| 1 | @c -*-texinfo-*- |
| 2 | @c This is part of the GNU Emacs Lisp Reference Manual. |
| 3 | @c Copyright (C) 1990, 1991, 1992, 1993 Free Software Foundation, Inc. |
| 4 | @c See the file elisp.texi for copying conditions. |
| 5 | @setfilename ../info/internals |
| 6 | @node GNU Emacs Internals, Standard Errors, Tips, Top |
| 7 | @comment node-name, next, previous, up |
| 8 | @appendix GNU Emacs Internals |
| 9 | |
| 10 | This chapter describes how the runnable Emacs executable is dumped with |
| 11 | the preloaded Lisp libraries in it, how storage is allocated, and some |
| 12 | internal aspects of GNU Emacs that may be of interest to C programmers. |
| 13 | |
| 14 | @menu |
| 15 | * Building Emacs:: How to preload Lisp libraries into Emacs. |
| 16 | * Pure Storage:: A kludge to make preloaded Lisp functions sharable. |
| 17 | * Garbage Collection:: Reclaiming space for Lisp objects no longer used. |
| 18 | * Writing Emacs Primitives:: Writing C code for Emacs. |
| 19 | * Object Internals:: Data formats of buffers, windows, processes. |
| 20 | @end menu |
| 21 | |
| 22 | @node Building Emacs, Pure Storage, GNU Emacs Internals, GNU Emacs Internals |
| 23 | @appendixsec Building Emacs |
| 24 | @cindex building Emacs |
| 25 | @pindex temacs |
| 26 | |
| 27 | This section explains the steps involved in building the Emacs |
| 28 | executable. You don't have to know this material to build and install |
| 29 | Emacs, since the makefiles do all these things automatically. This |
| 30 | information is pertinent to Emacs maintenance. |
| 31 | |
| 32 | Compilation of the C source files in the @file{src} directory |
| 33 | produces an executable file called @file{temacs}, also called a |
| 34 | @dfn{bare impure Emacs}. It contains the Emacs Lisp interpreter and I/O |
| 35 | routines, but not the editing commands. |
| 36 | |
| 37 | @cindex @file{loadup.el} |
| 38 | The command @w{@samp{temacs -l loadup}} uses @file{temacs} to create |
| 39 | the real runnable Emacs executable. These arguments direct |
| 40 | @file{temacs} to evaluate the Lisp files specified in the file |
| 41 | @file{loadup.el}. These files set up the normal Emacs editing |
| 42 | environment, resulting in an Emacs that is still impure but no longer |
| 43 | bare. |
| 44 | |
| 45 | It takes a substantial time to load the standard Lisp files. Luckily, |
| 46 | you don't have to do this each time you run Emacs; @file{temacs} can |
| 47 | dump out an executable program called @file{emacs} that has these files |
| 48 | preloaded. @file{emacs} starts more quickly because it does not need to |
| 49 | load the files. This is the Emacs executable that is normally |
| 50 | installed. |
| 51 | |
| 52 | To create @file{emacs}, use the command @samp{temacs -batch -l loadup |
| 53 | dump}. The purpose of @samp{-batch} here is to prevent @file{temacs} |
| 54 | from trying to initialize any of its data on the terminal; this ensures |
| 55 | that the tables of terminal information are empty in the dumped Emacs. |
| 56 | The argument @samp{dump} tells @file{loadup.el} to dump a new executable |
| 57 | named @file{emacs}. |
| 58 | |
| 59 | Some operating systems don't support dumping. On those systems, you |
| 60 | must start Emacs with the @samp{temacs -l loadup} command each time you |
| 61 | use it. This takes a substantial time, but since you need to start |
| 62 | Emacs once a day at most---or once a week if you never log out---the |
| 63 | extra time is not too severe a problem. |
| 64 | |
| 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 increase the value |
| 68 | of @code{PURESIZE}, in @file{src/puresize.h}, to make room for the |
| 69 | additional data. (Try adding increments of 20000 until it is big |
| 70 | enough.) However, the advantage of preloading additional files |
| 71 | decreases as machines get faster. On modern machines, it is usually not |
| 72 | advisable. |
| 73 | |
| 74 | After @file{loadup.el} reads @file{site-load.el}, it finds the |
| 75 | documentation strings for primitive and preloaded functions (and |
| 76 | variables) in the file @file{etc/DOC} where they are stored, by calling |
| 77 | @code{Snarf-documentation} (@pxref{Accessing Documentation}). |
| 78 | |
| 79 | @cindex @file{site-init.el} |
| 80 | You can specify other Lisp expressions to execute just before dumping |
| 81 | by putting them in a library named @file{site-init.el}. This file is |
| 82 | executed after the documentation strings are found. |
| 83 | |
| 84 | If you want to preload function or variable definitions, there are |
| 85 | three ways you can do this and make their documentation strings |
| 86 | accessible when you subsequently run Emacs: |
| 87 | |
| 88 | @itemize @bullet |
| 89 | @item |
| 90 | Arrange to scan these files when producing the @file{etc/DOC} file, |
| 91 | and load them with @file{site-load.el}. |
| 92 | |
| 93 | @item |
| 94 | Load the files with @file{site-init.el}, then copy the files into the |
| 95 | installation directory for Lisp files when you install Emacs. |
| 96 | |
| 97 | @item |
| 98 | Specify a non-@code{nil} value for |
| 99 | @code{byte-compile-dynamic-docstrings} as a local variable in each these |
| 100 | files, and load them with either @file{site-load.el} or |
| 101 | @file{site-init.el}. (This method has the drawback that the |
| 102 | documentation strings take up space in Emacs all the time.) |
| 103 | @end itemize |
| 104 | |
| 105 | It is not advisable to put anything in @file{site-load.el} or |
| 106 | @file{site-init.el} that would alter any of the features that users |
| 107 | expect in an ordinary unmodified Emacs. If you feel you must override |
