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++. | |
31d4b314 FP |
27 | * |
28 | * Francesco Potorti` is the current maintainer. 7.5 | |
c6d46f5f JB |
29 | */ |
30 | ||
72a339d7 | 31 | #ifdef HAVE_CONFIG_H |
18160b98 | 32 | #include <../src/config.h> |
1e134a5f RM |
33 | #endif |
34 | ||
918f9ad1 JB |
35 | #include <stdio.h> |
36 | #include <ctype.h> | |
37 | #include <sys/types.h> | |
38 | #include <sys/stat.h> | |
39 | ||
40 | #include "getopt.h" | |
41 | ||
c6d46f5f JB |
42 | extern char *malloc (), *realloc (); |
43 | extern char *getenv (); | |
c6d46f5f JB |
44 | extern char *strcpy (), *strncpy (); |
45 | extern int strcmp (); | |
46 | ||
8a6c8bcf RS |
47 | char *etags_index (), *etags_rindex (); |
48 | char *savenstr (); | |
c6d46f5f JB |
49 | |
50 | /* Define the symbol ETAGS to make the program "etags", | |
51 | which makes emacs-style tag tables by default. | |
52 | Define CTAGS to make the program "ctags" compatible with the usual one. | |
53 | Define neither one to get behavior that depends | |
54 | on the name with which the program is invoked | |
55 | (but we don't normally compile it that way). */ | |
56 | ||
57 | #if !defined(ETAGS) && !defined(CTAGS) | |
58 | /* If neither is defined, program can be run as either. */ | |
59 | #define ETAGS | |
60 | #define CTAGS | |
61 | #endif | |
62 | ||
63 | /* On VMS, CTAGS is not useful, so always do ETAGS. */ | |
64 | #ifdef VMS | |
65 | #ifndef ETAGS | |
66 | #define ETAGS | |
67 | #endif | |
68 | #endif | |
69 | ||
70 | /* Exit codes for success and failure. */ | |
71 | #ifdef VMS | |
72 | #define GOOD (1) | |
73 | #define BAD (0) | |
74 | #else | |
75 | #define GOOD (0) | |
76 | #define BAD (1) | |
77 | #endif | |
78 | ||
79 | /* | |
80 | * The FILEPOS abstract type, which represents a position in a file, | |
81 | * plus the following accessor functions: | |
82 | * | |
83 | * long GET_CHARNO (pos) | |
84 | * returns absolute char number. | |
c6d46f5f JB |
85 | * void SET_FILEPOS (pos, fp, charno) |
86 | * FILE *fp; long charno; | |
87 | * sets `pos' from the current file | |
88 | * position of `fp' and from `charno', | |
89 | * which must be the absolute character | |
90 | * number corresponding to the current | |
91 | * position of `fp'. | |
92 | * | |
93 | * The `pos' parameter is an lvalue expression of type FILEPOS. | |
94 | * Parameters to the accessor functions are evaluated 0 or more times, | |
95 | * and so must have no side effects. | |
96 | * | |
97 | * FILEPOS objects can also be assigned and passed to and from | |
98 | * functions in the normal C manner. | |
99 | * | |
100 | * Implementation notes: the `+ 0' is to enforce rvalue-ness. | |
101 | */ | |
c6d46f5f | 102 | |
c6d46f5f | 103 | #ifndef DEBUG |
13fde0cd | 104 | /* real implementation */ |
c6d46f5f JB |
105 | typedef long FILEPOS; |
106 | #define GET_CHARNO(pos) ((pos) + 0) | |
c6d46f5f JB |
107 | #define SET_FILEPOS(pos, fp, cno) ((void) ((pos) = (cno))) |
108 | #else | |
13fde0cd | 109 | /* debugging implementation */ |
c6d46f5f JB |
110 | typedef struct |
111 | { | |
112 | long charno; | |
113 | } FILEPOS; | |
114 | ||
115 | #define GET_CHARNO(pos) ((pos).charno + 0) | |
c6d46f5f JB |
116 | #define SET_FILEPOS(pos, fp, cno) \ |
117 | ((void) ((pos).charno = (cno), \ | |
118 | (cno) != ftell (fp) ? (error ("SET_FILEPOS inconsistency"), 0) \ | |
119 | : 0)) | |
120 | #endif | |
c6d46f5f JB |
121 | |
122 | #define streq(s, t) (strcmp (s, t) == 0) | |
123 | #define strneq(s, t, n) (strncmp (s, t, n) == 0) | |
13fde0cd | 124 | #define logical int |
c6d46f5f JB |
125 | |
126 | #define TRUE 1 | |
127 | #define FALSE 0 | |
128 | ||
129 | #define iswhite(arg) (_wht[arg]) /* T if char is white */ | |
130 | #define begtoken(arg) (_btk[arg]) /* T if char can start token */ | |
131 | #define intoken(arg) (_itk[arg]) /* T if char can be in token */ | |
132 | #define endtoken(arg) (_etk[arg]) /* T if char ends tokens */ | |
c6d46f5f JB |
133 | |
134 | #define max(I1,I2) ((I1) > (I2) ? (I1) : (I2)) | |
135 | ||
136 | struct nd_st | |
137 | { /* sorting structure */ | |
138 | char *name; /* function or type name */ | |
139 | char *file; /* file name */ | |
140 | logical is_func; /* use pattern or line no */ | |
141 | logical rewritten; /* list name separately */ | |
142 | logical been_warned; /* set if noticed dup */ | |
143 | int lno; /* line number tag is on */ | |
144 | long cno; /* character number line starts on */ | |
145 | char *pat; /* search pattern */ | |
146 | struct nd_st *left, *right; /* left and right sons */ | |
147 | }; | |
148 | ||
149 | long ftell (); | |
150 | typedef struct nd_st NODE; | |
151 | ||
13fde0cd RS |
152 | logical header_file; /* TRUE if .h file, FALSE o.w. */ |
153 | /* boolean "functions" (see init) */ | |
154 | logical _wht[0177], _etk[0177], _itk[0177], _btk[0177]; | |
c6d46f5f JB |
155 | |
156 | ||
157 | char *concat (); | |
158 | char *savenstr (); | |
159 | char *savestr (); | |
160 | char *xmalloc (); | |
161 | char *xrealloc (); | |
31d4b314 | 162 | int L_isdef (), L_isquote (); |
c6d46f5f JB |
163 | int PF_funcs (); |
164 | int total_size_of_entries (); | |
165 | logical consider_token (); | |
166 | logical tail (); | |
167 | long readline (); | |
168 | void Asm_funcs (); | |
169 | void C_entries (); | |
170 | void L_funcs (); | |
171 | void L_getit (); | |
172 | void PAS_funcs (); | |
173 | void Scheme_funcs (); | |
174 | void TEX_funcs (); | |
175 | void add_node (); | |
176 | void error (); | |
177 | void fatal (); | |
178 | void find_entries (); | |
179 | void free_tree (); | |
180 | void getit (); | |
181 | void getline (); | |
182 | void init (); | |
183 | void initbuffer (); | |
184 | void initbuffer (); | |
185 | void pfnote (); | |
186 | void process_file (); | |
187 | void put_entries (); | |
188 | void takeprec (); | |
189 | ||
190 | /* | |
191 | * MACRO | |
192 | * xnew -- allocate storage | |
193 | * | |
194 | * SYNOPSIS | |
195 | * Type *xnew (int n, Type); | |
196 | */ | |
197 | #define xnew(n, Type) ((Type *) xmalloc ((n) * sizeof (Type))) | |
198 | \f | |
199 | ||
200 | ||
201 | /* | |
202 | * Symbol table stuff. | |
203 | * | |
204 | * Should probably be implemented with hash table; linked list for now. | |
205 | */ | |
206 | ||
207 | enum sym_type | |
208 | { | |
209 | st_none, st_C_struct, st_C_enum, st_C_define, st_C_typedef, st_C_typespec | |
210 | }; | |
211 | ||
212 | struct stab_entry | |
213 | { | |
214 | char *sym; | |
215 | int symlen; | |
216 | enum sym_type type; | |
217 | struct stab_entry *next; | |
218 | }; | |
219 | ||
220 | typedef struct stab_entry Stab_entry; | |
221 | typedef Stab_entry *Stab; | |
222 | ||
223 | /* | |
224 | * NAME | |
225 | * Stab, Stab_entry, stab_create, stab_search, stab_find -- symbol table | |
226 | * | |
227 | * SYNOPSIS | |
228 | * Types: Stab, Stab_entry, enum sym_type | |
229 | * | |
230 | * Stab * stab_create () | |
231 | * | |
232 | * Stab_entry * stab_find (stab, sym) | |
233 | * Stab *stab; | |
234 | * char *sym; | |
235 | * | |
236 | * Stab_entry * stab_search (stab, sym) | |
237 | * Stab *stab; | |
238 | * char *sym; | |
239 | * | |
240 | * DESCRIPTION | |
241 | * stab_create creates a Stab, a symbol table object, and returns a | |
242 | * pointer to it. stab_find finds a symbol in a Stab; it returns a | |
243 | * pointer to the Stab_entry if found, otherwise NULL. stab_search | |
244 | * is like stab_find, except that it creates a new Stab_entry, | |
245 | * initialized with type = st_none, if one did not exist already | |
246 | * (it never returns NULL). | |
247 | * | |
248 | * A Stab_entry is a structure that contains at least the following | |
249 | * members: | |
250 | * | |
251 | * char *name; // must not be modified | |
252 | * enum sym_type type; // should be set | |
253 | * | |
254 | * The type field is initially set to st_none; it should be set to | |
255 | * something else by the caller of stab_search. Other possible values | |
256 | * of an enum sym_type can be added. | |
257 | */ | |
258 | ||
259 | Stab * | |
260 | stab_create () | |
261 | { | |
262 | Stab *sp; | |
263 | sp = xnew (1, Stab); | |
264 | *sp = NULL; /* a Stab starts out as a null Stab_entry* */ | |
265 | return sp; | |
266 | } | |
267 | ||
268 | Stab_entry * | |
269 | stab_find (stab, sym, symlen) | |
270 | Stab *stab; | |
271 | register char *sym; | |
272 | register int symlen; | |
273 | { | |
274 | register Stab_entry *se; | |
275 | for (se = *stab; se != NULL; se = se->next) | |
276 | { | |
277 | if (se->symlen == symlen && strneq (se->sym, sym, symlen)) | |
278 | return se; | |
279 | } | |
280 | ||
281 | return NULL; | |
282 | } | |
283 | ||
284 | Stab_entry * | |
285 | stab_search (stab, sym, symlen) | |
286 | register Stab *stab; | |
287 | char *sym; | |
288 | int symlen; | |
289 | { | |
290 | register Stab_entry *se; | |
291 | se = stab_find (stab, sym, symlen); | |
292 | ||
293 | if (se == NULL) | |
294 | { | |
295 | /* make a new one */ | |
296 | se = xnew (1, Stab_entry); | |
297 | se->sym = savenstr (sym, symlen); | |
298 | se->symlen = symlen; | |
299 | se->type = st_none; | |
300 | se->next = *stab; | |
301 | *stab = se; | |
302 | } | |
303 | ||
304 | return se; | |
305 | } | |
306 | ||
307 | /* | |
308 | * NAME | |
309 | * stab_type -- type of a symbol table entry | |
310 | * | |
311 | * SYNOPSIS | |
312 | * enum sym_type stab_type (Stab_entry *se); | |
313 | * | |
314 | * WARNING | |
315 | * May evaluate its argument more than once. | |
316 | */ | |
317 | ||
318 | #define stab_type(se) ((se)==NULL ? st_none : (se)->type) | |
319 | \f | |
320 | ||
321 | ||
322 | typedef int LINENO; | |
323 | ||
324 | typedef struct | |
325 | { | |
326 | char *p; | |
327 | int len; | |
c6d46f5f JB |
328 | LINENO lineno; |
329 | logical rewritten; | |
330 | } TOKEN; | |
c6d46f5f | 331 | |
13fde0cd | 332 | /* C extensions. |
c6d46f5f | 333 | */ |
13fde0cd RS |
334 | #define C_PLPL 0x00001 /* C++ */ |
335 | #define C_STAR 0x00003 /* C* */ | |
336 | #define YACC 0x10000 /* yacc file */ | |
c6d46f5f JB |
337 | |
338 | char searchar = '/'; /* use /.../ searches */ | |
339 | ||
340 | LINENO lineno; /* line number of current line */ | |
341 | long charno; /* current character number */ | |
c6d46f5f JB |
342 | |
343 | long linecharno; /* charno of start of line; not used by C, but | |
344 | * by every other language. | |
345 | */ | |
346 | ||
347 | char *curfile, /* current input file name */ | |
348 | *outfile, /* output file */ | |
349 | *white = " \f\t\n", /* white chars */ | |
13fde0cd RS |
350 | *endtk = " \t\n\"'#()[]{}=-+%*/&|^~!<>;,.:?", /* token ending chars */ |
351 | /* token starting chars */ | |
352 | *begtk = "ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz$", | |
353 | /* valid in-token chars */ | |
354 | *intk = "ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz$0123456789"; | |
c6d46f5f | 355 | |
c6d46f5f JB |
356 | int append_to_tagfile; /* -a: append to tags */ |
357 | int emacs_tags_format; /* emacs style output (no -e option any more) */ | |
358 | /* The following three default to 1 for etags, but to 0 for ctags. */ | |
359 | int typedefs; /* -t: create tags for typedefs */ | |
360 | int typedefs_and_cplusplus; /* -T: create tags for typedefs, level */ | |
361 | /* 0 struct/enum/union decls, and C++ */ | |
362 | /* member functions */ | |
363 | int constantypedefs; /* -d: create tags for C #define and enum */ | |
364 | /* constants. Default under etags. Enum */ | |
365 | /* constants not implemented. */ | |
366 | /* -D: opposite of -d. Default under ctags. */ | |
367 | int update; /* -u: update tags */ | |
368 | int vgrind_style; /* -v: create vgrind style index output */ | |
369 | int no_warnings; /* -w: suppress warnings */ | |
370 | int cxref_style; /* -x: create cxref style output */ | |
371 | int cplusplus; /* .[hc] means C++, not C */ | |
