1= number)))
18088 (let ((total 0)) (while (> number 0) (setq total ...)
18089 (setq number ...)) total)
18093 eval((triangle-bugged 4))
18094 eval-last-sexp-1(nil)
18095 eval-last-sexp(nil)
18096 call-interactively(eval-last-sexp)
18097 ---------- Buffer: *Backtrace* ----------
18102 (I have reformatted this example slightly; the debugger does not fold
18103 long lines. As usual, you can quit the debugger by typing @kbd{q} in
18104 the @file{*Backtrace*} buffer.)
18106 In practice, for a bug as simple as this, the Lisp error line will
18107 tell you what you need to know to correct the definition. The
18108 function @code{1=} is void.
18112 In GNU Emacs 20 and before, you will see:
18115 Symbol's function definition is void:@: 1=
18119 which has the same meaning as the @file{*Backtrace*} buffer line in
18123 However, suppose you are not quite certain what is going on?
18124 You can read the complete backtrace.
18126 In this case, you need to run a recent GNU Emacs, which automatically
18127 starts the debugger that puts you in the @file{*Backtrace*} buffer; or
18128 else, you need to start the debugger manually as described below.
18130 Read the @file{*Backtrace*} buffer from the bottom up; it tells you
18131 what Emacs did that led to the error. Emacs made an interactive call
18132 to @kbd{C-x C-e} (@code{eval-last-sexp}), which led to the evaluation
18133 of the @code{triangle-bugged} expression. Each line above tells you
18134 what the Lisp interpreter evaluated next.
18137 The third line from the top of the buffer is
18140 (setq number (1= number))
18144 Emacs tried to evaluate this expression; in order to do so, it tried
18145 to evaluate the inner expression shown on the second line from the
18154 This is where the error occurred; as the top line says:
18157 Debugger entered--Lisp error: (void-function 1=)
18161 You can correct the mistake, re-evaluate the function definition, and
18162 then run your test again.
18164 @node debug-on-entry
18165 @section @code{debug-on-entry}
18166 @findex debug-on-entry
18168 A recent GNU Emacs starts the debugger automatically when your
18169 function has an error.
18172 GNU Emacs version 20 and before did not; it simply
18173 presented you with an error message. You had to start the debugger
18177 Incidentally, you can start the debugger manually for all versions of
18178 Emacs; the advantage is that the debugger runs even if you do not have
18179 a bug in your code. Sometimes your code will be free of bugs!
18181 You can enter the debugger when you call the function by calling
18182 @code{debug-on-entry}.
18189 M-x debug-on-entry @key{RET} triangle-bugged @key{RET}
18194 Now, evaluate the following:
18197 (triangle-bugged 5)
18201 All versions of Emacs will create a @file{*Backtrace*} buffer and tell
18202 you that it is beginning to evaluate the @code{triangle-bugged}
18207 ---------- Buffer: *Backtrace* ----------
18208 Debugger entered--entering a function:
18209 * triangle-bugged(5)
18210 eval((triangle-bugged 5))
18213 eval-last-sexp-1(nil)
18214 eval-last-sexp(nil)
18215 call-interactively(eval-last-sexp)
18216 ---------- Buffer: *Backtrace* ----------
18220 In the @file{*Backtrace*} buffer, type @kbd{d}. Emacs will evaluate
18221 the first expression in @code{triangle-bugged}; the buffer will look
18226 ---------- Buffer: *Backtrace* ----------
18227 Debugger entered--beginning evaluation of function call form:
18228 * (let ((total 0)) (while (> number 0) (setq total ...)
18229 (setq number ...)) total)
18230 * triangle-bugged(5)
18231 eval((triangle-bugged 5))
18234 eval-last-sexp-1(nil)
18235 eval-last-sexp(nil)
18236 call-interactively(eval-last-sexp)
18237 ---------- Buffer: *Backtrace* ----------
18242 Now, type @kbd{d} again, eight times, slowly. Each time you type
18243 @kbd{d}, Emacs will evaluate another expression in the function
18247 Eventually, the buffer will look like this:
18251 ---------- Buffer: *Backtrace* ----------
18252 Debugger entered--beginning evaluation of function call form:
18253 * (setq number (1= number))
18254 * (while (> number 0) (setq total (+ total number))
18255 (setq number (1= number)))
18258 * (let ((total 0)) (while (> number 0) (setq total ...)
18259 (setq number ...)) total)
18260 * triangle-bugged(5)
18261 eval((triangle-bugged 5))
18264 eval-last-sexp-1(nil)
18265 eval-last-sexp(nil)
18266 call-interactively(eval-last-sexp)
18267 ---------- Buffer: *Backtrace* ----------
18273 Finally, after you type @kbd{d} two more times, Emacs will reach the
18274 error, and the top two lines of the @file{*Backtrace*} buffer will look
18279 ---------- Buffer: *Backtrace* ----------
18280 Debugger entered--Lisp error: (void-function 1=)
18283 ---------- Buffer: *Backtrace* ----------
18287 By typing @kbd{d}, you were able to step through the function.
18289 You can quit a @file{*Backtrace*} buffer by typing @kbd{q} in it; this
18290 quits the trace, but does not cancel @code{debug-on-entry}.
18292 @findex cancel-debug-on-entry
18293 To cancel the effect of @code{debug-on-entry}, call
18294 @code{cancel-debug-on-entry} and the name of the function, like this:
18297 M-x cancel-debug-on-entry @key{RET} triangle-bugged @key{RET}
18301 (If you are reading this in Info, cancel @code{debug-on-entry} now.)
18303 @node debug-on-quit
18304 @section @code{debug-on-quit} and @code{(debug)}
18306 In addition to setting @code{debug-on-error} or calling @code{debug-on-entry},
18307 there are two other ways to start @code{debug}.
18309 @findex debug-on-quit
18310 You can start @code{debug} whenever you type @kbd{C-g}
18311 (@code{keyboard-quit}) by setting the variable @code{debug-on-quit} to
18312 @code{t}. This is useful for debugging infinite loops.
18315 @cindex @code{(debug)} in code
18316 Or, you can insert a line that says @code{(debug)} into your code
18317 where you want the debugger to start, like this:
18321 (defun triangle-bugged (number)
18322 "Return sum of numbers 1 through NUMBER inclusive."
18324 (while (> number 0)
18325 (setq total (+ total number))
18326 (debug) ; @r{Start debugger.}
18327 (setq number (1= number))) ; @r{Error here.}
18332 The @code{debug} function is described in detail in @ref{Debugger, ,
18333 The Lisp Debugger, elisp, The GNU Emacs Lisp Reference Manual}.
18336 @section The @code{edebug} Source Level Debugger
18337 @cindex Source level debugger
18340 Edebug is a source level debugger. Edebug normally displays the
18341 source of the code you are debugging, with an arrow at the left that
18342 shows which line you are currently executing.
18344 You can walk through the execution of a function, line by line, or run
18345 quickly until reaching a @dfn{breakpoint} where execution stops.
18347 Edebug is described in @ref{Edebug, , , elisp, The GNU Emacs
18348 Lisp Reference Manual}.
18351 Here is a bugged function definition for @code{triangle-recursively}.
18352 @xref{Recursive triangle function, , Recursion in place of a counter},
18353 for a review of it.
18357 (defun triangle-recursively-bugged (number)
18358 "Return sum of numbers 1 through NUMBER inclusive.
18363 (triangle-recursively-bugged
18364 (1= number))))) ; @r{Error here.}
18369 Normally, you would install this definition by positioning your cursor
18370 after the function's closing parenthesis and typing @kbd{C-x C-e}
18371 (@code{eval-last-sexp}) or else by positioning your cursor within the
18372 definition and typing @kbd{C-M-x} (@code{eval-defun}). (By default,
18373 the @code{eval-defun} command works only in Emacs Lisp mode or in Lisp
18377 However, to prepare this function definition for Edebug, you must
18378 first @dfn{instrument} the code using a different command. You can do
18379 this by positioning your cursor within or just after the definition
18383 M-x edebug-defun @key{RET}
18387 This will cause Emacs to load Edebug automatically if it is not
18388 already loaded, and properly instrument the function.
18390 After instrumenting the function, place your cursor after the
18391 following expression and type @kbd{C-x C-e} (@code{eval-last-sexp}):
18394 (triangle-recursively-bugged 3)
18398 You will be jumped back to the source for
18399 @code{triangle-recursively-bugged} and the cursor positioned at the
18400 beginning of the @code{if} line of the function. Also, you will see
18401 an arrowhead at the left hand side of that line. The arrowhead marks
18402 the line where the function is executing. (In the following examples,
18403 we show the arrowhead with @samp{=>}; in a windowing system, you may
18404 see the arrowhead as a solid triangle in the window fringe.)
18407 =>@point{}(if (= number 1)
18412 In the example, the location of point is displayed with a star,
18413 @samp{@point{}} (in Info, it is displayed as @samp{-!-}).
18416 In the example, the location of point is displayed as @samp{@point{}}
18417 (in a printed book, it is displayed with a five pointed star).
18420 If you now press @key{SPC}, point will move to the next expression to
18421 be executed; the line will look like this:
18424 =>(if @point{}(= number 1)
18428 As you continue to press @key{SPC}, point will move from expression to
18429 expression. At the same time, whenever an expression returns a value,
18430 that value will be displayed in the echo area. For example, after you
18431 move point past @code{number}, you will see the following:
18434 Result: 3 (#o3, #x3, ?\C-c)
18438 This means the value of @code{number} is 3, which is octal three,
18439 hexadecimal three, and @sc{ascii} Control-C (the third letter of the
18440 alphabet, in case you need to know this information).
18442 You can continue moving through the code until you reach the line with
18443 the error. Before evaluation, that line looks like this:
18446 => @point{}(1= number))))) ; @r{Error here.}
18451 When you press @key{SPC} once again, you will produce an error message
18455 Symbol's function definition is void:@: 1=
18461 Press @kbd{q} to quit Edebug.
18463 To remove instrumentation from a function definition, simply
18464 re-evaluate it with a command that does not instrument it.
18465 For example, you could place your cursor after the definition's
18466 closing parenthesis and type @kbd{C-x C-e}.
18468 Edebug does a great deal more than walk with you through a function.
18469 You can set it so it races through on its own, stopping only at an
18470 error or at specified stopping points; you can cause it to display the
18471 changing values of various expressions; you can find out how many
18472 times a function is called, and more.
18474 Edebug is described in @ref{Edebug, , , elisp, The GNU Emacs
18475 Lisp Reference Manual}.
18478 @node Debugging Exercises
18479 @section Debugging Exercises
18483 Install the @code{@value{COUNT-WORDS}} function and then cause it to
18484 enter the built-in debugger when you call it. Run the command on a
18485 region containing two words. You will need to press @kbd{d} a
18486 remarkable number of times. On your system, is a hook called after
18487 the command finishes? (For information on hooks, see @ref{Command
18488 Overview, , Command Loop Overview, elisp, The GNU Emacs Lisp Reference
18492 Copy @code{@value{COUNT-WORDS}} into the @file{*scratch*} buffer,
18493 instrument the function for Edebug, and walk through its execution.
18494 The function does not need to have a bug, although you can introduce
18495 one if you wish. If the function lacks a bug, the walk-through
18496 completes without problems.
18499 While running Edebug, type @kbd{?} to see a list of all the Edebug commands.
18500 (The @code{global-edebug-prefix} is usually @kbd{C-x X}, i.e.,
18501 @kbd{@key{CTRL}-x} followed by an upper case @kbd{X}; use this prefix
18502 for commands made outside of the Edebug debugging buffer.)
18505 In the Edebug debugging buffer, use the @kbd{p}
18506 (@code{edebug-bounce-point}) command to see where in the region the
18507 @code{@value{COUNT-WORDS}} is working.
18510 Move point to some spot further down the function and then type the
18511 @kbd{h} (@code{edebug-goto-here}) command to jump to that location.
18514 Use the @kbd{t} (@code{edebug-trace-mode}) command to cause Edebug to
18515 walk through the function on its own; use an upper case @kbd{T} for
18516 @code{edebug-Trace-fast-mode}.
18519 Set a breakpoint, then run Edebug in Trace mode until it reaches the
18524 @chapter Conclusion
18526 We have now reached the end of this Introduction. You have now
18527 learned enough about programming in Emacs Lisp to set values, to write
18528 simple @file{.emacs} files for yourself and your friends, and write
18529 simple customizations and extensions to Emacs.
