2 @c This is part of the GNU Emacs Lisp Reference Manual.
3 @c Copyright (C) 1990-1993, 1998-1999, 2001-2013 Free Software
5 @c See the file elisp.texi for copying conditions.
6 @node GNU Emacs Internals
7 @appendix GNU Emacs Internals
9 This chapter describes how the runnable Emacs executable is dumped with
10 the preloaded Lisp libraries in it, how storage is allocated, and some
11 internal aspects of GNU Emacs that may be of interest to C programmers.
14 * Building Emacs:: How the dumped Emacs is made.
15 * Pure Storage:: Kludge to make preloaded Lisp functions shareable.
16 * Garbage Collection:: Reclaiming space for Lisp objects no longer used.
17 * Memory Usage:: Info about total size of Lisp objects made so far.
18 * Writing Emacs Primitives:: Writing C code for Emacs.
19 * Object Internals:: Data formats of buffers, windows, processes.
20 * C Integer Types:: How C integer types are used inside Emacs.
24 @section Building Emacs
25 @cindex building Emacs
28 This section explains the steps involved in building the Emacs
29 executable. You don't have to know this material to build and install
30 Emacs, since the makefiles do all these things automatically. This
31 information is pertinent to Emacs developers.
33 Compilation of the C source files in the @file{src} directory
34 produces an executable file called @file{temacs}, also called a
35 @dfn{bare impure Emacs}. It contains the Emacs Lisp interpreter and
36 I/O routines, but not the editing commands.
38 @cindex @file{loadup.el}
39 The command @w{@command{temacs -l loadup}} would run @file{temacs}
40 and direct it to load @file{loadup.el}. The @code{loadup} library
41 loads additional Lisp libraries, which set up the normal Emacs editing
42 environment. After this step, the Emacs executable is no longer
46 Because it takes some time to load the standard Lisp files, the
47 @file{temacs} executable usually isn't run directly by users.
48 Instead, as one of the last steps of building Emacs, the command
49 @samp{temacs -batch -l loadup dump} is run. The special @samp{dump}
50 argument causes @command{temacs} to dump out an executable program,
51 called @file{emacs}, which has all the standard Lisp files preloaded.
52 (The @samp{-batch} argument prevents @file{temacs} from trying to
53 initialize any of its data on the terminal, so that the tables of
54 terminal information are empty in the dumped Emacs.)
56 @cindex preloaded Lisp files
57 @vindex preloaded-file-list
58 The dumped @file{emacs} executable (also called a @dfn{pure} Emacs)
59 is the one which is installed. The variable
60 @code{preloaded-file-list} stores a list of the Lisp files preloaded
61 into the dumped Emacs. If you port Emacs to a new operating system,
62 and are not able to implement dumping, then Emacs must load
63 @file{loadup.el} each time it starts.
65 @cindex @file{site-load.el}
66 You can specify additional files to preload by writing a library named
67 @file{site-load.el} that loads them. You may need to rebuild Emacs
68 with an added definition
71 #define SITELOAD_PURESIZE_EXTRA @var{n}
75 to make @var{n} added bytes of pure space to hold the additional files;
76 see @file{src/puresize.h}.
77 (Try adding increments of 20000 until it is big enough.) However, the
78 advantage of preloading additional files decreases as machines get
79 faster. On modern machines, it is usually not advisable.
81 After @file{loadup.el} reads @file{site-load.el}, it finds the
82 documentation strings for primitive and preloaded functions (and
83 variables) in the file @file{etc/DOC} where they are stored, by
84 calling @code{Snarf-documentation} (@pxref{Definition of
85 Snarf-documentation,, Accessing Documentation}).
87 @cindex @file{site-init.el}
88 @cindex preloading additional functions and variables
89 You can specify other Lisp expressions to execute just before dumping
90 by putting them in a library named @file{site-init.el}. This file is
91 executed after the documentation strings are found.
93 If you want to preload function or variable definitions, there are
94 three ways you can do this and make their documentation strings
95 accessible when you subsequently run Emacs:
99 Arrange to scan these files when producing the @file{etc/DOC} file,
100 and load them with @file{site-load.el}.
103 Load the files with @file{site-init.el}, then copy the files into the
104 installation directory for Lisp files when you install Emacs.
107 Specify a @code{nil} value for @code{byte-compile-dynamic-docstrings}
108 as a local variable in each of these files, and load them with either
109 @file{site-load.el} or @file{site-init.el}. (This method has the
110 drawback that the documentation strings take up space in Emacs all the
114 It is not advisable to put anything in @file{site-load.el} or
115 @file{site-init.el} that would alter any of the features that users
116 expect in an ordinary unmodified Emacs. If you feel you must override
117 normal features for your site, do it with @file{default.el}, so that
118 users can override your changes if they wish. @xref{Startup Summary}.
120 In a package that can be preloaded, it is sometimes necessary (or
121 useful) to delay certain evaluations until Emacs subsequently starts
122 up. The vast majority of such cases relate to the values of
123 customizable variables. For example, @code{tutorial-directory} is a
124 variable defined in @file{startup.el}, which is preloaded. The default
125 value is set based on @code{data-directory}. The variable needs to
126 access the value of @code{data-directory} when Emacs starts, not when
127 it is dumped, because the Emacs executable has probably been installed
128 in a different location since it was dumped.
130 @defun custom-initialize-delay symbol value
131 This function delays the initialization of @var{symbol} to the next
132 Emacs start. You normally use this function by specifying it as the
133 @code{:initialize} property of a customizable variable. (The argument
134 @var{value} is unused, and is provided only for compatibility with the
135 form Custom expects.)
138 In the unlikely event that you need a more general functionality than
139 @code{custom-initialize-delay} provides, you can use
140 @code{before-init-hook} (@pxref{Startup Summary}).
142 @defun dump-emacs to-file from-file
144 This function dumps the current state of Emacs into an executable file
145 @var{to-file}. It takes symbols from @var{from-file} (this is normally
146 the executable file @file{temacs}).
148 If you want to use this function in an Emacs that was already dumped,
149 you must run Emacs with @samp{-batch}.
153 @section Pure Storage
156 Emacs Lisp uses two kinds of storage for user-created Lisp objects:
157 @dfn{normal storage} and @dfn{pure storage}. Normal storage is where
158 all the new data created during an Emacs session are kept
159 (@pxref{Garbage Collection}). Pure storage is used for certain data
160 in the preloaded standard Lisp files---data that should never change
161 during actual use of Emacs.
