240
|
1 /* Tags file maker to go with GNU Emacs
|
621
|
2 Copyright (C) 1984, 1987, 1988, 1989, 1992 Free Software Foundation, Inc. and Ken Arnold
|
240
|
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
|
621
|
34 #include "getopt.h"
|
|
35
|
401
|
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
|
240
|
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
|
|
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
|
|
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
|
|
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() \
|
|
413 (level==0 || c_ext && level==1 && structdef==sinbody)
|
|
414
|
621
|
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
|
240
|
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
|
|
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 */
|
|
464
|
|
465 /* Name this program was invoked with. */
|
|
466 char *progname;
|
|
467
|
621
|
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
|
240
|
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
|
|
510
|
|
511 void
|
621
|
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
|
|
617 void
|
240
|
618 main (argc, argv)
|
|
619 int argc;
|
|
620 char *argv[];
|
|
621 {
|
|
622 char cmd[100];
|
|
623 int i;
|
401
|
624 unsigned int nincluded_files = 0;
|
462
|
625 char **included_files = (char **) alloca (argc * sizeof (char *));
|
240
|
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
|
621
|
649 for (;;)
|
240
|
650 {
|
621
|
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)
|
240
|
658 {
|
621
|
659 case '\0':
|
|
660 /* If getopt returns '\0', then it has already processed a
|
|
661 long-named option. We should do nothing. */
|
|
662 break;
|
240
|
663
|
621
|
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)
|
|
679 {
|
|
680 fprintf (stderr,
|
|
681 "%s: -o flag may only be given once\n", progname);
|
240
|
682 goto usage;
|
|
683 }
|
621
|
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;
|
240
|
742 }
|
|
743 }
|
|
744
|
621
|
745 if (optind == argc)
|
240
|
746 {
|
621
|
747 fprintf (stderr, "%s: No input files specified.\n", progname);
|
|
748
|
240
|
749 usage:
|
621
|
750 fprintf (stderr, "%s: Try '%s --help' for a complete list of options.\n",
|
|
751 progname, progname);
|
240
|
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
|
|
781 #ifdef VMS
|
621
|
782 argc -= optind;
|
|
783 argv += optind;
|
|
784 while (gfnames (&argc, &argv, &got_err) != NULL)
|
240
|
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
|
621
|
799 for (; optind < argc; optind++)
|
240
|
800 {
|
621
|
801 this_file = argv[optind];
|
240
|
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 {
|
401
|
823 while (nincluded_files-- > 0)
|
|
824 fprintf (outf, "\f\n%s,include\n", *included_files++);
|
|
825
|
240
|
826 (void) fclose (outf);
|
|
827 exit (0);
|
|
828 }
|
|
829
|
|
830 if (cxref_style)
|
|
831 {
|
|
832 put_entries (head);
|
|
833 exit (GOOD);
|
|
834 }
|
621
|
835 if (update)
|
240
|
836 {
|
621
|
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++)
|
240
|
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 }
|
|
886 if (emacs_tags_format)
|
|
887 {
|
|
888 char *cp = rindex (file, '/');
|
|
889 if (cp)
|
|
890 ++cp;
|
|
891 else
|
|
892 cp = file;
|
|
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 */
|
621
|
1036 else if (cp && ((cp[1] != 'c'
|
|
1037 && cp[1] != 'h'
|
|
1038 && cp[1] != 'y')
|
240
|
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
|
|
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
|
|
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
|
|
1286 fprintf (stdout, "%-16s%4d %-16s %s\n",
|
|
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
|
|
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
|
|
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;
|
621
|
1441 tydef = none;
|
|
1442 next_token_is_func = 0;
|
240
|
1443
|
|
1444 C_create_stabs ();
|
|
1445
|
|
1446 while (!feof (inf))
|
|
1447 {
|
|
1448 c = *lp++;
|
|
1449 if (c == '\\')
|
|
1450 {
|
621
|
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. */
|
401
|
1454 if (*lp == 0)
|
|
1455 continue;
|
|
1456 lp++;
|
240
|
1457 c = ' ';
|
|
1458 }
|
|
1459 else if (incomm)
|
|
1460 {
|
401
|
1461 if (c == '*' && *lp == '/')
|
240
|
1462 {
|
401
|
1463 c = *lp++;
|
|
1464 incomm = FALSE;
|
240
|
1465 }
|
|
1466 }
|
|
1467 else if (inquote)
|
|
1468 {
|
|
1469 if (c == '"')
|
|
1470 inquote = FALSE;
|
|
1471 continue;
|
|
1472 }
|
|
1473 else if (inchar)
|
|
1474 {
|
|
1475 if (c == '\'')
|
|
1476 inchar = FALSE;
|
|
1477 continue;
|
|
1478 }
|
|
1479 else
|
|
1480 switch (c)
|
|
1481 {
|
|
1482 case '"':
|
|
1483 inquote = TRUE;
|
|
1484 continue;
|
|
1485 case '\'':
|
|
1486 inchar = TRUE;
|
|
1487 continue;
|
|
1488 case '/':
|
|
1489 if (*lp == '*')
|
|
1490 {
|
|
1491 lp++;
|
|
1492 incomm = TRUE;
|
|
1493 }
|
|
1494 else if (c_ext && *lp == '/')
|
|
1495 {
|
401
|
1496 c = 0; /* C++ comment: skip rest of line */
|
240
|
1497 }
|
|
1498 continue;
|
|
1499 case '#':
|
|
1500 if (lp == lb.buffer + 1 && definedef == dnone)
|
|
1501 definedef = dsharpseen;
|
|
1502 continue;
|
|
1503
|
|
1504 /*
|
|
1505 * The next two are to help the strucdef state machine.
