2 @c This is part of the GNU Emacs Lisp Reference Manual.
3 @c Copyright (C) 1990, 1991, 1992, 1993, 1998, 1999, 2001, 2002, 2003,
4 @c 2004, 2005, 2006, 2007, 2008 Free Software Foundation, Inc.
5 @c See the file elisp.texi for copying conditions.
6 @setfilename ../../info/internals
7 @node GNU Emacs Internals, Standard Errors, Tips, Top
8 @comment node-name, next, previous, up
9 @appendix GNU Emacs Internals
11 This chapter describes how the runnable Emacs executable is dumped with
12 the preloaded Lisp libraries in it, how storage is allocated, and some
13 internal aspects of GNU Emacs that may be of interest to C programmers.
16 * Building Emacs:: How the dumped Emacs is made.
17 * Pure Storage:: A kludge to make preloaded Lisp functions sharable.
18 * Garbage Collection:: Reclaiming space for Lisp objects no longer used.
19 * Memory Usage:: Info about total size of Lisp objects made so far.
20 * Writing Emacs Primitives:: Writing C code for Emacs.
21 * Object Internals:: Data formats of buffers, windows, processes.
25 @appendixsec Building Emacs
26 @cindex building Emacs
29 This section explains the steps involved in building the Emacs
30 executable. You don't have to know this material to build and install
31 Emacs, since the makefiles do all these things automatically. This
32 information is pertinent to Emacs maintenance.
34 Compilation of the C source files in the @file{src} directory
35 produces an executable file called @file{temacs}, also called a
36 @dfn{bare impure Emacs}. It contains the Emacs Lisp interpreter and I/O
37 routines, but not the editing commands.
39 @cindex @file{loadup.el}
40 The command @w{@samp{temacs -l loadup}} uses @file{temacs} to create
41 the real runnable Emacs executable. These arguments direct
42 @file{temacs} to evaluate the Lisp files specified in the file
43 @file{loadup.el}. These files set up the normal Emacs editing
44 environment, resulting in an Emacs that is still impure but no longer
48 It takes a substantial time to load the standard Lisp files. Luckily,
49 you don't have to do this each time you run Emacs; @file{temacs} can
50 dump out an executable program called @file{emacs} that has these files
51 preloaded. @file{emacs} starts more quickly because it does not need to
52 load the files. This is the Emacs executable that is normally
55 To create @file{emacs}, use the command @samp{temacs -batch -l loadup
56 dump}. The purpose of @samp{-batch} here is to prevent @file{temacs}
57 from trying to initialize any of its data on the terminal; this ensures
58 that the tables of terminal information are empty in the dumped Emacs.
59 The argument @samp{dump} tells @file{loadup.el} to dump a new executable
62 Some operating systems don't support dumping. On those systems, you
63 must start Emacs with the @samp{temacs -l loadup} command each time you
64 use it. This takes a substantial time, but since you need to start
65 Emacs once a day at most---or once a week if you never log out---the
66 extra time is not too severe a problem.
68 @cindex @file{site-load.el}
70 You can specify additional files to preload by writing a library named
71 @file{site-load.el} that loads them. You may need to add a definition
74 #define SITELOAD_PURESIZE_EXTRA @var{n}
78 to make @var{n} added bytes of pure space to hold the additional files.
79 (Try adding increments of 20000 until it is big enough.) However, the
80 advantage of preloading additional files decreases as machines get
81 faster. On modern machines, it is usually not advisable.
83 After @file{loadup.el} reads @file{site-load.el}, it finds the
84 documentation strings for primitive and preloaded functions (and
85 variables) in the file @file{etc/DOC} where they are stored, by
86 calling @code{Snarf-documentation} (@pxref{Definition of
87 Snarf-documentation,, Accessing Documentation}).
89 @cindex @file{site-init.el}
90 @cindex preloading additional functions and variables
91 You can specify other Lisp expressions to execute just before dumping
92 by putting them in a library named @file{site-init.el}. This file is
93 executed after the documentation strings are found.
95 If you want to preload function or variable definitions, there are
96 three ways you can do this and make their documentation strings
97 accessible when you subsequently run Emacs:
101 Arrange to scan these files when producing the @file{etc/DOC} file,
102 and load them with @file{site-load.el}.
105 Load the files with @file{site-init.el}, then copy the files into the
106 installation directory for Lisp files when you install Emacs.
109 Specify a non-@code{nil} value for
110 @code{byte-compile-dynamic-docstrings} as a local variable in each of these
111 files, and load them with either @file{site-load.el} or
112 @file{site-init.el}. (This method has the drawback that the
113 documentation strings take up space in Emacs all the time.)
116 It is not advisable to put anything in @file{site-load.el} or
117 @file{site-init.el} that would alter any of the features that users
118 expect in an ordinary unmodified Emacs. If you feel you must override
119 normal features for your site, do it with @file{default.el}, so that
120 users can override your changes if they wish. @xref{Startup Summary}.
122 In a package that can be preloaded, it is sometimes useful to
123 specify a computation to be done when Emacs subsequently starts up.
124 For this, use @code{eval-at-startup}:
126 @defmac eval-at-startup body@dots{}
127 This evaluates the @var{body} forms, either immediately if running in
128 an Emacs that has already started up, or later when Emacs does start
129 up. Since the value of the @var{body} forms is not necessarily
130 available when the @code{eval-at-startup} form is run, that form
131 always returns @code{nil}.
134 @defun dump-emacs to-file from-file
136 This function dumps the current state of Emacs into an executable file
137 @var{to-file}. It takes symbols from @var{from-file} (this is normally
138 the executable file @file{temacs}).
140 If you want to use this function in an Emacs that was already dumped,
141 you must run Emacs with @samp{-batch}.
145 @appendixsec Pure Storage
148 Emacs Lisp uses two kinds of storage for user-created Lisp objects:
149 @dfn{normal storage} and @dfn{pure storage}. Normal storage is where
150 all the new data created during an Emacs session are kept; see the
151 following section for information on normal storage. Pure storage is
152 used for certain data in the preloaded standard Lisp files---data that
153 should never change during actual use of Emacs.