18531 This is a place to stop. Or, if you wish, you can now go onward, and
18534 You have learned some of the basic nuts and bolts of programming. But
18535 only some. There are a great many more brackets and hinges that are
18536 easy to use that we have not touched.
18538 A path you can follow right now lies among the sources to GNU Emacs
18541 @cite{The GNU Emacs Lisp Reference Manual}.
18544 @ref{Top, , The GNU Emacs Lisp Reference Manual, elisp, The GNU
18545 Emacs Lisp Reference Manual}.
18548 The Emacs Lisp sources are an adventure. When you read the sources and
18549 come across a function or expression that is unfamiliar, you need to
18550 figure out or find out what it does.
18552 Go to the Reference Manual. It is a thorough, complete, and fairly
18553 easy-to-read description of Emacs Lisp. It is written not only for
18554 experts, but for people who know what you know. (The @cite{Reference
18555 Manual} comes with the standard GNU Emacs distribution. Like this
18556 introduction, it comes as a Texinfo source file, so you can read it
18557 on your computer and as a typeset, printed book.)
18559 Go to the other built-in help that is part of GNU Emacs: the built-in
18560 documentation for all functions and variables, and
18561 @code{xref-find-definitions}, the program that takes you to sources.
18563 Here is an example of how I explore the sources. Because of its name,
18564 @file{simple.el} is the file I looked at first, a long time ago. As
18565 it happens some of the functions in @file{simple.el} are complicated,
18566 or at least look complicated at first sight. The @code{open-line}
18567 function, for example, looks complicated.
18569 You may want to walk through this function slowly, as we did with the
18570 @code{forward-sentence} function. (@xref{forward-sentence, The
18571 @code{forward-sentence} function}.) Or you may want to skip that
18572 function and look at another, such as @code{split-line}. You don't
18573 need to read all the functions. According to
18574 @code{count-words-in-defun}, the @code{split-line} function contains
18575 102 words and symbols.
18577 Even though it is short, @code{split-line} contains expressions
18578 we have not studied: @code{skip-chars-forward}, @code{indent-to},
18579 @code{current-column} and @code{insert-and-inherit}.
18581 Consider the @code{skip-chars-forward} function.
18582 In GNU Emacs, you can find out more about @code{skip-chars-forward} by
18583 typing @kbd{C-h f} (@code{describe-function}) and the name of the
18584 function. This gives you the function documentation.
18586 You may be able to guess what is done by a well named function such as
18587 @code{indent-to}; or you can look it up, too. Incidentally, the
18588 @code{describe-function} function itself is in @file{help.el}; it is
18589 one of those long, but decipherable functions. You can look up
18590 @code{describe-function} using the @kbd{C-h f} command!
18592 In this instance, since the code is Lisp, the @file{*Help*} buffer
18593 contains the name of the library containing the function's source.
18594 You can put point over the name of the library and press the @key{RET} key,
18595 which in this situation is bound to @code{help-follow}, and be taken
18596 directly to the source, in the same way as @kbd{M-.}
18597 (@code{xref-find-definitions}).
18599 The definition for @code{describe-function} illustrates how to
18600 customize the @code{interactive} expression without using the standard
18601 character codes; and it shows how to create a temporary buffer.
18603 (The @code{indent-to} function is written in C rather than Emacs Lisp;
18604 it is a built-in function. @code{help-follow} takes you to its
18605 source as does @code{xref-find-definitions}, when properly set up.)
18607 You can look at a function's source using
18608 @code{xref-find-definitions}, which is bound to @kbd{M-.} Finally,
18609 you can find out what the Reference Manual has to say by visiting the
18610 manual in Info, and typing @kbd{i} (@code{Info-index}) and the name of
18611 the function, or by looking up the function in the index to a printed
18612 copy of the manual.
18614 Similarly, you can find out what is meant by
18615 @code{insert-and-inherit}.
18617 Other interesting source files include @file{paragraphs.el},
18618 @file{loaddefs.el}, and @file{loadup.el}. The @file{paragraphs.el}
18619 file includes short, easily understood functions as well as longer
18620 ones. The @file{loaddefs.el} file contains the many standard
18621 autoloads and many keymaps. I have never looked at it all; only at
18622 parts. @file{loadup.el} is the file that loads the standard parts of
18623 Emacs; it tells you a great deal about how Emacs is built.
18624 (@xref{Building Emacs, , Building Emacs, elisp, The GNU Emacs Lisp
18625 Reference Manual}, for more about building.)
18627 As I said, you have learned some nuts and bolts; however, and very
18628 importantly, we have hardly touched major aspects of programming; I
18629 have said nothing about how to sort information, except to use the
18630 predefined @code{sort} function; I have said nothing about how to store
18631 information, except to use variables and lists; I have said nothing
18632 about how to write programs that write programs. These are topics for
18633 another, and different kind of book, a different kind of learning.
18635 What you have done is learn enough for much practical work with GNU
18636 Emacs. What you have done is get started. This is the end of a
18639 @c ================ Appendix ================
18642 @appendix The @code{the-the} Function
18644 @cindex Duplicated words function
18645 @cindex Words, duplicated
18647 Sometimes when you you write text, you duplicate words---as with ``you
18648 you'' near the beginning of this sentence. I find that most
18649 frequently, I duplicate ``the''; hence, I call the function for
18650 detecting duplicated words, @code{the-the}.
18653 As a first step, you could use the following regular expression to
18654 search for duplicates:
18657 \\(\\w+[ \t\n]+\\)\\1
18661 This regexp matches one or more word-constituent characters followed
18662 by one or more spaces, tabs, or newlines. However, it does not detect
18663 duplicated words on different lines, since the ending of the first
18664 word, the end of the line, is different from the ending of the second
18665 word, a space. (For more information about regular expressions, see
18666 @ref{Regexp Search, , Regular Expression Searches}, as well as
18667 @ref{Regexps, , Syntax of Regular Expressions, emacs, The GNU Emacs
18668 Manual}, and @ref{Regular Expressions, , Regular Expressions, elisp,
18669 The GNU Emacs Lisp Reference Manual}.)
18671 You might try searching just for duplicated word-constituent
18672 characters but that does not work since the pattern detects doubles
18673 such as the two occurrences of ``th'' in ``with the''.
18675 Another possible regexp searches for word-constituent characters
18676 followed by non-word-constituent characters, reduplicated. Here,
18677 @w{@samp{\\w+}} matches one or more word-constituent characters and
18678 @w{@samp{\\W*}} matches zero or more non-word-constituent characters.
18681 \\(\\(\\w+\\)\\W*\\)\\1
18687 Here is the pattern that I use. It is not perfect, but good enough.
18688 @w{@samp{\\b}} matches the empty string, provided it is at the beginning
18689 or end of a word; @w{@samp{[^@@ \n\t]+}} matches one or more occurrences of
18690 any characters that are @emph{not} an @@-sign, space, newline, or tab.
18693 \\b\\([^@@ \n\t]+\\)[ \n\t]+\\1\\b
18696 One can write more complicated expressions, but I found that this
18697 expression is good enough, so I use it.
18699 Here is the @code{the-the} function, as I include it in my
18700 @file{.emacs} file, along with a handy global key binding:
18705 "Search forward for for a duplicated word."
18707 (message "Searching for for duplicated words ...")
18711 ;; This regexp is not perfect
18712 ;; but is fairly good over all:
18713 (if (re-search-forward
18714 "\\b\\([^@@ \n\t]+\\)[ \n\t]+\\1\\b" nil 'move)
18715 (message "Found duplicated word.")
18716 (message "End of buffer")))
18720 ;; Bind 'the-the' to C-c \
18721 (global-set-key "\C-c\\" 'the-the)
18730 one two two three four five
18735 You can substitute the other regular expressions shown above in the
18736 function definition and try each of them on this list.
18739 @appendix Handling the Kill Ring
18740 @cindex Kill ring handling
18741 @cindex Handling the kill ring
18742 @cindex Ring, making a list like a
18744 The kill ring is a list that is transformed into a ring by the
18745 workings of the @code{current-kill} function. The @code{yank} and
18746 @code{yank-pop} commands use the @code{current-kill} function.
18748 This appendix describes the @code{current-kill} function as well as
18749 both the @code{yank} and the @code{yank-pop} commands, but first,
18750 consider the workings of the kill ring.
18753 * What the Kill Ring Does::
18755 * yank:: Paste a copy of a clipped element.
18756 * yank-pop:: Insert element pointed to.
18761 @node What the Kill Ring Does
18762 @unnumberedsec What the Kill Ring Does
18766 The kill ring has a default maximum length of sixty items; this number
18767 is too large for an explanation. Instead, set it to four. Please
18768 evaluate the following:
18772 (setq old-kill-ring-max kill-ring-max)
18773 (setq kill-ring-max 4)
18778 Then, please copy each line of the following indented example into the
18779 kill ring. You may kill each line with @kbd{C-k} or mark it and copy
18783 (In a read-only buffer, such as the @file{*info*} buffer, the kill
18784 command, @kbd{C-k} (@code{kill-line}), will not remove the text,
18785 merely copy it to the kill ring. However, your machine may beep at
18786 you. Alternatively, for silence, you may copy the region of each line
18787 with the @kbd{M-w} (@code{kill-ring-save}) command. You must mark
18788 each line for this command to succeed, but it does not matter at which
18789 end you put point or mark.)
18793 Please invoke the calls in order, so that five elements attempt to
18794 fill the kill ring:
18799 second piece of text
18801 fourth line of text
18808 Then find the value of @code{kill-ring} by evaluating
18820 ("fifth bit of text" "fourth line of text"
18821 "third line" "second piece of text")
18826 The first element, @samp{first some text}, was dropped.
18829 To return to the old value for the length of the kill ring, evaluate:
18832 (setq kill-ring-max old-kill-ring-max)
18836 @appendixsec The @code{current-kill} Function
18837 @findex current-kill
18839 The @code{current-kill} function changes the element in the kill ring
18840 to which @code{kill-ring-yank-pointer} points. (Also, the
18841 @code{kill-new} function sets @code{kill-ring-yank-pointer} to point
18842 to the latest element of the kill ring. The @code{kill-new}
18843 function is used directly or indirectly by @code{kill-append},
18844 @code{copy-region-as-kill}, @code{kill-ring-save}, @code{kill-line},
18845 and @code{kill-region}.)
18848 * Code for current-kill::
18849 * Understanding current-kill::
18853 @node Code for current-kill
18854 @unnumberedsubsec The code for @code{current-kill}
18859 The @code{current-kill} function is used by @code{yank} and by
18860 @code{yank-pop}. Here is the code for @code{current-kill}:
18864 (defun current-kill (n &optional do-not-move)
18865 "Rotate the yanking point by N places, and then return that kill.
18866 If N is zero and `interprogram-paste-function' is set to a
18867 function that returns a string or a list of strings, and if that
18868 function doesn't return nil, then that string (or list) is added
18869 to the front of the kill ring and the string (or first string in
18870 the list) is returned as the latest kill.
18873 If N is not zero, and if `yank-pop-change-selection' is
18874 non-nil, use `interprogram-cut-function' to transfer the
18875 kill at the new yank point into the window system selection.
18878 If optional arg DO-NOT-MOVE is non-nil, then don't actually
18879 move the yanking point; just return the Nth kill forward."
18881 (let ((interprogram-paste (and (= n 0)
18882 interprogram-paste-function
18883 (funcall interprogram-paste-function))))
18886 (if interprogram-paste
18888 ;; Disable the interprogram cut function when we add the new
18889 ;; text to the kill ring, so Emacs doesn't try to own the
18890 ;; selection, with identical text.
18891 (let ((interprogram-cut-function nil))
18892 (if (listp interprogram-paste)
18893 (mapc 'kill-new (nreverse interprogram-paste))
18894 (kill-new interprogram-paste)))
18898 (or kill-ring (error "Kill ring is empty"))
18899 (let ((ARGth-kill-element
18900 (nthcdr (mod (- n (length kill-ring-yank-pointer))
18901 (length kill-ring))
18903 (unless do-not-move
18904 (setq kill-ring-yank-pointer ARGth-kill-element)
18905 (when (and yank-pop-change-selection
18907 interprogram-cut-function)
18908 (funcall interprogram-cut-function (car ARGth-kill-element))))
18909 (car ARGth-kill-element)))))
18913 Remember also that the @code{kill-new} function sets
18914 @code{kill-ring-yank-pointer} to the latest element of the kill
18915 ring, which means that all the functions that call it set the value
18916 indirectly: @code{kill-append}, @code{copy-region-as-kill},
18917 @code{kill-ring-save}, @code{kill-line}, and @code{kill-region}.
