Commit | Line | Data |
---|---|---|
c6d46f5f | 1 | /* Tags file maker to go with GNU Emacs |
13fde0cd | 2 | Copyright (C) 1984, 1987, 1988, 1989, 1993 Free Software Foundation, Inc. and Ken Arnold |
c6d46f5f | 3 | |
ea6cd314 | 4 | This file is not considered part of GNU Emacs. |
c6d46f5f | 5 | |
ea6cd314 | 6 | This program is free software; you can redistribute it and/or modify |
c6d46f5f | 7 | it under the terms of the GNU General Public License as published by |
ea6cd314 RS |
8 | the Free Software Foundation; either version 2 of the License, or |
9 | (at your option) any later version. | |
c6d46f5f | 10 | |
ea6cd314 | 11 | This program is distributed in the hope that it will be useful, |
c6d46f5f JB |
12 | but WITHOUT ANY WARRANTY; without even the implied warranty of |
13 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
14 | GNU General Public License for more details. | |
15 | ||
16 | You should have received a copy of the GNU General Public License | |
ea6cd314 RS |
17 | along with this program; if not, write to the Free Software |
18 | Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */ | |
c6d46f5f JB |
19 | |
20 | /* | |
21 | * Authors: | |
22 | * Ctags originally by Ken Arnold. | |
23 | * FORTRAN added by Jim Kleckner. | |
24 | * Ed Pelegri-Llopart added C typedefs. | |
25 | * Gnu Emacs TAGS format and modifications by RMS? | |
26 | * Sam Kendall added C++. | |
27 | */ | |
28 | ||
29 | #include <stdio.h> | |
30 | #include <ctype.h> | |
31 | #include <sys/types.h> | |
32 | #include <sys/stat.h> | |
33 | ||
b827c6c1 RS |
34 | #include "../src/config.h" |
35 | ||
4746118a JB |
36 | #include "getopt.h" |
37 | ||
1e134a5f RM |
38 | #ifdef __GNUC__ |
39 | #define alloca __builtin_alloca | |
40 | #else | |
41 | #ifdef sparc | |
42 | #include <alloca.h> | |
43 | #else | |
44 | extern char *alloca (); | |
45 | #endif | |
46 | #endif | |
47 | ||
c6d46f5f JB |
48 | extern char *malloc (), *realloc (); |
49 | extern char *getenv (); | |
c6d46f5f JB |
50 | extern char *strcpy (), *strncpy (); |
51 | extern int strcmp (); | |
52 | ||
8a6c8bcf RS |
53 | char *etags_index (), *etags_rindex (); |
54 | char *savenstr (); | |
c6d46f5f JB |
55 | |
56 | /* Define the symbol ETAGS to make the program "etags", | |
57 | which makes emacs-style tag tables by default. | |
58 | Define CTAGS to make the program "ctags" compatible with the usual one. | |
59 | Define neither one to get behavior that depends | |
60 | on the name with which the program is invoked | |
61 | (but we don't normally compile it that way). */ | |
62 | ||
63 | #if !defined(ETAGS) && !defined(CTAGS) | |
64 | /* If neither is defined, program can be run as either. */ | |
65 | #define ETAGS | |
66 | #define CTAGS | |
67 | #endif | |
68 | ||
69 | /* On VMS, CTAGS is not useful, so always do ETAGS. */ | |
70 | #ifdef VMS | |
71 | #ifndef ETAGS | |
72 | #define ETAGS | |
73 | #endif | |
74 | #endif | |
75 | ||
76 | /* Exit codes for success and failure. */ | |
77 | #ifdef VMS | |
78 | #define GOOD (1) | |
79 | #define BAD (0) | |
80 | #else | |
81 | #define GOOD (0) | |
82 | #define BAD (1) | |
83 | #endif | |
84 | ||
85 | /* | |
86 | * The FILEPOS abstract type, which represents a position in a file, | |
87 | * plus the following accessor functions: | |
88 | * | |
89 | * long GET_CHARNO (pos) | |
90 | * returns absolute char number. | |
c6d46f5f JB |
91 | * void SET_FILEPOS (pos, fp, charno) |
92 | * FILE *fp; long charno; | |
93 | * sets `pos' from the current file | |
94 | * position of `fp' and from `charno', | |
95 | * which must be the absolute character | |
96 | * number corresponding to the current | |
97 | * position of `fp'. | |
98 | * | |
99 | * The `pos' parameter is an lvalue expression of type FILEPOS. | |
100 | * Parameters to the accessor functions are evaluated 0 or more times, | |
101 | * and so must have no side effects. | |
102 | * | |
103 | * FILEPOS objects can also be assigned and passed to and from | |
104 | * functions in the normal C manner. | |
105 | * | |
106 | * Implementation notes: the `+ 0' is to enforce rvalue-ness. | |
107 | */ | |
c6d46f5f | 108 | |
c6d46f5f | 109 | #ifndef DEBUG |
13fde0cd | 110 | /* real implementation */ |
c6d46f5f JB |
111 | typedef long FILEPOS; |
112 | #define GET_CHARNO(pos) ((pos) + 0) | |
c6d46f5f JB |
113 | #define SET_FILEPOS(pos, fp, cno) ((void) ((pos) = (cno))) |
114 | #else | |
13fde0cd | 115 | /* debugging implementation */ |
c6d46f5f JB |
116 | typedef struct |
117 | { | |
118 | long charno; | |
119 | } FILEPOS; | |
120 | ||
121 | #define GET_CHARNO(pos) ((pos).charno + 0) | |
c6d46f5f JB |
122 | #define SET_FILEPOS(pos, fp, cno) \ |
123 | ((void) ((pos).charno = (cno), \ | |
124 | (cno) != ftell (fp) ? (error ("SET_FILEPOS inconsistency"), 0) \ | |
125 | : 0)) | |
126 | #endif | |
c6d46f5f JB |
127 | |
128 | #define streq(s, t) (strcmp (s, t) == 0) | |
129 | #define strneq(s, t, n) (strncmp (s, t, n) == 0) | |
13fde0cd | 130 | #define logical int |
c6d46f5f JB |
131 | |
132 | #define TRUE 1 | |
133 | #define FALSE 0 | |
134 | ||
135 | #define iswhite(arg) (_wht[arg]) /* T if char is white */ | |
136 | #define begtoken(arg) (_btk[arg]) /* T if char can start token */ | |
137 | #define intoken(arg) (_itk[arg]) /* T if char can be in token */ | |
138 | #define endtoken(arg) (_etk[arg]) /* T if char ends tokens */ | |
c6d46f5f JB |
139 | |
140 | #define max(I1,I2) ((I1) > (I2) ? (I1) : (I2)) | |
141 | ||
142 | struct nd_st | |
143 | { /* sorting structure */ | |
144 | char *name; /* function or type name */ | |
145 | char *file; /* file name */ | |
146 | logical is_func; /* use pattern or line no */ | |
147 | logical rewritten; /* list name separately */ | |
148 | logical been_warned; /* set if noticed dup */ | |
149 | int lno; /* line number tag is on */ | |
150 | long cno; /* character number line starts on */ | |
151 | char *pat; /* search pattern */ | |
152 | struct nd_st *left, *right; /* left and right sons */ | |
153 | }; | |
154 | ||
155 | long ftell (); | |
156 | typedef struct nd_st NODE; | |
157 | ||
13fde0cd RS |
158 | logical header_file; /* TRUE if .h file, FALSE o.w. */ |
159 | /* boolean "functions" (see init) */ | |
160 | logical _wht[0177], _etk[0177], _itk[0177], _btk[0177]; | |
c6d46f5f JB |
161 | |
162 | ||
163 | char *concat (); | |
164 | char *savenstr (); | |
165 | char *savestr (); | |
166 | char *xmalloc (); | |
167 | char *xrealloc (); | |
168 | int L_isdef (); | |
169 | int PF_funcs (); | |
170 | int total_size_of_entries (); | |
171 | logical consider_token (); | |
172 | logical tail (); | |
173 | long readline (); | |
174 | void Asm_funcs (); | |
175 | void C_entries (); | |
176 | void L_funcs (); | |
177 | void L_getit (); | |
178 | void PAS_funcs (); | |
179 | void Scheme_funcs (); | |
180 | void TEX_funcs (); | |
181 | void add_node (); | |
182 | void error (); | |
183 | void fatal (); | |
184 | void find_entries (); | |
185 | void free_tree (); | |
186 | void getit (); | |
187 | void getline (); | |
188 | void init (); | |
189 | void initbuffer (); | |
190 | void initbuffer (); | |
191 | void pfnote (); | |
192 | void process_file (); | |
193 | void put_entries (); | |
194 | void takeprec (); | |
195 | ||
196 | /* | |
197 | * MACRO | |
198 | * xnew -- allocate storage | |
199 | * | |
200 | * SYNOPSIS | |
201 | * Type *xnew (int n, Type); | |
202 | */ | |
203 | #define xnew(n, Type) ((Type *) xmalloc ((n) * sizeof (Type))) | |
204 | \f | |
205 | ||
206 | ||
207 | /* | |
208 | * Symbol table stuff. | |
209 | * | |
210 | * Should probably be implemented with hash table; linked list for now. | |
211 | */ | |
212 | ||
213 | enum sym_type | |
214 | { | |
215 | st_none, st_C_struct, st_C_enum, st_C_define, st_C_typedef, st_C_typespec | |
216 | }; | |
217 | ||
218 | struct stab_entry | |
219 | { | |
220 | char *sym; | |
221 | int symlen; | |
222 | enum sym_type type; | |
223 | struct stab_entry *next; | |
224 | }; | |
225 | ||
226 | typedef struct stab_entry Stab_entry; | |
227 | typedef Stab_entry *Stab; | |
228 | ||
229 | /* | |
230 | * NAME | |
231 | * Stab, Stab_entry, stab_create, stab_search, stab_find -- symbol table | |
232 | * | |
233 | * SYNOPSIS | |
234 | * Types: Stab, Stab_entry, enum sym_type | |
235 | * | |
236 | * Stab * stab_create () | |
237 | * | |
238 | * Stab_entry * stab_find (stab, sym) | |
239 | * Stab *stab; | |
240 | * char *sym; | |
241 | * | |
242 | * Stab_entry * stab_search (stab, sym) | |
243 | * Stab *stab; | |
244 | * char *sym; | |
245 | * | |
246 | * DESCRIPTION | |
247 | * stab_create creates a Stab, a symbol table object, and returns a | |
248 | * pointer to it. stab_find finds a symbol in a Stab; it returns a | |
249 | * pointer to the Stab_entry if found, otherwise NULL. stab_search | |
250 | * is like stab_find, except that it creates a new Stab_entry, | |
251 | * initialized with type = st_none, if one did not exist already | |
252 | * (it never returns NULL). | |
253 | * | |
254 | * A Stab_entry is a structure that contains at least the following | |
255 | * members: | |
256 | * | |
257 | * char *name; // must not be modified | |
258 | * enum sym_type type; // should be set | |
259 | * | |
260 | * The type field is initially set to st_none; it should be set to | |
261 | * something else by the caller of stab_search. Other possible values | |
262 | * of an enum sym_type can be added. | |
263 | */ | |
264 | ||
265 | Stab * | |
266 | stab_create () | |
267 | { | |
268 | Stab *sp; | |
269 | sp = xnew (1, Stab); | |
270 | *sp = NULL; /* a Stab starts out as a null Stab_entry* */ | |
271 | return sp; | |
272 | } | |
273 | ||
274 | Stab_entry * | |
275 | stab_find (stab, sym, symlen) | |
276 | Stab *stab; | |
277 | register char *sym; | |
278 | register int symlen; | |
279 | { | |
280 | register Stab_entry *se; | |
281 | for (se = *stab; se != NULL; se = se->next) | |
282 | { | |
283 | if (se->symlen == symlen && strneq (se->sym, sym, symlen)) | |
284 | return se; | |
285 | } | |
286 | ||
287 | return NULL; | |
288 | } | |
289 | ||
290 | Stab_entry * | |
291 | stab_search (stab, sym, symlen) | |
292 | register Stab *stab; | |
293 | char *sym; | |
294 | int symlen; | |
295 | { | |
296 | register Stab_entry *se; | |
297 | se = stab_find (stab, sym, symlen); | |
298 | ||
299 | if (se == NULL) | |
300 | { | |
301 | /* make a new one */ | |
302 | se = xnew (1, Stab_entry); | |
303 | se->sym = savenstr (sym, symlen); | |
304 | se->symlen = symlen; | |
305 | se->type = st_none; | |
306 | se->next = *stab; | |
307 | *stab = se; | |
308 | } | |
309 | ||
310 | return se; | |
311 | } | |
312 | ||
313 | /* | |
314 | * NAME | |
315 | * stab_type -- type of a symbol table entry | |
316 | * | |
317 | * SYNOPSIS | |
318 | * enum sym_type stab_type (Stab_entry *se); | |
319 | * | |
320 | * WARNING | |
321 | * May evaluate its argument more than once. | |
322 | */ | |
323 | ||
324 | #define stab_type(se) ((se)==NULL ? st_none : (se)->type) | |
325 | \f | |
326 | ||
327 | ||
328 | typedef int LINENO; | |
329 | ||
330 | typedef struct | |
331 | { | |
332 | char *p; | |
333 | int len; | |
c6d46f5f JB |
334 | LINENO lineno; |
335 | logical rewritten; | |
336 | } TOKEN; | |
c6d46f5f | 337 | |
13fde0cd | 338 | /* C extensions. |
c6d46f5f | 339 | */ |
13fde0cd RS |
340 | #define C_PLPL 0x00001 /* C++ */ |
341 | #define C_STAR 0x00003 /* C* */ | |
342 | #define YACC 0x10000 /* yacc file */ | |
c6d46f5f JB |
343 | |
344 | char searchar = '/'; /* use /.../ searches */ | |
345 | ||
346 | LINENO lineno; /* line number of current line */ | |
347 | long charno; /* current character number */ | |
c6d46f5f JB |
348 | |
349 | long linecharno; /* charno of start of line; not used by C, but | |
350 | * by every other language. | |
351 | */ | |