163 Pure storage is allocated only while @command{temacs} is loading the
164 standard preloaded Lisp libraries. In the file @file{emacs}, it is
165 marked as read-only (on operating systems that permit this), so that
166 the memory space can be shared by all the Emacs jobs running on the
167 machine at once. Pure storage is not expandable; a fixed amount is
168 allocated when Emacs is compiled, and if that is not sufficient for
169 the preloaded libraries, @file{temacs} allocates dynamic memory for
170 the part that didn't fit. The resulting image will work, but garbage
171 collection (@pxref{Garbage Collection}) is disabled in this situation,
172 causing a memory leak. Such an overflow normally won't happen unless
173 you try to preload additional libraries or add features to the
174 standard ones. Emacs will display a warning about the overflow when
175 it starts. If this happens, you should increase the compilation
176 parameter @code{SYSTEM_PURESIZE_EXTRA} in the file
177 @file{src/puresize.h} and rebuild Emacs.
179 @defun purecopy object
180 This function makes a copy in pure storage of @var{object}, and returns
181 it. It copies a string by simply making a new string with the same
182 characters, but without text properties, in pure storage. It
183 recursively copies the contents of vectors and cons cells. It does
184 not make copies of other objects such as symbols, but just returns
185 them unchanged. It signals an error if asked to copy markers.
187 This function is a no-op except while Emacs is being built and dumped;
188 it is usually called only in preloaded Lisp files.
191 @defvar pure-bytes-used
192 The value of this variable is the number of bytes of pure storage
193 allocated so far. Typically, in a dumped Emacs, this number is very
194 close to the total amount of pure storage available---if it were not,
195 we would preallocate less.
199 This variable determines whether @code{defun} should make a copy of the
200 function definition in pure storage. If it is non-@code{nil}, then the
201 function definition is copied into pure storage.
203 This flag is @code{t} while loading all of the basic functions for
204 building Emacs initially (allowing those functions to be shareable and
205 non-collectible). Dumping Emacs as an executable always writes
206 @code{nil} in this variable, regardless of the value it actually has
207 before and after dumping.
209 You should not change this flag in a running Emacs.
212 @node Garbage Collection
213 @section Garbage Collection
215 @cindex memory allocation
216 When a program creates a list or the user defines a new function
217 (such as by loading a library), that data is placed in normal storage.
218 If normal storage runs low, then Emacs asks the operating system to
219 allocate more memory. Different types of Lisp objects, such as
220 symbols, cons cells, small vectors, markers, etc., are segregated in
221 distinct blocks in memory. (Large vectors, long strings, buffers and
222 certain other editing types, which are fairly large, are allocated in
223 individual blocks, one per object; small strings are packed into blocks
224 of 8k bytes, and small vectors are packed into blocks of 4k bytes).
226 @cindex vector-like objects, storage
227 @cindex storage of vector-like Lisp objects
228 Beyond the basic vector, a lot of objects like window, buffer, and
229 frame are managed as if they were vectors. The corresponding C data
230 structures include the @code{struct vectorlike_header} field whose
231 @code{size} member contains the subtype enumerated by @code{enum pvec_type}
232 and an information about how many @code{Lisp_Object} fields this structure
233 contains and what the size of the rest data is. This information is
234 needed to calculate the memory footprint of an object, and used
235 by the vector allocation code while iterating over the vector blocks.
237 @cindex garbage collection
238 It is quite common to use some storage for a while, then release it
239 by (for example) killing a buffer or deleting the last pointer to an
240 object. Emacs provides a @dfn{garbage collector} to reclaim this
241 abandoned storage. The garbage collector operates by finding and
242 marking all Lisp objects that are still accessible to Lisp programs.
243 To begin with, it assumes all the symbols, their values and associated
244 function definitions, and any data presently on the stack, are
245 accessible. Any objects that can be reached indirectly through other
246 accessible objects are also accessible.
248 When marking is finished, all objects still unmarked are garbage. No
249 matter what the Lisp program or the user does, it is impossible to refer
250 to them, since there is no longer a way to reach them. Their space
251 might as well be reused, since no one will miss them. The second
252 (``sweep'') phase of the garbage collector arranges to reuse them.
254 @c ??? Maybe add something describing weak hash tables here?
257 The sweep phase puts unused cons cells onto a @dfn{free list}
258 for future allocation; likewise for symbols and markers. It compacts
259 the accessible strings so they occupy fewer 8k blocks; then it frees the
260 other 8k blocks. Unreachable vectors from vector blocks are coalesced
261 to create largest possible free areas; if a free area spans a complete
262 4k block, that block is freed. Otherwise, the free area is recorded
263 in a free list array, where each entry corresponds to a free list
264 of areas of the same size. Large vectors, buffers, and other large
265 objects are allocated and freed individually.
267 @cindex CL note---allocate more storage
269 @b{Common Lisp note:} Unlike other Lisps, GNU Emacs Lisp does not
270 call the garbage collector when the free list is empty. Instead, it
271 simply requests the operating system to allocate more storage, and
272 processing continues until @code{gc-cons-threshold} bytes have been
275 This means that you can make sure that the garbage collector will not
276 run during a certain portion of a Lisp program by calling the garbage
277 collector explicitly just before it (provided that portion of the
278 program does not use so much space as to force a second garbage
282 @deffn Command garbage-collect
283 This command runs a garbage collection, and returns information on
284 the amount of space in use. (Garbage collection can also occur
285 spontaneously if you use more than @code{gc-cons-threshold} bytes of
286 Lisp data since the previous garbage collection.)
288 @code{garbage-collect} returns a list with information on amount of space in
289 use, where each entry has the form @samp{(@var{name} @var{size} @var{used})}
290 or @samp{(@var{name} @var{size} @var{used} @var{free})}. In the entry,
291 @var{name} is a symbol describing the kind of objects this entry represents,
292 @var{size} is the number of bytes used by each one, @var{used} is the number
293 of those objects that were found live in the heap, and optional @var{free} is
294 the number of those objects that are not live but that Emacs keeps around for
295 future allocations. So an overall result is:
298 ((@code{conses} @var{cons-size} @var{used-conses} @var{free-conses})
299 (@code{symbols} @var{symbol-size} @var{used-symbols} @var{free-symbols})
300 (@code{miscs} @var{misc-size} @var{used-miscs} @var{free-miscs})
301 (@code{strings} @var{string-size} @var{used-strings} @var{free-strings})
302 (@code{string-bytes} @var{byte-size} @var{used-bytes})
303 (@code{vectors} @var{vector-size} @var{used-vectors})
304 (@code{vector-slots} @var{slot-size} @var{used-slots} @var{free-slots})
305 (@code{floats} @var{float-size} @var{used-floats} @var{free-floats})
306 (@code{intervals} @var{interval-size} @var{used-intervals} @var{free-intervals})
307 (@code{buffers} @var{buffer-size} @var{used-buffers})
308 (@code{heap} @var{unit-size} @var{total-size} @var{free-size}))
315 @result{} ((conses 16 49126 8058) (symbols 48 14607 0)
316 (miscs 40 34 56) (strings 32 2942 2607)
317 (string-bytes 1 78607) (vectors 16 7247)
318 (vector-slots 8 341609 29474) (floats 8 71 102)
319 (intervals 56 27 26) (buffers 944 8)
320 (heap 1024 11715 2678))
323 Below is a table explaining each element. Note that last @code{heap} entry
324 is optional and present only if an underlying @code{malloc} implementation
325 provides @code{mallinfo} function.
