Commit | Line | Data |
---|---|---|
c6d46f5f | 1 | /* Tags file maker to go with GNU Emacs |
4746118a | 2 | Copyright (C) 1984, 1987, 1988, 1989, 1992 Free Software Foundation, Inc. and Ken Arnold |
c6d46f5f JB |
3 | |
4 | This file is part of GNU Emacs. | |
5 | ||
6 | GNU Emacs is free software; you can redistribute it and/or modify | |
7 | it under the terms of the GNU General Public License as published by | |
8 | the Free Software Foundation; either version 1, or (at your option) | |
9 | any later version. | |
10 | ||
11 | GNU Emacs is distributed in the hope that it will be useful, | |
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 | |
17 | along with GNU Emacs; see the file COPYING. If not, write to | |
18 | the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ | |
19 | ||
20 | /* | |
21 | * Authors: | |
22 | * Ctags originally by Ken Arnold. | |
23 | * FORTRAN added by Jim Kleckner. | |
24 | * Ed Pelegri-Llopart added C typedefs. | |
25 | * Gnu Emacs TAGS format and modifications by RMS? | |
26 | * Sam Kendall added C++. | |
27 | */ | |
28 | ||
29 | #include <stdio.h> | |
30 | #include <ctype.h> | |
31 | #include <sys/types.h> | |
32 | #include <sys/stat.h> | |
33 | ||
4746118a JB |
34 | #include "getopt.h" |
35 | ||
1e134a5f RM |
36 | #ifdef __GNUC__ |
37 | #define alloca __builtin_alloca | |
38 | #else | |
39 | #ifdef sparc | |
40 | #include <alloca.h> | |
41 | #else | |
42 | extern char *alloca (); | |
43 | #endif | |
44 | #endif | |
45 | ||
c6d46f5f JB |
46 | extern char *malloc (), *realloc (); |
47 | extern char *getenv (); | |
48 | extern char *index (), *rindex (); | |
49 | extern char *strcpy (), *strncpy (); | |
50 | extern int strcmp (); | |
51 | ||
52 | #ifdef hpux | |
53 | #define notdef | |
54 | #endif | |
55 | ||
56 | /* Define the symbol ETAGS to make the program "etags", | |
57 | which makes emacs-style tag tables by default. | |
58 | Define CTAGS to make the program "ctags" compatible with the usual one. | |
59 | Define neither one to get behavior that depends | |
60 | on the name with which the program is invoked | |
61 | (but we don't normally compile it that way). */ | |
62 | ||
63 | #if !defined(ETAGS) && !defined(CTAGS) | |
64 | /* If neither is defined, program can be run as either. */ | |
65 | #define ETAGS | |
66 | #define CTAGS | |
67 | #endif | |
68 | ||
69 | /* On VMS, CTAGS is not useful, so always do ETAGS. */ | |
70 | #ifdef VMS | |
71 | #ifndef ETAGS | |
72 | #define ETAGS | |
73 | #endif | |
74 | #endif | |
75 | ||
76 | /* Exit codes for success and failure. */ | |
77 | #ifdef VMS | |
78 | #define GOOD (1) | |
79 | #define BAD (0) | |
80 | #else | |
81 | #define GOOD (0) | |
82 | #define BAD (1) | |
83 | #endif | |
84 | ||
85 | /* | |
86 | * The FILEPOS abstract type, which represents a position in a file, | |
87 | * plus the following accessor functions: | |
88 | * | |
89 | * long GET_CHARNO (pos) | |
90 | * returns absolute char number. | |
91 | * long GET_COOKIE (pos) | |
92 | * returns ftell () cookie. | |
93 | * void SET_FILEPOS (pos, fp, charno) | |
94 | * FILE *fp; long charno; | |
95 | * sets `pos' from the current file | |
96 | * position of `fp' and from `charno', | |
97 | * which must be the absolute character | |
98 | * number corresponding to the current | |
99 | * position of `fp'. | |
100 | * | |
101 | * The `pos' parameter is an lvalue expression of type FILEPOS. | |
102 | * Parameters to the accessor functions are evaluated 0 or more times, | |
103 | * and so must have no side effects. | |
104 | * | |
105 | * FILEPOS objects can also be assigned and passed to and from | |
106 | * functions in the normal C manner. | |
107 | * | |
108 | * Implementation notes: the `+ 0' is to enforce rvalue-ness. | |
109 | */ | |
110 | #ifdef VMS | |
111 | typedef struct | |
112 | { | |
113 | long cookie; | |
114 | long charno; | |
115 | } FILEPOS; | |
116 | ||
117 | #define GET_CHARNO(pos) ((pos).charno + 0) | |
118 | #define GET_COOKIE(pos) ((pos).cookie + 0) | |
119 | #define SET_FILEPOS(pos, fp, cno) \ | |
120 | ((void) ((pos).cookie = ftell (fp), (pos).charno = (cno))) | |
121 | #else | |
122 | #ifndef DEBUG | |
123 | /* UNIX real implementation */ | |
124 | typedef long FILEPOS; | |
125 | #define GET_CHARNO(pos) ((pos) + 0) | |
126 | #define GET_COOKIE(pos) GET_CHARNO (pos) | |
127 | #define SET_FILEPOS(pos, fp, cno) ((void) ((pos) = (cno))) | |
128 | #else | |
129 | /* UNIX debugging implementation */ | |
130 | typedef struct | |
131 | { | |
132 | long charno; | |
133 | } FILEPOS; | |
134 | ||
135 | #define GET_CHARNO(pos) ((pos).charno + 0) | |
136 | #define GET_COOKIE(pos) GET_CHARNO (pos) | |
137 | #define SET_FILEPOS(pos, fp, cno) \ | |
138 | ((void) ((pos).charno = (cno), \ | |
139 | (cno) != ftell (fp) ? (error ("SET_FILEPOS inconsistency"), 0) \ | |
140 | : 0)) | |
141 | #endif | |
142 | #endif | |
143 | ||
144 | #define streq(s, t) (strcmp (s, t) == 0) | |
145 | #define strneq(s, t, n) (strncmp (s, t, n) == 0) | |
146 | #define reg register | |
147 | #define logical char | |
148 | ||
149 | #define TRUE 1 | |
150 | #define FALSE 0 | |
151 | ||
152 | #define iswhite(arg) (_wht[arg]) /* T if char is white */ | |
153 | #define begtoken(arg) (_btk[arg]) /* T if char can start token */ | |
154 | #define intoken(arg) (_itk[arg]) /* T if char can be in token */ | |
155 | #define endtoken(arg) (_etk[arg]) /* T if char ends tokens */ | |
156 | #define isgood(arg) (_gd[arg]) /* T if char can be after ')' */ | |
157 | ||
158 | #define max(I1,I2) ((I1) > (I2) ? (I1) : (I2)) | |
159 | ||
160 | struct nd_st | |
161 | { /* sorting structure */ | |
162 | char *name; /* function or type name */ | |
163 | char *file; /* file name */ | |
164 | logical is_func; /* use pattern or line no */ | |
165 | logical rewritten; /* list name separately */ | |
166 | logical been_warned; /* set if noticed dup */ | |
167 | int lno; /* line number tag is on */ | |
168 | long cno; /* character number line starts on */ | |
169 | char *pat; /* search pattern */ | |
170 | struct nd_st *left, *right; /* left and right sons */ | |
171 | }; | |
172 | ||
173 | long ftell (); | |
174 | typedef struct nd_st NODE; | |
175 | ||
176 | logical gotone, /* found a func already on line */ | |
177 | /* boolean "func" (see init) */ | |
178 | header_file, /* TRUE if .h file, FALSE o.w. */ | |
179 | _wht[0177], _etk[0177], _itk[0177], _btk[0177], _gd[0177]; | |
180 | ||
181 | ||
182 | char *concat (); | |
183 | char *savenstr (); | |
184 | char *savestr (); | |
185 | char *xmalloc (); | |
186 | char *xrealloc (); | |
187 | int L_isdef (); | |
188 | int PF_funcs (); | |
189 | int total_size_of_entries (); | |
190 | logical consider_token (); | |
191 | logical tail (); | |
192 | long readline (); | |
193 | void Asm_funcs (); | |
194 | void C_entries (); | |
195 | void L_funcs (); | |
196 | void L_getit (); | |
197 | void PAS_funcs (); | |
198 | void Scheme_funcs (); | |
199 | void TEX_funcs (); | |
200 | void add_node (); | |
201 | void error (); | |
202 | void fatal (); | |
203 | void find_entries (); | |
204 | void free_tree (); | |
205 | void getit (); | |
206 | void getline (); | |
207 | void init (); | |
208 | void initbuffer (); | |
209 | void initbuffer (); | |
210 | void pfnote (); | |
211 | void process_file (); | |
212 | void put_entries (); | |
213 | void takeprec (); | |
214 | ||
215 | /* | |
216 | * MACRO | |
217 | * xnew -- allocate storage | |
218 | * | |
219 | * SYNOPSIS | |
220 | * Type *xnew (int n, Type); | |
221 | */ | |
222 | #define xnew(n, Type) ((Type *) xmalloc ((n) * sizeof (Type))) | |
223 | \f | |
224 | ||
225 | ||
226 | /* | |
227 | * Symbol table stuff. | |
228 | * | |
229 | * Should probably be implemented with hash table; linked list for now. | |
230 | */ | |
231 | ||
232 | enum sym_type | |
233 | { | |
234 | st_none, st_C_struct, st_C_enum, st_C_define, st_C_typedef, st_C_typespec | |
235 | }; | |
236 | ||
237 | struct stab_entry | |
238 | { | |
239 | char *sym; | |
240 | int symlen; | |
241 | enum sym_type type; | |
242 | struct stab_entry *next; | |
243 | }; | |
244 | ||
245 | typedef struct stab_entry Stab_entry; | |
246 | typedef Stab_entry *Stab; | |
247 | ||
248 | /* | |
249 | * NAME | |
250 | * Stab, Stab_entry, stab_create, stab_search, stab_find -- symbol table | |
251 | * | |
252 | * SYNOPSIS | |
253 | * Types: Stab, Stab_entry, enum sym_type | |
254 | * | |
255 | * Stab * stab_create () | |
256 | * | |
257 | * Stab_entry * stab_find (stab, sym) | |
258 | * Stab *stab; | |
259 | * char *sym; | |
260 | * | |
261 | * Stab_entry * stab_search (stab, sym) | |
262 | * Stab *stab; | |
263 | * char *sym; | |
264 | * | |
265 | * DESCRIPTION | |
266 | * stab_create creates a Stab, a symbol table object, and returns a | |
267 | * pointer to it. stab_find finds a symbol in a Stab; it returns a | |
268 | * pointer to the Stab_entry if found, otherwise NULL. stab_search | |
269 | * is like stab_find, except that it creates a new Stab_entry, | |
270 | * initialized with type = st_none, if one did not exist already | |
271 | * (it never returns NULL). | |
272 | * | |
273 | * A Stab_entry is a structure that contains at least the following | |
274 | * members: | |
275 | * | |
276 | * char *name; // must not be modified | |
277 | * enum sym_type type; // should be set | |
278 | * | |
279 | * The type field is initially set to st_none; it should be set to | |
280 | * something else by the caller of stab_search. Other possible values | |
281 | * of an enum sym_type can be added. | |
282 | */ | |
283 | ||
284 | Stab * | |
285 | stab_create () | |
286 | { | |
287 | Stab *sp; | |
288 | sp = xnew (1, Stab); | |
289 | *sp = NULL; /* a Stab starts out as a null Stab_entry* */ | |
290 | return sp; | |
291 | } | |
292 | ||
293 | Stab_entry * | |
294 | stab_find (stab, sym, symlen) | |
295 | Stab *stab; | |
296 | register char *sym; | |
297 | register int symlen; | |
298 | { | |
299 | register Stab_entry *se; | |
300 | for (se = *stab; se != NULL; se = se->next) | |
301 | { | |
302 | if (se->symlen == symlen && strneq (se->sym, sym, symlen)) | |
303 | return se; | |
304 | } | |
305 | ||
306 | return NULL; | |
307 | } | |
308 | ||
309 | Stab_entry * | |
310 | stab_search (stab, sym, symlen) | |
311 | register Stab *stab; | |
312 | char *sym; | |
313 | int symlen; | |
314 | { | |
315 | register Stab_entry *se; | |
316 | se = stab_find (stab, sym, symlen); | |
317 | ||
318 | if (se == NULL) | |
319 | { | |
320 | /* make a new one */ | |
321 | se = xnew (1, Stab_entry); | |
322 | se->sym = savenstr (sym, symlen); | |
323 | se->symlen = symlen; | |
324 | se->type = st_none; | |
325 | se->next = *stab; | |
326 | *stab = se; | |
327 | } | |
328 | ||
329 | return se; | |
330 | } | |
331 | ||
332 | /* | |
333 | * NAME | |
334 | * stab_type -- type of a symbol table entry | |
335 | * | |
336 | * SYNOPSIS | |
337 | * enum sym_type stab_type (Stab_entry *se); | |
338 | * | |
339 | * WARNING | |
340 | * May evaluate its argument more than once. | |
341 | */ | |
342 | ||
343 | #define stab_type(se) ((se)==NULL ? st_none : (se)->type) | |
344 | \f | |
345 | ||
346 | ||
347 | typedef int LINENO; | |
348 | ||
349 | typedef struct | |
350 | { | |
351 | char *p; | |
352 | int len; | |
353 | FILEPOS linestart; | |
354 | LINENO lineno; | |
355 | logical rewritten; | |
356 | } TOKEN; | |
357 | \f | |
358 | ||
359 | /* typedefs are recognized using a simple finite automaton. | |
360 | * tydef is its state variable. | |
361 | */ | |
362 | typedef enum | |
363 | { | |