18920 Here is the line in @code{kill-new}, which is explained in
18921 @ref{kill-new function, , The @code{kill-new} function}.
18924 (setq kill-ring-yank-pointer kill-ring)
18928 @node Understanding current-kill
18929 @unnumberedsubsec @code{current-kill} in Outline
18932 The @code{current-kill} function looks complex, but as usual, it can
18933 be understood by taking it apart piece by piece. First look at it in
18938 (defun current-kill (n &optional do-not-move)
18939 "Rotate the yanking point by N places, and then return that kill."
18945 This function takes two arguments, one of which is optional. It has a
18946 documentation string. It is @emph{not} interactive.
18949 * Body of current-kill::
18950 * Digression concerning error:: How to mislead humans, but not computers.
18951 * Determining the Element::
18955 @node Body of current-kill
18956 @unnumberedsubsubsec The Body of @code{current-kill}
18959 The body of the function definition is a @code{let} expression, which
18960 itself has a body as well as a @var{varlist}.
18962 The @code{let} expression declares a variable that will be only usable
18963 within the bounds of this function. This variable is called
18964 @code{interprogram-paste} and is for copying to another program. It
18965 is not for copying within this instance of GNU Emacs. Most window
18966 systems provide a facility for interprogram pasting. Sadly, that
18967 facility usually provides only for the last element. Most windowing
18968 systems have not adopted a ring of many possibilities, even though
18969 Emacs has provided it for decades.
18971 The @code{if} expression has two parts, one if there exists
18972 @code{interprogram-paste} and one if not.
18975 Let us consider the else-part of the @code{current-kill}
18976 function. (The then-part uses the @code{kill-new} function, which
18977 we have already described. @xref{kill-new function, , The
18978 @code{kill-new} function}.)
18982 (or kill-ring (error "Kill ring is empty"))
18983 (let ((ARGth-kill-element
18984 (nthcdr (mod (- n (length kill-ring-yank-pointer))
18985 (length kill-ring))
18988 (setq kill-ring-yank-pointer ARGth-kill-element))
18989 (car ARGth-kill-element))
18994 The code first checks whether the kill ring has content; otherwise it
18998 Note that the @code{or} expression is very similar to testing length
19005 (if (zerop (length kill-ring)) ; @r{if-part}
19006 (error "Kill ring is empty")) ; @r{then-part}
19012 If there is not anything in the kill ring, its length must be zero and
19013 an error message sent to the user: @samp{Kill ring is empty}. The
19014 @code{current-kill} function uses an @code{or} expression which is
19015 simpler. But an @code{if} expression reminds us what goes on.
19017 This @code{if} expression uses the function @code{zerop} which returns
19018 true if the value it is testing is zero. When @code{zerop} tests
19019 true, the then-part of the @code{if} is evaluated. The then-part is a
19020 list starting with the function @code{error}, which is a function that
19021 is similar to the @code{message} function
19022 (@pxref{message, , The @code{message} Function}) in that
19023 it prints a one-line message in the echo area. However, in addition
19024 to printing a message, @code{error} also stops evaluation of the
19025 function within which it is embedded. This means that the rest of the
19026 function will not be evaluated if the length of the kill ring is zero.
19028 Then the @code{current-kill} function selects the element to return.
19029 The selection depends on the number of places that @code{current-kill}
19030 rotates and on where @code{kill-ring-yank-pointer} points.
19032 Next, either the optional @code{do-not-move} argument is true or the
19033 current value of @code{kill-ring-yank-pointer} is set to point to the
19034 list. Finally, another expression returns the first element of the
19035 list even if the @code{do-not-move} argument is true.
19038 @node Digression concerning error
19039 @unnumberedsubsubsec Digression about the word ``error''
19042 In my opinion, it is slightly misleading, at least to humans, to use
19043 the term ``error'' as the name of the @code{error} function. A better
19044 term would be ``cancel''. Strictly speaking, of course, you cannot
19045 point to, much less rotate a pointer to a list that has no length, so
19046 from the point of view of the computer, the word ``error'' is correct.
19047 But a human expects to attempt this sort of thing, if only to find out
19048 whether the kill ring is full or empty. This is an act of
19051 From the human point of view, the act of exploration and discovery is
19052 not necessarily an error, and therefore should not be labeled as one,
19053 even in the bowels of a computer. As it is, the code in Emacs implies
19054 that a human who is acting virtuously, by exploring his or her
19055 environment, is making an error. This is bad. Even though the computer
19056 takes the same steps as it does when there is an error, a term such as
19057 ``cancel'' would have a clearer connotation.
19060 @node Determining the Element
19061 @unnumberedsubsubsec Determining the Element
19064 Among other actions, the else-part of the @code{if} expression sets
19065 the value of @code{kill-ring-yank-pointer} to
19066 @code{ARGth-kill-element} when the kill ring has something in it and
19067 the value of @code{do-not-move} is @code{nil}.
19070 The code looks like this:
19074 (nthcdr (mod (- n (length kill-ring-yank-pointer))
19075 (length kill-ring))
19080 This needs some examination. Unless it is not supposed to move the
19081 pointer, the @code{current-kill} function changes where
19082 @code{kill-ring-yank-pointer} points.
19084 @w{@code{(setq kill-ring-yank-pointer ARGth-kill-element))}}
19085 expression does. Also, clearly, @code{ARGth-kill-element} is being
19086 set to be equal to some @sc{cdr} of the kill ring, using the
19087 @code{nthcdr} function that is described in an earlier section.
19088 (@xref{copy-region-as-kill}.) How does it do this?
19090 As we have seen before (@pxref{nthcdr}), the @code{nthcdr} function
19091 works by repeatedly taking the @sc{cdr} of a list---it takes the
19092 @sc{cdr} of the @sc{cdr} of the @sc{cdr} @dots{}
19095 The two following expressions produce the same result:
19099 (setq kill-ring-yank-pointer (cdr kill-ring))
19101 (setq kill-ring-yank-pointer (nthcdr 1 kill-ring))
19105 However, the @code{nthcdr} expression is more complicated. It uses
19106 the @code{mod} function to determine which @sc{cdr} to select.
19108 (You will remember to look at inner functions first; indeed, we will
19109 have to go inside the @code{mod}.)
19111 The @code{mod} function returns the value of its first argument modulo
19112 the second; that is to say, it returns the remainder after dividing
19113 the first argument by the second. The value returned has the same
19114 sign as the second argument.
19122 @result{} 0 ;; @r{because there is no remainder}
19129 In this case, the first argument is often smaller than the second.
19141 We can guess what the @code{-} function does. It is like @code{+} but
19142 subtracts instead of adds; the @code{-} function subtracts its second
19143 argument from its first. Also, we already know what the @code{length}
19144 function does (@pxref{length}). It returns the length of a list.
19146 And @code{n} is the name of the required argument to the
19147 @code{current-kill} function.
19150 So when the first argument to @code{nthcdr} is zero, the @code{nthcdr}
19151 expression returns the whole list, as you can see by evaluating the
19156 ;; kill-ring-yank-pointer @r{and} kill-ring @r{have a length of four}
19157 ;; @r{and} (mod (- 0 4) 4) @result{} 0
19158 (nthcdr (mod (- 0 4) 4)
19159 '("fourth line of text"
19161 "second piece of text"
19162 "first some text"))
19167 When the first argument to the @code{current-kill} function is one,
19168 the @code{nthcdr} expression returns the list without its first
19173 (nthcdr (mod (- 1 4) 4)
19174 '("fourth line of text"
19176 "second piece of text"
19177 "first some text"))
19181 @cindex @samp{global variable} defined
19182 @cindex @samp{variable, global}, defined
19183 Incidentally, both @code{kill-ring} and @code{kill-ring-yank-pointer}
19184 are @dfn{global variables}. That means that any expression in Emacs
19185 Lisp can access them. They are not like the local variables set by
19186 @code{let} or like the symbols in an argument list.
19187 Local variables can only be accessed
19188 within the @code{let} that defines them or the function that specifies
19189 them in an argument list (and within expressions called by them).
19191 @c texi2dvi fails when the name of the section is within ifnottex ...
19193 (@xref{Prevent confusion, , @code{let} Prevents Confusion}, and
19196 (@xref{Permanent Installation, , @code{let} Prevents Confusion}, and
19198 @ref{defun, , The @code{defun} Macro}.)
19201 @appendixsec @code{yank}
19204 After learning about @code{current-kill}, the code for the
19205 @code{yank} function is almost easy.
19207 The @code{yank} function does not use the
19208 @code{kill-ring-yank-pointer} variable directly. It calls
19209 @code{insert-for-yank} which calls @code{current-kill} which sets the
19210 @code{kill-ring-yank-pointer} variable.
19213 The code looks like this:
19218 (defun yank (&optional arg)
19219 "Reinsert (\"paste\") the last stretch of killed text.
19220 More precisely, reinsert the stretch of killed text most recently
19221 killed OR yanked. Put point at end, and set mark at beginning.
19222 With just \\[universal-argument] as argument, same but put point at beginning (and mark at end).
19223 With argument N, reinsert the Nth most recently killed stretch of killed
19226 When this command inserts killed text into the buffer, it honors
19227 `yank-excluded-properties' and `yank-handler' as described in the
19228 doc string for `insert-for-yank-1', which see.
19230 See also the command `yank-pop' (\\[yank-pop])."
19234 (setq yank-window-start (window-start))
19235 ;; If we don't get all the way thru, make last-command indicate that
19236 ;; for the following command.
19237 (setq this-command t)
19238 (push-mark (point))
19241 (insert-for-yank (current-kill (cond
19246 ;; This is like exchange-point-and-mark, but doesn't activate the mark.
19247 ;; It is cleaner to avoid activation, even though the command
19248 ;; loop would deactivate the mark because we inserted text.
19249 (goto-char (prog1 (mark t)
19250 (set-marker (mark-marker) (point) (current-buffer)))))
19253 ;; If we do get all the way thru, make this-command indicate that.
19254 (if (eq this-command t)
19255 (setq this-command 'yank))
19260 The key expression is @code{insert-for-yank}, which inserts the string
19261 returned by @code{current-kill}, but removes some text properties from
19264 However, before getting to that expression, the function sets the value
19265 of @code{yank-window-start} to the position returned by the
19266 @code{(window-start)} expression, the position at which the display
19267 currently starts. The @code{yank} function also sets
19268 @code{this-command} and pushes the mark.
19270 After it yanks the appropriate element, if the optional argument is a
19271 @sc{cons} rather than a number or nothing, it puts point at beginning
19272 of the yanked text and mark at its end.
19274 (The @code{prog1} function is like @code{progn} but returns the value
19275 of its first argument rather than the value of its last argument. Its
19276 first argument is forced to return the buffer's mark as an integer.
19277 You can see the documentation for these functions by placing point
19278 over them in this buffer and then typing @kbd{C-h f}
19279 (@code{describe-function}) followed by a @kbd{RET}; the default is the
19282 The last part of the function tells what to do when it succeeds.
19285 @appendixsec @code{yank-pop}
19288 After understanding @code{yank} and @code{current-kill}, you know how
19289 to approach the @code{yank-pop} function. Leaving out the
19290 documentation to save space, it looks like this:
19295 (defun yank-pop (&optional arg)
19298 (if (not (eq last-command 'yank))
19299 (error "Previous command was not a yank"))
19302 (setq this-command 'yank)
19303 (unless arg (setq arg 1))
19304 (let ((inhibit-read-only t)
19305 (before (< (point) (mark t))))
19309 (funcall (or yank-undo-function 'delete-region) (point) (mark t))
19310 (funcall (or yank-undo-function 'delete-region) (mark t) (point)))
19311 (setq yank-undo-function nil)
19314 (set-marker (mark-marker) (point) (current-buffer))
19315 (insert-for-yank (current-kill arg))
19316 ;; Set the window start back where it was in the yank command,
19318 (set-window-start (selected-window) yank-window-start t)
19322 ;; This is like exchange-point-and-mark,
19323 ;; but doesn't activate the mark.
19324 ;; It is cleaner to avoid activation, even though the command
19325 ;; loop would deactivate the mark because we inserted text.
19326 (goto-char (prog1 (mark t)
19327 (set-marker (mark-marker)
19329 (current-buffer))))))
19334 The function is interactive with a small @samp{p} so the prefix
19335 argument is processed and passed to the function. The command can
19336 only be used after a previous yank; otherwise an error message is
19337 sent. This check uses the variable @code{last-command} which is set
19338 by @code{yank} and is discussed elsewhere.