352 | ||
353 | char *curfile, /* current input file name */ | |
354 | *outfile, /* output file */ | |
355 | *white = " \f\t\n", /* white chars */ | |
13fde0cd RS |
356 | *endtk = " \t\n\"'#()[]{}=-+%*/&|^~!<>;,.:?", /* token ending chars */ |
357 | /* token starting chars */ | |
358 | *begtk = "ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz$", | |
359 | /* valid in-token chars */ | |
360 | *intk = "ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz$0123456789"; | |
c6d46f5f | 361 | |
c6d46f5f JB |
362 | int append_to_tagfile; /* -a: append to tags */ |
363 | int emacs_tags_format; /* emacs style output (no -e option any more) */ | |
364 | /* The following three default to 1 for etags, but to 0 for ctags. */ | |
365 | int typedefs; /* -t: create tags for typedefs */ | |
366 | int typedefs_and_cplusplus; /* -T: create tags for typedefs, level */ | |
367 | /* 0 struct/enum/union decls, and C++ */ | |
368 | /* member functions */ | |
369 | int constantypedefs; /* -d: create tags for C #define and enum */ | |
370 | /* constants. Default under etags. Enum */ | |
371 | /* constants not implemented. */ | |
372 | /* -D: opposite of -d. Default under ctags. */ | |
373 | int update; /* -u: update tags */ | |
374 | int vgrind_style; /* -v: create vgrind style index output */ | |
375 | int no_warnings; /* -w: suppress warnings */ | |
376 | int cxref_style; /* -x: create cxref style output */ | |
377 | int cplusplus; /* .[hc] means C++, not C */ | |
378 | int noindentypedefs; /* -S: ignore indentation in C */ | |
c6d46f5f JB |
379 | |
380 | /* Name this program was invoked with. */ | |
381 | char *progname; | |
382 | ||
4746118a JB |
383 | struct option longopts[] = { |
384 | { "append", no_argument, NULL, 'a' }, | |
385 | { "backward-search", no_argument, NULL, 'B' }, | |
386 | { "c++", no_argument, NULL, 'C' }, | |
387 | { "cxref", no_argument, NULL, 'x' }, | |
388 | { "defines", no_argument, NULL, 'd' }, | |
389 | { "forward-search", no_argument, NULL, 'F' }, | |
390 | { "help", no_argument, NULL, 'H' }, | |
391 | { "ignore-indentation", no_argument, NULL, 'S' }, | |
392 | { "include", required_argument, NULL, 'i' }, | |
393 | { "no-defines", no_argument, NULL, 'D' }, | |
394 | { "no-warn", no_argument, NULL, 'w' }, | |
395 | { "output", required_argument, NULL, 'o' }, | |
396 | { "typedefs", no_argument, NULL, 't' }, | |
397 | { "typedefs-and-c++", no_argument, NULL, 'T' }, | |
398 | { "update", no_argument, NULL, 'u' }, | |
399 | { "version", no_argument, NULL, 'V' }, | |
400 | { "vgrind", no_argument, NULL, 'v' }, | |
401 | { 0 } | |
402 | }; | |
403 | ||
c6d46f5f JB |
404 | FILE *inf, /* ioptr for current input file */ |
405 | *outf; /* ioptr for tags file */ | |
406 | ||
407 | NODE *head; /* the head of the binary tree of tags */ | |
408 | ||
409 | int permit_duplicates = 1; /* Nonzero means allow duplicate tags. */ | |
410 | ||
411 | /* A `struct linebuffer' is a structure which holds a line of text. | |
412 | `readline' reads a line from a stream into a linebuffer | |
413 | and works regardless of the length of the line. */ | |
414 | ||
415 | struct linebuffer | |
416 | { | |
417 | long size; | |
418 | char *buffer; | |
419 | }; | |
420 | ||
421 | struct linebuffer lb; /* the current line */ | |
c6d46f5f | 422 | struct linebuffer filename_lb; /* used to read in filenames */ |
13fde0cd RS |
423 | struct |
424 | { | |
425 | FILEPOS linepos; | |
426 | struct linebuffer lb; /* used by C_entries instead of lb */ | |
427 | } lbs[2]; | |
c6d46f5f | 428 | \f |
4746118a JB |
429 | void |
430 | print_version () | |
431 | { | |
432 | #ifdef CTAGS | |
433 | printf ("CTAGS "); | |
434 | #ifdef ETAGS | |
435 | printf ("and "); | |
436 | #endif | |
437 | #endif | |
438 | #ifdef ETAGS | |
439 | printf ("ETAGS "); | |
440 | #endif | |
441 | printf ("for Emacs version 19.0.\n"); | |
442 | ||
443 | exit (0); | |
444 | } | |
445 | ||
446 | void | |
447 | print_help () | |
448 | { | |
449 | printf ("These are the options accepted by %s. You may use unambiguous\n\ | |
450 | abbreviations for the long option names.\n\n", progname); | |
451 | ||
452 | fputs ("\ | |
453 | -a, --append\n\ | |
454 | Append tag entries to existing tags file.\n\ | |
455 | -C, --c++\n\ | |
456 | Treat files with `.c' and `.h' extensions as C++ code, not C\n\ | |
457 | code. Files with `.C', `.H', `.cxx', `.hxx', or `.cc'\n\ | |
458 | extensions are always assumed to be C++ code.\n\ | |
459 | -d, --defines\n\ | |
460 | Create tag entries for #defines, too.", stdout); | |
461 | ||
462 | #ifdef ETAGS | |
463 | fputs (" This is the default\n\ | |
464 | behavior.", stdout); | |
465 | #endif | |
466 | ||
467 | fputs ("\n\ | |
468 | -D, --no-defines\n\ | |
469 | Don't create tag entries for #defines.", stdout); | |
470 | ||
471 | #ifdef CTAGS | |
472 | fputs (" This is the default\n\ | |
473 | behavior.", stdout); | |
474 | #endif | |
475 | ||
476 | puts ("\n\ | |
477 | -o FILE, --output=FILE\n\ | |
478 | Write the tags to FILE.\n\ | |
479 | -S, --ignore-indentation\n\ | |
480 | Don't rely on indentation quite as much as normal. Currently,\n\ | |
481 | this means not to assume that a closing brace in the first\n\ | |
482 | column is the final brace of a function or structure\n\ | |
483 | definition.\n\ | |
484 | -t, --typedefs\n\ | |
485 | Generate tag entries for typedefs. This is the default\n\ | |
486 | behavior.\n\ | |
487 | -T, --typedefs-and-c++\n\ | |
488 | Generate tag entries for typedefs, struct/enum/union tags, and\n\ | |
489 | C++ member functions."); | |
490 | ||
491 | #ifdef ETAGS | |
492 | puts ("-i FILE, --include=FILE\n\ | |
493 | Include a note in tag file indicating that, when searching for\n\ | |
494 | a tag, one should also consult the tags file FILE after\n\ | |
495 | checking the current file."); | |
496 | #endif | |
497 | ||
498 | #ifdef CTAGS | |
499 | puts ("-B, --backward-search\n\ | |
500 | Write the search commands for the tag entries using '?', the\n\ | |
501 | backward-search command.\n\ | |
502 | -F, --forward-search\n\ | |
503 | Write the search commands for the tag entries using '/', the\n\ | |
504 | forward-search command.\n\ | |
505 | -u, --update\n\ | |
506 | Update the tag entries for the given files, leaving tag\n\ | |
507 | entries for other files in place. Currently, this is\n\ | |
508 | implemented by deleting the existing entries for the given\n\ | |
509 | files and then rewriting the new entries at the end of the\n\ | |
510 | tags file. It is often faster to simply rebuild the entire\n\ | |
511 | tag file than to use this.\n\ | |
512 | -v, --vgrind\n\ | |
513 | Generates an index of items intended for human consumption,\n\ | |
514 | similar to the output of vgrind. The index is sorted, and\n\ | |
515 | gives the page number of each item.\n\ | |
516 | -x, --cxref\n\ | |
517 | Like --vgrind, but in the style of cxref, rather than vgrind.\n\ | |
518 | The output uses line numbers instead of page numbers, but\n\ | |
519 | beyond that the differences are cosmetic; try both to see\n\ | |
520 | which you like.\n\ | |
521 | -w, --no-warn\n\ | |
522 | Suppress warning messages about entries defined in multiple\n\ | |
523 | files."); | |
524 | #endif | |
525 | ||
526 | puts ("-V, --version\n\ | |
527 | Print the version of the program.\n\ | |
528 | -H, --help\n\ | |
529 | Print this help message."); | |
530 | ||
531 | exit (0); | |
532 | } | |
533 | ||
534 | \f | |
c6d46f5f JB |
535 | void |
536 | main (argc, argv) | |
537 | int argc; | |
538 | char *argv[]; | |
539 | { | |
540 | char cmd[100]; | |
541 | int i; | |
1e134a5f | 542 | unsigned int nincluded_files = 0; |
b960e009 | 543 | char **included_files = (char **) alloca (argc * sizeof (char *)); |
c6d46f5f JB |
544 | char *this_file; |
545 | #ifdef VMS | |
546 | char got_err; | |
547 | ||
548 | extern char *gfnames (); | |
549 | extern char *massage_name (); | |
550 | #endif | |
551 | ||
552 | progname = argv[0]; | |
553 | ||
554 | #ifndef CTAGS | |
555 | emacs_tags_format = 1; | |
556 | #else | |
557 | emacs_tags_format = 0; | |
558 | #endif | |
559 | ||
560 | /* | |
561 | * If etags, always find typedefs and structure tags. Why not? | |
562 | * Also default is to find macro constants. | |
563 | */ | |
564 | if (emacs_tags_format) | |
565 | typedefs = typedefs_and_cplusplus = constantypedefs = 1; | |
566 | ||
4746118a | 567 | for (;;) |
c6d46f5f | 568 | { |
4746118a JB |
569 | int opt; |
570 | opt = getopt_long (argc, argv, "aCdDo:StTi:BFuvxwVH", longopts, 0); | |
571 | ||
572 | if (opt == EOF) | |
573 | break; | |
574 | ||
575 | switch (opt) | |
c6d46f5f | 576 | { |
4746118a JB |
577 | case '\0': |
578 | /* If getopt returns '\0', then it has already processed a | |
579 | long-named option. We should do nothing. */ | |
580 | break; | |
581 | ||
582 | /* Common options. */ | |
583 | case 'a': | |
584 | append_to_tagfile++; | |
585 | break; | |
586 | case 'C': | |
587 | cplusplus = 1; | |
588 | break; | |
589 | case 'd': | |
590 | constantypedefs = 1; | |
591 | break; | |
592 | case 'D': | |
593 | constantypedefs = 0; | |
594 | break; | |
595 | case 'o': | |
596 | if (outfile) | |
c6d46f5f | 597 | { |
4746118a JB |
598 | fprintf (stderr, |
599 | "%s: -o flag may only be given once\n", progname); | |
c6d46f5f JB |
600 | goto usage; |
601 | } | |
4746118a JB |
602 | outfile = optarg; |
603 | break; | |
604 | case 'S': | |
605 | noindentypedefs++; | |
606 | break; | |
607 | case 't': | |
608 | typedefs++; | |
609 | break; | |
610 | case 'T': | |
611 | typedefs++; | |
612 | typedefs_and_cplusplus++; | |
613 | break; | |
614 | case 'V': | |
615 | print_version (); | |
616 | break; | |
617 | case 'H': | |
618 | print_help (); | |
619 | break; | |
620 | ||
621 | /* Etags options */ | |
622 | case 'i': | |
623 | if (!emacs_tags_format) | |
624 | goto usage; | |
625 | included_files[nincluded_files++] = optarg; | |
626 | break; | |
627 | ||
628 | /* Ctags options. */ | |
629 | case 'B': | |
630 | searchar = '?'; | |
631 | if (emacs_tags_format) | |
632 | goto usage; | |
633 | break; | |
634 | case 'F': | |
635 | searchar = '/'; | |
636 | if (emacs_tags_format) | |
637 | goto usage; | |
638 | break; | |
639 | case 'u': | |
640 | update++; | |
641 | if (emacs_tags_format) | |
642 | goto usage; | |
643 | break; | |
644 | case 'v': | |
645 | vgrind_style++; | |
646 | /*FALLTHRU*/ | |
647 | case 'x': | |
648 | cxref_style++; | |
649 | if (emacs_tags_format) | |
650 | goto usage; | |
651 | break; | |
652 | case 'w': | |
653 | no_warnings++; | |
654 | if (emacs_tags_format) | |
655 | goto usage; | |
656 | break; | |
657 | ||
658 | default: | |
659 | goto usage; | |
c6d46f5f | 660 | } |
c6d46f5f JB |
661 | } |
662 | ||
0e5ad25f | 663 | if (optind == argc && nincluded_files == 0) |
c6d46f5f | 664 | { |
4746118a JB |
665 | fprintf (stderr, "%s: No input files specified.\n", progname); |
666 | ||
c6d46f5f | 667 | usage: |
0e5ad25f | 668 | fprintf (stderr, "%s: Try `%s --help' for a complete list of options.\n", |
4746118a | 669 | progname, progname); |
c6d46f5f JB |
670 | exit (BAD); |
671 | } | |
672 | ||
673 | if (outfile == 0) | |
674 | { | |
675 | outfile = emacs_tags_format ? "TAGS" : "tags"; | |
676 | } | |
677 | ||
678 | init (); /* set up boolean "functions" */ | |
679 | ||
680 | initbuffer (&lb); | |
13fde0cd RS |
681 | initbuffer (&lbs[0].lb); |
682 | initbuffer (&lbs[1].lb); | |
c6d46f5f JB |
683 | initbuffer (&filename_lb); |
684 | /* | |
685 | * loop through files finding functions | |
686 | */ | |
687 | if (emacs_tags_format) | |
688 | { | |
689 | if (streq (outfile, "-")) | |
690 | outf = stdout; | |
691 | else | |
692 | outf = fopen (outfile, append_to_tagfile ? "a" : "w"); | |
693 | if (!outf) | |
694 | { | |
695 | perror (outfile); | |