364 | none, begin, middle, end | |
365 | } TYST; | |
366 | ||
367 | TYST tydef = none; | |
368 | ||
369 | ||
370 | /* struct tags for C++ are recognized using another simple | |
371 | * finite automaton. `structdef' is its state variable. | |
372 | * This machinery is only invoked for C++; otherwise structdef | |
373 | * should remain snone. However, this machinery can easily be | |
374 | * adapted to find structure tags in normal C code. | |
375 | */ | |
376 | typedef enum | |
377 | { | |
378 | snone, /* nothing seen yet */ | |
379 | skeyseen, /* struct-like keyword seen */ | |
380 | stagseen, /* struct-like tag seen */ | |
381 | scolonseen, /* colon seen after struct-like tag */ | |
382 | sinbody /* in a class body: recognize member func defs */ | |
383 | } STRUCTST; | |
384 | STRUCTST structdef = snone; | |
385 | /* | |
386 | * When structdef is stagseen, scolonseen, or sinbody, structtag is the | |
387 | * struct tag, and structkey is the preceding struct-like keyword. | |
388 | */ | |
389 | char structtag[512]; | |
390 | Stab_entry *structkey; | |
391 | ||
392 | /* | |
393 | * Yet another little state machine to deal with preprocessor lines. | |
394 | */ | |
395 | typedef enum | |
396 | { | |
397 | dnone, /* nothing seen */ | |
398 | dsharpseen, /* '#' seen as first char on line */ | |
399 | ddefineseen, /* '#' and 'define' seen */ | |
400 | dignorerest /* ignore rest of line */ | |
401 | } DEFINEST; | |
402 | DEFINEST definedef; | |
403 | ||
404 | /* | |
405 | * LEVEL_OK_FOR_FUNCDEF allows C++ function definition within class body. | |
406 | * Currently tydef and structdef stuff (typedefs and struct definitions) are | |
407 | * only noticed when level==0, but that may change. | |
408 | * | |
409 | * Note that this macro may only be evaluated inside C_entries(). It is | |
410 | * for self-documentation only. | |
411 | */ | |
412 | #define LEVEL_OK_FOR_FUNCDEF() \ | |
daa37602 | 413 | (level==0 || (c_ext && level==1 && structdef==sinbody)) |
c6d46f5f | 414 | |
4746118a JB |
415 | /* |
416 | * next_token_is_func | |
417 | * set this to TRUE, and the next token considered is called a function. | |
418 | */ | |
419 | logical next_token_is_func; | |
420 | ||
c6d46f5f JB |
421 | /* C extensions. Currently all listed extensions are C++ dialects, so |
422 | * `c_ext' is used as an abbreviation for `c_ext&C_PLPL'. If a non-C++ | |
423 | * dialect is added, this must change. | |
424 | */ | |
425 | #define C_PLPL 0x1 /* C++ */ | |
426 | #define C_STAR 0x3 /* C* */ | |
427 | ||
428 | char searchar = '/'; /* use /.../ searches */ | |
429 | ||
430 | LINENO lineno; /* line number of current line */ | |
431 | long charno; /* current character number */ | |
432 | FILEPOS linepos; /* start of line (C only) */ | |
433 | FILEPOS prev_linepos; /* start of previous line (C only) */ | |
434 | ||
435 | long linecharno; /* charno of start of line; not used by C, but | |
436 | * by every other language. | |
437 | */ | |
438 | ||
439 | char *curfile, /* current input file name */ | |
440 | *outfile, /* output file */ | |
441 | *white = " \f\t\n", /* white chars */ | |
442 | *endtk = " \t\n\"'#()[]{}=-+%*/&|^~!<>;,.:?", /* token ending chars */ | |
443 | *begtk = "ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz$", /* token starting chars */ | |
444 | *intk = "ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz$0123456789", /* valid in-token chars */ | |
445 | *notgd = ",;"; /* non-valid after-function chars */ | |
446 | ||
c6d46f5f JB |
447 | int append_to_tagfile; /* -a: append to tags */ |
448 | int emacs_tags_format; /* emacs style output (no -e option any more) */ | |
449 | /* The following three default to 1 for etags, but to 0 for ctags. */ | |
450 | int typedefs; /* -t: create tags for typedefs */ | |
451 | int typedefs_and_cplusplus; /* -T: create tags for typedefs, level */ | |
452 | /* 0 struct/enum/union decls, and C++ */ | |
453 | /* member functions */ | |
454 | int constantypedefs; /* -d: create tags for C #define and enum */ | |
455 | /* constants. Default under etags. Enum */ | |
456 | /* constants not implemented. */ | |
457 | /* -D: opposite of -d. Default under ctags. */ | |
458 | int update; /* -u: update tags */ | |
459 | int vgrind_style; /* -v: create vgrind style index output */ | |
460 | int no_warnings; /* -w: suppress warnings */ | |
461 | int cxref_style; /* -x: create cxref style output */ | |
462 | int cplusplus; /* .[hc] means C++, not C */ | |
463 | int noindentypedefs; /* -S: ignore indentation in C */ | |
c6d46f5f JB |
464 | |
465 | /* Name this program was invoked with. */ | |
466 | char *progname; | |
467 | ||
4746118a JB |
468 | struct option longopts[] = { |
469 | { "append", no_argument, NULL, 'a' }, | |
470 | { "backward-search", no_argument, NULL, 'B' }, | |
471 | { "c++", no_argument, NULL, 'C' }, | |
472 | { "cxref", no_argument, NULL, 'x' }, | |
473 | { "defines", no_argument, NULL, 'd' }, | |
474 | { "forward-search", no_argument, NULL, 'F' }, | |
475 | { "help", no_argument, NULL, 'H' }, | |
476 | { "ignore-indentation", no_argument, NULL, 'S' }, | |
477 | { "include", required_argument, NULL, 'i' }, | |
478 | { "no-defines", no_argument, NULL, 'D' }, | |
479 | { "no-warn", no_argument, NULL, 'w' }, | |
480 | { "output", required_argument, NULL, 'o' }, | |
481 | { "typedefs", no_argument, NULL, 't' }, | |
482 | { "typedefs-and-c++", no_argument, NULL, 'T' }, | |
483 | { "update", no_argument, NULL, 'u' }, | |
484 | { "version", no_argument, NULL, 'V' }, | |
485 | { "vgrind", no_argument, NULL, 'v' }, | |
486 | { 0 } | |
487 | }; | |
488 | ||
c6d46f5f JB |
489 | FILE *inf, /* ioptr for current input file */ |
490 | *outf; /* ioptr for tags file */ | |
491 | ||
492 | NODE *head; /* the head of the binary tree of tags */ | |
493 | ||
494 | int permit_duplicates = 1; /* Nonzero means allow duplicate tags. */ | |
495 | ||
496 | /* A `struct linebuffer' is a structure which holds a line of text. | |
497 | `readline' reads a line from a stream into a linebuffer | |
498 | and works regardless of the length of the line. */ | |
499 | ||
500 | struct linebuffer | |
501 | { | |
502 | long size; | |
503 | char *buffer; | |
504 | }; | |
505 | ||
506 | struct linebuffer lb; /* the current line */ | |
507 | struct linebuffer lb1; /* sometimes, a previous line in which a token lies */ | |
508 | struct linebuffer filename_lb; /* used to read in filenames */ | |
509 | \f | |
510 | ||
4746118a JB |
511 | void |
512 | print_version () | |
513 | { | |
514 | #ifdef CTAGS | |
515 | printf ("CTAGS "); | |
516 | #ifdef ETAGS | |
517 | printf ("and "); | |
518 | #endif | |
519 | #endif | |
520 | #ifdef ETAGS | |
521 | printf ("ETAGS "); | |
522 | #endif | |
523 | printf ("for Emacs version 19.0.\n"); | |
524 | ||
525 | exit (0); | |
526 | } | |
527 | ||
528 | void | |
529 | print_help () | |
530 | { | |
531 | printf ("These are the options accepted by %s. You may use unambiguous\n\ | |
532 | abbreviations for the long option names.\n\n", progname); | |
533 | ||
534 | fputs ("\ | |
535 | -a, --append\n\ | |
536 | Append tag entries to existing tags file.\n\ | |
537 | -C, --c++\n\ | |
538 | Treat files with `.c' and `.h' extensions as C++ code, not C\n\ | |
539 | code. Files with `.C', `.H', `.cxx', `.hxx', or `.cc'\n\ | |
540 | extensions are always assumed to be C++ code.\n\ | |
541 | -d, --defines\n\ | |
542 | Create tag entries for #defines, too.", stdout); | |
543 | ||
544 | #ifdef ETAGS | |
545 | fputs (" This is the default\n\ | |
546 | behavior.", stdout); | |
547 | #endif | |
548 | ||
549 | fputs ("\n\ | |
550 | -D, --no-defines\n\ | |
551 | Don't create tag entries for #defines.", stdout); | |
552 | ||
553 | #ifdef CTAGS | |
554 | fputs (" This is the default\n\ | |
555 | behavior.", stdout); | |
556 | #endif | |
557 | ||
558 | puts ("\n\ | |
559 | -o FILE, --output=FILE\n\ | |
560 | Write the tags to FILE.\n\ | |
561 | -S, --ignore-indentation\n\ | |
562 | Don't rely on indentation quite as much as normal. Currently,\n\ | |
563 | this means not to assume that a closing brace in the first\n\ | |
564 | column is the final brace of a function or structure\n\ | |
565 | definition.\n\ | |
566 | -t, --typedefs\n\ | |
567 | Generate tag entries for typedefs. This is the default\n\ | |
568 | behavior.\n\ | |
569 | -T, --typedefs-and-c++\n\ | |
570 | Generate tag entries for typedefs, struct/enum/union tags, and\n\ | |
571 | C++ member functions."); | |
572 | ||
573 | #ifdef ETAGS | |
574 | puts ("-i FILE, --include=FILE\n\ | |
575 | Include a note in tag file indicating that, when searching for\n\ | |
576 | a tag, one should also consult the tags file FILE after\n\ | |
577 | checking the current file."); | |
578 | #endif | |
579 | ||
580 | #ifdef CTAGS | |
581 | puts ("-B, --backward-search\n\ | |
582 | Write the search commands for the tag entries using '?', the\n\ | |
583 | backward-search command.\n\ | |
584 | -F, --forward-search\n\ | |
585 | Write the search commands for the tag entries using '/', the\n\ | |
586 | forward-search command.\n\ | |
587 | -u, --update\n\ | |
588 | Update the tag entries for the given files, leaving tag\n\ | |
589 | entries for other files in place. Currently, this is\n\ | |
590 | implemented by deleting the existing entries for the given\n\ | |
591 | files and then rewriting the new entries at the end of the\n\ | |
592 | tags file. It is often faster to simply rebuild the entire\n\ | |
593 | tag file than to use this.\n\ | |
594 | -v, --vgrind\n\ | |
595 | Generates an index of items intended for human consumption,\n\ | |
596 | similar to the output of vgrind. The index is sorted, and\n\ | |
597 | gives the page number of each item.\n\ | |
598 | -x, --cxref\n\ | |
599 | Like --vgrind, but in the style of cxref, rather than vgrind.\n\ | |
600 | The output uses line numbers instead of page numbers, but\n\ | |
601 | beyond that the differences are cosmetic; try both to see\n\ | |
602 | which you like.\n\ | |
603 | -w, --no-warn\n\ | |
604 | Suppress warning messages about entries defined in multiple\n\ | |
605 | files."); | |
606 | #endif | |
607 | ||
608 | puts ("-V, --version\n\ | |
609 | Print the version of the program.\n\ | |
610 | -H, --help\n\ | |
611 | Print this help message."); | |
612 | ||
613 | exit (0); | |
614 | } | |
615 | ||
616 | \f | |
c6d46f5f JB |
617 | void |
618 | main (argc, argv) | |
619 | int argc; | |
620 | char *argv[]; | |
621 | { | |
622 | char cmd[100]; | |
623 | int i; | |
1e134a5f | 624 | unsigned int nincluded_files = 0; |
b960e009 | 625 | char **included_files = (char **) alloca (argc * sizeof (char *)); |
c6d46f5f JB |
626 | char *this_file; |
627 | #ifdef VMS | |
628 | char got_err; | |
629 | ||
630 | extern char *gfnames (); | |
631 | extern char *massage_name (); | |
632 | #endif | |
633 | ||
634 | progname = argv[0]; | |
635 | ||
636 | #ifndef CTAGS | |
637 | emacs_tags_format = 1; | |
638 | #else | |
639 | emacs_tags_format = 0; | |
640 | #endif | |
641 | ||
642 | /* | |
643 | * If etags, always find typedefs and structure tags. Why not? | |
644 | * Also default is to find macro constants. | |
645 | */ | |
646 | if (emacs_tags_format) | |
647 | typedefs = typedefs_and_cplusplus = constantypedefs = 1; | |
648 | ||
4746118a | 649 | for (;;) |
c6d46f5f | 650 | { |
4746118a JB |
651 | int opt; |
652 | opt = getopt_long (argc, argv, "aCdDo:StTi:BFuvxwVH", longopts, 0); | |
653 | ||
654 | if (opt == EOF) | |
655 | break; | |
656 | ||
657 | switch (opt) | |
c6d46f5f | 658 | { |
4746118a JB |
659 | case '\0': |
