Commit | Line | Data |
---|---|---|
f2cc4248 | 1 | /* Generate doc-string file for GNU Emacs from source files. |
291c7e74 GM |
2 | Copyright (C) 1985, 86, 92, 93, 94, 97, 1999, 2000, 2001 |
3 | Free Software Foundation, Inc. | |
f2cc4248 RS |
4 | |
5 | This file is part of GNU Emacs. | |
6 | ||
93320c23 JA |
7 | GNU Emacs is free software; you can redistribute it and/or modify |
8 | it under the terms of the GNU General Public License as published by | |
e065a56e | 9 | the Free Software Foundation; either version 2, or (at your option) |
93320c23 JA |
10 | any later version. |
11 | ||
f2cc4248 | 12 | GNU Emacs is distributed in the hope that it will be useful, |
93320c23 JA |
13 | but WITHOUT ANY WARRANTY; without even the implied warranty of |
14 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
15 | GNU General Public License for more details. | |
f2cc4248 | 16 | |
93320c23 JA |
17 | You should have received a copy of the GNU General Public License |
18 | along with GNU Emacs; see the file COPYING. If not, write to | |
3b7ad313 EN |
19 | the Free Software Foundation, Inc., 59 Temple Place - Suite 330, |
20 | Boston, MA 02111-1307, USA. */ | |
f2cc4248 RS |
21 | |
22 | /* The arguments given to this program are all the C and Lisp source files | |
23 | of GNU Emacs. .elc and .el and .c files are allowed. | |
24 | A .o file can also be specified; the .c file it was made from is used. | |
25 | This helps the makefile pass the correct list of files. | |
26 | ||
27 | The results, which go to standard output or to a file | |
28 | specified with -a or -o (-a to append, -o to start from nothing), | |
29 | are entries containing function or variable names and their documentation. | |
30 | Each entry starts with a ^_ character. | |
31 | Then comes F for a function or V for a variable. | |
32 | Then comes the function or variable name, terminated with a newline. | |
33 | Then comes the documentation for that function or variable. | |
34 | */ | |
35 | ||
34b4ece5 | 36 | #define NO_SHORTNAMES /* Tell config not to load remap.h */ |
433d333d RS |
37 | #include <config.h> |
38 | ||
39 | /* defined to be emacs_main, sys_fopen, etc. in config.h */ | |
40 | #undef main | |
41 | #undef fopen | |
42 | #undef chdir | |
34b4ece5 | 43 | |
f2cc4248 | 44 | #include <stdio.h> |
e3938952 RS |
45 | #ifdef MSDOS |
46 | #include <fcntl.h> | |
47 | #endif /* MSDOS */ | |
86b0513a RS |
48 | #ifdef WINDOWSNT |
49 | #include <stdlib.h> | |
50 | #include <fcntl.h> | |
51 | #include <direct.h> | |
52 | #endif /* WINDOWSNT */ | |
e3938952 | 53 | |
86b0513a | 54 | #ifdef DOS_NT |
e3938952 RS |
55 | #define READ_TEXT "rt" |
56 | #define READ_BINARY "rb" | |
86b0513a | 57 | #else /* not DOS_NT */ |
e3938952 RS |
58 | #define READ_TEXT "r" |
59 | #define READ_BINARY "r" | |
86b0513a | 60 | #endif /* not DOS_NT */ |
f2cc4248 | 61 | |
340ff9de DM |
62 | int scan_file (); |
63 | int scan_lisp_file (); | |
64 | int scan_c_file (); | |
65 | ||
2d1985a2 KH |
66 | #ifdef MSDOS |
67 | /* s/msdos.h defines this as sys_chdir, but we're not linking with the | |
68 | file where that function is defined. */ | |
69 | #undef chdir | |
70 | #endif | |
71 | ||
a0613c61 AS |
72 | #ifdef HAVE_UNISTD_H |
73 | #include <unistd.h> | |
74 | #endif | |
75 | ||
b5ff43cc | 76 | /* Stdio stream for output to the DOC file. */ |
f2cc4248 RS |
77 | FILE *outfile; |
78 | ||
b5ff43cc RS |
79 | /* Name this program was invoked with. */ |
80 | char *progname; | |
81 | ||
82 | /* Print error message. `s1' is printf control string, `s2' is arg for it. */ | |
83 | ||
84 | /* VARARGS1 */ | |
85 | void | |
86 | error (s1, s2) | |
87 | char *s1, *s2; | |
88 | { | |
89 | fprintf (stderr, "%s: ", progname); | |
90 | fprintf (stderr, s1, s2); | |
91 | fprintf (stderr, "\n"); | |
92 | } | |
93 | ||
94 | /* Print error message and exit. */ | |
95 | ||
96 | /* VARARGS1 */ | |
97 | void | |
98 | fatal (s1, s2) | |
99 | char *s1, *s2; | |
100 | { | |
101 | error (s1, s2); | |
102 | exit (1); | |
103 | } | |
104 | ||
105 | /* Like malloc but get fatal error if memory is exhausted. */ | |
106 | ||
34b4ece5 | 107 | long * |
b5ff43cc RS |
108 | xmalloc (size) |
109 | unsigned int size; | |
110 | { | |