329 Internal size of a cons cell, i.e., @code{sizeof (struct Lisp_Cons)}.
332 The number of cons cells in use.
335 The number of cons cells for which space has been obtained from
336 the operating system, but that are not currently being used.
339 Internal size of a symbol, i.e., @code{sizeof (struct Lisp_Symbol)}.
342 The number of symbols in use.
345 The number of symbols for which space has been obtained from
346 the operating system, but that are not currently being used.
349 Internal size of a miscellaneous entity, i.e.,
350 @code{sizeof (union Lisp_Misc)}, which is a size of the
351 largest type enumerated in @code{enum Lisp_Misc_Type}.
354 The number of miscellaneous objects in use. These include markers
355 and overlays, plus certain objects not visible to users.
358 The number of miscellaneous objects for which space has been obtained
359 from the operating system, but that are not currently being used.
362 Internal size of a string header, i.e., @code{sizeof (struct Lisp_String)}.
365 The number of string headers in use.
368 The number of string headers for which space has been obtained
369 from the operating system, but that are not currently being used.
372 This is used for convenience and equals to @code{sizeof (char)}.
375 The total size of all string data in bytes.
378 Internal size of a vector header, i.e., @code{sizeof (struct Lisp_Vector)}.
381 The number of vector headers allocated from the vector blocks.
384 Internal size of a vector slot, always equal to @code{sizeof (Lisp_Object)}.
387 The number of slots in all used vectors.
390 The number of free slots in all vector blocks.
393 Internal size of a float object, i.e., @code{sizeof (struct Lisp_Float)}.
394 (Do not confuse it with the native platform @code{float} or @code{double}.)
397 The number of floats in use.
400 The number of floats for which space has been obtained from
401 the operating system, but that are not currently being used.
404 Internal size of an interval object, i.e., @code{sizeof (struct interval)}.
407 The number of intervals in use.
410 The number of intervals for which space has been obtained from
411 the operating system, but that are not currently being used.
414 Internal size of a buffer, i.e., @code{sizeof (struct buffer)}.
415 (Do not confuse with the value returned by @code{buffer-size} function.)
418 The number of buffer objects in use. This includes killed buffers
419 invisible to users, i.e., all buffers in @code{all_buffers} list.
422 The unit of heap space measurement, always equal to 1024 bytes.
425 Total heap size, in @var{unit-size} units.
428 Heap space which is not currently used, in @var{unit-size} units.
431 If there was overflow in pure space (@pxref{Pure Storage}),
432 @code{garbage-collect} returns @code{nil}, because a real garbage
433 collection cannot be done.
436 @defopt garbage-collection-messages
437 If this variable is non-@code{nil}, Emacs displays a message at the
438 beginning and end of garbage collection. The default value is
443 This is a normal hook that is run at the end of garbage collection.
444 Garbage collection is inhibited while the hook functions run, so be
445 careful writing them.
448 @defopt gc-cons-threshold
449 The value of this variable is the number of bytes of storage that must
450 be allocated for Lisp objects after one garbage collection in order to
451 trigger another garbage collection. You can use the result returned by
452 @code{garbage-collect} to get an information about size of the particular
453 object type; space allocated to the contents of buffers does not count.
454 Note that the subsequent garbage collection does not happen immediately
455 when the threshold is exhausted, but only the next time the Lisp interpreter
458 The initial threshold value is @code{GC_DEFAULT_THRESHOLD}, defined in
459 @file{alloc.c}. Since it's defined in @code{word_size} units, the value
460 is 400,000 for the default 32-bit configuration and 800,000 for the 64-bit
461 one. If you specify a larger value, garbage collection will happen less
462 often. This reduces the amount of time spent garbage collecting, but
463 increases total memory use. You may want to do this when running a program
464 that creates lots of Lisp data.
466 You can make collections more frequent by specifying a smaller value, down
467 to 1/10th of @code{GC_DEFAULT_THRESHOLD}. A value less than this minimum
468 will remain in effect only until the subsequent garbage collection, at which
469 time @code{garbage-collect} will set the threshold back to the minimum.
472 @defopt gc-cons-percentage
473 The value of this variable specifies the amount of consing before a
474 garbage collection occurs, as a fraction of the current heap size.
475 This criterion and @code{gc-cons-threshold} apply in parallel, and
476 garbage collection occurs only when both criteria are satisfied.
478 As the heap size increases, the time to perform a garbage collection
479 increases. Thus, it can be desirable to do them less frequently in
483 The value returned by @code{garbage-collect} describes the amount of
484 memory used by Lisp data, broken down by data type. By contrast, the
485 function @code{memory-limit} provides information on the total amount of
486 memory Emacs is currently using.
489 This function returns the address of the last byte Emacs has allocated,
490 divided by 1024. We divide the value by 1024 to make sure it fits in a
493 You can use this to get a general idea of how your actions affect the
498 This variable is @code{t} if Emacs is nearly out of memory for Lisp
499 objects, and @code{nil} otherwise.
502 @defun memory-use-counts
503 This returns a list of numbers that count the number of objects
504 created in this Emacs session. Each of these counters increments for
505 a certain kind of object. See the documentation string for details.
509 This variable contains the total number of garbage collections
510 done so far in this Emacs session.
514 This variable contains the total number of seconds of elapsed time
515 during garbage collection so far in this Emacs session, as a floating
520 @section Memory Usage
523 These functions and variables give information about the total amount
524 of memory allocation that Emacs has done, broken down by data type.
525 Note the difference between these and the values returned by
526 @code{garbage-collect}; those count objects that currently exist, but
527 these count the number or size of all allocations, including those for
528 objects that have since been freed.
530 @defvar cons-cells-consed
531 The total number of cons cells that have been allocated so far
532 in this Emacs session.
535 @defvar floats-consed
536 The total number of floats that have been allocated so far
537 in this Emacs session.
540 @defvar vector-cells-consed
541 The total number of vector cells that have been allocated so far
542 in this Emacs session.
545 @defvar symbols-consed
546 The total number of symbols that have been allocated so far
547 in this Emacs session.
550 @defvar string-chars-consed
551 The total number of string characters that have been allocated so far
555 @defvar misc-objects-consed
556 The total number of miscellaneous objects that have been allocated so
557 far in this session. These include markers and overlays, plus
558 certain objects not visible to users.
561 @defvar intervals-consed
562 The total number of intervals that have been allocated so far
563 in this Emacs session.
566 @defvar strings-consed
567 The total number of strings that have been allocated so far in this
571 @node Writing Emacs Primitives
572 @section Writing Emacs Primitives
573 @cindex primitive function internals
574 @cindex writing Emacs primitives
576 Lisp primitives are Lisp functions implemented in C@. The details of
577 interfacing the C function so that Lisp can call it are handled by a few
578 C macros. The only way to really understand how to write new C code is
579 to read the source, but we can explain some things here.