660 | /* If getopt returns '\0', then it has already processed a | |
661 | long-named option. We should do nothing. */ | |
662 | break; | |
663 | ||
664 | /* Common options. */ | |
665 | case 'a': | |
666 | append_to_tagfile++; | |
667 | break; | |
668 | case 'C': | |
669 | cplusplus = 1; | |
670 | break; | |
671 | case 'd': | |
672 | constantypedefs = 1; | |
673 | break; | |
674 | case 'D': | |
675 | constantypedefs = 0; | |
676 | break; | |
677 | case 'o': | |
678 | if (outfile) | |
c6d46f5f | 679 | { |
4746118a JB |
680 | fprintf (stderr, |
681 | "%s: -o flag may only be given once\n", progname); | |
c6d46f5f JB |
682 | goto usage; |
683 | } | |
4746118a JB |
684 | outfile = optarg; |
685 | break; | |
686 | case 'S': | |
687 | noindentypedefs++; | |
688 | break; | |
689 | case 't': | |
690 | typedefs++; | |
691 | break; | |
692 | case 'T': | |
693 | typedefs++; | |
694 | typedefs_and_cplusplus++; | |
695 | break; | |
696 | case 'V': | |
697 | print_version (); | |
698 | break; | |
699 | case 'H': | |
700 | print_help (); | |
701 | break; | |
702 | ||
703 | /* Etags options */ | |
704 | case 'i': | |
705 | if (!emacs_tags_format) | |
706 | goto usage; | |
707 | included_files[nincluded_files++] = optarg; | |
708 | break; | |
709 | ||
710 | /* Ctags options. */ | |
711 | case 'B': | |
712 | searchar = '?'; | |
713 | if (emacs_tags_format) | |
714 | goto usage; | |
715 | break; | |
716 | case 'F': | |
717 | searchar = '/'; | |
718 | if (emacs_tags_format) | |
719 | goto usage; | |
720 | break; | |
721 | case 'u': | |
722 | update++; | |
723 | if (emacs_tags_format) | |
724 | goto usage; | |
725 | break; | |
726 | case 'v': | |
727 | vgrind_style++; | |
728 | /*FALLTHRU*/ | |
729 | case 'x': | |
730 | cxref_style++; | |
731 | if (emacs_tags_format) | |
732 | goto usage; | |
733 | break; | |
734 | case 'w': | |
735 | no_warnings++; | |
736 | if (emacs_tags_format) | |
737 | goto usage; | |
738 | break; | |
739 | ||
740 | default: | |
741 | goto usage; | |
c6d46f5f | 742 | } |
c6d46f5f JB |
743 | } |
744 | ||
4746118a | 745 | if (optind == argc) |
c6d46f5f | 746 | { |
4746118a JB |
747 | fprintf (stderr, "%s: No input files specified.\n", progname); |
748 | ||
c6d46f5f | 749 | usage: |
4746118a JB |
750 | fprintf (stderr, "%s: Try '%s --help' for a complete list of options.\n", |
751 | progname, progname); | |
c6d46f5f JB |
752 | exit (BAD); |
753 | } | |
754 | ||
755 | if (outfile == 0) | |
756 | { | |
757 | outfile = emacs_tags_format ? "TAGS" : "tags"; | |
758 | } | |
759 | ||
760 | init (); /* set up boolean "functions" */ | |
761 | ||
762 | initbuffer (&lb); | |
763 | initbuffer (&lb1); | |
764 | initbuffer (&filename_lb); | |
765 | /* | |
766 | * loop through files finding functions | |
767 | */ | |
768 | if (emacs_tags_format) | |
769 | { | |
770 | if (streq (outfile, "-")) | |
771 | outf = stdout; | |
772 | else | |
773 | outf = fopen (outfile, append_to_tagfile ? "a" : "w"); | |
774 | if (!outf) | |
775 | { | |
776 | perror (outfile); | |
777 | exit (1); | |
778 | } | |
779 | } | |
780 | ||
c6d46f5f | 781 | #ifdef VMS |
4746118a JB |
782 | argc -= optind; |
783 | argv += optind; | |
784 | while (gfnames (&argc, &argv, &got_err) != NULL) | |
c6d46f5f JB |
785 | { |
786 | if (got_err) | |
787 | { | |
788 | error ("Can't find file %s\n", this_file); | |
789 | argc--, argv++; | |
790 | } | |
791 | else | |
792 | { | |
793 | this_file = massage_name (this_file); | |
794 | #if 0 | |
795 | } | |
796 | } /* solely to balance out the ifdef'd parens above */ | |
797 | #endif | |
798 | #else | |
4746118a | 799 | for (; optind < argc; optind++) |
c6d46f5f | 800 | { |
4746118a | 801 | this_file = argv[optind]; |
c6d46f5f JB |
802 | if (1) |
803 | { | |
804 | #endif | |
805 | /* Input file named "-" means read file names from stdin | |
806 | and use them. */ | |
807 | if (streq (this_file, "-")) | |
808 | { | |
809 | while (!feof (stdin)) | |
810 | { | |
811 | (void) readline (&filename_lb, stdin); | |
812 | if (strlen (filename_lb.buffer) > 0) | |
813 | process_file (filename_lb.buffer); | |
814 | } | |
815 | } | |
816 | else | |
817 | process_file (this_file); | |
818 | } | |
819 | } | |
820 | ||
821 | if (emacs_tags_format) | |
822 | { | |
1e134a5f RM |
823 | while (nincluded_files-- > 0) |
824 | fprintf (outf, "\f\n%s,include\n", *included_files++); | |
825 | ||
c6d46f5f JB |
826 | (void) fclose (outf); |
827 | exit (0); | |
828 | } | |
829 | ||
830 | if (cxref_style) | |
831 | { | |
832 | put_entries (head); | |
833 | exit (GOOD); | |
834 | } | |
4746118a | 835 | if (update) |
c6d46f5f | 836 | { |
4746118a JB |
837 | /* update cannot be set under VMS, so we may assume that argc |
838 | and argv have not been munged. */ | |
839 | for (i = optind; i < argc; i++) | |
c6d46f5f JB |
840 | { |
841 | sprintf (cmd, | |
842 | "mv %s OTAGS;fgrep -v '\t%s\t' OTAGS >%s;rm OTAGS", | |
843 | outfile, argv[i], outfile); | |
844 | (void) system (cmd); | |
845 | } | |
846 | append_to_tagfile++; | |
847 | } | |
848 | outf = fopen (outfile, append_to_tagfile ? "a" : "w"); | |
849 | if (outf == NULL) | |
850 | { | |
851 | perror (outfile); | |
852 | exit (GOOD); | |
853 | } | |
854 | put_entries (head); | |
855 | (void) fclose (outf); | |
856 | if (update) | |
857 | { | |
858 | sprintf (cmd, "sort %s -o %s", outfile, outfile); | |
859 | (void) system (cmd); | |
860 | } | |
861 | exit (GOOD); | |
862 | } | |
863 | ||
864 | ||
865 | /* | |
866 | * This routine is called on each file argument. | |
867 | */ | |
868 | void | |
869 | process_file (file) | |
870 | char *file; | |
871 | { | |
872 | struct stat stat_buf; | |
873 | ||
874 | stat (file, &stat_buf); | |
875 | if (!(stat_buf.st_mode & S_IFREG) || !(stat_buf.st_mode & S_IFLNK)) | |
876 | { | |
877 | fprintf (stderr, "Skipping %s: it is not a regular file.\n", file); | |
878 | return; | |
879 | } | |
880 | ||
881 | if (streq (file, outfile) && !streq (outfile, "-")) | |
882 | { | |
883 | fprintf (stderr, "Skipping inclusion of %s in self.\n", file); | |
884 | return; | |
885 | } | |
c6d46f5f JB |
886 | if (emacs_tags_format) |
887 | { | |
888 | char *cp = rindex (file, '/'); | |
889 | if (cp) | |
890 | ++cp; | |
891 | else | |
892 | cp = file; | |
c6d46f5f JB |
893 | } |
894 | find_entries (file); | |
895 | if (emacs_tags_format) | |
896 | { | |
897 | fprintf (outf, "\f\n%s,%d\n", | |
898 | file, total_size_of_entries (head)); | |
899 | put_entries (head); | |
900 | free_tree (head); | |
901 | head = NULL; | |
902 | } | |
903 | } | |
904 | ||
905 | /* | |
906 | * This routine sets up the boolean psuedo-functions which work | |
907 | * by seting boolean flags dependent upon the corresponding character | |
908 | * Every char which is NOT in that string is not a white char. Therefore, | |
909 | * all of the array "_wht" is set to FALSE, and then the elements | |
910 | * subscripted by the chars in "white" are set to TRUE. Thus "_wht" | |
911 | * of a char is TRUE if it is the string "white", else FALSE. | |
912 | */ | |
913 | void | |
914 | init () | |
915 | { | |
916 | reg char *sp; | |
917 | reg int i; | |
918 | ||
919 | for (i = 0; i < 0177; i++) | |
920 | { | |
921 | _wht[i] = _etk[i] = _itk[i] = _btk[i] = FALSE; | |
922 | _gd[i] = TRUE; | |
923 | } | |
924 | for (sp = white; *sp; sp++) | |
925 | _wht[*sp] = TRUE; | |
926 | for (sp = endtk; *sp; sp++) | |
927 | _etk[*sp] = TRUE; | |
928 | for (sp = intk; *sp; sp++) | |
929 | _itk[*sp] = TRUE; | |
930 | for (sp = begtk; *sp; sp++) | |
931 | _btk[*sp] = TRUE; | |
932 | for (sp = notgd; *sp; sp++) | |
933 | _gd[*sp] = FALSE; | |
934 | _wht[0] = _wht['\n']; | |
935 | _etk[0] = _etk['\n']; | |
936 | _btk[0] = _btk['\n']; | |
937 | _itk[0] = _itk['\n']; | |
938 | _gd[0] = _gd['\n']; | |
939 | } | |
940 | ||
941 | /* | |
942 | * This routine opens the specified file and calls the function | |
943 | * which finds the function and type definitions. | |
944 | */ | |
945 | void | |
946 | find_entries (file) | |
947 | char *file; | |
948 | { | |
949 | char *cp; | |
950 | void prolog_funcs (); | |
951 | ||
952 | inf = fopen (file, "r"); | |
953 | if (inf == NULL) | |
954 | { | |
955 | perror (file); | |
956 | return; | |
957 | } | |
958 | curfile = savestr (file); | |
959 | cp = rindex (file, '.'); | |
960 | ||
961 | header_file = (cp && (streq (cp + 1, "h"))); | |
962 | ||
963 | /* .tex, .aux or .bbl implies LaTeX source code */ | |
964 | if (cp && (streq (cp + 1, "tex") || streq (cp + 1, "aux") | |
965 | || streq (cp + 1, "bbl"))) | |
966 | { | |
967 | TEX_funcs (inf); | |
968 | goto close_and_return; | |
969 | } | |
970 | /* .l or .el or .lisp (or .cl or .clisp or ...) implies lisp source code */ | |
971 | if (cp && (streq (cp + 1, "l") | |
972 | || streq (cp + 1, "el") | |
973 | || streq (cp + 1, "lsp") | |
974 | || streq (cp + 1, "lisp") | |
975 | || streq (cp + 1, "cl") | |
976 | || streq (cp + 1, "clisp"))) | |
977 | { | |
978 | L_funcs (inf); | |
979 | goto close_and_return; | |
980 | } | |
981 | /* .scm or .sm or .scheme or ... implies scheme source code */ | |
982 | if (cp && (streq (cp + 1, "sm") | |
983 | || streq (cp + 1, "scm") | |
984 | || streq (cp + 1, "scheme") | |
985 | || streq (cp + 1, "t") | |
986 | || streq (cp + 1, "sch") | |
987 | || streq (cp + 1, "SM") | |
988 | || streq (cp + 1, "SCM") | |
989 | /* The `SCM' or `scm' prefix with a version number */ | |
990 | || (cp[-1] == 'm' && cp[-2] == 'c' && cp[-3] == 's' | |
991 | && string_numeric_p (cp + 1)) | |
992 | || (cp[-1] == 'M' && cp[-2] == 'C' && cp[-3] == 'S' | |
993 | && string_numeric_p (cp + 1)))) | |
994 | { | |
995 | Scheme_funcs (inf); | |
996 | fclose (inf); | |
997 | return; | |
998 | } | |
999 | /* Assume that ".s" or ".a" is assembly code. -wolfgang. */ | |
1000 | if (cp && (cp[1] == 's' || cp[1] == 'a') && cp[2] == '\0') | |
1001 | { | |
1002 | Asm_funcs (inf); | |
1003 | fclose (inf); | |
1004 | return; | |
1005 | } | |
1006 | /* .C or .H or .cxx or .hxx or .cc: a C++ file */ | |
1007 | if (cp && (streq (cp + 1, "C") | |
1008 | || streq (cp + 1, "H") | |
1009 | || streq (cp + 1, "cxx") | |
1010 | || streq (cp + 1, "hxx") | |
1011 | || streq (cp + 1, "cc"))) | |
1012 | { | |
1013 | C_entries (C_PLPL); /* C++ */ | |
1014 | goto close_and_return; | |
1015 | } | |
1016 | /* .cs or .hs: a C* file */ | |
1017 | if (cp && (cp[1] == 'c' || cp[1] == 'h') && cp[2] == 's' && cp[3] == '\0') | |
1018 | { | |
1019 | C_entries (C_STAR); | |
1020 | goto close_and_return; | |
1021 | } | |
1022 | /* .pl implies prolog source code */ | |
1023 | if (cp && !strcmp (cp + 1, "pl")) | |
1024 | { | |
1025 | prolog_funcs (inf); | |
1026 | goto close_and_return; | |
1027 | } | |
1028 | /* .p or .pas: a Pascal file */ | |
1029 | if (cp && (streq (cp + 1, "p") | |
1030 | || streq (cp + 1, "pas"))) | |
1031 | { | |
1032 | PAS_funcs (inf); | |
1033 | goto close_and_return; | |
1034 | } | |
1035 | /* if not a .c or .h or .y file, try fortran */ | |
4746118a JB |
1036 | else if (cp && ((cp[1] != 'c' |
1037 | && cp[1] != 'h' | |
1038 | && cp[1] != 'y') | |
c6d46f5f JB |
1039 | || (cp[1] != 0 && cp[2] != 0))) |
1040 | { | |
1041 | if (PF_funcs (inf) != 0) | |
1042 | goto close_and_return; | |
1043 | rewind (inf); /* no fortran tags found, try C */ | |
1044 | } | |
1045 | C_entries (cplusplus ? C_PLPL : 0); | |
1046 | ||
1047 | close_and_return: | |
1048 | (void) fclose (inf); | |
1049 | } | |
1050 | ||
1051 | /* Nonzero if string STR is composed of digits. */ | |
1052 | ||
1053 | int | |
1054 | string_numeric_p (str) | |
1055 | char *str; | |
1056 | { | |