155 Pure storage is allocated only while @file{temacs} is loading the
156 standard preloaded Lisp libraries. In the file @file{emacs}, it is
157 marked as read-only (on operating systems that permit this), so that
158 the memory space can be shared by all the Emacs jobs running on the
159 machine at once. Pure storage is not expandable; a fixed amount is
160 allocated when Emacs is compiled, and if that is not sufficient for
161 the preloaded libraries, @file{temacs} allocates dynamic memory for
162 the part that didn't fit. If that happens, you should increase the
163 compilation parameter @code{PURESIZE} in the file
164 @file{src/puresize.h} and rebuild Emacs, even though the resulting
165 image will work: garbage collection is disabled in this situation,
166 causing a memory leak. Such an overflow normally won't happen unless you
167 try to preload additional libraries or add features to the standard
168 ones. Emacs will display a warning about the overflow when it
171 @defun purecopy object
172 This function makes a copy in pure storage of @var{object}, and returns
173 it. It copies a string by simply making a new string with the same
174 characters, but without text properties, in pure storage. It
175 recursively copies the contents of vectors and cons cells. It does
176 not make copies of other objects such as symbols, but just returns
177 them unchanged. It signals an error if asked to copy markers.
179 This function is a no-op except while Emacs is being built and dumped;
180 it is usually called only in the file @file{emacs/lisp/loaddefs.el}, but
181 a few packages call it just in case you decide to preload them.
184 @defvar pure-bytes-used
185 The value of this variable is the number of bytes of pure storage
186 allocated so far. Typically, in a dumped Emacs, this number is very
187 close to the total amount of pure storage available---if it were not,
188 we would preallocate less.
192 This variable determines whether @code{defun} should make a copy of the
193 function definition in pure storage. If it is non-@code{nil}, then the
194 function definition is copied into pure storage.
196 This flag is @code{t} while loading all of the basic functions for
197 building Emacs initially (allowing those functions to be sharable and
198 non-collectible). Dumping Emacs as an executable always writes
199 @code{nil} in this variable, regardless of the value it actually has
200 before and after dumping.
202 You should not change this flag in a running Emacs.
205 @node Garbage Collection
206 @appendixsec Garbage Collection
207 @cindex garbage collection
209 @cindex memory allocation
210 When a program creates a list or the user defines a new function (such
211 as by loading a library), that data is placed in normal storage. If
212 normal storage runs low, then Emacs asks the operating system to
213 allocate more memory in blocks of 1k bytes. Each block is used for one
214 type of Lisp object, so symbols, cons cells, markers, etc., are
215 segregated in distinct blocks in memory. (Vectors, long strings,
216 buffers and certain other editing types, which are fairly large, are
217 allocated in individual blocks, one per object, while small strings are
218 packed into blocks of 8k bytes.)
220 It is quite common to use some storage for a while, then release it by
221 (for example) killing a buffer or deleting the last pointer to an
222 object. Emacs provides a @dfn{garbage collector} to reclaim this
223 abandoned storage. (This name is traditional, but ``garbage recycler''
224 might be a more intuitive metaphor for this facility.)
226 The garbage collector operates by finding and marking all Lisp objects
227 that are still accessible to Lisp programs. To begin with, it assumes
228 all the symbols, their values and associated function definitions, and
229 any data presently on the stack, are accessible. Any objects that can
230 be reached indirectly through other accessible objects are also
233 When marking is finished, all objects still unmarked are garbage. No
234 matter what the Lisp program or the user does, it is impossible to refer
235 to them, since there is no longer a way to reach them. Their space
236 might as well be reused, since no one will miss them. The second
237 (``sweep'') phase of the garbage collector arranges to reuse them.
239 @c ??? Maybe add something describing weak hash tables here?
242 The sweep phase puts unused cons cells onto a @dfn{free list}
243 for future allocation; likewise for symbols and markers. It compacts
244 the accessible strings so they occupy fewer 8k blocks; then it frees the
245 other 8k blocks. Vectors, buffers, windows, and other large objects are
246 individually allocated and freed using @code{malloc} and @code{free}.
248 @cindex CL note---allocate more storage
250 @b{Common Lisp note:} Unlike other Lisps, GNU Emacs Lisp does not
251 call the garbage collector when the free list is empty. Instead, it
252 simply requests the operating system to allocate more storage, and
253 processing continues until @code{gc-cons-threshold} bytes have been
256 This means that you can make sure that the garbage collector will not
257 run during a certain portion of a Lisp program by calling the garbage
258 collector explicitly just before it (provided that portion of the
259 program does not use so much space as to force a second garbage
263 @deffn Command garbage-collect
264 This command runs a garbage collection, and returns information on
265 the amount of space in use. (Garbage collection can also occur
266 spontaneously if you use more than @code{gc-cons-threshold} bytes of
267 Lisp data since the previous garbage collection.)
269 @code{garbage-collect} returns a list containing the following
274 ((@var{used-conses} . @var{free-conses})
275 (@var{used-syms} . @var{free-syms})
277 (@var{used-miscs} . @var{free-miscs})
278 @var{used-string-chars}
279 @var{used-vector-slots}
280 (@var{used-floats} . @var{free-floats})
281 (@var{used-intervals} . @var{free-intervals})
282 (@var{used-strings} . @var{free-strings}))
290 @result{} ((106886 . 13184) (9769 . 0)
291 (7731 . 4651) 347543 121628
292 (31 . 94) (1273 . 168)
297 Here is a table explaining each element:
301 The number of cons cells in use.
304 The number of cons cells for which space has been obtained from the
305 operating system, but that are not currently being used.
308 The number of symbols in use.
311 The number of symbols for which space has been obtained from the
312 operating system, but that are not currently being used.