696 | exit (1); | |
697 | } | |
698 | } | |
699 | ||
c6d46f5f | 700 | #ifdef VMS |
4746118a JB |
701 | argc -= optind; |
702 | argv += optind; | |
703 | while (gfnames (&argc, &argv, &got_err) != NULL) | |
c6d46f5f JB |
704 | { |
705 | if (got_err) | |
706 | { | |
707 | error ("Can't find file %s\n", this_file); | |
708 | argc--, argv++; | |
709 | } | |
710 | else | |
711 | { | |
712 | this_file = massage_name (this_file); | |
713 | #if 0 | |
714 | } | |
13fde0cd | 715 | } /* solely to balance out the ifdef'd parens above */ |
c6d46f5f JB |
716 | #endif |
717 | #else | |
4746118a | 718 | for (; optind < argc; optind++) |
c6d46f5f | 719 | { |
4746118a | 720 | this_file = argv[optind]; |
c6d46f5f JB |
721 | if (1) |
722 | { | |
723 | #endif | |
724 | /* Input file named "-" means read file names from stdin | |
725 | and use them. */ | |
726 | if (streq (this_file, "-")) | |
727 | { | |
728 | while (!feof (stdin)) | |
729 | { | |
730 | (void) readline (&filename_lb, stdin); | |
731 | if (strlen (filename_lb.buffer) > 0) | |
732 | process_file (filename_lb.buffer); | |
733 | } | |
734 | } | |
735 | else | |
736 | process_file (this_file); | |
737 | } | |
738 | } | |
739 | ||
740 | if (emacs_tags_format) | |
741 | { | |
1e134a5f RM |
742 | while (nincluded_files-- > 0) |
743 | fprintf (outf, "\f\n%s,include\n", *included_files++); | |
744 | ||
c6d46f5f JB |
745 | (void) fclose (outf); |
746 | exit (0); | |
747 | } | |
748 | ||
749 | if (cxref_style) | |
750 | { | |
751 | put_entries (head); | |
752 | exit (GOOD); | |
753 | } | |
4746118a | 754 | if (update) |
c6d46f5f | 755 | { |
4746118a JB |
756 | /* update cannot be set under VMS, so we may assume that argc |
757 | and argv have not been munged. */ | |
758 | for (i = optind; i < argc; i++) | |
c6d46f5f JB |
759 | { |
760 | sprintf (cmd, | |
761 | "mv %s OTAGS;fgrep -v '\t%s\t' OTAGS >%s;rm OTAGS", | |
762 | outfile, argv[i], outfile); | |
763 | (void) system (cmd); | |
764 | } | |
765 | append_to_tagfile++; | |
766 | } | |
767 | outf = fopen (outfile, append_to_tagfile ? "a" : "w"); | |
768 | if (outf == NULL) | |
769 | { | |
770 | perror (outfile); | |
771 | exit (GOOD); | |
772 | } | |
773 | put_entries (head); | |
774 | (void) fclose (outf); | |
775 | if (update) | |
776 | { | |
777 | sprintf (cmd, "sort %s -o %s", outfile, outfile); | |
778 | (void) system (cmd); | |
779 | } | |
780 | exit (GOOD); | |
781 | } | |
782 | ||
783 | ||
784 | /* | |
785 | * This routine is called on each file argument. | |
786 | */ | |
787 | void | |
788 | process_file (file) | |
789 | char *file; | |
790 | { | |
791 | struct stat stat_buf; | |
792 | ||
793 | stat (file, &stat_buf); | |
8a6c8bcf RS |
794 | if (!(stat_buf.st_mode & S_IFREG) |
795 | #ifdef S_IFLNK | |
796 | || !(stat_buf.st_mode & S_IFLNK) | |
797 | #endif | |
798 | ) | |
c6d46f5f JB |
799 | { |
800 | fprintf (stderr, "Skipping %s: it is not a regular file.\n", file); | |
801 | return; | |
802 | } | |
803 | ||
804 | if (streq (file, outfile) && !streq (outfile, "-")) | |
805 | { | |
806 | fprintf (stderr, "Skipping inclusion of %s in self.\n", file); | |
807 | return; | |
808 | } | |
c6d46f5f JB |
809 | if (emacs_tags_format) |
810 | { | |
8a6c8bcf | 811 | char *cp = etags_rindex (file, '/'); |
c6d46f5f JB |
812 | if (cp) |
813 | ++cp; | |
814 | else | |
815 | cp = file; | |
c6d46f5f JB |
816 | } |
817 | find_entries (file); | |
818 | if (emacs_tags_format) | |
819 | { | |
820 | fprintf (outf, "\f\n%s,%d\n", | |
821 | file, total_size_of_entries (head)); | |
822 | put_entries (head); | |
823 | free_tree (head); | |
824 | head = NULL; | |
825 | } | |
826 | } | |
827 | ||
828 | /* | |
829 | * This routine sets up the boolean psuedo-functions which work | |
99e0a2e0 | 830 | * by setting boolean flags dependent upon the corresponding character |
c6d46f5f JB |
831 | * Every char which is NOT in that string is not a white char. Therefore, |
832 | * all of the array "_wht" is set to FALSE, and then the elements | |
833 | * subscripted by the chars in "white" are set to TRUE. Thus "_wht" | |
834 | * of a char is TRUE if it is the string "white", else FALSE. | |
835 | */ | |
836 | void | |
837 | init () | |
838 | { | |
13fde0cd RS |
839 | register char *sp; |
840 | register int i; | |
c6d46f5f JB |
841 | |
842 | for (i = 0; i < 0177; i++) | |
13fde0cd | 843 | _wht[i] = _etk[i] = _itk[i] = _btk[i] = FALSE; |
c6d46f5f JB |
844 | for (sp = white; *sp; sp++) |
845 | _wht[*sp] = TRUE; | |
846 | for (sp = endtk; *sp; sp++) | |
847 | _etk[*sp] = TRUE; | |
848 | for (sp = intk; *sp; sp++) | |
849 | _itk[*sp] = TRUE; | |
850 | for (sp = begtk; *sp; sp++) | |
851 | _btk[*sp] = TRUE; | |
c6d46f5f JB |
852 | _wht[0] = _wht['\n']; |
853 | _etk[0] = _etk['\n']; | |
854 | _btk[0] = _btk['\n']; | |
855 | _itk[0] = _itk['\n']; | |
c6d46f5f JB |
856 | } |
857 | ||
858 | /* | |
859 | * This routine opens the specified file and calls the function | |
860 | * which finds the function and type definitions. | |
861 | */ | |
862 | void | |
863 | find_entries (file) | |
864 | char *file; | |
865 | { | |
866 | char *cp; | |
867 | void prolog_funcs (); | |
868 | ||
869 | inf = fopen (file, "r"); | |
870 | if (inf == NULL) | |
871 | { | |
872 | perror (file); | |
873 | return; | |
874 | } | |
875 | curfile = savestr (file); | |
8a6c8bcf | 876 | cp = etags_rindex (file, '.'); |
c6d46f5f JB |
877 | |
878 | header_file = (cp && (streq (cp + 1, "h"))); | |
879 | ||
880 | /* .tex, .aux or .bbl implies LaTeX source code */ | |
881 | if (cp && (streq (cp + 1, "tex") || streq (cp + 1, "aux") | |
882 | || streq (cp + 1, "bbl"))) | |
883 | { | |
884 | TEX_funcs (inf); | |
885 | goto close_and_return; | |
886 | } | |
887 | /* .l or .el or .lisp (or .cl or .clisp or ...) implies lisp source code */ | |
888 | if (cp && (streq (cp + 1, "l") | |
889 | || streq (cp + 1, "el") | |
890 | || streq (cp + 1, "lsp") | |
891 | || streq (cp + 1, "lisp") | |
892 | || streq (cp + 1, "cl") | |
893 | || streq (cp + 1, "clisp"))) | |
894 | { | |
895 | L_funcs (inf); | |
896 | goto close_and_return; | |
897 | } | |
898 | /* .scm or .sm or .scheme or ... implies scheme source code */ | |
899 | if (cp && (streq (cp + 1, "sm") | |
900 | || streq (cp + 1, "scm") | |
901 | || streq (cp + 1, "scheme") | |
902 | || streq (cp + 1, "t") | |
903 | || streq (cp + 1, "sch") | |
904 | || streq (cp + 1, "SM") | |
905 | || streq (cp + 1, "SCM") | |
906 | /* The `SCM' or `scm' prefix with a version number */ | |
907 | || (cp[-1] == 'm' && cp[-2] == 'c' && cp[-3] == 's' | |
908 | && string_numeric_p (cp + 1)) | |
909 | || (cp[-1] == 'M' && cp[-2] == 'C' && cp[-3] == 'S' | |
910 | && string_numeric_p (cp + 1)))) | |
911 | { | |
912 | Scheme_funcs (inf); | |
913 | fclose (inf); | |
914 | return; | |
915 | } | |
13fde0cd RS |
916 | /* Assume that ".s" or ".a" is assembly code. -wolfgang. |
917 | Or even ".sa". */ | |
918 | if (cp && (streq (cp + 1, "s") | |
919 | || streq (cp + 1, "a") | |
920 | || streq (cp + 1, "sa"))) | |
c6d46f5f JB |
921 | { |
922 | Asm_funcs (inf); | |
923 | fclose (inf); | |
924 | return; | |
925 | } | |
926 | /* .C or .H or .cxx or .hxx or .cc: a C++ file */ | |
927 | if (cp && (streq (cp + 1, "C") | |
928 | || streq (cp + 1, "H") | |
929 | || streq (cp + 1, "cxx") | |
930 | || streq (cp + 1, "hxx") | |
931 | || streq (cp + 1, "cc"))) | |
932 | { | |
933 | C_entries (C_PLPL); /* C++ */ | |
934 | goto close_and_return; | |
935 | } | |
936 | /* .cs or .hs: a C* file */ | |
13fde0cd RS |
937 | if (cp && (streq (cp + 1, "cs") |
938 | || streq (cp + 1, "hs"))) | |
c6d46f5f JB |
939 | { |
940 | C_entries (C_STAR); | |
941 | goto close_and_return; | |
942 | } | |
13fde0cd RS |
943 | /* .y: a yacc file */ |
944 | if (cp && (streq (cp + 1, "y"))) | |
945 | { | |
946 | C_entries (YACC); | |
947 | goto close_and_return; | |
948 | } | |
c6d46f5f | 949 | /* .pl implies prolog source code */ |
13fde0cd | 950 | if (cp && streq (cp + 1, "pl")) |
c6d46f5f JB |
951 | { |
952 | prolog_funcs (inf); | |
953 | goto close_and_return; | |
954 | } | |
955 | /* .p or .pas: a Pascal file */ | |
956 | if (cp && (streq (cp + 1, "p") | |
957 | || streq (cp + 1, "pas"))) | |
958 | { | |
959 | PAS_funcs (inf); | |
960 | goto close_and_return; | |
961 | } | |
1e8e1162 | 962 | /* If .f or .for, assume it is fortran or nothing. */ |
13fde0cd RS |
963 | if (cp && (streq (cp + 1, "f") |
964 | || streq (cp + 1, "for"))) | |
1e8e1162 RS |
965 | { |
966 | PF_funcs (inf); | |
967 | goto close_and_return; | |
968 | } | |
c6d46f5f | 969 | /* if not a .c or .h or .y file, try fortran */ |
1e8e1162 RS |
970 | if (cp && ((cp[1] != 'c' |
971 | && cp[1] != 'h' | |
972 | && cp[1] != 'y') | |
973 | || (cp[1] != 0 && cp[2] != 0))) | |
c6d46f5f JB |
974 | { |
975 | if (PF_funcs (inf) != 0) | |
976 | goto close_and_return; | |
977 | rewind (inf); /* no fortran tags found, try C */ | |
978 | } | |
979 | C_entries (cplusplus ? C_PLPL : 0); | |
980 | ||
981 | close_and_return: | |
982 | (void) fclose (inf); | |
983 | } | |
984 | ||
985 | /* Nonzero if string STR is composed of digits. */ | |
986 | ||
987 | int | |
988 | string_numeric_p (str) | |
989 | char *str; | |
990 | { | |
991 | while (*str) | |
992 | { | |
993 | if (*str < '0' || *str > '9') | |
994 | return 0; | |
995 | } | |
996 | return 1; | |
997 | } | |
998 | \f | |
999 | /* Record a tag. */ | |
1000 | /* Should take a TOKEN* instead!! */ | |
c6d46f5f JB |
1001 | void |
1002 | pfnote (name, is_func, rewritten, linestart, linelen, lno, cno) | |
1003 | char *name; /* tag name */ | |
1004 | logical is_func; /* function or type name? */ | |
1005 | logical rewritten; /* tag different from text of definition? */ | |
1006 | char *linestart; | |
1007 | int linelen; | |
1008 | int lno; | |
1009 | long cno; | |
1010 | { | |
1011 | register char *fp; | |
1012 | register NODE *np; | |
1013 | char tem[51]; | |
1014 | char c; | |
1015 | ||
1016 | np = (NODE *) malloc (sizeof (NODE)); | |
1017 | if (np == NULL) | |
1018 | { | |
1019 | if (!emacs_tags_format) | |
1020 | { | |
1021 | /* It's okay to output early in etags -- it only disrupts the | |
1022 | * character count of the tag entries, which is no longer used | |
1023 | * by tags.el anyway. | |
1024 | */ | |
1025 | error ("too many entries to sort"); | |
1026 | } | |
1027 | put_entries (head); | |
1028 | free_tree (head); | |
1029 | head = NULL; | |
1030 | np = xnew (1, NODE); | |
1031 | } | |
1032 | /* If ctags mode, change name "main" to M<thisfilename>. */ | |
1033 | if (!emacs_tags_format && !cxref_style && streq (name, "main")) | |
1034 | { | |
8a6c8bcf | 1035 | fp = etags_rindex (curfile, '/'); |
c6d46f5f | 1036 | name = concat ("M", fp == 0 ? curfile : fp + 1, ""); |
8a6c8bcf | 1037 | fp = etags_rindex (name, '.'); |
c6d46f5f JB |
1038 | if (fp && fp[1] != '\0' && fp[2] == '\0') |
1039 | *fp = 0; | |
1040 | rewritten = TRUE; | |
1041 | } | |
1042 | np->name = savestr (name); | |
1043 | np->file = curfile; | |
1044 | np->is_func = is_func; | |
1045 | np->rewritten = rewritten; | |
1046 | np->lno = lno; | |
1047 | /* UNCOMMENT THE +1 HERE: */ | |
1048 | np->cno = cno /* + 1 */ ; /* our char numbers are 0-base; emacs's are 1-base */ | |
1049 | np->left = np->right = 0; | |
1050 | if (emacs_tags_format) | |
1051 | { | |
1052 | c = linestart[linelen]; | |
1053 | linestart[linelen] = 0; | |
1054 | } | |
1055 | else if (cxref_style == 0) | |
1056 | { | |
1057 | sprintf (tem, strlen (linestart) < 50 ? "%s$" : "%.50s", linestart); | |
1058 | linestart = tem; | |
1059 | } | |
1060 | np->pat = savestr (linestart); | |
1061 | if (emacs_tags_format) | |
1062 | { | |
1063 | linestart[linelen] = c; | |
1064 | } | |
1065 | ||
1066 | add_node (np, &head); | |
1067 | } | |
1068 | ||
1069 | /* | |
1070 | * free_tree () | |