1057 | while (*str) | |
1058 | { | |
1059 | if (*str < '0' || *str > '9') | |
1060 | return 0; | |
1061 | } | |
1062 | return 1; | |
1063 | } | |
1064 | \f | |
1065 | /* Record a tag. */ | |
1066 | /* Should take a TOKEN* instead!! */ | |
1067 | ||
1068 | void | |
1069 | pfnote (name, is_func, rewritten, linestart, linelen, lno, cno) | |
1070 | char *name; /* tag name */ | |
1071 | logical is_func; /* function or type name? */ | |
1072 | logical rewritten; /* tag different from text of definition? */ | |
1073 | char *linestart; | |
1074 | int linelen; | |
1075 | int lno; | |
1076 | long cno; | |
1077 | { | |
1078 | register char *fp; | |
1079 | register NODE *np; | |
1080 | char tem[51]; | |
1081 | char c; | |
1082 | ||
1083 | np = (NODE *) malloc (sizeof (NODE)); | |
1084 | if (np == NULL) | |
1085 | { | |
1086 | if (!emacs_tags_format) | |
1087 | { | |
1088 | /* It's okay to output early in etags -- it only disrupts the | |
1089 | * character count of the tag entries, which is no longer used | |
1090 | * by tags.el anyway. | |
1091 | */ | |
1092 | error ("too many entries to sort"); | |
1093 | } | |
1094 | put_entries (head); | |
1095 | free_tree (head); | |
1096 | head = NULL; | |
1097 | np = xnew (1, NODE); | |
1098 | } | |
1099 | /* If ctags mode, change name "main" to M<thisfilename>. */ | |
1100 | if (!emacs_tags_format && !cxref_style && streq (name, "main")) | |
1101 | { | |
1102 | fp = rindex (curfile, '/'); | |
1103 | name = concat ("M", fp == 0 ? curfile : fp + 1, ""); | |
1104 | fp = rindex (name, '.'); | |
1105 | if (fp && fp[1] != '\0' && fp[2] == '\0') | |
1106 | *fp = 0; | |
1107 | rewritten = TRUE; | |
1108 | } | |
1109 | np->name = savestr (name); | |
1110 | np->file = curfile; | |
1111 | np->is_func = is_func; | |
1112 | np->rewritten = rewritten; | |
1113 | np->lno = lno; | |
1114 | /* UNCOMMENT THE +1 HERE: */ | |
1115 | np->cno = cno /* + 1 */ ; /* our char numbers are 0-base; emacs's are 1-base */ | |
1116 | np->left = np->right = 0; | |
1117 | if (emacs_tags_format) | |
1118 | { | |
1119 | c = linestart[linelen]; | |
1120 | linestart[linelen] = 0; | |
1121 | } | |
1122 | else if (cxref_style == 0) | |
1123 | { | |
1124 | sprintf (tem, strlen (linestart) < 50 ? "%s$" : "%.50s", linestart); | |
1125 | linestart = tem; | |
1126 | } | |
1127 | np->pat = savestr (linestart); | |
1128 | if (emacs_tags_format) | |
1129 | { | |
1130 | linestart[linelen] = c; | |
1131 | } | |
1132 | ||
1133 | add_node (np, &head); | |
1134 | } | |
1135 | ||
1136 | /* | |
1137 | * free_tree () | |
1138 | * recurse on left children, iterate on right children. | |
1139 | */ | |
1140 | void | |
1141 | free_tree (node) | |
1142 | register NODE *node; | |
1143 | { | |
1144 | while (node) | |
1145 | { | |
1146 | register NODE *node_right = node->right; | |
1147 | free_tree (node->left); | |
1148 | free (node->name); | |
1149 | free (node->pat); | |
1150 | free ((char *) node); | |
1151 | node = node_right; | |
1152 | } | |
1153 | } | |
1154 | ||
1155 | /* | |
1156 | * add_node () | |
1157 | * Adds a node to the tree of nodes. In etags mode, we don't keep | |
1158 | * it sorted; we just keep a linear list. In ctags mode, maintain | |
1159 | * an ordered tree, with no attempt at balancing. | |
1160 | * | |
1161 | * add_node is the only function allowed to add nodes, so it can | |
1162 | * maintain state. | |
1163 | */ | |
1164 | void | |
1165 | add_node (node, cur_node_p) | |
1166 | NODE *node, **cur_node_p; | |
1167 | { | |
1168 | register int dif; | |
1169 | register NODE *cur_node = *cur_node_p; | |
1170 | static NODE *last_node = NULL;/* careful */ | |
1171 | ||
1172 | if (cur_node == NULL) | |
1173 | { | |
1174 | *cur_node_p = node; | |
1175 | last_node = node; | |
1176 | return; | |
1177 | } | |
1178 | ||
1179 | if (emacs_tags_format) | |
1180 | { | |
1181 | /* Etags Mode */ | |
1182 | if (!last_node) | |
1183 | fatal ("internal error in add_node"); | |
1184 | last_node->right = node; | |
1185 | last_node = node; | |
1186 | } | |
1187 | else | |
1188 | { | |
1189 | /* Ctags Mode */ | |
1190 | dif = strcmp (node->name, cur_node->name); | |
1191 | ||
1192 | /* | |
1193 | * If this tag name matches an existing one, then | |
1194 | * do not add the node, but maybe print a warning. | |
1195 | */ | |
1196 | if (!dif) | |
1197 | { | |
1198 | if (node->file == cur_node->file) | |
1199 | { | |
1200 | if (!no_warnings) | |
1201 | { | |
1202 | fprintf (stderr, "Duplicate entry in file %s, line %d: %s\n", | |
1203 | node->file, lineno, node->name); | |
1204 | fprintf (stderr, "Second entry ignored\n"); | |
1205 | } | |
1206 | return; | |
1207 | } | |
1208 | if (!cur_node->been_warned && !no_warnings) | |
1209 | { | |
1210 | fprintf (stderr, | |
1211 | "Duplicate entry in files %s and %s: %s (Warning only)\n", | |
1212 | node->file, cur_node->file, node->name); | |
1213 | } | |
1214 | cur_node->been_warned = TRUE; | |
1215 | return; | |
1216 | } | |
1217 | ||
1218 | /* Maybe refuse to add duplicate nodes. */ | |
1219 | if (!permit_duplicates) | |
1220 | { | |
1221 | if (!strcmp (node->name, cur_node->name) | |
1222 | && !strcmp (node->file, cur_node->file)) | |
1223 | return; | |
1224 | } | |
1225 | ||
1226 | /* Actually add the node */ | |
1227 | add_node (node, dif < 0 ? &cur_node->left : &cur_node->right); | |
1228 | } | |
1229 | } | |
1230 | \f | |
1231 | void | |
1232 | put_entries (node) | |
1233 | reg NODE *node; | |
1234 | { | |
1235 | reg char *sp; | |
1236 | ||
1237 | if (node == NULL) | |
1238 | return; | |
1239 | ||
1240 | /* Output subentries that precede this one */ | |
1241 | put_entries (node->left); | |
1242 | ||
1243 | /* Output this entry */ | |
1244 | ||
1245 | if (emacs_tags_format) | |
1246 | { | |
1247 | if (node->rewritten) | |
1248 | { | |
1249 | fprintf (outf, "%s\177%s\001%d,%d\n", | |
1250 | node->name, node->pat, node->lno, node->cno); | |
1251 | } | |
1252 | else | |
1253 | { | |
1254 | fprintf (outf, "%s\177%d,%d\n", | |
1255 | node->pat, node->lno, node->cno); | |
1256 | } | |
1257 | } | |
1258 | else if (!cxref_style) | |
1259 | { | |
1260 | fprintf (outf, "%s\t%s\t", | |
1261 | node->name, node->file); | |
1262 | ||
1263 | if (node->is_func) | |
1264 | { /* a function */ | |
1265 | putc (searchar, outf); | |
1266 | putc ('^', outf); | |
1267 | ||
1268 | for (sp = node->pat; *sp; sp++) | |
1269 | { | |
1270 | if (*sp == '\\' || *sp == searchar) | |
1271 | putc ('\\', outf); | |
1272 | putc (*sp, outf); | |
1273 | } | |
1274 | putc (searchar, outf); | |
1275 | } | |
1276 | else | |
1277 | { /* a typedef; text pattern inadequate */ | |
1278 | fprintf (outf, "%d", node->lno); | |
1279 | } | |
1280 | putc ('\n', outf); | |
1281 | } | |
1282 | else if (vgrind_style) | |
1283 | fprintf (stdout, "%s %s %d\n", | |
1284 | node->name, node->file, (node->lno + 63) / 64); | |
1285 | else | |
daa37602 | 1286 | fprintf (stdout, "%-16s %3d %-16s %s\n", |
c6d46f5f JB |
1287 | node->name, node->lno, node->file, node->pat); |
1288 | ||
1289 | /* Output subentries that follow this one */ | |
1290 | put_entries (node->right); | |
1291 | } | |
1292 | ||
1293 | /* Length of a number's decimal representation. */ | |
1294 | int | |
1295 | number_len (num) | |
1296 | long num; | |
1297 | { | |
1298 | int len = 0; | |
1299 | if (!num) | |
1300 | return 1; | |
1301 | for (; num; num /= 10) | |
1302 | ++len; | |
1303 | return len; | |
1304 | } | |
1305 | ||
1306 | /* | |
1307 | * Return total number of characters that put_entries will output for | |
1308 | * the nodes in the subtree of the specified node. Works only if emacs_tags_format | |
1309 | * is set, but called only in that case. This count is irrelevant with | |
1310 | * the new tags.el, but is still supplied for backward compatibility. | |
1311 | */ | |
1312 | int | |
1313 | total_size_of_entries (node) | |
1314 | reg NODE *node; | |
1315 | { | |
1316 | reg int total; | |
1317 | ||
1318 | if (node == NULL) | |
1319 | return 0; | |
1320 | ||
1321 | total = 0; | |
1322 | for (; node; node = node->right) | |
1323 | { | |
1324 | /* Count left subentries. */ | |
1325 | total += total_size_of_entries (node->left); | |
1326 | ||
1327 | /* Count this entry */ | |
1328 | total += strlen (node->pat) + 1; | |
1329 | total += number_len ((long) node->lno) + 1 + number_len (node->cno) + 1; | |
1330 | if (node->rewritten) | |
1331 | total += 1 + strlen (node->name); /* \001name */ | |
1332 | } | |
1333 | ||
1334 | return total; | |
1335 | } | |
1336 | \f | |
1337 | /* | |
1338 | * The C symbol tables. | |
1339 | */ | |
1340 | ||
1341 | Stab *C_stab, *C_PLPL_stab, *C_STAR_stab; | |
1342 | ||
1343 | /* | |
1344 | * SYNOPSIS | |
1345 | * Stab *get_C_stab (int c_ext); | |
1346 | */ | |
1347 | #define get_C_stab(c_ext) ((c_ext&C_STAR) ? C_STAR_stab : \ | |
1348 | c_ext ? C_PLPL_stab : \ | |
1349 | C_stab) | |
1350 | ||
1351 | void | |
1352 | add_keyword (stab, sym, type) | |
1353 | Stab *stab; | |
1354 | char *sym; | |
1355 | enum sym_type type; | |
1356 | { | |
1357 | stab_search (stab, sym, strlen (sym))->type = type; | |
1358 | } | |
1359 | ||
1360 | Stab * | |
1361 | C_create_stab (c_ext) | |
1362 | int c_ext; | |
1363 | { | |
1364 | Stab *stab; | |
1365 | ||
1366 | stab = stab_create (); | |
1367 | ||
1368 | /* C, C++ and C* */ | |
1369 | if (c_ext & C_PLPL) | |
1370 | add_keyword (stab, "class", st_C_struct); | |
1371 | if (c_ext & C_STAR) | |
1372 | add_keyword (stab, "domain", st_C_struct); | |
1373 | add_keyword (stab, "union", st_C_struct); | |
1374 | add_keyword (stab, "struct", st_C_struct); | |
1375 | add_keyword (stab, "enum", st_C_enum); | |
1376 | add_keyword (stab, "typedef", st_C_typedef); | |
1377 | add_keyword (stab, "define", st_C_define); | |
1378 | add_keyword (stab, "long", st_C_typespec); | |
1379 | add_keyword (stab, "short", st_C_typespec); | |
1380 | add_keyword (stab, "int", st_C_typespec); | |
1381 | add_keyword (stab, "char", st_C_typespec); | |
1382 | add_keyword (stab, "float", st_C_typespec); | |
1383 | add_keyword (stab, "double", st_C_typespec); | |
1384 | add_keyword (stab, "signed", st_C_typespec); | |
1385 | add_keyword (stab, "unsigned", st_C_typespec); | |
1386 | add_keyword (stab, "const", st_C_typespec); | |
1387 | add_keyword (stab, "volatile", st_C_typespec); | |
1388 | ||
1389 | return stab; | |
1390 | } | |
1391 | ||
1392 | void | |
1393 | C_create_stabs () | |
1394 | { | |
1395 | C_stab = C_create_stab (0); | |
1396 | C_PLPL_stab = C_create_stab (C_PLPL); | |
1397 | C_STAR_stab = C_create_stab (C_STAR | C_PLPL); | |
1398 | } | |
1399 | \f | |
1400 | /* | |
1401 | * C_entries () | |
1402 | * This routine finds functions and typedefs in C syntax and adds them | |
1403 | * to the list. | |
1404 | */ | |
1405 | ||
1406 | #define CNL_SAVE_DEFINEDEF \ | |
1407 | { \ | |
1408 | prev_linepos = linepos; \ | |
1409 | SET_FILEPOS (linepos, inf, charno); \ | |
1410 | lineno++; \ | |
1411 | charno += readline (&lb, inf); \ | |
1412 | lp = lb.buffer; \ | |
1413 | } | |
1414 | ||
1415 | #define CNL \ | |
1416 | { \ | |
1417 | CNL_SAVE_DEFINEDEF; \ | |
1418 | definedef = dnone; \ | |
1419 | } | |
1420 | ||
1421 | void | |
1422 | C_entries (c_ext) | |
1423 | int c_ext; /* extension of C? */ | |
1424 | { | |
1425 | register int c; /* latest char read; '\0' for end of line */ | |
1426 | register int tokoff; /* offset in line of beginning of latest token */ | |
1427 | register int toklen; /* length of latest token */ | |
1428 | register char *lp; /* pointer one beyond the character `c' */ | |
1429 | logical incomm, inquote, inchar, midtoken; | |