315 The number of miscellaneous objects in use. These include markers and
316 overlays, plus certain objects not visible to users.
319 The number of miscellaneous objects for which space has been obtained
320 from the operating system, but that are not currently being used.
322 @item used-string-chars
323 The total size of all strings, in characters.
325 @item used-vector-slots
326 The total number of elements of existing vectors.
330 The number of floats in use.
334 The number of floats for which space has been obtained from the
335 operating system, but that are not currently being used.
338 The number of intervals in use. Intervals are an internal
339 data structure used for representing text properties.
342 The number of intervals for which space has been obtained
343 from the operating system, but that are not currently being used.
346 The number of strings in use.
349 The number of string headers for which the space was obtained from the
350 operating system, but which are currently not in use. (A string
351 object consists of a header and the storage for the string text
352 itself; the latter is only allocated when the string is created.)
355 If there was overflow in pure space (see the previous section),
356 @code{garbage-collect} returns @code{nil}, because a real garbage
357 collection can not be done in this situation.
360 @defopt garbage-collection-messages
361 If this variable is non-@code{nil}, Emacs displays a message at the
362 beginning and end of garbage collection. The default value is
363 @code{nil}, meaning there are no such messages.
367 This is a normal hook that is run at the end of garbage collection.
368 Garbage collection is inhibited while the hook functions run, so be
369 careful writing them.
372 @defopt gc-cons-threshold
373 The value of this variable is the number of bytes of storage that must
374 be allocated for Lisp objects after one garbage collection in order to
375 trigger another garbage collection. A cons cell counts as eight bytes,
376 a string as one byte per character plus a few bytes of overhead, and so
377 on; space allocated to the contents of buffers does not count. Note
378 that the subsequent garbage collection does not happen immediately when
379 the threshold is exhausted, but only the next time the Lisp evaluator is
382 The initial threshold value is 400,000. If you specify a larger
383 value, garbage collection will happen less often. This reduces the
384 amount of time spent garbage collecting, but increases total memory use.
385 You may want to do this when running a program that creates lots of
388 You can make collections more frequent by specifying a smaller value,
389 down to 10,000. A value less than 10,000 will remain in effect only
390 until the subsequent garbage collection, at which time
391 @code{garbage-collect} will set the threshold back to 10,000.
394 @defopt gc-cons-percentage
395 The value of this variable specifies the amount of consing before a
396 garbage collection occurs, as a fraction of the current heap size.
397 This criterion and @code{gc-cons-threshold} apply in parallel, and
398 garbage collection occurs only when both criteria are satisfied.
400 As the heap size increases, the time to perform a garbage collection
401 increases. Thus, it can be desirable to do them less frequently in
405 The value returned by @code{garbage-collect} describes the amount of
406 memory used by Lisp data, broken down by data type. By contrast, the
407 function @code{memory-limit} provides information on the total amount of
408 memory Emacs is currently using.
412 This function returns the address of the last byte Emacs has allocated,
413 divided by 1024. We divide the value by 1024 to make sure it fits in a
416 You can use this to get a general idea of how your actions affect the
421 This variable is @code{t} if Emacs is close to out of memory for Lisp
422 objects, and @code{nil} otherwise.
425 @defun memory-use-counts
426 This returns a list of numbers that count the number of objects
427 created in this Emacs session. Each of these counters increments for
428 a certain kind of object. See the documentation string for details.
432 This variable contains the total number of garbage collections
433 done so far in this Emacs session.
437 This variable contains the total number of seconds of elapsed time
438 during garbage collection so far in this Emacs session, as a floating
443 @section Memory Usage
446 These functions and variables give information about the total amount
447 of memory allocation that Emacs has done, broken down by data type.
448 Note the difference between these and the values returned by
449 @code{(garbage-collect)}; those count objects that currently exist, but
450 these count the number or size of all allocations, including those for
451 objects that have since been freed.
453 @defvar cons-cells-consed
454 The total number of cons cells that have been allocated so far
455 in this Emacs session.
458 @defvar floats-consed
459 The total number of floats that have been allocated so far
460 in this Emacs session.
463 @defvar vector-cells-consed
464 The total number of vector cells that have been allocated so far
465 in this Emacs session.
468 @defvar symbols-consed
469 The total number of symbols that have been allocated so far
470 in this Emacs session.
473 @defvar string-chars-consed
474 The total number of string characters that have been allocated so far
475 in this Emacs session.
478 @defvar misc-objects-consed
479 The total number of miscellaneous objects that have been allocated so
480 far in this Emacs session. These include markers and overlays, plus
481 certain objects not visible to users.
484 @defvar intervals-consed
485 The total number of intervals that have been allocated so far
486 in this Emacs session.
489 @defvar strings-consed
490 The total number of strings that have been allocated so far in this
494 @node Writing Emacs Primitives
495 @appendixsec Writing Emacs Primitives
496 @cindex primitive function internals
497 @cindex writing Emacs primitives
499 Lisp primitives are Lisp functions implemented in C. The details of
500 interfacing the C function so that Lisp can call it are handled by a few
501 C macros. The only way to really understand how to write new C code is
502 to read the source, but we can explain some things here.
504 An example of a special form is the definition of @code{or}, from
505 @file{eval.c}. (An ordinary function would have the same general
508 @cindex garbage collection protection
511 DEFUN ("or", For, Sor, 0, UNEVALLED, 0,
512 doc: /* Eval args until one of them yields non-nil, then return that
513 value. The remaining args are not evalled at all.
514 If all args return nil, return nil.
517 usage: (or CONDITIONS ...) */)
521 register Lisp_Object val = Qnil;
532 val = Feval (XCAR (args));
546 @cindex @code{DEFUN}, C macro to define Lisp primitives
547 Let's start with a precise explanation of the arguments to the
548 @code{DEFUN} macro. Here is a template for them:
551 DEFUN (@var{lname}, @var{fname}, @var{sname}, @var{min}, @var{max}, @var{interactive}, @var{doc})
556 This is the name of the Lisp symbol to define as the function name; in
557 the example above, it is @code{or}.