581 An example of a special form is the definition of @code{or}, from
582 @file{eval.c}. (An ordinary function would have the same general
585 @cindex garbage collection protection
588 DEFUN ("or", For, Sor, 0, UNEVALLED, 0,
589 doc: /* Eval args until one of them yields non-nil, then return
591 The remaining args are not evalled at all.
592 If all args return nil, return nil.
595 usage: (or CONDITIONS ...) */)
598 register Lisp_Object val = Qnil;
609 val = eval_sub (XCAR (args));
623 @cindex @code{DEFUN}, C macro to define Lisp primitives
624 Let's start with a precise explanation of the arguments to the
625 @code{DEFUN} macro. Here is a template for them:
628 DEFUN (@var{lname}, @var{fname}, @var{sname}, @var{min}, @var{max}, @var{interactive}, @var{doc})
633 This is the name of the Lisp symbol to define as the function name; in
634 the example above, it is @code{or}.
637 This is the C function name for this function. This is the name that
638 is used in C code for calling the function. The name is, by
639 convention, @samp{F} prepended to the Lisp name, with all dashes
640 (@samp{-}) in the Lisp name changed to underscores. Thus, to call
641 this function from C code, call @code{For}.
644 This is a C variable name to use for a structure that holds the data for
645 the subr object that represents the function in Lisp. This structure
646 conveys the Lisp symbol name to the initialization routine that will
647 create the symbol and store the subr object as its definition. By
648 convention, this name is always @var{fname} with @samp{F} replaced with
652 This is the minimum number of arguments that the function requires. The
653 function @code{or} allows a minimum of zero arguments.
656 This is the maximum number of arguments that the function accepts, if
657 there is a fixed maximum. Alternatively, it can be @code{UNEVALLED},
658 indicating a special form that receives unevaluated arguments, or
659 @code{MANY}, indicating an unlimited number of evaluated arguments (the
660 equivalent of @code{&rest}). Both @code{UNEVALLED} and @code{MANY} are
661 macros. If @var{max} is a number, it must be more than @var{min} but
665 This is an interactive specification, a string such as might be used as
666 the argument of @code{interactive} in a Lisp function. In the case of
667 @code{or}, it is 0 (a null pointer), indicating that @code{or} cannot be
668 called interactively. A value of @code{""} indicates a function that
669 should receive no arguments when called interactively. If the value
670 begins with a @samp{(}, the string is evaluated as a Lisp form.
671 For examples of the last two forms, see @code{widen} and
672 @code{narrow-to-region} in @file{editfns.c}.
675 This is the documentation string. It uses C comment syntax rather
676 than C string syntax because comment syntax requires nothing special
677 to include multiple lines. The @samp{doc:} identifies the comment
678 that follows as the documentation string. The @samp{/*} and @samp{*/}
679 delimiters that begin and end the comment are not part of the
680 documentation string.
682 If the last line of the documentation string begins with the keyword
683 @samp{usage:}, the rest of the line is treated as the argument list
684 for documentation purposes. This way, you can use different argument
685 names in the documentation string from the ones used in the C code.
686 @samp{usage:} is required if the function has an unlimited number of
689 All the usual rules for documentation strings in Lisp code
690 (@pxref{Documentation Tips}) apply to C code documentation strings
694 After the call to the @code{DEFUN} macro, you must write the
695 argument list for the C function, including the types for the
696 arguments. If the primitive accepts a fixed maximum number of Lisp
697 arguments, there must be one C argument for each Lisp argument, and
698 each argument must be of type @code{Lisp_Object}. (Various macros and
699 functions for creating values of type @code{Lisp_Object} are declared
700 in the file @file{lisp.h}.) If the primitive has no upper limit on
701 the number of Lisp arguments, it must have exactly two C arguments:
702 the first is the number of Lisp arguments, and the second is the
703 address of a block containing their values. These have types
704 @code{int} and @w{@code{Lisp_Object *}} respectively. Since
705 @code{Lisp_Object} can hold any Lisp object of any data type, you
706 can determine the actual data type only at run time; so if you want
707 a primitive to accept only a certain type of argument, you must check
708 the type explicitly using a suitable predicate (@pxref{Type Predicates}).
709 @cindex type checking internals
711 @cindex @code{GCPRO} and @code{UNGCPRO}
712 @cindex protect C variables from garbage collection
713 Within the function @code{For} itself, note the use of the macros
714 @code{GCPRO1} and @code{UNGCPRO}. These macros are defined for the
715 sake of the few platforms which do not use Emacs' default
716 stack-marking garbage collector. The @code{GCPRO1} macro ``protects''
717 a variable from garbage collection, explicitly informing the garbage
718 collector that that variable and all its contents must be as
719 accessible. GC protection is necessary in any function which can
720 perform Lisp evaluation by calling @code{eval_sub} or @code{Feval} as
721 a subroutine, either directly or indirectly.
723 It suffices to ensure that at least one pointer to each object is
724 GC-protected. Thus, a particular local variable can do without
725 protection if it is certain that the object it points to will be
726 preserved by some other pointer (such as another local variable that
727 has a @code{GCPRO}). Otherwise, the local variable needs a
730 The macro @code{GCPRO1} protects just one local variable. If you
731 want to protect two variables, use @code{GCPRO2} instead; repeating
732 @code{GCPRO1} will not work. Macros @code{GCPRO3}, @code{GCPRO4},
733 @code{GCPRO5}, and @code{GCPRO6} also exist. All these macros
734 implicitly use local variables such as @code{gcpro1}; you must declare
735 these explicitly, with type @code{struct gcpro}. Thus, if you use
736 @code{GCPRO2}, you must declare @code{gcpro1} and @code{gcpro2}.
738 @code{UNGCPRO} cancels the protection of the variables that are
739 protected in the current function. It is necessary to do this
742 You must not use C initializers for static or global variables unless
743 the variables are never written once Emacs is dumped. These variables
744 with initializers are allocated in an area of memory that becomes
745 read-only (on certain operating systems) as a result of dumping Emacs.
748 @cindex @code{defsubr}, Lisp symbol for a primitive
749 Defining the C function is not enough to make a Lisp primitive
750 available; you must also create the Lisp symbol for the primitive and
751 store a suitable subr object in its function cell. The code looks like
755 defsubr (&@var{sname});
759 Here @var{sname} is the name you used as the third argument to @code{DEFUN}.
761 If you add a new primitive to a file that already has Lisp primitives
762 defined in it, find the function (near the end of the file) named
763 @code{syms_of_@var{something}}, and add the call to @code{defsubr}
764 there. If the file doesn't have this function, or if you create a new
765 file, add to it a @code{syms_of_@var{filename}} (e.g.,
766 @code{syms_of_myfile}). Then find the spot in @file{emacs.c} where all
767 of these functions are called, and add a call to
768 @code{syms_of_@var{filename}} there.