1430 | int level; /* current curly brace level */ | |
1431 | char tokb[BUFSIZ]; | |
1432 | ||
1433 | lineno = 0; | |
1434 | charno = 0; | |
1435 | lp = lb.buffer; | |
1436 | *lp = 0; | |
1437 | ||
1438 | definedef = dnone; | |
1439 | gotone = midtoken = inquote = inchar = incomm = FALSE; | |
1440 | level = 0; | |
4746118a JB |
1441 | tydef = none; |
1442 | next_token_is_func = 0; | |
c6d46f5f JB |
1443 | |
1444 | C_create_stabs (); | |
1445 | ||
1446 | while (!feof (inf)) | |
1447 | { | |
1448 | c = *lp++; | |
c6d46f5f JB |
1449 | if (c == '\\') |
1450 | { | |
4746118a JB |
1451 | /* If we're at the end of the line, the next character is a |
1452 | '\0'; don't skip it, because it's the thing that tells us | |
1453 | to read the next line. */ | |
1e134a5f RM |
1454 | if (*lp == 0) |
1455 | continue; | |
1456 | lp++; | |
c6d46f5f JB |
1457 | c = ' '; |
1458 | } | |
1459 | else if (incomm) | |
1460 | { | |
1e134a5f | 1461 | if (c == '*' && *lp == '/') |
c6d46f5f | 1462 | { |
1e134a5f RM |
1463 | c = *lp++; |
1464 | incomm = FALSE; | |
c6d46f5f JB |
1465 | } |
1466 | } | |
1467 | else if (inquote) | |
1468 | { | |
c6d46f5f JB |
1469 | if (c == '"') |
1470 | inquote = FALSE; | |
daa37602 JB |
1471 | else if (c == '\\') |
1472 | c = *lp++; | |
c6d46f5f JB |
1473 | } |
1474 | else if (inchar) | |
1475 | { | |
1476 | if (c == '\'') | |
1477 | inchar = FALSE; | |
1478 | continue; | |
1479 | } | |
1480 | else | |
1481 | switch (c) | |
1482 | { | |
1483 | case '"': | |
1484 | inquote = TRUE; | |
1485 | continue; | |
1486 | case '\'': | |
1487 | inchar = TRUE; | |
1488 | continue; | |
1489 | case '/': | |
1490 | if (*lp == '*') | |
1491 | { | |
1492 | lp++; | |
1493 | incomm = TRUE; | |
1494 | } | |
1495 | else if (c_ext && *lp == '/') | |
1496 | { | |
daa37602 JB |
1497 | c = 0; |
1498 | break; | |
c6d46f5f JB |
1499 | } |
1500 | continue; | |
1501 | case '#': | |
1502 | if (lp == lb.buffer + 1 && definedef == dnone) | |
1503 | definedef = dsharpseen; | |
1504 | continue; | |
1505 | ||
1506 | /* | |
1507 | * The next two are to help the strucdef state machine. | |
1508 | * They break when they are finished, so they don't interfere | |
1509 | * with anything else that is going on. | |
1510 | */ | |
1511 | case ':': | |
1512 | if (structdef == stagseen) | |
1513 | structdef = scolonseen; | |
1514 | break; | |
1515 | /* Not a struct definition when semicolon seen in non-sinbody context. */ | |
1516 | case ';': | |
1517 | if (structdef != snone && structdef != sinbody) | |
1518 | { | |
1519 | structdef = snone; | |
1520 | (void) strcpy (structtag, "<error 1>"); | |
1521 | } | |
1522 | break; | |
1523 | ||
1524 | case '{': | |
1525 | if (tydef == begin) | |
1526 | { | |
1527 | tydef = middle; | |
1528 | } | |
1529 | switch (structdef) | |
1530 | { | |
1531 | case skeyseen: /* unnamed struct */ | |
1532 | structtag[0] = '\0'; | |
1533 | /* FALLTHRU */ | |
1534 | case stagseen: | |
1535 | case scolonseen: /* named struct */ | |
1536 | structdef = sinbody; | |
1537 | break; | |
1538 | } | |
1539 | level++; | |
1540 | continue; | |
1541 | case '}': | |
1542 | if (!noindentypedefs && lp == lb.buffer + 1) | |
1543 | level = 0; /* reset level if first column */ | |
1544 | else if (level > 0) | |
1545 | level--; | |
1546 | if (level == 0 && tydef == middle) | |
1547 | { | |
1548 | tydef = end; | |
1549 | } | |
1550 | if (level == 0) | |
1551 | { | |
1552 | structdef = snone; | |
1553 | (void) strcpy (structtag, "<error 2>"); | |
1554 | } | |
1555 | continue; | |
1556 | } | |
1557 | if (LEVEL_OK_FOR_FUNCDEF () && !inquote && !incomm && gotone == FALSE) | |
1558 | { | |
1559 | if (midtoken) | |
1560 | { | |
1561 | if (endtoken (c)) | |
1562 | { | |
1563 | if (c_ext && c == ':' && *lp == ':' && intoken (*(lp + 1))) | |
1564 | { | |
1565 | /* | |
1566 | * This handles :: in the middle, but not at beginning | |
1567 | * of an identifier. | |
1568 | */ | |
1569 | lp += 2; | |
1570 | toklen += 3; | |
1571 | } | |
1572 | else | |
1573 | { | |
1e134a5f RM |
1574 | /* The following is no longer true, |
1575 | now that we advance to the next line | |
1576 | at the end of processing the character. */ | |
c6d46f5f JB |
1577 | /* |
1578 | * We've just finished lexing an identifier. | |
1579 | * Note that if `c' is '\0', `lb' is the NEXT | |
1580 | * line, `lp' points to the beginning of it, and | |
1581 | * old pointers into `lb.buffer' may no longer be | |
1582 | * valid, since `lb.buffer' may have been | |
1583 | * reallocated. In this case (which corresponds | |
1584 | * to an identifier followed immediately by a | |
1585 | * newline), we re-read the line into lb1. | |
1586 | * | |
1587 | * This would be faster if the previous line's | |
1588 | * buffer were always saved. | |
1589 | */ | |
1590 | logical is_func; | |
1591 | char *tok_linebuf; | |
1592 | TOKEN tok; | |
1593 | logical bingo, tok_at_end_of_line; | |
1594 | char *lp_tmp; /* addressable */ | |
1595 | ||
1e134a5f | 1596 | #if 0 |
c6d46f5f JB |
1597 | if (c == '\0') |
1598 | { | |
1599 | getline (GET_COOKIE (prev_linepos)); | |
1600 | tok_linebuf = lb1.buffer; | |
1601 | tok_at_end_of_line = TRUE; | |
1602 | tok.linestart = prev_linepos; | |
1603 | tok.lineno = lineno - 1; | |
1604 | } | |
1605 | else | |
1e134a5f | 1606 | #endif |
c6d46f5f JB |
1607 | { |
1608 | tok_linebuf = lb.buffer; | |
1609 | tok_at_end_of_line = FALSE; | |
1610 | tok.linestart = linepos; | |
1611 | tok.lineno = lineno; | |
1612 | } | |
1613 | tok.p = tok_linebuf + tokoff; | |
1614 | tok.len = toklen; | |
1615 | tok.rewritten = FALSE; | |
1616 | lp_tmp = lp; | |
1617 | bingo = consider_token (c, &lp_tmp, &tok, | |
1618 | &is_func, c_ext, level); | |
1619 | lp = lp_tmp; | |
1620 | if (bingo) | |
1621 | { | |
1622 | if (GET_CHARNO (tok.linestart) != GET_CHARNO (linepos) | |
1623 | && !tok_at_end_of_line) | |
1624 | { | |
1625 | /* | |
1626 | * Resynchronize tok.p to point into the right | |
1627 | * linebuffer. | |
1628 | */ | |
1629 | getline (GET_COOKIE (tok.linestart)); | |
1630 | if (!tok.rewritten) | |
1631 | tok.p = lb1.buffer + (tok.p - tok_linebuf); | |
1632 | tok_linebuf = lb1.buffer; | |
1633 | } | |
1634 | if (structdef == sinbody && definedef == dnone && is_func) | |
1635 | { /* function defined in C++ class body */ | |
1636 | sprintf (tokb, "%s::%.*s", | |
1637 | structtag[0] == '\0' ? "_anonymous_" | |
1638 | : structtag, | |
1639 | tok.len, tok.p); | |
1640 | tok.rewritten = TRUE; | |
1641 | } | |
1642 | else | |
1643 | { | |
1644 | sprintf (tokb, "%.*s", tok.len, tok.p); | |
1645 | } | |
1646 | pfnote (tokb, is_func, tok.rewritten, tok_linebuf, | |
1647 | tokoff + toklen + (tok_at_end_of_line ? 0 : 1), | |
1648 | tok.lineno, GET_CHARNO (tok.linestart)); | |
1649 | gotone = is_func; /* function */ | |
1650 | } | |
1651 | midtoken = FALSE; | |
1652 | } | |
1653 | } | |
1654 | else if (intoken (c)) | |
1655 | toklen++; | |
1656 | } | |
1657 | else if (begtoken (c)) | |
1658 | { | |
1659 | tokoff = lp - 1 - lb.buffer; | |
1660 | toklen = 1; | |
1661 | midtoken = TRUE; | |
1662 | } | |
1663 | } | |
1e134a5f RM |
1664 | /* Detect end of line, after having handled the last token on the line. */ |
1665 | if (c == 0) | |
1666 | { | |
1667 | CNL; | |
1668 | gotone = FALSE; | |
1669 | } | |
c6d46f5f JB |
1670 | if (c == ';' && tydef == end) /* clean with typedefs */ |
1671 | tydef = none; | |
1672 | } | |
1673 | } | |
1674 | ||
1675 | /* | |
1676 | * consider_token () | |
1677 | * checks to see if the current token is at the start of a | |
1678 | * function, or corresponds to a typedef. It updates the input | |
1679 | * line pointer *LPP so that the '(' will be in it when it returns. | |
1680 | * | |
1681 | * *IS_FUNC gets TRUE iff the token is a function. | |
1682 | * C_EXT is which language we are looking at. | |
1683 | * | |
1684 | * In the future we will need some way to adjust where the end of | |
1685 | * the token is; for instance, implementing the C++ keyword | |
1686 | * `operator' properly will adjust the end of the token to be after | |
1687 | * whatever follows `operator'. | |
1688 | * | |
1689 | * Globals | |
1690 | * structdef IN OUT | |
1691 | * definedef IN OUT | |
1692 | * tydef IN OUT | |
1693 | */ | |
1694 | ||
1695 | logical | |
1696 | consider_token (c, lpp, tokp, is_func, c_ext, level) | |
1697 | reg char c; /* IN: first char after the token */ | |
1698 | char **lpp; /* IN OUT: *lpp points to 2nd char after the token */ | |
1699 | reg TOKEN *tokp; /* IN */ | |
1700 | logical *is_func; /* OUT */ | |
1701 | int c_ext; /* IN */ | |
1702 | int level; /* IN */ | |
1703 | { | |
1704 | reg char *lp = *lpp; | |
c6d46f5f JB |
1705 | logical firsttok; /* TRUE if have seen first token in ()'s */ |
1706 | Stab_entry *tokse = stab_find (get_C_stab (c_ext), tokp->p, tokp->len); | |
1707 | enum sym_type toktype = stab_type (tokse); | |
1708 | ||
1709 | *is_func = TRUE; /* a function */ | |
1710 | ||
1711 | /* | |
1712 | * Advance the definedef state machine. We set `gotone' for good measure; | |
1713 | * it's redundant. | |
1714 | */ | |
1715 | switch (definedef) | |
1716 | { | |
1717 | case dnone: | |
1718 | /* We're not on a preprocessor line. */ | |
1719 | break; | |
1720 | case dsharpseen: | |
1721 | if (toktype == st_C_define) | |
1722 | { | |
1723 | definedef = ddefineseen; | |
1724 | gotone = FALSE; | |
1725 | } | |
1726 | else | |
1727 | { | |
1728 | definedef = dignorerest; | |
1729 | gotone = TRUE; | |
1730 | } | |
1731 | goto badone; | |
1732 | case ddefineseen: | |
1733 | /* | |
1734 | * Make a tag for any macro. | |
1735 | * This will flub up if there is a newline immediately following | |
1736 | * the macro name. | |
1737 | */ | |
1738 | *is_func = (c == '('); | |
1739 | definedef = dignorerest; | |
1740 | gotone = TRUE; | |
1741 | if (!*is_func && !constantypedefs) | |
1742 | goto badone; | |
1743 | goto goodone; | |
1744 | case dignorerest: | |
1745 | goto badone; | |
1746 | default: | |
1747 | error ("internal error: definedef value"); | |
1748 | } | |
1749 | ||
1750 | /* | |
1751 | * Skip whitespace and comments after the token. This loop should | |
1752 | * also skip C++ comments. | |
1753 | */ | |
1754 | while (1) | |
1755 | { | |
1756 | /* At whitespace => skip it. */ | |
1757 | if (iswhite (c)) | |
1758 | { | |
1759 | c = *lp++; | |
1760 | } | |
1761 | /* At a comment => skip to end of comment. */ | |
1762 | else if (c == '/' && *lp == '*') | |
1763 | { | |
1764 | /* If we find a comment, skip it. */ | |
1765 | while (!(c == '*' && *lp == '/')) | |
1766 | { | |
1767 | c = *lp++; | |
1768 | if (c == 0) | |
1769 | { | |
1e134a5f RM |
1770 | lp--; |
1771 | break; | |
c6d46f5f JB |
1772 | } |
1773 | } | |
1774 | if (c == '*' && *lp == '/') | |
1775 | { | |
1776 | lp++; /* lp now points past the '/' */ | |
1777 | c = *lp++; /* c is now the --whatever-- after the '/' */ | |
1778 | } | |
1779 | } | |
1780 | else | |
1781 | break; | |
1782 | ||
1783 | /* If we arrived at eof or eol, decide which one it is. | |
1784 | If it's eol, advance to the next line. */ | |
1785 | ||
1786 | if (c == 0) | |
1787 | { | |
1e134a5f RM |
1788 | lp--; |
1789 | break; | |
c6d46f5f JB |
1790 | } |
1791 | } | |
1792 | ||
1793 | /* | |
1794 | * If you have custom token types, or when configuration files can | |
1795 | * define custom token types, this switch will be larger. | |
1796 | */ | |
1797 | switch (toktype) | |
1798 | { | |
1799 | case st_C_typedef: | |
1800 | if (typedefs) | |
1801 | { | |
1802 | tydef = begin; | |
1803 | goto badone; | |