34b4ece5 | 111 | long *result = (long *) malloc (size); |
b5ff43cc RS |
112 | if (result == NULL) |
113 | fatal ("virtual memory exhausted", 0); | |
114 | return result; | |
115 | } | |
116 | \f | |
340ff9de | 117 | int |
f2cc4248 RS |
118 | main (argc, argv) |
119 | int argc; | |
120 | char **argv; | |
121 | { | |
122 | int i; | |
123 | int err_count = 0; | |
a27897c9 | 124 | int first_infile; |
f2cc4248 | 125 | |
b5ff43cc RS |
126 | progname = argv[0]; |
127 | ||
4e043ed3 RS |
128 | outfile = stdout; |
129 | ||
86b0513a | 130 | /* Don't put CRs in the DOC file. */ |
e3938952 | 131 | #ifdef MSDOS |
5281dea4 | 132 | _fmode = O_BINARY; |
4e043ed3 RS |
133 | #if 0 /* Suspicion is that this causes hanging. |
134 | So instead we require people to use -o on MSDOS. */ | |
e3938952 RS |
135 | (stdout)->_flag &= ~_IOTEXT; |
136 | _setmode (fileno (stdout), O_BINARY); | |
4e043ed3 RS |
137 | #endif |
138 | outfile = 0; | |
e3938952 | 139 | #endif /* MSDOS */ |
86b0513a RS |
140 | #ifdef WINDOWSNT |
141 | _fmode = O_BINARY; | |
142 | _setmode (fileno (stdout), O_BINARY); | |
143 | #endif /* WINDOWSNT */ | |
144 | ||
f2cc4248 RS |
145 | /* If first two args are -o FILE, output to FILE. */ |
146 | i = 1; | |
147 | if (argc > i + 1 && !strcmp (argv[i], "-o")) | |
148 | { | |
149 | outfile = fopen (argv[i + 1], "w"); | |
150 | i += 2; | |
151 | } | |
152 | if (argc > i + 1 && !strcmp (argv[i], "-a")) | |
153 | { | |
154 | outfile = fopen (argv[i + 1], "a"); | |
155 | i += 2; | |
156 | } | |
d2d92f7a JB |
157 | if (argc > i + 1 && !strcmp (argv[i], "-d")) |
158 | { | |
159 | chdir (argv[i + 1]); | |
160 | i += 2; | |
161 | } | |
f2cc4248 | 162 | |
4e043ed3 RS |
163 | if (outfile == 0) |
164 | fatal ("No output file specified", ""); | |
165 | ||
a27897c9 | 166 | first_infile = i; |
f2cc4248 | 167 | for (; i < argc; i++) |
a27897c9 RS |
168 | { |
169 | int j; | |
170 | /* Don't process one file twice. */ | |
171 | for (j = first_infile; j < i; j++) | |
172 | if (! strcmp (argv[i], argv[j])) | |
173 | break; | |
174 | if (j == i) | |
175 | err_count += scan_file (argv[i]); | |
176 | } | |
f2cc4248 | 177 | #ifndef VMS |
a27897c9 | 178 | exit (err_count > 0); |
a706b808 | 179 | #endif /* VMS */ |
a27897c9 | 180 | return err_count > 0; |
f2cc4248 RS |
181 | } |
182 | ||
a8a7afbe | 183 | /* Read file FILENAME and output its doc strings to outfile. */ |
f2cc4248 RS |
184 | /* Return 1 if file is not found, 0 if it is found. */ |
185 | ||
340ff9de | 186 | int |
f2cc4248 RS |
187 | scan_file (filename) |
188 | char *filename; | |
189 | { | |
190 | int len = strlen (filename); | |
dc61cb9d | 191 | if (len > 4 && !strcmp (filename + len - 4, ".elc")) |
e3938952 | 192 | return scan_lisp_file (filename, READ_BINARY); |
dc61cb9d | 193 | else if (len > 3 && !strcmp (filename + len - 3, ".el")) |
e3938952 | 194 | return scan_lisp_file (filename, READ_TEXT); |
f2cc4248 | 195 | else |
e3938952 | 196 | return scan_c_file (filename, READ_TEXT); |
f2cc4248 RS |
197 | } |
198 | \f | |
199 | char buf[128]; | |
200 | ||
74c55c82 GM |
201 | /* Skip a C string or C-style comment from INFILE, and return the |
202 | character that follows. COMMENT non-zero means skip a comment. If | |
203 | PRINTFLAG is positive, output string contents to outfile. If it is | |
204 | negative, store contents in buf. Convert escape sequences \n and | |
205 | \t to newline and tab; discard \ followed by newline. */ | |
f2cc4248 | 206 | |
340ff9de | 207 | int |
74c55c82 | 208 | read_c_string_or_comment (infile, printflag, comment) |
f2cc4248 RS |
209 | FILE *infile; |
210 | int printflag; | |
211 | { | |
212 | register int c; | |
213 | char *p = buf; | |
214 | ||
74c55c82 GM |
215 | if (comment) |
216 | { | |
217 | while ((c = getc (infile)) != EOF | |
218 | && (c == '\n' || c == '\r' || c == '\t' || c == ' ')) | |
219 | ; | |
220 | } | |
221 | else | |
222 | c = getc (infile); | |
223 | ||
f2cc4248 RS |
224 | while (c != EOF) |
225 | { | |
74c55c82 | 226 | while (c != EOF && (comment ? c != '*' : c != '"')) |
f2cc4248 RS |
227 | { |
228 | if (c == '\\') | |
229 | { | |
230 | c = getc (infile); | |
433d333d | 231 | if (c == '\n' || c == '\r') |
f2cc4248 RS |
232 | { |
233 | c = getc (infile); | |