770 @anchor{Defining Lisp variables in C}
771 @vindex byte-boolean-vars
772 @cindex defining Lisp variables in C
773 @cindex @code{DEFVAR_INT}, @code{DEFVAR_LISP}, @code{DEFVAR_BOOL}
774 The function @code{syms_of_@var{filename}} is also the place to define
775 any C variables that are to be visible as Lisp variables.
776 @code{DEFVAR_LISP} makes a C variable of type @code{Lisp_Object} visible
777 in Lisp. @code{DEFVAR_INT} makes a C variable of type @code{int}
778 visible in Lisp with a value that is always an integer.
779 @code{DEFVAR_BOOL} makes a C variable of type @code{int} visible in Lisp
780 with a value that is either @code{t} or @code{nil}. Note that variables
781 defined with @code{DEFVAR_BOOL} are automatically added to the list
782 @code{byte-boolean-vars} used by the byte compiler.
784 @cindex defining customization variables in C
785 If you want to make a Lisp variables that is defined in C behave
786 like one declared with @code{defcustom}, add an appropriate entry to
789 @cindex @code{staticpro}, protection from GC
790 If you define a file-scope C variable of type @code{Lisp_Object},
791 you must protect it from garbage-collection by calling @code{staticpro}
792 in @code{syms_of_@var{filename}}, like this:
795 staticpro (&@var{variable});
798 Here is another example function, with more complicated arguments.
799 This comes from the code in @file{window.c}, and it demonstrates the use
800 of macros and functions to manipulate Lisp objects.
804 DEFUN ("coordinates-in-window-p", Fcoordinates_in_window_p,
805 Scoordinates_in_window_p, 2, 2, 0,
806 doc: /* Return non-nil if COORDINATES are in WINDOW.
810 or `right-margin' is returned. */)
811 (register Lisp_Object coordinates, Lisp_Object window)
820 CHECK_LIVE_WINDOW (window);
821 w = XWINDOW (window);
822 f = XFRAME (w->frame);
823 CHECK_CONS (coordinates);
824 lx = Fcar (coordinates);
825 ly = Fcdr (coordinates);
826 CHECK_NUMBER_OR_FLOAT (lx);
827 CHECK_NUMBER_OR_FLOAT (ly);
828 x = FRAME_PIXEL_X_FROM_CANON_X (f, lx) + FRAME_INTERNAL_BORDER_WIDTH(f);
829 y = FRAME_PIXEL_Y_FROM_CANON_Y (f, ly) + FRAME_INTERNAL_BORDER_WIDTH(f);
833 switch (coordinates_in_window (w, x, y))
835 case ON_NOTHING: /* NOT in window at all. */
842 case ON_MODE_LINE: /* In mode line of window. */
849 case ON_SCROLL_BAR: /* On scroll-bar of window. */
850 /* Historically we are supposed to return nil in this case. */
862 Note that C code cannot call functions by name unless they are defined
863 in C@. The way to call a function written in Lisp is to use
864 @code{Ffuncall}, which embodies the Lisp function @code{funcall}. Since
865 the Lisp function @code{funcall} accepts an unlimited number of
866 arguments, in C it takes two: the number of Lisp-level arguments, and a
867 one-dimensional array containing their values. The first Lisp-level
868 argument is the Lisp function to call, and the rest are the arguments to
869 pass to it. Since @code{Ffuncall} can call the evaluator, you must
870 protect pointers from garbage collection around the call to
873 The C functions @code{call0}, @code{call1}, @code{call2}, and so on,
874 provide handy ways to call a Lisp function conveniently with a fixed
875 number of arguments. They work by calling @code{Ffuncall}.
877 @file{eval.c} is a very good file to look through for examples;
878 @file{lisp.h} contains the definitions for some important macros and
881 If you define a function which is side-effect free, update the code
882 in @file{byte-opt.el} that binds @code{side-effect-free-fns} and
883 @code{side-effect-and-error-free-fns} so that the compiler optimizer
886 @node Object Internals
887 @section Object Internals
888 @cindex object internals
890 Emacs Lisp provides a rich set of the data types. Some of them, like cons
891 cells, integers and strings, are common to nearly all Lisp dialects. Some
892 others, like markers and buffers, are quite special and needed to provide
893 the basic support to write editor commands in Lisp. To implement such
894 a variety of object types and provide an efficient way to pass objects between
895 the subsystems of an interpreter, there is a set of C data structures and
896 a special type to represent the pointers to all of them, which is known as
897 @dfn{tagged pointer}.
899 In C, the tagged pointer is an object of type @code{Lisp_Object}. Any
900 initialized variable of such a type always holds the value of one of the
901 following basic data types: integer, symbol, string, cons cell, float,
902 vectorlike or miscellaneous object. Each of these data types has the
903 corresponding tag value. All tags are enumerated by @code{enum Lisp_Type}
904 and placed into a 3-bit bitfield of the @code{Lisp_Object}. The rest of the
905 bits is the value itself. Integer values are immediate, i.e., directly
906 represented by those @dfn{value bits}, and all other objects are represented
907 by the C pointers to a corresponding object allocated from the heap. Width
908 of the @code{Lisp_Object} is platform- and configuration-dependent: usually
909 it's equal to the width of an underlying platform pointer (i.e., 32-bit on
910 a 32-bit machine and 64-bit on a 64-bit one), but also there is a special
911 configuration where @code{Lisp_Object} is 64-bit but all pointers are 32-bit.
912 The latter trick was designed to overcome the limited range of values for
913 Lisp integers on a 32-bit system by using 64-bit @code{long long} type for
916 The following C data structures are defined in @file{lisp.h} to represent
917 the basic data types beyond integers:
920 @item struct Lisp_Cons
921 Cons cell, an object used to construct lists.
923 @item struct Lisp_String
924 String, the basic object to represent a sequence of characters.
926 @item struct Lisp_Vector
927 Array, a fixed-size set of Lisp objects which may be accessed by an index.
929 @item struct Lisp_Symbol
930 Symbol, the unique-named entity commonly used as an identifier.
932 @item struct Lisp_Float
933 Floating point value.
935 @item union Lisp_Misc
936 Miscellaneous kinds of objects which don't fit into any of the above.
939 These types are the first-class citizens of an internal type system.
940 Since the tag space is limited, all other types are the subtypes of either
941 @code{Lisp_Vectorlike} or @code{Lisp_Misc}. Vector subtypes are enumerated
942 by @code{enum pvec_type}, and nearly all complex objects like windows, buffers,
943 frames, and processes fall into this category. The rest of special types,
944 including markers and overlays, are enumerated by @code{enum Lisp_Misc_Type}
945 and form the set of subtypes of @code{Lisp_Misc}.