1071 | * recurse on left children, iterate on right children. | |
1072 | */ | |
1073 | void | |
1074 | free_tree (node) | |
1075 | register NODE *node; | |
1076 | { | |
1077 | while (node) | |
1078 | { | |
1079 | register NODE *node_right = node->right; | |
1080 | free_tree (node->left); | |
1081 | free (node->name); | |
1082 | free (node->pat); | |
1083 | free ((char *) node); | |
1084 | node = node_right; | |
1085 | } | |
1086 | } | |
1087 | ||
1088 | /* | |
1089 | * add_node () | |
1090 | * Adds a node to the tree of nodes. In etags mode, we don't keep | |
1091 | * it sorted; we just keep a linear list. In ctags mode, maintain | |
1092 | * an ordered tree, with no attempt at balancing. | |
1093 | * | |
1094 | * add_node is the only function allowed to add nodes, so it can | |
1095 | * maintain state. | |
1096 | */ | |
1097 | void | |
1098 | add_node (node, cur_node_p) | |
1099 | NODE *node, **cur_node_p; | |
1100 | { | |
1101 | register int dif; | |
1102 | register NODE *cur_node = *cur_node_p; | |
1103 | static NODE *last_node = NULL;/* careful */ | |
1104 | ||
1105 | if (cur_node == NULL) | |
1106 | { | |
1107 | *cur_node_p = node; | |
1108 | last_node = node; | |
1109 | return; | |
1110 | } | |
1111 | ||
1112 | if (emacs_tags_format) | |
1113 | { | |
1114 | /* Etags Mode */ | |
1115 | if (!last_node) | |
1116 | fatal ("internal error in add_node"); | |
1117 | last_node->right = node; | |
1118 | last_node = node; | |
1119 | } | |
1120 | else | |
1121 | { | |
1122 | /* Ctags Mode */ | |
1123 | dif = strcmp (node->name, cur_node->name); | |
1124 | ||
1125 | /* | |
1126 | * If this tag name matches an existing one, then | |
1127 | * do not add the node, but maybe print a warning. | |
1128 | */ | |
1129 | if (!dif) | |
1130 | { | |
1131 | if (node->file == cur_node->file) | |
1132 | { | |
1133 | if (!no_warnings) | |
1134 | { | |
1135 | fprintf (stderr, "Duplicate entry in file %s, line %d: %s\n", | |
1136 | node->file, lineno, node->name); | |
1137 | fprintf (stderr, "Second entry ignored\n"); | |
1138 | } | |
1139 | return; | |
1140 | } | |
1141 | if (!cur_node->been_warned && !no_warnings) | |
1142 | { | |
1143 | fprintf (stderr, | |
1144 | "Duplicate entry in files %s and %s: %s (Warning only)\n", | |
1145 | node->file, cur_node->file, node->name); | |
1146 | } | |
1147 | cur_node->been_warned = TRUE; | |
1148 | return; | |
1149 | } | |
1150 | ||
1151 | /* Maybe refuse to add duplicate nodes. */ | |
1152 | if (!permit_duplicates) | |
1153 | { | |
1154 | if (!strcmp (node->name, cur_node->name) | |
1155 | && !strcmp (node->file, cur_node->file)) | |
1156 | return; | |
1157 | } | |
1158 | ||
1159 | /* Actually add the node */ | |
1160 | add_node (node, dif < 0 ? &cur_node->left : &cur_node->right); | |
1161 | } | |
1162 | } | |
1163 | \f | |
1164 | void | |
1165 | put_entries (node) | |
13fde0cd | 1166 | register NODE *node; |
c6d46f5f | 1167 | { |
13fde0cd | 1168 | register char *sp; |
c6d46f5f JB |
1169 | |
1170 | if (node == NULL) | |
1171 | return; | |
1172 | ||
1173 | /* Output subentries that precede this one */ | |
1174 | put_entries (node->left); | |
1175 | ||
1176 | /* Output this entry */ | |
1177 | ||
1178 | if (emacs_tags_format) | |
1179 | { | |
1180 | if (node->rewritten) | |
1181 | { | |
1182 | fprintf (outf, "%s\177%s\001%d,%d\n", | |
1183 | node->name, node->pat, node->lno, node->cno); | |
1184 | } | |
1185 | else | |
1186 | { | |
1187 | fprintf (outf, "%s\177%d,%d\n", | |
1188 | node->pat, node->lno, node->cno); | |
1189 | } | |
1190 | } | |
1191 | else if (!cxref_style) | |
1192 | { | |
1193 | fprintf (outf, "%s\t%s\t", | |
1194 | node->name, node->file); | |
1195 | ||
1196 | if (node->is_func) | |
1197 | { /* a function */ | |
1198 | putc (searchar, outf); | |
1199 | putc ('^', outf); | |
1200 | ||
1201 | for (sp = node->pat; *sp; sp++) | |
1202 | { | |
1203 | if (*sp == '\\' || *sp == searchar) | |
1204 | putc ('\\', outf); | |
1205 | putc (*sp, outf); | |
1206 | } | |
1207 | putc (searchar, outf); | |
1208 | } | |
1209 | else | |
1210 | { /* a typedef; text pattern inadequate */ | |
1211 | fprintf (outf, "%d", node->lno); | |
1212 | } | |
1213 | putc ('\n', outf); | |
1214 | } | |
1215 | else if (vgrind_style) | |
1216 | fprintf (stdout, "%s %s %d\n", | |
1217 | node->name, node->file, (node->lno + 63) / 64); | |
1218 | else | |
daa37602 | 1219 | fprintf (stdout, "%-16s %3d %-16s %s\n", |
c6d46f5f JB |
1220 | node->name, node->lno, node->file, node->pat); |
1221 | ||
1222 | /* Output subentries that follow this one */ | |
1223 | put_entries (node->right); | |
1224 | } | |
1225 | ||
1226 | /* Length of a number's decimal representation. */ | |
1227 | int | |
1228 | number_len (num) | |
1229 | long num; | |
1230 | { | |
1231 | int len = 0; | |
1232 | if (!num) | |
1233 | return 1; | |
1234 | for (; num; num /= 10) | |
1235 | ++len; | |
1236 | return len; | |
1237 | } | |
1238 | ||
1239 | /* | |
1240 | * Return total number of characters that put_entries will output for | |
1241 | * the nodes in the subtree of the specified node. Works only if emacs_tags_format | |
1242 | * is set, but called only in that case. This count is irrelevant with | |
1243 | * the new tags.el, but is still supplied for backward compatibility. | |
1244 | */ | |
1245 | int | |
1246 | total_size_of_entries (node) | |
13fde0cd | 1247 | register NODE *node; |
c6d46f5f | 1248 | { |
13fde0cd | 1249 | register int total; |
c6d46f5f JB |
1250 | |
1251 | if (node == NULL) | |
1252 | return 0; | |
1253 | ||
1254 | total = 0; | |
1255 | for (; node; node = node->right) | |
1256 | { | |
1257 | /* Count left subentries. */ | |
1258 | total += total_size_of_entries (node->left); | |
1259 | ||
1260 | /* Count this entry */ | |
1261 | total += strlen (node->pat) + 1; | |
1262 | total += number_len ((long) node->lno) + 1 + number_len (node->cno) + 1; | |
1263 | if (node->rewritten) | |
1264 | total += 1 + strlen (node->name); /* \001name */ | |
1265 | } | |
1266 | ||
1267 | return total; | |
1268 | } | |
1269 | \f | |
1270 | /* | |
1271 | * The C symbol tables. | |
1272 | */ | |
1273 | ||
1274 | Stab *C_stab, *C_PLPL_stab, *C_STAR_stab; | |
1275 | ||
1276 | /* | |
1277 | * SYNOPSIS | |
1278 | * Stab *get_C_stab (int c_ext); | |
1279 | */ | |
13fde0cd RS |
1280 | #define get_C_stab(c_ext) ((c_ext & C_STAR) ? C_STAR_stab : \ |
1281 | (c_ext & C_PLPL) ? C_PLPL_stab : \ | |
c6d46f5f JB |
1282 | C_stab) |
1283 | ||
1284 | void | |
1285 | add_keyword (stab, sym, type) | |
1286 | Stab *stab; | |
1287 | char *sym; | |
1288 | enum sym_type type; | |
1289 | { | |
1290 | stab_search (stab, sym, strlen (sym))->type = type; | |
1291 | } | |
1292 | ||
1293 | Stab * | |
1294 | C_create_stab (c_ext) | |
1295 | int c_ext; | |
1296 | { | |
1297 | Stab *stab; | |
1298 | ||
1299 | stab = stab_create (); | |
1300 | ||
1301 | /* C, C++ and C* */ | |
1302 | if (c_ext & C_PLPL) | |
1303 | add_keyword (stab, "class", st_C_struct); | |
1304 | if (c_ext & C_STAR) | |
1305 | add_keyword (stab, "domain", st_C_struct); | |
1306 | add_keyword (stab, "union", st_C_struct); | |
1307 | add_keyword (stab, "struct", st_C_struct); | |
1308 | add_keyword (stab, "enum", st_C_enum); | |
1309 | add_keyword (stab, "typedef", st_C_typedef); | |
1310 | add_keyword (stab, "define", st_C_define); | |
1311 | add_keyword (stab, "long", st_C_typespec); | |
1312 | add_keyword (stab, "short", st_C_typespec); | |
1313 | add_keyword (stab, "int", st_C_typespec); | |
1314 | add_keyword (stab, "char", st_C_typespec); | |
1315 | add_keyword (stab, "float", st_C_typespec); | |
1316 | add_keyword (stab, "double", st_C_typespec); | |
1317 | add_keyword (stab, "signed", st_C_typespec); | |
1318 | add_keyword (stab, "unsigned", st_C_typespec); | |
13fde0cd RS |
1319 | add_keyword (stab, "auto", st_C_typespec); |
1320 | add_keyword (stab, "void", st_C_typespec); | |
1321 | add_keyword (stab, "extern", st_C_typespec); | |
1322 | add_keyword (stab, "static", st_C_typespec); | |
c6d46f5f JB |
1323 | add_keyword (stab, "const", st_C_typespec); |
1324 | add_keyword (stab, "volatile", st_C_typespec); | |
1325 | ||
1326 | return stab; | |
1327 | } | |
1328 | ||
1329 | void | |
1330 | C_create_stabs () | |
1331 | { | |
1332 | C_stab = C_create_stab (0); | |
1333 | C_PLPL_stab = C_create_stab (C_PLPL); | |
1334 | C_STAR_stab = C_create_stab (C_STAR | C_PLPL); | |
1335 | } | |
1336 | \f | |
13fde0cd RS |
1337 | /* |
1338 | * etags.c 4.2 1993/03/22 12:13:40 pot Exp | |
1339 | * C functions are recognized using a simple finite automaton. | |
1340 | * funcdef is its state variable. | |
1341 | */ | |
1342 | typedef enum | |
1343 | { | |
1344 | fnone, ftagseen, finlist, flistseen | |
1345 | } FUNCST; | |
1346 | FUNCST funcdef; | |
1347 | ||
1348 | ||
1349 | /* typedefs are recognized using a simple finite automaton. | |
1350 | * typeddef is its state variable. | |
1351 | */ | |
1352 | typedef enum | |
1353 | { | |
1354 | tnone, ttypedseen, tinbody, tend | |
1355 | } TYPEDST; | |
1356 | TYPEDST typdef; | |
1357 | ||
1358 | ||
1359 | /* struct tags for C++ are recognized using another simple | |
1360 | * finite automaton. `structdef' is its state variable. | |
1361 | * This machinery is only invoked for C++; otherwise structdef | |
1362 | * should remain snone. However, this machinery can easily be | |
1363 | * adapted to find structure tags in normal C code. | |
1364 | */ | |
1365 | typedef enum | |
1366 | { | |
1367 | snone, /* nothing seen yet */ | |
1368 | skeyseen, /* struct-like keyword seen */ | |
1369 | stagseen, /* struct-like tag seen */ | |
1370 | scolonseen, /* colon seen after struct-like tag */ | |
1371 | sinbody /* in class body: recognize member func defs */ | |
1372 | } STRUCTST; | |
1373 | STRUCTST structdef; | |
1374 | /* | |
1375 | * When structdef is stagseen, scolonseen, or sinbody, structtag is the | |
1376 | * struct tag, and structkey is the preceding struct-like keyword. | |
1377 | */ | |
1378 | char structtag[BUFSIZ]; | |
1379 | Stab_entry *structkey; | |
1380 | ||
1381 | /* | |
1382 | * Yet another little state machine to deal with preprocessor lines. | |
1383 | */ | |
1384 | typedef enum | |
1385 | { | |
1386 | dnone, /* nothing seen */ | |
1387 | dsharpseen, /* '#' seen as first char on line */ | |
1388 | ddefineseen, /* '#' and 'define' seen */ | |
1389 | dignorerest /* ignore rest of line */ | |
1390 | } DEFINEST; | |
1391 | DEFINEST definedef; | |
1392 | ||
1393 | /* | |
1394 | * Set this to TRUE, and the next token considered is called a function. | |
1395 | * Used only for GNUmacs's function-defining macros. | |
1396 | */ | |
1397 | logical next_token_is_func; | |
1398 | ||
1399 | /* | |
1400 | * TRUE in the rules part of a yacc file, FALSE outside (parse as C). | |
1401 | */ | |
1402 | logical yacc_rules; | |
1403 | ||
c6d46f5f JB |
1404 | /* |
1405 | * C_entries () | |
13fde0cd RS |
1406 | * This routine finds functions, typedefs, #define's and |
1407 | * struct/union/enum definitions in C syntax and adds them | |
c6d46f5f JB |
1408 | * to the list. |
1409 | */ | |
1410 | ||
13fde0cd RS |
1411 | /* |
1412 | * LEVEL_OK_FOR_FUNCDEF allows C++ function definition within class body. | |
1413 | * Currently typdef and structdef stuff (typedefs and struct | |
1414 | * definitions) are only noticed when level==0, but that may change. | |
1415 | */ | |
1416 | #define LEVEL_OK_FOR_FUNCDEF() \ | |
1417 | (level==0 || (cplpl && level==1 && structdef==sinbody)) | |
1418 | ||
1419 | #define curlb (lbs[curndx].lb) | |
1420 | #define othlb (lbs[1-curndx].lb) | |
1421 | #define newlb (lbs[newndx].lb) | |
1422 | #define curlinepos (lbs[curndx].linepos) | |
1423 | #define othlinepos (lbs[1-curndx].linepos) | |
1424 | #define newlinepos (lbs[newndx].linepos) | |