234 | continue; | |
235 | } | |
236 | if (c == 'n') | |
237 | c = '\n'; | |
238 | if (c == 't') | |
239 | c = '\t'; | |
240 | } | |
74c55c82 | 241 | |
f2cc4248 RS |
242 | if (printflag > 0) |
243 | putc (c, outfile); | |
244 | else if (printflag < 0) | |
245 | *p++ = c; | |
246 | c = getc (infile); | |
247 | } | |
74c55c82 | 248 | |
f2cc4248 | 249 | c = getc (infile); |
f2cc4248 | 250 | |
74c55c82 GM |
251 | if (comment) |
252 | { | |
253 | if (c == '/') | |
254 | { | |
255 | c = getc (infile); | |
256 | break; | |
257 | } | |
258 | } | |
259 | else | |
260 | { | |
261 | if (c != '"') | |
262 | break; | |
263 | ||
264 | /* If we had a "", concatenate the two strings. */ | |
265 | c = getc (infile); | |
266 | } | |
267 | } | |
268 | ||
f2cc4248 RS |
269 | if (printflag < 0) |
270 | *p = 0; | |
271 | ||
272 | return c; | |
273 | } | |
74c55c82 GM |
274 | |
275 | ||
f2cc4248 | 276 | \f |
069ad9ea | 277 | /* Write to file OUT the argument names of function FUNC, whose text is in BUF. |
f2cc4248 RS |
278 | MINARGS and MAXARGS are the minimum and maximum number of arguments. */ |
279 | ||
340ff9de | 280 | void |
069ad9ea | 281 | write_c_args (out, func, buf, minargs, maxargs) |
f2cc4248 | 282 | FILE *out; |
069ad9ea | 283 | char *func, *buf; |
f2cc4248 RS |
284 | int minargs, maxargs; |
285 | { | |
f125a9e8 | 286 | register char *p; |
30e4c427 JB |
287 | int in_ident = 0; |
288 | int just_spaced = 0; | |
069ad9ea | 289 | int need_space = 1; |
f2cc4248 | 290 | |
069ad9ea RM |
291 | fprintf (out, "(%s", func); |
292 | ||
293 | if (*buf == '(') | |
294 | ++buf; | |
f2cc4248 | 295 | |
f125a9e8 | 296 | for (p = buf; *p; p++) |
f2cc4248 | 297 | { |
30e4c427 | 298 | char c = *p; |
3941a179 | 299 | int ident_start = 0; |
30e4c427 JB |
300 | |
301 | /* Notice when we start printing a new identifier. */ | |
302 | if ((('A' <= c && c <= 'Z') | |
303 | || ('a' <= c && c <= 'z') | |
304 | || ('0' <= c && c <= '9') | |
305 | || c == '_') | |
306 | != in_ident) | |
f2cc4248 | 307 | { |
30e4c427 JB |
308 | if (!in_ident) |
309 | { | |
310 | in_ident = 1; | |
3941a179 | 311 | ident_start = 1; |
f125a9e8 | 312 | |
069ad9ea RM |
313 | if (need_space) |
314 | putc (' ', out); | |
315 | ||
30e4c427 JB |
316 | if (minargs == 0 && maxargs > 0) |
317 | fprintf (out, "&optional "); | |
318 | just_spaced = 1; | |
f125a9e8 | 319 | |
30e4c427 JB |
320 | minargs--; |
321 | maxargs--; | |
322 | } | |
323 | else | |
324 | in_ident = 0; | |
f2cc4248 | 325 | } |
30e4c427 JB |
326 | |
327 | /* Print the C argument list as it would appear in lisp: | |
291c7e74 GM |
328 | print underscores as hyphens, and print commas and newlines |
329 | as spaces. Collapse adjacent spaces into one. */ | |
330 | if (c == '_') | |
331 | c = '-'; | |
332 | else if (c == ',' || c == '\n') | |
333 | c = ' '; | |
30e4c427 | 334 | |
3941a179 JB |
335 | /* In C code, `default' is a reserved word, so we spell it |
336 | `defalt'; unmangle that here. */ | |
337 | if (ident_start | |
338 | && strncmp (p, "defalt", 6) == 0 | |
339 | && ! (('A' <= p[6] && p[6] <= 'Z') | |
340 | || ('a' <= p[6] && p[6] <= 'z') | |
341 | || ('0' <= p[6] && p[6] <= '9') | |
342 | || p[6] == '_')) | |
343 | { | |
4607ea99 | 344 | fprintf (out, "DEFAULT"); |
3941a179 JB |
345 | p += 5; |
346 | in_ident = 0; | |
347 | just_spaced = 0; | |
348 | } | |
291c7e74 | 349 | else if (c != ' ' || !just_spaced) |
069ad9ea RM |
350 | { |
351 | if (c >= 'a' && c <= 'z') | |
352 | /* Upcase the letter. */ | |
353 | c += 'A' - 'a'; | |
354 | putc (c, out); | |
355 | } | |
30e4c427 | 356 | |
291c7e74 | 357 | just_spaced = c == ' '; |
069ad9ea | 358 | need_space = 0; |
f2cc4248 | 359 | } |
f2cc4248 RS |
360 | } |
361 | \f | |
362 | /* Read through a c file. If a .o file is named, | |
363 | the corresponding .c file is read instead. | |
364 | Looks for DEFUN constructs such as are defined in ../src/lisp.h. | |
365 | Accepts any word starting DEF... so it finds DEFSIMPLE and DEFPRED. */ | |
366 | ||
340ff9de | 367 | int |
e3938952 RS |
368 | scan_c_file (filename, mode) |
369 | char *filename, *mode; | |
f2cc4248 RS |
370 | { |
371 | FILE *infile; | |
372 | register int c; | |
373 | register int commas; | |