| 108 | normal features for your site, do it with @file{default.el}, so that |
| 109 | users can override your changes if they wish. @xref{Start-up Summary}. |
| 110 | |
| 111 | @defun dump-emacs to-file from-file |
| 112 | @cindex unexec |
| 113 | This function dumps the current state of Emacs into an executable file |
| 114 | @var{to-file}. It takes symbols from @var{from-file} (this is normally |
| 115 | the executable file @file{temacs}). |
| 116 | |
| 117 | If you want to use this function in an Emacs that was already dumped, |
| 118 | you must run Emacs with @samp{-batch}. |
| 119 | @end defun |
| 120 | |
| 121 | @deffn Command emacs-version |
| 122 | This function returns a string describing the version of Emacs that is |
| 123 | running. It is useful to include this string in bug reports. |
| 124 | |
| 125 | @example |
| 126 | @group |
| 127 | (emacs-version) |
| 128 | @result{} "GNU Emacs 19.29.1 (i386-debian-linux) \ |
| 129 | of Tue Jun 6 1995 on balloon" |
| 130 | @end group |
| 131 | @end example |
| 132 | |
| 133 | Called interactively, the function prints the same information in the |
| 134 | echo area. |
| 135 | @end deffn |
| 136 | |
| 137 | @defvar emacs-build-time |
| 138 | The value of this variable is the time at which Emacs was built at the |
| 139 | local site. |
| 140 | |
| 141 | @example |
| 142 | @group |
| 143 | emacs-build-time |
| 144 | @result{} "Tue Jun 6 14:55:57 1995" |
| 145 | @end group |
| 146 | @end example |
| 147 | @end defvar |
| 148 | |
| 149 | @defvar emacs-version |
| 150 | The value of this variable is the version of Emacs being run. It is a |
| 151 | string such as @code{"19.29.1"}. |
| 152 | @end defvar |
| 153 | |
| 154 | The following two variables did not exist before Emacs version 19.23, |
| 155 | which reduces their usefulness at present, but we hope they will be |
| 156 | convenient in the future. |
| 157 | |
| 158 | @defvar emacs-major-version |
| 159 | The major version number of Emacs, as an integer. For Emacs version |
| 160 | 19.29, the value is 19. |
| 161 | @end defvar |
| 162 | |
| 163 | @defvar emacs-minor-version |
| 164 | The minor version number of Emacs, as an integer. For Emacs version |
| 165 | 19.29, the value is 29. |
| 166 | @end defvar |
| 167 | |
| 168 | @node Pure Storage, Garbage Collection, Building Emacs, GNU Emacs Internals |
| 169 | @appendixsec Pure Storage |
| 170 | @cindex pure storage |
| 171 | |
| 172 | Emacs Lisp uses two kinds of storage for user-created Lisp objects: |
| 173 | @dfn{normal storage} and @dfn{pure storage}. Normal storage is where |
| 174 | all the new data created during an Emacs session is kept; see the |
| 175 | following section for information on normal storage. Pure storage is |
| 176 | used for certain data in the preloaded standard Lisp files---data that |
| 177 | should never change during actual use of Emacs. |
| 178 | |
| 179 | Pure storage is allocated only while @file{temacs} is loading the |
| 180 | standard preloaded Lisp libraries. In the file @file{emacs}, it is |
| 181 | marked as read-only (on operating systems that permit this), so that |
| 182 | the memory space can be shared by all the Emacs jobs running on the |
| 183 | machine at once. Pure storage is not expandable; a fixed amount is |
| 184 | allocated when Emacs is compiled, and if that is not sufficient for the |
| 185 | preloaded libraries, @file{temacs} crashes. If that happens, you must |
| 186 | increase the compilation parameter @code{PURESIZE} in the file |
| 187 | @file{src/puresize.h}. This normally won't happen unless you try to |
| 188 | preload additional libraries or add features to the standard ones. |
| 189 | |
| 190 | @defun purecopy object |
| 191 | This function makes a copy of @var{object} in pure storage and returns |
| 192 | it. It copies strings by simply making a new string with the same |
| 193 | characters in pure storage. It recursively copies the contents of |
| 194 | vectors and cons cells. It does not make copies of other objects such |
| 195 | as symbols, but just returns them unchanged. It signals an error if |
| 196 | asked to copy markers. |
| 197 | |
| 198 | This function is a no-op except while Emacs is being built and dumped; |
| 199 | it is usually called only in the file @file{emacs/lisp/loaddefs.el}, but |
| 200 | a few packages call it just in case you decide to preload them. |
| 201 | @end defun |
| 202 | |
| 203 | @defvar pure-bytes-used |
| 204 | The value of this variable is the number of bytes of pure storage |
| 205 | allocated so far. Typically, in a dumped Emacs, this number is very |
| 206 | close to the total amount of pure storage available---if it were not, |
| 207 | we would preallocate less. |
| 208 | @end defvar |
| 209 | |
| 210 | @defvar purify-flag |
| 211 | This variable determines whether @code{defun} should make a copy of the |
| 212 | function definition in pure storage. If it is non-@code{nil}, then the |
| 213 | function definition is copied into pure storage. |
| 214 | |
| 215 | This flag is @code{t} while loading all of the basic functions for |
| 216 | building Emacs initially (allowing those functions to be sharable and |
| 217 | non-collectible). Dumping Emacs as an executable always writes |
| 218 | @code{nil} in this variable, regardless of the value it actually has |
| 219 | before and after dumping. |
| 220 | |
| 221 | You should not change this flag in a running Emacs. |
| 222 | @end defvar |
| 223 | |
| 224 | @node Garbage Collection, Writing Emacs Primitives, Pure Storage, GNU Emacs Internals |
| 225 | @appendixsec Garbage Collection |