1425 | ||
c6d46f5f | 1426 | #define CNL_SAVE_DEFINEDEF \ |
13fde0cd RS |
1427 | do { \ |
1428 | SET_FILEPOS (curlinepos, inf, charno); \ | |
c6d46f5f | 1429 | lineno++; \ |
13fde0cd RS |
1430 | charno += readline (&curlb, inf); \ |
1431 | lp = curlb.buffer; \ | |
1432 | quotednl = FALSE; \ | |
1433 | newndx = curndx; \ | |
1434 | } while (FALSE) | |
c6d46f5f JB |
1435 | |
1436 | #define CNL \ | |
13fde0cd | 1437 | do { \ |
c6d46f5f JB |
1438 | CNL_SAVE_DEFINEDEF; \ |
1439 | definedef = dnone; \ | |
13fde0cd RS |
1440 | } while (FALSE) |
1441 | ||
1442 | #define MAKE_TAG_FROM_NEW_LB(isfun) pfnote (tokb, isfun, tok.rewritten, \ | |
1443 | newlb.buffer, tokoff + toklen + 1, tok.lineno, GET_CHARNO (newlinepos)) | |
1444 | #define MAKE_TAG_FROM_OTH_LB(isfun) pfnote (tokb, isfun, tok.rewritten, \ | |
1445 | othlb.buffer, tokoff + toklen + 1, tok.lineno, GET_CHARNO (othlinepos)) | |
c6d46f5f JB |
1446 | |
1447 | void | |
1448 | C_entries (c_ext) | |
1449 | int c_ext; /* extension of C? */ | |
1450 | { | |
13fde0cd | 1451 | register char c; /* latest char read; '\0' for end of line */ |
c6d46f5f | 1452 | register char *lp; /* pointer one beyond the character `c' */ |
13fde0cd RS |
1453 | int curndx, newndx; /* indices for current and new lb */ |
1454 | TOKEN tok; /* latest token read for funcdef & structdef */ | |
1455 | char tokb[BUFSIZ]; /* latest token name for funcdef & structdef */ | |
1456 | register int tokoff; /* offset in line of start of latest token */ | |
1457 | register int toklen; /* length of latest token */ | |
c6d46f5f | 1458 | int level; /* current curly brace level */ |
13fde0cd RS |
1459 | logical incomm, inquote, inchar, quotednl, midtoken; |
1460 | logical cplpl; | |
c6d46f5f | 1461 | |
13fde0cd | 1462 | curndx = newndx = 0; |
c6d46f5f JB |
1463 | lineno = 0; |
1464 | charno = 0; | |
13fde0cd | 1465 | lp = curlb.buffer; |
c6d46f5f JB |
1466 | *lp = 0; |
1467 | ||
13fde0cd RS |
1468 | definedef = dnone; funcdef = fnone; typdef= tnone; structdef= snone; |
1469 | next_token_is_func = yacc_rules = FALSE; | |
1470 | midtoken = inquote = inchar = incomm = quotednl = FALSE; | |
c6d46f5f | 1471 | level = 0; |
13fde0cd | 1472 | cplpl = c_ext & C_PLPL; |
c6d46f5f JB |
1473 | |
1474 | C_create_stabs (); | |
1475 | ||
1476 | while (!feof (inf)) | |
1477 | { | |
1478 | c = *lp++; | |
c6d46f5f JB |
1479 | if (c == '\\') |
1480 | { | |
4746118a JB |
1481 | /* If we're at the end of the line, the next character is a |
1482 | '\0'; don't skip it, because it's the thing that tells us | |
1483 | to read the next line. */ | |
13fde0cd | 1484 | if (*lp == '\0') |
99e0a2e0 | 1485 | { |
13fde0cd | 1486 | quotednl = TRUE; |
99e0a2e0 RS |
1487 | continue; |
1488 | } | |
1e134a5f | 1489 | lp++; |
c6d46f5f JB |
1490 | c = ' '; |
1491 | } | |
1492 | else if (incomm) | |
1493 | { | |
13fde0cd | 1494 | switch (c) |
c6d46f5f | 1495 | { |
13fde0cd RS |
1496 | case '*': |
1497 | if (*lp == '/') | |
1498 | { | |
1499 | c = *lp++; | |
1500 | incomm = FALSE; | |
1501 | } | |
1502 | break; | |
1503 | case '\0': | |
1504 | /* Newlines inside comments do not end macro definitions in | |
1505 | traditional cpp. */ | |
1506 | CNL_SAVE_DEFINEDEF; | |
1507 | break; | |
c6d46f5f | 1508 | } |
13fde0cd | 1509 | continue; |
c6d46f5f JB |
1510 | } |
1511 | else if (inquote) | |
1512 | { | |
13fde0cd RS |
1513 | switch (c) |
1514 | { | |
1515 | case '"': | |
1516 | inquote = FALSE; | |
1517 | break; | |
1518 | case '\0': | |
1519 | /* Newlines inside strings, do not end macro definitions | |
1520 | in traditional cpp, even though compilers don't | |
1521 | usually accept them. */ | |
1522 | CNL_SAVE_DEFINEDEF; | |
1523 | break; | |
1524 | } | |
1525 | continue; | |
c6d46f5f JB |
1526 | } |
1527 | else if (inchar) | |
1528 | { | |
1529 | if (c == '\'') | |
1530 | inchar = FALSE; | |
1531 | continue; | |
1532 | } | |
13fde0cd | 1533 | else |
c6d46f5f JB |
1534 | switch (c) |
1535 | { | |
1536 | case '"': | |
1537 | inquote = TRUE; | |
1538 | continue; | |
1539 | case '\'': | |
1540 | inchar = TRUE; | |
1541 | continue; | |
1542 | case '/': | |
1543 | if (*lp == '*') | |
1544 | { | |
1545 | lp++; | |
1546 | incomm = TRUE; | |
13fde0cd | 1547 | continue; |
c6d46f5f | 1548 | } |
13fde0cd | 1549 | else if (cplpl && *lp == '/') |
c6d46f5f | 1550 | { |
daa37602 JB |
1551 | c = 0; |
1552 | break; | |
c6d46f5f JB |
1553 | } |
1554 | continue; | |
13fde0cd RS |
1555 | case '%': |
1556 | if ((c_ext & YACC) && *lp == '%') | |
1557 | { | |
1558 | /* entering or exiting rules section in yacc file */ | |
1559 | lp++; | |
1560 | definedef = dnone; funcdef = fnone; | |
1561 | typdef= tnone; structdef= snone; | |
1562 | next_token_is_func = FALSE; | |
1563 | midtoken = inquote = inchar = incomm = quotednl = FALSE; | |
1564 | level = 0; | |
1565 | yacc_rules = !yacc_rules; | |
1566 | continue; | |
1567 | } | |
c6d46f5f | 1568 | case '#': |
13fde0cd | 1569 | if (lp == newlb.buffer + 1 && definedef == dnone) |
c6d46f5f JB |
1570 | definedef = dsharpseen; |
1571 | continue; | |
13fde0cd | 1572 | } /* switch (c) */ |
c6d46f5f | 1573 | |
c6d46f5f | 1574 | |
13fde0cd RS |
1575 | if (LEVEL_OK_FOR_FUNCDEF () |
1576 | && definedef != dignorerest | |
1577 | && structdef != scolonseen | |
1578 | && funcdef != finlist) | |
c6d46f5f JB |
1579 | { |
1580 | if (midtoken) | |
1581 | { | |
1582 | if (endtoken (c)) | |
1583 | { | |
13fde0cd | 1584 | if (cplpl && c == ':' && *lp == ':' && intoken (*(lp + 1))) |
c6d46f5f JB |
1585 | { |
1586 | /* | |
1587 | * This handles :: in the middle, but not at beginning | |
1588 | * of an identifier. | |
1589 | */ | |
1590 | lp += 2; | |
1591 | toklen += 3; | |
1592 | } | |
1593 | else | |
1594 | { | |
c6d46f5f | 1595 | logical is_func; |
c6d46f5f | 1596 | |
13fde0cd RS |
1597 | tok.lineno = lineno; |
1598 | tok.p = newlb.buffer + tokoff; | |
c6d46f5f JB |
1599 | tok.len = toklen; |
1600 | tok.rewritten = FALSE; | |
13fde0cd RS |
1601 | if (yacc_rules |
1602 | || consider_token (c, lp, &tok, | |
1603 | c_ext, level, &is_func)) | |
c6d46f5f | 1604 | { |
99e0a2e0 RS |
1605 | if (structdef == sinbody |
1606 | && definedef == dnone && is_func) | |
c6d46f5f JB |
1607 | { /* function defined in C++ class body */ |
1608 | sprintf (tokb, "%s::%.*s", | |
13fde0cd RS |
1609 | ((structtag[0] == '\0') |
1610 | ? "_anonymous_" : structtag), | |
c6d46f5f JB |
1611 | tok.len, tok.p); |
1612 | tok.rewritten = TRUE; | |
1613 | } | |
1614 | else | |
1615 | { | |
1616 | sprintf (tokb, "%.*s", tok.len, tok.p); | |
1617 | } | |
13fde0cd RS |
1618 | |
1619 | if (funcdef == ftagseen || structdef == stagseen) | |
1620 | { | |
1621 | if (newndx == curndx) | |
1622 | curndx = 1 - curndx; /* switch line buffers */ | |
1623 | } | |
1624 | else | |
1625 | MAKE_TAG_FROM_NEW_LB (is_func); | |
c6d46f5f JB |
1626 | } |
1627 | midtoken = FALSE; | |
1628 | } | |
13fde0cd | 1629 | } /* if (endtoken (c)) */ |
c6d46f5f | 1630 | else if (intoken (c)) |
13fde0cd RS |
1631 | { |
1632 | toklen++; | |
1633 | continue; | |
1634 | } | |
1635 | } /* if (midtoken) */ | |
c6d46f5f JB |
1636 | else if (begtoken (c)) |
1637 | { | |
13fde0cd RS |
1638 | switch (funcdef) |
1639 | { | |
1640 | case flistseen: | |
1641 | MAKE_TAG_FROM_OTH_LB (TRUE); | |
1642 | /* FALLTHRU */ | |
1643 | case ftagseen: | |
1644 | funcdef = fnone; | |
1645 | break; | |
1646 | } | |
1647 | if (structdef == stagseen) | |
1648 | structdef = snone; | |
1649 | if (!yacc_rules || lp == newlb.buffer + 1) | |
1650 | { | |
1651 | tokoff = lp - 1 - newlb.buffer; | |
1652 | toklen = 1; | |
1653 | midtoken = TRUE; | |
1654 | } | |
1655 | continue; | |
c6d46f5f | 1656 | } |
13fde0cd RS |
1657 | } /* if must look at token */ |
1658 | ||
1659 | ||
1660 | /* Detect end of line, colon, comma, semicolon and various braces | |
1661 | after having handled the last token on the line.*/ | |
1662 | switch (c) | |
1e134a5f | 1663 | { |
13fde0cd RS |
1664 | case ':': |
1665 | if (structdef == stagseen) | |
1666 | structdef = scolonseen; | |
1667 | else if (yacc_rules && funcdef == ftagseen) | |
57e83cfe | 1668 | { |
13fde0cd RS |
1669 | MAKE_TAG_FROM_OTH_LB (FALSE); |
1670 | funcdef == fnone; | |
57e83cfe | 1671 | } |
13fde0cd RS |
1672 | break; |
1673 | case ';': | |
1674 | funcdef = fnone; | |
1675 | /* FALLTHRU */ | |
1676 | case ',': | |
1677 | if (funcdef != finlist) | |
1678 | funcdef = fnone; | |
1679 | if (level == 0 && typdef == tend) | |
1680 | typdef = tnone; | |
1681 | /* FALLTHRU */ | |
1682 | case '[': | |
1683 | if (funcdef != finlist) | |
1684 | funcdef = fnone; | |
1685 | if (structdef == stagseen) | |
1686 | structdef = snone; | |
1687 | break; | |
1688 | case '(': | |
1689 | switch (funcdef) | |
57e83cfe | 1690 | { |
13fde0cd RS |
1691 | case ftagseen: |
1692 | funcdef = finlist; | |
1693 | break; | |
1694 | case finlist: | |
1695 | case flistseen: | |
1696 | funcdef = fnone; | |
1697 | break; | |
57e83cfe | 1698 | } |
13fde0cd RS |
1699 | break; |
1700 | case ')': | |
1701 | if (funcdef == finlist) | |
1702 | funcdef = flistseen; | |
1703 | break; | |
1704 | case '{': | |
1705 | if (typdef == ttypedseen) | |
1706 | typdef = tinbody; | |
1707 | switch (structdef) | |
1708 | { | |
1709 | case skeyseen: /* unnamed struct */ | |
1710 | structtag[0] = '\0'; | |
1711 | structdef = sinbody; | |
1712 | break; | |
1713 | case stagseen: | |
1714 | case scolonseen: /* named struct */ | |
1715 | structdef = sinbody; | |
1716 | MAKE_TAG_FROM_OTH_LB (FALSE); | |
1717 | break; | |
1718 | } | |
1719 | level++; | |
1720 | /* FALLTHRU */ | |
1721 | case '*': | |
1722 | if (funcdef == flistseen) | |
1723 | { | |
1724 | MAKE_TAG_FROM_OTH_LB (TRUE); | |
1725 | funcdef = fnone; | |
1726 | } | |
1727 | break; | |
1728 | case '}': | |
1729 | if (!noindentypedefs && lp == newlb.buffer + 1) | |
1730 | level = 0; /* reset level if first column */ | |
1731 | else if (level > 0) | |
1732 | level--; | |
1733 | if (level == 0) | |
1734 | { | |
1735 | if (typdef == tinbody) | |
1736 | typdef = tend; | |
1737 | structdef = snone; | |
1738 | (void) strcpy (structtag, "<error 2>"); | |
1739 | } | |
1740 | break; | |
1741 | case '\0': | |
1742 | /* If a macro spans multiple lines don't reset its state. */ | |
1743 | if (quotednl) | |
1744 | CNL_SAVE_DEFINEDEF; | |
1745 | else | |
1746 | CNL; | |
1747 | break; | |
1748 | } /* switch (c) */ | |
1749 | ||
1750 | } /* while not eof */ | |
c6d46f5f JB |
1751 | } |
1752 | ||
1753 | /* | |
1754 | * consider_token () | |
1755 | * checks to see if the current token is at the start of a | |
13fde0cd RS |
1756 | * function, or corresponds to a typedef, or is a struct/union/enum |
1757 | * tag. | |
c6d46f5f | 1758 | * |
13fde0cd | 1759 | * *IS_FUNC gets TRUE iff the token is a function or macro with args. |
c6d46f5f JB |
1760 | * C_EXT is which language we are looking at. |
1761 | * | |
1762 | * In the future we will need some way to adjust where the end of | |
1763 | * the token is; for instance, implementing the C++ keyword | |
1764 | * `operator' properly will adjust the end of the token to be after | |
1765 | * whatever follows `operator'. | |
1766 | * | |
1767 | * Globals | |
13fde0cd RS |
1768 | * funcdef IN OUT |
1769 | * structdef IN OUT | |
1770 | * definedef IN OUT | |
1771 | * typdef IN OUT | |
1772 | * next_token_is_func IN OUT | |
c6d46f5f JB |
1773 | */ |
1774 | ||
1775 | logical | |
13fde0cd RS |
1776 | consider_token (c, lp, tokp, c_ext, level, is_func) |
1777 | register char c; /* IN: first char after the token */ | |
1778 | register char *lp; /* IN: lp points to 2nd char after the token */ | |
1779 | register TOKEN *tokp; /* IN */ | |
c6d46f5f JB |
1780 | int c_ext; /* IN */ |
1781 | int level; /* IN */ | |
13fde0cd | 1782 | logical *is_func; /* OUT */ |