374 | register int defunflag; | |
84128dee | 375 | register int defvarperbufferflag; |
f2cc4248 RS |
376 | register int defvarflag; |
377 | int minargs, maxargs; | |
1feb8ae1 | 378 | int extension = filename[strlen (filename) - 1]; |
f2cc4248 | 379 | |
1feb8ae1 | 380 | if (extension == 'o') |
f2cc4248 RS |
381 | filename[strlen (filename) - 1] = 'c'; |
382 | ||
e3938952 | 383 | infile = fopen (filename, mode); |
f2cc4248 RS |
384 | |
385 | /* No error if non-ex input file */ | |
386 | if (infile == NULL) | |
387 | { | |
388 | perror (filename); | |
389 | return 0; | |
390 | } | |
391 | ||
1feb8ae1 RS |
392 | /* Reset extension to be able to detect duplicate files. */ |
393 | filename[strlen (filename) - 1] = extension; | |
394 | ||
f2cc4248 RS |
395 | c = '\n'; |
396 | while (!feof (infile)) | |
397 | { | |
433d333d | 398 | if (c != '\n' && c != '\r') |
f2cc4248 RS |
399 | { |
400 | c = getc (infile); | |
401 | continue; | |
402 | } | |
403 | c = getc (infile); | |
404 | if (c == ' ') | |
405 | { | |
406 | while (c == ' ') | |
407 | c = getc (infile); | |
408 | if (c != 'D') | |
409 | continue; | |
410 | c = getc (infile); | |
411 | if (c != 'E') | |
412 | continue; | |
413 | c = getc (infile); | |
414 | if (c != 'F') | |
415 | continue; | |
416 | c = getc (infile); | |
417 | if (c != 'V') | |
418 | continue; | |
84128dee JB |
419 | c = getc (infile); |
420 | if (c != 'A') | |
421 | continue; | |
422 | c = getc (infile); | |
423 | if (c != 'R') | |
424 | continue; | |
425 | c = getc (infile); | |
426 | if (c != '_') | |
427 | continue; | |
428 | ||
f2cc4248 RS |
429 | defvarflag = 1; |
430 | defunflag = 0; | |
84128dee JB |
431 | |
432 | c = getc (infile); | |
433 | defvarperbufferflag = (c == 'P'); | |
434 | ||
f2cc4248 RS |
435 | c = getc (infile); |
436 | } | |
437 | else if (c == 'D') | |
438 | { | |
439 | c = getc (infile); | |
440 | if (c != 'E') | |
441 | continue; | |
442 | c = getc (infile); | |
443 | if (c != 'F') | |
444 | continue; | |
445 | c = getc (infile); | |
446 | defunflag = c == 'U'; | |
447 | defvarflag = 0; | |
448 | } | |
449 | else continue; | |
450 | ||
451 | while (c != '(') | |
452 | { | |
453 | if (c < 0) | |
454 | goto eof; | |
455 | c = getc (infile); | |
456 | } | |
457 | ||
74c55c82 | 458 | /* Lisp variable or function name. */ |
f2cc4248 RS |
459 | c = getc (infile); |
460 | if (c != '"') | |
461 | continue; | |
74c55c82 GM |
462 | c = read_c_string_or_comment (infile, -1, 0); |
463 | ||
464 | /* DEFVAR_LISP ("name", addr /\* doc *\/) | |
465 | DEFVAR_LISP ("name", addr, doc) */ | |
f2cc4248 RS |
466 | |
467 | if (defunflag) | |
468 | commas = 5; | |
84128dee JB |
469 | else if (defvarperbufferflag) |
470 | commas = 2; | |
f2cc4248 RS |
471 | else if (defvarflag) |
472 | commas = 1; | |
473 | else /* For DEFSIMPLE and DEFPRED */ | |
474 | commas = 2; | |
475 | ||
476 | while (commas) | |
477 | { | |
478 | if (c == ',') | |
479 | { | |
480 | commas--; | |
74c55c82 | 481 | |
f2cc4248 RS |
482 | if (defunflag && (commas == 1 || commas == 2)) |
483 | { | |
484 | do | |
485 | c = getc (infile); | |
433d333d | 486 | while (c == ' ' || c == '\n' || c == '\r' || c == '\t'); |
f2cc4248 RS |
487 | if (c < 0) |
488 | goto eof; | |
489 | ungetc (c, infile); | |
490 | if (commas == 2) /* pick up minargs */ | |
491 | fscanf (infile, "%d", &minargs); | |
492 | else /* pick up maxargs */ | |
493 | if (c == 'M' || c == 'U') /* MANY || UNEVALLED */ | |
494 | maxargs = -1; | |
495 | else | |
496 | fscanf (infile, "%d", &maxargs); | |
497 | } | |
498 | } | |
74c55c82 GM |
499 | |
500 | if (c == EOF) | |
f2cc4248 RS |
501 | goto eof; |
502 | c = getc (infile); | |
503 | } | |
74c55c82 | 504 | |
433d333d | 505 | while (c == ' ' || c == '\n' || c == '\r' || c == '\t') |
f2cc4248 | 506 | c = getc (infile); |
74c55c82 | 507 | |
f2cc4248 | 508 | if (c == '"') |
74c55c82 GM |
509 | c = read_c_string_or_comment (infile, 0, 0); |
510 | ||
511 | while (c != EOF && c != ',' && c != '/') | |
f2cc4248 | 512 | c = getc (infile); |
74c55c82 GM |
513 | if (c == ',') |
514 | { | |
515 | c = getc (infile); | |
516 | while (c == ' ' || c == '\n' || c == '\r' || c == '\t') | |
517 | c = getc (infile); | |
518 | } | |
f2cc4248 | 519 | |