| 226 | @cindex garbage collector |
| 227 | |
| 228 | @cindex memory allocation |
| 229 | When a program creates a list or the user defines a new function (such |
| 230 | as by loading a library), that data is placed in normal storage. If |
| 231 | normal storage runs low, then Emacs asks the operating system to |
| 232 | allocate more memory in blocks of 1k bytes. Each block is used for one |
| 233 | type of Lisp object, so symbols, cons cells, markers, etc., are |
| 234 | segregated in distinct blocks in memory. (Vectors, long strings, |
| 235 | buffers and certain other editing types, which are fairly large, are |
| 236 | allocated in individual blocks, one per object, while small strings are |
| 237 | packed into blocks of 8k bytes.) |
| 238 | |
| 239 | It is quite common to use some storage for a while, then release it by |
| 240 | (for example) killing a buffer or deleting the last pointer to an |
| 241 | object. Emacs provides a @dfn{garbage collector} to reclaim this |
| 242 | abandoned storage. (This name is traditional, but ``garbage recycler'' |
| 243 | might be a more intuitive metaphor for this facility.) |
| 244 | |
| 245 | The garbage collector operates by finding and marking all Lisp objects |
| 246 | that are still accessible to Lisp programs. To begin with, it assumes |
| 247 | all the symbols, their values and associated function definitions, and |
| 248 | any data presently on the stack, are accessible. Any objects that can |
| 249 | be reached indirectly through other accessible objects are also |
| 250 | accessible. |
| 251 | |
| 252 | When marking is finished, all objects still unmarked are garbage. No |
| 253 | matter what the Lisp program or the user does, it is impossible to refer |
| 254 | to them, since there is no longer a way to reach them. Their space |
| 255 | might as well be reused, since no one will miss them. The second |
| 256 | (``sweep'') phase of the garbage collector arranges to reuse them. |
| 257 | |
| 258 | @cindex free list |
| 259 | The sweep phase puts unused cons cells onto a @dfn{free list} |
| 260 | for future allocation; likewise for symbols and markers. It compacts |
| 261 | the accessible strings so they occupy fewer 8k blocks; then it frees the |
| 262 | other 8k blocks. Vectors, buffers, windows, and other large objects are |
| 263 | individually allocated and freed using @code{malloc} and @code{free}. |
| 264 | |
| 265 | @cindex CL note---allocate more storage |
| 266 | @quotation |
| 267 | @b{Common Lisp note:} Unlike other Lisps, GNU Emacs Lisp does not |
| 268 | call the garbage collector when the free list is empty. Instead, it |
| 269 | simply requests the operating system to allocate more storage, and |
| 270 | processing continues until @code{gc-cons-threshold} bytes have been |
| 271 | used. |
| 272 | |
| 273 | This means that you can make sure that the garbage collector will not |
| 274 | run during a certain portion of a Lisp program by calling the garbage |
| 275 | collector explicitly just before it (provided that portion of the |
| 276 | program does not use so much space as to force a second garbage |
| 277 | collection). |
| 278 | @end quotation |
| 279 | |
| 280 | @deffn Command garbage-collect |
| 281 | This command runs a garbage collection, and returns information on |
| 282 | the amount of space in use. (Garbage collection can also occur |
| 283 | spontaneously if you use more than @code{gc-cons-threshold} bytes of |
| 284 | Lisp data since the previous garbage collection.) |
| 285 | |
| 286 | @code{garbage-collect} returns a list containing the following |
| 287 | information: |
| 288 | |
| 289 | @example |
| 290 | @group |
| 291 | ((@var{used-conses} . @var{free-conses}) |
| 292 | (@var{used-syms} . @var{free-syms}) |
| 293 | @end group |
| 294 | (@var{used-markers} . @var{free-markers}) |
| 295 | @var{used-string-chars} |
| 296 | @var{used-vector-slots} |
| 297 | (@var{used-floats} . @var{free-floats})) |
| 298 | |
| 299 | @group |
| 300 | (garbage-collect) |
| 301 | @result{} ((3435 . 2332) (1688 . 0) |
| 302 | (57 . 417) 24510 3839 (4 . 1)) |
| 303 | @end group |
| 304 | @end example |
| 305 | |
| 306 | Here is a table explaining each element: |
| 307 | |
| 308 | @table @var |
| 309 | @item used-conses |
| 310 | The number of cons cells in use. |
| 311 | |
| 312 | @item free-conses |
| 313 | The number of cons cells for which space has been obtained from the |
| 314 | operating system, but that are not currently being used. |
| 315 | |
| 316 | @item used-syms |
| 317 | The number of symbols in use. |
| 318 | |
| 319 | @item free-syms |
| 320 | The number of symbols for which space has been obtained from the |
| 321 | operating system, but that are not currently being used. |
| 322 | |
| 323 | @item used-markers |
| 324 | The number of markers in use. |
| 325 | |
| 326 | @item free-markers |
| 327 | The number of markers for which space has been obtained from the |
| 328 | operating system, but that are not currently being used. |
| 329 | |
| 330 | @item used-string-chars |
| 331 | The total size of all strings, in characters. |
| 332 | |
| 333 | @item used-vector-slots |
| 334 | The total number of elements of existing vectors. |
| 335 | |
| 336 | @item used-floats |
| 337 | @c Emacs 19 feature |
| 338 | The number of floats in use. |
| 339 | |
| 340 | @item free-floats |
| 341 | @c Emacs 19 feature |
| 342 | The number of floats for which space has been obtained from the |
| 343 | operating system, but that are not currently being used. |
| 344 | @end table |
| 345 | @end deffn |
| 346 | |
| 347 | @defopt garbage-collection-messages |