372 | int noindentypedefs; /* -S: ignore indentation in C */ | |
c6d46f5f JB |
373 | |
374 | /* Name this program was invoked with. */ | |
375 | char *progname; | |
376 | ||
4746118a JB |
377 | struct option longopts[] = { |
378 | { "append", no_argument, NULL, 'a' }, | |
379 | { "backward-search", no_argument, NULL, 'B' }, | |
380 | { "c++", no_argument, NULL, 'C' }, | |
381 | { "cxref", no_argument, NULL, 'x' }, | |
382 | { "defines", no_argument, NULL, 'd' }, | |
383 | { "forward-search", no_argument, NULL, 'F' }, | |
384 | { "help", no_argument, NULL, 'H' }, | |
385 | { "ignore-indentation", no_argument, NULL, 'S' }, | |
386 | { "include", required_argument, NULL, 'i' }, | |
387 | { "no-defines", no_argument, NULL, 'D' }, | |
388 | { "no-warn", no_argument, NULL, 'w' }, | |
389 | { "output", required_argument, NULL, 'o' }, | |
390 | { "typedefs", no_argument, NULL, 't' }, | |
391 | { "typedefs-and-c++", no_argument, NULL, 'T' }, | |
392 | { "update", no_argument, NULL, 'u' }, | |
393 | { "version", no_argument, NULL, 'V' }, | |
394 | { "vgrind", no_argument, NULL, 'v' }, | |
395 | { 0 } | |
396 | }; | |
397 | ||
c6d46f5f JB |
398 | FILE *inf, /* ioptr for current input file */ |
399 | *outf; /* ioptr for tags file */ | |
400 | ||
401 | NODE *head; /* the head of the binary tree of tags */ | |
402 | ||
403 | int permit_duplicates = 1; /* Nonzero means allow duplicate tags. */ | |
404 | ||
405 | /* A `struct linebuffer' is a structure which holds a line of text. | |
406 | `readline' reads a line from a stream into a linebuffer | |
407 | and works regardless of the length of the line. */ | |
408 | ||
409 | struct linebuffer | |
410 | { | |
411 | long size; | |
412 | char *buffer; | |
413 | }; | |
414 | ||
415 | struct linebuffer lb; /* the current line */ | |
c6d46f5f | 416 | struct linebuffer filename_lb; /* used to read in filenames */ |
13fde0cd RS |
417 | struct |
418 | { | |
419 | FILEPOS linepos; | |
420 | struct linebuffer lb; /* used by C_entries instead of lb */ | |
421 | } lbs[2]; | |
c6d46f5f | 422 | \f |
4746118a JB |
423 | void |
424 | print_version () | |
425 | { | |
426 | #ifdef CTAGS | |
427 | printf ("CTAGS "); | |
428 | #ifdef ETAGS | |
429 | printf ("and "); | |
430 | #endif | |
431 | #endif | |
432 | #ifdef ETAGS | |
433 | printf ("ETAGS "); | |
434 | #endif | |
435 | printf ("for Emacs version 19.0.\n"); | |
436 | ||
437 | exit (0); | |
438 | } | |
439 | ||
440 | void | |
441 | print_help () | |
442 | { | |
443 | printf ("These are the options accepted by %s. You may use unambiguous\n\ | |
444 | abbreviations for the long option names.\n\n", progname); | |
445 | ||
52cc7c59 JB |
446 | puts ("-a, --append\n\ |
447 | Append tag entries to existing tags file."); | |
448 | puts ("-C, --c++\n\ | |
4746118a JB |
449 | Treat files with `.c' and `.h' extensions as C++ code, not C\n\ |
450 | code. Files with `.C', `.H', `.cxx', `.hxx', or `.cc'\n\ | |
52cc7c59 JB |
451 | extensions are always assumed to be C++ code."); |
452 | fputs ("-d, --defines\n\ | |
4746118a JB |
453 | Create tag entries for #defines, too.", stdout); |
454 | ||
455 | #ifdef ETAGS | |
456 | fputs (" This is the default\n\ | |
457 | behavior.", stdout); | |
458 | #endif | |
459 | ||
460 | fputs ("\n\ | |
461 | -D, --no-defines\n\ | |
462 | Don't create tag entries for #defines.", stdout); | |
463 | ||
464 | #ifdef CTAGS | |
465 | fputs (" This is the default\n\ | |
466 | behavior.", stdout); | |
467 | #endif | |
468 | ||
469 | puts ("\n\ | |
470 | -o FILE, --output=FILE\n\ | |
471 | Write the tags to FILE.\n\ | |
472 | -S, --ignore-indentation\n\ | |
473 | Don't rely on indentation quite as much as normal. Currently,\n\ | |
474 | this means not to assume that a closing brace in the first\n\ | |
475 | column is the final brace of a function or structure\n\ | |
52cc7c59 JB |
476 | definition."); |
477 | puts ("-t, --typedefs\n\ | |
4746118a | 478 | Generate tag entries for typedefs. This is the default\n\ |
52cc7c59 JB |
479 | behavior."); |
480 | puts ("-T, --typedefs-and-c++\n\ | |
4746118a JB |
481 | Generate tag entries for typedefs, struct/enum/union tags, and\n\ |
482 | C++ member functions."); | |
483 | ||
484 | #ifdef ETAGS | |
485 | puts ("-i FILE, --include=FILE\n\ | |
486 | Include a note in tag file indicating that, when searching for\n\ | |
487 | a tag, one should also consult the tags file FILE after\n\ | |
488 | checking the current file."); | |
489 | #endif | |
490 | ||
491 | #ifdef CTAGS | |
492 | puts ("-B, --backward-search\n\ | |
493 | Write the search commands for the tag entries using '?', the\n\ | |
52cc7c59 JB |
494 | backward-search command."); |
495 | puts ("-F, --forward-search\n\ | |
4746118a | 496 | Write the search commands for the tag entries using '/', the\n\ |
52cc7c59 JB |
497 | forward-search command."); |
498 | puts ("-u, --update\n\ | |
4746118a JB |
499 | Update the tag entries for the given files, leaving tag\n\ |
500 | entries for other files in place. Currently, this is\n\ | |
501 | implemented by deleting the existing entries for the given\n\ | |
502 | files and then rewriting the new entries at the end of the\n\ | |
503 | tags file. It is often faster to simply rebuild the entire\n\ | |
52cc7c59 JB |
504 | tag file than to use this."); |
505 | puts ("-v, --vgrind\n\ | |
4746118a JB |
506 | Generates an index of items intended for human consumption,\n\ |
507 | similar to the output of vgrind. The index is sorted, and\n\ | |
52cc7c59 JB |
508 | gives the page number of each item."); |
509 | puts ("-x, --cxref\n\ | |
4746118a JB |
510 | Like --vgrind, but in the style of cxref, rather than vgrind.\n\ |
511 | The output uses line numbers instead of page numbers, but\n\ | |
512 | beyond that the differences are cosmetic; try both to see\n\ | |
52cc7c59 JB |
513 | which you like."); |
514 | puts ("-w, --no-warn\n\ | |
4746118a JB |
515 | Suppress warning messages about entries defined in multiple\n\ |
516 | files."); | |
517 | #endif | |
518 | ||
519 | puts ("-V, --version\n\ | |
520 | Print the version of the program.\n\ | |
521 | -H, --help\n\ | |
522 | Print this help message."); | |
523 | ||
524 | exit (0); | |
525 | } | |
526 | ||
527 | \f | |
c6d46f5f JB |
528 | void |
529 | main (argc, argv) | |
530 | int argc; | |
531 | char *argv[]; | |
532 | { | |
533 | char cmd[100]; | |
534 | int i; | |
1e134a5f | 535 | unsigned int nincluded_files = 0; |
72a339d7 | 536 | char **included_files = xnew (argc, char *); |
c6d46f5f JB |
537 | char *this_file; |
538 | #ifdef VMS | |
539 | char got_err; | |
540 | ||
541 | extern char *gfnames (); | |
542 | extern char *massage_name (); | |
543 | #endif | |
544 | ||
545 | progname = argv[0]; | |
546 | ||
547 | #ifndef CTAGS | |
548 | emacs_tags_format = 1; | |
549 | #else | |
550 | emacs_tags_format = 0; | |
551 | #endif | |
552 | ||
553 | /* | |
554 | * If etags, always find typedefs and structure tags. Why not? | |
555 | * Also default is to find macro constants. | |
556 | */ | |
557 | if (emacs_tags_format) | |
558 | typedefs = typedefs_and_cplusplus = constantypedefs = 1; | |
559 | ||
4746118a | 560 | for (;;) |
c6d46f5f | 561 | { |
4746118a | 562 | int opt; |
3ed79319 | 563 | opt = getopt_long (argc, argv, "aCdDo:f:StTi:BFuvxwVH", longopts, 0); |
4746118a JB |
564 | |
565 | if (opt == EOF) | |
566 | break; | |
567 | ||
568 | switch (opt) | |
c6d46f5f | 569 | { |
4746118a JB |
570 | case '\0': |
571 | /* If getopt returns '\0', then it has already processed a | |
572 | long-named option. We should do nothing. */ | |
573 | break; | |
574 | ||
575 | /* Common options. */ | |
576 | case 'a': | |
577 | append_to_tagfile++; | |
578 | break; | |
579 | case 'C': | |
580 | cplusplus = 1; | |
581 | break; | |
582 | case 'd': | |
583 | constantypedefs = 1; | |
584 | break; | |
585 | case 'D': | |
586 | constantypedefs = 0; | |
587 | break; | |
3ed79319 | 588 | case 'f': |
4746118a JB |
589 | case 'o': |
590 | if (outfile) | |
c6d46f5f | 591 | { |
4746118a | 592 | fprintf (stderr, |
3ed79319 | 593 | "%s: -%c flag may only be given once\n", progname, opt); |
c6d46f5f JB |
594 | goto usage; |
595 | } | |
4746118a JB |
596 | outfile = optarg; |
597 | break; | |
598 | case 'S': | |
599 | noindentypedefs++; | |
600 | break; | |
601 | case 't': | |
602 | typedefs++; | |
603 | break; | |
604 | case 'T': | |
605 | typedefs++; | |
606 | typedefs_and_cplusplus++; | |
607 | break; | |
608 | case 'V': | |
609 | print_version (); | |
610 | break; | |
611 | case 'H': | |
612 | print_help (); | |
613 | break; | |
614 | ||
615 | /* Etags options */ | |
616 | case 'i': | |
617 | if (!emacs_tags_format) | |
618 | goto usage; | |
619 | included_files[nincluded_files++] = optarg; | |
620 | break; | |
621 | ||
622 | /* Ctags options. */ | |
623 | case 'B': | |
624 | searchar = '?'; | |
625 | if (emacs_tags_format) | |
626 | goto usage; | |
627 | break; | |
628 | case 'F': | |
629 | searchar = '/'; | |
630 | if (emacs_tags_format) | |
631 | goto usage; | |
632 | break; | |
633 | case 'u': | |
634 | update++; | |
635 | if (emacs_tags_format) | |
636 | goto usage; | |
637 | break; | |
638 | case 'v': | |
639 | vgrind_style++; | |
640 | /*FALLTHRU*/ | |
641 | case 'x': | |
642 | cxref_style++; | |
643 | if (emacs_tags_format) | |
644 | goto usage; | |
645 | break; | |
646 | case 'w': | |
647 | no_warnings++; | |
648 | if (emacs_tags_format) | |
649 | goto usage; | |
650 | break; | |
651 | ||
652 | default: | |
653 | goto usage; | |
c6d46f5f | 654 | } |
c6d46f5f JB |
655 | } |
656 | ||
0e5ad25f | 657 | if (optind == argc && nincluded_files == 0) |
c6d46f5f | 658 | { |
4746118a JB |
659 | fprintf (stderr, "%s: No input files specified.\n", progname); |
660 | ||
c6d46f5f | 661 | usage: |
0e5ad25f | 662 | fprintf (stderr, "%s: Try `%s --help' for a complete list of options.\n", |
4746118a | 663 | progname, progname); |
c6d46f5f JB |
664 | exit (BAD); |
665 | } | |
666 | ||
667 | if (outfile == 0) | |
668 | { | |
669 | outfile = emacs_tags_format ? "TAGS" : "tags"; | |
670 | } | |
671 | ||
672 | init (); /* set up boolean "functions" */ | |
673 | ||
674 | initbuffer (&lb); | |
13fde0cd RS |
675 | initbuffer (&lbs[0].lb); |
676 | initbuffer (&lbs[1].lb); | |
c6d46f5f JB |
677 | initbuffer (&filename_lb); |
678 | /* | |
679 | * loop through files finding functions | |
680 | */ | |
681 | if (emacs_tags_format) | |
682 | { | |
683 | if (streq (outfile, "-")) | |
684 | outf = stdout; | |
685 | else | |
686 | outf = fopen (outfile, append_to_tagfile ? "a" : "w"); | |
687 | if (!outf) | |
688 | { | |
689 | perror (outfile); | |
690 | exit (1); | |
691 | } | |
692 | } | |
693 | ||
c6d46f5f | 694 | #ifdef VMS |
4746118a JB |
695 | argc -= optind; |
696 | argv += optind; | |
697 | while (gfnames (&argc, &argv, &got_err) != NULL) | |
c6d46f5f JB |
698 | { |
699 | if (got_err) | |
700 | { | |
701 | error ("Can't find file %s\n", this_file); | |
702 | argc--, argv++; | |
703 | } | |
704 | else | |
705 | { | |
706 | this_file = massage_name (this_file); | |
707 | #if 0 | |
708 | } | |
13fde0cd | 709 | } /* solely to balance out the ifdef'd parens above */ |
c6d46f5f JB |
710 | #endif |
711 | #else | |
4746118a | 712 | for (; optind < argc; optind++) |
c6d46f5f | 713 | { |
4746118a | 714 | this_file = argv[optind]; |
c6d46f5f JB |
715 | if (1) |
716 | { | |
717 | #endif | |
718 | /* Input file named "-" means read file names from stdin | |
719 | and use them. */ | |
720 | if (streq (this_file, "-")) | |
721 | { | |
722 | while (!feof (stdin)) | |
723 | { | |
724 | (void) readline (&filename_lb, stdin); | |
725 | if (strlen (filename_lb.buffer) > 0) | |
726 | process_file (filename_lb.buffer); | |
727 | } | |
728 | } | |
729 | else | |
730 | process_file (this_file); | |
731 | } | |
732 | } | |
733 | ||
734 | if (emacs_tags_format) | |