1804 | } | |
1805 | break; | |
1806 | case st_C_typespec: | |
1807 | if (tydef == begin || tydef == end) | |
1808 | { | |
1809 | tydef = end; | |
1810 | goto badone; | |
1811 | } | |
1812 | break; | |
1813 | } | |
1814 | ||
1815 | /* | |
1816 | * This structdef business is currently only invoked when level==0. | |
1817 | * It should be recursively invoked whatever the level, and a stack of | |
1818 | * states kept, to allow for definitions of structs within structs. | |
1819 | * | |
1820 | * This structdef business is NOT invoked when we are ctags and the | |
1821 | * file is plain C. This is because a struct tag may have the same | |
1822 | * name as another tag, and this loses with ctags. | |
1823 | * | |
1824 | * This if statement deals with the tydef state machine as follows: if | |
1825 | * tydef==begin and token is struct/union/class/enum, goto badone. | |
1826 | * All the other code here is for the structdef state machine. | |
1827 | */ | |
1828 | switch (toktype) | |
1829 | { | |
1830 | case st_C_struct: | |
1831 | case st_C_enum: | |
1832 | if (tydef == begin || (typedefs_and_cplusplus && level == 0 && structdef == snone)) | |
1833 | { | |
1834 | structdef = skeyseen; | |
1835 | structkey = tokse; | |
1836 | } | |
1837 | goto badone; | |
1838 | } | |
1839 | ||
1840 | if (structdef == skeyseen) | |
1841 | { | |
1842 | /* If next char is '{' or (for C++) ':', found a structure tag. */ | |
1843 | if (c == '{' || (c_ext && c == ':')) | |
1844 | { | |
1845 | /* | |
1846 | * We should do this slightly differently for straight C: | |
1847 | * instead of defining `tag', as we now do, we should define | |
1848 | * `struct tag'. (Do this only if the find-tag defaulting is | |
1849 | * done on a sophisticated per-mode basis, so that if the user | |
1850 | * says meta-. anywhere in `struct foo', the default comes out | |
1851 | * `struct foo', not `struct' or `foo'.) This will require | |
1852 | * remembering which keyword (struct/union/class/enum) we saw, as a | |
1853 | * Stab_entry* -- this will also make it possible to merge the | |
1854 | * skeyseen and senumseen states, if we want. | |
1855 | */ | |
1856 | if (stab_type (structkey) == st_C_struct) | |
1857 | { | |
1858 | (void) strncpy (structtag, tokp->p, tokp->len); | |
1859 | structtag[tokp->len] = '\0'; /* for struct/union/class */ | |
1860 | structdef = stagseen; | |
1861 | } | |
1862 | else | |
1863 | { | |
1864 | structtag[0] = '\0'; /* for enum */ | |
1865 | } | |
1866 | *is_func = FALSE; /* not a function */ | |
1867 | goto goodone; | |
1868 | } | |
1869 | else | |
1870 | { | |
1871 | /* Not a definition: reset structdef */ | |
1872 | structdef = snone; | |
1873 | (void) strcpy (structtag, "<error 3>"); | |
1874 | } | |
1875 | /* Now what? And how does/should this stuff interact with tydef?? */ | |
1876 | /* Also maybe reset lp to *lpp for benefit of the function finding code. */ | |
1877 | } | |
1878 | if (tydef == begin) | |
1879 | { | |
1880 | tydef = end; | |
1881 | goto badone; | |
1882 | } | |
1883 | if (tydef == end) | |
1884 | { | |
1885 | *is_func = 0; | |
1886 | goto goodone; | |
1887 | } | |
1888 | /* Detect GNUmacs's function-defining macros. */ | |
4746118a | 1889 | if (definedef == dnone) |
c6d46f5f | 1890 | { |
daa37602 JB |
1891 | if (strneq (tokp->p, "DEF", 3) |
1892 | || strneq (tokp->p, "ENTRY", 5) | |
1893 | || strneq (tokp->p, "SYSCALL", 7) | |
1894 | || strneq (tokp->p, "PSEUDO", 6)) | |
4746118a JB |
1895 | { |
1896 | next_token_is_func = TRUE; | |
1897 | goto badone; | |
1898 | } | |
1899 | else if (strneq (tokp->p, "EXFUN", 5)) | |
1900 | { | |
1901 | next_token_is_func = FALSE; | |
1902 | goto badone; | |
1903 | } | |
c6d46f5f JB |
1904 | } |
1905 | if (next_token_is_func) | |
1906 | { | |
1907 | next_token_is_func = FALSE; | |
1908 | goto goodone; | |
1909 | } | |
1910 | if (c != '(') | |
1911 | goto badone; | |
1912 | firsttok = FALSE; | |
1913 | while ((c = *lp++) != ')') | |
1914 | { | |
1915 | if (c == 0) | |
1916 | { | |
1e134a5f RM |
1917 | lp--; |
1918 | break; | |
c6d46f5f JB |
1919 | } |
1920 | /* | |
1921 | * This line used to confuse ctags: | |
1922 | * int (*oldhup)(); | |
1923 | * This fixes it. A nonwhite char before the first | |
1924 | * token, other than a / (in case of a comment in there) | |
1925 | * makes this not a declaration. | |
1926 | */ | |
1927 | if (begtoken (c) || c == '/') | |
1928 | firsttok++; | |
1929 | else if (!iswhite (c) && !firsttok) | |
1930 | goto badone; | |
1931 | } | |
1932 | while (iswhite (c = *lp++)) | |
1933 | { | |
1934 | if (c == 0) | |
1935 | { | |
1e134a5f RM |
1936 | lp--; |
1937 | break; | |
c6d46f5f JB |
1938 | } |
1939 | } | |
1940 | if (!isgood (c)) | |
1941 | goto badone; | |
1942 | ||
1943 | goodone: | |
1944 | *lpp = lp - 1; | |
1945 | return TRUE; | |
1946 | ||
1947 | badone: | |
1948 | *lpp = lp - 1; | |
1949 | return FALSE; | |
1950 | } | |
1951 | ||
1952 | void | |
1953 | getline (atcookie) | |
1954 | long atcookie; | |
1955 | { | |
1956 | long saveftell = ftell (inf); | |
1957 | ||
1958 | (void) fseek (inf, atcookie, 0); | |
1959 | (void) readline (&lb1, inf); | |
1960 | (void) fseek (inf, saveftell, 0); | |
1961 | } | |
1962 | \f | |
1963 | /* Fortran parsing */ | |
1964 | ||
1965 | char *dbp; | |
1966 | int pfcnt; | |
1967 | ||
1968 | int | |
1969 | PF_funcs (fi) | |
1970 | FILE *fi; | |
1971 | { | |
1972 | lineno = 0; | |
1973 | charno = 0; | |
1974 | pfcnt = 0; | |
1975 | ||
1976 | while (!feof (fi)) | |
1977 | { | |
1978 | lineno++; | |
1979 | linecharno = charno; | |
1980 | charno += readline (&lb, fi); | |
1981 | dbp = lb.buffer; | |
1982 | if (*dbp == '%') | |
1983 | dbp++; /* Ratfor escape to fortran */ | |
1984 | while (isspace (*dbp)) | |
1985 | dbp++; | |
1986 | if (*dbp == 0) | |
1987 | continue; | |
1988 | switch (*dbp | ' ') | |
1989 | { | |
1990 | case 'i': | |
1991 | if (tail ("integer")) | |
1992 | takeprec (); | |
1993 | break; | |
1994 | case 'r': | |
1995 | if (tail ("real")) | |
1996 | takeprec (); | |
1997 | break; | |
1998 | case 'l': | |
1999 | if (tail ("logical")) | |
2000 | takeprec (); | |
2001 | break; | |
2002 | case 'c': | |
2003 | if (tail ("complex") || tail ("character")) | |
2004 | takeprec (); | |
2005 | break; | |
2006 | case 'd': | |
2007 | if (tail ("double")) | |
2008 | { | |
2009 | while (isspace (*dbp)) | |
2010 | dbp++; | |
2011 | if (*dbp == 0) | |
2012 | continue; | |
2013 | if (tail ("precision")) | |
2014 | break; | |
2015 | continue; | |
2016 | } | |
2017 | break; | |
2018 | } | |
2019 | while (isspace (*dbp)) | |
2020 | dbp++; | |
2021 | if (*dbp == 0) | |
2022 | continue; | |
2023 | switch (*dbp | ' ') | |
2024 | { | |
2025 | case 'f': | |
2026 | if (tail ("function")) | |
2027 | getit (); | |
2028 | continue; | |
2029 | case 's': | |
2030 | if (tail ("subroutine")) | |
2031 | getit (); | |
2032 | continue; | |
2033 | case 'p': | |
2034 | if (tail ("program")) | |
2035 | { | |
2036 | getit (); | |
2037 | continue; | |
2038 | } | |
2039 | if (tail ("procedure")) | |
2040 | getit (); | |
2041 | continue; | |
2042 | } | |
2043 | } | |
2044 | return (pfcnt); | |
2045 | } | |
2046 | ||
2047 | logical | |
2048 | tail (cp) | |
2049 | char *cp; | |
2050 | { | |
2051 | register int len = 0; | |
2052 | ||
2053 | while (*cp && (*cp & ~' ') == ((*(dbp + len)) & ~' ')) | |
2054 | cp++, len++; | |
2055 | if (*cp == 0) | |
2056 | { | |
2057 | dbp += len; | |
2058 | return (1); | |
2059 | } | |
2060 | return (0); | |
2061 | } | |
2062 | ||
2063 | void | |
2064 | takeprec () | |
2065 | { | |
2066 | while (isspace (*dbp)) | |
2067 | dbp++; | |
2068 | if (*dbp != '*') | |
2069 | return; | |
2070 | dbp++; | |
2071 | while (isspace (*dbp)) | |
2072 | dbp++; | |
2073 | if (!isdigit (*dbp)) | |
2074 | { | |
2075 | --dbp; /* force failure */ | |
2076 | return; | |
2077 | } | |
2078 | do | |
2079 | dbp++; | |
2080 | while (isdigit (*dbp)); | |
2081 | } | |
2082 | ||
2083 | void | |
2084 | getit () | |
2085 | { | |
2086 | register char *cp; | |
2087 | char c; | |
2088 | char nambuf[BUFSIZ]; | |
2089 | ||
2090 | while (isspace (*dbp)) | |
2091 | dbp++; | |
daa37602 JB |
2092 | if (*dbp == 0 |
2093 | || (!isalpha (*dbp) | |
2094 | && *dbp != '_' | |
2095 | && *dbp != '$')) | |
c6d46f5f JB |
2096 | return; |
2097 | for (cp = dbp + 1; *cp && (isalpha (*cp) || isdigit (*cp) | |
2098 | || (*cp == '_') || (*cp == '$')); cp++) | |
2099 | continue; | |
2100 | c = cp[0]; | |
2101 | cp[0] = 0; | |
2102 | (void) strcpy (nambuf, dbp); | |
2103 | cp[0] = c; | |
2104 | pfnote (nambuf, TRUE, FALSE, lb.buffer, cp - lb.buffer + 1, lineno, linecharno); | |
2105 | pfcnt++; | |
2106 | } | |
2107 | ||
2108 | /* Handle a file of assembler code. */ | |
2109 | ||
2110 | void | |
2111 | Asm_funcs (fi) | |
2112 | FILE *fi; | |
2113 | { | |
2114 | int i; | |
2115 | register char c; | |
2116 | ||
2117 | lineno = 0; | |
2118 | charno = 0; | |
2119 | pfcnt = 0; | |
2120 | ||
2121 | while (!feof (fi)) | |
2122 | { | |
2123 | lineno++; | |
2124 | linecharno = charno; | |
2125 | charno += readline (&lb, fi); | |
2126 | dbp = lb.buffer; | |
2127 | ||
2128 | for (i = 0; ((c = dbp[i]) && !isspace (c)) && (c != ':'); i++) | |
2129 | ; | |
2130 | ||
2131 | if ((i > 0) && (c == ':')) | |
2132 | getit (); | |
2133 | } | |
2134 | } | |
2135 | \f | |
2136 | /* Added by Mosur Mohan, 4/22/88 */ | |
2137 | /* Pascal parsing */ | |
2138 | ||
2139 | #define GET_NEW_LINE \ | |
2140 | { \ | |
2141 | linecharno = charno; lineno++; \ | |
2142 | charno += 1 + readline (&lb, inf); \ | |
2143 | dbp = lb.buffer; \ | |
2144 | } | |
2145 | ||
2146 | /* Locates tags for procedures & functions. | |
2147 | * Doesn't do any type- or var-definitions. | |
2148 | * It does look for the keyword "extern" or "forward" | |
2149 | * immediately following the procedure statement; | |
2150 | * if found, the tag is skipped. | |
2151 | */ | |
2152 | ||
2153 | void | |
2154 | PAS_funcs (fi) | |
2155 | FILE *fi; | |
2156 | { | |
2157 | struct linebuffer tline; /* mostly copied from C_entries */ | |
2158 | long save_lcno; | |
2159 | int save_lineno; | |
2160 | char c, *cp; | |
2161 | char nambuf[BUFSIZ]; | |
2162 | ||
2163 | logical /* each of these flags is TRUE iff: */ | |
2164 | incomm1, /* point is inside {..} comment */ | |
2165 | incomm2, /* point is inside (*..*) comment */ | |
2166 | inquote, /* point is inside '..' string */ | |
2167 | get_tagname, /* point is after PROCEDURE/FUNCTION */ | |
2168 | /* keyword, so next item = potential tag */ | |
2169 | found_tag, /* point is after a potential tag */ | |
2170 | inparms, /* point is within parameter-list */ | |
2171 | verify_tag; /* point has passed the parm-list, so the */ | |
2172 | /* next token will determine whether */ | |
2173 | /* this is a FORWARD/EXTERN to be */ | |
2174 | /* ignored, or whether it is a real tag */ | |
2175 | ||
2176 | lineno = 0; | |
2177 | charno = 0; | |
2178 | dbp = lb.buffer; | |
2179 | *dbp = 0; | |
2180 | initbuffer (&tline); | |
2181 | ||
2182 | incomm1 = incomm2 = inquote = FALSE; | |
2183 | found_tag = FALSE; /* have a proc name; check if extern */ | |
2184 | get_tagname = FALSE; /* have found "procedure" keyword */ | |
2185 | inparms = FALSE; /* found '(' after "proc" */ | |
2186 | verify_tag = FALSE; /* check if "extern" is ahead */ | |
2187 | ||
2188 | /* long main loop to get next char */ | |
2189 | while (!feof (fi)) | |
2190 | { | |
2191 | c = *dbp++; | |
2192 | if (c == 0) /* if end of line */ | |
2193 | { | |
2194 | GET_NEW_LINE; | |
2195 | if (*dbp == 0) | |
2196 | continue; | |