19339 (@xref{copy-region-as-kill}.)
19341 The @code{let} clause sets the variable @code{before} to true or false
19342 depending whether point is before or after mark and then the region
19343 between point and mark is deleted. This is the region that was just
19344 inserted by the previous yank and it is this text that will be
19347 @code{funcall} calls its first argument as a function, passing
19348 remaining arguments to it. The first argument is whatever the
19349 @code{or} expression returns. The two remaining arguments are the
19350 positions of point and mark set by the preceding @code{yank} command.
19352 There is more, but that is the hardest part.
19355 @appendixsec The @file{ring.el} File
19356 @cindex @file{ring.el} file
19358 Interestingly, GNU Emacs posses a file called @file{ring.el} that
19359 provides many of the features we just discussed. But functions such
19360 as @code{kill-ring-yank-pointer} do not use this library, possibly
19361 because they were written earlier.
19364 @appendix A Graph with Labeled Axes
19366 Printed axes help you understand a graph. They convey scale. In an
19367 earlier chapter (@pxref{Readying a Graph, , Readying a Graph}), we
19368 wrote the code to print the body of a graph. Here we write the code
19369 for printing and labeling vertical and horizontal axes, along with the
19373 * Labeled Example::
19374 * print-graph Varlist:: @code{let} expression in @code{print-graph}.
19375 * print-Y-axis:: Print a label for the vertical axis.
19376 * print-X-axis:: Print a horizontal label.
19377 * Print Whole Graph:: The function to print a complete graph.
19381 @node Labeled Example
19382 @unnumberedsec Labeled Example Graph
19385 Since insertions fill a buffer to the right and below point, the new
19386 graph printing function should first print the Y or vertical axis,
19387 then the body of the graph, and finally the X or horizontal axis.
19388 This sequence lays out for us the contents of the function:
19398 Print body of graph.
19405 Here is an example of how a finished graph should look:
19418 1 - ****************
19425 In this graph, both the vertical and the horizontal axes are labeled
19426 with numbers. However, in some graphs, the horizontal axis is time
19427 and would be better labeled with months, like this:
19441 Indeed, with a little thought, we can easily come up with a variety of
19442 vertical and horizontal labeling schemes. Our task could become
19443 complicated. But complications breed confusion. Rather than permit
19444 this, it is better choose a simple labeling scheme for our first
19445 effort, and to modify or replace it later.
19448 These considerations suggest the following outline for the
19449 @code{print-graph} function:
19453 (defun print-graph (numbers-list)
19454 "@var{documentation}@dots{}"
19455 (let ((height @dots{}
19459 (print-Y-axis height @dots{} )
19460 (graph-body-print numbers-list)
19461 (print-X-axis @dots{} )))
19465 We can work on each part of the @code{print-graph} function definition
19468 @node print-graph Varlist
19469 @appendixsec The @code{print-graph} Varlist
19470 @cindex @code{print-graph} varlist
19472 In writing the @code{print-graph} function, the first task is to write
19473 the varlist in the @code{let} expression. (We will leave aside for the
19474 moment any thoughts about making the function interactive or about the
19475 contents of its documentation string.)
19477 The varlist should set several values. Clearly, the top of the label
19478 for the vertical axis must be at least the height of the graph, which
19479 means that we must obtain this information here. Note that the
19480 @code{print-graph-body} function also requires this information. There
19481 is no reason to calculate the height of the graph in two different
19482 places, so we should change @code{print-graph-body} from the way we
19483 defined it earlier to take advantage of the calculation.
19485 Similarly, both the function for printing the X axis labels and the
19486 @code{print-graph-body} function need to learn the value of the width of
19487 each symbol. We can perform the calculation here and change the
19488 definition for @code{print-graph-body} from the way we defined it in the
19491 The length of the label for the horizontal axis must be at least as long
19492 as the graph. However, this information is used only in the function
19493 that prints the horizontal axis, so it does not need to be calculated here.
19495 These thoughts lead us directly to the following form for the varlist
19496 in the @code{let} for @code{print-graph}:
19500 (let ((height (apply 'max numbers-list)) ; @r{First version.}
19501 (symbol-width (length graph-blank)))
19506 As we shall see, this expression is not quite right.
19510 @appendixsec The @code{print-Y-axis} Function
19511 @cindex Axis, print vertical
19512 @cindex Y axis printing
19513 @cindex Vertical axis printing
19514 @cindex Print vertical axis
19516 The job of the @code{print-Y-axis} function is to print a label for
19517 the vertical axis that looks like this:
19535 The function should be passed the height of the graph, and then should
19536 construct and insert the appropriate numbers and marks.
19539 * print-Y-axis in Detail::
19540 * Height of label:: What height for the Y axis?
19541 * Compute a Remainder:: How to compute the remainder of a division.
19542 * Y Axis Element:: Construct a line for the Y axis.
19543 * Y-axis-column:: Generate a list of Y axis labels.
19544 * print-Y-axis Penultimate:: A not quite final version.
19548 @node print-Y-axis in Detail
19549 @unnumberedsubsec The @code{print-Y-axis} Function in Detail
19552 It is easy enough to see in the figure what the Y axis label should
19553 look like; but to say in words, and then to write a function
19554 definition to do the job is another matter. It is not quite true to
19555 say that we want a number and a tic every five lines: there are only
19556 three lines between the @samp{1} and the @samp{5} (lines 2, 3, and 4),
19557 but four lines between the @samp{5} and the @samp{10} (lines 6, 7, 8,
19558 and 9). It is better to say that we want a number and a tic mark on
19559 the base line (number 1) and then that we want a number and a tic on
19560 the fifth line from the bottom and on every line that is a multiple of
19564 @node Height of label
19565 @unnumberedsubsec What height should the label be?
19568 The next issue is what height the label should be? Suppose the maximum
19569 height of tallest column of the graph is seven. Should the highest
19570 label on the Y axis be @samp{5 -}, and should the graph stick up above
19571 the label? Or should the highest label be @samp{7 -}, and mark the peak
19572 of the graph? Or should the highest label be @code{10 -}, which is a
19573 multiple of five, and be higher than the topmost value of the graph?
19575 The latter form is preferred. Most graphs are drawn within rectangles
19576 whose sides are an integral number of steps long---5, 10, 15, and so
19577 on for a step distance of five. But as soon as we decide to use a
19578 step height for the vertical axis, we discover that the simple
19579 expression in the varlist for computing the height is wrong. The
19580 expression is @code{(apply 'max numbers-list)}. This returns the
19581 precise height, not the maximum height plus whatever is necessary to
19582 round up to the nearest multiple of five. A more complex expression
19585 As usual in cases like this, a complex problem becomes simpler if it is
19586 divided into several smaller problems.
19588 First, consider the case when the highest value of the graph is an
19589 integral multiple of five---when it is 5, 10, 15, or some higher
19590 multiple of five. We can use this value as the Y axis height.
19592 A fairly simply way to determine whether a number is a multiple of
19593 five is to divide it by five and see if the division results in a
19594 remainder. If there is no remainder, the number is a multiple of
19595 five. Thus, seven divided by five has a remainder of two, and seven
19596 is not an integral multiple of five. Put in slightly different
19597 language, more reminiscent of the classroom, five goes into seven
19598 once, with a remainder of two. However, five goes into ten twice,
19599 with no remainder: ten is an integral multiple of five.
19601 @node Compute a Remainder
19602 @appendixsubsec Side Trip: Compute a Remainder
19604 @findex % @r{(remainder function)}
19605 @cindex Remainder function, @code{%}
19606 In Lisp, the function for computing a remainder is @code{%}. The
19607 function returns the remainder of its first argument divided by its
19608 second argument. As it happens, @code{%} is a function in Emacs Lisp
19609 that you cannot discover using @code{apropos}: you find nothing if you
19610 type @kbd{M-x apropos @key{RET} remainder @key{RET}}. The only way to
19611 learn of the existence of @code{%} is to read about it in a book such
19612 as this or in the Emacs Lisp sources.
19614 You can try the @code{%} function by evaluating the following two
19626 The first expression returns 2 and the second expression returns 0.
19628 To test whether the returned value is zero or some other number, we
19629 can use the @code{zerop} function. This function returns @code{t} if
19630 its argument, which must be a number, is zero.
19642 Thus, the following expression will return @code{t} if the height
19643 of the graph is evenly divisible by five:
19646 (zerop (% height 5))
19650 (The value of @code{height}, of course, can be found from @code{(apply
19651 'max numbers-list)}.)
19653 On the other hand, if the value of @code{height} is not a multiple of
19654 five, we want to reset the value to the next higher multiple of five.
19655 This is straightforward arithmetic using functions with which we are
19656 already familiar. First, we divide the value of @code{height} by five
19657 to determine how many times five goes into the number. Thus, five
19658 goes into twelve twice. If we add one to this quotient and multiply by
19659 five, we will obtain the value of the next multiple of five that is
19660 larger than the height. Five goes into twelve twice. Add one to two,
19661 and multiply by five; the result is fifteen, which is the next multiple
19662 of five that is higher than twelve. The Lisp expression for this is:
19665 (* (1+ (/ height 5)) 5)
19669 For example, if you evaluate the following, the result is 15:
19672 (* (1+ (/ 12 5)) 5)
19675 All through this discussion, we have been using 5 as the value
19676 for spacing labels on the Y axis; but we may want to use some other
19677 value. For generality, we should replace 5 with a variable to
19678 which we can assign a value. The best name I can think of for this
19679 variable is @code{Y-axis-label-spacing}.
19682 Using this term, and an @code{if} expression, we produce the
19687 (if (zerop (% height Y-axis-label-spacing))
19690 (* (1+ (/ height Y-axis-label-spacing))
19691 Y-axis-label-spacing))
19696 This expression returns the value of @code{height} itself if the height
19697 is an even multiple of the value of the @code{Y-axis-label-spacing} or
19698 else it computes and returns a value of @code{height} that is equal to
19699 the next higher multiple of the value of the @code{Y-axis-label-spacing}.
19701 We can now include this expression in the @code{let} expression of the
19702 @code{print-graph} function (after first setting the value of
19703 @code{Y-axis-label-spacing}):
19704 @vindex Y-axis-label-spacing
19708 (defvar Y-axis-label-spacing 5
19709 "Number of lines from one Y axis label to next.")
19714 (let* ((height (apply 'max numbers-list))
19715 (height-of-top-line
19716 (if (zerop (% height Y-axis-label-spacing))
19721 (* (1+ (/ height Y-axis-label-spacing))
19722 Y-axis-label-spacing)))
19723 (symbol-width (length graph-blank))))
19729 (Note use of the @code{let*} function: the initial value of height is
19730 computed once by the @code{(apply 'max numbers-list)} expression and
19731 then the resulting value of @code{height} is used to compute its
19732 final value. @xref{fwd-para let, , The @code{let*} expression}, for
19733 more about @code{let*}.)
19735 @node Y Axis Element
19736 @appendixsubsec Construct a Y Axis Element
19738 When we print the vertical axis, we want to insert strings such as
19739 @w{@samp{5 -}} and @w{@samp{10 - }} every five lines.
19740 Moreover, we want the numbers and dashes to line up, so shorter
19741 numbers must be padded with leading spaces. If some of the strings
19742 use two digit numbers, the strings with single digit numbers must
19743 include a leading blank space before the number.
19745 @findex number-to-string
19746 To figure out the length of the number, the @code{length} function is
19747 used. But the @code{length} function works only with a string, not with
19748 a number. So the number has to be converted from being a number to
19749 being a string. This is done with the @code{number-to-string} function.
19754 (length (number-to-string 35))
19757 (length (number-to-string 100))
19763 (@code{number-to-string} is also called @code{int-to-string}; you will
19764 see this alternative name in various sources.)
19766 In addition, in each label, each number is followed by a string such
19767 as @w{@samp{ - }}, which we will call the @code{Y-axis-tic} marker.
19768 This variable is defined with @code{defvar}:
19773 (defvar Y-axis-tic " - "
19774 "String that follows number in a Y axis label.")
19778 The length of the Y label is the sum of the length of the Y axis tic
19779 mark and the length of the number of the top of the graph.
19782 (length (concat (number-to-string height) Y-axis-tic)))
19785 This value will be calculated by the @code{print-graph} function in
19786 its varlist as @code{full-Y-label-width} and passed on. (Note that we
19787 did not think to include this in the varlist when we first proposed it.)