735 | { | |
1e134a5f RM |
736 | while (nincluded_files-- > 0) |
737 | fprintf (outf, "\f\n%s,include\n", *included_files++); | |
738 | ||
c6d46f5f JB |
739 | (void) fclose (outf); |
740 | exit (0); | |
741 | } | |
742 | ||
743 | if (cxref_style) | |
744 | { | |
745 | put_entries (head); | |
746 | exit (GOOD); | |
747 | } | |
4746118a | 748 | if (update) |
c6d46f5f | 749 | { |
4746118a JB |
750 | /* update cannot be set under VMS, so we may assume that argc |
751 | and argv have not been munged. */ | |
752 | for (i = optind; i < argc; i++) | |
c6d46f5f JB |
753 | { |
754 | sprintf (cmd, | |
755 | "mv %s OTAGS;fgrep -v '\t%s\t' OTAGS >%s;rm OTAGS", | |
756 | outfile, argv[i], outfile); | |
757 | (void) system (cmd); | |
758 | } | |
759 | append_to_tagfile++; | |
760 | } | |
761 | outf = fopen (outfile, append_to_tagfile ? "a" : "w"); | |
762 | if (outf == NULL) | |
763 | { | |
764 | perror (outfile); | |
765 | exit (GOOD); | |
766 | } | |
767 | put_entries (head); | |
768 | (void) fclose (outf); | |
769 | if (update) | |
770 | { | |
771 | sprintf (cmd, "sort %s -o %s", outfile, outfile); | |
772 | (void) system (cmd); | |
773 | } | |
774 | exit (GOOD); | |
775 | } | |
776 | ||
777 | ||
778 | /* | |
779 | * This routine is called on each file argument. | |
780 | */ | |
781 | void | |
782 | process_file (file) | |
783 | char *file; | |
784 | { | |
785 | struct stat stat_buf; | |
786 | ||
787 | stat (file, &stat_buf); | |
8a6c8bcf RS |
788 | if (!(stat_buf.st_mode & S_IFREG) |
789 | #ifdef S_IFLNK | |
790 | || !(stat_buf.st_mode & S_IFLNK) | |
791 | #endif | |
792 | ) | |
c6d46f5f JB |
793 | { |
794 | fprintf (stderr, "Skipping %s: it is not a regular file.\n", file); | |
795 | return; | |
796 | } | |
797 | ||
798 | if (streq (file, outfile) && !streq (outfile, "-")) | |
799 | { | |
800 | fprintf (stderr, "Skipping inclusion of %s in self.\n", file); | |
801 | return; | |
802 | } | |
c6d46f5f JB |
803 | if (emacs_tags_format) |
804 | { | |
8a6c8bcf | 805 | char *cp = etags_rindex (file, '/'); |
c6d46f5f JB |
806 | if (cp) |
807 | ++cp; | |
808 | else | |
809 | cp = file; | |
c6d46f5f JB |
810 | } |
811 | find_entries (file); | |
812 | if (emacs_tags_format) | |
813 | { | |
814 | fprintf (outf, "\f\n%s,%d\n", | |
815 | file, total_size_of_entries (head)); | |
816 | put_entries (head); | |
817 | free_tree (head); | |
818 | head = NULL; | |
819 | } | |
820 | } | |
821 | ||
822 | /* | |
eb8c3be9 | 823 | * This routine sets up the boolean pseudo-functions which work |
99e0a2e0 | 824 | * by setting boolean flags dependent upon the corresponding character |
c6d46f5f JB |
825 | * Every char which is NOT in that string is not a white char. Therefore, |
826 | * all of the array "_wht" is set to FALSE, and then the elements | |
827 | * subscripted by the chars in "white" are set to TRUE. Thus "_wht" | |
828 | * of a char is TRUE if it is the string "white", else FALSE. | |
829 | */ | |
830 | void | |
831 | init () | |
832 | { | |
13fde0cd RS |
833 | register char *sp; |
834 | register int i; | |
c6d46f5f JB |
835 | |
836 | for (i = 0; i < 0177; i++) | |
13fde0cd | 837 | _wht[i] = _etk[i] = _itk[i] = _btk[i] = FALSE; |
c6d46f5f JB |
838 | for (sp = white; *sp; sp++) |
839 | _wht[*sp] = TRUE; | |
840 | for (sp = endtk; *sp; sp++) | |
841 | _etk[*sp] = TRUE; | |
842 | for (sp = intk; *sp; sp++) | |
843 | _itk[*sp] = TRUE; | |
844 | for (sp = begtk; *sp; sp++) | |
845 | _btk[*sp] = TRUE; | |
c6d46f5f JB |
846 | _wht[0] = _wht['\n']; |
847 | _etk[0] = _etk['\n']; | |
848 | _btk[0] = _btk['\n']; | |
849 | _itk[0] = _itk['\n']; | |
c6d46f5f JB |
850 | } |
851 | ||
852 | /* | |
853 | * This routine opens the specified file and calls the function | |
854 | * which finds the function and type definitions. | |
855 | */ | |
856 | void | |
857 | find_entries (file) | |
858 | char *file; | |
859 | { | |
860 | char *cp; | |
861 | void prolog_funcs (); | |
862 | ||
863 | inf = fopen (file, "r"); | |
864 | if (inf == NULL) | |
865 | { | |
866 | perror (file); | |
867 | return; | |
868 | } | |
869 | curfile = savestr (file); | |
8a6c8bcf | 870 | cp = etags_rindex (file, '.'); |
c6d46f5f JB |
871 | |
872 | header_file = (cp && (streq (cp + 1, "h"))); | |
873 | ||
874 | /* .tex, .aux or .bbl implies LaTeX source code */ | |
875 | if (cp && (streq (cp + 1, "tex") || streq (cp + 1, "aux") | |
876 | || streq (cp + 1, "bbl"))) | |
877 | { | |
878 | TEX_funcs (inf); | |
879 | goto close_and_return; | |
880 | } | |
881 | /* .l or .el or .lisp (or .cl or .clisp or ...) implies lisp source code */ | |
882 | if (cp && (streq (cp + 1, "l") | |
883 | || streq (cp + 1, "el") | |
884 | || streq (cp + 1, "lsp") | |
885 | || streq (cp + 1, "lisp") | |
886 | || streq (cp + 1, "cl") | |
887 | || streq (cp + 1, "clisp"))) | |
888 | { | |
889 | L_funcs (inf); | |
890 | goto close_and_return; | |
891 | } | |
892 | /* .scm or .sm or .scheme or ... implies scheme source code */ | |
893 | if (cp && (streq (cp + 1, "sm") | |
894 | || streq (cp + 1, "scm") | |
895 | || streq (cp + 1, "scheme") | |
896 | || streq (cp + 1, "t") | |
897 | || streq (cp + 1, "sch") | |
898 | || streq (cp + 1, "SM") | |
899 | || streq (cp + 1, "SCM") | |
900 | /* The `SCM' or `scm' prefix with a version number */ | |
901 | || (cp[-1] == 'm' && cp[-2] == 'c' && cp[-3] == 's' | |
902 | && string_numeric_p (cp + 1)) | |
903 | || (cp[-1] == 'M' && cp[-2] == 'C' && cp[-3] == 'S' | |
904 | && string_numeric_p (cp + 1)))) | |
905 | { | |
906 | Scheme_funcs (inf); | |
907 | fclose (inf); | |
908 | return; | |
909 | } | |
13fde0cd RS |
910 | /* Assume that ".s" or ".a" is assembly code. -wolfgang. |
911 | Or even ".sa". */ | |
912 | if (cp && (streq (cp + 1, "s") | |
913 | || streq (cp + 1, "a") | |
914 | || streq (cp + 1, "sa"))) | |
c6d46f5f JB |
915 | { |
916 | Asm_funcs (inf); | |
917 | fclose (inf); | |
918 | return; | |
919 | } | |
920 | /* .C or .H or .cxx or .hxx or .cc: a C++ file */ | |
921 | if (cp && (streq (cp + 1, "C") | |
922 | || streq (cp + 1, "H") | |
923 | || streq (cp + 1, "cxx") | |
924 | || streq (cp + 1, "hxx") | |
925 | || streq (cp + 1, "cc"))) | |
926 | { | |
927 | C_entries (C_PLPL); /* C++ */ | |
928 | goto close_and_return; | |
929 | } | |
930 | /* .cs or .hs: a C* file */ | |
13fde0cd RS |
931 | if (cp && (streq (cp + 1, "cs") |
932 | || streq (cp + 1, "hs"))) | |
c6d46f5f JB |
933 | { |
934 | C_entries (C_STAR); | |
935 | goto close_and_return; | |
936 | } | |
13fde0cd RS |
937 | /* .y: a yacc file */ |
938 | if (cp && (streq (cp + 1, "y"))) | |
939 | { | |
940 | C_entries (YACC); | |
941 | goto close_and_return; | |
942 | } | |
c6d46f5f | 943 | /* .pl implies prolog source code */ |
13fde0cd | 944 | if (cp && streq (cp + 1, "pl")) |
c6d46f5f JB |
945 | { |
946 | prolog_funcs (inf); | |
947 | goto close_and_return; | |
948 | } | |
949 | /* .p or .pas: a Pascal file */ | |
950 | if (cp && (streq (cp + 1, "p") | |
951 | || streq (cp + 1, "pas"))) | |
952 | { | |
953 | PAS_funcs (inf); | |
954 | goto close_and_return; | |
955 | } | |
1e8e1162 | 956 | /* If .f or .for, assume it is fortran or nothing. */ |
13fde0cd RS |
957 | if (cp && (streq (cp + 1, "f") |
958 | || streq (cp + 1, "for"))) | |
1e8e1162 RS |
959 | { |
960 | PF_funcs (inf); | |
961 | goto close_and_return; | |
962 | } | |
c6d46f5f | 963 | /* if not a .c or .h or .y file, try fortran */ |
1e8e1162 RS |
964 | if (cp && ((cp[1] != 'c' |
965 | && cp[1] != 'h' | |
966 | && cp[1] != 'y') | |
967 | || (cp[1] != 0 && cp[2] != 0))) | |
c6d46f5f JB |
968 | { |
969 | if (PF_funcs (inf) != 0) | |
970 | goto close_and_return; | |
971 | rewind (inf); /* no fortran tags found, try C */ | |
972 | } | |
973 | C_entries (cplusplus ? C_PLPL : 0); | |
974 | ||
975 | close_and_return: | |
976 | (void) fclose (inf); | |
977 | } | |
978 | ||
979 | /* Nonzero if string STR is composed of digits. */ | |
980 | ||
981 | int | |
982 | string_numeric_p (str) | |
983 | char *str; | |
984 | { | |
985 | while (*str) | |
986 | { | |
987 | if (*str < '0' || *str > '9') | |
988 | return 0; | |
989 | } | |
990 | return 1; | |
991 | } | |
992 | \f | |
993 | /* Record a tag. */ | |
994 | /* Should take a TOKEN* instead!! */ | |
c6d46f5f JB |
995 | void |
996 | pfnote (name, is_func, rewritten, linestart, linelen, lno, cno) | |
997 | char *name; /* tag name */ | |
998 | logical is_func; /* function or type name? */ | |
999 | logical rewritten; /* tag different from text of definition? */ | |
1000 | char *linestart; | |
1001 | int linelen; | |
1002 | int lno; | |
1003 | long cno; | |
1004 | { | |
1005 | register char *fp; | |
1006 | register NODE *np; | |
1007 | char tem[51]; | |
1008 | char c; | |
1009 | ||
1010 | np = (NODE *) malloc (sizeof (NODE)); | |
1011 | if (np == NULL) | |
1012 | { | |
1013 | if (!emacs_tags_format) | |
1014 | { | |
1015 | /* It's okay to output early in etags -- it only disrupts the | |
1016 | * character count of the tag entries, which is no longer used | |
1017 | * by tags.el anyway. | |
1018 | */ | |
1019 | error ("too many entries to sort"); | |
1020 | } | |
1021 | put_entries (head); | |
1022 | free_tree (head); | |
1023 | head = NULL; | |
1024 | np = xnew (1, NODE); | |
1025 | } | |
1026 | /* If ctags mode, change name "main" to M<thisfilename>. */ | |
1027 | if (!emacs_tags_format && !cxref_style && streq (name, "main")) | |
1028 | { | |
8a6c8bcf | 1029 | fp = etags_rindex (curfile, '/'); |
c6d46f5f | 1030 | name = concat ("M", fp == 0 ? curfile : fp + 1, ""); |
8a6c8bcf | 1031 | fp = etags_rindex (name, '.'); |
c6d46f5f JB |
1032 | if (fp && fp[1] != '\0' && fp[2] == '\0') |
1033 | *fp = 0; | |
1034 | rewritten = TRUE; | |
1035 | } | |
1036 | np->name = savestr (name); | |
1037 | np->file = curfile; | |
1038 | np->is_func = is_func; | |
1039 | np->rewritten = rewritten; | |
1040 | np->lno = lno; | |
1041 | /* UNCOMMENT THE +1 HERE: */ | |
1042 | np->cno = cno /* + 1 */ ; /* our char numbers are 0-base; emacs's are 1-base */ | |
1043 | np->left = np->right = 0; | |
1044 | if (emacs_tags_format) | |
1045 | { | |
1046 | c = linestart[linelen]; | |
1047 | linestart[linelen] = 0; | |
1048 | } | |
1049 | else if (cxref_style == 0) | |
1050 | { | |
1051 | sprintf (tem, strlen (linestart) < 50 ? "%s$" : "%.50s", linestart); | |
1052 | linestart = tem; | |
1053 | } | |
1054 | np->pat = savestr (linestart); | |
1055 | if (emacs_tags_format) | |
1056 | { | |
1057 | linestart[linelen] = c; | |
1058 | } | |
1059 | ||
1060 | add_node (np, &head); | |
1061 | } | |
1062 | ||
1063 | /* | |
1064 | * free_tree () | |
1065 | * recurse on left children, iterate on right children. | |
1066 | */ | |
1067 | void | |
1068 | free_tree (node) | |
1069 | register NODE *node; | |
1070 | { | |
1071 | while (node) | |
1072 | { | |
1073 | register NODE *node_right = node->right; | |
1074 | free_tree (node->left); | |
1075 | free (node->name); | |
1076 | free (node->pat); | |
1077 | free ((char *) node); | |
1078 | node = node_right; | |
1079 | } | |
1080 | } | |
1081 | ||
1082 | /* | |
1083 | * add_node () | |
1084 | * Adds a node to the tree of nodes. In etags mode, we don't keep | |
1085 | * it sorted; we just keep a linear list. In ctags mode, maintain | |
1086 | * an ordered tree, with no attempt at balancing. | |
1087 | * | |
1088 | * add_node is the only function allowed to add nodes, so it can | |
1089 | * maintain state. | |
1090 | */ | |
8f53e1ee RS |
1091 | /* Must avoid static vars within functions since some systems |
1092 | #define static as nothing. */ | |
1093 | static NODE *last_node = NULL; | |
1094 | ||
c6d46f5f JB |
1095 | void |
1096 | add_node (node, cur_node_p) | |
1097 | NODE *node, **cur_node_p; | |
1098 | { | |
1099 | register int dif; | |
1100 | register NODE *cur_node = *cur_node_p; | |
c6d46f5f JB |
1101 | |
1102 | if (cur_node == NULL) | |
1103 | { | |