2197 | if (!((found_tag && verify_tag) || | |
2198 | get_tagname)) | |
2199 | c = *dbp++; /* only if don't need *dbp pointing */ | |
2200 | /* to the beginning of the name of */ | |
2201 | /* the procedure or function */ | |
2202 | } | |
2203 | if (incomm1) /* within { - } comments */ | |
2204 | { | |
2205 | if (c == '}') | |
2206 | incomm1 = FALSE; | |
2207 | continue; | |
2208 | } | |
2209 | else if (incomm2) /* within (* - *) comments */ | |
2210 | { | |
2211 | if (c == '*') | |
2212 | { | |
2213 | while ((c = *dbp++) == '*') | |
2214 | continue; | |
2215 | if (c == 0) | |
2216 | GET_NEW_LINE; | |
2217 | if (c == ')') | |
2218 | incomm2 = FALSE; | |
2219 | } | |
2220 | continue; | |
2221 | } | |
2222 | else if (inquote) | |
2223 | { | |
2224 | if (c == '\'') | |
2225 | inquote = FALSE; | |
2226 | continue; | |
2227 | } | |
2228 | else | |
2229 | switch (c) | |
2230 | { | |
2231 | case '\'': | |
2232 | inquote = TRUE; /* found first quote */ | |
2233 | continue; | |
2234 | case '{': /* found open-{-comment */ | |
2235 | incomm1 = TRUE; | |
2236 | continue; | |
2237 | case '(': | |
2238 | if (*dbp == '*') /* found open-(*-comment */ | |
2239 | { | |
2240 | incomm2 = TRUE; | |
2241 | dbp++; | |
2242 | } | |
2243 | else if (found_tag) /* found '(' after tag, i.e., parm-list */ | |
2244 | inparms = TRUE; | |
2245 | continue; | |
2246 | case ')': /* end of parms list */ | |
2247 | if (inparms) | |
2248 | inparms = FALSE; | |
2249 | continue; | |
2250 | case ';': | |
2251 | if ((found_tag) && (!inparms)) /* end of proc or fn stmt */ | |
2252 | { | |
2253 | verify_tag = TRUE; | |
2254 | break; | |
2255 | } | |
2256 | continue; | |
2257 | } | |
2258 | if ((found_tag) && (verify_tag) && (*dbp != ' ')) | |
2259 | { | |
2260 | /* check if this is an "extern" declaration */ | |
2261 | if (*dbp == 0) | |
2262 | continue; | |
2263 | if ((*dbp == 'e') || (*dbp == 'E')) | |
2264 | { | |
2265 | if (tail ("extern")) /* superfluous, really! */ | |
2266 | { | |
2267 | found_tag = FALSE; | |
2268 | verify_tag = FALSE; | |
2269 | } | |
2270 | } | |
2271 | else if ((*dbp == 'f') || (*dbp == 'F')) | |
2272 | { | |
2273 | if (tail ("forward")) /* check for forward reference */ | |
2274 | { | |
2275 | found_tag = FALSE; | |
2276 | verify_tag = FALSE; | |
2277 | } | |
2278 | } | |
2279 | if ((found_tag) && (verify_tag)) /* not external proc, so make tag */ | |
2280 | { | |
2281 | found_tag = FALSE; | |
2282 | verify_tag = FALSE; | |
2283 | pfnote (nambuf, TRUE, FALSE, | |
2284 | tline.buffer, cp - tline.buffer + 1, | |
2285 | save_lineno, save_lcno); | |
2286 | continue; | |
2287 | } | |
2288 | } | |
2289 | if (get_tagname) /* grab name of proc or fn */ | |
2290 | { | |
2291 | if (*dbp == 0) | |
2292 | continue; | |
2293 | ||
2294 | /* save all values for later tagging */ | |
2295 | tline.size = lb.size; | |
2296 | strcpy (tline.buffer, lb.buffer); | |
2297 | save_lineno = lineno; | |
2298 | save_lcno = linecharno; | |
2299 | ||
2300 | /* grab block name */ | |
2301 | for (cp = dbp + 1; *cp && (!endtoken (*cp)); cp++) | |
2302 | continue; | |
2303 | c = cp[0]; | |
2304 | cp[0] = 0; | |
2305 | strcpy (nambuf, dbp); | |
2306 | cp[0] = c; | |
2307 | dbp = cp; /* restore dbp to e-o-token */ | |
2308 | get_tagname = FALSE; | |
2309 | found_tag = TRUE; | |
2310 | continue; | |
2311 | ||
2312 | /* and proceed to check for "extern" */ | |
2313 | } | |
2314 | if ((!incomm1) && (!incomm2) && (!inquote) && | |
2315 | (!found_tag) && (!get_tagname)) | |
2316 | { | |
2317 | /* check for proc/fn keywords */ | |
2318 | switch (c | ' ') | |
2319 | { | |
2320 | case 'p': | |
2321 | if (tail ("rocedure")) /* c = 'p', dbp has advanced */ | |
2322 | get_tagname = TRUE; | |
2323 | continue; | |
2324 | case 'f': | |
2325 | if (tail ("unction")) | |
2326 | get_tagname = TRUE; | |
2327 | continue; | |
2328 | } | |
2329 | } | |
2330 | } /* while not e-o-f */ | |
2331 | } | |
2332 | \f | |
2333 | /* | |
2334 | * lisp tag functions | |
2335 | * just look for (def or (DEF | |
2336 | */ | |
2337 | ||
2338 | void | |
2339 | L_funcs (fi) | |
2340 | FILE *fi; | |
2341 | { | |
2342 | lineno = 0; | |
2343 | charno = 0; | |
2344 | pfcnt = 0; | |
2345 | ||
2346 | while (!feof (fi)) | |
2347 | { | |
2348 | lineno++; | |
2349 | linecharno = charno; | |
2350 | charno += readline (&lb, fi); | |
2351 | dbp = lb.buffer; | |
2352 | if (dbp[0] == '(') | |
2353 | { | |
2354 | if (L_isdef (dbp)) | |
2355 | { | |
2356 | while (!isspace (*dbp)) | |
2357 | dbp++; | |
2358 | while (isspace (*dbp)) | |
2359 | dbp++; | |
2360 | L_getit (); | |
2361 | } | |
2362 | else | |
2363 | { | |
2364 | /* Check for (foo::defmumble name-defined ... */ | |
2365 | while (*dbp && *dbp != ':' && !isspace (*dbp) | |
2366 | && *dbp != '(' && *dbp != ')') | |
2367 | dbp++; | |
2368 | if (*dbp == ':') | |
2369 | { | |
2370 | while (*dbp == ':') | |
2371 | dbp++; | |
2372 | ||
2373 | if (L_isdef (dbp)) | |
2374 | { | |
2375 | while (!isspace (*dbp)) | |
2376 | dbp++; | |
2377 | while (isspace (*dbp)) | |
2378 | dbp++; | |
2379 | L_getit (); | |
2380 | } | |
2381 | } | |
2382 | } | |
2383 | } | |
2384 | } | |
2385 | } | |
2386 | ||
2387 | int | |
2388 | L_isdef (dbp) | |
2389 | char *dbp; | |
2390 | { | |
2391 | return ((dbp[1] == 'D' || dbp[1] == 'd') && | |
2392 | (dbp[2] == 'E' || dbp[2] == 'e') && | |
2393 | (dbp[3] == 'F' || dbp[3] == 'f')); | |
2394 | } | |
2395 | ||
2396 | void | |
2397 | L_getit () | |
2398 | { | |
2399 | register char *cp; | |
2400 | char c; | |
2401 | char nambuf[BUFSIZ]; | |
2402 | ||
2403 | if (*dbp == 0) | |
2404 | return; | |
2405 | for (cp = dbp + 1; *cp && *cp != '(' && *cp != ' '; cp++) | |
2406 | continue; | |
2407 | c = cp[0]; | |
2408 | cp[0] = 0; | |
2409 | (void) strcpy (nambuf, dbp); | |
2410 | cp[0] = c; | |
2411 | pfnote (nambuf, TRUE, FALSE, lb.buffer, cp - lb.buffer + 1, lineno, linecharno); | |
2412 | pfcnt++; | |
2413 | } | |
2414 | \f | |
2415 | /* | |
2416 | * Scheme tag functions | |
2417 | * look for (def... xyzzy | |
2418 | * look for (def... (xyzzy | |
2419 | * look for (def ... ((...(xyzzy .... | |
2420 | * look for (set! xyzzy | |
2421 | */ | |
2422 | ||
2423 | static void get_scheme (); | |
2424 | ||
2425 | void | |
2426 | Scheme_funcs (fi) | |
2427 | FILE *fi; | |
2428 | { | |
2429 | lineno = 0; | |
2430 | charno = 0; | |
2431 | pfcnt = 0; | |
2432 | ||
2433 | while (!feof (fi)) | |
2434 | { | |
2435 | lineno++; | |
2436 | linecharno = charno; | |
2437 | charno += readline (&lb, fi); | |
2438 | dbp = lb.buffer; | |
2439 | if (dbp[0] == '(' && | |
2440 | (dbp[1] == 'D' || dbp[1] == 'd') && | |
2441 | (dbp[2] == 'E' || dbp[2] == 'e') && | |
2442 | (dbp[3] == 'F' || dbp[3] == 'f')) | |
2443 | { | |
2444 | while (!isspace (*dbp)) | |
2445 | dbp++; | |
2446 | /* Skip over open parens and white space */ | |
2447 | while (*dbp && (isspace (*dbp) || *dbp == '(')) | |
2448 | dbp++; | |
2449 | get_scheme (); | |
2450 | } | |
2451 | if (dbp[0] == '(' && | |
2452 | (dbp[1] == 'S' || dbp[1] == 's') && | |
2453 | (dbp[2] == 'E' || dbp[2] == 'e') && | |
2454 | (dbp[3] == 'T' || dbp[3] == 't') && | |
2455 | (dbp[4] == '!' || dbp[4] == '!') && | |
2456 | (isspace (dbp[5]))) | |
2457 | { | |
2458 | while (!isspace (*dbp)) | |
2459 | dbp++; | |
2460 | /* Skip over white space */ | |
2461 | while (isspace (*dbp)) | |
2462 | dbp++; | |
2463 | get_scheme (); | |
2464 | } | |
2465 | } | |
2466 | } | |
2467 | ||
2468 | static void | |
2469 | get_scheme () | |
2470 | { | |
2471 | register char *cp; | |
2472 | char c; | |
2473 | char nambuf[BUFSIZ]; | |
2474 | ||
2475 | if (*dbp == 0) | |
2476 | return; | |
2477 | /* Go till you get to white space or a syntactic break */ | |
2478 | for (cp = dbp + 1; *cp && *cp != '(' && *cp != ')' && !isspace (*cp); cp++) | |
2479 | continue; | |
2480 | /* Null terminate the string there. */ | |
2481 | c = cp[0]; | |
2482 | cp[0] = 0; | |
2483 | /* Copy the string */ | |
2484 | strcpy (nambuf, dbp); | |
2485 | /* Unterminate the string */ | |
2486 | cp[0] = c; | |
2487 | /* Announce the change */ | |
2488 | pfnote (nambuf, TRUE, FALSE, lb.buffer, cp - lb.buffer + 1, lineno, linecharno); | |
2489 | pfcnt++; | |
2490 | } | |
2491 | \f | |
2492 | /* Find tags in TeX and LaTeX input files. */ | |
2493 | ||
2494 | /* TEX_toktab is a table of TeX control sequences that define tags. | |
2495 | Each TEX_tabent records one such control sequence. | |
2496 | CONVERT THIS TO USE THE Stab TYPE!! */ | |
2497 | ||
2498 | struct TEX_tabent | |
2499 | { | |
2500 | char *name; | |
2501 | int len; | |
2502 | }; | |
2503 | ||
2504 | struct TEX_tabent *TEX_toktab = NULL; /* Table with tag tokens */ | |
2505 | ||
2506 | /* Default set of control sequences to put into TEX_toktab. | |
2507 | The value of environment var TEXTAGS is prepended to this. */ | |
2508 | ||
2509 | static char *TEX_defenv = | |
2510 | ":chapter:section:subsection:subsubsection:eqno:label:ref:cite:bibitem:typeout"; | |
2511 | ||
2512 | void TEX_mode (); | |
2513 | struct TEX_tabent *TEX_decode_env (); | |
2514 | void TEX_getit (); | |
2515 | int TEX_Token (); | |
2516 | ||
2517 | static char TEX_esc = '\\'; | |
2518 | static char TEX_opgrp = '{'; | |
2519 | static char TEX_clgrp = '}'; | |
2520 | ||
2521 | /* | |
2522 | * TeX/LaTeX scanning loop. | |
2523 | */ | |
2524 | ||
2525 | void | |
2526 | TEX_funcs (fi) | |
2527 | FILE *fi; | |
2528 | { | |
2529 | char *lasthit; | |
2530 | ||
2531 | lineno = 0; | |
2532 | charno = 0; | |
2533 | pfcnt = 0; | |
2534 | ||
2535 | /* Select either \ or ! as escape character. */ | |
2536 | TEX_mode (fi); | |
2537 | ||
2538 | /* Initialize token table once from environment. */ | |
2539 | if (!TEX_toktab) | |
2540 | TEX_toktab = TEX_decode_env ("TEXTAGS", TEX_defenv); | |
2541 | ||
2542 | while (!feof (fi)) | |
2543 | { | |
2544 | lineno++; | |
2545 | linecharno = charno; | |
2546 | charno += readline (&lb, fi); | |
2547 | dbp = lb.buffer; | |
2548 | lasthit = dbp; | |
2549 | ||
2550 | while (!feof (fi)) | |
2551 | { /* Scan each line in file */ | |
2552 | lineno++; | |
2553 | linecharno = charno; | |
2554 | charno += readline (&lb, fi); | |
2555 | dbp = lb.buffer; | |
2556 | lasthit = dbp; | |
2557 | while (dbp = index (dbp, TEX_esc)) /* Look at each escape in line */ | |
2558 | { | |
2559 | register int i; | |
2560 | ||
2561 | if (!*(++dbp)) | |
2562 | break; | |
2563 | linecharno += dbp - lasthit; | |
2564 | lasthit = dbp; | |
2565 | i = TEX_Token (lasthit); | |
2566 | if (0 <= i) | |
2567 | { | |
2568 | TEX_getit (lasthit, TEX_toktab[i].len); | |
2569 | break; /* We only save a line once */ | |
2570 | } | |
2571 | } | |
2572 | } | |
2573 | } | |
2574 | } | |
2575 | ||
2576 | #define TEX_LESC '\\' | |
2577 | #define TEX_SESC '!' | |
2578 | #define TEX_cmt '%' | |
2579 | ||
2580 | /* Figure out whether TeX's escapechar is '\\' or '!' and set grouping */ | |
2581 | /* chars accordingly. */ | |
2582 | ||
2583 | void | |
2584 | TEX_mode (f) | |
2585 | FILE *f; | |
2586 | { | |
2587 | int c; | |
2588 | ||
2589 | while ((c = getc (f)) != EOF) | |
2590 | { | |
2591 | /* Skip to next line if we hit the TeX comment char. */ | |
2592 | if (c == TEX_cmt) | |
2593 | while (c != '\n') | |
2594 | c = getc (f); | |
2595 | else if (c == TEX_LESC || c == TEX_SESC ) | |
2596 | break; | |
2597 | } | |
2598 | ||
2599 | if (c == TEX_LESC) | |
2600 | { | |
2601 | TEX_esc = TEX_LESC; | |
2602 | TEX_opgrp = '{'; | |
2603 | TEX_clgrp = '}'; | |
2604 | } | |
2605 | else | |
2606 | { | |
2607 | TEX_esc = TEX_SESC; | |
2608 | TEX_opgrp = '<'; | |
