| 1 | /* Generate doc-string file for GNU Emacs from source files. |
| 2 | Copyright (C) 1985, 86, 92, 93, 94, 97, 1999, 2000, 2001 |
| 3 | Free Software Foundation, Inc. |
| 4 | |
| 5 | This file is part of GNU Emacs. |
| 6 | |
| 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 |
| 9 | the Free Software Foundation; either version 2, or (at your option) |
| 10 | any later version. |
| 11 | |
| 12 | GNU Emacs is distributed in the hope that it will be useful, |
| 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. |
| 16 | |
| 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 |
| 19 | the Free Software Foundation, Inc., 59 Temple Place - Suite 330, |
| 20 | Boston, MA 02111-1307, USA. */ |
| 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 | |
| 36 | #define NO_SHORTNAMES /* Tell config not to load remap.h */ |
| 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 |
| 43 | |
| 44 | #include <stdio.h> |
| 45 | #ifdef MSDOS |
| 46 | #include <fcntl.h> |
| 47 | #endif /* MSDOS */ |
| 48 | #ifdef WINDOWSNT |
| 49 | #include <stdlib.h> |
| 50 | #include <fcntl.h> |
| 51 | #include <direct.h> |
| 52 | #endif /* WINDOWSNT */ |
| 53 | |
| 54 | #ifdef DOS_NT |
| 55 | #define READ_TEXT "rt" |
| 56 | #define READ_BINARY "rb" |
| 57 | #else /* not DOS_NT */ |
| 58 | #define READ_TEXT "r" |
| 59 | #define READ_BINARY "r" |
| 60 | #endif /* not DOS_NT */ |
| 61 | |
| 62 | int scan_file (); |
| 63 | int scan_lisp_file (); |
| 64 | int scan_c_file (); |
| 65 | |
| 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 | |
| 72 | #ifdef HAVE_UNISTD_H |
| 73 | #include <unistd.h> |
| 74 | #endif |
| 75 | |
| 76 | /* Stdio stream for output to the DOC file. */ |
| 77 | FILE *outfile; |
| 78 | |
| 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 | |
| 107 | long * |
| 108 | xmalloc (size) |
| 109 | unsigned int size; |
| 110 | { |
| 111 | long *result = (long *) malloc (size); |
| 112 | if (result == NULL) |
| 113 | fatal ("virtual memory exhausted", 0); |
| 114 | return result; |
| 115 | } |
| 116 | \f |
| 117 | int |
| 118 | main (argc, argv) |
| 119 | int argc; |
| 120 | char **argv; |
| 121 | { |
| 122 | int i; |
| 123 | int err_count = 0; |
| 124 | int first_infile; |
| 125 | |
| 126 | progname = argv[0]; |
| 127 | |
| 128 | outfile = stdout; |
| 129 | |
| 130 | /* Don't put CRs in the DOC file. */ |
| 131 | #ifdef MSDOS |
| 132 | _fmode = O_BINARY; |
| 133 | #if 0 /* Suspicion is that this causes hanging. |
| 134 | So instead we require people to use -o on MSDOS. */ |
| 135 | (stdout)->_flag &= ~_IOTEXT; |
| 136 | _setmode (fileno (stdout), O_BINARY); |
| 137 | #endif |
| 138 | outfile = 0; |
| 139 | #endif /* MSDOS */ |
| 140 | #ifdef WINDOWSNT |
| 141 | _fmode = O_BINARY; |
| 142 | _setmode (fileno (stdout), O_BINARY); |
| 143 | #endif /* WINDOWSNT */ |
| 144 | |
| 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 | } |
| 157 | if (argc > i + 1 && !strcmp (argv[i], "-d")) |
| 158 | { |
| 159 | chdir (argv[i + 1]); |
| 160 | i += 2; |
| 161 | } |
| 162 | |
| 163 | if (outfile == 0) |
| 164 | fatal ("No output file specified", ""); |
| 165 | |
| 166 | first_infile = i; |
| 167 | for (; i < argc; i++) |
| 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 | } |
| 177 | #ifndef VMS |
| 178 | exit (err_count > 0); |
| 179 | #endif /* VMS */ |
| 180 | return err_count > 0; |
| 181 | } |
| 182 | |
| 183 | /* Read file FILENAME and output its doc strings to outfile. */ |
| 184 | /* Return 1 if file is not found, 0 if it is found. */ |
| 185 | |
| 186 | int |
| 187 | scan_file (filename) |
| 188 | char *filename; |
| 189 | { |
| 190 | int len = strlen (filename); |
| 191 | if (len > 4 && !strcmp (filename + len - 4, ".elc")) |
| 192 | return scan_lisp_file (filename, READ_BINARY); |
| 193 | else if (len > 3 && !strcmp (filename + len - 3, ".el")) |
| 194 | return scan_lisp_file (filename, READ_TEXT); |
| 195 | else |
| 196 | return scan_c_file (filename, READ_TEXT); |
| 197 | } |
| 198 | \f |
| 199 | char buf[128]; |
| 200 | |
| 201 | /* Some state during the execution of `read_c_string_or_comment'. */ |
| 202 | struct rcsoc_state |
| 203 | { |
| 204 | /* A count of spaces and newlines that have been read, but not output. */ |