19789 To make a complete vertical axis label, a tic mark is concatenated
19790 with a number; and the two together may be preceded by one or more
19791 spaces depending on how long the number is. The label consists of
19792 three parts: the (optional) leading spaces, the number, and the tic
19793 mark. The function is passed the value of the number for the specific
19794 row, and the value of the width of the top line, which is calculated
19795 (just once) by @code{print-graph}.
19799 (defun Y-axis-element (number full-Y-label-width)
19800 "Construct a NUMBERed label element.
19801 A numbered element looks like this ` 5 - ',
19802 and is padded as needed so all line up with
19803 the element for the largest number."
19806 (let* ((leading-spaces
19807 (- full-Y-label-width
19809 (concat (number-to-string number)
19814 (make-string leading-spaces ? )
19815 (number-to-string number)
19820 The @code{Y-axis-element} function concatenates together the leading
19821 spaces, if any; the number, as a string; and the tic mark.
19823 To figure out how many leading spaces the label will need, the
19824 function subtracts the actual length of the label---the length of the
19825 number plus the length of the tic mark---from the desired label width.
19827 @findex make-string
19828 Blank spaces are inserted using the @code{make-string} function. This
19829 function takes two arguments: the first tells it how long the string
19830 will be and the second is a symbol for the character to insert, in a
19831 special format. The format is a question mark followed by a blank
19832 space, like this, @samp{? }. @xref{Character Type, , Character Type,
19833 elisp, The GNU Emacs Lisp Reference Manual}, for a description of the
19834 syntax for characters. (Of course, you might want to replace the
19835 blank space by some other character @dots{} You know what to do.)
19837 The @code{number-to-string} function is used in the concatenation
19838 expression, to convert the number to a string that is concatenated
19839 with the leading spaces and the tic mark.
19841 @node Y-axis-column
19842 @appendixsubsec Create a Y Axis Column
19844 The preceding functions provide all the tools needed to construct a
19845 function that generates a list of numbered and blank strings to insert
19846 as the label for the vertical axis:
19848 @findex Y-axis-column
19851 (defun Y-axis-column (height width-of-label)
19852 "Construct list of Y axis labels and blank strings.
19853 For HEIGHT of line above base and WIDTH-OF-LABEL."
19857 (while (> height 1)
19858 (if (zerop (% height Y-axis-label-spacing))
19859 ;; @r{Insert label.}
19862 (Y-axis-element height width-of-label)
19866 ;; @r{Else, insert blanks.}
19869 (make-string width-of-label ? )
19871 (setq height (1- height)))
19872 ;; @r{Insert base line.}
19874 (cons (Y-axis-element 1 width-of-label) Y-axis))
19875 (nreverse Y-axis)))
19879 In this function, we start with the value of @code{height} and
19880 repetitively subtract one from its value. After each subtraction, we
19881 test to see whether the value is an integral multiple of the
19882 @code{Y-axis-label-spacing}. If it is, we construct a numbered label
19883 using the @code{Y-axis-element} function; if not, we construct a
19884 blank label using the @code{make-string} function. The base line
19885 consists of the number one followed by a tic mark.
19888 @node print-Y-axis Penultimate
19889 @appendixsubsec The Not Quite Final Version of @code{print-Y-axis}
19891 The list constructed by the @code{Y-axis-column} function is passed to
19892 the @code{print-Y-axis} function, which inserts the list as a column.
19894 @findex print-Y-axis
19897 (defun print-Y-axis (height full-Y-label-width)
19898 "Insert Y axis using HEIGHT and FULL-Y-LABEL-WIDTH.
19899 Height must be the maximum height of the graph.
19900 Full width is the width of the highest label element."
19901 ;; Value of height and full-Y-label-width
19902 ;; are passed by print-graph.
19905 (let ((start (point)))
19907 (Y-axis-column height full-Y-label-width))
19908 ;; @r{Place point ready for inserting graph.}
19910 ;; @r{Move point forward by value of} full-Y-label-width
19911 (forward-char full-Y-label-width)))
19915 The @code{print-Y-axis} uses the @code{insert-rectangle} function to
19916 insert the Y axis labels created by the @code{Y-axis-column} function.
19917 In addition, it places point at the correct position for printing the body of
19920 You can test @code{print-Y-axis}:
19928 Y-axis-label-spacing
19937 Copy the following expression:
19940 (print-Y-axis 12 5)
19944 Switch to the @file{*scratch*} buffer and place the cursor where you
19945 want the axis labels to start.
19948 Type @kbd{M-:} (@code{eval-expression}).
19951 Yank the @code{graph-body-print} expression into the minibuffer
19952 with @kbd{C-y} (@code{yank)}.
19955 Press @key{RET} to evaluate the expression.
19958 Emacs will print labels vertically, the top one being @w{@samp{10 -@w{
19959 }}}. (The @code{print-graph} function will pass the value of
19960 @code{height-of-top-line}, which in this case will end up as 15,
19961 thereby getting rid of what might appear as a bug.)
19965 @appendixsec The @code{print-X-axis} Function
19966 @cindex Axis, print horizontal
19967 @cindex X axis printing
19968 @cindex Print horizontal axis
19969 @cindex Horizontal axis printing
19971 X axis labels are much like Y axis labels, except that the ticks are on a
19972 line above the numbers. Labels should look like this:
19981 The first tic is under the first column of the graph and is preceded by
19982 several blank spaces. These spaces provide room in rows above for the Y
19983 axis labels. The second, third, fourth, and subsequent ticks are all
19984 spaced equally, according to the value of @code{X-axis-label-spacing}.
19986 The second row of the X axis consists of numbers, preceded by several
19987 blank spaces and also separated according to the value of the variable
19988 @code{X-axis-label-spacing}.
19990 The value of the variable @code{X-axis-label-spacing} should itself be
19991 measured in units of @code{symbol-width}, since you may want to change
19992 the width of the symbols that you are using to print the body of the
19993 graph without changing the ways the graph is labeled.
19996 * Similarities differences:: Much like @code{print-Y-axis}, but not exactly.
19997 * X Axis Tic Marks:: Create tic marks for the horizontal axis.
20001 @node Similarities differences
20002 @unnumberedsubsec Similarities and differences
20005 The @code{print-X-axis} function is constructed in more or less the
20006 same fashion as the @code{print-Y-axis} function except that it has
20007 two lines: the line of tic marks and the numbers. We will write a
20008 separate function to print each line and then combine them within the
20009 @code{print-X-axis} function.
20011 This is a three step process:
20015 Write a function to print the X axis tic marks, @code{print-X-axis-tic-line}.
20018 Write a function to print the X numbers, @code{print-X-axis-numbered-line}.
20021 Write a function to print both lines, the @code{print-X-axis} function,
20022 using @code{print-X-axis-tic-line} and
20023 @code{print-X-axis-numbered-line}.
20026 @node X Axis Tic Marks
20027 @appendixsubsec X Axis Tic Marks
20029 The first function should print the X axis tic marks. We must specify
20030 the tic marks themselves and their spacing:
20034 (defvar X-axis-label-spacing
20035 (if (boundp 'graph-blank)
20036 (* 5 (length graph-blank)) 5)
20037 "Number of units from one X axis label to next.")
20042 (Note that the value of @code{graph-blank} is set by another
20043 @code{defvar}. The @code{boundp} predicate checks whether it has
20044 already been set; @code{boundp} returns @code{nil} if it has not. If
20045 @code{graph-blank} were unbound and we did not use this conditional
20046 construction, in a recent GNU Emacs, we would enter the debugger and
20047 see an error message saying @samp{@w{Debugger entered--Lisp error:}
20048 @w{(void-variable graph-blank)}}.)
20051 Here is the @code{defvar} for @code{X-axis-tic-symbol}:
20055 (defvar X-axis-tic-symbol "|"
20056 "String to insert to point to a column in X axis.")
20061 The goal is to make a line that looks like this:
20067 The first tic is indented so that it is under the first column, which is
20068 indented to provide space for the Y axis labels.
20070 A tic element consists of the blank spaces that stretch from one tic to
20071 the next plus a tic symbol. The number of blanks is determined by the
20072 width of the tic symbol and the @code{X-axis-label-spacing}.
20075 The code looks like this:
20079 ;;; X-axis-tic-element
20083 ;; @r{Make a string of blanks.}
20084 (- (* symbol-width X-axis-label-spacing)
20085 (length X-axis-tic-symbol))
20087 ;; @r{Concatenate blanks with tic symbol.}
20093 Next, we determine how many blanks are needed to indent the first tic
20094 mark to the first column of the graph. This uses the value of
20095 @code{full-Y-label-width} passed it by the @code{print-graph} function.
20098 The code to make @code{X-axis-leading-spaces}
20103 ;; X-axis-leading-spaces
20105 (make-string full-Y-label-width ? )
20110 We also need to determine the length of the horizontal axis, which is
20111 the length of the numbers list, and the number of ticks in the horizontal
20118 (length numbers-list)
20124 (* symbol-width X-axis-label-spacing)
20128 ;; number-of-X-ticks
20129 (if (zerop (% (X-length tic-width)))
20130 (/ (X-length tic-width))
20131 (1+ (/ (X-length tic-width))))
20136 All this leads us directly to the function for printing the X axis tic line:
20138 @findex print-X-axis-tic-line
20141 (defun print-X-axis-tic-line
20142 (number-of-X-tics X-axis-leading-spaces X-axis-tic-element)
20143 "Print ticks for X axis."
20144 (insert X-axis-leading-spaces)
20145 (insert X-axis-tic-symbol) ; @r{Under first column.}
20148 ;; @r{Insert second tic in the right spot.}
20151 (- (* symbol-width X-axis-label-spacing)
20152 ;; @r{Insert white space up to second tic symbol.}
20153 (* 2 (length X-axis-tic-symbol)))
20155 X-axis-tic-symbol))
20158 ;; @r{Insert remaining ticks.}
20159 (while (> number-of-X-tics 1)
20160 (insert X-axis-tic-element)
20161 (setq number-of-X-tics (1- number-of-X-tics))))
20165 The line of numbers is equally straightforward:
20168 First, we create a numbered element with blank spaces before each number:
20170 @findex X-axis-element
20173 (defun X-axis-element (number)
20174 "Construct a numbered X axis element."
20175 (let ((leading-spaces
20176 (- (* symbol-width X-axis-label-spacing)
20177 (length (number-to-string number)))))
20178 (concat (make-string leading-spaces ? )
20179 (number-to-string number))))
20183 Next, we create the function to print the numbered line, starting with
20184 the number 1 under the first column:
20186 @findex print-X-axis-numbered-line
20189 (defun print-X-axis-numbered-line
20190 (number-of-X-tics X-axis-leading-spaces)
20191 "Print line of X-axis numbers"
20192 (let ((number X-axis-label-spacing))
20193 (insert X-axis-leading-spaces)
20199 ;; @r{Insert white space up to next number.}
20200 (- (* symbol-width X-axis-label-spacing) 2)
20202 (number-to-string number)))
20205 ;; @r{Insert remaining numbers.}
20206 (setq number (+ number X-axis-label-spacing))
20207 (while (> number-of-X-tics 1)
20208 (insert (X-axis-element number))
20209 (setq number (+ number X-axis-label-spacing))
20210 (setq number-of-X-tics (1- number-of-X-tics)))))
20214 Finally, we need to write the @code{print-X-axis} that uses
20215 @code{print-X-axis-tic-line} and
20216 @code{print-X-axis-numbered-line}.
20218 The function must determine the local values of the variables used by both
20219 @code{print-X-axis-tic-line} and @code{print-X-axis-numbered-line}, and
20220 then it must call them. Also, it must print the carriage return that
20221 separates the two lines.
20223 The function consists of a varlist that specifies five local variables,
20224 and calls to each of the two line printing functions:
20226 @findex print-X-axis
20229 (defun print-X-axis (numbers-list)
20230 "Print X axis labels to length of NUMBERS-LIST."
20231 (let* ((leading-spaces
20232 (make-string full-Y-label-width ? ))
20235 ;; symbol-width @r{is provided by} graph-body-print
20236 (tic-width (* symbol-width X-axis-label-spacing))
20237 (X-length (length numbers-list))
20245 ;; @r{Make a string of blanks.}
20246 (- (* symbol-width X-axis-label-spacing)
20247 (length X-axis-tic-symbol))
20251 ;; @r{Concatenate blanks with tic symbol.}
20252 X-axis-tic-symbol))
20256 (if (zerop (% X-length tic-width))
20257 (/ X-length tic-width)
20258 (1+ (/ X-length tic-width)))))
20261 (print-X-axis-tic-line tic-number leading-spaces X-tic)
20263 (print-X-axis-numbered-line tic-number leading-spaces)))
20268 You can test @code{print-X-axis}:
20272 Install @code{X-axis-tic-symbol}, @code{X-axis-label-spacing},
20273 @code{print-X-axis-tic-line}, as well as @code{X-axis-element},
20274 @code{print-X-axis-numbered-line}, and @code{print-X-axis}.