|
|
1506 * They break when they are finished, so they don't interfere
|
|
1507 * with anything else that is going on.
|
|
1508 */
|
|
1509 case ':':
|
|
1510 if (structdef == stagseen)
|
|
1511 structdef = scolonseen;
|
|
1512 break;
|
|
1513 /* Not a struct definition when semicolon seen in non-sinbody context. */
|
|
1514 case ';':
|
|
1515 if (structdef != snone && structdef != sinbody)
|
|
1516 {
|
|
1517 structdef = snone;
|
|
1518 (void) strcpy (structtag, "<error 1>");
|
|
1519 }
|
|
1520 break;
|
|
1521
|
|
1522 case '{':
|
|
1523 if (tydef == begin)
|
|
1524 {
|
|
1525 tydef = middle;
|
|
1526 }
|
|
1527 switch (structdef)
|
|
1528 {
|
|
1529 case skeyseen: /* unnamed struct */
|
|
1530 structtag[0] = '\0';
|
|
1531 /* FALLTHRU */
|
|
1532 case stagseen:
|
|
1533 case scolonseen: /* named struct */
|
|
1534 structdef = sinbody;
|
|
1535 break;
|
|
1536 }
|
|
1537 level++;
|
|
1538 continue;
|
|
1539 case '}':
|
|
1540 if (!noindentypedefs && lp == lb.buffer + 1)
|
|
1541 level = 0; /* reset level if first column */
|
|
1542 else if (level > 0)
|
|
1543 level--;
|
|
1544 if (level == 0 && tydef == middle)
|
|
1545 {
|
|
1546 tydef = end;
|
|
1547 }
|
|
1548 if (level == 0)
|
|
1549 {
|
|
1550 structdef = snone;
|
|
1551 (void) strcpy (structtag, "<error 2>");
|
|
1552 }
|
|
1553 continue;
|
|
1554 }
|
|
1555 if (LEVEL_OK_FOR_FUNCDEF () && !inquote && !incomm && gotone == FALSE)
|
|
1556 {
|
|
1557 if (midtoken)
|
|
1558 {
|
|
1559 if (endtoken (c))
|
|
1560 {
|
|
1561 if (c_ext && c == ':' && *lp == ':' && intoken (*(lp + 1)))
|
|
1562 {
|
|
1563 /*
|
|
1564 * This handles :: in the middle, but not at beginning
|
|
1565 * of an identifier.
|
|
1566 */
|
|
1567 lp += 2;
|
|
1568 toklen += 3;
|
|
1569 }
|
|
1570 else
|
|
1571 {
|
401
|
1572 /* The following is no longer true,
|
|
1573 now that we advance to the next line
|
|
1574 at the end of processing the character. */
|
240
|
1575 /*
|
|
1576 * We've just finished lexing an identifier.
|
|
1577 * Note that if `c' is '\0', `lb' is the NEXT
|
|
1578 * line, `lp' points to the beginning of it, and
|
|
1579 * old pointers into `lb.buffer' may no longer be
|
|
1580 * valid, since `lb.buffer' may have been
|
|
1581 * reallocated. In this case (which corresponds
|
|
1582 * to an identifier followed immediately by a
|
|
1583 * newline), we re-read the line into lb1.
|
|
1584 *
|
|
1585 * This would be faster if the previous line's
|
|
1586 * buffer were always saved.
|
|
1587 */
|
|
1588 logical is_func;
|
|
1589 char *tok_linebuf;
|
|
1590 TOKEN tok;
|
|
1591 logical bingo, tok_at_end_of_line;
|
|
1592 char *lp_tmp; /* addressable */
|
|
1593
|
401
|
1594 #if 0
|
240
|
1595 if (c == '\0')
|
|
1596 {
|
|
1597 getline (GET_COOKIE (prev_linepos));
|
|
1598 tok_linebuf = lb1.buffer;
|
|
1599 tok_at_end_of_line = TRUE;
|
|
1600 tok.linestart = prev_linepos;
|
|
1601 tok.lineno = lineno - 1;
|
|
1602 }
|
|
1603 else
|
401
|
1604 #endif
|
240
|
1605 {
|
|
1606 tok_linebuf = lb.buffer;
|
|
1607 tok_at_end_of_line = FALSE;
|
|
1608 tok.linestart = linepos;
|
|
1609 tok.lineno = lineno;
|
|
1610 }
|
|
1611 tok.p = tok_linebuf + tokoff;
|
|
1612 tok.len = toklen;
|
|
1613 tok.rewritten = FALSE;
|
|
1614 lp_tmp = lp;
|
|
1615 bingo = consider_token (c, &lp_tmp, &tok,
|
|
1616 &is_func, c_ext, level);
|
|
1617 lp = lp_tmp;
|
|
1618 if (bingo)
|
|
1619 {
|
|
1620 if (GET_CHARNO (tok.linestart) != GET_CHARNO (linepos)
|
|
1621 && !tok_at_end_of_line)
|
|
1622 {
|
|
1623 /*
|
|
1624 * Resynchronize tok.p to point into the right
|
|
1625 * linebuffer.