| 205 | unsigned pending_spaces, pending_newlines; |
| 206 | |
| 207 | /* Where we're reading from. */ |
| 208 | FILE *in_file; |
| 209 | |
| 210 | /* If non-zero, a buffer into which to copy characters. */ |
| 211 | char *buf_ptr; |
| 212 | /* If non-zero, a file into which to copy characters. */ |
| 213 | FILE *out_file; |
| 214 | |
| 215 | /* A keyword we look for at the beginning of lines. If found, it is |
| 216 | not copied, and SAW_KEYWORD is set to true. */ |
| 217 | char *keyword; |
| 218 | /* The current point we've reached in an occurance of KEYWORD in |
| 219 | the input stream. */ |
| 220 | char *cur_keyword_ptr; |
| 221 | /* Set to true if we saw an occurance of KEYWORD. */ |
| 222 | int saw_keyword; |
| 223 | }; |
| 224 | |
| 225 | /* Output CH to the file or buffer in STATE. Any pending newlines or |
| 226 | spaces are output first. */ |
| 227 | |
| 228 | static INLINE void |
| 229 | put_char (ch, state) |
| 230 | int ch; |
| 231 | struct rcsoc_state *state; |
| 232 | { |
| 233 | int out_ch; |
| 234 | do |
| 235 | { |
| 236 | if (state->pending_newlines > 0) |
| 237 | { |
| 238 | state->pending_newlines--; |
| 239 | out_ch = '\n'; |
| 240 | } |
| 241 | else if (state->pending_spaces > 0) |
| 242 | { |
| 243 | state->pending_spaces--; |
| 244 | out_ch = ' '; |
| 245 | } |
| 246 | else |
| 247 | out_ch = ch; |
| 248 | |
| 249 | if (state->out_file) |
| 250 | putc (out_ch, state->out_file); |
| 251 | if (state->buf_ptr) |
| 252 | *state->buf_ptr++ = out_ch; |
| 253 | } |
| 254 | while (out_ch != ch); |
| 255 | } |
| 256 | |
| 257 | /* If in the middle of scanning a keyword, continue scanning with |
| 258 | character CH, otherwise output CH to the file or buffer in STATE. |
| 259 | Any pending newlines or spaces are output first, as well as any |
| 260 | previously scanned characters that were thought to be part of a |
| 261 | keyword, but were in fact not. */ |
| 262 | |
| 263 | static void |
| 264 | scan_keyword_or_put_char (ch, state) |
| 265 | int ch; |
| 266 | struct rcsoc_state *state; |
| 267 | { |
| 268 | if (state->keyword |
| 269 | && *state->cur_keyword_ptr == ch |
| 270 | && (state->cur_keyword_ptr > state->keyword |
| 271 | || state->pending_newlines > 0)) |
| 272 | /* We might be looking at STATE->keyword at some point. |
| 273 | Keep looking until we know for sure. */ |
| 274 | { |
| 275 | if (*++state->cur_keyword_ptr == '\0') |
| 276 | /* Saw the whole keyword. Set SAW_KEYWORD flag to true. */ |
| 277 | { |
| 278 | state->saw_keyword = 1; |
| 279 | |
| 280 | /* Reset the scanning pointer. */ |
| 281 | state->cur_keyword_ptr = state->keyword; |
| 282 | |
| 283 | /* Canonicalize whitespace preceding a usage string. */ |
| 284 | state->pending_newlines = 2; |
| 285 | state->pending_spaces = 0; |
| 286 | |
| 287 | /* Skip any whitespace between the keyword and the |
| 288 | usage string. */ |
| 289 | do |
| 290 | ch = getc (state->in_file); |
| 291 | while (ch == ' ' || ch == '\n'); |
| 292 | |
| 293 | /* Output the open-paren we just read. */ |
| 294 | put_char (ch, state); |
| 295 | |
| 296 | /* Skip the function name and replace it with `fn'. */ |
| 297 | do |
| 298 | ch = getc (state->in_file); |
| 299 | while (ch != ' ' && ch != ')'); |
| 300 | put_char ('f', state); |
| 301 | put_char ('n', state); |
| 302 | |
| 303 | /* Put back the last character. */ |
| 304 | ungetc (ch, state->in_file); |
| 305 | } |
| 306 | } |
| 307 | else |
| 308 | { |
| 309 | if (state->keyword && state->cur_keyword_ptr > state->keyword) |
| 310 | /* We scanned the beginning of a potential usage |
| 311 | keyword, but it was a false alarm. Output the |
| 312 | part we scanned. */ |
| 313 | { |
| 314 | char *p; |
| 315 | |
| 316 | for (p = state->keyword; p < state->cur_keyword_ptr; p++) |
| 317 | put_char (*p, state); |
| 318 | |
| 319 | state->cur_keyword_ptr = state->keyword; |
| 320 | } |
| 321 | |
| 322 | put_char (ch, state); |
| 323 | } |
| 324 | } |
| 325 | |
| 326 | |
| 327 | /* Skip a C string or C-style comment from INFILE, and return the |
| 328 | character that follows. COMMENT non-zero means skip a comment. If |
| 329 | PRINTFLAG is positive, output string contents to outfile. If it is |