560 This is the C function name for this function. This is
561 the name that is used in C code for calling the function. The name is,
562 by convention, @samp{F} prepended to the Lisp name, with all dashes
563 (@samp{-}) in the Lisp name changed to underscores. Thus, to call this
564 function from C code, call @code{For}. Remember that the arguments must
565 be of type @code{Lisp_Object}; various macros and functions for creating
566 values of type @code{Lisp_Object} are declared in the file
570 This is a C variable name to use for a structure that holds the data for
571 the subr object that represents the function in Lisp. This structure
572 conveys the Lisp symbol name to the initialization routine that will
573 create the symbol and store the subr object as its definition. By
574 convention, this name is always @var{fname} with @samp{F} replaced with
578 This is the minimum number of arguments that the function requires. The
579 function @code{or} allows a minimum of zero arguments.
582 This is the maximum number of arguments that the function accepts, if
583 there is a fixed maximum. Alternatively, it can be @code{UNEVALLED},
584 indicating a special form that receives unevaluated arguments, or
585 @code{MANY}, indicating an unlimited number of evaluated arguments (the
586 equivalent of @code{&rest}). Both @code{UNEVALLED} and @code{MANY} are
587 macros. If @var{max} is a number, it may not be less than @var{min} and
588 it may not be greater than eight.
591 This is an interactive specification, a string such as might be used as
592 the argument of @code{interactive} in a Lisp function. In the case of
593 @code{or}, it is 0 (a null pointer), indicating that @code{or} cannot be
594 called interactively. A value of @code{""} indicates a function that
595 should receive no arguments when called interactively. If the value
596 begins with a @samp{(}, the string is evaluated as a Lisp form.
599 This is the documentation string. It uses C comment syntax rather
600 than C string syntax because comment syntax requires nothing special
601 to include multiple lines. The @samp{doc:} identifies the comment
602 that follows as the documentation string. The @samp{/*} and @samp{*/}
603 delimiters that begin and end the comment are not part of the
604 documentation string.
606 If the last line of the documentation string begins with the keyword
607 @samp{usage:}, the rest of the line is treated as the argument list
608 for documentation purposes. This way, you can use different argument
609 names in the documentation string from the ones used in the C code.
610 @samp{usage:} is required if the function has an unlimited number of
613 All the usual rules for documentation strings in Lisp code
614 (@pxref{Documentation Tips}) apply to C code documentation strings
618 After the call to the @code{DEFUN} macro, you must write the argument
619 name list that every C function must have, followed by ordinary C
620 declarations for the arguments. For a function with a fixed maximum
621 number of arguments, declare a C argument for each Lisp argument, and
622 give them all type @code{Lisp_Object}. When a Lisp function has no
623 upper limit on the number of arguments, its implementation in C actually
624 receives exactly two arguments: the first is the number of Lisp
625 arguments, and the second is the address of a block containing their
626 values. They have types @code{int} and @w{@code{Lisp_Object *}}.
628 @cindex @code{GCPRO} and @code{UNGCPRO}
629 @cindex protect C variables from garbage collection
630 Within the function @code{For} itself, note the use of the macros
631 @code{GCPRO1} and @code{UNGCPRO}. @code{GCPRO1} is used to
632 ``protect'' a variable from garbage collection---to inform the garbage
633 collector that it must look in that variable and regard its contents
634 as an accessible object. GC protection is necessary whenever you call
635 @code{Feval} or anything that can directly or indirectly call
636 @code{Feval}. At such a time, any Lisp object that this function may
637 refer to again must be protected somehow.
639 It suffices to ensure that at least one pointer to each object is
640 GC-protected; that way, the object cannot be recycled, so all pointers
641 to it remain valid. Thus, a particular local variable can do without
642 protection if it is certain that the object it points to will be
643 preserved by some other pointer (such as another local variable which
644 has a @code{GCPRO})@footnote{Formerly, strings were a special
645 exception; in older Emacs versions, every local variable that might
646 point to a string needed a @code{GCPRO}.}. Otherwise, the local
647 variable needs a @code{GCPRO}.
649 The macro @code{GCPRO1} protects just one local variable. If you
650 want to protect two variables, use @code{GCPRO2} instead; repeating
651 @code{GCPRO1} will not work. Macros @code{GCPRO3}, @code{GCPRO4},
652 @code{GCPRO5}, and @code{GCPRO6} also exist. All these macros
653 implicitly use local variables such as @code{gcpro1}; you must declare
654 these explicitly, with type @code{struct gcpro}. Thus, if you use
655 @code{GCPRO2}, you must declare @code{gcpro1} and @code{gcpro2}.
656 Alas, we can't explain all the tricky details here.
658 @code{UNGCPRO} cancels the protection of the variables that are
659 protected in the current function. It is necessary to do this
662 Built-in functions that take a variable number of arguments actually
663 accept two arguments at the C level: the number of Lisp arguments, and
664 a @code{Lisp_Object *} pointer to a C vector containing those Lisp
665 arguments. This C vector may be part of a Lisp vector, but it need
666 not be. The responsibility for using @code{GCPRO} to protect the Lisp
667 arguments from GC if necessary rests with the caller in this case,
668 since the caller allocated or found the storage for them.
670 You must not use C initializers for static or global variables unless
671 the variables are never written once Emacs is dumped. These variables
672 with initializers are allocated in an area of memory that becomes
673 read-only (on certain operating systems) as a result of dumping Emacs.
676 Do not use static variables within functions---place all static
677 variables at top level in the file. This is necessary because Emacs on
678 some operating systems defines the keyword @code{static} as a null
679 macro. (This definition is used because those systems put all variables
680 declared static in a place that becomes read-only after dumping, whether
681 they have initializers or not.)