| 348 | If this variable is non-@code{nil}, Emacs displays a message at the |
| 349 | beginning and end of garbage collection. The default value is |
| 350 | @code{nil}, meaning there are no such messages. |
| 351 | @end defopt |
| 352 | |
| 353 | @defopt gc-cons-threshold |
| 354 | The value of this variable is the number of bytes of storage that must |
| 355 | be allocated for Lisp objects after one garbage collection in order to |
| 356 | trigger another garbage collection. A cons cell counts as eight bytes, |
| 357 | a string as one byte per character plus a few bytes of overhead, and so |
| 358 | on; space allocated to the contents of buffers does not count. Note |
| 359 | that the subsequent garbage collection does not happen immediately when |
| 360 | the threshold is exhausted, but only the next time the Lisp evaluator is |
| 361 | called. |
| 362 | |
| 363 | The initial threshold value is 300,000. If you specify a larger |
| 364 | value, garbage collection will happen less often. This reduces the |
| 365 | amount of time spent garbage collecting, but increases total memory use. |
| 366 | You may want to do this when running a program that creates lots of |
| 367 | Lisp data. |
| 368 | |
| 369 | You can make collections more frequent by specifying a smaller value, |
| 370 | down to 10,000. A value less than 10,000 will remain in effect only |
| 371 | until the subsequent garbage collection, at which time |
| 372 | @code{garbage-collect} will set the threshold back to 10,000. |
| 373 | @end defopt |
| 374 | |
| 375 | @c Emacs 19 feature |
| 376 | @defun memory-limit |
| 377 | This function returns the address of the last byte Emacs has allocated, |
| 378 | divided by 1024. We divide the value by 1024 to make sure it fits in a |
| 379 | Lisp integer. |
| 380 | |
| 381 | You can use this to get a general idea of how your actions affect the |
| 382 | memory usage. |
| 383 | @end defun |
| 384 | |
| 385 | @node Writing Emacs Primitives, Object Internals, Garbage Collection, GNU Emacs Internals |
| 386 | @appendixsec Writing Emacs Primitives |
| 387 | @cindex primitive function internals |
| 388 | |
| 389 | Lisp primitives are Lisp functions implemented in C. The details of |
| 390 | interfacing the C function so that Lisp can call it are handled by a few |
| 391 | C macros. The only way to really understand how to write new C code is |
| 392 | to read the source, but we can explain some things here. |
| 393 | |
| 394 | An example of a special form is the definition of @code{or}, from |
| 395 | @file{eval.c}. (An ordinary function would have the same general |
| 396 | appearance.) |
| 397 | |
| 398 | @cindex garbage collection protection |
| 399 | @smallexample |
| 400 | @group |
| 401 | DEFUN ("or", For, Sor, 0, UNEVALLED, 0, |
| 402 | "Eval args until one of them yields non-nil; return that value.\n\ |
| 403 | The remaining args are not evalled at all.\n\ |
| 404 | @end group |
| 405 | @group |
| 406 | If all args return nil, return nil.") |
| 407 | (args) |
| 408 | Lisp_Object args; |
| 409 | @{ |
| 410 | register Lisp_Object val; |
| 411 | Lisp_Object args_left; |
| 412 | struct gcpro gcpro1; |
| 413 | @end group |
| 414 | |
| 415 | @group |
| 416 | if (NULL (args)) |
| 417 | return Qnil; |
| 418 | |
| 419 | args_left = args; |
| 420 | GCPRO1 (args_left); |
| 421 | @end group |
| 422 | |
| 423 | @group |
| 424 | do |
| 425 | @{ |
| 426 | val = Feval (Fcar (args_left)); |
| 427 | if (!NULL (val)) |
| 428 | break; |
| 429 | args_left = Fcdr (args_left); |
| 430 | @} |
| 431 | while (!NULL (args_left)); |
| 432 | @end group |
| 433 | |
| 434 | @group |
| 435 | UNGCPRO; |
| 436 | return val; |
| 437 | @} |
| 438 | @end group |
| 439 | @end smallexample |
| 440 | |
| 441 | Let's start with a precise explanation of the arguments to the |
| 442 | @code{DEFUN} macro. Here is a template for them: |
| 443 | |
| 444 | @example |
| 445 | DEFUN (@var{lname}, @var{fname}, @var{sname}, @var{min}, @var{max}, @var{interactive}, @var{doc}) |
| 446 | @end example |
| 447 | |
| 448 | @table @var |
| 449 | @item lname |
| 450 | This is the name of the Lisp symbol to define as the function name; in |
| 451 | the example above, it is @code{or}. |
| 452 | |
| 453 | @item fname |
| 454 | This is the C function name for this function. This is |
| 455 | the name that is used in C code for calling the function. The name is, |
| 456 | by convention, @samp{F} prepended to the Lisp name, with all dashes |
| 457 | (@samp{-}) in the Lisp name changed to underscores. Thus, to call this |
| 458 | function from C code, call @code{For}. Remember that the arguments must |
| 459 | be of type @code{Lisp_Object}; various macros and functions for creating |
| 460 | values of type @code{Lisp_Object} are declared in the file |
| 461 | @file{lisp.h}. |
| 462 | |
| 463 | @item sname |
| 464 | This is a C variable name to use for a structure that holds the data for |
| 465 | the subr object that represents the function in Lisp. This structure |
| 466 | conveys the Lisp symbol name to the initialization routine that will |
| 467 | create the symbol and store the subr object as its definition. By |
| 468 | convention, this name is always @var{fname} with @samp{F} replaced with |
| 469 | @samp{S}. |
| 470 | |
| 471 | @item min |
| 472 | This is the minimum number of arguments that the function requires. The |
| 473 | function @code{or} allows a minimum of zero arguments. |
| 474 | |
| 475 | @item max |
| 476 | This is the maximum number of arguments that the function accepts, if |
| 477 | there is a fixed maximum. Alternatively, it can be @code{UNEVALLED}, |