| 330 | negative, store contents in buf. Convert escape sequences \n and |
| 331 | \t to newline and tab; discard \ followed by newline. |
| 332 | If SAW_USAGE is non-zero, then any occurances of the string `usage:' |
| 333 | at the beginning of a line will be removed, and *SAW_USAGE set to |
| 334 | true if any were encountered. */ |
| 335 | |
| 336 | int |
| 337 | read_c_string_or_comment (infile, printflag, comment, saw_usage) |
| 338 | FILE *infile; |
| 339 | int printflag; |
| 340 | int *saw_usage; |
| 341 | { |
| 342 | register int c; |
| 343 | struct rcsoc_state state; |
| 344 | |
| 345 | state.in_file = infile; |
| 346 | state.buf_ptr = (printflag < 0 ? buf : 0); |
| 347 | state.out_file = (printflag > 0 ? outfile : 0); |
| 348 | state.pending_spaces = 0; |
| 349 | state.pending_newlines = 0; |
| 350 | state.keyword = (saw_usage ? "usage:" : 0); |
| 351 | state.cur_keyword_ptr = state.keyword; |
| 352 | state.saw_keyword = 0; |
| 353 | |
| 354 | c = getc (infile); |
| 355 | if (comment) |
| 356 | while (c == '\n' || c == '\r' || c == '\t' || c == ' ') |
| 357 | c = getc (infile); |
| 358 | |
| 359 | while (c != EOF) |
| 360 | { |
| 361 | while (c != EOF && (comment ? c != '*' : c != '"')) |
| 362 | { |
| 363 | if (c == '\\') |
| 364 | { |
| 365 | c = getc (infile); |
| 366 | if (c == '\n' || c == '\r') |
| 367 | { |
| 368 | c = getc (infile); |
| 369 | continue; |
| 370 | } |
| 371 | if (c == 'n') |
| 372 | c = '\n'; |
| 373 | if (c == 't') |
| 374 | c = '\t'; |
| 375 | } |
| 376 | |
| 377 | if (c == ' ') |
| 378 | state.pending_spaces++; |
| 379 | else if (c == '\n') |
| 380 | { |
| 381 | state.pending_newlines++; |
| 382 | state.pending_spaces = 0; |
| 383 | } |
| 384 | else |
| 385 | scan_keyword_or_put_char (c, &state); |
| 386 | |
| 387 | c = getc (infile); |
| 388 | } |
| 389 | |
| 390 | if (c != EOF) |
| 391 | c = getc (infile); |
| 392 | |
| 393 | if (comment) |
| 394 | { |
| 395 | if (c == '/') |
| 396 | { |
| 397 | c = getc (infile); |
| 398 | break; |
| 399 | } |
| 400 | |
| 401 | scan_keyword_or_put_char ('*', &state); |
| 402 | } |
| 403 | else |
| 404 | { |
| 405 | if (c != '"') |
| 406 | break; |
| 407 | |
| 408 | /* If we had a "", concatenate the two strings. */ |
| 409 | c = getc (infile); |
| 410 | } |
| 411 | } |
| 412 | |
| 413 | if (printflag < 0) |
| 414 | *state.buf_ptr = 0; |
| 415 | |
| 416 | if (saw_usage) |
| 417 | *saw_usage = state.saw_keyword; |
| 418 | |
| 419 | return c; |
| 420 | } |
| 421 | |
| 422 | |
| 423 | \f |
| 424 | /* Write to file OUT the argument names of function FUNC, whose text is in BUF. |
| 425 | MINARGS and MAXARGS are the minimum and maximum number of arguments. */ |
| 426 | |
| 427 | void |
| 428 | write_c_args (out, func, buf, minargs, maxargs) |
| 429 | FILE *out; |
| 430 | char *func, *buf; |
| 431 | int minargs, maxargs; |
| 432 | { |
| 433 | register char *p; |
| 434 | int in_ident = 0; |
| 435 | int just_spaced = 0; |
| 436 | int need_space = 1; |
| 437 | |
| 438 | fprintf (out, "(fn"); |
| 439 | |
| 440 | if (*buf == '(') |
| 441 | ++buf; |
| 442 | |
| 443 | for (p = buf; *p; p++) |
| 444 | { |
| 445 | char c = *p; |
| 446 | int ident_start = 0; |
| 447 | |
| 448 | /* Notice when we start printing a new identifier. */ |
| 449 | if ((('A' <= c && c <= 'Z') |
| 450 | || ('a' <= c && c <= 'z') |
| 451 | || ('0' <= c && c <= '9') |
| 452 | || c == '_') |
| 453 | != in_ident) |
| 454 | { |
| 455 | if (!in_ident) |
| 456 | { |
| 457 | in_ident = 1; |
| 458 | ident_start = 1; |
| 459 | |
| 460 | if (need_space) |
| 461 | putc (' ', out); |
| 462 | |
| 463 | if (minargs == 0 && maxargs > 0) |
| 464 | fprintf (out, "&optional "); |
| 465 | just_spaced = 1; |
| 466 | |
| 467 | minargs--; |
| 468 | maxargs--; |
| 469 | } |
| 470 | else |
| 471 | in_ident = 0; |
| 472 | } |
| 473 | |
| 474 | /* Print the C argument list as it would appear in lisp: |
| 475 | print underscores as hyphens, and print commas and newlines |
| 476 | as spaces. Collapse adjacent spaces into one. */ |
| 477 | if (c == '_') |
| 478 | c = '-'; |
| 479 | else if (c == ',' || c == '\n') |
| 480 | c = ' '; |
| 481 | |
| 482 | /* In C code, `default' is a reserved word, so we spell it |