683 @cindex @code{defsubr}, Lisp symbol for a primitive
684 Defining the C function is not enough to make a Lisp primitive
685 available; you must also create the Lisp symbol for the primitive and
686 store a suitable subr object in its function cell. The code looks like
690 defsubr (&@var{subr-structure-name});
694 Here @var{subr-structure-name} is the name you used as the third
695 argument to @code{DEFUN}.
697 If you add a new primitive to a file that already has Lisp primitives
698 defined in it, find the function (near the end of the file) named
699 @code{syms_of_@var{something}}, and add the call to @code{defsubr}
700 there. If the file doesn't have this function, or if you create a new
701 file, add to it a @code{syms_of_@var{filename}} (e.g.,
702 @code{syms_of_myfile}). Then find the spot in @file{emacs.c} where all
703 of these functions are called, and add a call to
704 @code{syms_of_@var{filename}} there.
706 @anchor{Defining Lisp variables in C}
707 @vindex byte-boolean-vars
708 @cindex defining Lisp variables in C
709 @cindex @code{DEFVAR_INT}, @code{DEFVAR_LISP}, @code{DEFVAR_BOOL}
710 The function @code{syms_of_@var{filename}} is also the place to define
711 any C variables that are to be visible as Lisp variables.
712 @code{DEFVAR_LISP} makes a C variable of type @code{Lisp_Object} visible
713 in Lisp. @code{DEFVAR_INT} makes a C variable of type @code{int}
714 visible in Lisp with a value that is always an integer.
715 @code{DEFVAR_BOOL} makes a C variable of type @code{int} visible in Lisp
716 with a value that is either @code{t} or @code{nil}. Note that variables
717 defined with @code{DEFVAR_BOOL} are automatically added to the list
718 @code{byte-boolean-vars} used by the byte compiler.
720 @cindex @code{staticpro}, protection from GC
721 If you define a file-scope C variable of type @code{Lisp_Object},
722 you must protect it from garbage-collection by calling @code{staticpro}
723 in @code{syms_of_@var{filename}}, like this:
726 staticpro (&@var{variable});
729 Here is another example function, with more complicated arguments.
730 This comes from the code in @file{window.c}, and it demonstrates the use
731 of macros and functions to manipulate Lisp objects.
735 DEFUN ("coordinates-in-window-p", Fcoordinates_in_window_p,
736 Scoordinates_in_window_p, 2, 2,
737 "xSpecify coordinate pair: \nXExpression which evals to window: ",
738 "Return non-nil if COORDINATES is in WINDOW.\n\
739 COORDINATES is a cons of the form (X . Y), X and Y being distances\n\
743 If they are on the border between WINDOW and its right sibling,\n\
744 `vertical-line' is returned.")
745 (coordinates, window)
746 register Lisp_Object coordinates, window;
752 CHECK_LIVE_WINDOW (window, 0);
753 CHECK_CONS (coordinates, 1);
754 x = XINT (Fcar (coordinates));
755 y = XINT (Fcdr (coordinates));
759 switch (coordinates_in_window (XWINDOW (window), &x, &y))
761 case 0: /* NOT in window at all. */
766 case 1: /* In text part of window. */
767 return Fcons (make_number (x), make_number (y));
771 case 2: /* In mode line of window. */
776 case 3: /* On right border of window. */
777 return Qvertical_line;
788 Note that C code cannot call functions by name unless they are defined
789 in C. The way to call a function written in Lisp is to use
790 @code{Ffuncall}, which embodies the Lisp function @code{funcall}. Since
791 the Lisp function @code{funcall} accepts an unlimited number of
792 arguments, in C it takes two: the number of Lisp-level arguments, and a
793 one-dimensional array containing their values. The first Lisp-level
794 argument is the Lisp function to call, and the rest are the arguments to
795 pass to it. Since @code{Ffuncall} can call the evaluator, you must
796 protect pointers from garbage collection around the call to
799 The C functions @code{call0}, @code{call1}, @code{call2}, and so on,
800 provide handy ways to call a Lisp function conveniently with a fixed
801 number of arguments. They work by calling @code{Ffuncall}.
803 @file{eval.c} is a very good file to look through for examples;
804 @file{lisp.h} contains the definitions for some important macros and
807 If you define a function which is side-effect free, update the code
808 in @file{byte-opt.el} which binds @code{side-effect-free-fns} and
809 @code{side-effect-and-error-free-fns} so that the compiler optimizer
812 @node Object Internals
813 @appendixsec Object Internals
814 @cindex object internals
816 GNU Emacs Lisp manipulates many different types of data. The actual
817 data are stored in a heap and the only access that programs have to it
818 is through pointers. Pointers are thirty-two bits wide in most
819 implementations. Depending on the operating system and type of machine
820 for which you compile Emacs, twenty-nine bits are used to address the
821 object, and the remaining three bits are used for the tag that
822 identifies the object's type.
824 Because Lisp objects are represented as tagged pointers, it is always
825 possible to determine the Lisp data type of any object. The C data type
826 @code{Lisp_Object} can hold any Lisp object of any data type. Ordinary
827 variables have type @code{Lisp_Object}, which means they can hold any
828 type of Lisp value; you can determine the actual data type only at run
829 time. The same is true for function arguments; if you want a function
830 to accept only a certain type of argument, you must check the type
831 explicitly using a suitable predicate (@pxref{Type Predicates}).
832 @cindex type checking internals
835 * Buffer Internals:: Components of a buffer structure.
836 * Window Internals:: Components of a window structure.
837 * Process Internals:: Components of a process structure.
840 @node Buffer Internals
841 @appendixsubsec Buffer Internals
842 @cindex internals, of buffer
843 @cindex buffer internals
845 Buffers contain fields not directly accessible by the Lisp programmer.
846 We describe them here, naming them by the names used in the C code.
847 Many are accessible indirectly in Lisp programs via Lisp primitives.
849 Two structures are used to represent buffers in C. The
850 @code{buffer_text} structure contains fields describing the text of a
851 buffer; the @code{buffer} structure holds other fields. In the case
852 of indirect buffers, two or more @code{buffer} structures reference
853 the same @code{buffer_text} structure.