74c55c82 GM |
520 | if (c == '"' |
521 | || (c == '/' | |
522 | && (c = getc (infile), | |
523 | ungetc (c, infile), | |
524 | c == '*'))) | |
f2cc4248 | 525 | { |
74c55c82 GM |
526 | int comment = c != '"'; |
527 | ||
f2cc4248 RS |
528 | putc (037, outfile); |
529 | putc (defvarflag ? 'V' : 'F', outfile); | |
530 | fprintf (outfile, "%s\n", buf); | |
74c55c82 GM |
531 | |
532 | if (comment) | |
533 | getc (infile); /* Skip past `*' */ | |
534 | c = read_c_string_or_comment (infile, 1, comment); | |
772e2009 JB |
535 | |
536 | /* If this is a defun, find the arguments and print them. If | |
537 | this function takes MANY or UNEVALLED args, then the C source | |
538 | won't give the names of the arguments, so we shouldn't bother | |
74c55c82 GM |
539 | trying to find them. |
540 | ||
541 | Old: DEFUN (..., "DOC") (args) | |
542 | New: DEFUN (..., /\* DOC *\/ (args)) */ | |
772e2009 | 543 | if (defunflag && maxargs != -1) |
f2cc4248 RS |
544 | { |
545 | char argbuf[1024], *p = argbuf; | |
74c55c82 GM |
546 | |
547 | if (!comment) | |
548 | while (c != ')') | |
549 | { | |
550 | if (c < 0) | |
551 | goto eof; | |
552 | c = getc (infile); | |
553 | } | |
554 | ||
f2cc4248 RS |
555 | /* Skip into arguments. */ |
556 | while (c != '(') | |
557 | { | |
558 | if (c < 0) | |
559 | goto eof; | |
560 | c = getc (infile); | |
561 | } | |
562 | /* Copy arguments into ARGBUF. */ | |
563 | *p++ = c; | |
564 | do | |
565 | *p++ = c = getc (infile); | |
566 | while (c != ')'); | |
567 | *p = '\0'; | |
568 | /* Output them. */ | |
569 | fprintf (outfile, "\n\n"); | |
069ad9ea | 570 | write_c_args (outfile, buf, argbuf, minargs, maxargs); |
f2cc4248 RS |
571 | } |
572 | } | |
573 | } | |
574 | eof: | |
575 | fclose (infile); | |
576 | return 0; | |
577 | } | |
578 | \f | |
579 | /* Read a file of Lisp code, compiled or interpreted. | |
580 | Looks for | |
581 | (defun NAME ARGS DOCSTRING ...) | |
23d6b5a6 | 582 | (defmacro NAME ARGS DOCSTRING ...) |
34e778a6 | 583 | (defsubst NAME ARGS DOCSTRING ...) |
23d6b5a6 | 584 | (autoload (quote NAME) FILE DOCSTRING ...) |
f2cc4248 RS |
585 | (defvar NAME VALUE DOCSTRING) |
586 | (defconst NAME VALUE DOCSTRING) | |
23d6b5a6 JB |
587 | (fset (quote NAME) (make-byte-code ... DOCSTRING ...)) |
588 | (fset (quote NAME) #[... DOCSTRING ...]) | |
2d6e2619 | 589 | (defalias (quote NAME) #[... DOCSTRING ...]) |
3fe77f98 | 590 | (custom-declare-variable (quote NAME) VALUE DOCSTRING ...) |
f2cc4248 | 591 | starting in column zero. |
23d6b5a6 | 592 | (quote NAME) may appear as 'NAME as well. |
b5ff43cc RS |
593 | |
594 | We also look for #@LENGTH CONTENTS^_ at the beginning of the line. | |
595 | When we find that, we save it for the following defining-form, | |
596 | and we use that instead of reading a doc string within that defining-form. | |
597 | ||
eb8c3be9 | 598 | For defvar, defconst, and fset we skip to the docstring with a kludgy |
23d6b5a6 JB |
599 | formatting convention: all docstrings must appear on the same line as the |
600 | initial open-paren (the one in column zero) and must contain a backslash | |
b0f08a24 | 601 | and a newline immediately after the initial double-quote. No newlines |
23d6b5a6 | 602 | must appear between the beginning of the form and the first double-quote. |
b0f08a24 DL |
603 | For defun, defmacro, and autoload, we know how to skip over the |
604 | arglist, but the doc string must still have a backslash and newline | |
605 | immediately after the double quote. | |
606 | The only source files that must follow this convention are preloaded | |
607 | uncompiled ones like loaddefs.el and bindings.el; aside | |
23d6b5a6 JB |
608 | from that, it is always the .elc file that we look at, and they are no |
609 | problem because byte-compiler output follows this convention. | |
f2cc4248 RS |
610 | The NAME and DOCSTRING are output. |
611 | NAME is preceded by `F' for a function or `V' for a variable. | |
612 | An entry is output only if DOCSTRING has \ newline just after the opening " | |
613 | */ | |
614 | ||
23d6b5a6 JB |
615 | void |
616 | skip_white (infile) | |
617 | FILE *infile; | |
618 | { | |
619 | char c = ' '; | |
433d333d | 620 | while (c == ' ' || c == '\t' || c == '\n' || c == '\r') |