2609 | TEX_clgrp = '>'; | |
2610 | } | |
2611 | rewind (f); | |
2612 | } | |
2613 | ||
2614 | /* Read environment and prepend it to the default string. */ | |
2615 | /* Build token table. */ | |
2616 | ||
2617 | struct TEX_tabent * | |
2618 | TEX_decode_env (evarname, defenv) | |
2619 | char *evarname; | |
2620 | char *defenv; | |
2621 | { | |
2622 | register char *env, *p; | |
2623 | extern char *savenstr (), *index (); | |
2624 | ||
2625 | struct TEX_tabent *tab; | |
2626 | int size, i; | |
2627 | ||
2628 | /* Append default string to environment. */ | |
2629 | env = getenv (evarname); | |
2630 | if (!env) | |
2631 | env = defenv; | |
2632 | else | |
2633 | env = concat (env, defenv, ""); | |
2634 | ||
2635 | /* Allocate a token table */ | |
2636 | for (size = 1, p = env; p;) | |
2637 | if ((p = index (p, ':')) && *(++p)) | |
2638 | size++; | |
2639 | tab = xnew (size, struct TEX_tabent); | |
2640 | ||
2641 | /* Unpack environment string into token table. Be careful about */ | |
2642 | /* zero-length strings (leading ':', "::" and trailing ':') */ | |
2643 | for (i = 0; *env;) | |
2644 | { | |
2645 | p = index (env, ':'); | |
2646 | if (!p) /* End of environment string. */ | |
2647 | p = env + strlen (env); | |
2648 | if (p - env > 0) | |
2649 | { /* Only non-zero strings. */ | |
2650 | tab[i].name = savenstr (env, p - env); | |
2651 | tab[i].len = strlen (tab[i].name); | |
2652 | i++; | |
2653 | } | |
2654 | if (*p) | |
2655 | env = p + 1; | |
2656 | else | |
2657 | { | |
2658 | tab[i].name = NULL; /* Mark end of table. */ | |
2659 | tab[i].len = 0; | |
2660 | break; | |
2661 | } | |
2662 | } | |
2663 | return tab; | |
2664 | } | |
2665 | ||
2666 | /* Record a tag defined by a TeX command of length LEN and starting at NAME. | |
2667 | The name being defined actually starts at (NAME + LEN + 1). | |
2668 | But we seem to include the TeX command in the tag name. */ | |
2669 | ||
2670 | void | |
2671 | TEX_getit (name, len) | |
2672 | char *name; | |
2673 | int len; | |
2674 | { | |
2675 | char *p = name + len; | |
2676 | char nambuf[BUFSIZ]; | |
2677 | ||
2678 | if (*name == 0) | |
2679 | return; | |
2680 | ||
2681 | /* Let tag name extend to next group close (or end of line) */ | |
2682 | while (*p && *p != TEX_clgrp) | |
2683 | p++; | |
2684 | (void) strncpy (nambuf, name, p - name); | |
2685 | nambuf[p - name] = 0; | |
2686 | ||
2687 | pfnote (nambuf, TRUE, FALSE, lb.buffer, strlen (lb.buffer), lineno, linecharno); | |
2688 | pfcnt++; | |
2689 | } | |
2690 | ||
2691 | /* If the text at CP matches one of the tag-defining TeX command names, | |
2692 | return the index of that command in TEX_toktab. | |
2693 | Otherwise return -1. */ | |
2694 | ||
2695 | /* Keep the capital `T' in `Token' for dumb truncating compilers | |
2696 | (this distinguishes it from `TEX_toktab' */ | |
2697 | int | |
2698 | TEX_Token (cp) | |
2699 | char *cp; | |
2700 | { | |
2701 | int i; | |
2702 | ||
2703 | for (i = 0; TEX_toktab[i].len > 0; i++) | |
2704 | if (strncmp (TEX_toktab[i].name, cp, TEX_toktab[i].len) == 0) | |
2705 | return i; | |
2706 | return -1; | |
2707 | } | |
2708 | \f | |
2709 | /* Support for Prolog. */ | |
2710 | ||
2711 | /* whole head (not only functor, but also arguments) | |
2712 | is gotten in compound term. */ | |
2713 | ||
2714 | void | |
2715 | prolog_getit (s, lineno, linecharno) | |
2716 | char *s; | |
2717 | int lineno; | |
2718 | long linecharno; | |
2719 | { | |
2720 | char nambuf[BUFSIZ], *save_s, tmpc; | |
2721 | int insquote, npar; | |
2722 | ||
2723 | save_s = s; | |
2724 | insquote = FALSE; | |
2725 | npar = 0; | |
2726 | while (1) | |
2727 | { | |
2728 | if (*s == '\0') /* syntax error. */ | |
2729 | return; | |
2730 | else if (insquote && *s == '\'' && *(s + 1) == '\'') | |
2731 | s += 2; | |
2732 | else if (*s == '\'') | |
2733 | { | |
2734 | insquote = !insquote; | |
2735 | s++; | |
2736 | } | |
2737 | else if (!insquote && *s == '(') | |
2738 | { | |
2739 | npar++; | |
2740 | s++; | |
2741 | } | |
2742 | else if (!insquote && *s == ')') | |
2743 | { | |
2744 | npar--; | |
2745 | s++; | |
2746 | if (npar == 0) | |
2747 | break; | |
2748 | else if (npar < 0) /* syntax error. */ | |
2749 | return; | |
2750 | } | |
2751 | else if (!insquote && *s == '.' && (isspace (*(s + 1)) || *(s + 1) == '\0')) | |
2752 | { /* fullstop. */ | |
2753 | if (npar != 0) /* syntax error. */ | |
2754 | return; | |
2755 | s++; | |
2756 | break; | |
2757 | } | |
2758 | else | |
2759 | s++; | |
2760 | } | |
2761 | tmpc = *s; | |
2762 | *s = '\0'; | |
2763 | strcpy (nambuf, save_s); | |
2764 | *s = tmpc; | |
2765 | pfnote (nambuf, TRUE, save_s, strlen (nambuf), lineno, linecharno); | |
2766 | } | |
2767 | ||
2768 | /* It is assumed that prolog predicate starts from column 0. */ | |
2769 | ||
2770 | void | |
2771 | prolog_funcs (fi) | |
2772 | FILE *fi; | |
2773 | { | |
2774 | void skip_comment (), prolog_getit (); | |
2775 | ||
2776 | lineno = linecharno = charno = 0; | |
2777 | while (!feof (fi)) | |
2778 | { | |
2779 | lineno++; | |
2780 | linecharno += charno; | |
2781 | charno = readline (&lb, fi) + 1; /* 1 for newline. */ | |
2782 | dbp = lb.buffer; | |
2783 | if (isspace (dbp[0])) /* not predicate header. */ | |
2784 | continue; | |
2785 | else if (dbp[0] == '%') /* comment. */ | |
2786 | continue; | |
2787 | else if (dbp[0] == '/' && dbp[1] == '*') /* comment. */ | |
2788 | skip_comment (&lb, fi, &lineno, &linecharno); | |
2789 | else /* found. */ | |
2790 | prolog_getit (dbp, lineno, linecharno); | |
2791 | } | |
2792 | } | |
2793 | ||
2794 | void | |
2795 | skip_comment (plb, fi, plineno, plinecharno) | |
2796 | struct linebuffer *plb; | |
2797 | FILE *fi; | |
2798 | int *plineno; /* result */ | |
2799 | long *plinecharno; /* result */ | |
2800 | { | |
2801 | while (!substr ("*/", plb->buffer)) | |
2802 | { | |
2803 | (*plineno)++; | |
2804 | *plinecharno += readline (plb, fi) + 1; | |
2805 | } /* 1 for newline. */ | |
2806 | } | |
2807 | ||
2808 | /* Return TRUE if 'sub' exists somewhere in 's'. */ | |
2809 | ||
2810 | int | |
2811 | substr (sub, s) | |
2812 | char *sub; | |
2813 | char *s; | |
2814 | { | |
2815 | while (*s && (s = index (s, *sub))) | |
2816 | if (prestr (sub, s)) | |
2817 | return (TRUE); | |
2818 | else | |
2819 | s++; | |
2820 | return (FALSE); | |
2821 | } | |
2822 | ||
2823 | /* Return TRUE if 'pre' is prefix of string 's'. */ | |
2824 | ||
2825 | int | |
2826 | prestr (pre, s) | |
2827 | char *pre; | |
2828 | char *s; | |
2829 | { | |
2830 | if (*pre == '\0') | |
2831 | return (TRUE); | |
2832 | else if (*pre == *s) | |
2833 | return (prestr (pre + 1, s + 1)); | |
2834 | else | |
2835 | return (FALSE); | |
2836 | } | |
2837 | \f | |
2838 | /* Initialize a linebuffer for use */ | |
2839 | ||
2840 | void | |
2841 | initbuffer (linebuffer) | |
2842 | struct linebuffer *linebuffer; | |
2843 | { | |
2844 | linebuffer->size = 200; | |
2845 | linebuffer->buffer = xnew (200, char); | |
2846 | } | |
2847 | ||
2848 | /* | |
2849 | * Read a line of text from `stream' into `linebuffer'. | |
2850 | * Return the number of characters read from `stream', | |
2851 | * which is the length of the line including the newline, if any. | |
2852 | */ | |
2853 | long | |
2854 | readline (linebuffer, stream) | |
2855 | struct linebuffer *linebuffer; | |
2856 | register FILE *stream; | |
2857 | { | |
2858 | char *buffer = linebuffer->buffer; | |
2859 | register char *p = linebuffer->buffer; | |
2860 | register char *pend; | |
2861 | int newline; /* 1 if ended with newline, 0 if ended with EOF */ | |
2862 | ||
2863 | pend = p + linebuffer->size; /* Separate to avoind 386/IX compiler bug. */ | |
2864 | ||
2865 | while (1) | |
2866 | { | |
2867 | register int c = getc (stream); | |
2868 | if (p == pend) | |
2869 | { | |
2870 | linebuffer->size *= 2; | |
2871 | buffer = (char *) xrealloc (buffer, linebuffer->size); | |
2872 | p += buffer - linebuffer->buffer; | |
2873 | pend = buffer + linebuffer->size; | |
2874 | linebuffer->buffer = buffer; | |
2875 | } | |
2876 | if (c < 0 || c == '\n') | |
2877 | { | |
2878 | *p = 0; | |
2879 | newline = (c == '\n' ? 1 : 0); | |
2880 | break; | |
2881 | } | |
2882 | *p++ = c; | |
2883 | } | |
2884 | ||
2885 | return p - buffer + newline; | |
2886 | } | |
2887 | \f | |
2888 | char * | |
2889 | savestr (cp) | |
2890 | char *cp; | |
2891 | { | |
2892 | return savenstr (cp, strlen (cp)); | |
2893 | } | |
2894 | ||
2895 | char * | |
2896 | savenstr (cp, len) | |
2897 | char *cp; | |
2898 | int len; | |
2899 | { | |
2900 | register char *dp; | |
2901 | ||
2902 | dp = xnew (len + 1, char); | |
2903 | (void) strncpy (dp, cp, len); | |
2904 | dp[len] = '\0'; | |
2905 | return dp; | |
2906 | } | |
2907 | ||
2908 | #ifdef notdef | |
2909 | /* | |
2910 | * Return the ptr in sp at which the character c last | |
2911 | * appears; NULL if not found | |
2912 | * | |
2913 | * Identical to v7 rindex, included for portability. | |
2914 | */ | |
2915 | ||
2916 | char * | |
2917 | rindex (sp, c) | |
2918 | register char *sp, c; | |
2919 | { | |
2920 | register char *r; | |
2921 | ||
2922 | r = NULL; | |
2923 | do | |
2924 | { | |
2925 | if (*sp == c) | |
2926 | r = sp; | |
2927 | } while (*sp++); | |
2928 | return (r); | |
2929 | } | |
2930 | ||
2931 | /* | |
2932 | * Return the ptr in sp at which the character c first | |
2933 | * appears; NULL if not found | |
2934 | * | |
2935 | * Identical to v7 index, included for portability. | |
2936 | */ | |
2937 | ||
2938 | char * | |
2939 | index (sp, c) | |
2940 | register char *sp, c; | |
2941 | { | |
2942 | do | |
2943 | { | |
2944 | if (*sp == c) | |
2945 | return (sp); | |
2946 | } while (*sp++); | |
2947 | return (NULL); | |
2948 | } | |
2949 | ||
2950 | #endif /* notdef */ | |
2951 | ||
2952 | /* Print error message and exit. */ | |
2953 | ||
2954 | /* VARARGS1 */ | |
2955 | void | |
2956 | fatal (s1, s2) | |
2957 | char *s1, *s2; | |
2958 | { | |
2959 | error (s1, s2); | |
2960 | exit (1); | |
2961 | } | |
2962 | ||
2963 | /* Print error message. `s1' is printf control string, `s2' is arg for it. */ | |
2964 | ||
2965 | /* VARARGS1 */ | |
2966 | void | |
2967 | error (s1, s2) | |
2968 | char *s1, *s2; | |
2969 | { | |
2970 | fprintf (stderr, "%s: ", progname); | |
2971 | fprintf (stderr, s1, s2); | |
2972 | fprintf (stderr, "\n"); | |
2973 | } | |
2974 | ||
2975 | /* Return a newly-allocated string whose contents concatenate those of s1, s2, s3. */ | |
2976 | ||
2977 | char * | |
2978 | concat (s1, s2, s3) | |
2979 | char *s1, *s2, *s3; | |
2980 | { | |
2981 | int len1 = strlen (s1), len2 = strlen (s2), len3 = strlen (s3); | |
2982 | char *result = xnew (len1 + len2 + len3 + 1, char); | |
2983 | ||
2984 | (void) strcpy (result, s1); | |
2985 | (void) strcpy (result + len1, s2); | |
2986 | (void) strcpy (result + len1 + len2, s3); | |
2987 | *(result + len1 + len2 + len3) = 0; | |
2988 | ||
2989 | return result; | |
2990 | } | |
2991 | ||
2992 | /* Like malloc but get fatal error if memory is exhausted. */ | |
2993 | ||
2994 | char * | |
2995 | xmalloc (size) | |
2996 | int size; | |
2997 | { | |
2998 | char *result = malloc (size); | |
2999 | if (!result) | |
3000 | fatal ("virtual memory exhausted", 0); | |
3001 | return result; | |
3002 | } | |
3003 | ||
3004 | char * | |
3005 | xrealloc (ptr, size) | |
3006 | char *ptr; | |
3007 | int size; | |
3008 | { | |
3009 | char *result = realloc (ptr, size); | |
3010 | if (!result) | |
3011 | fatal ("virtual memory exhausted"); | |
3012 | return result; | |
3013 | } |