| 478 | indicating a special form that receives unevaluated arguments, or |
| 479 | @code{MANY}, indicating an unlimited number of evaluated arguments (the |
| 480 | equivalent of @code{&rest}). Both @code{UNEVALLED} and @code{MANY} are |
| 481 | macros. If @var{max} is a number, it may not be less than @var{min} and |
| 482 | it may not be greater than seven. |
| 483 | |
| 484 | @item interactive |
| 485 | This is an interactive specification, a string such as might be used as |
| 486 | the argument of @code{interactive} in a Lisp function. In the case of |
| 487 | @code{or}, it is 0 (a null pointer), indicating that @code{or} cannot be |
| 488 | called interactively. A value of @code{""} indicates a function that |
| 489 | should receive no arguments when called interactively. |
| 490 | |
| 491 | @item doc |
| 492 | This is the documentation string. It is written just like a |
| 493 | documentation string for a function defined in Lisp, except you must |
| 494 | write @samp{\n\} at the end of each line. In particular, the first line |
| 495 | should be a single sentence. |
| 496 | @end table |
| 497 | |
| 498 | After the call to the @code{DEFUN} macro, you must write the argument |
| 499 | name list that every C function must have, followed by ordinary C |
| 500 | declarations for the arguments. For a function with a fixed maximum |
| 501 | number of arguments, declare a C argument for each Lisp argument, and |
| 502 | give them all type @code{Lisp_Object}. When a Lisp function has no |
| 503 | upper limit on the number of arguments, its implementation in C actually |
| 504 | receives exactly two arguments: the first is the number of Lisp |
| 505 | arguments, and the second is the address of a block containing their |
| 506 | values. They have types @code{int} and @w{@code{Lisp_Object *}}. |
| 507 | |
| 508 | Within the function @code{For} itself, note the use of the macros |
| 509 | @code{GCPRO1} and @code{UNGCPRO}. @code{GCPRO1} is used to ``protect'' |
| 510 | a variable from garbage collection---to inform the garbage collector that |
| 511 | it must look in that variable and regard its contents as an accessible |
| 512 | object. This is necessary whenever you call @code{Feval} or anything |
| 513 | that can directly or indirectly call @code{Feval}. At such a time, any |
| 514 | Lisp object that you intend to refer to again must be protected somehow. |
| 515 | @code{UNGCPRO} cancels the protection of the variables that are |
| 516 | protected in the current function. It is necessary to do this explicitly. |
| 517 | |
| 518 | For most data types, it suffices to protect at least one pointer to |
| 519 | the object; as long as the object is not recycled, all pointers to it |
| 520 | remain valid. This is not so for strings, because the garbage collector |
| 521 | can move them. When the garbage collector moves a string, it relocates |
| 522 | all the pointers it knows about; any other pointers become invalid. |
| 523 | Therefore, you must protect all pointers to strings across any point |
| 524 | where garbage collection may be possible. |
| 525 | |
| 526 | The macro @code{GCPRO1} protects just one local variable. If you want |
| 527 | to protect two, use @code{GCPRO2} instead; repeating @code{GCPRO1} will |
| 528 | not work. Macros @code{GCPRO3} and @code{GCPRO4} also exist. |
| 529 | |
| 530 | These macros implicitly use local variables such as @code{gcpro1}; you |
| 531 | must declare these explicitly, with type @code{struct gcpro}. Thus, if |
| 532 | you use @code{GCPRO2}, you must declare @code{gcpro1} and @code{gcpro2}. |
| 533 | Alas, we can't explain all the tricky details here. |
| 534 | |
| 535 | You must not use C initializers for static or global variables unless |
| 536 | they are never written once Emacs is dumped. These variables with |
| 537 | initializers are allocated in an area of memory that becomes read-only |
| 538 | (on certain operating systems) as a result of dumping Emacs. @xref{Pure |
| 539 | Storage}. |
| 540 | |
| 541 | Do not use static variables within functions---place all static |
| 542 | variables at top level in the file. This is necessary because Emacs on |
| 543 | some operating systems defines the keyword @code{static} as a null |
| 544 | macro. (This definition is used because those systems put all variables |
| 545 | declared static in a place that becomes read-only after dumping, whether |
| 546 | they have initializers or not.) |
| 547 | |
| 548 | Defining the C function is not enough to make a Lisp primitive |
| 549 | available; you must also create the Lisp symbol for the primitive and |
| 550 | store a suitable subr object in its function cell. The code looks like |
| 551 | this: |
| 552 | |
| 553 | @example |
| 554 | defsubr (&@var{subr-structure-name}); |
| 555 | @end example |
| 556 | |
| 557 | @noindent |
| 558 | Here @var{subr-structure-name} is the name you used as the third |
| 559 | argument to @code{DEFUN}. |
| 560 | |
| 561 | If you add a new primitive to a file that already has Lisp primitives |
| 562 | defined in it, find the function (near the end of the file) named |
| 563 | @code{syms_of_@var{something}}, and add the call to @code{defsubr} |
| 564 | there. If the file doesn't have this function, or if you create a new |
| 565 | file, add to it a @code{syms_of_@var{filename}} (e.g., |
| 566 | @code{syms_of_myfile}). Then find the spot in @file{emacs.c} where all |
| 567 | of these functions are called, and add a call to |
| 568 | @code{syms_of_@var{filename}} there. |
| 569 | |
| 570 | The function @code{syms_of_@var{filename}} is also the place to define |