c6d46f5f | 1783 | { |
c6d46f5f JB |
1784 | logical firsttok; /* TRUE if have seen first token in ()'s */ |
1785 | Stab_entry *tokse = stab_find (get_C_stab (c_ext), tokp->p, tokp->len); | |
1786 | enum sym_type toktype = stab_type (tokse); | |
1787 | ||
13fde0cd | 1788 | *is_func = FALSE; /* not a function */ |
c6d46f5f JB |
1789 | |
1790 | /* | |
13fde0cd | 1791 | * Advance the definedef state machine. |
c6d46f5f JB |
1792 | */ |
1793 | switch (definedef) | |
1794 | { | |
1795 | case dnone: | |
1796 | /* We're not on a preprocessor line. */ | |
1797 | break; | |
1798 | case dsharpseen: | |
1799 | if (toktype == st_C_define) | |
1800 | { | |
1801 | definedef = ddefineseen; | |
c6d46f5f JB |
1802 | } |
1803 | else | |
1804 | { | |
1805 | definedef = dignorerest; | |
c6d46f5f | 1806 | } |
13fde0cd | 1807 | return (FALSE); |
c6d46f5f JB |
1808 | case ddefineseen: |
1809 | /* | |
1810 | * Make a tag for any macro. | |
c6d46f5f | 1811 | */ |
c6d46f5f | 1812 | definedef = dignorerest; |
13fde0cd | 1813 | *is_func = (c == '('); |
c6d46f5f | 1814 | if (!*is_func && !constantypedefs) |
13fde0cd RS |
1815 | return (FALSE); |
1816 | else | |
1817 | return (TRUE); | |
c6d46f5f | 1818 | case dignorerest: |
13fde0cd | 1819 | return (FALSE); |
c6d46f5f JB |
1820 | default: |
1821 | error ("internal error: definedef value"); | |
1822 | } | |
1823 | ||
1824 | /* | |
13fde0cd | 1825 | * Now typedefs |
c6d46f5f | 1826 | */ |
13fde0cd | 1827 | switch (typdef) |
c6d46f5f | 1828 | { |
13fde0cd RS |
1829 | case tnone: |
1830 | if (toktype == st_C_typedef) | |
c6d46f5f | 1831 | { |
13fde0cd RS |
1832 | if (typedefs) |
1833 | typdef = ttypedseen; | |
1834 | return (FALSE); | |
c6d46f5f | 1835 | } |
13fde0cd RS |
1836 | break; |
1837 | case ttypedseen: | |
1838 | switch (toktype) | |
c6d46f5f | 1839 | { |
13fde0cd RS |
1840 | case st_none: |
1841 | case st_C_typespec: | |
1842 | typdef = tend; | |
1843 | break; | |
1844 | case st_C_struct: | |
1845 | case st_C_enum: | |
1e134a5f | 1846 | break; |
c6d46f5f | 1847 | } |
13fde0cd | 1848 | /* Do not return here, so the structdef stuff has a chance. */ |
c6d46f5f | 1849 | break; |
13fde0cd RS |
1850 | case tend: |
1851 | switch (toktype) | |
c6d46f5f | 1852 | { |
13fde0cd RS |
1853 | case st_C_typespec: |
1854 | case st_C_struct: | |
1855 | case st_C_enum: | |
1856 | return (FALSE); | |
c6d46f5f | 1857 | } |
13fde0cd | 1858 | return (TRUE); |
c6d46f5f JB |
1859 | } |
1860 | ||
1861 | /* | |
1862 | * This structdef business is currently only invoked when level==0. | |
1863 | * It should be recursively invoked whatever the level, and a stack of | |
1864 | * states kept, to allow for definitions of structs within structs. | |
1865 | * | |
1866 | * This structdef business is NOT invoked when we are ctags and the | |
1867 | * file is plain C. This is because a struct tag may have the same | |
1868 | * name as another tag, and this loses with ctags. | |
1869 | * | |
13fde0cd RS |
1870 | * This if statement deals with the typdef state machine as |
1871 | * follows: if typdef==ttypedseen and token is struct/union/class/enum, | |
1872 | * return (FALSE). All the other code here is for the structdef | |
1873 | * state machine. | |
c6d46f5f JB |
1874 | */ |
1875 | switch (toktype) | |
1876 | { | |
1877 | case st_C_struct: | |
1878 | case st_C_enum: | |
13fde0cd | 1879 | if (typdef == ttypedseen |
99e0a2e0 | 1880 | || (typedefs_and_cplusplus && level == 0 && structdef == snone)) |
c6d46f5f JB |
1881 | { |
1882 | structdef = skeyseen; | |
1883 | structkey = tokse; | |
1884 | } | |
13fde0cd | 1885 | return (FALSE); |
c6d46f5f | 1886 | } |
c6d46f5f JB |
1887 | if (structdef == skeyseen) |
1888 | { | |
13fde0cd | 1889 | if (stab_type (structkey) == st_C_struct) |
c6d46f5f | 1890 | { |
13fde0cd RS |
1891 | (void) strncpy (structtag, tokp->p, tokp->len); |
1892 | structtag[tokp->len] = '\0'; /* for struct/union/class */ | |
c6d46f5f JB |
1893 | } |
1894 | else | |
1895 | { | |
13fde0cd | 1896 | structtag[0] = '\0'; /* for enum (why is it treated differently?) */ |
c6d46f5f | 1897 | } |
13fde0cd RS |
1898 | structdef = stagseen; |
1899 | return (TRUE); | |
c6d46f5f | 1900 | } |
13fde0cd RS |
1901 | |
1902 | /* Avoid entering funcdef stuff if typdef is going on. */ | |
1903 | if (typdef != tnone) | |
c6d46f5f | 1904 | { |
13fde0cd RS |
1905 | definedef = dnone; |
1906 | return (FALSE); | |
c6d46f5f | 1907 | } |
13fde0cd | 1908 | |
c6d46f5f | 1909 | /* Detect GNUmacs's function-defining macros. */ |
4746118a | 1910 | if (definedef == dnone) |
c6d46f5f | 1911 | { |
daa37602 JB |
1912 | if (strneq (tokp->p, "DEF", 3) |
1913 | || strneq (tokp->p, "ENTRY", 5) | |
1914 | || strneq (tokp->p, "SYSCALL", 7) | |
1915 | || strneq (tokp->p, "PSEUDO", 6)) | |
4746118a JB |
1916 | { |
1917 | next_token_is_func = TRUE; | |
13fde0cd | 1918 | return (FALSE); |
4746118a | 1919 | } |
13fde0cd | 1920 | if (strneq (tokp->p, "EXFUN", 5)) |
4746118a JB |
1921 | { |
1922 | next_token_is_func = FALSE; | |
13fde0cd | 1923 | return (FALSE); |
4746118a | 1924 | } |
c6d46f5f JB |
1925 | } |
1926 | if (next_token_is_func) | |
1927 | { | |
1928 | next_token_is_func = FALSE; | |
13fde0cd | 1929 | return (TRUE); |
c6d46f5f | 1930 | } |
13fde0cd RS |
1931 | |
1932 | /* A function? */ | |
1933 | switch (toktype) | |
c6d46f5f | 1934 | { |
13fde0cd RS |
1935 | case st_C_typespec: |
1936 | funcdef == fnone; /* should be useless */ | |
1937 | return (FALSE); | |
1938 | default: | |
1939 | funcdef = ftagseen; | |
1940 | *is_func = TRUE; | |
1941 | return (TRUE); | |
c6d46f5f | 1942 | } |
c6d46f5f JB |
1943 | } |
1944 | \f | |
1945 | /* Fortran parsing */ | |
1946 | ||
1947 | char *dbp; | |
1948 | int pfcnt; | |
1949 | ||
1950 | int | |
1951 | PF_funcs (fi) | |
1952 | FILE *fi; | |
1953 | { | |
1954 | lineno = 0; | |
1955 | charno = 0; | |
1956 | pfcnt = 0; | |
1957 | ||
1958 | while (!feof (fi)) | |
1959 | { | |
1960 | lineno++; | |
1961 | linecharno = charno; | |
1962 | charno += readline (&lb, fi); | |
1963 | dbp = lb.buffer; | |
1964 | if (*dbp == '%') | |
1965 | dbp++; /* Ratfor escape to fortran */ | |
1966 | while (isspace (*dbp)) | |
1967 | dbp++; | |
1968 | if (*dbp == 0) | |
1969 | continue; | |
1970 | switch (*dbp | ' ') | |
1971 | { | |
1972 | case 'i': | |
1973 | if (tail ("integer")) | |
1974 | takeprec (); | |
1975 | break; | |
1976 | case 'r': | |
1977 | if (tail ("real")) | |
1978 | takeprec (); | |
1979 | break; | |
1980 | case 'l': | |
1981 | if (tail ("logical")) | |
1982 | takeprec (); | |
1983 | break; | |
1984 | case 'c': | |
1985 | if (tail ("complex") || tail ("character")) | |
1986 | takeprec (); | |
1987 | break; | |
1988 | case 'd': | |
1989 | if (tail ("double")) | |
1990 | { | |
1991 | while (isspace (*dbp)) | |
1992 | dbp++; | |
1993 | if (*dbp == 0) | |
1994 | continue; | |
1995 | if (tail ("precision")) | |
1996 | break; | |
1997 | continue; | |
1998 | } | |
1999 | break; | |
2000 | } | |
2001 | while (isspace (*dbp)) | |
2002 | dbp++; | |
2003 | if (*dbp == 0) | |
2004 | continue; | |
2005 | switch (*dbp | ' ') | |
2006 | { | |
2007 | case 'f': | |
2008 | if (tail ("function")) | |
2009 | getit (); | |
2010 | continue; | |
2011 | case 's': | |
2012 | if (tail ("subroutine")) | |
2013 | getit (); | |
2014 | continue; | |
8a6c8bcf RS |
2015 | case 'e': |
2016 | if (tail ("entry")) | |
2017 | getit (); | |
2018 | continue; | |
c6d46f5f JB |
2019 | case 'p': |
2020 | if (tail ("program")) | |
2021 | { | |
2022 | getit (); | |
2023 | continue; | |
2024 | } | |
2025 | if (tail ("procedure")) | |
2026 | getit (); | |
2027 | continue; | |
2028 | } | |
2029 | } | |
2030 | return (pfcnt); | |
2031 | } | |
2032 | ||
2033 | logical | |
2034 | tail (cp) | |
2035 | char *cp; | |
2036 | { | |
2037 | register int len = 0; | |
2038 | ||
2039 | while (*cp && (*cp & ~' ') == ((*(dbp + len)) & ~' ')) | |
2040 | cp++, len++; | |
2041 | if (*cp == 0) | |
2042 | { | |
2043 | dbp += len; | |
2044 | return (1); | |
2045 | } | |
2046 | return (0); | |
2047 | } | |
2048 | ||
2049 | void | |
2050 | takeprec () | |
2051 | { | |
2052 | while (isspace (*dbp)) | |
2053 | dbp++; | |
2054 | if (*dbp != '*') | |
2055 | return; | |
2056 | dbp++; | |
2057 | while (isspace (*dbp)) | |
2058 | dbp++; | |
2059 | if (!isdigit (*dbp)) | |
2060 | { | |
2061 | --dbp; /* force failure */ | |
2062 | return; | |
2063 | } | |
2064 | do | |
2065 | dbp++; | |
2066 | while (isdigit (*dbp)); | |
2067 | } | |
2068 | ||
2069 | void | |
2070 | getit () | |
2071 | { | |
2072 | register char *cp; | |
2073 | char c; | |
2074 | char nambuf[BUFSIZ]; | |
2075 | ||
2076 | while (isspace (*dbp)) | |
2077 | dbp++; | |
daa37602 JB |
2078 | if (*dbp == 0 |
2079 | || (!isalpha (*dbp) | |
2080 | && *dbp != '_' | |
2081 | && *dbp != '$')) | |
c6d46f5f JB |
2082 | return; |
2083 | for (cp = dbp + 1; *cp && (isalpha (*cp) || isdigit (*cp) | |
2084 | || (*cp == '_') || (*cp == '$')); cp++) | |
2085 | continue; | |
2086 | c = cp[0]; | |
2087 | cp[0] = 0; | |
2088 | (void) strcpy (nambuf, dbp); | |
2089 | cp[0] = c; | |
2090 | pfnote (nambuf, TRUE, FALSE, lb.buffer, cp - lb.buffer + 1, lineno, linecharno); | |
2091 | pfcnt++; | |
2092 | } | |
2093 | ||
2094 | /* Handle a file of assembler code. */ | |
2095 | ||
2096 | void | |
2097 | Asm_funcs (fi) | |
2098 | FILE *fi; | |
2099 | { | |
2100 | int i; | |
2101 | register char c; | |
2102 | ||
2103 | lineno = 0; | |
2104 | charno = 0; | |
2105 | pfcnt = 0; | |
2106 | ||
2107 | while (!feof (fi)) | |
2108 | { | |
2109 | lineno++; | |
2110 | linecharno = charno; | |
2111 | charno += readline (&lb, fi); | |
2112 | dbp = lb.buffer; | |
2113 | ||
2114 | for (i = 0; ((c = dbp[i]) && !isspace (c)) && (c != ':'); i++) | |
2115 | ; | |
2116 | ||
2117 | if ((i > 0) && (c == ':')) | |
2118 | getit (); | |
2119 | } | |
2120 | } | |
2121 | \f | |
2122 | /* Added by Mosur Mohan, 4/22/88 */ | |
2123 | /* Pascal parsing */ | |
2124 | ||
2125 | #define GET_NEW_LINE \ | |
2126 | { \ | |
2127 | linecharno = charno; lineno++; \ | |
2128 | charno += 1 + readline (&lb, inf); \ | |
2129 | dbp = lb.buffer; \ | |
2130 | } | |
2131 | ||
2132 | /* Locates tags for procedures & functions. | |
2133 | * Doesn't do any type- or var-definitions. | |
2134 | * It does look for the keyword "extern" or "forward" | |
2135 | * immediately following the procedure statement; | |
2136 | * if found, the tag is skipped. | |
2137 | */ | |
2138 | ||
2139 | void | |
2140 | PAS_funcs (fi) | |
2141 | FILE *fi; | |
2142 | { | |
2143 | struct linebuffer tline; /* mostly copied from C_entries */ | |
2144 | long save_lcno; | |
2145 | int save_lineno; | |
2146 | char c, *cp; | |
2147 | char nambuf[BUFSIZ]; | |
2148 | ||
2149 | logical /* each of these flags is TRUE iff: */ | |
2150 | incomm1, /* point is inside {..} comment */ | |
2151 | incomm2, /* point is inside (*..*) comment */ | |
2152 | inquote, /* point is inside '..' string */ | |
2153 | get_tagname, /* point is after PROCEDURE/FUNCTION */ | |
2154 | /* keyword, so next item = potential tag */ | |
2155 | found_tag, /* point is after a potential tag */ | |
2156 | inparms, /* point is within parameter-list */ | |
2157 | verify_tag; /* point has passed the parm-list, so the */ | |
2158 | /* next token will determine whether */ | |
2159 | /* this is a FORWARD/EXTERN to be */ | |
2160 | /* ignored, or whether it is a real tag */ | |
2161 | ||
2162 | lineno = 0; | |
2163 | charno = 0; | |
2164 | dbp = lb.buffer; | |
2165 | *dbp = 0; | |
2166 | initbuffer (&tline); | |
2167 | ||
2168 | incomm1 = incomm2 = inquote = FALSE; | |
2169 | found_tag = FALSE; /* have a proc name; check if extern */ | |
2170 | get_tagname = FALSE; /* have found "procedure" keyword */ | |
2171 | inparms = FALSE; /* found '(' after "proc" */ | |