| 483 | `defalt'; unmangle that here. */ |
| 484 | if (ident_start |
| 485 | && strncmp (p, "defalt", 6) == 0 |
| 486 | && ! (('A' <= p[6] && p[6] <= 'Z') |
| 487 | || ('a' <= p[6] && p[6] <= 'z') |
| 488 | || ('0' <= p[6] && p[6] <= '9') |
| 489 | || p[6] == '_')) |
| 490 | { |
| 491 | fprintf (out, "DEFAULT"); |
| 492 | p += 5; |
| 493 | in_ident = 0; |
| 494 | just_spaced = 0; |
| 495 | } |
| 496 | else if (c != ' ' || !just_spaced) |
| 497 | { |
| 498 | if (c >= 'a' && c <= 'z') |
| 499 | /* Upcase the letter. */ |
| 500 | c += 'A' - 'a'; |
| 501 | putc (c, out); |
| 502 | } |
| 503 | |
| 504 | just_spaced = c == ' '; |
| 505 | need_space = 0; |
| 506 | } |
| 507 | } |
| 508 | \f |
| 509 | /* Read through a c file. If a .o file is named, |
| 510 | the corresponding .c file is read instead. |
| 511 | Looks for DEFUN constructs such as are defined in ../src/lisp.h. |
| 512 | Accepts any word starting DEF... so it finds DEFSIMPLE and DEFPRED. */ |
| 513 | |
| 514 | int |
| 515 | scan_c_file (filename, mode) |
| 516 | char *filename, *mode; |
| 517 | { |
| 518 | FILE *infile; |
| 519 | register int c; |
| 520 | register int commas; |
| 521 | register int defunflag; |
| 522 | register int defvarperbufferflag; |
| 523 | register int defvarflag; |
| 524 | int minargs, maxargs; |
| 525 | int extension = filename[strlen (filename) - 1]; |
| 526 | |
| 527 | if (extension == 'o') |
| 528 | filename[strlen (filename) - 1] = 'c'; |
| 529 | |
| 530 | infile = fopen (filename, mode); |
| 531 | |
| 532 | /* No error if non-ex input file */ |
| 533 | if (infile == NULL) |
| 534 | { |
| 535 | perror (filename); |
| 536 | return 0; |
| 537 | } |
| 538 | |
| 539 | /* Reset extension to be able to detect duplicate files. */ |
| 540 | filename[strlen (filename) - 1] = extension; |
| 541 | |
| 542 | c = '\n'; |
| 543 | while (!feof (infile)) |
| 544 | { |
| 545 | int doc_keyword = 0; |
| 546 | |
| 547 | if (c != '\n' && c != '\r') |
| 548 | { |
| 549 | c = getc (infile); |
| 550 | continue; |
| 551 | } |
| 552 | c = getc (infile); |
| 553 | if (c == ' ') |
| 554 | { |
| 555 | while (c == ' ') |
| 556 | c = getc (infile); |
| 557 | if (c != 'D') |
| 558 | continue; |
| 559 | c = getc (infile); |
| 560 | if (c != 'E') |
| 561 | continue; |
| 562 | c = getc (infile); |
| 563 | if (c != 'F') |
| 564 | continue; |
| 565 | c = getc (infile); |
| 566 | if (c != 'V') |
| 567 | continue; |
| 568 | c = getc (infile); |
| 569 | if (c != 'A') |
| 570 | continue; |
| 571 | c = getc (infile); |
| 572 | if (c != 'R') |
| 573 | continue; |
| 574 | c = getc (infile); |
| 575 | if (c != '_') |
| 576 | continue; |
| 577 | |
| 578 | defvarflag = 1; |
| 579 | defunflag = 0; |
| 580 | |
| 581 | c = getc (infile); |
| 582 | defvarperbufferflag = (c == 'P'); |
| 583 | |
| 584 | c = getc (infile); |
| 585 | } |
| 586 | else if (c == 'D') |
| 587 | { |
| 588 | c = getc (infile); |
| 589 | if (c != 'E') |
| 590 | continue; |
| 591 | c = getc (infile); |
| 592 | if (c != 'F') |
| 593 | continue; |
| 594 | c = getc (infile); |
| 595 | defunflag = c == 'U'; |
| 596 | defvarflag = 0; |
| 597 | } |
| 598 | else continue; |
| 599 | |
| 600 | while (c != '(') |
| 601 | { |
| 602 | if (c < 0) |
| 603 | goto eof; |
| 604 | c = getc (infile); |
| 605 | } |
| 606 | |
| 607 | /* Lisp variable or function name. */ |
| 608 | c = getc (infile); |
| 609 | if (c != '"') |
| 610 | continue; |
| 611 | c = read_c_string_or_comment (infile, -1, 0, 0); |
| 612 | |
| 613 | /* DEFVAR_LISP ("name", addr, "doc") |
| 614 | DEFVAR_LISP ("name", addr /\* doc *\/) |
| 615 | DEFVAR_LISP ("name", addr, doc: /\* doc *\/) */ |
| 616 | |
| 617 | if (defunflag) |
| 618 | commas = 5; |
| 619 | else if (defvarperbufferflag) |
| 620 | commas = 2; |
| 621 | else if (defvarflag) |
| 622 | commas = 1; |
| 623 | else /* For DEFSIMPLE and DEFPRED */ |
| 624 | commas = 2; |
| 625 | |
| 626 | while (commas) |
| 627 | { |
| 628 | if (c == ',') |
| 629 | { |
| 630 | commas--; |
| 631 | |
| 632 | if (defunflag && (commas == 1 || commas == 2)) |
| 633 | { |
| 634 | do |
| 635 | c = getc (infile); |
| 636 | while (c == ' ' || c == '\n' || c == '\r' || c == '\t'); |
| 637 | if (c < 0) |
| 638 | goto eof; |
| 639 | ungetc (c, infile); |