20277 Copy the following expression:
20282 (let ((full-Y-label-width 5)
20285 '(1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16))))
20290 Switch to the @file{*scratch*} buffer and place the cursor where you
20291 want the axis labels to start.
20294 Type @kbd{M-:} (@code{eval-expression}).
20297 Yank the test expression into the minibuffer
20298 with @kbd{C-y} (@code{yank)}.
20301 Press @key{RET} to evaluate the expression.
20305 Emacs will print the horizontal axis like this:
20315 @node Print Whole Graph
20316 @appendixsec Printing the Whole Graph
20317 @cindex Printing the whole graph
20318 @cindex Whole graph printing
20319 @cindex Graph, printing all
20321 Now we are nearly ready to print the whole graph.
20323 The function to print the graph with the proper labels follows the
20324 outline we created earlier (@pxref{Full Graph, , A Graph with Labeled
20325 Axes}), but with additions.
20328 Here is the outline:
20332 (defun print-graph (numbers-list)
20333 "@var{documentation}@dots{}"
20334 (let ((height @dots{}
20338 (print-Y-axis height @dots{} )
20339 (graph-body-print numbers-list)
20340 (print-X-axis @dots{} )))
20345 * The final version:: A few changes.
20346 * Test print-graph:: Run a short test.
20347 * Graphing words in defuns:: Executing the final code.
20348 * lambda:: How to write an anonymous function.
20349 * mapcar:: Apply a function to elements of a list.
20350 * Another Bug:: Yet another bug @dots{} most insidious.
20351 * Final printed graph:: The graph itself!
20355 @node The final version
20356 @unnumberedsubsec Changes for the Final Version
20359 The final version is different from what we planned in two ways:
20360 first, it contains additional values calculated once in the varlist;
20361 second, it carries an option to specify the labels' increment per row.
20362 This latter feature turns out to be essential; otherwise, a graph may
20363 have more rows than fit on a display or on a sheet of paper.
20366 This new feature requires a change to the @code{Y-axis-column}
20367 function, to add @code{vertical-step} to it. The function looks like
20370 @findex Y-axis-column @r{Final version.}
20373 ;;; @r{Final version.}
20374 (defun Y-axis-column
20375 (height width-of-label &optional vertical-step)
20376 "Construct list of labels for Y axis.
20377 HEIGHT is maximum height of graph.
20378 WIDTH-OF-LABEL is maximum width of label.
20379 VERTICAL-STEP, an option, is a positive integer
20380 that specifies how much a Y axis label increments
20381 for each line. For example, a step of 5 means
20382 that each line is five units of the graph."
20386 (number-per-line (or vertical-step 1)))
20387 (while (> height 1)
20388 (if (zerop (% height Y-axis-label-spacing))
20391 ;; @r{Insert label.}
20395 (* height number-per-line)
20400 ;; @r{Else, insert blanks.}
20403 (make-string width-of-label ? )
20405 (setq height (1- height)))
20408 ;; @r{Insert base line.}
20409 (setq Y-axis (cons (Y-axis-element
20410 (or vertical-step 1)
20413 (nreverse Y-axis)))
20417 The values for the maximum height of graph and the width of a symbol
20418 are computed by @code{print-graph} in its @code{let} expression; so
20419 @code{graph-body-print} must be changed to accept them.
20421 @findex graph-body-print @r{Final version.}
20424 ;;; @r{Final version.}
20425 (defun graph-body-print (numbers-list height symbol-width)
20426 "Print a bar graph of the NUMBERS-LIST.
20427 The numbers-list consists of the Y-axis values.
20428 HEIGHT is maximum height of graph.
20429 SYMBOL-WIDTH is number of each column."
20432 (let (from-position)
20433 (while numbers-list
20434 (setq from-position (point))
20436 (column-of-graph height (car numbers-list)))
20437 (goto-char from-position)
20438 (forward-char symbol-width)
20441 ;; @r{Draw graph column by column.}
20443 (setq numbers-list (cdr numbers-list)))
20444 ;; @r{Place point for X axis labels.}
20445 (forward-line height)
20451 Finally, the code for the @code{print-graph} function:
20453 @findex print-graph @r{Final version.}
20456 ;;; @r{Final version.}
20458 (numbers-list &optional vertical-step)
20459 "Print labeled bar graph of the NUMBERS-LIST.
20460 The numbers-list consists of the Y-axis values.
20464 Optionally, VERTICAL-STEP, a positive integer,
20465 specifies how much a Y axis label increments for
20466 each line. For example, a step of 5 means that
20467 each row is five units."
20470 (let* ((symbol-width (length graph-blank))
20471 ;; @code{height} @r{is both the largest number}
20472 ;; @r{and the number with the most digits.}
20473 (height (apply 'max numbers-list))
20476 (height-of-top-line
20477 (if (zerop (% height Y-axis-label-spacing))
20480 (* (1+ (/ height Y-axis-label-spacing))
20481 Y-axis-label-spacing)))
20484 (vertical-step (or vertical-step 1))
20485 (full-Y-label-width
20491 (* height-of-top-line vertical-step))
20497 height-of-top-line full-Y-label-width vertical-step)
20501 numbers-list height-of-top-line symbol-width)
20502 (print-X-axis numbers-list)))
20506 @node Test print-graph
20507 @appendixsubsec Testing @code{print-graph}
20510 We can test the @code{print-graph} function with a short list of numbers:
20514 Install the final versions of @code{Y-axis-column},
20515 @code{graph-body-print}, and @code{print-graph} (in addition to the
20519 Copy the following expression:
20522 (print-graph '(3 2 5 6 7 5 3 4 6 4 3 2 1))
20526 Switch to the @file{*scratch*} buffer and place the cursor where you
20527 want the axis labels to start.
20530 Type @kbd{M-:} (@code{eval-expression}).
20533 Yank the test expression into the minibuffer
20534 with @kbd{C-y} (@code{yank)}.
20537 Press @key{RET} to evaluate the expression.
20541 Emacs will print a graph that looks like this:
20562 On the other hand, if you pass @code{print-graph} a
20563 @code{vertical-step} value of 2, by evaluating this expression:
20566 (print-graph '(3 2 5 6 7 5 3 4 6 4 3 2 1) 2)
20571 The graph looks like this:
20592 (A question: is the @samp{2} on the bottom of the vertical axis a bug or a
20593 feature? If you think it is a bug, and should be a @samp{1} instead, (or
20594 even a @samp{0}), you can modify the sources.)
20596 @node Graphing words in defuns
20597 @appendixsubsec Graphing Numbers of Words and Symbols
20599 Now for the graph for which all this code was written: a graph that
20600 shows how many function definitions contain fewer than 10 words and
20601 symbols, how many contain between 10 and 19 words and symbols, how
20602 many contain between 20 and 29 words and symbols, and so on.
20604 This is a multi-step process. First make sure you have loaded all the
20608 It is a good idea to reset the value of @code{top-of-ranges} in case
20609 you have set it to some different value. You can evaluate the
20614 (setq top-of-ranges
20617 110 120 130 140 150
20618 160 170 180 190 200
20619 210 220 230 240 250
20620 260 270 280 290 300)
20625 Next create a list of the number of words and symbols in each range.
20629 Evaluate the following:
20633 (setq list-for-graph
20636 (recursive-lengths-list-many-files
20637 (directory-files "/usr/local/emacs/lisp"
20645 On my old machine, this took about an hour. It looked though 303 Lisp
20646 files in my copy of Emacs version 19.23. After all that computing,
20647 the @code{list-for-graph} had this value:
20651 (537 1027 955 785 594 483 349 292 224 199 166 120 116 99
20652 90 80 67 48 52 45 41 33 28 26 25 20 12 28 11 13 220)
20657 This means that my copy of Emacs had 537 function definitions with
20658 fewer than 10 words or symbols in them, 1,027 function definitions
20659 with 10 to 19 words or symbols in them, 955 function definitions with
20660 20 to 29 words or symbols in them, and so on.
20662 Clearly, just by looking at this list we can see that most function
20663 definitions contain ten to thirty words and symbols.
20665 Now for printing. We do @emph{not} want to print a graph that is
20666 1,030 lines high @dots{} Instead, we should print a graph that is
20667 fewer than twenty-five lines high. A graph that height can be
20668 displayed on almost any monitor, and easily printed on a sheet of paper.
20670 This means that each value in @code{list-for-graph} must be reduced to
20671 one-fiftieth its present value.
20673 Here is a short function to do just that, using two functions we have
20674 not yet seen, @code{mapcar} and @code{lambda}.
20678 (defun one-fiftieth (full-range)
20679 "Return list, each number one-fiftieth of previous."
20680 (mapcar (lambda (arg) (/ arg 50)) full-range))
20685 @appendixsubsec A @code{lambda} Expression: Useful Anonymity
20686 @cindex Anonymous function
20689 @code{lambda} is the symbol for an anonymous function, a function
20690 without a name. Every time you use an anonymous function, you need to
20691 include its whole body.
20698 (lambda (arg) (/ arg 50))
20702 is a function that returns the value resulting from
20703 dividing whatever is passed to it as @code{arg} by 50.
20706 Earlier, for example, we had a function @code{multiply-by-seven}; it
20707 multiplied its argument by 7. This function is similar, except it
20708 divides its argument by 50; and, it has no name. The anonymous
20709 equivalent of @code{multiply-by-seven} is:
20712 (lambda (number) (* 7 number))
20716 (@xref{defun, , The @code{defun} Macro}.)
20720 If we want to multiply 3 by 7, we can write:
20722 @c clear print-postscript-figures
20723 @c lambda example diagram #1
20727 (multiply-by-seven 3)
20728 \_______________/ ^
20734 @ifset print-postscript-figures
20737 @center @image{lambda-1}
20741 @ifclear print-postscript-figures
20745 (multiply-by-seven 3)
20746 \_______________/ ^
20755 This expression returns 21.
20759 Similarly, we can write:
20761 @c lambda example diagram #2
20765 ((lambda (number) (* 7 number)) 3)
20766 \____________________________/ ^
20768 anonymous function argument
20772 @ifset print-postscript-figures
20775 @center @image{lambda-2}
20779 @ifclear print-postscript-figures
20783 ((lambda (number) (* 7 number)) 3)
20784 \____________________________/ ^
20786 anonymous function argument
20794 If we want to divide 100 by 50, we can write:
20796 @c lambda example diagram #3
20800 ((lambda (arg) (/ arg 50)) 100)
20801 \______________________/ \_/
20803 anonymous function argument
20807 @ifset print-postscript-figures
20810 @center @image{lambda-3}
20814 @ifclear print-postscript-figures
20818 ((lambda (arg) (/ arg 50)) 100)
20819 \______________________/ \_/
20821 anonymous function argument
20828 This expression returns 2. The 100 is passed to the function, which
20829 divides that number by 50.
20831 @xref{Lambda Expressions, , Lambda Expressions, elisp, The GNU Emacs
20832 Lisp Reference Manual}, for more about @code{lambda}. Lisp and lambda
20833 expressions derive from the Lambda Calculus.
20836 @appendixsubsec The @code{mapcar} Function
20839 @code{mapcar} is a function that calls its first argument with each
20840 element of its second argument, in turn. The second argument must be
20843 The @samp{map} part of the name comes from the mathematical phrase,
20844 ``mapping over a domain'', meaning to apply a function to each of the
20845 elements in a domain. The mathematical phrase is based on the
20846 metaphor of a surveyor walking, one step at a time, over an area he is
20847 mapping. And @samp{car}, of course, comes from the Lisp notion of the
20856 (mapcar '1+ '(2 4 6))
20862 The function @code{1+} which adds one to its argument, is executed on
20863 @emph{each} element of the list, and a new list is returned.
20865 Contrast this with @code{apply}, which applies its first argument to
20867 (@xref{Readying a Graph, , Readying a Graph}, for an explanation of
20871 In the definition of @code{one-fiftieth}, the first argument is the
20872 anonymous function:
20875 (lambda (arg) (/ arg 50))
20879 and the second argument is @code{full-range}, which will be bound to
20880 @code{list-for-graph}.