1104 | *cur_node_p = node; | |
1105 | last_node = node; | |
1106 | return; | |
1107 | } | |
1108 | ||
1109 | if (emacs_tags_format) | |
1110 | { | |
1111 | /* Etags Mode */ | |
1112 | if (!last_node) | |
1113 | fatal ("internal error in add_node"); | |
1114 | last_node->right = node; | |
1115 | last_node = node; | |
1116 | } | |
1117 | else | |
1118 | { | |
1119 | /* Ctags Mode */ | |
1120 | dif = strcmp (node->name, cur_node->name); | |
1121 | ||
1122 | /* | |
1123 | * If this tag name matches an existing one, then | |
1124 | * do not add the node, but maybe print a warning. | |
1125 | */ | |
1126 | if (!dif) | |
1127 | { | |
1128 | if (node->file == cur_node->file) | |
1129 | { | |
1130 | if (!no_warnings) | |
1131 | { | |
1132 | fprintf (stderr, "Duplicate entry in file %s, line %d: %s\n", | |
1133 | node->file, lineno, node->name); | |
1134 | fprintf (stderr, "Second entry ignored\n"); | |
1135 | } | |
1136 | return; | |
1137 | } | |
1138 | if (!cur_node->been_warned && !no_warnings) | |
1139 | { | |
1140 | fprintf (stderr, | |
1141 | "Duplicate entry in files %s and %s: %s (Warning only)\n", | |
1142 | node->file, cur_node->file, node->name); | |
1143 | } | |
1144 | cur_node->been_warned = TRUE; | |
1145 | return; | |
1146 | } | |
1147 | ||
1148 | /* Maybe refuse to add duplicate nodes. */ | |
1149 | if (!permit_duplicates) | |
1150 | { | |
1151 | if (!strcmp (node->name, cur_node->name) | |
1152 | && !strcmp (node->file, cur_node->file)) | |
1153 | return; | |
1154 | } | |
1155 | ||
1156 | /* Actually add the node */ | |
1157 | add_node (node, dif < 0 ? &cur_node->left : &cur_node->right); | |
1158 | } | |
1159 | } | |
1160 | \f | |
1161 | void | |
1162 | put_entries (node) | |
13fde0cd | 1163 | register NODE *node; |
c6d46f5f | 1164 | { |
13fde0cd | 1165 | register char *sp; |
c6d46f5f JB |
1166 | |
1167 | if (node == NULL) | |
1168 | return; | |
1169 | ||
1170 | /* Output subentries that precede this one */ | |
1171 | put_entries (node->left); | |
1172 | ||
1173 | /* Output this entry */ | |
1174 | ||
1175 | if (emacs_tags_format) | |
1176 | { | |
1177 | if (node->rewritten) | |
1178 | { | |
1179 | fprintf (outf, "%s\177%s\001%d,%d\n", | |
cc6d6e58 RM |
1180 | node->pat, node->name, |
1181 | node->lno, node->cno); | |
c6d46f5f JB |
1182 | } |
1183 | else | |
1184 | { | |
1185 | fprintf (outf, "%s\177%d,%d\n", | |
cc6d6e58 RM |
1186 | node->pat, |
1187 | node->lno, node->cno); | |
c6d46f5f JB |
1188 | } |
1189 | } | |
1190 | else if (!cxref_style) | |
1191 | { | |
1192 | fprintf (outf, "%s\t%s\t", | |
1193 | node->name, node->file); | |
1194 | ||
1195 | if (node->is_func) | |
1196 | { /* a function */ | |
1197 | putc (searchar, outf); | |
1198 | putc ('^', outf); | |
1199 | ||
1200 | for (sp = node->pat; *sp; sp++) | |
1201 | { | |
1202 | if (*sp == '\\' || *sp == searchar) | |
1203 | putc ('\\', outf); | |
1204 | putc (*sp, outf); | |
1205 | } | |
1206 | putc (searchar, outf); | |
1207 | } | |
1208 | else | |
1209 | { /* a typedef; text pattern inadequate */ | |
1210 | fprintf (outf, "%d", node->lno); | |
1211 | } | |
1212 | putc ('\n', outf); | |
1213 | } | |
1214 | else if (vgrind_style) | |
1215 | fprintf (stdout, "%s %s %d\n", | |
1216 | node->name, node->file, (node->lno + 63) / 64); | |
1217 | else | |
daa37602 | 1218 | fprintf (stdout, "%-16s %3d %-16s %s\n", |
c6d46f5f JB |
1219 | node->name, node->lno, node->file, node->pat); |
1220 | ||
1221 | /* Output subentries that follow this one */ | |
1222 | put_entries (node->right); | |
1223 | } | |
1224 | ||
1225 | /* Length of a number's decimal representation. */ | |
1226 | int | |
1227 | number_len (num) | |
1228 | long num; | |
1229 | { | |
1230 | int len = 0; | |
1231 | if (!num) | |
1232 | return 1; | |
1233 | for (; num; num /= 10) | |
1234 | ++len; | |
1235 | return len; | |
1236 | } | |
1237 | ||
1238 | /* | |
1239 | * Return total number of characters that put_entries will output for | |
1240 | * the nodes in the subtree of the specified node. Works only if emacs_tags_format | |
1241 | * is set, but called only in that case. This count is irrelevant with | |
1242 | * the new tags.el, but is still supplied for backward compatibility. | |
1243 | */ | |
1244 | int | |
1245 | total_size_of_entries (node) | |
13fde0cd | 1246 | register NODE *node; |
c6d46f5f | 1247 | { |
13fde0cd | 1248 | register int total; |
c6d46f5f JB |
1249 | |
1250 | if (node == NULL) | |
1251 | return 0; | |
1252 | ||
1253 | total = 0; | |
1254 | for (; node; node = node->right) | |
1255 | { | |
1256 | /* Count left subentries. */ | |
1257 | total += total_size_of_entries (node->left); | |
1258 | ||
1259 | /* Count this entry */ | |
1260 | total += strlen (node->pat) + 1; | |
1261 | total += number_len ((long) node->lno) + 1 + number_len (node->cno) + 1; | |
1262 | if (node->rewritten) | |
1263 | total += 1 + strlen (node->name); /* \001name */ | |
1264 | } | |
1265 | ||
1266 | return total; | |
1267 | } | |
1268 | \f | |
1269 | /* | |
1270 | * The C symbol tables. | |
1271 | */ | |
1272 | ||
1273 | Stab *C_stab, *C_PLPL_stab, *C_STAR_stab; | |
1274 | ||
1275 | /* | |
1276 | * SYNOPSIS | |
1277 | * Stab *get_C_stab (int c_ext); | |
1278 | */ | |
13fde0cd RS |
1279 | #define get_C_stab(c_ext) ((c_ext & C_STAR) ? C_STAR_stab : \ |
1280 | (c_ext & C_PLPL) ? C_PLPL_stab : \ | |
c6d46f5f JB |
1281 | C_stab) |
1282 | ||
1283 | void | |
1284 | add_keyword (stab, sym, type) | |
1285 | Stab *stab; | |
1286 | char *sym; | |
1287 | enum sym_type type; | |
1288 | { | |
1289 | stab_search (stab, sym, strlen (sym))->type = type; | |
1290 | } | |
1291 | ||
1292 | Stab * | |
1293 | C_create_stab (c_ext) | |
1294 | int c_ext; | |
1295 | { | |
1296 | Stab *stab; | |
1297 | ||
1298 | stab = stab_create (); | |
1299 | ||
1300 | /* C, C++ and C* */ | |
1301 | if (c_ext & C_PLPL) | |
1302 | add_keyword (stab, "class", st_C_struct); | |
1303 | if (c_ext & C_STAR) | |
1304 | add_keyword (stab, "domain", st_C_struct); | |
1305 | add_keyword (stab, "union", st_C_struct); | |
1306 | add_keyword (stab, "struct", st_C_struct); | |
1307 | add_keyword (stab, "enum", st_C_enum); | |
1308 | add_keyword (stab, "typedef", st_C_typedef); | |
1309 | add_keyword (stab, "define", st_C_define); | |
1310 | add_keyword (stab, "long", st_C_typespec); | |
1311 | add_keyword (stab, "short", st_C_typespec); | |
1312 | add_keyword (stab, "int", st_C_typespec); | |
1313 | add_keyword (stab, "char", st_C_typespec); | |
1314 | add_keyword (stab, "float", st_C_typespec); | |
1315 | add_keyword (stab, "double", st_C_typespec); | |
1316 | add_keyword (stab, "signed", st_C_typespec); | |
1317 | add_keyword (stab, "unsigned", st_C_typespec); | |
13fde0cd RS |
1318 | add_keyword (stab, "auto", st_C_typespec); |
1319 | add_keyword (stab, "void", st_C_typespec); | |
1320 | add_keyword (stab, "extern", st_C_typespec); | |
1321 | add_keyword (stab, "static", st_C_typespec); | |
c6d46f5f JB |
1322 | add_keyword (stab, "const", st_C_typespec); |
1323 | add_keyword (stab, "volatile", st_C_typespec); | |
1324 | ||
1325 | return stab; | |
1326 | } | |
1327 | ||
1328 | void | |
1329 | C_create_stabs () | |
1330 | { | |
1331 | C_stab = C_create_stab (0); | |
1332 | C_PLPL_stab = C_create_stab (C_PLPL); | |
1333 | C_STAR_stab = C_create_stab (C_STAR | C_PLPL); | |
1334 | } | |
1335 | \f | |
13fde0cd | 1336 | /* |
13fde0cd RS |
1337 | * C functions are recognized using a simple finite automaton. |
1338 | * funcdef is its state variable. | |
1339 | */ | |
1340 | typedef enum | |
1341 | { | |
31d4b314 FP |
1342 | fnone, /* nothing seen */ |
1343 | ftagseen, /* function-like tag seen */ | |
1344 | finlist, /* in parameter list */ | |
1345 | flistseen, /* after parameter list */ | |
1346 | fignore, /* before open brace */ | |
13fde0cd RS |
1347 | } FUNCST; |
1348 | FUNCST funcdef; | |
1349 | ||
1350 | ||
1351 | /* typedefs are recognized using a simple finite automaton. | |
1352 | * typeddef is its state variable. | |
1353 | */ | |
1354 | typedef enum | |
1355 | { | |
31d4b314 FP |
1356 | tnone, /* nothing seen */ |
1357 | ttypedseen, /* typedef keyword seen */ | |
1358 | tinbody, /* inside typedef body */ | |
1359 | tend, /* just before typedef tag */ | |
13fde0cd RS |
1360 | } TYPEDST; |
1361 | TYPEDST typdef; | |
1362 | ||
1363 | ||
1364 | /* struct tags for C++ are recognized using another simple | |
1365 | * finite automaton. `structdef' is its state variable. | |
1366 | * This machinery is only invoked for C++; otherwise structdef | |
1367 | * should remain snone. However, this machinery can easily be | |
1368 | * adapted to find structure tags in normal C code. | |
1369 | */ | |
1370 | typedef enum | |
1371 | { | |
1372 | snone, /* nothing seen yet */ | |
1373 | skeyseen, /* struct-like keyword seen */ | |
1374 | stagseen, /* struct-like tag seen */ | |
1375 | scolonseen, /* colon seen after struct-like tag */ | |
31d4b314 | 1376 | sinbody, /* in struct body: recognize member func defs*/ |
13fde0cd RS |
1377 | } STRUCTST; |
1378 | STRUCTST structdef; | |
1379 | /* | |
1380 | * When structdef is stagseen, scolonseen, or sinbody, structtag is the | |
1381 | * struct tag, and structkey is the preceding struct-like keyword. | |
1382 | */ | |
1383 | char structtag[BUFSIZ]; | |
1384 | Stab_entry *structkey; | |
1385 | ||
1386 | /* | |
1387 | * Yet another little state machine to deal with preprocessor lines. | |
1388 | */ | |
1389 | typedef enum | |
1390 | { | |
1391 | dnone, /* nothing seen */ | |
1392 | dsharpseen, /* '#' seen as first char on line */ | |
1393 | ddefineseen, /* '#' and 'define' seen */ | |
31d4b314 | 1394 | dignorerest, /* ignore rest of line */ |
13fde0cd RS |
1395 | } DEFINEST; |
1396 | DEFINEST definedef; | |
1397 | ||
1398 | /* | |
1399 | * Set this to TRUE, and the next token considered is called a function. | |
1400 | * Used only for GNUmacs's function-defining macros. | |
1401 | */ | |
1402 | logical next_token_is_func; | |
1403 | ||
1404 | /* | |
1405 | * TRUE in the rules part of a yacc file, FALSE outside (parse as C). | |
1406 | */ | |
1407 | logical yacc_rules; | |
1408 | ||
c6d46f5f JB |
1409 | /* |
1410 | * C_entries () | |
13fde0cd RS |
1411 | * This routine finds functions, typedefs, #define's and |
1412 | * struct/union/enum definitions in C syntax and adds them | |
c6d46f5f JB |
1413 | * to the list. |
1414 | */ | |
1415 | ||
13fde0cd RS |
1416 | #define curlb (lbs[curndx].lb) |
1417 | #define othlb (lbs[1-curndx].lb) | |
1418 | #define newlb (lbs[newndx].lb) | |
1419 | #define curlinepos (lbs[curndx].linepos) | |
1420 | #define othlinepos (lbs[1-curndx].linepos) | |
1421 | #define newlinepos (lbs[newndx].linepos) | |
1422 | ||
c6d46f5f | 1423 | #define CNL_SAVE_DEFINEDEF \ |
13fde0cd RS |
1424 | do { \ |
1425 | SET_FILEPOS (curlinepos, inf, charno); \ | |
c6d46f5f | 1426 | lineno++; \ |
13fde0cd RS |
1427 | charno += readline (&curlb, inf); \ |
1428 | lp = curlb.buffer; \ | |
1429 | quotednl = FALSE; \ | |
1430 | newndx = curndx; \ | |
1431 | } while (FALSE) | |
c6d46f5f JB |
1432 | |
1433 | #define CNL \ | |
13fde0cd | 1434 | do { \ |
c6d46f5f JB |
1435 | CNL_SAVE_DEFINEDEF; \ |
1436 | definedef = dnone; \ | |
13fde0cd RS |
1437 | } while (FALSE) |
1438 | ||
1439 | #define MAKE_TAG_FROM_NEW_LB(isfun) pfnote (tokb, isfun, tok.rewritten, \ | |
1440 | newlb.buffer, tokoff + toklen + 1, tok.lineno, GET_CHARNO (newlinepos)) | |
1441 | #define MAKE_TAG_FROM_OTH_LB(isfun) pfnote (tokb, isfun, tok.rewritten, \ | |
1442 | othlb.buffer, tokoff + toklen + 1, tok.lineno, GET_CHARNO (othlinepos)) | |
c6d46f5f JB |
1443 | |
1444 | void | |
1445 | C_entries (c_ext) | |
1446 | int c_ext; /* extension of C? */ | |
1447 | { | |
13fde0cd | 1448 | register char c; /* latest char read; '\0' for end of line */ |
c6d46f5f | 1449 | register char *lp; /* pointer one beyond the character `c' */ |