| 640 | if (commas == 2) /* pick up minargs */ |
| 641 | fscanf (infile, "%d", &minargs); |
| 642 | else /* pick up maxargs */ |
| 643 | if (c == 'M' || c == 'U') /* MANY || UNEVALLED */ |
| 644 | maxargs = -1; |
| 645 | else |
| 646 | fscanf (infile, "%d", &maxargs); |
| 647 | } |
| 648 | } |
| 649 | |
| 650 | if (c == EOF) |
| 651 | goto eof; |
| 652 | c = getc (infile); |
| 653 | } |
| 654 | |
| 655 | while (c == ' ' || c == '\n' || c == '\r' || c == '\t') |
| 656 | c = getc (infile); |
| 657 | |
| 658 | if (c == '"') |
| 659 | c = read_c_string_or_comment (infile, 0, 0, 0); |
| 660 | |
| 661 | while (c != EOF && c != ',' && c != '/') |
| 662 | c = getc (infile); |
| 663 | if (c == ',') |
| 664 | { |
| 665 | c = getc (infile); |
| 666 | while (c == ' ' || c == '\n' || c == '\r' || c == '\t') |
| 667 | c = getc (infile); |
| 668 | while ((c >= 'a' && c <= 'z') || (c >= 'Z' && c <= 'Z')) |
| 669 | c = getc (infile); |
| 670 | if (c == ':') |
| 671 | { |
| 672 | doc_keyword = 1; |
| 673 | c = getc (infile); |
| 674 | while (c == ' ' || c == '\n' || c == '\r' || c == '\t') |
| 675 | c = getc (infile); |
| 676 | } |
| 677 | } |
| 678 | |
| 679 | if (c == '"' |
| 680 | || (c == '/' |
| 681 | && (c = getc (infile), |
| 682 | ungetc (c, infile), |
| 683 | c == '*'))) |
| 684 | { |
| 685 | int comment = c != '"'; |
| 686 | int saw_usage; |
| 687 | |
| 688 | putc (037, outfile); |
| 689 | putc (defvarflag ? 'V' : 'F', outfile); |
| 690 | fprintf (outfile, "%s\n", buf); |
| 691 | |
| 692 | if (comment) |
| 693 | getc (infile); /* Skip past `*' */ |
| 694 | c = read_c_string_or_comment (infile, 1, comment, &saw_usage); |
| 695 | |
| 696 | /* If this is a defun, find the arguments and print them. If |
| 697 | this function takes MANY or UNEVALLED args, then the C source |
| 698 | won't give the names of the arguments, so we shouldn't bother |
| 699 | trying to find them. |
| 700 | |
| 701 | Various doc-string styles: |
| 702 | 0: DEFUN (..., "DOC") (args) [!comment] |
| 703 | 1: DEFUN (..., /\* DOC *\/ (args)) [comment && !doc_keyword] |
| 704 | 2: DEFUN (..., doc: /\* DOC *\/) (args) [comment && doc_keyword] |
| 705 | */ |
| 706 | if (defunflag && maxargs != -1 && !saw_usage) |
| 707 | { |
| 708 | char argbuf[1024], *p = argbuf; |
| 709 | |
| 710 | if (!comment || doc_keyword) |
| 711 | while (c != ')') |
| 712 | { |
| 713 | if (c < 0) |
| 714 | goto eof; |
| 715 | c = getc (infile); |
| 716 | } |
| 717 | |
| 718 | /* Skip into arguments. */ |
| 719 | while (c != '(') |
| 720 | { |
| 721 | if (c < 0) |
| 722 | goto eof; |
| 723 | c = getc (infile); |
| 724 | } |
| 725 | /* Copy arguments into ARGBUF. */ |
| 726 | *p++ = c; |
| 727 | do |
| 728 | *p++ = c = getc (infile); |
| 729 | while (c != ')'); |
| 730 | *p = '\0'; |
| 731 | /* Output them. */ |
| 732 | fprintf (outfile, "\n\n"); |
| 733 | write_c_args (outfile, buf, argbuf, minargs, maxargs); |
| 734 | } |
| 735 | else if (defunflag && maxargs == -1 && !saw_usage) |
| 736 | /* The DOC should provide the usage form. */ |
| 737 | fprintf (stderr, "Missing `usage' for function `%s'.\n", buf); |
| 738 | } |
| 739 | } |
| 740 | eof: |
| 741 | fclose (infile); |
| 742 | return 0; |
| 743 | } |
| 744 | \f |
| 745 | /* Read a file of Lisp code, compiled or interpreted. |
| 746 | Looks for |
| 747 | (defun NAME ARGS DOCSTRING ...) |
| 748 | (defmacro NAME ARGS DOCSTRING ...) |
| 749 | (defsubst NAME ARGS DOCSTRING ...) |
| 750 | (autoload (quote NAME) FILE DOCSTRING ...) |
| 751 | (defvar NAME VALUE DOCSTRING) |
| 752 | (defconst NAME VALUE DOCSTRING) |
| 753 | (fset (quote NAME) (make-byte-code ... DOCSTRING ...)) |
| 754 | (fset (quote NAME) #[... DOCSTRING ...]) |
| 755 | (defalias (quote NAME) #[... DOCSTRING ...]) |
| 756 | (custom-declare-variable (quote NAME) VALUE DOCSTRING ...) |
| 757 | starting in column zero. |
| 758 | (quote NAME) may appear as 'NAME as well. |
| 759 | |
| 760 | We also look for #@LENGTH CONTENTS^_ at the beginning of the line. |
| 761 | When we find that, we save it for the following defining-form, |
| 762 | and we use that instead of reading a doc string within that defining-form. |
| 763 | |
| 764 | For defvar, defconst, and fset we skip to the docstring with a kludgy |