20883 The whole expression looks like this:
20886 (mapcar (lambda (arg) (/ arg 50)) full-range))
20889 @xref{Mapping Functions, , Mapping Functions, elisp, The GNU Emacs
20890 Lisp Reference Manual}, for more about @code{mapcar}.
20892 Using the @code{one-fiftieth} function, we can generate a list in
20893 which each element is one-fiftieth the size of the corresponding
20894 element in @code{list-for-graph}.
20898 (setq fiftieth-list-for-graph
20899 (one-fiftieth list-for-graph))
20904 The resulting list looks like this:
20908 (10 20 19 15 11 9 6 5 4 3 3 2 2
20909 1 1 1 1 0 1 0 0 0 0 0 0 0 0 0 0 0 4)
20914 This, we are almost ready to print! (We also notice the loss of
20915 information: many of the higher ranges are 0, meaning that fewer than
20916 50 defuns had that many words or symbols---but not necessarily meaning
20917 that none had that many words or symbols.)
20920 @appendixsubsec Another Bug @dots{} Most Insidious
20921 @cindex Bug, most insidious type
20922 @cindex Insidious type of bug
20924 I said ``almost ready to print''! Of course, there is a bug in the
20925 @code{print-graph} function @dots{} It has a @code{vertical-step}
20926 option, but not a @code{horizontal-step} option. The
20927 @code{top-of-range} scale goes from 10 to 300 by tens. But the
20928 @code{print-graph} function will print only by ones.
20930 This is a classic example of what some consider the most insidious
20931 type of bug, the bug of omission. This is not the kind of bug you can
20932 find by studying the code, for it is not in the code; it is an omitted
20933 feature. Your best actions are to try your program early and often;
20934 and try to arrange, as much as you can, to write code that is easy to
20935 understand and easy to change. Try to be aware, whenever you can,
20936 that whatever you have written, @emph{will} be rewritten, if not soon,
20937 eventually. A hard maxim to follow.
20939 It is the @code{print-X-axis-numbered-line} function that needs the
20940 work; and then the @code{print-X-axis} and the @code{print-graph}
20941 functions need to be adapted. Not much needs to be done; there is one
20942 nicety: the numbers ought to line up under the tic marks. This takes
20946 Here is the corrected @code{print-X-axis-numbered-line}:
20950 (defun print-X-axis-numbered-line
20951 (number-of-X-tics X-axis-leading-spaces
20952 &optional horizontal-step)
20953 "Print line of X-axis numbers"
20954 (let ((number X-axis-label-spacing)
20955 (horizontal-step (or horizontal-step 1)))
20958 (insert X-axis-leading-spaces)
20959 ;; @r{Delete extra leading spaces.}
20962 (length (number-to-string horizontal-step)))))
20967 ;; @r{Insert white space.}
20969 X-axis-label-spacing)
20972 (number-to-string horizontal-step)))
20976 (* number horizontal-step))))
20979 ;; @r{Insert remaining numbers.}
20980 (setq number (+ number X-axis-label-spacing))
20981 (while (> number-of-X-tics 1)
20982 (insert (X-axis-element
20983 (* number horizontal-step)))
20984 (setq number (+ number X-axis-label-spacing))
20985 (setq number-of-X-tics (1- number-of-X-tics)))))
20990 If you are reading this in Info, you can see the new versions of
20991 @code{print-X-axis} @code{print-graph} and evaluate them. If you are
20992 reading this in a printed book, you can see the changed lines here
20993 (the full text is too much to print).
20998 (defun print-X-axis (numbers-list horizontal-step)
21000 (print-X-axis-numbered-line
21001 tic-number leading-spaces horizontal-step))
21009 &optional vertical-step horizontal-step)
21011 (print-X-axis numbers-list horizontal-step))
21019 (defun print-X-axis (numbers-list horizontal-step)
21020 "Print X axis labels to length of NUMBERS-LIST.
21021 Optionally, HORIZONTAL-STEP, a positive integer,
21022 specifies how much an X axis label increments for
21026 ;; Value of symbol-width and full-Y-label-width
21027 ;; are passed by print-graph.
21028 (let* ((leading-spaces
21029 (make-string full-Y-label-width ? ))
21030 ;; symbol-width @r{is provided by} graph-body-print
21031 (tic-width (* symbol-width X-axis-label-spacing))
21032 (X-length (length numbers-list))
21038 ;; @r{Make a string of blanks.}
21039 (- (* symbol-width X-axis-label-spacing)
21040 (length X-axis-tic-symbol))
21044 ;; @r{Concatenate blanks with tic symbol.}
21045 X-axis-tic-symbol))
21047 (if (zerop (% X-length tic-width))
21048 (/ X-length tic-width)
21049 (1+ (/ X-length tic-width)))))
21053 (print-X-axis-tic-line
21054 tic-number leading-spaces X-tic)
21056 (print-X-axis-numbered-line
21057 tic-number leading-spaces horizontal-step)))
21064 (numbers-list &optional vertical-step horizontal-step)
21065 "Print labeled bar graph of the NUMBERS-LIST.
21066 The numbers-list consists of the Y-axis values.
21070 Optionally, VERTICAL-STEP, a positive integer,
21071 specifies how much a Y axis label increments for
21072 each line. For example, a step of 5 means that
21073 each row is five units.
21077 Optionally, HORIZONTAL-STEP, a positive integer,
21078 specifies how much an X axis label increments for
21080 (let* ((symbol-width (length graph-blank))
21081 ;; @code{height} @r{is both the largest number}
21082 ;; @r{and the number with the most digits.}
21083 (height (apply 'max numbers-list))
21086 (height-of-top-line
21087 (if (zerop (% height Y-axis-label-spacing))
21090 (* (1+ (/ height Y-axis-label-spacing))
21091 Y-axis-label-spacing)))
21094 (vertical-step (or vertical-step 1))
21095 (full-Y-label-width
21099 (* height-of-top-line vertical-step))
21104 height-of-top-line full-Y-label-width vertical-step)
21106 numbers-list height-of-top-line symbol-width)
21107 (print-X-axis numbers-list horizontal-step)))
21114 Graphing Definitions Re-listed
21117 Here are all the graphing definitions in their final form:
21121 (defvar top-of-ranges
21124 110 120 130 140 150
21125 160 170 180 190 200
21126 210 220 230 240 250)
21127 "List specifying ranges for `defuns-per-range'.")
21131 (defvar graph-symbol "*"
21132 "String used as symbol in graph, usually an asterisk.")
21136 (defvar graph-blank " "
21137 "String used as blank in graph, usually a blank space.
21138 graph-blank must be the same number of columns wide
21143 (defvar Y-axis-tic " - "
21144 "String that follows number in a Y axis label.")
21148 (defvar Y-axis-label-spacing 5
21149 "Number of lines from one Y axis label to next.")
21153 (defvar X-axis-tic-symbol "|"
21154 "String to insert to point to a column in X axis.")
21158 (defvar X-axis-label-spacing
21159 (if (boundp 'graph-blank)
21160 (* 5 (length graph-blank)) 5)
21161 "Number of units from one X axis label to next.")
21167 (defun count-words-in-defun ()
21168 "Return the number of words and symbols in a defun."
21169 (beginning-of-defun)
21171 (end (save-excursion (end-of-defun) (point))))
21176 (and (< (point) end)
21178 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*"
21180 (setq count (1+ count)))
21187 (defun lengths-list-file (filename)
21188 "Return list of definitions' lengths within FILE.
21189 The returned list is a list of numbers.
21190 Each number is the number of words or
21191 symbols in one function definition."
21195 (message "Working on `%s' ... " filename)
21197 (let ((buffer (find-file-noselect filename))
21199 (set-buffer buffer)
21200 (setq buffer-read-only t)
21202 (goto-char (point-min))
21206 (while (re-search-forward "^(defun" nil t)
21208 (cons (count-words-in-defun) lengths-list)))
21209 (kill-buffer buffer)
21216 (defun lengths-list-many-files (list-of-files)
21217 "Return list of lengths of defuns in LIST-OF-FILES."
21218 (let (lengths-list)
21219 ;;; @r{true-or-false-test}
21220 (while list-of-files
21226 ;;; @r{Generate a lengths' list.}
21228 (expand-file-name (car list-of-files)))))
21229 ;;; @r{Make files' list shorter.}
21230 (setq list-of-files (cdr list-of-files)))
21231 ;;; @r{Return final value of lengths' list.}
21238 (defun defuns-per-range (sorted-lengths top-of-ranges)
21239 "SORTED-LENGTHS defuns in each TOP-OF-RANGES range."
21240 (let ((top-of-range (car top-of-ranges))
21241 (number-within-range 0)
21242 defuns-per-range-list)
21247 (while top-of-ranges
21251 ;; @r{Need number for numeric test.}
21252 (car sorted-lengths)
21253 (< (car sorted-lengths) top-of-range))
21255 ;; @r{Count number of definitions within current range.}
21256 (setq number-within-range (1+ number-within-range))
21257 (setq sorted-lengths (cdr sorted-lengths)))
21261 ;; @r{Exit inner loop but remain within outer loop.}
21263 (setq defuns-per-range-list
21264 (cons number-within-range defuns-per-range-list))
21265 (setq number-within-range 0) ; @r{Reset count to zero.}
21267 ;; @r{Move to next range.}
21268 (setq top-of-ranges (cdr top-of-ranges))
21269 ;; @r{Specify next top of range value.}
21270 (setq top-of-range (car top-of-ranges)))
21274 ;; @r{Exit outer loop and count the number of defuns larger than}
21275 ;; @r{ the largest top-of-range value.}
21276 (setq defuns-per-range-list
21278 (length sorted-lengths)
21279 defuns-per-range-list))
21281 ;; @r{Return a list of the number of definitions within each range,}
21282 ;; @r{ smallest to largest.}
21283 (nreverse defuns-per-range-list)))
21289 (defun column-of-graph (max-graph-height actual-height)
21290 "Return list of MAX-GRAPH-HEIGHT strings;
21291 ACTUAL-HEIGHT are graph-symbols.
21292 The graph-symbols are contiguous entries at the end
21294 The list will be inserted as one column of a graph.
21295 The strings are either graph-blank or graph-symbol."
21299 (let ((insert-list nil)
21300 (number-of-top-blanks
21301 (- max-graph-height actual-height)))
21303 ;; @r{Fill in @code{graph-symbols}.}
21304 (while (> actual-height 0)
21305 (setq insert-list (cons graph-symbol insert-list))
21306 (setq actual-height (1- actual-height)))
21310 ;; @r{Fill in @code{graph-blanks}.}
21311 (while (> number-of-top-blanks 0)
21312 (setq insert-list (cons graph-blank insert-list))
21313 (setq number-of-top-blanks
21314 (1- number-of-top-blanks)))
21316 ;; @r{Return whole list.}
21323 (defun Y-axis-element (number full-Y-label-width)
21324 "Construct a NUMBERed label element.
21325 A numbered element looks like this ` 5 - ',
21326 and is padded as needed so all line up with
21327 the element for the largest number."
21330 (let* ((leading-spaces
21331 (- full-Y-label-width
21333 (concat (number-to-string number)
21338 (make-string leading-spaces ? )
21339 (number-to-string number)
21346 (defun print-Y-axis
21347 (height full-Y-label-width &optional vertical-step)
21348 "Insert Y axis by HEIGHT and FULL-Y-LABEL-WIDTH.
21349 Height must be the maximum height of the graph.
21350 Full width is the width of the highest label element.
21351 Optionally, print according to VERTICAL-STEP."
21354 ;; Value of height and full-Y-label-width
21355 ;; are passed by 'print-graph'.
21356 (let ((start (point)))
21358 (Y-axis-column height full-Y-label-width vertical-step))
21361 ;; @r{Place point ready for inserting graph.}
21363 ;; @r{Move point forward by value of} full-Y-label-width
21364 (forward-char full-Y-label-width)))
21370 (defun print-X-axis-tic-line
21371 (number-of-X-tics X-axis-leading-spaces X-axis-tic-element)
21372 "Print ticks for X axis."
21373 (insert X-axis-leading-spaces)
21374 (insert X-axis-tic-symbol) ; @r{Under first column.}
21377 ;; @r{Insert second tic in the right spot.}
21380 (- (* symbol-width X-axis-label-spacing)
21381 ;; @r{Insert white space up to second tic symbol.}
21382 (* 2 (length X-axis-tic-symbol)))
21384 X-axis-tic-symbol))
21387 ;; @r{Insert remaining ticks.}
21388 (while (> number-of-X-tics 1)
21389 (insert X-axis-tic-element)
21390 (setq number-of-X-tics (1- number-of-X-tics))))
21396 (defun X-axis-element (number)
21397 "Construct a numbered X axis element."