947 Below there is a description of a few subtypes of @code{Lisp_Vectorlike}.
948 Buffer object represents the text to display and edit. Window is the part
949 of display structure which shows the buffer or used as a container to
950 recursively place other windows on the same frame. (Do not confuse Emacs Lisp
951 window object with the window as an entity managed by the user interface
952 system like X; in Emacs terminology, the latter is called frame.) Finally,
953 process object is used to manage the subprocesses.
956 * Buffer Internals:: Components of a buffer structure.
957 * Window Internals:: Components of a window structure.
958 * Process Internals:: Components of a process structure.
961 @node Buffer Internals
962 @subsection Buffer Internals
963 @cindex internals, of buffer
964 @cindex buffer internals
966 Two structures (see @file{buffer.h}) are used to represent buffers
967 in C@. The @code{buffer_text} structure contains fields describing the
968 text of a buffer; the @code{buffer} structure holds other fields. In
969 the case of indirect buffers, two or more @code{buffer} structures
970 reference the same @code{buffer_text} structure.
972 Here are some of the fields in @code{struct buffer_text}:
976 The address of the buffer contents.
980 The character and byte positions of the buffer gap. @xref{Buffer
985 The character and byte positions of the end of the buffer text.
988 The size of buffer's gap. @xref{Buffer Gap}.
993 @itemx overlay_modiff
994 These fields count the number of buffer-modification events performed
995 in this buffer. @code{modiff} is incremented after each
996 buffer-modification event, and is never otherwise changed;
997 @code{save_modiff} contains the value of @code{modiff} the last time
998 the buffer was visited or saved; @code{chars_modiff} counts only
999 modifications to the characters in the buffer, ignoring all other
1000 kinds of changes; and @code{overlay_modiff} counts only modifications
1004 @itemx end_unchanged
1005 The number of characters at the start and end of the text that are
1006 known to be unchanged since the last complete redisplay.
1008 @item unchanged_modified
1009 @itemx overlay_unchanged_modified
1010 The values of @code{modiff} and @code{overlay_modiff}, respectively,
1011 after the last complete redisplay. If their current values match
1012 @code{modiff} or @code{overlay_modiff}, that means
1013 @code{beg_unchanged} and @code{end_unchanged} contain no useful
1017 The markers that refer to this buffer. This is actually a single
1018 marker, and successive elements in its marker @code{chain} are the other
1019 markers referring to this buffer text.
1022 The interval tree which records the text properties of this buffer.
1025 Some of the fields of @code{struct buffer} are:
1029 A header of type @code{struct vectorlike_header} is common to all
1033 A @code{struct buffer_text} structure that ordinarily holds the buffer
1034 contents. In indirect buffers, this field is not used.
1037 A pointer to the @code{buffer_text} structure for this buffer. In an
1038 ordinary buffer, this is the @code{own_text} field above. In an
1039 indirect buffer, this is the @code{own_text} field of the base buffer.
1042 A pointer to the next buffer, in the chain of all buffers, including
1043 killed buffers. This chain is used only for allocation and garbage
1044 collection, in order to collect killed buffers properly.
1048 The character and byte positions of point in a buffer.
1052 The character and byte positions of the beginning of the accessible
1053 range of text in the buffer.
1057 The character and byte positions of the end of the accessible range of
1061 In an indirect buffer, this points to the base buffer. In an ordinary
1065 This field contains flags indicating that certain variables are local
1066 in this buffer. Such variables are declared in the C code using
1067 @code{DEFVAR_PER_BUFFER}, and their buffer-local bindings are stored
1068 in fields in the buffer structure itself. (Some of these fields are
1069 described in this table.)
1072 The modification time of the visited file. It is set when the file is
1073 written or read. Before writing the buffer into a file, this field is
1074 compared to the modification time of the file to see if the file has
1075 changed on disk. @xref{Buffer Modification}.
1077 @item auto_save_modified
1078 The time when the buffer was last auto-saved.
1080 @item last_window_start
1081 The @code{window-start} position in the buffer as of the last time the
1082 buffer was displayed in a window.
1085 This flag indicates that narrowing has changed in the buffer.
1088 @item prevent_redisplay_optimizations_p
1089 This flag indicates that redisplay optimizations should not be used to
1090 display this buffer.
1092 @item overlay_center
1093 This field holds the current overlay center position. @xref{Managing
1096 @item overlays_before
1097 @itemx overlays_after
1098 These fields hold, respectively, a list of overlays that end at or
1099 before the current overlay center, and a list of overlays that end
1100 after the current overlay center. @xref{Managing Overlays}.
1101 @code{overlays_before} is sorted in order of decreasing end position,
1102 and @code{overlays_after} is sorted in order of increasing beginning
1105 @c FIXME? the following are now all Lisp_Object BUFFER_INTERNAL_FIELD (foo).
1108 A Lisp string that names the buffer. It is guaranteed to be unique.
1109 @xref{Buffer Names}.
1112 The length of the file this buffer is visiting, when last read or
1113 saved. This and other fields concerned with saving are not kept in
1114 the @code{buffer_text} structure because indirect buffers are never
1118 The directory for expanding relative file names. This is the value of
1119 the buffer-local variable @code{default-directory} (@pxref{File Name Expansion}).
1122 The name of the file visited in this buffer, or @code{nil}. This is
1123 the value of the buffer-local variable @code{buffer-file-name}
1124 (@pxref{Buffer File Name}).
1128 @itemx auto_save_file_name
1129 @itemx auto_save_file_format
1132 @itemx file_truename
1133 @itemx invisibility_spec
1134 @itemx display_count
1136 These fields store the values of Lisp variables that are automatically
1137 buffer-local (@pxref{Buffer-Local Variables}), whose corresponding
1138 variable names have the additional prefix @code{buffer-} and have
1139 underscores replaced with dashes. For instance, @code{undo_list}
1140 stores the value of @code{buffer-undo-list}.
1143 The mark for the buffer. The mark is a marker, hence it is also
1144 included on the list @code{markers}. @xref{The Mark}.
1146 @item local_var_alist
1147 The association list describing the buffer-local variable bindings of
1148 this buffer, not including the built-in buffer-local bindings that
1149 have special slots in the buffer object. (Those slots are omitted
1150 from this table.) @xref{Buffer-Local Variables}.
1153 Symbol naming the major mode of this buffer, e.g., @code{lisp-mode}.
1156 Pretty name of the major mode, e.g., @code{"Lisp"}.
1161 @itemx category_table
1162 @itemx display_table
1163 These fields store the buffer's local keymap (@pxref{Keymaps}), abbrev
1164 table (@pxref{Abbrev Tables}), syntax table (@pxref{Syntax Tables}),
1165 category table (@pxref{Categories}), and display table (@pxref{Display
1168 @item downcase_table
1170 @itemx case_canon_table
1171 These fields store the conversion tables for converting text to lower
1172 case, upper case, and for canonicalizing text for case-fold search.