| 571 | any C variables that are to be visible as Lisp variables. |
| 572 | @code{DEFVAR_LISP} makes a C variable of type @code{Lisp_Object} visible |
| 573 | in Lisp. @code{DEFVAR_INT} makes a C variable of type @code{int} |
| 574 | visible in Lisp with a value that is always an integer. |
| 575 | @code{DEFVAR_BOOL} makes a C variable of type @code{int} visible in Lisp |
| 576 | with a value that is either @code{t} or @code{nil}. |
| 577 | |
| 578 | Here is another example function, with more complicated arguments. |
| 579 | This comes from the code for the X Window System, and it demonstrates |
| 580 | the use of macros and functions to manipulate Lisp objects. |
| 581 | |
| 582 | @smallexample |
| 583 | @group |
| 584 | DEFUN ("coordinates-in-window-p", Fcoordinates_in_window_p, |
| 585 | Scoordinates_in_window_p, 2, 2, |
| 586 | "xSpecify coordinate pair: \nXExpression which evals to window: ", |
| 587 | "Return non-nil if POSITIONS is in WINDOW.\n\ |
| 588 | \(POSITIONS is a list, (SCREEN-X SCREEN-Y)\)\n\ |
| 589 | @end group |
| 590 | @group |
| 591 | Returned value is list of positions expressed\n\ |
| 592 | relative to window upper left corner.") |
| 593 | (coordinate, window) |
| 594 | register Lisp_Object coordinate, window; |
| 595 | @{ |
| 596 | register Lisp_Object xcoord, ycoord; |
| 597 | @end group |
| 598 | |
| 599 | @group |
| 600 | if (!CONSP (coordinate)) wrong_type_argument (Qlistp, coordinate); |
| 601 | CHECK_WINDOW (window, 2); |
| 602 | xcoord = Fcar (coordinate); |
| 603 | ycoord = Fcar (Fcdr (coordinate)); |
| 604 | CHECK_NUMBER (xcoord, 0); |
| 605 | CHECK_NUMBER (ycoord, 1); |
| 606 | @end group |
| 607 | @group |
| 608 | if ((XINT (xcoord) < XINT (XWINDOW (window)->left)) |
| 609 | || (XINT (xcoord) >= (XINT (XWINDOW (window)->left) |
| 610 | + XINT (XWINDOW (window)->width)))) |
| 611 | return Qnil; |
| 612 | XFASTINT (xcoord) -= XFASTINT (XWINDOW (window)->left); |
| 613 | @end group |
| 614 | @group |
| 615 | if (XINT (ycoord) == (screen_height - 1)) |
| 616 | return Qnil; |
| 617 | @end group |
| 618 | @group |
| 619 | if ((XINT (ycoord) < XINT (XWINDOW (window)->top)) |
| 620 | || (XINT (ycoord) >= (XINT (XWINDOW (window)->top) |
| 621 | + XINT (XWINDOW (window)->height)) - 1)) |
| 622 | return Qnil; |
| 623 | @end group |
| 624 | @group |
| 625 | XFASTINT (ycoord) -= XFASTINT (XWINDOW (window)->top); |
| 626 | return (Fcons (xcoord, Fcons (ycoord, Qnil))); |
| 627 | @} |
| 628 | @end group |
| 629 | @end smallexample |
| 630 | |
| 631 | Note that C code cannot call functions by name unless they are defined |
| 632 | in C. The way to call a function written in Lisp is to use |
| 633 | @code{Ffuncall}, which embodies the Lisp function @code{funcall}. Since |
| 634 | the Lisp function @code{funcall} accepts an unlimited number of |
| 635 | arguments, in C it takes two: the number of Lisp-level arguments, and a |
| 636 | one-dimensional array containing their values. The first Lisp-level |
| 637 | argument is the Lisp function to call, and the rest are the arguments to |
| 638 | pass to it. Since @code{Ffuncall} can call the evaluator, you must |
| 639 | protect pointers from garbage collection around the call to |
| 640 | @code{Ffuncall}. |
| 641 | |
| 642 | The C functions @code{call0}, @code{call1}, @code{call2}, and so on, |
| 643 | provide handy ways to call a Lisp function conveniently with a fixed |
| 644 | number of arguments. They work by calling @code{Ffuncall}. |
| 645 | |
| 646 | @file{eval.c} is a very good file to look through for examples; |
| 647 | @file{lisp.h} contains the definitions for some important macros and |
| 648 | functions. |
| 649 | |
| 650 | @node Object Internals, , Writing Emacs Primitives, GNU Emacs Internals |
| 651 | @appendixsec Object Internals |
| 652 | @cindex object internals |
| 653 | |
| 654 | GNU Emacs Lisp manipulates many different types of data. The actual |
| 655 | data are stored in a heap and the only access that programs have to it is |
| 656 | through pointers. Pointers are thirty-two bits wide in most |
| 657 | implementations. Depending on the operating system and type of machine |
| 658 | for which you compile Emacs, twenty-four to twenty-six bits are used to |
| 659 | address the object, and the remaining six to eight bits are used for a |
| 660 | tag that identifies the object's type. |
| 661 | |
| 662 | Because Lisp objects are represented as tagged pointers, it is always |
| 663 | possible to determine the Lisp data type of any object. The C data type |
| 664 | @code{Lisp_Object} can hold any Lisp object of any data type. Ordinary |
| 665 | variables have type @code{Lisp_Object}, which means they can hold any |
| 666 | type of Lisp value; you can determine the actual data type only at run |
| 667 | time. The same is true for function arguments; if you want a function |
| 668 | to accept only a certain type of argument, you must check the type |
| 669 | explicitly using a suitable predicate (@pxref{Type Predicates}). |
| 670 | @cindex type checking internals |
| 671 | |
| 672 | @menu |
| 673 | * Buffer Internals:: Components of a buffer structure. |
| 674 | * Window Internals:: Components of a window structure. |
| 675 | * Process Internals:: Components of a process structure. |
| 676 | @end menu |
| 677 | |
| 678 | @node Buffer Internals, Window Internals, Object Internals, Object Internals |
| 679 | @appendixsubsec Buffer Internals |
| 680 | @cindex internals, of buffer |
| 681 | @cindex buffer internals |
| 682 | |
| 683 | Buffers contain fields not directly accessible by the Lisp programmer. |
| 684 | We describe them here, naming them by the names used in the C code. |