23d6b5a6 JB |
621 | c = getc (infile); |
622 | ungetc (c, infile); | |
623 | } | |
624 | ||
625 | void | |
626 | read_lisp_symbol (infile, buffer) | |
627 | FILE *infile; | |
628 | char *buffer; | |
629 | { | |
630 | char c; | |
631 | char *fillp = buffer; | |
632 | ||
633 | skip_white (infile); | |
634 | while (1) | |
635 | { | |
636 | c = getc (infile); | |
637 | if (c == '\\') | |
638 | *(++fillp) = getc (infile); | |
433d333d | 639 | else if (c == ' ' || c == '\t' || c == '\n' || c == '\r' || c == '(' || c == ')') |
23d6b5a6 JB |
640 | { |
641 | ungetc (c, infile); | |
642 | *fillp = 0; | |
643 | break; | |
644 | } | |
645 | else | |
646 | *fillp++ = c; | |
647 | } | |
648 | ||
649 | if (! buffer[0]) | |
650 | fprintf (stderr, "## expected a symbol, got '%c'\n", c); | |
651 | ||
652 | skip_white (infile); | |
653 | } | |
654 | ||
340ff9de | 655 | int |
e3938952 RS |
656 | scan_lisp_file (filename, mode) |
657 | char *filename, *mode; | |
f2cc4248 RS |
658 | { |
659 | FILE *infile; | |
660 | register int c; | |
b5ff43cc | 661 | char *saved_string = 0; |
f2cc4248 | 662 | |
e3938952 | 663 | infile = fopen (filename, mode); |
f2cc4248 RS |
664 | if (infile == NULL) |
665 | { | |
666 | perror (filename); | |
667 | return 0; /* No error */ | |
668 | } | |
669 | ||
670 | c = '\n'; | |
671 | while (!feof (infile)) | |
672 | { | |
b5ff43cc | 673 | char buffer[BUFSIZ]; |
23d6b5a6 JB |
674 | char type; |
675 | ||
66f54605 | 676 | /* If not at end of line, skip till we get to one. */ |
433d333d | 677 | if (c != '\n' && c != '\r') |
f2cc4248 RS |
678 | { |
679 | c = getc (infile); | |
680 | continue; | |
681 | } | |
66f54605 | 682 | /* Skip the line break. */ |
7e6972e9 | 683 | while (c == '\n' || c == '\r') |
66f54605 | 684 | c = getc (infile); |
b5ff43cc RS |
685 | /* Detect a dynamic doc string and save it for the next expression. */ |
686 | if (c == '#') | |
687 | { | |
688 | c = getc (infile); | |
689 | if (c == '@') | |
690 | { | |
691 | int length = 0; | |
692 | int i; | |
693 | ||
694 | /* Read the length. */ | |
695 | while ((c = getc (infile), | |
696 | c >= '0' && c <= '9')) | |
697 | { | |
698 | length *= 10; | |
699 | length += c - '0'; | |
700 | } | |
701 | ||
702 | /* The next character is a space that is counted in the length | |
703 | but not part of the doc string. | |
704 | We already read it, so just ignore it. */ | |
705 | length--; | |
706 | ||
707 | /* Read in the contents. */ | |
708 | if (saved_string != 0) | |
709 | free (saved_string); | |
710 | saved_string = (char *) malloc (length); | |
711 | for (i = 0; i < length; i++) | |
712 | saved_string[i] = getc (infile); | |
713 | /* The last character is a ^_. | |
714 | That is needed in the .elc file | |
715 | but it is redundant in DOC. So get rid of it here. */ | |
716 | saved_string[length - 1] = 0; | |
66f54605 PR |
717 | /* Skip the line break. */ |
718 | while (c == '\n' && c == '\r') | |
719 | c = getc (infile); | |
720 | /* Skip the following line. */ | |
433d333d | 721 | while (c != '\n' && c != '\r') |
b5ff43cc RS |
722 | c = getc (infile); |
723 | } | |
724 | continue; | |
725 | } | |
726 | ||
f2cc4248 RS |
727 | if (c != '(') |
728 | continue; | |
a8a7afbe | 729 | |
23d6b5a6 JB |
730 | read_lisp_symbol (infile, buffer); |
731 | ||
66f54605 | 732 | if (! strcmp (buffer, "defun") |
34e778a6 AS |
733 | || ! strcmp (buffer, "defmacro") |
734 | || ! strcmp (buffer, "defsubst")) | |
f2cc4248 | 735 | { |
23d6b5a6 JB |
736 | type = 'F'; |
737 | read_lisp_symbol (infile, buffer); | |
f2cc4248 | 738 | |
23d6b5a6 | 739 | /* Skip the arguments: either "nil" or a list in parens */ |
f2cc4248 | 740 | |
23d6b5a6 JB |
741 | c = getc (infile); |
742 | if (c == 'n') /* nil */ | |
f2cc4248 | 743 | { |
66f54605 PR |
744 | if ((c = getc (infile)) != 'i' |
745 | || (c = getc (infile)) != 'l') | |
23d6b5a6 JB |
746 | { |
747 | fprintf (stderr, "## unparsable arglist in %s (%s)\n", | |
748 | buffer, filename); | |
749 | continue; | |
750 | } | |
f2cc4248 | 751 | } |
23d6b5a6 | 752 | else if (c != '(') |
f2cc4248 | 753 | { |
23d6b5a6 JB |
754 | fprintf (stderr, "## unparsable arglist in %s (%s)\n", |
755 | buffer, filename); | |
756 | continue; | |