2172 | verify_tag = FALSE; /* check if "extern" is ahead */ | |
2173 | ||
2174 | /* long main loop to get next char */ | |
2175 | while (!feof (fi)) | |
2176 | { | |
2177 | c = *dbp++; | |
2178 | if (c == 0) /* if end of line */ | |
2179 | { | |
2180 | GET_NEW_LINE; | |
2181 | if (*dbp == 0) | |
2182 | continue; | |
2183 | if (!((found_tag && verify_tag) || | |
2184 | get_tagname)) | |
2185 | c = *dbp++; /* only if don't need *dbp pointing */ | |
2186 | /* to the beginning of the name of */ | |
2187 | /* the procedure or function */ | |
2188 | } | |
2189 | if (incomm1) /* within { - } comments */ | |
2190 | { | |
2191 | if (c == '}') | |
2192 | incomm1 = FALSE; | |
2193 | continue; | |
2194 | } | |
2195 | else if (incomm2) /* within (* - *) comments */ | |
2196 | { | |
2197 | if (c == '*') | |
2198 | { | |
2199 | while ((c = *dbp++) == '*') | |
2200 | continue; | |
2201 | if (c == 0) | |
2202 | GET_NEW_LINE; | |
2203 | if (c == ')') | |
2204 | incomm2 = FALSE; | |
2205 | } | |
2206 | continue; | |
2207 | } | |
2208 | else if (inquote) | |
2209 | { | |
2210 | if (c == '\'') | |
2211 | inquote = FALSE; | |
2212 | continue; | |
2213 | } | |
2214 | else | |
2215 | switch (c) | |
2216 | { | |
2217 | case '\'': | |
2218 | inquote = TRUE; /* found first quote */ | |
2219 | continue; | |
2220 | case '{': /* found open-{-comment */ | |
2221 | incomm1 = TRUE; | |
2222 | continue; | |
2223 | case '(': | |
2224 | if (*dbp == '*') /* found open-(*-comment */ | |
2225 | { | |
2226 | incomm2 = TRUE; | |
2227 | dbp++; | |
2228 | } | |
2229 | else if (found_tag) /* found '(' after tag, i.e., parm-list */ | |
2230 | inparms = TRUE; | |
2231 | continue; | |
2232 | case ')': /* end of parms list */ | |
2233 | if (inparms) | |
2234 | inparms = FALSE; | |
2235 | continue; | |
2236 | case ';': | |
2237 | if ((found_tag) && (!inparms)) /* end of proc or fn stmt */ | |
2238 | { | |
2239 | verify_tag = TRUE; | |
2240 | break; | |
2241 | } | |
2242 | continue; | |
2243 | } | |
2244 | if ((found_tag) && (verify_tag) && (*dbp != ' ')) | |
2245 | { | |
2246 | /* check if this is an "extern" declaration */ | |
2247 | if (*dbp == 0) | |
2248 | continue; | |
2249 | if ((*dbp == 'e') || (*dbp == 'E')) | |
2250 | { | |
2251 | if (tail ("extern")) /* superfluous, really! */ | |
2252 | { | |
2253 | found_tag = FALSE; | |
2254 | verify_tag = FALSE; | |
2255 | } | |
2256 | } | |
2257 | else if ((*dbp == 'f') || (*dbp == 'F')) | |
2258 | { | |
2259 | if (tail ("forward")) /* check for forward reference */ | |
2260 | { | |
2261 | found_tag = FALSE; | |
2262 | verify_tag = FALSE; | |
2263 | } | |
2264 | } | |
2265 | if ((found_tag) && (verify_tag)) /* not external proc, so make tag */ | |
2266 | { | |
2267 | found_tag = FALSE; | |
2268 | verify_tag = FALSE; | |
2269 | pfnote (nambuf, TRUE, FALSE, | |
2270 | tline.buffer, cp - tline.buffer + 1, | |
2271 | save_lineno, save_lcno); | |
2272 | continue; | |
2273 | } | |
2274 | } | |
2275 | if (get_tagname) /* grab name of proc or fn */ | |
2276 | { | |
2277 | if (*dbp == 0) | |
2278 | continue; | |
2279 | ||
2280 | /* save all values for later tagging */ | |
2281 | tline.size = lb.size; | |
2282 | strcpy (tline.buffer, lb.buffer); | |
2283 | save_lineno = lineno; | |
2284 | save_lcno = linecharno; | |
2285 | ||
2286 | /* grab block name */ | |
2287 | for (cp = dbp + 1; *cp && (!endtoken (*cp)); cp++) | |
2288 | continue; | |
2289 | c = cp[0]; | |
2290 | cp[0] = 0; | |
2291 | strcpy (nambuf, dbp); | |
2292 | cp[0] = c; | |
2293 | dbp = cp; /* restore dbp to e-o-token */ | |
2294 | get_tagname = FALSE; | |
2295 | found_tag = TRUE; | |
2296 | continue; | |
2297 | ||
2298 | /* and proceed to check for "extern" */ | |
2299 | } | |
2300 | if ((!incomm1) && (!incomm2) && (!inquote) && | |
2301 | (!found_tag) && (!get_tagname)) | |
2302 | { | |
2303 | /* check for proc/fn keywords */ | |
2304 | switch (c | ' ') | |
2305 | { | |
2306 | case 'p': | |
2307 | if (tail ("rocedure")) /* c = 'p', dbp has advanced */ | |
2308 | get_tagname = TRUE; | |
2309 | continue; | |
2310 | case 'f': | |
2311 | if (tail ("unction")) | |
2312 | get_tagname = TRUE; | |
2313 | continue; | |
2314 | } | |
2315 | } | |
2316 | } /* while not e-o-f */ | |
2317 | } | |
2318 | \f | |
2319 | /* | |
2320 | * lisp tag functions | |
2321 | * just look for (def or (DEF | |
2322 | */ | |
2323 | ||
2324 | void | |
2325 | L_funcs (fi) | |
2326 | FILE *fi; | |
2327 | { | |
2328 | lineno = 0; | |
2329 | charno = 0; | |
2330 | pfcnt = 0; | |
2331 | ||
2332 | while (!feof (fi)) | |
2333 | { | |
2334 | lineno++; | |
2335 | linecharno = charno; | |
2336 | charno += readline (&lb, fi); | |
2337 | dbp = lb.buffer; | |
2338 | if (dbp[0] == '(') | |
2339 | { | |
2340 | if (L_isdef (dbp)) | |
2341 | { | |
2342 | while (!isspace (*dbp)) | |
2343 | dbp++; | |
2344 | while (isspace (*dbp)) | |
2345 | dbp++; | |
2346 | L_getit (); | |
2347 | } | |
2348 | else | |
2349 | { | |
2350 | /* Check for (foo::defmumble name-defined ... */ | |
2351 | while (*dbp && *dbp != ':' && !isspace (*dbp) | |
2352 | && *dbp != '(' && *dbp != ')') | |
2353 | dbp++; | |
2354 | if (*dbp == ':') | |
2355 | { | |
2356 | while (*dbp == ':') | |
2357 | dbp++; | |
2358 | ||
2359 | if (L_isdef (dbp)) | |
2360 | { | |
2361 | while (!isspace (*dbp)) | |
2362 | dbp++; | |
2363 | while (isspace (*dbp)) | |
2364 | dbp++; | |
2365 | L_getit (); | |
2366 | } | |
2367 | } | |
2368 | } | |
2369 | } | |
2370 | } | |
2371 | } | |
2372 | ||
2373 | int | |
2374 | L_isdef (dbp) | |
2375 | char *dbp; | |
2376 | { | |
2377 | return ((dbp[1] == 'D' || dbp[1] == 'd') && | |
2378 | (dbp[2] == 'E' || dbp[2] == 'e') && | |
2379 | (dbp[3] == 'F' || dbp[3] == 'f')); | |
2380 | } | |
2381 | ||
2382 | void | |
2383 | L_getit () | |
2384 | { | |
2385 | register char *cp; | |
2386 | char c; | |
2387 | char nambuf[BUFSIZ]; | |
2388 | ||
2389 | if (*dbp == 0) | |
2390 | return; | |
2391 | for (cp = dbp + 1; *cp && *cp != '(' && *cp != ' '; cp++) | |
2392 | continue; | |
2393 | c = cp[0]; | |
2394 | cp[0] = 0; | |
2395 | (void) strcpy (nambuf, dbp); | |
2396 | cp[0] = c; | |
2397 | pfnote (nambuf, TRUE, FALSE, lb.buffer, cp - lb.buffer + 1, lineno, linecharno); | |
2398 | pfcnt++; | |
2399 | } | |
2400 | \f | |
2401 | /* | |
2402 | * Scheme tag functions | |
2403 | * look for (def... xyzzy | |
2404 | * look for (def... (xyzzy | |
2405 | * look for (def ... ((...(xyzzy .... | |
2406 | * look for (set! xyzzy | |
2407 | */ | |
2408 | ||
2409 | static void get_scheme (); | |
2410 | ||
2411 | void | |
2412 | Scheme_funcs (fi) | |
2413 | FILE *fi; | |
2414 | { | |
2415 | lineno = 0; | |
2416 | charno = 0; | |
2417 | pfcnt = 0; | |
2418 | ||
2419 | while (!feof (fi)) | |
2420 | { | |
2421 | lineno++; | |
2422 | linecharno = charno; | |
2423 | charno += readline (&lb, fi); | |
2424 | dbp = lb.buffer; | |
2425 | if (dbp[0] == '(' && | |
2426 | (dbp[1] == 'D' || dbp[1] == 'd') && | |
2427 | (dbp[2] == 'E' || dbp[2] == 'e') && | |
2428 | (dbp[3] == 'F' || dbp[3] == 'f')) | |
2429 | { | |
2430 | while (!isspace (*dbp)) | |
2431 | dbp++; | |
2432 | /* Skip over open parens and white space */ | |
2433 | while (*dbp && (isspace (*dbp) || *dbp == '(')) | |
2434 | dbp++; | |
2435 | get_scheme (); | |
2436 | } | |
2437 | if (dbp[0] == '(' && | |
2438 | (dbp[1] == 'S' || dbp[1] == 's') && | |
2439 | (dbp[2] == 'E' || dbp[2] == 'e') && | |
2440 | (dbp[3] == 'T' || dbp[3] == 't') && | |
2441 | (dbp[4] == '!' || dbp[4] == '!') && | |
2442 | (isspace (dbp[5]))) | |
2443 | { | |
2444 | while (!isspace (*dbp)) | |
2445 | dbp++; | |
2446 | /* Skip over white space */ | |
2447 | while (isspace (*dbp)) | |
2448 | dbp++; | |
2449 | get_scheme (); | |
2450 | } | |
2451 | } | |
2452 | } | |
2453 | ||
2454 | static void | |
2455 | get_scheme () | |
2456 | { | |
2457 | register char *cp; | |
2458 | char c; | |
2459 | char nambuf[BUFSIZ]; | |
2460 | ||
2461 | if (*dbp == 0) | |
2462 | return; | |
2463 | /* Go till you get to white space or a syntactic break */ | |
2464 | for (cp = dbp + 1; *cp && *cp != '(' && *cp != ')' && !isspace (*cp); cp++) | |
2465 | continue; | |
2466 | /* Null terminate the string there. */ | |
2467 | c = cp[0]; | |
2468 | cp[0] = 0; | |
2469 | /* Copy the string */ | |
2470 | strcpy (nambuf, dbp); | |
2471 | /* Unterminate the string */ | |
2472 | cp[0] = c; | |
2473 | /* Announce the change */ | |
2474 | pfnote (nambuf, TRUE, FALSE, lb.buffer, cp - lb.buffer + 1, lineno, linecharno); | |
2475 | pfcnt++; | |
2476 | } | |
2477 | \f | |
2478 | /* Find tags in TeX and LaTeX input files. */ | |
2479 | ||
2480 | /* TEX_toktab is a table of TeX control sequences that define tags. | |
2481 | Each TEX_tabent records one such control sequence. | |
2482 | CONVERT THIS TO USE THE Stab TYPE!! */ | |
2483 | ||
2484 | struct TEX_tabent | |
2485 | { | |
2486 | char *name; | |
2487 | int len; | |
2488 | }; | |
2489 | ||
2490 | struct TEX_tabent *TEX_toktab = NULL; /* Table with tag tokens */ | |
2491 | ||
2492 | /* Default set of control sequences to put into TEX_toktab. | |
2493 | The value of environment var TEXTAGS is prepended to this. */ | |
2494 | ||
2495 | static char *TEX_defenv = | |
2496 | ":chapter:section:subsection:subsubsection:eqno:label:ref:cite:bibitem:typeout"; | |
2497 | ||
2498 | void TEX_mode (); | |
2499 | struct TEX_tabent *TEX_decode_env (); | |
2500 | void TEX_getit (); | |
2501 | int TEX_Token (); | |
2502 | ||
2503 | static char TEX_esc = '\\'; | |
2504 | static char TEX_opgrp = '{'; | |
2505 | static char TEX_clgrp = '}'; | |
2506 | ||
2507 | /* | |
2508 | * TeX/LaTeX scanning loop. | |
2509 | */ | |
2510 | ||
2511 | void | |
2512 | TEX_funcs (fi) | |
2513 | FILE *fi; | |
2514 | { | |
2515 | char *lasthit; | |
2516 | ||
2517 | lineno = 0; | |
2518 | charno = 0; | |
2519 | pfcnt = 0; | |
2520 | ||
2521 | /* Select either \ or ! as escape character. */ | |
2522 | TEX_mode (fi); | |
2523 | ||
2524 | /* Initialize token table once from environment. */ | |
2525 | if (!TEX_toktab) | |
2526 | TEX_toktab = TEX_decode_env ("TEXTAGS", TEX_defenv); | |
2527 | ||
2528 | while (!feof (fi)) | |
d2729198 | 2529 | { /* Scan each line in file */ |
c6d46f5f JB |
2530 | lineno++; |
2531 | linecharno = charno; | |
2532 | charno += readline (&lb, fi); | |
2533 | dbp = lb.buffer; | |
2534 | lasthit = dbp; | |
d2729198 | 2535 | while (dbp = etags_index (dbp, TEX_esc)) /* Look at each escape in line */ |
8a6c8bcf RS |
2536 | { |
2537 | register int i; | |
c6d46f5f | 2538 | |
8a6c8bcf RS |
2539 | if (!*(++dbp)) |
2540 | break; | |
2541 | linecharno += dbp - lasthit; | |
c6d46f5f | 2542 | lasthit = dbp; |
8a6c8bcf RS |
2543 | i = TEX_Token (lasthit); |
2544 | if (0 <= i) | |
c6d46f5f | 2545 | { |
8a6c8bcf | 2546 | TEX_getit (lasthit, TEX_toktab[i].len); |
d2729198 | 2547 | break; /* We only save a line once */ |
c6d46f5f JB |
2548 | } |
2549 | } | |
2550 | } | |
2551 | } | |
2552 | ||
2553 | #define TEX_LESC '\\' | |
2554 | #define TEX_SESC '!' | |
2555 | #define TEX_cmt '%' | |
2556 | ||
2557 | /* Figure out whether TeX's escapechar is '\\' or '!' and set grouping */ | |
2558 | /* chars accordingly. */ | |
2559 | ||
2560 | void | |
2561 | TEX_mode (f) | |
2562 | FILE *f; | |
2563 | { | |
2564 | int c; | |
2565 | ||
2566 | while ((c = getc (f)) != EOF) | |
2567 | { | |
2568 | /* Skip to next line if we hit the TeX comment char. */ | |
2569 | if (c == TEX_cmt) | |
2570 | while (c != '\n') | |
2571 | c = getc (f); | |
2572 | else if (c == TEX_LESC || c == TEX_SESC ) | |
2573 | break; | |
2574 | } | |
2575 | ||
2576 | if (c == TEX_LESC) | |
2577 | { | |
2578 | TEX_esc = TEX_LESC; | |
2579 | TEX_opgrp = '{'; | |
2580 | TEX_clgrp = '}'; | |
2581 | } | |
2582 | else | |
2583 | { | |
2584 | TEX_esc = TEX_SESC; | |
2585 | TEX_opgrp = '<'; | |