| 685 | Many are accessible indirectly in Lisp programs via Lisp primitives. |
| 686 | |
| 687 | @table @code |
| 688 | @item name |
| 689 | The buffer name is a string that names the buffer. It is guaranteed to |
| 690 | be unique. @xref{Buffer Names}. |
| 691 | |
| 692 | @item save_modified |
| 693 | This field contains the time when the buffer was last saved, as an integer. |
| 694 | @xref{Buffer Modification}. |
| 695 | |
| 696 | @item modtime |
| 697 | This field contains the modification time of the visited file. It is |
| 698 | set when the file is written or read. Every time the buffer is written |
| 699 | to the file, this field is compared to the modification time of the |
| 700 | file. @xref{Buffer Modification}. |
| 701 | |
| 702 | @item auto_save_modified |
| 703 | This field contains the time when the buffer was last auto-saved. |
| 704 | |
| 705 | @item last_window_start |
| 706 | This field contains the @code{window-start} position in the buffer as of |
| 707 | the last time the buffer was displayed in a window. |
| 708 | |
| 709 | @item undo_list |
| 710 | This field points to the buffer's undo list. @xref{Undo}. |
| 711 | |
| 712 | @item syntax_table_v |
| 713 | This field contains the syntax table for the buffer. @xref{Syntax Tables}. |
| 714 | |
| 715 | @item downcase_table |
| 716 | This field contains the conversion table for converting text to lower case. |
| 717 | @xref{Case Table}. |
| 718 | |
| 719 | @item upcase_table |
| 720 | This field contains the conversion table for converting text to upper case. |
| 721 | @xref{Case Table}. |
| 722 | |
| 723 | @item case_canon_table |
| 724 | This field contains the conversion table for canonicalizing text for |
| 725 | case-folding search. @xref{Case Table}. |
| 726 | |
| 727 | @item case_eqv_table |
| 728 | This field contains the equivalence table for case-folding search. |
| 729 | @xref{Case Table}. |
| 730 | |
| 731 | @item display_table |
| 732 | This field contains the buffer's display table, or @code{nil} if it doesn't |
| 733 | have one. @xref{Display Tables}. |
| 734 | |
| 735 | @item markers |
| 736 | This field contains the chain of all markers that currently point into |
| 737 | the buffer. Deletion of text in the buffer, and motion of the buffer's |
| 738 | gap, must check each of these markers and perhaps update it. |
| 739 | @xref{Markers}. |
| 740 | |
| 741 | @item backed_up |
| 742 | This field is a flag that tells whether a backup file has been made |
| 743 | for the visited file of this buffer. |
| 744 | |
| 745 | @item mark |
| 746 | This field contains the mark for the buffer. The mark is a marker, |
| 747 | hence it is also included on the list @code{markers}. @xref{The Mark}. |
| 748 | |
| 749 | @item mark_active |
| 750 | This field is non-@code{nil} if the buffer's mark is active. |
| 751 | |
| 752 | @item local_var_alist |
| 753 | This field contains the association list describing the variables local |
| 754 | in this buffer, and their values, with the exception of local variables |
| 755 | that have special slots in the buffer object. (Those slots are omitted |
| 756 | from this table.) @xref{Buffer-Local Variables}. |
| 757 | |
| 758 | @item base_buffer |
| 759 | This field holds the buffer's base buffer (if it is an indirect buffer), |
| 760 | or @code{nil}. |
| 761 | |
| 762 | @item keymap |
| 763 | This field holds the buffer's local keymap. @xref{Keymaps}. |
| 764 | |
| 765 | @item overlay_center |
| 766 | This field holds the current overlay center position. @xref{Overlays}. |
| 767 | |
| 768 | @item overlays_before |
| 769 | This field holds a list of the overlays in this buffer that end at or |
| 770 | before the current overlay center position. They are sorted in order of |
| 771 | decreasing end position. |
| 772 | |
| 773 | @item overlays_after |
| 774 | This field holds a list of the overlays in this buffer that end after |
| 775 | the current overlay center position. They are sorted in order of |
| 776 | increasing beginning position. |
| 777 | @end table |
| 778 | |
| 779 | @node Window Internals, Process Internals, Buffer Internals, Object Internals |
| 780 | @appendixsubsec Window Internals |
| 781 | @cindex internals, of window |
| 782 | @cindex window internals |
| 783 | |
| 784 | Windows have the following accessible fields: |
| 785 | |
| 786 | @table @code |
| 787 | @item frame |
| 788 | The frame that this window is on. |
| 789 | |
| 790 | @item mini_p |
| 791 | Non-@code{nil} if this window is a minibuffer window. |
| 792 | |
| 793 | @item buffer |
| 794 | The buffer that the window is displaying. This may change often during |
| 795 | the life of the window. |
| 796 | |
| 797 | @item dedicated |
| 798 | Non-@code{nil} if this window is dedicated to its buffer. |
| 799 | |
| 800 | @item pointm |
| 801 | @cindex window point internals |
| 802 | This is the value of point in the current buffer when this window is |
| 803 | selected; when it is not selected, it retains its previous value. |
| 804 | |
| 805 | @item start |
| 806 | The position in the buffer that is the first character to be displayed |
| 807 | in the window. |
| 808 | |
| 809 | @item force_start |
| 810 | If this flag is non-@code{nil}, it says that the window has been |
| 811 | scrolled explicitly by the Lisp program. This affects what the next |
| 812 | redisplay does if point is off the screen: instead of scrolling the |
| 813 | window to show the text around point, it moves point to a location that |
| 814 | is on the screen. |
| 815 | |
| 816 | @item last_modified |
| 817 | The @code{modified} field of the window's buffer, as of the last time |