| 765 | formatting convention: all docstrings must appear on the same line as the |
| 766 | initial open-paren (the one in column zero) and must contain a backslash |
| 767 | and a newline immediately after the initial double-quote. No newlines |
| 768 | must appear between the beginning of the form and the first double-quote. |
| 769 | For defun, defmacro, and autoload, we know how to skip over the |
| 770 | arglist, but the doc string must still have a backslash and newline |
| 771 | immediately after the double quote. |
| 772 | The only source files that must follow this convention are preloaded |
| 773 | uncompiled ones like loaddefs.el and bindings.el; aside |
| 774 | from that, it is always the .elc file that we look at, and they are no |
| 775 | problem because byte-compiler output follows this convention. |
| 776 | The NAME and DOCSTRING are output. |
| 777 | NAME is preceded by `F' for a function or `V' for a variable. |
| 778 | An entry is output only if DOCSTRING has \ newline just after the opening " |
| 779 | */ |
| 780 | |
| 781 | void |
| 782 | skip_white (infile) |
| 783 | FILE *infile; |
| 784 | { |
| 785 | char c = ' '; |
| 786 | while (c == ' ' || c == '\t' || c == '\n' || c == '\r') |
| 787 | c = getc (infile); |
| 788 | ungetc (c, infile); |
| 789 | } |
| 790 | |
| 791 | void |
| 792 | read_lisp_symbol (infile, buffer) |
| 793 | FILE *infile; |
| 794 | char *buffer; |
| 795 | { |
| 796 | char c; |
| 797 | char *fillp = buffer; |
| 798 | |
| 799 | skip_white (infile); |
| 800 | while (1) |
| 801 | { |
| 802 | c = getc (infile); |
| 803 | if (c == '\\') |
| 804 | *(++fillp) = getc (infile); |
| 805 | else if (c == ' ' || c == '\t' || c == '\n' || c == '\r' || c == '(' || c == ')') |
| 806 | { |
| 807 | ungetc (c, infile); |
| 808 | *fillp = 0; |
| 809 | break; |
| 810 | } |
| 811 | else |
| 812 | *fillp++ = c; |
| 813 | } |
| 814 | |
| 815 | if (! buffer[0]) |
| 816 | fprintf (stderr, "## expected a symbol, got '%c'\n", c); |
| 817 | |
| 818 | skip_white (infile); |
| 819 | } |
| 820 | |
| 821 | int |
| 822 | scan_lisp_file (filename, mode) |
| 823 | char *filename, *mode; |
| 824 | { |
| 825 | FILE *infile; |
| 826 | register int c; |
| 827 | char *saved_string = 0; |
| 828 | |
| 829 | infile = fopen (filename, mode); |
| 830 | if (infile == NULL) |
| 831 | { |
| 832 | perror (filename); |
| 833 | return 0; /* No error */ |
| 834 | } |
| 835 | |
| 836 | c = '\n'; |
| 837 | while (!feof (infile)) |
| 838 | { |
| 839 | char buffer[BUFSIZ]; |
| 840 | char type; |
| 841 | |
| 842 | /* If not at end of line, skip till we get to one. */ |
| 843 | if (c != '\n' && c != '\r') |
| 844 | { |
| 845 | c = getc (infile); |
| 846 | continue; |
| 847 | } |
| 848 | /* Skip the line break. */ |
| 849 | while (c == '\n' || c == '\r') |
| 850 | c = getc (infile); |
| 851 | /* Detect a dynamic doc string and save it for the next expression. */ |
| 852 | if (c == '#') |
| 853 | { |
| 854 | c = getc (infile); |
| 855 | if (c == '@') |
| 856 | { |
| 857 | int length = 0; |
| 858 | int i; |
| 859 | |
| 860 | /* Read the length. */ |
| 861 | while ((c = getc (infile), |
| 862 | c >= '0' && c <= '9')) |
| 863 | { |
| 864 | length *= 10; |
| 865 | length += c - '0'; |
| 866 | } |
| 867 | |
| 868 | /* The next character is a space that is counted in the length |
| 869 | but not part of the doc string. |
| 870 | We already read it, so just ignore it. */ |
| 871 | length--; |
| 872 | |
| 873 | /* Read in the contents. */ |
| 874 | if (saved_string != 0) |
| 875 | free (saved_string); |
| 876 | saved_string = (char *) malloc (length); |
| 877 | for (i = 0; i < length; i++) |
| 878 | saved_string[i] = getc (infile); |
| 879 | /* The last character is a ^_. |
| 880 | That is needed in the .elc file |
| 881 | but it is redundant in DOC. So get rid of it here. */ |
| 882 | saved_string[length - 1] = 0; |
| 883 | /* Skip the line break. */ |
| 884 | while (c == '\n' && c == '\r') |
| 885 | c = getc (infile); |
| 886 | /* Skip the following line. */ |
| 887 | while (c != '\n' && c != '\r') |
| 888 | c = getc (infile); |
| 889 | } |
| 890 | continue; |
| 891 | } |
| 892 | |
| 893 | if (c != '(') |