855 Here is a list of the @code{struct buffer_text} fields:
859 This field contains the actual address of the buffer contents.
862 This holds the character position of the gap in the buffer.
866 This field contains the character position of the end of the buffer
870 Contains the byte position of the gap.
873 Holds the byte position of the end of the buffer text.
876 Contains the size of buffer's gap. @xref{Buffer Gap}.
879 This field counts buffer-modification events for this buffer. It is
880 incremented for each such event, and never otherwise changed.
883 Contains the previous value of @code{modiff}, as of the last time a
884 buffer was visited or saved in a file.
887 Counts modifications to overlays analogous to @code{modiff}.
890 Holds the number of characters at the start of the text that are known
891 to be unchanged since the last redisplay that finished.
894 Holds the number of characters at the end of the text that are known to
895 be unchanged since the last redisplay that finished.
897 @item unchanged_modified
898 Contains the value of @code{modiff} at the time of the last redisplay
899 that finished. If this value matches @code{modiff},
900 @code{beg_unchanged} and @code{end_unchanged} contain no useful
903 @item overlay_unchanged_modified
904 Contains the value of @code{overlay_modiff} at the time of the last
905 redisplay that finished. If this value matches @code{overlay_modiff},
906 @code{beg_unchanged} and @code{end_unchanged} contain no useful
910 The markers that refer to this buffer. This is actually a single
911 marker, and successive elements in its marker @code{chain} are the other
912 markers referring to this buffer text.
915 Contains the interval tree which records the text properties of this
919 The fields of @code{struct buffer} are:
923 Points to the next buffer, in the chain of all buffers including killed
924 buffers. This chain is used only for garbage collection, in order to
925 collect killed buffers properly. Note that vectors, and most kinds of
926 objects allocated as vectors, are all on one chain, but buffers are on a
927 separate chain of their own.
930 This is a @code{struct buffer_text} structure. In an ordinary buffer,
931 it holds the buffer contents. In indirect buffers, this field is not
935 This points to the @code{buffer_text} structure that is used for this
936 buffer. In an ordinary buffer, this is the @code{own_text} field above.
937 In an indirect buffer, this is the @code{own_text} field of the base
941 Contains the character position of point in a buffer.
944 Contains the byte position of point in a buffer.
947 This field contains the character position of the beginning of the
948 accessible range of text in the buffer.
951 This field contains the byte position of the beginning of the
952 accessible range of text in the buffer.
955 This field contains the character position of the end of the
956 accessible range of text in the buffer.
959 This field contains the byte position of the end of the
960 accessible range of text in the buffer.
963 In an indirect buffer, this points to the base buffer. In an ordinary
966 @item local_var_flags
967 This field contains flags indicating that certain variables are local in
968 this buffer. Such variables are declared in the C code using
969 @code{DEFVAR_PER_BUFFER}, and their buffer-local bindings are stored in
970 fields in the buffer structure itself. (Some of these fields are
971 described in this table.)
974 This field contains the modification time of the visited file. It is
975 set when the file is written or read. Before writing the buffer into a
976 file, this field is compared to the modification time of the file to see
977 if the file has changed on disk. @xref{Buffer Modification}.
979 @item auto_save_modified
980 This field contains the time when the buffer was last auto-saved.
982 @item auto_save_failure_time
983 The time at which we detected a failure to auto-save, or -1 if we didn't
986 @item last_window_start
987 This field contains the @code{window-start} position in the buffer as of
988 the last time the buffer was displayed in a window.
991 This flag is set when narrowing changes in a buffer.
993 @item prevent_redisplay_optimizations_p
994 this flag indicates that redisplay optimizations should not be used
995 to display this buffer.
998 This field points to the buffer's undo list. @xref{Undo}.
1001 The buffer name is a string that names the buffer. It is guaranteed to
1002 be unique. @xref{Buffer Names}.
1005 The name of the file visited in this buffer, or @code{nil}.
1008 The directory for expanding relative file names.
1011 Length of the file this buffer is visiting, when last read or saved.
1012 This and other fields concerned with saving are not kept in the
1013 @code{buffer_text} structure because indirect buffers are never saved.
1015 @item auto_save_file_name
1016 File name used for auto-saving this buffer. This is not in the
1017 @code{buffer_text} because it's not used in indirect buffers at all.
1020 Non-@code{nil} means this buffer is read-only.
1023 This field contains the mark for the buffer. The mark is a marker,
1024 hence it is also included on the list @code{markers}. @xref{The Mark}.
1026 @item local_var_alist
1027 This field contains the association list describing the buffer-local
1028 variable bindings of this buffer, not including the built-in
1029 buffer-local bindings that have special slots in the buffer object.
1030 (Those slots are omitted from this table.) @xref{Buffer-Local
1034 Symbol naming the major mode of this buffer, e.g., @code{lisp-mode}.
1037 Pretty name of major mode, e.g., @code{"Lisp"}.
1039 @item mode_line_format
1040 Mode line element that controls the format of the mode line. If this
1041 is @code{nil}, no mode line will be displayed.
1043 @item header_line_format
1044 This field is analogous to @code{mode_line_format} for the mode
1045 line displayed at the top of windows.
1048 This field holds the buffer's local keymap. @xref{Keymaps}.
1051 This buffer's local abbrevs.
1054 This field contains the syntax table for the buffer. @xref{Syntax Tables}.
1056 @item category_table
1057 This field contains the category table for the buffer.
1059 @item case_fold_search
1060 The value of @code{case-fold-search} in this buffer.
1063 The value of @code{tab-width} in this buffer.
1066 The value of @code{fill-column} in this buffer.
1069 The value of @code{left-margin} in this buffer.
1071 @item auto_fill_function
1072 The value of @code{auto-fill-function} in this buffer.
1074 @item downcase_table
1075 This field contains the conversion table for converting text to lower case.
1079 This field contains the conversion table for converting text to upper case.
1082 @item case_canon_table
1083 This field contains the conversion table for canonicalizing text for
1084 case-folding search. @xref{Case Tables}.