| 818 | a redisplay completed in this window. |
| 819 | |
| 820 | @item last_point |
| 821 | The buffer's value of point, as of the last time |
| 822 | a redisplay completed in this window. |
| 823 | |
| 824 | @item left |
| 825 | This is the left-hand edge of the window, measured in columns. (The |
| 826 | leftmost column on the screen is @w{column 0}.) |
| 827 | |
| 828 | @item top |
| 829 | This is the top edge of the window, measured in lines. (The top line on |
| 830 | the screen is @w{line 0}.) |
| 831 | |
| 832 | @item height |
| 833 | The height of the window, measured in lines. |
| 834 | |
| 835 | @item width |
| 836 | The width of the window, measured in columns. |
| 837 | |
| 838 | @item next |
| 839 | This is the window that is the next in the chain of siblings. It is |
| 840 | @code{nil} in a window that is the rightmost or bottommost of a group of |
| 841 | siblings. |
| 842 | |
| 843 | @item prev |
| 844 | This is the window that is the previous in the chain of siblings. It is |
| 845 | @code{nil} in a window that is the leftmost or topmost of a group of |
| 846 | siblings. |
| 847 | |
| 848 | @item parent |
| 849 | Internally, Emacs arranges windows in a tree; each group of siblings has |
| 850 | a parent window whose area includes all the siblings. This field points |
| 851 | to a window's parent. |
| 852 | |
| 853 | Parent windows do not display buffers, and play little role in display |
| 854 | except to shape their child windows. Emacs Lisp programs usually have |
| 855 | no access to the parent windows; they operate on the windows at the |
| 856 | leaves of the tree, which actually display buffers. |
| 857 | |
| 858 | @item hscroll |
| 859 | This is the number of columns that the display in the window is scrolled |
| 860 | horizontally to the left. Normally, this is 0. |
| 861 | |
| 862 | @item use_time |
| 863 | This is the last time that the window was selected. The function |
| 864 | @code{get-lru-window} uses this field. |
| 865 | |
| 866 | @item display_table |
| 867 | The window's display table, or @code{nil} if none is specified for it. |
| 868 | |
| 869 | @item update_mode_line |
| 870 | Non-@code{nil} means this window's mode line needs to be updated. |
| 871 | |
| 872 | @item base_line_number |
| 873 | The line number of a certain position in the buffer, or @code{nil}. |
| 874 | This is used for displaying the line number of point in the mode line. |
| 875 | |
| 876 | @item base_line_pos |
| 877 | The position in the buffer for which the line number is known, or |
| 878 | @code{nil} meaning none is known. |
| 879 | |
| 880 | @item region_showing |
| 881 | If the region (or part of it) is highlighted in this window, this field |
| 882 | holds the mark position that made one end of that region. Otherwise, |
| 883 | this field is @code{nil}. |
| 884 | @end table |
| 885 | |
| 886 | @node Process Internals, , Window Internals, Object Internals |
| 887 | @appendixsubsec Process Internals |
| 888 | @cindex internals, of process |
| 889 | @cindex process internals |
| 890 | |
| 891 | The fields of a process are: |
| 892 | |
| 893 | @table @code |
| 894 | @item name |
| 895 | A string, the name of the process. |
| 896 | |
| 897 | @item command |
| 898 | A list containing the command arguments that were used to start this |
| 899 | process. |
| 900 | |
| 901 | @item filter |
| 902 | A function used to accept output from the process instead of a buffer, |
| 903 | or @code{nil}. |
| 904 | |
| 905 | @item sentinel |
| 906 | A function called whenever the process receives a signal, or @code{nil}. |
| 907 | |
| 908 | @item buffer |
| 909 | The associated buffer of the process. |
| 910 | |
| 911 | @item pid |
| 912 | An integer, the Unix process @sc{id}. |
| 913 | |
| 914 | @item childp |
| 915 | A flag, non-@code{nil} if this is really a child process. |
| 916 | It is @code{nil} for a network connection. |
| 917 | |
| 918 | @item mark |
| 919 | A marker indicating the position of the end of the last output from this |
| 920 | process inserted into the buffer. This is often but not always the end |
| 921 | of the buffer. |
| 922 | |
| 923 | @item kill_without_query |
| 924 | If this is non-@code{nil}, killing Emacs while this process is still |
| 925 | running does not ask for confirmation about killing the process. |
| 926 | |
| 927 | @item raw_status_low |
| 928 | @itemx raw_status_high |
| 929 | These two fields record 16 bits each of the process status returned by |
| 930 | the @code{wait} system call. |
| 931 | |
| 932 | @item status |
| 933 | The process status, as @code{process-status} should return it. |
| 934 | |
| 935 | @item tick |
| 936 | @itemx update_tick |
| 937 | If these two fields are not equal, a change in the status of the process |
| 938 | needs to be reported, either by running the sentinel or by inserting a |
| 939 | message in the process buffer. |
| 940 | |
| 941 | @item pty_flag |
| 942 | Non-@code{nil} if communication with the subprocess uses a @sc{pty}; |
| 943 | @code{nil} if it uses a pipe. |
| 944 | |
| 945 | @item infd |
| 946 | The file descriptor for input from the process. |
| 947 | |
| 948 | @item outfd |
| 949 | The file descriptor for output to the process. |
| 950 | |
| 951 | @item subtty |
| 952 | The file descriptor for the terminal that the subprocess is using. (On |
| 953 | some systems, there is no need to record this, so the value is |
| 954 | @code{nil}.) |
| 955 | |
| 956 | @item tty_name |
| 957 | The name of the terminal that the subprocess is using, |
| 958 | or @code{nil} if it is using pipes. |
| 959 | @end table |