1176 An alist of the minor modes of this buffer.
1181 These fields are only used in an indirect buffer, or in a buffer that
1182 is the base of an indirect buffer. Each holds a marker that records
1183 @code{pt}, @code{begv}, and @code{zv} respectively, for this buffer
1184 when the buffer is not current.
1186 @item mode_line_format
1187 @itemx header_line_format
1188 @itemx case_fold_search
1192 @itemx auto_fill_function
1193 @itemx truncate_lines
1196 @itemx bidi_display_reordering
1197 @itemx bidi_paragraph_direction
1198 @itemx selective_display
1199 @itemx selective_display_ellipses
1200 @itemx overwrite_mode
1203 @itemx enable_multibyte_characters
1204 @itemx buffer_file_coding_system
1205 @itemx cache_long_line_scans
1206 @itemx point_before_scroll
1207 @itemx left_fringe_width
1208 @itemx right_fringe_width
1209 @itemx fringes_outside_margins
1210 @itemx scroll_bar_width
1211 @itemx indicate_empty_lines
1212 @itemx indicate_buffer_boundaries
1213 @itemx fringe_indicator_alist
1214 @itemx fringe_cursor_alist
1215 @itemx scroll_up_aggressively
1216 @itemx scroll_down_aggressively
1218 @itemx cursor_in_non_selected_windows
1219 These fields store the values of Lisp variables that are automatically
1220 buffer-local (@pxref{Buffer-Local Variables}), whose corresponding
1221 variable names have underscores replaced with dashes. For instance,
1222 @code{mode_line_format} stores the value of @code{mode-line-format}.
1224 @item last_selected_window
1225 This is the last window that was selected with this buffer in it, or @code{nil}
1226 if that window no longer displays this buffer.
1229 @node Window Internals
1230 @subsection Window Internals
1231 @cindex internals, of window
1232 @cindex window internals
1234 The fields of a window (for a complete list, see the definition of
1235 @code{struct window} in @file{window.h}) include:
1239 The frame that this window is on.
1242 Non-@code{nil} if this window is a minibuffer window.
1245 Internally, Emacs arranges windows in a tree; each group of siblings has
1246 a parent window whose area includes all the siblings. This field points
1247 to a window's parent.
1249 Parent windows do not display buffers, and play little role in display
1250 except to shape their child windows. Emacs Lisp programs usually have
1251 no access to the parent windows; they operate on the windows at the
1252 leaves of the tree, which actually display buffers.
1256 These fields contain the window's leftmost child and its topmost child
1257 respectively. @code{hchild} is used if the window is subdivided
1258 horizontally by child windows, and @code{vchild} if it is subdivided
1259 vertically. In a live window, only one of @code{hchild}, @code{vchild},
1260 and @code{buffer} (q.v.@:) is non-@code{nil}.
1264 The next sibling and previous sibling of this window. @code{next} is
1265 @code{nil} if the window is the right-most or bottom-most in its group;
1266 @code{prev} is @code{nil} if it is the left-most or top-most in its
1270 The left-hand edge of the window, measured in columns, relative to the
1271 leftmost column in the frame (column 0).
1274 The top edge of the window, measured in lines, relative to the topmost
1275 line in the frame (line 0).
1279 The width and height of the window, measured in columns and lines
1280 respectively. The width includes the scroll bar and fringes, and/or
1281 the separator line on the right of the window (if any).
1284 The buffer that the window is displaying.
1287 A marker pointing to the position in the buffer that is the first
1288 character displayed in the window.
1291 @cindex window point internals
1292 This is the value of point in the current buffer when this window is
1293 selected; when it is not selected, it retains its previous value.
1296 If this flag is non-@code{nil}, it says that the window has been
1297 scrolled explicitly by the Lisp program. This affects what the next
1298 redisplay does if point is off the screen: instead of scrolling the
1299 window to show the text around point, it moves point to a location that
1302 @item frozen_window_start_p
1303 This field is set temporarily to 1 to indicate to redisplay that
1304 @code{start} of this window should not be changed, even if point
1307 @item start_at_line_beg
1308 Non-@code{nil} means current value of @code{start} was the beginning of a line
1312 This is the last time that the window was selected. The function
1313 @code{get-lru-window} uses this field.
1315 @item sequence_number
1316 A unique number assigned to this window when it was created.
1319 The @code{modiff} field of the window's buffer, as of the last time
1320 a redisplay completed in this window.
1322 @item last_overlay_modified
1323 The @code{overlay_modiff} field of the window's buffer, as of the last
1324 time a redisplay completed in this window.
1327 The buffer's value of point, as of the last time a redisplay completed
1331 A non-@code{nil} value means the window's buffer was ``modified'' when the
1332 window was last updated.
1334 @item vertical_scroll_bar
1335 This window's vertical scroll bar.
1337 @item left_margin_cols
1338 @itemx right_margin_cols
1339 The widths of the left and right margins in this window. A value of
1340 @code{nil} means no margin.
1342 @item left_fringe_width
1343 @itemx right_fringe_width
1344 The widths of the left and right fringes in this window. A value of
1345 @code{nil} or @code{t} means use the values of the frame.
1347 @item fringes_outside_margins
1348 A non-@code{nil} value means the fringes outside the display margins;
1349 othersize they are between the margin and the text.
1351 @item window_end_pos
1352 This is computed as @code{z} minus the buffer position of the last glyph
1353 in the current matrix of the window. The value is only valid if
1354 @code{window_end_valid} is not @code{nil}.
1356 @item window_end_bytepos
1357 The byte position corresponding to @code{window_end_pos}.
1359 @item window_end_vpos
1360 The window-relative vertical position of the line containing
1361 @code{window_end_pos}.
1363 @item window_end_valid
1364 This field is set to a non-@code{nil} value if @code{window_end_pos} is truly
1365 valid. This is @code{nil} if nontrivial redisplay is pre-empted, since in that
1366 case the display that @code{window_end_pos} was computed for did not get
1370 A structure describing where the cursor is in this window.
1373 The value of @code{cursor} as of the last redisplay that finished.
1376 A structure describing where the cursor of this window physically is.
1378 @item phys_cursor_type
1379 @c FIXME What is this?
1380 @c itemx phys_cursor_ascent
1381 @itemx phys_cursor_height
1382 @itemx phys_cursor_width
1383 The type, height, and width of the cursor that was last displayed on
1386 @item phys_cursor_on_p
1387 This field is non-zero if the cursor is physically on.
1390 Non-zero means the cursor in this window is logically off. This is
1391 used for blinking the cursor.
1393 @item last_cursor_off_p
1394 This field contains the value of @code{cursor_off_p} as of the time of
1397 @item must_be_updated_p
1398 This is set to 1 during redisplay when this window must be updated.
1401 This is the number of columns that the display in the window is scrolled
1402 horizontally to the left. Normally, this is 0.