13fde0cd RS |
1450 | int curndx, newndx; /* indices for current and new lb */ |
1451 | TOKEN tok; /* latest token read for funcdef & structdef */ | |
1452 | char tokb[BUFSIZ]; /* latest token name for funcdef & structdef */ | |
1453 | register int tokoff; /* offset in line of start of latest token */ | |
1454 | register int toklen; /* length of latest token */ | |
591fa824 | 1455 | int cblev; /* current curly brace level */ |
13fde0cd RS |
1456 | logical incomm, inquote, inchar, quotednl, midtoken; |
1457 | logical cplpl; | |
c6d46f5f | 1458 | |
13fde0cd | 1459 | curndx = newndx = 0; |
c6d46f5f JB |
1460 | lineno = 0; |
1461 | charno = 0; | |
13fde0cd | 1462 | lp = curlb.buffer; |
c6d46f5f JB |
1463 | *lp = 0; |
1464 | ||
13fde0cd RS |
1465 | definedef = dnone; funcdef = fnone; typdef= tnone; structdef= snone; |
1466 | next_token_is_func = yacc_rules = FALSE; | |
1467 | midtoken = inquote = inchar = incomm = quotednl = FALSE; | |
591fa824 | 1468 | cblev = 0; |
13fde0cd | 1469 | cplpl = c_ext & C_PLPL; |
c6d46f5f JB |
1470 | |
1471 | C_create_stabs (); | |
1472 | ||
1473 | while (!feof (inf)) | |
1474 | { | |
1475 | c = *lp++; | |
c6d46f5f JB |
1476 | if (c == '\\') |
1477 | { | |
4746118a JB |
1478 | /* If we're at the end of the line, the next character is a |
1479 | '\0'; don't skip it, because it's the thing that tells us | |
1480 | to read the next line. */ | |
13fde0cd | 1481 | if (*lp == '\0') |
99e0a2e0 | 1482 | { |
13fde0cd | 1483 | quotednl = TRUE; |
99e0a2e0 RS |
1484 | continue; |
1485 | } | |
1e134a5f | 1486 | lp++; |
c6d46f5f JB |
1487 | c = ' '; |
1488 | } | |
1489 | else if (incomm) | |
1490 | { | |
13fde0cd | 1491 | switch (c) |
c6d46f5f | 1492 | { |
13fde0cd RS |
1493 | case '*': |
1494 | if (*lp == '/') | |
1495 | { | |
1496 | c = *lp++; | |
1497 | incomm = FALSE; | |
1498 | } | |
1499 | break; | |
1500 | case '\0': | |
1501 | /* Newlines inside comments do not end macro definitions in | |
1502 | traditional cpp. */ | |
1503 | CNL_SAVE_DEFINEDEF; | |
1504 | break; | |
c6d46f5f | 1505 | } |
13fde0cd | 1506 | continue; |
c6d46f5f JB |
1507 | } |
1508 | else if (inquote) | |
1509 | { | |
13fde0cd RS |
1510 | switch (c) |
1511 | { | |
1512 | case '"': | |
1513 | inquote = FALSE; | |
1514 | break; | |
1515 | case '\0': | |
1516 | /* Newlines inside strings, do not end macro definitions | |
1517 | in traditional cpp, even though compilers don't | |
1518 | usually accept them. */ | |
1519 | CNL_SAVE_DEFINEDEF; | |
1520 | break; | |
1521 | } | |
1522 | continue; | |
c6d46f5f JB |
1523 | } |
1524 | else if (inchar) | |
1525 | { | |
1526 | if (c == '\'') | |
1527 | inchar = FALSE; | |
1528 | continue; | |
1529 | } | |
13fde0cd | 1530 | else |
c6d46f5f JB |
1531 | switch (c) |
1532 | { | |
1533 | case '"': | |
1534 | inquote = TRUE; | |
1535 | continue; | |
1536 | case '\'': | |
1537 | inchar = TRUE; | |
1538 | continue; | |
1539 | case '/': | |
1540 | if (*lp == '*') | |
1541 | { | |
1542 | lp++; | |
1543 | incomm = TRUE; | |
13fde0cd | 1544 | continue; |
c6d46f5f | 1545 | } |
13fde0cd | 1546 | else if (cplpl && *lp == '/') |
c6d46f5f | 1547 | { |
daa37602 JB |
1548 | c = 0; |
1549 | break; | |
c6d46f5f JB |
1550 | } |
1551 | continue; | |
13fde0cd RS |
1552 | case '%': |
1553 | if ((c_ext & YACC) && *lp == '%') | |
1554 | { | |
1555 | /* entering or exiting rules section in yacc file */ | |
1556 | lp++; | |
1557 | definedef = dnone; funcdef = fnone; | |
1558 | typdef= tnone; structdef= snone; | |
1559 | next_token_is_func = FALSE; | |
1560 | midtoken = inquote = inchar = incomm = quotednl = FALSE; | |
591fa824 | 1561 | cblev = 0; |
13fde0cd RS |
1562 | yacc_rules = !yacc_rules; |
1563 | continue; | |
591fa824 | 1564 | } |
c6d46f5f | 1565 | case '#': |
13fde0cd | 1566 | if (lp == newlb.buffer + 1 && definedef == dnone) |
c6d46f5f JB |
1567 | definedef = dsharpseen; |
1568 | continue; | |
13fde0cd | 1569 | } /* switch (c) */ |
c6d46f5f | 1570 | |
c6d46f5f | 1571 | |
591fa824 RS |
1572 | /* Consider token only if some complicated conditions are satisfied. */ |
1573 | if (((cblev == 0 && structdef != scolonseen) | |
1574 | || (cblev == 1 && cplpl && structdef == sinbody)) | |
13fde0cd | 1575 | && definedef != dignorerest |
13fde0cd | 1576 | && funcdef != finlist) |
c6d46f5f JB |
1577 | { |
1578 | if (midtoken) | |
1579 | { | |
1580 | if (endtoken (c)) | |
1581 | { | |
13fde0cd | 1582 | if (cplpl && c == ':' && *lp == ':' && intoken (*(lp + 1))) |
c6d46f5f JB |
1583 | { |
1584 | /* | |
1585 | * This handles :: in the middle, but not at beginning | |
1586 | * of an identifier. | |
1587 | */ | |
1588 | lp += 2; | |
1589 | toklen += 3; | |
1590 | } | |
1591 | else | |
1592 | { | |
c6d46f5f | 1593 | logical is_func; |
c6d46f5f | 1594 | |
13fde0cd RS |
1595 | tok.lineno = lineno; |
1596 | tok.p = newlb.buffer + tokoff; | |
c6d46f5f JB |
1597 | tok.len = toklen; |
1598 | tok.rewritten = FALSE; | |
13fde0cd RS |
1599 | if (yacc_rules |
1600 | || consider_token (c, lp, &tok, | |
591fa824 | 1601 | c_ext, cblev, &is_func)) |
c6d46f5f | 1602 | { |
99e0a2e0 RS |
1603 | if (structdef == sinbody |
1604 | && definedef == dnone && is_func) | |
c6d46f5f JB |
1605 | { /* function defined in C++ class body */ |
1606 | sprintf (tokb, "%s::%.*s", | |
13fde0cd RS |
1607 | ((structtag[0] == '\0') |
1608 | ? "_anonymous_" : structtag), | |
c6d46f5f JB |
1609 | tok.len, tok.p); |
1610 | tok.rewritten = TRUE; | |
1611 | } | |
1612 | else | |
1613 | { | |
1614 | sprintf (tokb, "%.*s", tok.len, tok.p); | |
1615 | } | |
13fde0cd | 1616 | |
591fa824 RS |
1617 | if (funcdef == ftagseen |
1618 | || structdef == stagseen | |
1619 | || typdef == tend) | |
13fde0cd RS |
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); | |
31d4b314 FP |
1642 | funcdef = fignore; |
1643 | break; | |
13fde0cd RS |
1644 | case ftagseen: |
1645 | funcdef = fnone; | |
1646 | break; | |
1647 | } | |
1648 | if (structdef == stagseen) | |
1649 | structdef = snone; | |
1650 | if (!yacc_rules || lp == newlb.buffer + 1) | |
1651 | { | |
1652 | tokoff = lp - 1 - newlb.buffer; | |
1653 | toklen = 1; | |
1654 | midtoken = TRUE; | |
1655 | } | |
1656 | continue; | |
c6d46f5f | 1657 | } |
13fde0cd RS |
1658 | } /* if must look at token */ |
1659 | ||
1660 | ||
1661 | /* Detect end of line, colon, comma, semicolon and various braces | |
1662 | after having handled the last token on the line.*/ | |
1663 | switch (c) | |
1e134a5f | 1664 | { |
13fde0cd RS |
1665 | case ':': |
1666 | if (structdef == stagseen) | |
1667 | structdef = scolonseen; | |
1668 | else if (yacc_rules && funcdef == ftagseen) | |
57e83cfe | 1669 | { |
13fde0cd | 1670 | MAKE_TAG_FROM_OTH_LB (FALSE); |
31d4b314 | 1671 | funcdef = fignore; |
57e83cfe | 1672 | } |
13fde0cd RS |
1673 | break; |
1674 | case ';': | |
591fa824 RS |
1675 | if (cblev == 0 && typdef == tend) |
1676 | { | |
1677 | typdef = tnone; | |
1678 | MAKE_TAG_FROM_OTH_LB (FALSE); | |
1679 | } | |
31d4b314 FP |
1680 | if (funcdef != fignore) |
1681 | funcdef = fnone; | |
13fde0cd RS |
1682 | /* FALLTHRU */ |
1683 | case ',': | |
13fde0cd RS |
1684 | /* FALLTHRU */ |
1685 | case '[': | |
31d4b314 | 1686 | if (funcdef != finlist && funcdef != fignore) |
13fde0cd RS |
1687 | funcdef = fnone; |
1688 | if (structdef == stagseen) | |
1689 | structdef = snone; | |
1690 | break; | |
1691 | case '(': | |
1692 | switch (funcdef) | |
57e83cfe | 1693 | { |
13fde0cd RS |
1694 | case ftagseen: |
1695 | funcdef = finlist; | |
1696 | break; | |
1697 | case finlist: | |
1698 | case flistseen: | |
1699 | funcdef = fnone; | |
1700 | break; | |
57e83cfe | 1701 | } |
13fde0cd RS |
1702 | break; |
1703 | case ')': | |
1704 | if (funcdef == finlist) | |
1705 | funcdef = flistseen; | |
1706 | break; | |
1707 | case '{': | |
1708 | if (typdef == ttypedseen) | |
1709 | typdef = tinbody; | |
1710 | switch (structdef) | |
1711 | { | |
1712 | case skeyseen: /* unnamed struct */ | |
1713 | structtag[0] = '\0'; | |
1714 | structdef = sinbody; | |
1715 | break; | |
1716 | case stagseen: | |
1717 | case scolonseen: /* named struct */ | |
1718 | structdef = sinbody; | |
1719 | MAKE_TAG_FROM_OTH_LB (FALSE); | |
1720 | break; | |
1721 | } | |
31d4b314 FP |
1722 | switch (funcdef) |
1723 | { | |
1724 | case flistseen: | |
1725 | MAKE_TAG_FROM_OTH_LB (TRUE); | |
1726 | /* FALLTHRU */ | |
1727 | case fignore: | |
1728 | funcdef = fnone; | |
1729 | } | |
591fa824 | 1730 | cblev++; |
31d4b314 | 1731 | break; |
13fde0cd RS |
1732 | case '*': |
1733 | if (funcdef == flistseen) | |
1734 | { | |
1735 | MAKE_TAG_FROM_OTH_LB (TRUE); | |
31d4b314 | 1736 | funcdef = fignore; |
13fde0cd RS |
1737 | } |
1738 | break; | |
1739 | case '}': | |
1740 | if (!noindentypedefs && lp == newlb.buffer + 1) | |
591fa824 RS |
1741 | cblev = 0; /* reset curly brace level if first column */ |
1742 | else if (cblev > 0) | |
1743 | cblev--; | |
1744 | if (cblev == 0) | |
13fde0cd RS |
1745 | { |
1746 | if (typdef == tinbody) | |
1747 | typdef = tend; | |
1748 | structdef = snone; | |
1749 | (void) strcpy (structtag, "<error 2>"); | |
1750 | } | |
1751 | break; | |
1752 | case '\0': | |
1753 | /* If a macro spans multiple lines don't reset its state. */ | |
1754 | if (quotednl) | |
1755 | CNL_SAVE_DEFINEDEF; | |
1756 | else | |
1757 | CNL; | |
1758 | break; | |
1759 | } /* switch (c) */ | |
1760 | ||
1761 | } /* while not eof */ | |
c6d46f5f JB |
1762 | } |
1763 | ||
1764 | /* | |
1765 | * consider_token () | |
1766 | * checks to see if the current token is at the start of a | |
13fde0cd RS |
1767 | * function, or corresponds to a typedef, or is a struct/union/enum |
1768 | * tag. | |
c6d46f5f | 1769 | * |
13fde0cd | 1770 | * *IS_FUNC gets TRUE iff the token is a function or macro with args. |
c6d46f5f JB |
1771 | * C_EXT is which language we are looking at. |
1772 | * | |
1773 | * In the future we will need some way to adjust where the end of | |
1774 | * the token is; for instance, implementing the C++ keyword | |
1775 | * `operator' properly will adjust the end of the token to be after | |
1776 | * whatever follows `operator'. | |
1777 | * | |
1778 | * Globals | |
13fde0cd RS |
1779 | * funcdef IN OUT |
1780 | * structdef IN OUT | |
1781 | * definedef IN OUT | |
1782 | * typdef IN OUT | |
1783 | * next_token_is_func IN OUT | |
c6d46f5f JB |
1784 | */ |
1785 | ||
1786 | logical | |
591fa824 | 1787 | consider_token (c, lp, tokp, c_ext, cblev, is_func) |
13fde0cd RS |
1788 | register char c; /* IN: first char after the token */ |
1789 | register char *lp; /* IN: lp points to 2nd char after the token */ | |
591fa824 RS |
1790 | register TOKEN *tokp; /* IN: token pointer */ |
1791 | int c_ext; /* IN: C extensions mask */ | |
1792 | int cblev; /* IN: curly brace level */ | |
13fde0cd | 1793 | logical *is_func; /* OUT */ |
c6d46f5f | 1794 | { |
c6d46f5f JB |
1795 | logical firsttok; /* TRUE if have seen first token in ()'s */ |
1796 | Stab_entry *tokse = stab_find (get_C_stab (c_ext), tokp->p, tokp->len); | |
1797 | enum sym_type toktype = stab_type (tokse); | |
1798 | ||
13fde0cd | 1799 | *is_func = FALSE; /* not a function */ |
c6d46f5f JB |
1800 | |
1801 | /* | |
13fde0cd | 1802 | * Advance the definedef state machine. |
c6d46f5f JB |
1803 | */ |
1804 | switch (definedef) | |
1805 | { | |
1806 | case dnone: | |
1807 | /* We're not on a preprocessor line. */ | |
1808 | break; | |
1809 | case dsharpseen: | |
1810 | if (toktype == st_C_define) | |
1811 | { | |
1812 | definedef = ddefineseen; | |
c6d46f5f JB |
1813 | } |
1814 | else | |
1815 | { | |
1816 | definedef = dignorerest; | |
c6d46f5f | 1817 | } |
13fde0cd | 1818 | return (FALSE); |
c6d46f5f JB |
1819 | case ddefineseen: |
1820 | /* | |
1821 | * Make a tag for any macro. | |
c6d46f5f | 1822 | */ |
c6d46f5f | 1823 | definedef = dignorerest; |
13fde0cd | 1824 | *is_func = (c == '('); |