1086 @item case_eqv_table
1087 This field contains the equivalence table for case-folding search.
1090 @item truncate_lines
1091 The value of @code{truncate-lines} in this buffer.
1094 The value of @code{ctl-arrow} in this buffer.
1096 @item selective_display
1097 The value of @code{selective-display} in this buffer.
1099 @item selective_display_ellipsis
1100 The value of @code{selective-display-ellipsis} in this buffer.
1103 An alist of the minor modes of this buffer.
1105 @item overwrite_mode
1106 The value of @code{overwrite_mode} in this buffer.
1109 The value of @code{abbrev-mode} in this buffer.
1112 This field contains the buffer's display table, or @code{nil} if it doesn't
1113 have one. @xref{Display Tables}.
1116 This field contains the time when the buffer was last saved, as an integer.
1117 @xref{Buffer Modification}.
1120 This field is non-@code{nil} if the buffer's mark is active.
1122 @item overlays_before
1123 This field holds a list of the overlays in this buffer that end at or
1124 before the current overlay center position. They are sorted in order of
1125 decreasing end position.
1127 @item overlays_after
1128 This field holds a list of the overlays in this buffer that end after
1129 the current overlay center position. They are sorted in order of
1130 increasing beginning position.
1132 @item overlay_center
1133 This field holds the current overlay center position. @xref{Overlays}.
1135 @item enable_multibyte_characters
1136 This field holds the buffer's local value of
1137 @code{enable-multibyte-characters}---either @code{t} or @code{nil}.
1139 @item buffer_file_coding_system
1140 The value of @code{buffer-file-coding-system} in this buffer.
1143 The value of @code{buffer-file-format} in this buffer.
1145 @item auto_save_file_format
1146 The value of @code{buffer-auto-save-file-format} in this buffer.
1149 In an indirect buffer, or a buffer that is the base of an indirect
1150 buffer, this holds a marker that records point for this buffer when the
1151 buffer is not current.
1154 In an indirect buffer, or a buffer that is the base of an indirect
1155 buffer, this holds a marker that records @code{begv} for this buffer
1156 when the buffer is not current.
1159 In an indirect buffer, or a buffer that is the base of an indirect
1160 buffer, this holds a marker that records @code{zv} for this buffer when
1161 the buffer is not current.
1164 The truename of the visited file, or @code{nil}.
1166 @item invisibility_spec
1167 The value of @code{buffer-invisibility-spec} in this buffer.
1169 @item last_selected_window
1170 This is the last window that was selected with this buffer in it, or @code{nil}
1171 if that window no longer displays this buffer.
1174 This field is incremented each time the buffer is displayed in a window.
1176 @item left_margin_width
1177 The value of @code{left-margin-width} in this buffer.
1179 @item right_margin_width
1180 The value of @code{right-margin-width} in this buffer.
1182 @item indicate_empty_lines
1183 Non-@code{nil} means indicate empty lines (lines with no text) with a
1184 small bitmap in the fringe, when using a window system that can do it.
1187 This holds a time stamp that is updated each time this buffer is
1188 displayed in a window.
1190 @item scroll_up_aggressively
1191 The value of @code{scroll-up-aggressively} in this buffer.
1193 @item scroll_down_aggressively
1194 The value of @code{scroll-down-aggressively} in this buffer.
1197 @node Window Internals
1198 @appendixsubsec Window Internals
1199 @cindex internals, of window
1200 @cindex window internals
1202 Windows have the following accessible fields:
1206 The frame that this window is on.
1209 Non-@code{nil} if this window is a minibuffer window.
1212 Internally, Emacs arranges windows in a tree; each group of siblings has
1213 a parent window whose area includes all the siblings. This field points
1214 to a window's parent.
1216 Parent windows do not display buffers, and play little role in display
1217 except to shape their child windows. Emacs Lisp programs usually have
1218 no access to the parent windows; they operate on the windows at the
1219 leaves of the tree, which actually display buffers.
1221 The following four fields also describe the window tree structure.
1224 In a window subdivided horizontally by child windows, the leftmost child.
1225 Otherwise, @code{nil}.
1228 In a window subdivided vertically by child windows, the topmost child.
1229 Otherwise, @code{nil}.
1232 The next sibling of this window. It is @code{nil} in a window that is
1233 the rightmost or bottommost of a group of siblings.
1236 The previous sibling of this window. It is @code{nil} in a window that
1237 is the leftmost or topmost of a group of siblings.
1240 This is the left-hand edge of the window, measured in columns. (The
1241 leftmost column on the screen is @w{column 0}.)
1244 This is the top edge of the window, measured in lines. (The top line on
1245 the screen is @w{line 0}.)
1248 The height of the window, measured in lines.
1251 The width of the window, measured in columns. This width includes the
1252 scroll bar and fringes, and/or the separator line on the right of the
1256 The buffer that the window is displaying. This may change often during
1257 the life of the window.
1260 The position in the buffer that is the first character to be displayed
1264 @cindex window point internals
1265 This is the value of point in the current buffer when this window is
1266 selected; when it is not selected, it retains its previous value.
1269 If this flag is non-@code{nil}, it says that the window has been
1270 scrolled explicitly by the Lisp program. This affects what the next
1271 redisplay does if point is off the screen: instead of scrolling the
1272 window to show the text around point, it moves point to a location that
1275 @item frozen_window_start_p
1276 This field is set temporarily to 1 to indicate to redisplay that
1277 @code{start} of this window should not be changed, even if point
1280 @item start_at_line_beg
1281 Non-@code{nil} means current value of @code{start} was the beginning of a line
1285 Non-@code{nil} means don't delete this window for becoming ``too small.''
1287 @item height_fixed_p
1288 This field is temporarily set to 1 to fix the height of the selected
1289 window when the echo area is resized.