| 894 | continue; |
| 895 | |
| 896 | read_lisp_symbol (infile, buffer); |
| 897 | |
| 898 | if (! strcmp (buffer, "defun") |
| 899 | || ! strcmp (buffer, "defmacro") |
| 900 | || ! strcmp (buffer, "defsubst")) |
| 901 | { |
| 902 | type = 'F'; |
| 903 | read_lisp_symbol (infile, buffer); |
| 904 | |
| 905 | /* Skip the arguments: either "nil" or a list in parens */ |
| 906 | |
| 907 | c = getc (infile); |
| 908 | if (c == 'n') /* nil */ |
| 909 | { |
| 910 | if ((c = getc (infile)) != 'i' |
| 911 | || (c = getc (infile)) != 'l') |
| 912 | { |
| 913 | fprintf (stderr, "## unparsable arglist in %s (%s)\n", |
| 914 | buffer, filename); |
| 915 | continue; |
| 916 | } |
| 917 | } |
| 918 | else if (c != '(') |
| 919 | { |
| 920 | fprintf (stderr, "## unparsable arglist in %s (%s)\n", |
| 921 | buffer, filename); |
| 922 | continue; |
| 923 | } |
| 924 | else |
| 925 | while (c != ')') |
| 926 | c = getc (infile); |
| 927 | skip_white (infile); |
| 928 | |
| 929 | /* If the next three characters aren't `dquote bslash newline' |
| 930 | then we're not reading a docstring. |
| 931 | */ |
| 932 | if ((c = getc (infile)) != '"' |
| 933 | || (c = getc (infile)) != '\\' |
| 934 | || ((c = getc (infile)) != '\n' && c != '\r')) |
| 935 | { |
| 936 | #ifdef DEBUG |
| 937 | fprintf (stderr, "## non-docstring in %s (%s)\n", |
| 938 | buffer, filename); |
| 939 | #endif |
| 940 | continue; |
| 941 | } |
| 942 | } |
| 943 | |
| 944 | else if (! strcmp (buffer, "defvar") |
| 945 | || ! strcmp (buffer, "defconst")) |
| 946 | { |
| 947 | char c1 = 0, c2 = 0; |
| 948 | type = 'V'; |
| 949 | read_lisp_symbol (infile, buffer); |
| 950 | |
| 951 | if (saved_string == 0) |
| 952 | { |
| 953 | |
| 954 | /* Skip until the end of line; remember two previous chars. */ |
| 955 | while (c != '\n' && c != '\r' && c >= 0) |
| 956 | { |
| 957 | c2 = c1; |
| 958 | c1 = c; |
| 959 | c = getc (infile); |
| 960 | } |
| 961 | |
| 962 | /* If two previous characters were " and \, |
| 963 | this is a doc string. Otherwise, there is none. */ |
| 964 | if (c2 != '"' || c1 != '\\') |
| 965 | { |
| 966 | #ifdef DEBUG |
| 967 | fprintf (stderr, "## non-docstring in %s (%s)\n", |
| 968 | buffer, filename); |
| 969 | #endif |
| 970 | continue; |
| 971 | } |
| 972 | } |
| 973 | } |
| 974 | |
| 975 | else if (! strcmp (buffer, "custom-declare-variable")) |
| 976 | { |
| 977 | char c1 = 0, c2 = 0; |
| 978 | type = 'V'; |
| 979 | |
| 980 | c = getc (infile); |
| 981 | if (c == '\'') |
| 982 | read_lisp_symbol (infile, buffer); |
| 983 | else |
| 984 | { |
| 985 | if (c != '(') |
| 986 | { |
| 987 | fprintf (stderr, |
| 988 | "## unparsable name in custom-declare-variable in %s\n", |
| 989 | filename); |
| 990 | continue; |
| 991 | } |
| 992 | read_lisp_symbol (infile, buffer); |
| 993 | if (strcmp (buffer, "quote")) |
| 994 | { |
| 995 | fprintf (stderr, |
| 996 | "## unparsable name in custom-declare-variable in %s\n", |
| 997 | filename); |
| 998 | continue; |
| 999 | } |
| 1000 | read_lisp_symbol (infile, buffer); |
| 1001 | c = getc (infile); |
| 1002 | if (c != ')') |
| 1003 | { |
| 1004 | fprintf (stderr, |
| 1005 | "## unparsable quoted name in custom-declare-variable in %s\n", |
| 1006 | filename); |
| 1007 | continue; |
| 1008 | } |
| 1009 | } |
| 1010 | |
| 1011 | if (saved_string == 0) |
| 1012 | { |
| 1013 | /* Skip to end of line; remember the two previous chars. */ |
| 1014 | while (c != '\n' && c != '\r' && c >= 0) |
| 1015 | { |
| 1016 | c2 = c1; |
| 1017 | c1 = c; |
| 1018 | c = getc (infile); |
| 1019 | } |
| 1020 | |
| 1021 | /* If two previous characters were " and \, |
| 1022 | this is a doc string. Otherwise, there is none. */ |
| 1023 | if (c2 != '"' || c1 != '\\') |
| 1024 | { |
| 1025 | #ifdef DEBUG |
| 1026 | fprintf (stderr, "## non-docstring in %s (%s)\n", |
| 1027 | buffer, filename); |
| 1028 | #endif |
| 1029 | continue; |
| 1030 | } |
| 1031 | } |
| 1032 | } |
| 1033 | |
| 1034 | else if (! strcmp (buffer, "fset") || ! strcmp (buffer, "defalias")) |
| 1035 | { |
| 1036 | char c1 = 0, c2 = 0; |
| 1037 | type = 'F'; |
| 1038 | |
| 1039 | c = getc (infile); |
| 1040 | if (c == '\'') |
| 1041 | read_lisp_symbol (infile, buffer); |
| 1042 | else |