c6d46f5f | 1825 | if (!*is_func && !constantypedefs) |
13fde0cd RS |
1826 | return (FALSE); |
1827 | else | |
1828 | return (TRUE); | |
c6d46f5f | 1829 | case dignorerest: |
13fde0cd | 1830 | return (FALSE); |
c6d46f5f JB |
1831 | default: |
1832 | error ("internal error: definedef value"); | |
1833 | } | |
1834 | ||
1835 | /* | |
13fde0cd | 1836 | * Now typedefs |
c6d46f5f | 1837 | */ |
13fde0cd | 1838 | switch (typdef) |
c6d46f5f | 1839 | { |
13fde0cd RS |
1840 | case tnone: |
1841 | if (toktype == st_C_typedef) | |
c6d46f5f | 1842 | { |
13fde0cd RS |
1843 | if (typedefs) |
1844 | typdef = ttypedseen; | |
1845 | return (FALSE); | |
c6d46f5f | 1846 | } |
13fde0cd RS |
1847 | break; |
1848 | case ttypedseen: | |
1849 | switch (toktype) | |
c6d46f5f | 1850 | { |
13fde0cd RS |
1851 | case st_none: |
1852 | case st_C_typespec: | |
1853 | typdef = tend; | |
1854 | break; | |
1855 | case st_C_struct: | |
1856 | case st_C_enum: | |
1e134a5f | 1857 | break; |
c6d46f5f | 1858 | } |
13fde0cd | 1859 | /* Do not return here, so the structdef stuff has a chance. */ |
c6d46f5f | 1860 | break; |
13fde0cd RS |
1861 | case tend: |
1862 | switch (toktype) | |
c6d46f5f | 1863 | { |
13fde0cd RS |
1864 | case st_C_typespec: |
1865 | case st_C_struct: | |
1866 | case st_C_enum: | |
1867 | return (FALSE); | |
c6d46f5f | 1868 | } |
13fde0cd | 1869 | return (TRUE); |
c6d46f5f JB |
1870 | } |
1871 | ||
1872 | /* | |
591fa824 RS |
1873 | * This structdef business is currently only invoked when cblev==0. |
1874 | * It should be recursively invoked whatever the curly brace level, | |
1875 | * and a stack of states kept, to allow for definitions of structs | |
1876 | * within structs. | |
c6d46f5f JB |
1877 | * |
1878 | * This structdef business is NOT invoked when we are ctags and the | |
1879 | * file is plain C. This is because a struct tag may have the same | |
1880 | * name as another tag, and this loses with ctags. | |
1881 | * | |
13fde0cd RS |
1882 | * This if statement deals with the typdef state machine as |
1883 | * follows: if typdef==ttypedseen and token is struct/union/class/enum, | |
1884 | * return (FALSE). All the other code here is for the structdef | |
1885 | * state machine. | |
c6d46f5f JB |
1886 | */ |
1887 | switch (toktype) | |
1888 | { | |
1889 | case st_C_struct: | |
1890 | case st_C_enum: | |
13fde0cd | 1891 | if (typdef == ttypedseen |
591fa824 | 1892 | || (typedefs_and_cplusplus && cblev == 0 && structdef == snone)) |
c6d46f5f JB |
1893 | { |
1894 | structdef = skeyseen; | |
1895 | structkey = tokse; | |
1896 | } | |
13fde0cd | 1897 | return (FALSE); |
c6d46f5f | 1898 | } |
c6d46f5f JB |
1899 | if (structdef == skeyseen) |
1900 | { | |
13fde0cd | 1901 | if (stab_type (structkey) == st_C_struct) |
c6d46f5f | 1902 | { |
13fde0cd RS |
1903 | (void) strncpy (structtag, tokp->p, tokp->len); |
1904 | structtag[tokp->len] = '\0'; /* for struct/union/class */ | |
c6d46f5f JB |
1905 | } |
1906 | else | |
1907 | { | |
13fde0cd | 1908 | structtag[0] = '\0'; /* for enum (why is it treated differently?) */ |
c6d46f5f | 1909 | } |
13fde0cd RS |
1910 | structdef = stagseen; |
1911 | return (TRUE); | |
c6d46f5f | 1912 | } |
13fde0cd RS |
1913 | |
1914 | /* Avoid entering funcdef stuff if typdef is going on. */ | |
1915 | if (typdef != tnone) | |
c6d46f5f | 1916 | { |
13fde0cd RS |
1917 | definedef = dnone; |
1918 | return (FALSE); | |
c6d46f5f | 1919 | } |
13fde0cd | 1920 | |
c6d46f5f | 1921 | /* Detect GNUmacs's function-defining macros. */ |
4746118a | 1922 | if (definedef == dnone) |
c6d46f5f | 1923 | { |
daa37602 JB |
1924 | if (strneq (tokp->p, "DEF", 3) |
1925 | || strneq (tokp->p, "ENTRY", 5) | |
1926 | || strneq (tokp->p, "SYSCALL", 7) | |
1927 | || strneq (tokp->p, "PSEUDO", 6)) | |
4746118a JB |
1928 | { |
1929 | next_token_is_func = TRUE; | |
13fde0cd | 1930 | return (FALSE); |
4746118a | 1931 | } |
13fde0cd | 1932 | if (strneq (tokp->p, "EXFUN", 5)) |
4746118a JB |
1933 | { |
1934 | next_token_is_func = FALSE; | |
13fde0cd | 1935 | return (FALSE); |
4746118a | 1936 | } |
c6d46f5f JB |
1937 | } |
1938 | if (next_token_is_func) | |
1939 | { | |
1940 | next_token_is_func = FALSE; | |
31d4b314 FP |
1941 | funcdef = fnone; |
1942 | *is_func = TRUE; /* to force search string in ctags */ | |
13fde0cd | 1943 | return (TRUE); |
c6d46f5f | 1944 | } |
13fde0cd RS |
1945 | |
1946 | /* A function? */ | |
1947 | switch (toktype) | |
c6d46f5f | 1948 | { |
13fde0cd | 1949 | case st_C_typespec: |
591fa824 | 1950 | funcdef = fnone; /* should be useless */ |
13fde0cd RS |
1951 | return (FALSE); |
1952 | default: | |
31d4b314 FP |
1953 | if (funcdef == fnone) |
1954 | { | |
1955 | funcdef = ftagseen; | |
1956 | *is_func = TRUE; | |
1957 | return (TRUE); | |
1958 | } | |
c6d46f5f | 1959 | } |
31d4b314 FP |
1960 | |
1961 | return (FALSE); | |
c6d46f5f JB |
1962 | } |
1963 | \f | |
1964 | /* Fortran parsing */ | |
1965 | ||
1966 | char *dbp; | |
1967 | int pfcnt; | |
1968 | ||
1969 | int | |
1970 | PF_funcs (fi) | |
1971 | FILE *fi; | |
1972 | { | |
1973 | lineno = 0; | |
1974 | charno = 0; | |
1975 | pfcnt = 0; | |
1976 | ||
1977 | while (!feof (fi)) | |
1978 | { | |
1979 | lineno++; | |
1980 | linecharno = charno; | |
1981 | charno += readline (&lb, fi); | |
1982 | dbp = lb.buffer; | |
1983 | if (*dbp == '%') | |
1984 | dbp++; /* Ratfor escape to fortran */ | |
1985 | while (isspace (*dbp)) | |
1986 | dbp++; | |
1987 | if (*dbp == 0) | |
1988 | continue; | |
1989 | switch (*dbp | ' ') | |
1990 | { | |
1991 | case 'i': | |
1992 | if (tail ("integer")) | |
1993 | takeprec (); | |
1994 | break; | |
1995 | case 'r': | |
1996 | if (tail ("real")) | |
1997 | takeprec (); | |
1998 | break; | |
1999 | case 'l': | |
2000 | if (tail ("logical")) | |
2001 | takeprec (); | |
2002 | break; | |
2003 | case 'c': | |
2004 | if (tail ("complex") || tail ("character")) | |
2005 | takeprec (); | |
2006 | break; | |
2007 | case 'd': | |
2008 | if (tail ("double")) | |
2009 | { | |
2010 | while (isspace (*dbp)) | |
2011 | dbp++; | |
2012 | if (*dbp == 0) | |
2013 | continue; | |
2014 | if (tail ("precision")) | |
2015 | break; | |
2016 | continue; | |
2017 | } | |
2018 | break; | |
2019 | } | |
2020 | while (isspace (*dbp)) | |
2021 | dbp++; | |
2022 | if (*dbp == 0) | |
2023 | continue; | |
2024 | switch (*dbp | ' ') | |
2025 | { | |
2026 | case 'f': | |
2027 | if (tail ("function")) | |
2028 | getit (); | |
2029 | continue; | |
2030 | case 's': | |
2031 | if (tail ("subroutine")) | |
2032 | getit (); | |
2033 | continue; | |
8a6c8bcf RS |
2034 | case 'e': |
2035 | if (tail ("entry")) | |
2036 | getit (); | |
2037 | continue; | |
c6d46f5f JB |
2038 | case 'p': |
2039 | if (tail ("program")) | |
2040 | { | |
2041 | getit (); | |
2042 | continue; | |
2043 | } | |
2044 | if (tail ("procedure")) | |
2045 | getit (); | |
2046 | continue; | |
2047 | } | |
2048 | } | |
2049 | return (pfcnt); | |
2050 | } | |
2051 | ||
2052 | logical | |
2053 | tail (cp) | |
2054 | char *cp; | |
2055 | { | |
2056 | register int len = 0; | |
2057 | ||
2058 | while (*cp && (*cp & ~' ') == ((*(dbp + len)) & ~' ')) | |
2059 | cp++, len++; | |
2060 | if (*cp == 0) | |
2061 | { | |
2062 | dbp += len; | |
2063 | return (1); | |
2064 | } | |
2065 | return (0); | |
2066 | } | |
2067 | ||
2068 | void | |
2069 | takeprec () | |
2070 | { | |
2071 | while (isspace (*dbp)) | |
2072 | dbp++; | |
2073 | if (*dbp != '*') | |
2074 | return; | |
2075 | dbp++; | |
2076 | while (isspace (*dbp)) | |
2077 | dbp++; | |
2078 | if (!isdigit (*dbp)) | |
2079 | { | |
2080 | --dbp; /* force failure */ | |
2081 | return; | |
2082 | } | |
2083 | do | |
2084 | dbp++; | |
2085 | while (isdigit (*dbp)); | |
2086 | } | |
2087 | ||
2088 | void | |
2089 | getit () | |
2090 | { | |
2091 | register char *cp; | |
2092 | char c; | |
2093 | char nambuf[BUFSIZ]; | |
2094 | ||
2095 | while (isspace (*dbp)) | |
2096 | dbp++; | |
daa37602 JB |
2097 | if (*dbp == 0 |
2098 | || (!isalpha (*dbp) | |
2099 | && *dbp != '_' | |
2100 | && *dbp != '$')) | |
c6d46f5f JB |
2101 | return; |
2102 | for (cp = dbp + 1; *cp && (isalpha (*cp) || isdigit (*cp) | |
2103 | || (*cp == '_') || (*cp == '$')); cp++) | |
2104 | continue; | |
2105 | c = cp[0]; | |
2106 | cp[0] = 0; | |
2107 | (void) strcpy (nambuf, dbp); | |
2108 | cp[0] = c; | |
2109 | pfnote (nambuf, TRUE, FALSE, lb.buffer, cp - lb.buffer + 1, lineno, linecharno); | |
2110 | pfcnt++; | |
2111 | } | |
2112 | ||
2113 | /* Handle a file of assembler code. */ | |
2114 | ||
2115 | void | |
2116 | Asm_funcs (fi) | |
2117 | FILE *fi; | |
2118 | { | |
2119 | int i; | |
2120 | register char c; | |
2121 | ||
2122 | lineno = 0; | |
2123 | charno = 0; | |
2124 | pfcnt = 0; | |
2125 | ||
2126 | while (!feof (fi)) | |
2127 | { | |
2128 | lineno++; | |
2129 | linecharno = charno; | |
2130 | charno += readline (&lb, fi); | |
2131 | dbp = lb.buffer; | |
2132 | ||
2133 | for (i = 0; ((c = dbp[i]) && !isspace (c)) && (c != ':'); i++) | |
2134 | ; | |
2135 | ||
2136 | if ((i > 0) && (c == ':')) | |
2137 | getit (); | |
2138 | } | |
2139 | } | |
2140 | \f | |
2141 | /* Added by Mosur Mohan, 4/22/88 */ | |
2142 | /* Pascal parsing */ | |
2143 | ||
2144 | #define GET_NEW_LINE \ | |
2145 | { \ | |
2146 | linecharno = charno; lineno++; \ | |
2147 | charno += 1 + readline (&lb, inf); \ | |
2148 | dbp = lb.buffer; \ | |
2149 | } | |
2150 | ||
2151 | /* Locates tags for procedures & functions. | |
2152 | * Doesn't do any type- or var-definitions. | |
2153 | * It does look for the keyword "extern" or "forward" | |
2154 | * immediately following the procedure statement; | |
2155 | * if found, the tag is skipped. | |
2156 | */ | |
2157 | ||
2158 | void | |
2159 | PAS_funcs (fi) | |
2160 | FILE *fi; | |
2161 | { | |
2162 | struct linebuffer tline; /* mostly copied from C_entries */ | |
2163 | long save_lcno; | |
2164 | int save_lineno; | |
2165 | char c, *cp; | |
2166 | char nambuf[BUFSIZ]; | |
2167 | ||
2168 | logical /* each of these flags is TRUE iff: */ | |
2169 | incomm1, /* point is inside {..} comment */ | |
2170 | incomm2, /* point is inside (*..*) comment */ | |
2171 | inquote, /* point is inside '..' string */ | |
2172 | get_tagname, /* point is after PROCEDURE/FUNCTION */ | |
2173 | /* keyword, so next item = potential tag */ | |
2174 | found_tag, /* point is after a potential tag */ | |
2175 | inparms, /* point is within parameter-list */ | |
2176 | verify_tag; /* point has passed the parm-list, so the */ | |
2177 | /* next token will determine whether */ | |
2178 | /* this is a FORWARD/EXTERN to be */ | |
2179 | /* ignored, or whether it is a real tag */ | |
2180 | ||
2181 | lineno = 0; | |
2182 | charno = 0; | |
2183 | dbp = lb.buffer; | |
2184 | *dbp = 0; | |
2185 | initbuffer (&tline); | |
2186 | ||
2187 | incomm1 = incomm2 = inquote = FALSE; | |
2188 | found_tag = FALSE; /* have a proc name; check if extern */ | |
2189 | get_tagname = FALSE; /* have found "procedure" keyword */ | |
2190 | inparms = FALSE; /* found '(' after "proc" */ | |
2191 | verify_tag = FALSE; /* check if "extern" is ahead */ | |
2192 | ||
2193 | /* long main loop to get next char */ | |
2194 | while (!feof (fi)) | |
2195 | { | |
2196 | c = *dbp++; | |
2197 | if (c == 0) /* if end of line */ | |
2198 | { | |
2199 | GET_NEW_LINE; | |
2200 | if (*dbp == 0) | |
2201 | continue; | |
2202 | if (!((found_tag && verify_tag) || | |
2203 | get_tagname)) | |
2204 | c = *dbp++; /* only if don't need *dbp pointing */ | |
2205 | /* to the beginning of the name of */ | |
2206 | /* the procedure or function */ | |
2207 | } | |
2208 | if (incomm1) /* within { - } comments */ | |
2209 | { | |
2210 | if (c == '}') | |
2211 | incomm1 = FALSE; | |
2212 | continue; | |
2213 | } | |
2214 | else if (incomm2) /* within (* - *) comments */ | |
2215 | { | |
2216 | if (c == '*') | |