f2cc4248 | 757 | } |
23d6b5a6 JB |
758 | else |
759 | while (c != ')') | |
f2cc4248 | 760 | c = getc (infile); |
23d6b5a6 JB |
761 | skip_white (infile); |
762 | ||
763 | /* If the next three characters aren't `dquote bslash newline' | |
764 | then we're not reading a docstring. | |
765 | */ | |
66f54605 PR |
766 | if ((c = getc (infile)) != '"' |
767 | || (c = getc (infile)) != '\\' | |
768 | || ((c = getc (infile)) != '\n' && c != '\r')) | |
f2cc4248 | 769 | { |
23d6b5a6 JB |
770 | #ifdef DEBUG |
771 | fprintf (stderr, "## non-docstring in %s (%s)\n", | |
772 | buffer, filename); | |
773 | #endif | |
774 | continue; | |
f2cc4248 | 775 | } |
f2cc4248 | 776 | } |
a8a7afbe | 777 | |
66f54605 PR |
778 | else if (! strcmp (buffer, "defvar") |
779 | || ! strcmp (buffer, "defconst")) | |
f2cc4248 | 780 | { |
23d6b5a6 JB |
781 | char c1 = 0, c2 = 0; |
782 | type = 'V'; | |
783 | read_lisp_symbol (infile, buffer); | |
a8a7afbe | 784 | |
b5ff43cc | 785 | if (saved_string == 0) |
f2cc4248 | 786 | { |
b5ff43cc | 787 | |
66f54605 | 788 | /* Skip until the end of line; remember two previous chars. */ |
433d333d | 789 | while (c != '\n' && c != '\r' && c >= 0) |
b5ff43cc RS |
790 | { |
791 | c2 = c1; | |
792 | c1 = c; | |
793 | c = getc (infile); | |
794 | } | |
23d6b5a6 | 795 | |
b5ff43cc RS |
796 | /* If two previous characters were " and \, |
797 | this is a doc string. Otherwise, there is none. */ | |
798 | if (c2 != '"' || c1 != '\\') | |
799 | { | |
23d6b5a6 | 800 | #ifdef DEBUG |
b5ff43cc RS |
801 | fprintf (stderr, "## non-docstring in %s (%s)\n", |
802 | buffer, filename); | |
23d6b5a6 | 803 | #endif |
b5ff43cc RS |
804 | continue; |
805 | } | |
f2cc4248 | 806 | } |
23d6b5a6 JB |
807 | } |
808 | ||
3fe77f98 RS |
809 | else if (! strcmp (buffer, "custom-declare-variable")) |
810 | { | |
811 | char c1 = 0, c2 = 0; | |
812 | type = 'V'; | |
813 | ||
814 | c = getc (infile); | |
815 | if (c == '\'') | |
816 | read_lisp_symbol (infile, buffer); | |
817 | else | |
818 | { | |
819 | if (c != '(') | |
820 | { | |
821 | fprintf (stderr, | |
822 | "## unparsable name in custom-declare-variable in %s\n", | |
823 | filename); | |
824 | continue; | |
825 | } | |
826 | read_lisp_symbol (infile, buffer); | |
827 | if (strcmp (buffer, "quote")) | |
828 | { | |
829 | fprintf (stderr, | |
830 | "## unparsable name in custom-declare-variable in %s\n", | |
831 | filename); | |
832 | continue; | |
833 | } | |
834 | read_lisp_symbol (infile, buffer); | |
835 | c = getc (infile); | |
836 | if (c != ')') | |
837 | { | |
838 | fprintf (stderr, | |
839 | "## unparsable quoted name in custom-declare-variable in %s\n", | |
840 | filename); | |
841 | continue; | |
842 | } | |
843 | } | |
844 | ||
845 | if (saved_string == 0) | |
846 | { | |
66f54605 | 847 | /* Skip to end of line; remember the two previous chars. */ |
433d333d | 848 | while (c != '\n' && c != '\r' && c >= 0) |
3fe77f98 RS |
849 | { |
850 | c2 = c1; | |
851 | c1 = c; | |
852 | c = getc (infile); | |
853 | } | |
854 | ||
855 | /* If two previous characters were " and \, | |
856 | this is a doc string. Otherwise, there is none. */ | |
857 | if (c2 != '"' || c1 != '\\') | |
858 | { | |
859 | #ifdef DEBUG | |
860 | fprintf (stderr, "## non-docstring in %s (%s)\n", | |
861 | buffer, filename); | |
862 | #endif | |
863 | continue; | |
864 | } | |
865 | } | |
866 | } | |
867 | ||
2d6e2619 | 868 | else if (! strcmp (buffer, "fset") || ! strcmp (buffer, "defalias")) |
23d6b5a6 JB |
869 | { |
870 | char c1 = 0, c2 = 0; | |
871 | type = 'F'; | |
a8a7afbe | 872 | |
23d6b5a6 JB |
873 | c = getc (infile); |
874 | if (c == '\'') | |
875 | read_lisp_symbol (infile, buffer); | |
876 | else | |
f2cc4248 | 877 | { |
23d6b5a6 JB |
878 | if (c != '(') |
879 | { | |
880 | fprintf (stderr, "## unparsable name in fset in %s\n", | |
881 | filename); | |
882 | continue; | |
883 | } | |
884 | read_lisp_symbol (infile, buffer); | |
885 | if (strcmp (buffer, "quote")) | |
886 | { | |
887 | fprintf (stderr, "## unparsable name in fset in %s\n", | |
888 | filename); | |
889 | continue; | |
890 | } | |
891 | read_lisp_symbol (infile, buffer); | |
f2cc4248 | 892 | c = getc (infile); |
23d6b5a6 JB |
893 | if (c != ')') |