21398 (let ((leading-spaces
21399 (- (* symbol-width X-axis-label-spacing)
21400 (length (number-to-string number)))))
21401 (concat (make-string leading-spaces ? )
21402 (number-to-string number))))
21408 (defun graph-body-print (numbers-list height symbol-width)
21409 "Print a bar graph of the NUMBERS-LIST.
21410 The numbers-list consists of the Y-axis values.
21411 HEIGHT is maximum height of graph.
21412 SYMBOL-WIDTH is number of each column."
21415 (let (from-position)
21416 (while numbers-list
21417 (setq from-position (point))
21419 (column-of-graph height (car numbers-list)))
21420 (goto-char from-position)
21421 (forward-char symbol-width)
21424 ;; @r{Draw graph column by column.}
21426 (setq numbers-list (cdr numbers-list)))
21427 ;; @r{Place point for X axis labels.}
21428 (forward-line height)
21435 (defun Y-axis-column
21436 (height width-of-label &optional vertical-step)
21437 "Construct list of labels for Y axis.
21438 HEIGHT is maximum height of graph.
21439 WIDTH-OF-LABEL is maximum width of label.
21442 VERTICAL-STEP, an option, is a positive integer
21443 that specifies how much a Y axis label increments
21444 for each line. For example, a step of 5 means
21445 that each line is five units of the graph."
21447 (number-per-line (or vertical-step 1)))
21450 (while (> height 1)
21451 (if (zerop (% height Y-axis-label-spacing))
21452 ;; @r{Insert label.}
21456 (* height number-per-line)
21461 ;; @r{Else, insert blanks.}
21464 (make-string width-of-label ? )
21466 (setq height (1- height)))
21469 ;; @r{Insert base line.}
21470 (setq Y-axis (cons (Y-axis-element
21471 (or vertical-step 1)
21474 (nreverse Y-axis)))
21480 (defun print-X-axis-numbered-line
21481 (number-of-X-tics X-axis-leading-spaces
21482 &optional horizontal-step)
21483 "Print line of X-axis numbers"
21484 (let ((number X-axis-label-spacing)
21485 (horizontal-step (or horizontal-step 1)))
21488 (insert X-axis-leading-spaces)
21490 (delete-char (- (1- (length (number-to-string horizontal-step)))))
21493 ;; @r{Insert white space up to next number.}
21494 (- (* symbol-width X-axis-label-spacing)
21495 (1- (length (number-to-string horizontal-step)))
21498 (number-to-string (* number horizontal-step))))
21501 ;; @r{Insert remaining numbers.}
21502 (setq number (+ number X-axis-label-spacing))
21503 (while (> number-of-X-tics 1)
21504 (insert (X-axis-element (* number horizontal-step)))
21505 (setq number (+ number X-axis-label-spacing))
21506 (setq number-of-X-tics (1- number-of-X-tics)))))
21512 (defun print-X-axis (numbers-list horizontal-step)
21513 "Print X axis labels to length of NUMBERS-LIST.
21514 Optionally, HORIZONTAL-STEP, a positive integer,
21515 specifies how much an X axis label increments for
21519 ;; Value of symbol-width and full-Y-label-width
21520 ;; are passed by 'print-graph'.
21521 (let* ((leading-spaces
21522 (make-string full-Y-label-width ? ))
21523 ;; symbol-width @r{is provided by} graph-body-print
21524 (tic-width (* symbol-width X-axis-label-spacing))
21525 (X-length (length numbers-list))
21531 ;; @r{Make a string of blanks.}
21532 (- (* symbol-width X-axis-label-spacing)
21533 (length X-axis-tic-symbol))
21537 ;; @r{Concatenate blanks with tic symbol.}
21538 X-axis-tic-symbol))
21540 (if (zerop (% X-length tic-width))
21541 (/ X-length tic-width)
21542 (1+ (/ X-length tic-width)))))
21546 (print-X-axis-tic-line
21547 tic-number leading-spaces X-tic)
21549 (print-X-axis-numbered-line
21550 tic-number leading-spaces horizontal-step)))
21556 (defun one-fiftieth (full-range)
21557 "Return list, each number of which is 1/50th previous."
21558 (mapcar (lambda (arg) (/ arg 50)) full-range))
21565 (numbers-list &optional vertical-step horizontal-step)
21566 "Print labeled bar graph of the NUMBERS-LIST.
21567 The numbers-list consists of the Y-axis values.
21571 Optionally, VERTICAL-STEP, a positive integer,
21572 specifies how much a Y axis label increments for
21573 each line. For example, a step of 5 means that
21574 each row is five units.
21578 Optionally, HORIZONTAL-STEP, a positive integer,
21579 specifies how much an X axis label increments for
21581 (let* ((symbol-width (length graph-blank))
21582 ;; @code{height} @r{is both the largest number}
21583 ;; @r{and the number with the most digits.}
21584 (height (apply 'max numbers-list))
21587 (height-of-top-line
21588 (if (zerop (% height Y-axis-label-spacing))
21591 (* (1+ (/ height Y-axis-label-spacing))
21592 Y-axis-label-spacing)))
21595 (vertical-step (or vertical-step 1))
21596 (full-Y-label-width
21600 (* height-of-top-line vertical-step))
21606 height-of-top-line full-Y-label-width vertical-step)
21608 numbers-list height-of-top-line symbol-width)
21609 (print-X-axis numbers-list horizontal-step)))
21616 @node Final printed graph
21617 @appendixsubsec The Printed Graph
21619 When made and installed, you can call the @code{print-graph} command
21625 (print-graph fiftieth-list-for-graph 50 10)
21655 50 - ***************** * *
21657 10 50 100 150 200 250 300 350
21664 The largest group of functions contain 10--19 words and symbols each.
21666 @node Free Software and Free Manuals
21667 @appendix Free Software and Free Manuals
21669 @strong{by Richard M. Stallman}
21672 The biggest deficiency in free operating systems is not in the
21673 software---it is the lack of good free manuals that we can include in
21674 these systems. Many of our most important programs do not come with
21675 full manuals. Documentation is an essential part of any software
21676 package; when an important free software package does not come with a
21677 free manual, that is a major gap. We have many such gaps today.
21679 Once upon a time, many years ago, I thought I would learn Perl. I got
21680 a copy of a free manual, but I found it hard to read. When I asked
21681 Perl users about alternatives, they told me that there were better
21682 introductory manuals---but those were not free.
21684 Why was this? The authors of the good manuals had written them for
21685 O'Reilly Associates, which published them with restrictive terms---no
21686 copying, no modification, source files not available---which exclude
21687 them from the free software community.
21689 That wasn't the first time this sort of thing has happened, and (to
21690 our community's great loss) it was far from the last. Proprietary
21691 manual publishers have enticed a great many authors to restrict their
21692 manuals since then. Many times I have heard a GNU user eagerly tell me
21693 about a manual that he is writing, with which he expects to help the
21694 GNU project---and then had my hopes dashed, as he proceeded to explain
21695 that he had signed a contract with a publisher that would restrict it
21696 so that we cannot use it.
21698 Given that writing good English is a rare skill among programmers, we
21699 can ill afford to lose manuals this way.
21701 Free documentation, like free software, is a matter of freedom, not
21702 price. The problem with these manuals was not that O'Reilly Associates
21703 charged a price for printed copies---that in itself is fine. The Free
21704 Software Foundation @uref{https://shop.fsf.org, sells printed copies} of
21705 free @uref{https://www.gnu.org/doc/doc.html, GNU manuals}, too.
21706 But GNU manuals are available in source code form, while these manuals
21707 are available only on paper. GNU manuals come with permission to copy
21708 and modify; the Perl manuals do not. These restrictions are the
21711 The criterion for a free manual is pretty much the same as for free
21712 software: it is a matter of giving all users certain
21713 freedoms. Redistribution (including commercial redistribution) must be
21714 permitted, so that the manual can accompany every copy of the program,
21715 on-line or on paper. Permission for modification is crucial too.
21717 As a general rule, I don't believe that it is essential for people to
21718 have permission to modify all sorts of articles and books. The issues
21719 for writings are not necessarily the same as those for software. For
21720 example, I don't think you or I are obliged to give permission to
21721 modify articles like this one, which describe our actions and our
21724 But there is a particular reason why the freedom to modify is crucial
21725 for documentation for free software. When people exercise their right
21726 to modify the software, and add or change its features, if they are
21727 conscientious they will change the manual too---so they can provide
21728 accurate and usable documentation with the modified program. A manual
21729 which forbids programmers to be conscientious and finish the job, or
21730 more precisely requires them to write a new manual from scratch if
21731 they change the program, does not fill our community's needs.
21733 While a blanket prohibition on modification is unacceptable, some
21734 kinds of limits on the method of modification pose no problem. For
21735 example, requirements to preserve the original author's copyright
21736 notice, the distribution terms, or the list of authors, are ok. It is
21737 also no problem to require modified versions to include notice that
21738 they were modified, even to have entire sections that may not be
21739 deleted or changed, as long as these sections deal with nontechnical
21740 topics. (Some GNU manuals have them.)
21742 These kinds of restrictions are not a problem because, as a practical
21743 matter, they don't stop the conscientious programmer from adapting the
21744 manual to fit the modified program. In other words, they don't block
21745 the free software community from making full use of the manual.
21747 However, it must be possible to modify all the technical content of
21748 the manual, and then distribute the result in all the usual media,
21749 through all the usual channels; otherwise, the restrictions do block
21750 the community, the manual is not free, and so we need another manual.
21752 Unfortunately, it is often hard to find someone to write another
21753 manual when a proprietary manual exists. The obstacle is that many
21754 users think that a proprietary manual is good enough---so they don't
21755 see the need to write a free manual. They do not see that the free
21756 operating system has a gap that needs filling.
21758 Why do users think that proprietary manuals are good enough? Some have
21759 not considered the issue. I hope this article will do something to
21762 Other users consider proprietary manuals acceptable for the same
21763 reason so many people consider proprietary software acceptable: they
21764 judge in purely practical terms, not using freedom as a
21765 criterion. These people are entitled to their opinions, but since
21766 those opinions spring from values which do not include freedom, they
21767 are no guide for those of us who do value freedom.
21769 Please spread the word about this issue. We continue to lose manuals
21770 to proprietary publishing. If we spread the word that proprietary
21771 manuals are not sufficient, perhaps the next person who wants to help
21772 GNU by writing documentation will realize, before it is too late, that
21773 he must above all make it free.
21775 We can also encourage commercial publishers to sell free, copylefted
21776 manuals instead of proprietary ones. One way you can help this is to
21777 check the distribution terms of a manual before you buy it, and prefer
21778 copylefted manuals to non-copylefted ones.
21782 Note: The Free Software Foundation maintains a page on its Web site
21783 that lists free books available from other publishers:@*
21784 @uref{https://www.gnu.org/doc/other-free-books.html}
21786 @node GNU Free Documentation License
21787 @appendix GNU Free Documentation License
21789 @cindex FDL, GNU Free Documentation License
21790 @include doclicense.texi
21796 MENU ENTRY: NODE NAME.
21802 @c Place biographical information on right-hand (verso) page
21805 \par\vfill\supereject
21807 \global\evenheadline={\hfil} \global\evenfootline={\hfil}
21808 \global\oddheadline={\hfil} \global\oddfootline={\hfil}
21811 % \par\vfill\supereject
21812 \global\evenheadline={\hfil} \global\evenfootline={\hfil}
21813 \global\oddheadline={\hfil} \global\oddfootline={\hfil}
21814 %\page\hbox{}%\page
21815 %\page\hbox{}%\page
21822 @c ================ Biographical information ================
21826 @center About the Author
21831 @node About the Author
21832 @unnumbered About the Author
21836 Robert J. Chassell has worked with GNU Emacs since 1985. He writes
21837 and edits, teaches Emacs and Emacs Lisp, and speaks throughout the
21838 world on software freedom. Chassell was a founding Director and
21839 Treasurer of the Free Software Foundation, Inc. He is co-author of
21840 the @cite{Texinfo} manual, and has edited more than a dozen other
21841 books. He graduated from Cambridge University, in England. He has an
21842 abiding interest in social and economic history and flies his own
21849 @c @c Prevent page number on blank verso, so eject it first.
21851 @c \par\vfill\supereject
21856 @c @evenheading @thispage @| @| @thistitle
21857 @c @oddheading @| @| @thispage