2217 | { | |
2218 | while ((c = *dbp++) == '*') | |
2219 | continue; | |
2220 | if (c == 0) | |
2221 | GET_NEW_LINE; | |
2222 | if (c == ')') | |
2223 | incomm2 = FALSE; | |
2224 | } | |
2225 | continue; | |
2226 | } | |
2227 | else if (inquote) | |
2228 | { | |
2229 | if (c == '\'') | |
2230 | inquote = FALSE; | |
2231 | continue; | |
2232 | } | |
2233 | else | |
2234 | switch (c) | |
2235 | { | |
2236 | case '\'': | |
2237 | inquote = TRUE; /* found first quote */ | |
2238 | continue; | |
2239 | case '{': /* found open-{-comment */ | |
2240 | incomm1 = TRUE; | |
2241 | continue; | |
2242 | case '(': | |
2243 | if (*dbp == '*') /* found open-(*-comment */ | |
2244 | { | |
2245 | incomm2 = TRUE; | |
2246 | dbp++; | |
2247 | } | |
2248 | else if (found_tag) /* found '(' after tag, i.e., parm-list */ | |
2249 | inparms = TRUE; | |
2250 | continue; | |
2251 | case ')': /* end of parms list */ | |
2252 | if (inparms) | |
2253 | inparms = FALSE; | |
2254 | continue; | |
2255 | case ';': | |
2256 | if ((found_tag) && (!inparms)) /* end of proc or fn stmt */ | |
2257 | { | |
2258 | verify_tag = TRUE; | |
2259 | break; | |
2260 | } | |
2261 | continue; | |
2262 | } | |
2263 | if ((found_tag) && (verify_tag) && (*dbp != ' ')) | |
2264 | { | |
2265 | /* check if this is an "extern" declaration */ | |
2266 | if (*dbp == 0) | |
2267 | continue; | |
2268 | if ((*dbp == 'e') || (*dbp == 'E')) | |
2269 | { | |
2270 | if (tail ("extern")) /* superfluous, really! */ | |
2271 | { | |
2272 | found_tag = FALSE; | |
2273 | verify_tag = FALSE; | |
2274 | } | |
2275 | } | |
2276 | else if ((*dbp == 'f') || (*dbp == 'F')) | |
2277 | { | |
2278 | if (tail ("forward")) /* check for forward reference */ | |
2279 | { | |
2280 | found_tag = FALSE; | |
2281 | verify_tag = FALSE; | |
2282 | } | |
2283 | } | |
2284 | if ((found_tag) && (verify_tag)) /* not external proc, so make tag */ | |
2285 | { | |
2286 | found_tag = FALSE; | |
2287 | verify_tag = FALSE; | |
2288 | pfnote (nambuf, TRUE, FALSE, | |
2289 | tline.buffer, cp - tline.buffer + 1, | |
2290 | save_lineno, save_lcno); | |
2291 | continue; | |
2292 | } | |
2293 | } | |
2294 | if (get_tagname) /* grab name of proc or fn */ | |
2295 | { | |
2296 | if (*dbp == 0) | |
2297 | continue; | |
2298 | ||
2299 | /* save all values for later tagging */ | |
2300 | tline.size = lb.size; | |
2301 | strcpy (tline.buffer, lb.buffer); | |
2302 | save_lineno = lineno; | |
2303 | save_lcno = linecharno; | |
2304 | ||
2305 | /* grab block name */ | |
2306 | for (cp = dbp + 1; *cp && (!endtoken (*cp)); cp++) | |
2307 | continue; | |
2308 | c = cp[0]; | |
2309 | cp[0] = 0; | |
2310 | strcpy (nambuf, dbp); | |
2311 | cp[0] = c; | |
2312 | dbp = cp; /* restore dbp to e-o-token */ | |
2313 | get_tagname = FALSE; | |
2314 | found_tag = TRUE; | |
2315 | continue; | |
2316 | ||
2317 | /* and proceed to check for "extern" */ | |
2318 | } | |
2319 | if ((!incomm1) && (!incomm2) && (!inquote) && | |
2320 | (!found_tag) && (!get_tagname)) | |
2321 | { | |
2322 | /* check for proc/fn keywords */ | |
2323 | switch (c | ' ') | |
2324 | { | |
2325 | case 'p': | |
2326 | if (tail ("rocedure")) /* c = 'p', dbp has advanced */ | |
2327 | get_tagname = TRUE; | |
2328 | continue; | |
2329 | case 'f': | |
2330 | if (tail ("unction")) | |
2331 | get_tagname = TRUE; | |
2332 | continue; | |
2333 | } | |
2334 | } | |
2335 | } /* while not e-o-f */ | |
2336 | } | |
2337 | \f | |
2338 | /* | |
2339 | * lisp tag functions | |
2340 | * just look for (def or (DEF | |
2341 | */ | |
2342 | ||
2343 | void | |
2344 | L_funcs (fi) | |
2345 | FILE *fi; | |
2346 | { | |
2347 | lineno = 0; | |
2348 | charno = 0; | |
2349 | pfcnt = 0; | |
2350 | ||
2351 | while (!feof (fi)) | |
2352 | { | |
2353 | lineno++; | |
2354 | linecharno = charno; | |
2355 | charno += readline (&lb, fi); | |
2356 | dbp = lb.buffer; | |
2357 | if (dbp[0] == '(') | |
2358 | { | |
2359 | if (L_isdef (dbp)) | |
2360 | { | |
2361 | while (!isspace (*dbp)) | |
2362 | dbp++; | |
2363 | while (isspace (*dbp)) | |
2364 | dbp++; | |
2365 | L_getit (); | |
2366 | } | |
2367 | else | |
2368 | { | |
2369 | /* Check for (foo::defmumble name-defined ... */ | |
31d4b314 | 2370 | do |
c6d46f5f | 2371 | dbp++; |
31d4b314 FP |
2372 | while (*dbp && !isspace (*dbp) |
2373 | && *dbp != ':' && *dbp != '(' && *dbp != ')'); | |
c6d46f5f JB |
2374 | if (*dbp == ':') |
2375 | { | |
31d4b314 | 2376 | do |
c6d46f5f | 2377 | dbp++; |
31d4b314 | 2378 | while (*dbp == ':'); |
c6d46f5f | 2379 | |
31d4b314 | 2380 | if (L_isdef (dbp - 1)) |
c6d46f5f JB |
2381 | { |
2382 | while (!isspace (*dbp)) | |
2383 | dbp++; | |
2384 | while (isspace (*dbp)) | |
2385 | dbp++; | |
2386 | L_getit (); | |
2387 | } | |
2388 | } | |
2389 | } | |
2390 | } | |
2391 | } | |
2392 | } | |
2393 | ||
2394 | int | |
2395 | L_isdef (dbp) | |
31d4b314 | 2396 | register char *dbp; |
c6d46f5f | 2397 | { |
31d4b314 FP |
2398 | return ((dbp[1] == 'd' || dbp[1] == 'D') |
2399 | && (dbp[2] == 'e' || dbp[2] == 'E') | |
2400 | && (dbp[3] == 'f' || dbp[3] == 'F')); | |
2401 | } | |
2402 | ||
2403 | int | |
2404 | L_isquote (dbp) | |
2405 | register char *dbp; | |
2406 | { | |
2407 | return ((*(++dbp) == 'q' || *dbp == 'Q') | |
2408 | && (*(++dbp) == 'u' || *dbp == 'U') | |
2409 | && (*(++dbp) == 'o' || *dbp == 'O') | |
2410 | && (*(++dbp) == 't' || *dbp == 'T') | |
2411 | && (*(++dbp) == 'e' || *dbp == 'E') | |
2412 | && isspace(*(++dbp))); | |
c6d46f5f JB |
2413 | } |
2414 | ||
2415 | void | |
2416 | L_getit () | |
2417 | { | |
2418 | register char *cp; | |
2419 | char c; | |
2420 | char nambuf[BUFSIZ]; | |
2421 | ||
31d4b314 FP |
2422 | if (*dbp == '\'') /* Skip prefix quote */ |
2423 | dbp++; | |
2424 | else if (*dbp == '(' && L_isquote (dbp)) /* Skip "(quote " */ | |
2425 | { | |
2426 | dbp += 7; | |
2427 | while (isspace(*dbp)) | |
2428 | dbp++; | |
2429 | } | |
2430 | for (cp = dbp /*+1*/; *cp && *cp != '(' && *cp != ' ' && *cp != ')'; cp++) | |
c6d46f5f | 2431 | continue; |
31d4b314 FP |
2432 | if (cp == dbp) |
2433 | return; | |
2434 | ||
c6d46f5f JB |
2435 | c = cp[0]; |
2436 | cp[0] = 0; | |
2437 | (void) strcpy (nambuf, dbp); | |
2438 | cp[0] = c; | |
591fa824 RS |
2439 | pfnote (nambuf, TRUE, FALSE, lb.buffer, |
2440 | cp - lb.buffer + 1, lineno, linecharno); | |
c6d46f5f JB |
2441 | pfcnt++; |
2442 | } | |
2443 | \f | |
2444 | /* | |
2445 | * Scheme tag functions | |
2446 | * look for (def... xyzzy | |
2447 | * look for (def... (xyzzy | |
2448 | * look for (def ... ((...(xyzzy .... | |
2449 | * look for (set! xyzzy | |
2450 | */ | |
2451 | ||
2452 | static void get_scheme (); | |
2453 | ||
2454 | void | |
2455 | Scheme_funcs (fi) | |
2456 | FILE *fi; | |
2457 | { | |
2458 | lineno = 0; | |
2459 | charno = 0; | |
2460 | pfcnt = 0; | |
2461 | ||
2462 | while (!feof (fi)) | |
2463 | { | |
2464 | lineno++; | |
2465 | linecharno = charno; | |
2466 | charno += readline (&lb, fi); | |
2467 | dbp = lb.buffer; | |
2468 | if (dbp[0] == '(' && | |
2469 | (dbp[1] == 'D' || dbp[1] == 'd') && | |
2470 | (dbp[2] == 'E' || dbp[2] == 'e') && | |
2471 | (dbp[3] == 'F' || dbp[3] == 'f')) | |
2472 | { | |
2473 | while (!isspace (*dbp)) | |
2474 | dbp++; | |
2475 | /* Skip over open parens and white space */ | |
2476 | while (*dbp && (isspace (*dbp) || *dbp == '(')) | |
2477 | dbp++; | |
2478 | get_scheme (); | |
2479 | } | |
2480 | if (dbp[0] == '(' && | |
2481 | (dbp[1] == 'S' || dbp[1] == 's') && | |
2482 | (dbp[2] == 'E' || dbp[2] == 'e') && | |
2483 | (dbp[3] == 'T' || dbp[3] == 't') && | |
2484 | (dbp[4] == '!' || dbp[4] == '!') && | |
2485 | (isspace (dbp[5]))) | |
2486 | { | |
2487 | while (!isspace (*dbp)) | |
2488 | dbp++; | |
2489 | /* Skip over white space */ | |
2490 | while (isspace (*dbp)) | |
2491 | dbp++; | |
2492 | get_scheme (); | |
2493 | } | |
2494 | } | |
2495 | } | |
2496 | ||
2497 | static void | |
2498 | get_scheme () | |
2499 | { | |
2500 | register char *cp; | |
2501 | char c; | |
2502 | char nambuf[BUFSIZ]; | |
2503 | ||
2504 | if (*dbp == 0) | |
2505 | return; | |
2506 | /* Go till you get to white space or a syntactic break */ | |
2507 | for (cp = dbp + 1; *cp && *cp != '(' && *cp != ')' && !isspace (*cp); cp++) | |
2508 | continue; | |
2509 | /* Null terminate the string there. */ | |
2510 | c = cp[0]; | |
2511 | cp[0] = 0; | |
2512 | /* Copy the string */ | |
2513 | strcpy (nambuf, dbp); | |
2514 | /* Unterminate the string */ | |
2515 | cp[0] = c; | |
2516 | /* Announce the change */ | |
2517 | pfnote (nambuf, TRUE, FALSE, lb.buffer, cp - lb.buffer + 1, lineno, linecharno); | |
2518 | pfcnt++; | |
2519 | } | |
2520 | \f | |
2521 | /* Find tags in TeX and LaTeX input files. */ | |
2522 | ||
2523 | /* TEX_toktab is a table of TeX control sequences that define tags. | |
2524 | Each TEX_tabent records one such control sequence. | |
2525 | CONVERT THIS TO USE THE Stab TYPE!! */ | |
2526 | ||
2527 | struct TEX_tabent | |
2528 | { | |
2529 | char *name; | |
2530 | int len; | |
2531 | }; | |
2532 | ||
2533 | struct TEX_tabent *TEX_toktab = NULL; /* Table with tag tokens */ | |
2534 | ||
2535 | /* Default set of control sequences to put into TEX_toktab. | |
2536 | The value of environment var TEXTAGS is prepended to this. */ | |
2537 | ||
2538 | static char *TEX_defenv = | |
2539 | ":chapter:section:subsection:subsubsection:eqno:label:ref:cite:bibitem:typeout"; | |
2540 | ||
2541 | void TEX_mode (); | |
2542 | struct TEX_tabent *TEX_decode_env (); | |
2543 | void TEX_getit (); | |
2544 | int TEX_Token (); | |
2545 | ||
2546 | static char TEX_esc = '\\'; | |
2547 | static char TEX_opgrp = '{'; | |
2548 | static char TEX_clgrp = '}'; | |
2549 | ||
2550 | /* | |
2551 | * TeX/LaTeX scanning loop. | |
2552 | */ | |
2553 | ||
2554 | void | |
2555 | TEX_funcs (fi) | |
2556 | FILE *fi; | |
2557 | { | |
2558 | char *lasthit; | |
2559 | ||
2560 | lineno = 0; | |
2561 | charno = 0; | |
2562 | pfcnt = 0; | |
2563 | ||
2564 | /* Select either \ or ! as escape character. */ | |
2565 | TEX_mode (fi); | |
2566 | ||
2567 | /* Initialize token table once from environment. */ | |
2568 | if (!TEX_toktab) | |
2569 | TEX_toktab = TEX_decode_env ("TEXTAGS", TEX_defenv); | |
2570 | ||
2571 | while (!feof (fi)) | |
d2729198 | 2572 | { /* Scan each line in file */ |
c6d46f5f JB |
2573 | lineno++; |
2574 | linecharno = charno; | |
2575 | charno += readline (&lb, fi); | |
2576 | dbp = lb.buffer; | |
2577 | lasthit = dbp; | |
d2729198 | 2578 | while (dbp = etags_index (dbp, TEX_esc)) /* Look at each escape in line */ |
8a6c8bcf RS |
2579 | { |
2580 | register int i; | |
c6d46f5f | 2581 | |
8a6c8bcf RS |
2582 | if (!*(++dbp)) |
2583 | break; | |
2584 | linecharno += dbp - lasthit; | |
c6d46f5f | 2585 | lasthit = dbp; |
8a6c8bcf RS |
2586 | i = TEX_Token (lasthit); |
2587 | if (0 <= i) | |
c6d46f5f | 2588 | { |
8a6c8bcf | 2589 | TEX_getit (lasthit, TEX_toktab[i].len); |
d2729198 | 2590 | break; /* We only save a line once */ |
c6d46f5f JB |
2591 | } |
2592 | } | |
2593 | } | |
2594 | } | |
2595 | ||
2596 | #define TEX_LESC '\\' | |
2597 | #define TEX_SESC '!' | |
2598 | #define TEX_cmt '%' | |
2599 | ||
2600 | /* Figure out whether TeX's escapechar is '\\' or '!' and set grouping */ | |
2601 | /* chars accordingly. */ | |
2602 | ||
2603 | void | |
2604 | TEX_mode (f) | |
2605 | FILE *f; | |
2606 | { | |
2607 | int c; | |
2608 | ||
2609 | while ((c = getc (f)) != EOF) | |
2610 | { | |
2611 | /* Skip to next line if we hit the TeX comment char. */ | |
2612 | if (c == TEX_cmt) | |
2613 | while (c != '\n') | |
2614 | c = getc (f); | |
2615 | else if (c == TEX_LESC || c == TEX_SESC ) | |
2616 | break; | |
2617 | } | |
2618 | ||
2619 | if (c == TEX_LESC) | |
2620 | { | |
2621 | TEX_esc = TEX_LESC; | |
2622 | TEX_opgrp = '{'; | |
2623 | TEX_clgrp = '}'; | |
2624 | } | |
2625 | else | |
2626 | { | |