894 | { | |
895 | fprintf (stderr, | |
896 | "## unparsable quoted name in fset in %s\n", | |
897 | filename); | |
898 | continue; | |
899 | } | |
f2cc4248 | 900 | } |
f2cc4248 | 901 | |
b5ff43cc | 902 | if (saved_string == 0) |
f2cc4248 | 903 | { |
66f54605 | 904 | /* Skip to end of line; remember the two previous chars. */ |
433d333d | 905 | while (c != '\n' && c != '\r' && c >= 0) |
b5ff43cc RS |
906 | { |
907 | c2 = c1; | |
908 | c1 = c; | |
909 | c = getc (infile); | |
910 | } | |
23d6b5a6 | 911 | |
b5ff43cc RS |
912 | /* If two previous characters were " and \, |
913 | this is a doc string. Otherwise, there is none. */ | |
914 | if (c2 != '"' || c1 != '\\') | |
915 | { | |
23d6b5a6 | 916 | #ifdef DEBUG |
b5ff43cc RS |
917 | fprintf (stderr, "## non-docstring in %s (%s)\n", |
918 | buffer, filename); | |
23d6b5a6 | 919 | #endif |
b5ff43cc RS |
920 | continue; |
921 | } | |
23d6b5a6 JB |
922 | } |
923 | } | |
f2cc4248 | 924 | |
23d6b5a6 JB |
925 | else if (! strcmp (buffer, "autoload")) |
926 | { | |
927 | type = 'F'; | |
928 | c = getc (infile); | |
929 | if (c == '\'') | |
930 | read_lisp_symbol (infile, buffer); | |
931 | else | |
f2cc4248 | 932 | { |
23d6b5a6 | 933 | if (c != '(') |
f2cc4248 | 934 | { |
23d6b5a6 JB |
935 | fprintf (stderr, "## unparsable name in autoload in %s\n", |
936 | filename); | |
937 | continue; | |
f2cc4248 | 938 | } |
23d6b5a6 JB |
939 | read_lisp_symbol (infile, buffer); |
940 | if (strcmp (buffer, "quote")) | |
f2cc4248 | 941 | { |
23d6b5a6 JB |
942 | fprintf (stderr, "## unparsable name in autoload in %s\n", |
943 | filename); | |
944 | continue; | |
f2cc4248 | 945 | } |
23d6b5a6 | 946 | read_lisp_symbol (infile, buffer); |
f2cc4248 | 947 | c = getc (infile); |
23d6b5a6 | 948 | if (c != ')') |
f2cc4248 | 949 | { |
23d6b5a6 JB |
950 | fprintf (stderr, |
951 | "## unparsable quoted name in autoload in %s\n", | |
952 | filename); | |
953 | continue; | |
f2cc4248 | 954 | } |
23d6b5a6 JB |
955 | } |
956 | skip_white (infile); | |
957 | if ((c = getc (infile)) != '\"') | |
958 | { | |
959 | fprintf (stderr, "## autoload of %s unparsable (%s)\n", | |
960 | buffer, filename); | |
f2cc4248 RS |
961 | continue; |
962 | } | |
74c55c82 | 963 | read_c_string_or_comment (infile, 0, 0); |
23d6b5a6 JB |
964 | skip_white (infile); |
965 | ||
b5ff43cc | 966 | if (saved_string == 0) |
a8a7afbe | 967 | { |
b5ff43cc RS |
968 | /* If the next three characters aren't `dquote bslash newline' |
969 | then we're not reading a docstring. */ | |
66f54605 PR |
970 | if ((c = getc (infile)) != '"' |
971 | || (c = getc (infile)) != '\\' | |
972 | || ((c = getc (infile)) != '\n' && c != '\r')) | |
b5ff43cc | 973 | { |
23d6b5a6 | 974 | #ifdef DEBUG |
b5ff43cc RS |
975 | fprintf (stderr, "## non-docstring in %s (%s)\n", |
976 | buffer, filename); | |
23d6b5a6 | 977 | #endif |
b5ff43cc RS |
978 | continue; |
979 | } | |
a8a7afbe | 980 | } |
a8a7afbe | 981 | } |
f2cc4248 | 982 | |
23d6b5a6 | 983 | #ifdef DEBUG |
66f54605 PR |
984 | else if (! strcmp (buffer, "if") |
985 | || ! strcmp (buffer, "byte-code")) | |
23d6b5a6 JB |
986 | ; |
987 | #endif | |
f2cc4248 | 988 | |
23d6b5a6 JB |
989 | else |
990 | { | |
991 | #ifdef DEBUG | |
992 | fprintf (stderr, "## unrecognised top-level form, %s (%s)\n", | |
993 | buffer, filename); | |
994 | #endif | |
995 | continue; | |
996 | } | |
f2cc4248 | 997 | |
b5ff43cc RS |
998 | /* At this point, we should either use the previous |
999 | dynamic doc string in saved_string | |
1000 | or gobble a doc string from the input file. | |
1001 | ||
1002 | In the latter case, the opening quote (and leading | |
1003 | backslash-newline) have already been read. */ | |
1004 | ||
f2cc4248 | 1005 | putc (037, outfile); |
23d6b5a6 JB |
1006 | putc (type, outfile); |
1007 | fprintf (outfile, "%s\n", buffer); | |
b5ff43cc RS |
1008 | if (saved_string) |
1009 | { | |
1010 | fputs (saved_string, outfile); | |
1011 | /* Don't use one dynamic doc string twice. */ | |
1012 | free (saved_string); | |
1013 | saved_string = 0; | |
1014 | } | |
1015 | else | |
74c55c82 | 1016 | read_c_string_or_comment (infile, 1, 0); |
f2cc4248 RS |
1017 | } |
1018 | fclose (infile); | |
1019 | return 0; | |
1020 | } |