2586 | TEX_clgrp = '>'; | |
2587 | } | |
2588 | rewind (f); | |
2589 | } | |
2590 | ||
2591 | /* Read environment and prepend it to the default string. */ | |
2592 | /* Build token table. */ | |
2593 | ||
2594 | struct TEX_tabent * | |
2595 | TEX_decode_env (evarname, defenv) | |
2596 | char *evarname; | |
2597 | char *defenv; | |
2598 | { | |
2599 | register char *env, *p; | |
c6d46f5f JB |
2600 | |
2601 | struct TEX_tabent *tab; | |
2602 | int size, i; | |
2603 | ||
2604 | /* Append default string to environment. */ | |
2605 | env = getenv (evarname); | |
2606 | if (!env) | |
2607 | env = defenv; | |
2608 | else | |
2609 | env = concat (env, defenv, ""); | |
2610 | ||
2611 | /* Allocate a token table */ | |
2612 | for (size = 1, p = env; p;) | |
8a6c8bcf | 2613 | if ((p = etags_index (p, ':')) && *(++p)) |
c6d46f5f | 2614 | size++; |
8a6c8bcf RS |
2615 | /* Add 1 to leave room for null terminator. */ |
2616 | tab = xnew (size + 1, struct TEX_tabent); | |
c6d46f5f JB |
2617 | |
2618 | /* Unpack environment string into token table. Be careful about */ | |
2619 | /* zero-length strings (leading ':', "::" and trailing ':') */ | |
2620 | for (i = 0; *env;) | |
2621 | { | |
8a6c8bcf | 2622 | p = etags_index (env, ':'); |
c6d46f5f JB |
2623 | if (!p) /* End of environment string. */ |
2624 | p = env + strlen (env); | |
2625 | if (p - env > 0) | |
2626 | { /* Only non-zero strings. */ | |
2627 | tab[i].name = savenstr (env, p - env); | |
2628 | tab[i].len = strlen (tab[i].name); | |
2629 | i++; | |
2630 | } | |
2631 | if (*p) | |
2632 | env = p + 1; | |
2633 | else | |
2634 | { | |
2635 | tab[i].name = NULL; /* Mark end of table. */ | |
2636 | tab[i].len = 0; | |
2637 | break; | |
2638 | } | |
2639 | } | |
2640 | return tab; | |
2641 | } | |
2642 | ||
2643 | /* Record a tag defined by a TeX command of length LEN and starting at NAME. | |
2644 | The name being defined actually starts at (NAME + LEN + 1). | |
2645 | But we seem to include the TeX command in the tag name. */ | |
2646 | ||
2647 | void | |
2648 | TEX_getit (name, len) | |
2649 | char *name; | |
2650 | int len; | |
2651 | { | |
2652 | char *p = name + len; | |
2653 | char nambuf[BUFSIZ]; | |
2654 | ||
2655 | if (*name == 0) | |
2656 | return; | |
2657 | ||
2658 | /* Let tag name extend to next group close (or end of line) */ | |
2659 | while (*p && *p != TEX_clgrp) | |
2660 | p++; | |
2661 | (void) strncpy (nambuf, name, p - name); | |
2662 | nambuf[p - name] = 0; | |
2663 | ||
2664 | pfnote (nambuf, TRUE, FALSE, lb.buffer, strlen (lb.buffer), lineno, linecharno); | |
2665 | pfcnt++; | |
2666 | } | |
2667 | ||
2668 | /* If the text at CP matches one of the tag-defining TeX command names, | |
8a6c8bcf | 2669 | return the etags_index of that command in TEX_toktab. |
c6d46f5f JB |
2670 | Otherwise return -1. */ |
2671 | ||
2672 | /* Keep the capital `T' in `Token' for dumb truncating compilers | |
2673 | (this distinguishes it from `TEX_toktab' */ | |
2674 | int | |
2675 | TEX_Token (cp) | |
2676 | char *cp; | |
2677 | { | |
2678 | int i; | |
2679 | ||
2680 | for (i = 0; TEX_toktab[i].len > 0; i++) | |
2681 | if (strncmp (TEX_toktab[i].name, cp, TEX_toktab[i].len) == 0) | |
2682 | return i; | |
2683 | return -1; | |
2684 | } | |
2685 | \f | |
2686 | /* Support for Prolog. */ | |
2687 | ||
2688 | /* whole head (not only functor, but also arguments) | |
2689 | is gotten in compound term. */ | |
2690 | ||
2691 | void | |
2692 | prolog_getit (s, lineno, linecharno) | |
2693 | char *s; | |
2694 | int lineno; | |
2695 | long linecharno; | |
2696 | { | |
2697 | char nambuf[BUFSIZ], *save_s, tmpc; | |
2698 | int insquote, npar; | |
2699 | ||
2700 | save_s = s; | |
2701 | insquote = FALSE; | |
2702 | npar = 0; | |
2703 | while (1) | |
2704 | { | |
2705 | if (*s == '\0') /* syntax error. */ | |
2706 | return; | |
2707 | else if (insquote && *s == '\'' && *(s + 1) == '\'') | |
2708 | s += 2; | |
2709 | else if (*s == '\'') | |
2710 | { | |
2711 | insquote = !insquote; | |
2712 | s++; | |
2713 | } | |
2714 | else if (!insquote && *s == '(') | |
2715 | { | |
2716 | npar++; | |
2717 | s++; | |
2718 | } | |
2719 | else if (!insquote && *s == ')') | |
2720 | { | |
2721 | npar--; | |
2722 | s++; | |
2723 | if (npar == 0) | |
2724 | break; | |
2725 | else if (npar < 0) /* syntax error. */ | |
2726 | return; | |
2727 | } | |
2728 | else if (!insquote && *s == '.' && (isspace (*(s + 1)) || *(s + 1) == '\0')) | |
2729 | { /* fullstop. */ | |
2730 | if (npar != 0) /* syntax error. */ | |
2731 | return; | |
2732 | s++; | |
2733 | break; | |
2734 | } | |
2735 | else | |
2736 | s++; | |
2737 | } | |
2738 | tmpc = *s; | |
2739 | *s = '\0'; | |
2740 | strcpy (nambuf, save_s); | |
2741 | *s = tmpc; | |
2742 | pfnote (nambuf, TRUE, save_s, strlen (nambuf), lineno, linecharno); | |
2743 | } | |
2744 | ||
2745 | /* It is assumed that prolog predicate starts from column 0. */ | |
2746 | ||
2747 | void | |
2748 | prolog_funcs (fi) | |
2749 | FILE *fi; | |
2750 | { | |
2751 | void skip_comment (), prolog_getit (); | |
2752 | ||
2753 | lineno = linecharno = charno = 0; | |
2754 | while (!feof (fi)) | |
2755 | { | |
2756 | lineno++; | |
2757 | linecharno += charno; | |
2758 | charno = readline (&lb, fi) + 1; /* 1 for newline. */ | |
2759 | dbp = lb.buffer; | |
2760 | if (isspace (dbp[0])) /* not predicate header. */ | |
2761 | continue; | |
2762 | else if (dbp[0] == '%') /* comment. */ | |
2763 | continue; | |
2764 | else if (dbp[0] == '/' && dbp[1] == '*') /* comment. */ | |
2765 | skip_comment (&lb, fi, &lineno, &linecharno); | |
2766 | else /* found. */ | |
2767 | prolog_getit (dbp, lineno, linecharno); | |
2768 | } | |
2769 | } | |
2770 | ||
2771 | void | |
2772 | skip_comment (plb, fi, plineno, plinecharno) | |
2773 | struct linebuffer *plb; | |
2774 | FILE *fi; | |
2775 | int *plineno; /* result */ | |
2776 | long *plinecharno; /* result */ | |
2777 | { | |
2778 | while (!substr ("*/", plb->buffer)) | |
2779 | { | |
2780 | (*plineno)++; | |
2781 | *plinecharno += readline (plb, fi) + 1; | |
2782 | } /* 1 for newline. */ | |
2783 | } | |
2784 | ||
2785 | /* Return TRUE if 'sub' exists somewhere in 's'. */ | |
2786 | ||
2787 | int | |
2788 | substr (sub, s) | |
2789 | char *sub; | |
2790 | char *s; | |
2791 | { | |
8a6c8bcf | 2792 | while (*s && (s = etags_index (s, *sub))) |
c6d46f5f JB |
2793 | if (prestr (sub, s)) |
2794 | return (TRUE); | |
2795 | else | |
2796 | s++; | |
2797 | return (FALSE); | |
2798 | } | |
2799 | ||
2800 | /* Return TRUE if 'pre' is prefix of string 's'. */ | |
2801 | ||
2802 | int | |
2803 | prestr (pre, s) | |
2804 | char *pre; | |
2805 | char *s; | |
2806 | { | |
2807 | if (*pre == '\0') | |
2808 | return (TRUE); | |
2809 | else if (*pre == *s) | |
2810 | return (prestr (pre + 1, s + 1)); | |
2811 | else | |
2812 | return (FALSE); | |
2813 | } | |
2814 | \f | |
2815 | /* Initialize a linebuffer for use */ | |
2816 | ||
2817 | void | |
2818 | initbuffer (linebuffer) | |
2819 | struct linebuffer *linebuffer; | |
2820 | { | |
2821 | linebuffer->size = 200; | |
2822 | linebuffer->buffer = xnew (200, char); | |
2823 | } | |
2824 | ||
2825 | /* | |
2826 | * Read a line of text from `stream' into `linebuffer'. | |
2827 | * Return the number of characters read from `stream', | |
2828 | * which is the length of the line including the newline, if any. | |
2829 | */ | |
2830 | long | |
2831 | readline (linebuffer, stream) | |
2832 | struct linebuffer *linebuffer; | |
2833 | register FILE *stream; | |
2834 | { | |
2835 | char *buffer = linebuffer->buffer; | |
2836 | register char *p = linebuffer->buffer; | |
2837 | register char *pend; | |
2838 | int newline; /* 1 if ended with newline, 0 if ended with EOF */ | |
2839 | ||
2840 | pend = p + linebuffer->size; /* Separate to avoind 386/IX compiler bug. */ | |
2841 | ||
2842 | while (1) | |
2843 | { | |
2844 | register int c = getc (stream); | |
2845 | if (p == pend) | |
2846 | { | |
2847 | linebuffer->size *= 2; | |
2848 | buffer = (char *) xrealloc (buffer, linebuffer->size); | |
2849 | p += buffer - linebuffer->buffer; | |
2850 | pend = buffer + linebuffer->size; | |
2851 | linebuffer->buffer = buffer; | |
2852 | } | |
2853 | if (c < 0 || c == '\n') | |
2854 | { | |
2855 | *p = 0; | |
2856 | newline = (c == '\n' ? 1 : 0); | |
2857 | break; | |
2858 | } | |
2859 | *p++ = c; | |
2860 | } | |
2861 | ||
2862 | return p - buffer + newline; | |
2863 | } | |
2864 | \f | |
2865 | char * | |
2866 | savestr (cp) | |
2867 | char *cp; | |
2868 | { | |
2869 | return savenstr (cp, strlen (cp)); | |
2870 | } | |
2871 | ||
2872 | char * | |
2873 | savenstr (cp, len) | |
2874 | char *cp; | |
2875 | int len; | |
2876 | { | |
2877 | register char *dp; | |
2878 | ||
2879 | dp = xnew (len + 1, char); | |
2880 | (void) strncpy (dp, cp, len); | |
2881 | dp[len] = '\0'; | |
2882 | return dp; | |
2883 | } | |
2884 | ||
c6d46f5f JB |
2885 | /* |
2886 | * Return the ptr in sp at which the character c last | |
2887 | * appears; NULL if not found | |
2888 | * | |
2889 | * Identical to v7 rindex, included for portability. | |
2890 | */ | |
2891 | ||
2892 | char * | |
8a6c8bcf | 2893 | etags_rindex (sp, c) |
c6d46f5f JB |
2894 | register char *sp, c; |
2895 | { | |
2896 | register char *r; | |
2897 | ||
2898 | r = NULL; | |
2899 | do | |
2900 | { | |
2901 | if (*sp == c) | |
2902 | r = sp; | |
2903 | } while (*sp++); | |
2904 | return (r); | |
2905 | } | |
2906 | ||
9d7ad1b3 | 2907 | |
c6d46f5f JB |
2908 | /* |
2909 | * Return the ptr in sp at which the character c first | |
2910 | * appears; NULL if not found | |
2911 | * | |
2912 | * Identical to v7 index, included for portability. | |
2913 | */ | |
2914 | ||
2915 | char * | |
8a6c8bcf | 2916 | etags_index (sp, c) |
c6d46f5f JB |
2917 | register char *sp, c; |
2918 | { | |
2919 | do | |
2920 | { | |
2921 | if (*sp == c) | |
2922 | return (sp); | |
2923 | } while (*sp++); | |
2924 | return (NULL); | |
2925 | } | |
2926 | ||
c6d46f5f JB |
2927 | /* Print error message and exit. */ |
2928 | ||
2929 | /* VARARGS1 */ | |
2930 | void | |
2931 | fatal (s1, s2) | |
2932 | char *s1, *s2; | |
2933 | { | |
2934 | error (s1, s2); | |
2935 | exit (1); | |
2936 | } | |
2937 | ||
2938 | /* Print error message. `s1' is printf control string, `s2' is arg for it. */ | |
2939 | ||
2940 | /* VARARGS1 */ | |
2941 | void | |
2942 | error (s1, s2) | |
2943 | char *s1, *s2; | |
2944 | { | |
2945 | fprintf (stderr, "%s: ", progname); | |
2946 | fprintf (stderr, s1, s2); | |
2947 | fprintf (stderr, "\n"); | |
2948 | } | |
2949 | ||
2950 | /* Return a newly-allocated string whose contents concatenate those of s1, s2, s3. */ | |
2951 | ||
2952 | char * | |
2953 | concat (s1, s2, s3) | |
2954 | char *s1, *s2, *s3; | |
2955 | { | |
2956 | int len1 = strlen (s1), len2 = strlen (s2), len3 = strlen (s3); | |
2957 | char *result = xnew (len1 + len2 + len3 + 1, char); | |
2958 | ||
2959 | (void) strcpy (result, s1); | |
2960 | (void) strcpy (result + len1, s2); | |
2961 | (void) strcpy (result + len1 + len2, s3); | |
2962 | *(result + len1 + len2 + len3) = 0; | |
2963 | ||
2964 | return result; | |
2965 | } | |
2966 | ||
2967 | /* Like malloc but get fatal error if memory is exhausted. */ | |
2968 | ||
2969 | char * | |
2970 | xmalloc (size) | |
2971 | int size; | |
2972 | { | |
2973 | char *result = malloc (size); | |
2974 | if (!result) | |
2975 | fatal ("virtual memory exhausted", 0); | |
2976 | return result; | |
2977 | } | |
2978 | ||
2979 | char * | |
2980 | xrealloc (ptr, size) | |
2981 | char *ptr; | |
2982 | int size; | |
2983 | { | |
2984 | char *result = realloc (ptr, size); | |
2985 | if (!result) | |
2986 | fatal ("virtual memory exhausted"); | |
2987 | return result; | |
2988 | } |