2627 | TEX_esc = TEX_SESC; | |
2628 | TEX_opgrp = '<'; | |
2629 | TEX_clgrp = '>'; | |
2630 | } | |
2631 | rewind (f); | |
2632 | } | |
2633 | ||
2634 | /* Read environment and prepend it to the default string. */ | |
2635 | /* Build token table. */ | |
2636 | ||
2637 | struct TEX_tabent * | |
2638 | TEX_decode_env (evarname, defenv) | |
2639 | char *evarname; | |
2640 | char *defenv; | |
2641 | { | |
2642 | register char *env, *p; | |
c6d46f5f JB |
2643 | |
2644 | struct TEX_tabent *tab; | |
2645 | int size, i; | |
2646 | ||
2647 | /* Append default string to environment. */ | |
2648 | env = getenv (evarname); | |
2649 | if (!env) | |
2650 | env = defenv; | |
2651 | else | |
2652 | env = concat (env, defenv, ""); | |
2653 | ||
2654 | /* Allocate a token table */ | |
2655 | for (size = 1, p = env; p;) | |
8a6c8bcf | 2656 | if ((p = etags_index (p, ':')) && *(++p)) |
c6d46f5f | 2657 | size++; |
8a6c8bcf RS |
2658 | /* Add 1 to leave room for null terminator. */ |
2659 | tab = xnew (size + 1, struct TEX_tabent); | |
c6d46f5f JB |
2660 | |
2661 | /* Unpack environment string into token table. Be careful about */ | |
2662 | /* zero-length strings (leading ':', "::" and trailing ':') */ | |
2663 | for (i = 0; *env;) | |
2664 | { | |
8a6c8bcf | 2665 | p = etags_index (env, ':'); |
c6d46f5f JB |
2666 | if (!p) /* End of environment string. */ |
2667 | p = env + strlen (env); | |
2668 | if (p - env > 0) | |
2669 | { /* Only non-zero strings. */ | |
2670 | tab[i].name = savenstr (env, p - env); | |
2671 | tab[i].len = strlen (tab[i].name); | |
2672 | i++; | |
2673 | } | |
2674 | if (*p) | |
2675 | env = p + 1; | |
2676 | else | |
2677 | { | |
2678 | tab[i].name = NULL; /* Mark end of table. */ | |
2679 | tab[i].len = 0; | |
2680 | break; | |
2681 | } | |
2682 | } | |
2683 | return tab; | |
2684 | } | |
2685 | ||
2686 | /* Record a tag defined by a TeX command of length LEN and starting at NAME. | |
2687 | The name being defined actually starts at (NAME + LEN + 1). | |
2688 | But we seem to include the TeX command in the tag name. */ | |
2689 | ||
2690 | void | |
2691 | TEX_getit (name, len) | |
2692 | char *name; | |
2693 | int len; | |
2694 | { | |
2695 | char *p = name + len; | |
2696 | char nambuf[BUFSIZ]; | |
2697 | ||
2698 | if (*name == 0) | |
2699 | return; | |
2700 | ||
2701 | /* Let tag name extend to next group close (or end of line) */ | |
2702 | while (*p && *p != TEX_clgrp) | |
2703 | p++; | |
2704 | (void) strncpy (nambuf, name, p - name); | |
2705 | nambuf[p - name] = 0; | |
2706 | ||
2707 | pfnote (nambuf, TRUE, FALSE, lb.buffer, strlen (lb.buffer), lineno, linecharno); | |
2708 | pfcnt++; | |
2709 | } | |
2710 | ||
2711 | /* If the text at CP matches one of the tag-defining TeX command names, | |
8a6c8bcf | 2712 | return the etags_index of that command in TEX_toktab. |
c6d46f5f JB |
2713 | Otherwise return -1. */ |
2714 | ||
2715 | /* Keep the capital `T' in `Token' for dumb truncating compilers | |
2716 | (this distinguishes it from `TEX_toktab' */ | |
2717 | int | |
2718 | TEX_Token (cp) | |
2719 | char *cp; | |
2720 | { | |
2721 | int i; | |
2722 | ||
2723 | for (i = 0; TEX_toktab[i].len > 0; i++) | |
2724 | if (strncmp (TEX_toktab[i].name, cp, TEX_toktab[i].len) == 0) | |
2725 | return i; | |
2726 | return -1; | |
2727 | } | |
2728 | \f | |
2729 | /* Support for Prolog. */ | |
2730 | ||
2731 | /* whole head (not only functor, but also arguments) | |
2732 | is gotten in compound term. */ | |
2733 | ||
2734 | void | |
2735 | prolog_getit (s, lineno, linecharno) | |
2736 | char *s; | |
2737 | int lineno; | |
2738 | long linecharno; | |
2739 | { | |
2740 | char nambuf[BUFSIZ], *save_s, tmpc; | |
2741 | int insquote, npar; | |
2742 | ||
2743 | save_s = s; | |
2744 | insquote = FALSE; | |
2745 | npar = 0; | |
2746 | while (1) | |
2747 | { | |
2748 | if (*s == '\0') /* syntax error. */ | |
2749 | return; | |
2750 | else if (insquote && *s == '\'' && *(s + 1) == '\'') | |
2751 | s += 2; | |
2752 | else if (*s == '\'') | |
2753 | { | |
2754 | insquote = !insquote; | |
2755 | s++; | |
2756 | } | |
2757 | else if (!insquote && *s == '(') | |
2758 | { | |
2759 | npar++; | |
2760 | s++; | |
2761 | } | |
2762 | else if (!insquote && *s == ')') | |
2763 | { | |
2764 | npar--; | |
2765 | s++; | |
2766 | if (npar == 0) | |
2767 | break; | |
2768 | else if (npar < 0) /* syntax error. */ | |
2769 | return; | |
2770 | } | |
2771 | else if (!insquote && *s == '.' && (isspace (*(s + 1)) || *(s + 1) == '\0')) | |
2772 | { /* fullstop. */ | |
2773 | if (npar != 0) /* syntax error. */ | |
2774 | return; | |
2775 | s++; | |
2776 | break; | |
2777 | } | |
2778 | else | |
2779 | s++; | |
2780 | } | |
2781 | tmpc = *s; | |
2782 | *s = '\0'; | |
2783 | strcpy (nambuf, save_s); | |
2784 | *s = tmpc; | |
2785 | pfnote (nambuf, TRUE, save_s, strlen (nambuf), lineno, linecharno); | |
2786 | } | |
2787 | ||
2788 | /* It is assumed that prolog predicate starts from column 0. */ | |
2789 | ||
2790 | void | |
2791 | prolog_funcs (fi) | |
2792 | FILE *fi; | |
2793 | { | |
2794 | void skip_comment (), prolog_getit (); | |
2795 | ||
2796 | lineno = linecharno = charno = 0; | |
2797 | while (!feof (fi)) | |
2798 | { | |
2799 | lineno++; | |
2800 | linecharno += charno; | |
2801 | charno = readline (&lb, fi) + 1; /* 1 for newline. */ | |
2802 | dbp = lb.buffer; | |
2803 | if (isspace (dbp[0])) /* not predicate header. */ | |
2804 | continue; | |
2805 | else if (dbp[0] == '%') /* comment. */ | |
2806 | continue; | |
2807 | else if (dbp[0] == '/' && dbp[1] == '*') /* comment. */ | |
2808 | skip_comment (&lb, fi, &lineno, &linecharno); | |
2809 | else /* found. */ | |
2810 | prolog_getit (dbp, lineno, linecharno); | |
2811 | } | |
2812 | } | |
2813 | ||
2814 | void | |
2815 | skip_comment (plb, fi, plineno, plinecharno) | |
2816 | struct linebuffer *plb; | |
2817 | FILE *fi; | |
2818 | int *plineno; /* result */ | |
2819 | long *plinecharno; /* result */ | |
2820 | { | |
2821 | while (!substr ("*/", plb->buffer)) | |
2822 | { | |
2823 | (*plineno)++; | |
2824 | *plinecharno += readline (plb, fi) + 1; | |
2825 | } /* 1 for newline. */ | |
2826 | } | |
2827 | ||
2828 | /* Return TRUE if 'sub' exists somewhere in 's'. */ | |
2829 | ||
2830 | int | |
2831 | substr (sub, s) | |
2832 | char *sub; | |
2833 | char *s; | |
2834 | { | |
8a6c8bcf | 2835 | while (*s && (s = etags_index (s, *sub))) |
c6d46f5f JB |
2836 | if (prestr (sub, s)) |
2837 | return (TRUE); | |
2838 | else | |
2839 | s++; | |
2840 | return (FALSE); | |
2841 | } | |
2842 | ||
2843 | /* Return TRUE if 'pre' is prefix of string 's'. */ | |
2844 | ||
2845 | int | |
2846 | prestr (pre, s) | |
2847 | char *pre; | |
2848 | char *s; | |
2849 | { | |
2850 | if (*pre == '\0') | |
2851 | return (TRUE); | |
2852 | else if (*pre == *s) | |
2853 | return (prestr (pre + 1, s + 1)); | |
2854 | else | |
2855 | return (FALSE); | |
2856 | } | |
2857 | \f | |
2858 | /* Initialize a linebuffer for use */ | |
2859 | ||
2860 | void | |
2861 | initbuffer (linebuffer) | |
2862 | struct linebuffer *linebuffer; | |
2863 | { | |
2864 | linebuffer->size = 200; | |
2865 | linebuffer->buffer = xnew (200, char); | |
2866 | } | |
2867 | ||
2868 | /* | |
2869 | * Read a line of text from `stream' into `linebuffer'. | |
2870 | * Return the number of characters read from `stream', | |
2871 | * which is the length of the line including the newline, if any. | |
2872 | */ | |
2873 | long | |
2874 | readline (linebuffer, stream) | |
2875 | struct linebuffer *linebuffer; | |
2876 | register FILE *stream; | |
2877 | { | |
2878 | char *buffer = linebuffer->buffer; | |
2879 | register char *p = linebuffer->buffer; | |
2880 | register char *pend; | |
2881 | int newline; /* 1 if ended with newline, 0 if ended with EOF */ | |
2882 | ||
eb8c3be9 | 2883 | pend = p + linebuffer->size; /* Separate to avoid 386/IX compiler bug. */ |
c6d46f5f JB |
2884 | |
2885 | while (1) | |
2886 | { | |
2887 | register int c = getc (stream); | |
2888 | if (p == pend) | |
2889 | { | |
2890 | linebuffer->size *= 2; | |
2891 | buffer = (char *) xrealloc (buffer, linebuffer->size); | |
2892 | p += buffer - linebuffer->buffer; | |
2893 | pend = buffer + linebuffer->size; | |
2894 | linebuffer->buffer = buffer; | |
2895 | } | |
2896 | if (c < 0 || c == '\n') | |
2897 | { | |
2898 | *p = 0; | |
2899 | newline = (c == '\n' ? 1 : 0); | |
2900 | break; | |
2901 | } | |
2902 | *p++ = c; | |
2903 | } | |
2904 | ||
2905 | return p - buffer + newline; | |
2906 | } | |
2907 | \f | |
2908 | char * | |
2909 | savestr (cp) | |
2910 | char *cp; | |
2911 | { | |
2912 | return savenstr (cp, strlen (cp)); | |
2913 | } | |
2914 | ||
2915 | char * | |
2916 | savenstr (cp, len) | |
2917 | char *cp; | |
2918 | int len; | |
2919 | { | |
2920 | register char *dp; | |
2921 | ||
2922 | dp = xnew (len + 1, char); | |
2923 | (void) strncpy (dp, cp, len); | |
2924 | dp[len] = '\0'; | |
2925 | return dp; | |
2926 | } | |
2927 | ||
c6d46f5f JB |
2928 | /* |
2929 | * Return the ptr in sp at which the character c last | |
2930 | * appears; NULL if not found | |
2931 | * | |
2932 | * Identical to v7 rindex, included for portability. | |
2933 | */ | |
2934 | ||
2935 | char * | |
8a6c8bcf | 2936 | etags_rindex (sp, c) |
c6d46f5f JB |
2937 | register char *sp, c; |
2938 | { | |
2939 | register char *r; | |
2940 | ||
2941 | r = NULL; | |
2942 | do | |
2943 | { | |
2944 | if (*sp == c) | |
2945 | r = sp; | |
2946 | } while (*sp++); | |
2947 | return (r); | |
2948 | } | |
2949 | ||
9d7ad1b3 | 2950 | |
c6d46f5f JB |
2951 | /* |
2952 | * Return the ptr in sp at which the character c first | |
2953 | * appears; NULL if not found | |
2954 | * | |
2955 | * Identical to v7 index, included for portability. | |
2956 | */ | |
2957 | ||
2958 | char * | |
8a6c8bcf | 2959 | etags_index (sp, c) |
c6d46f5f JB |
2960 | register char *sp, c; |
2961 | { | |
2962 | do | |
2963 | { | |
2964 | if (*sp == c) | |
2965 | return (sp); | |
2966 | } while (*sp++); | |
2967 | return (NULL); | |
2968 | } | |
2969 | ||
c6d46f5f JB |
2970 | /* Print error message and exit. */ |
2971 | ||
2972 | /* VARARGS1 */ | |
2973 | void | |
2974 | fatal (s1, s2) | |
2975 | char *s1, *s2; | |
2976 | { | |
2977 | error (s1, s2); | |
2978 | exit (1); | |
2979 | } | |
2980 | ||
2981 | /* Print error message. `s1' is printf control string, `s2' is arg for it. */ | |
2982 | ||
2983 | /* VARARGS1 */ | |
2984 | void | |
2985 | error (s1, s2) | |
2986 | char *s1, *s2; | |
2987 | { | |
2988 | fprintf (stderr, "%s: ", progname); | |
2989 | fprintf (stderr, s1, s2); | |
2990 | fprintf (stderr, "\n"); | |
2991 | } | |
2992 | ||
2993 | /* Return a newly-allocated string whose contents concatenate those of s1, s2, s3. */ | |
2994 | ||
2995 | char * | |
2996 | concat (s1, s2, s3) | |
2997 | char *s1, *s2, *s3; | |
2998 | { | |
2999 | int len1 = strlen (s1), len2 = strlen (s2), len3 = strlen (s3); | |
3000 | char *result = xnew (len1 + len2 + len3 + 1, char); | |
3001 | ||
3002 | (void) strcpy (result, s1); | |
3003 | (void) strcpy (result + len1, s2); | |
3004 | (void) strcpy (result + len1 + len2, s3); | |
3005 | *(result + len1 + len2 + len3) = 0; | |
3006 | ||
3007 | return result; | |
3008 | } | |
3009 | ||
3010 | /* Like malloc but get fatal error if memory is exhausted. */ | |
3011 | ||
3012 | char * | |
3013 | xmalloc (size) | |
3014 | int size; | |
3015 | { | |
3016 | char *result = malloc (size); | |
3017 | if (!result) | |
3018 | fatal ("virtual memory exhausted", 0); | |
3019 | return result; | |
3020 | } | |
3021 | ||
3022 | char * | |
3023 | xrealloc (ptr, size) | |
3024 | char *ptr; | |
3025 | int size; | |
3026 | { | |
3027 | char *result = realloc (ptr, size); | |
3028 | if (!result) | |
3029 | fatal ("virtual memory exhausted"); | |
3030 | return result; | |
3031 | } |