1292 This is the last time that the window was selected. The function
1293 @code{get-lru-window} uses this field.
1295 @item sequence_number
1296 A unique number assigned to this window when it was created.
1299 The @code{modiff} field of the window's buffer, as of the last time
1300 a redisplay completed in this window.
1302 @item last_overlay_modified
1303 The @code{overlay_modiff} field of the window's buffer, as of the last
1304 time a redisplay completed in this window.
1307 The buffer's value of point, as of the last time a redisplay completed
1311 A non-@code{nil} value means the window's buffer was ``modified'' when the
1312 window was last updated.
1314 @item vertical_scroll_bar
1315 This window's vertical scroll bar.
1317 @item left_margin_width
1318 The width of the left margin in this window, or @code{nil} not to
1319 specify it (in which case the buffer's value of @code{left-margin-width}
1322 @item right_margin_width
1323 Likewise for the right margin.
1331 @item window_end_pos
1332 This is computed as @code{z} minus the buffer position of the last glyph
1333 in the current matrix of the window. The value is only valid if
1334 @code{window_end_valid} is not @code{nil}.
1336 @item window_end_bytepos
1337 The byte position corresponding to @code{window_end_pos}.
1339 @item window_end_vpos
1340 The window-relative vertical position of the line containing
1341 @code{window_end_pos}.
1343 @item window_end_valid
1344 This field is set to a non-@code{nil} value if @code{window_end_pos} is truly
1345 valid. This is @code{nil} if nontrivial redisplay is preempted since in that
1346 case the display that @code{window_end_pos} was computed for did not get
1349 @item redisplay_end_trigger
1350 If redisplay in this window goes beyond this buffer position, it runs
1351 the @code{redisplay-end-trigger-hook}.
1356 ??? Are temporary storage areas.
1360 A structure describing where the cursor is in this window.
1363 The value of @code{cursor} as of the last redisplay that finished.
1366 A structure describing where the cursor of this window physically is.
1368 @item phys_cursor_type
1369 The type of cursor that was last displayed on this window.
1371 @item phys_cursor_on_p
1372 This field is non-zero if the cursor is physically on.
1375 Non-zero means the cursor in this window is logically on.
1377 @item last_cursor_off_p
1378 This field contains the value of @code{cursor_off_p} as of the time of
1381 @item must_be_updated_p
1382 This is set to 1 during redisplay when this window must be updated.
1385 This is the number of columns that the display in the window is scrolled
1386 horizontally to the left. Normally, this is 0.
1389 Vertical scroll amount, in pixels. Normally, this is 0.
1392 Non-@code{nil} if this window is dedicated to its buffer.
1395 The window's display table, or @code{nil} if none is specified for it.
1397 @item update_mode_line
1398 Non-@code{nil} means this window's mode line needs to be updated.
1400 @item base_line_number
1401 The line number of a certain position in the buffer, or @code{nil}.
1402 This is used for displaying the line number of point in the mode line.
1405 The position in the buffer for which the line number is known, or
1406 @code{nil} meaning none is known.
1408 @item region_showing
1409 If the region (or part of it) is highlighted in this window, this field
1410 holds the mark position that made one end of that region. Otherwise,
1411 this field is @code{nil}.
1413 @item column_number_displayed
1414 The column number currently displayed in this window's mode line, or @code{nil}
1415 if column numbers are not being displayed.
1417 @item current_matrix
1418 A glyph matrix describing the current display of this window.
1420 @item desired_matrix
1421 A glyph matrix describing the desired display of this window.
1424 @node Process Internals
1425 @appendixsubsec Process Internals
1426 @cindex internals, of process
1427 @cindex process internals
1429 The fields of a process are:
1433 A string, the name of the process.
1436 A list containing the command arguments that were used to start this
1437 process. For a network or serial process, it is @code{nil} if the
1438 process is running or @code{t} if the process is stopped.
1441 A function used to accept output from the process instead of a buffer,
1445 A function called whenever the process receives a signal, or @code{nil}.
1448 The associated buffer of the process.
1451 An integer, the operating system's process @acronym{ID}.
1455 A flag, non-@code{nil} if this is really a child process.
1456 It is @code{nil} for a network or serial connection.
1459 A marker indicating the position of the end of the last output from this
1460 process inserted into the buffer. This is often but not always the end
1463 @item kill_without_query
1464 If this is non-@code{nil}, killing Emacs while this process is still
1465 running does not ask for confirmation about killing the process.
1467 @item raw_status_low
1468 @itemx raw_status_high
1469 These two fields record 16 bits each of the process status returned by
1470 the @code{wait} system call.
1473 The process status, as @code{process-status} should return it.
1477 If these two fields are not equal, a change in the status of the process
1478 needs to be reported, either by running the sentinel or by inserting a
1479 message in the process buffer.
1482 Non-@code{nil} if communication with the subprocess uses a @acronym{PTY};
1483 @code{nil} if it uses a pipe.
1486 The file descriptor for input from the process.
1489 The file descriptor for output to the process.
1492 The file descriptor for the terminal that the subprocess is using. (On
1493 some systems, there is no need to record this, so the value is
1497 The name of the terminal that the subprocess is using,
1498 or @code{nil} if it is using pipes.
1500 @item decode_coding_system
1501 Coding-system for decoding the input from this process.
1504 A working buffer for decoding.
1506 @item decoding_carryover
1507 Size of carryover in decoding.
1509 @item encode_coding_system
1510 Coding-system for encoding the output to this process.
1513 A working buffer for encoding.
1515 @item encoding_carryover
1516 Size of carryover in encoding.
1518 @item inherit_coding_system_flag
1519 Flag to set @code{coding-system} of the process buffer from the
1520 coding system used to decode process output.
1523 Symbol indicating the type of process: @code{real}, @code{network},
1529 arch-tag: 4b2c33bc-d7e4-43f5-bc20-27c0db52a53e