1405 Vertical scroll amount, in pixels. Normally, this is 0.
1408 Non-@code{nil} if this window is dedicated to its buffer.
1411 The window's display table, or @code{nil} if none is specified for it.
1413 @item update_mode_line
1414 Non-@code{nil} means this window's mode line needs to be updated.
1416 @item base_line_number
1417 The line number of a certain position in the buffer, or @code{nil}.
1418 This is used for displaying the line number of point in the mode line.
1421 The position in the buffer for which the line number is known, or
1422 @code{nil} meaning none is known. If it is a buffer, don't display
1423 the line number as long as the window shows that buffer.
1425 @item region_showing
1426 If the region (or part of it) is highlighted in this window, this field
1427 holds the mark position that made one end of that region. Otherwise,
1428 this field is @code{nil}.
1430 @item column_number_displayed
1431 The column number currently displayed in this window's mode line, or @code{nil}
1432 if column numbers are not being displayed.
1434 @item current_matrix
1435 @itemx desired_matrix
1436 Glyph matrices describing the current and desired display of this window.
1439 @node Process Internals
1440 @subsection Process Internals
1441 @cindex internals, of process
1442 @cindex process internals
1444 The fields of a process (for a complete list, see the definition of
1445 @code{struct Lisp_Process} in @file{process.h}) include:
1449 A string, the name of the process.
1452 A list containing the command arguments that were used to start this
1453 process. For a network or serial process, it is @code{nil} if the
1454 process is running or @code{t} if the process is stopped.
1457 If non-@code{nil}, a function used to accept output from the process
1458 instead of a buffer.
1461 If non-@code{nil}, a function called whenever the state of the process
1465 The associated buffer of the process.
1468 An integer, the operating system's process @acronym{ID}.
1469 Pseudo-processes such as network or serial connections use a value of 0.
1472 A flag, @code{t} if this is really a child process. For a network or
1473 serial connection, it is a plist based on the arguments to
1474 @code{make-network-process} or @code{make-serial-process}.
1477 A marker indicating the position of the end of the last output from this
1478 process inserted into the buffer. This is often but not always the end
1481 @item kill_without_query
1482 If this is non-zero, killing Emacs while this process is still running
1483 does not ask for confirmation about killing the process.
1486 The raw process status, as returned by the @code{wait} system call.
1489 The process status, as @code{process-status} should return it.
1493 If these two fields are not equal, a change in the status of the process
1494 needs to be reported, either by running the sentinel or by inserting a
1495 message in the process buffer.
1498 Non-@code{nil} if communication with the subprocess uses a pty;
1499 @code{nil} if it uses a pipe.
1502 The file descriptor for input from the process.
1505 The file descriptor for output to the process.
1508 The name of the terminal that the subprocess is using,
1509 or @code{nil} if it is using pipes.
1511 @item decode_coding_system
1512 Coding-system for decoding the input from this process.
1515 A working buffer for decoding.
1517 @item decoding_carryover
1518 Size of carryover in decoding.
1520 @item encode_coding_system
1521 Coding-system for encoding the output to this process.
1524 A working buffer for encoding.
1526 @item inherit_coding_system_flag
1527 Flag to set @code{coding-system} of the process buffer from the
1528 coding system used to decode process output.
1531 Symbol indicating the type of process: @code{real}, @code{network},
1536 @node C Integer Types
1537 @section C Integer Types
1538 @cindex integer types (C programming language)
1540 Here are some guidelines for use of integer types in the Emacs C
1541 source code. These guidelines sometimes give competing advice; common
1546 Avoid arbitrary limits. For example, avoid @code{int len = strlen
1547 (s);} unless the length of @code{s} is required for other reasons to
1548 fit in @code{int} range.
1551 Do not assume that signed integer arithmetic wraps around on overflow.
1552 This is no longer true of Emacs porting targets: signed integer
1553 overflow has undefined behavior in practice, and can dump core or
1554 even cause earlier or later code to behave ``illogically''. Unsigned
1555 overflow does wrap around reliably, modulo a power of two.
1558 Prefer signed types to unsigned, as code gets confusing when signed
1559 and unsigned types are combined. Many other guidelines assume that
1560 types are signed; in the rarer cases where unsigned types are needed,
1561 similar advice may apply to the unsigned counterparts (e.g.,
1562 @code{size_t} instead of @code{ptrdiff_t}, or @code{uintptr_t} instead
1563 of @code{intptr_t}).
1566 Prefer @code{int} for Emacs character codes, in the range 0 ..@: 0x3FFFFF.
1569 Prefer @code{ptrdiff_t} for sizes, i.e., for integers bounded by the
1570 maximum size of any individual C object or by the maximum number of
1571 elements in any C array. This is part of Emacs's general preference
1572 for signed types. Using @code{ptrdiff_t} limits objects to
1573 @code{PTRDIFF_MAX} bytes, but larger objects would cause trouble
1574 anyway since they would break pointer subtraction, so this does not
1575 impose an arbitrary limit.
1578 Prefer @code{intptr_t} for internal representations of pointers, or
1579 for integers bounded only by the number of objects that can exist at
1580 any given time or by the total number of bytes that can be allocated.
1581 Currently Emacs sometimes uses other types when @code{intptr_t} would
1582 be better; fixing this is lower priority, as the code works as-is on
1583 Emacs's current porting targets.
1586 Prefer the Emacs-defined type @code{EMACS_INT} for representing values
1587 converted to or from Emacs Lisp fixnums, as fixnum arithmetic is based
1588 on @code{EMACS_INT}.
1591 When representing a system value (such as a file size or a count of
1592 seconds since the Epoch), prefer the corresponding system type (e.g.,
1593 @code{off_t}, @code{time_t}). Do not assume that a system type is
1594 signed, unless this assumption is known to be safe. For example,
1595 although @code{off_t} is always signed, @code{time_t} need not be.
1598 Prefer the Emacs-defined type @code{printmax_t} for representing
1599 values that might be any signed integer value that can be printed,
1600 using a @code{printf}-family function.
1603 Prefer @code{intmax_t} for representing values that might be any
1604 signed integer value.
1607 In bitfields, prefer @code{unsigned int} or @code{signed int} to
1608 @code{int}, as @code{int} is less portable: it might be signed, and
1609 might not be. Single-bit bit fields are invariably @code{unsigned
1610 int} so that their values are 0 and 1.
1613 In C, Emacs commonly uses @code{bool}, 1, and 0 for boolean values.
1614 Using @code{bool} for booleans can make programs easier to read and a
1615 bit faster than using @code{int}. Although it is also OK to use
1616 @code{int}, this older style is gradually being phased out. When
1617 using @code{bool}, respect the limitations of the replacement
1618 implementation of @code{bool}, as documented in the source file
1619 @file{lib/stdbool.in.h}, so that Emacs remains portable to pre-C99
1623 @c FIXME Mention src/globals.h somewhere in this file?