| 1043 | { |
| 1044 | if (c != '(') |
| 1045 | { |
| 1046 | fprintf (stderr, "## unparsable name in fset in %s\n", |
| 1047 | filename); |
| 1048 | continue; |
| 1049 | } |
| 1050 | read_lisp_symbol (infile, buffer); |
| 1051 | if (strcmp (buffer, "quote")) |
| 1052 | { |
| 1053 | fprintf (stderr, "## unparsable name in fset in %s\n", |
| 1054 | filename); |
| 1055 | continue; |
| 1056 | } |
| 1057 | read_lisp_symbol (infile, buffer); |
| 1058 | c = getc (infile); |
| 1059 | if (c != ')') |
| 1060 | { |
| 1061 | fprintf (stderr, |
| 1062 | "## unparsable quoted name in fset in %s\n", |
| 1063 | filename); |
| 1064 | continue; |
| 1065 | } |
| 1066 | } |
| 1067 | |
| 1068 | if (saved_string == 0) |
| 1069 | { |
| 1070 | /* Skip to end of line; remember the two previous chars. */ |
| 1071 | while (c != '\n' && c != '\r' && c >= 0) |
| 1072 | { |
| 1073 | c2 = c1; |
| 1074 | c1 = c; |
| 1075 | c = getc (infile); |
| 1076 | } |
| 1077 | |
| 1078 | /* If two previous characters were " and \, |
| 1079 | this is a doc string. Otherwise, there is none. */ |
| 1080 | if (c2 != '"' || c1 != '\\') |
| 1081 | { |
| 1082 | #ifdef DEBUG |
| 1083 | fprintf (stderr, "## non-docstring in %s (%s)\n", |
| 1084 | buffer, filename); |
| 1085 | #endif |
| 1086 | continue; |
| 1087 | } |
| 1088 | } |
| 1089 | } |
| 1090 | |
| 1091 | else if (! strcmp (buffer, "autoload")) |
| 1092 | { |
| 1093 | type = 'F'; |
| 1094 | c = getc (infile); |
| 1095 | if (c == '\'') |
| 1096 | read_lisp_symbol (infile, buffer); |
| 1097 | else |
| 1098 | { |
| 1099 | if (c != '(') |
| 1100 | { |
| 1101 | fprintf (stderr, "## unparsable name in autoload in %s\n", |
| 1102 | filename); |
| 1103 | continue; |
| 1104 | } |
| 1105 | read_lisp_symbol (infile, buffer); |
| 1106 | if (strcmp (buffer, "quote")) |
| 1107 | { |
| 1108 | fprintf (stderr, "## unparsable name in autoload in %s\n", |
| 1109 | filename); |
| 1110 | continue; |
| 1111 | } |
| 1112 | read_lisp_symbol (infile, buffer); |
| 1113 | c = getc (infile); |
| 1114 | if (c != ')') |
| 1115 | { |
| 1116 | fprintf (stderr, |
| 1117 | "## unparsable quoted name in autoload in %s\n", |
| 1118 | filename); |
| 1119 | continue; |
| 1120 | } |
| 1121 | } |
| 1122 | skip_white (infile); |
| 1123 | if ((c = getc (infile)) != '\"') |
| 1124 | { |
| 1125 | fprintf (stderr, "## autoload of %s unparsable (%s)\n", |
| 1126 | buffer, filename); |
| 1127 | continue; |
| 1128 | } |
| 1129 | read_c_string_or_comment (infile, 0, 0, 0); |
| 1130 | skip_white (infile); |
| 1131 | |
| 1132 | if (saved_string == 0) |
| 1133 | { |
| 1134 | /* If the next three characters aren't `dquote bslash newline' |
| 1135 | then we're not reading a docstring. */ |
| 1136 | if ((c = getc (infile)) != '"' |
| 1137 | || (c = getc (infile)) != '\\' |
| 1138 | || ((c = getc (infile)) != '\n' && c != '\r')) |
| 1139 | { |
| 1140 | #ifdef DEBUG |
| 1141 | fprintf (stderr, "## non-docstring in %s (%s)\n", |
| 1142 | buffer, filename); |
| 1143 | #endif |
| 1144 | continue; |
| 1145 | } |
| 1146 | } |
| 1147 | } |
| 1148 | |
| 1149 | #ifdef DEBUG |
| 1150 | else if (! strcmp (buffer, "if") |
| 1151 | || ! strcmp (buffer, "byte-code")) |
| 1152 | ; |
| 1153 | #endif |
| 1154 | |
| 1155 | else |
| 1156 | { |
| 1157 | #ifdef DEBUG |
| 1158 | fprintf (stderr, "## unrecognised top-level form, %s (%s)\n", |
| 1159 | buffer, filename); |
| 1160 | #endif |
| 1161 | continue; |
| 1162 | } |
| 1163 | |
| 1164 | /* At this point, we should either use the previous |
| 1165 | dynamic doc string in saved_string |
| 1166 | or gobble a doc string from the input file. |
| 1167 | |
| 1168 | In the latter case, the opening quote (and leading |
| 1169 | backslash-newline) have already been read. */ |
| 1170 | |
| 1171 | putc (037, outfile); |
| 1172 | putc (type, outfile); |
| 1173 | fprintf (outfile, "%s\n", buffer); |
| 1174 | if (saved_string) |
| 1175 | { |
| 1176 | fputs (saved_string, outfile); |
| 1177 | /* Don't use one dynamic doc string twice. */ |
| 1178 | free (saved_string); |
| 1179 | saved_string = 0; |
| 1180 | } |
| 1181 | else |
| 1182 | read_c_string_or_comment (infile, 1, 0, 0); |
| 1183 | } |
| 1184 | fclose (infile); |
| 1185 | return 0; |
| 1186 | } |