|
|
1626 */
|
|
1627 getline (GET_COOKIE (tok.linestart));
|
|
1628 if (!tok.rewritten)
|
|
1629 tok.p = lb1.buffer + (tok.p - tok_linebuf);
|
|
1630 tok_linebuf = lb1.buffer;
|
|
1631 }
|
|
1632 if (structdef == sinbody && definedef == dnone && is_func)
|
|
1633 { /* function defined in C++ class body */
|
|
1634 sprintf (tokb, "%s::%.*s",
|
|
1635 structtag[0] == '\0' ? "_anonymous_"
|
|
1636 : structtag,
|
|
1637 tok.len, tok.p);
|
|
1638 tok.rewritten = TRUE;
|
|
1639 }
|
|
1640 else
|
|
1641 {
|
|
1642 sprintf (tokb, "%.*s", tok.len, tok.p);
|
|
1643 }
|
|
1644 pfnote (tokb, is_func, tok.rewritten, tok_linebuf,
|
|
1645 tokoff + toklen + (tok_at_end_of_line ? 0 : 1),
|
|
1646 tok.lineno, GET_CHARNO (tok.linestart));
|
|
1647 gotone = is_func; /* function */
|
|
1648 }
|
|
1649 midtoken = FALSE;
|
|
1650 }
|
|
1651 }
|
|
1652 else if (intoken (c))
|
|
1653 toklen++;
|
|
1654 }
|
|
1655 else if (begtoken (c))
|
|
1656 {
|
|
1657 tokoff = lp - 1 - lb.buffer;
|
|
1658 toklen = 1;
|
|
1659 midtoken = TRUE;
|
|
1660 }
|
|
1661 }
|
401
|
1662 /* Detect end of line, after having handled the last token on the line. */
|
|
1663 if (c == 0)
|
|
1664 {
|
|
1665 CNL;
|
|
1666 gotone = FALSE;
|
|
1667 }
|
240
|
1668 if (c == ';' && tydef == end) /* clean with typedefs */
|
|
1669 tydef = none;
|
|
1670 }
|
|
1671 }
|
|
1672
|
|
1673 /*
|
|
1674 * consider_token ()
|
|
1675 * checks to see if the current token is at the start of a
|
|
1676 * function, or corresponds to a typedef. It updates the input
|
|
1677 * line pointer *LPP so that the '(' will be in it when it returns.
|
|
1678 *
|
|
1679 * *IS_FUNC gets TRUE iff the token is a function.
|
|
1680 * C_EXT is which language we are looking at.
|
|
1681 *
|
|
1682 * In the future we will need some way to adjust where the end of
|
|
1683 * the token is; for instance, implementing the C++ keyword
|
|
1684 * `operator' properly will adjust the end of the token to be after
|
|
1685 * whatever follows `operator'.
|
|
1686 *
|
|
1687 * Globals
|
|
1688 * structdef IN OUT
|
|
1689 * definedef IN OUT
|
|
1690 * tydef IN OUT
|
|
1691 */
|
|
1692
|
|
1693 logical
|
|
1694 consider_token (c, lpp, tokp, is_func, c_ext, level)
|
|
1695 reg char c; /* IN: first char after the token */
|
|
1696 char **lpp; /* IN OUT: *lpp points to 2nd char after the token */
|
|
1697 reg TOKEN *tokp; /* IN */
|
|
1698 logical *is_func; /* OUT */
|
|
1699 int c_ext; /* IN */
|
|
1700 int level; /* IN */
|
|
1701 {
|
|
1702 reg char *lp = *lpp;
|
|
1703 logical firsttok; /* TRUE if have seen first token in ()'s */
|
|
1704 Stab_entry *tokse = stab_find (get_C_stab (c_ext), tokp->p, tokp->len);
|
|
1705 enum sym_type toktype = stab_type (tokse);
|
|
1706
|
|
1707 *is_func = TRUE; /* a function */
|
|
1708
|
|
1709 /*
|
|
1710 * Advance the definedef state machine. We set `gotone' for good measure;
|
|
1711 * it's redundant.
|
|
1712 */
|
|
1713 switch (definedef)
|
|
1714 {
|
|
1715 case dnone:
|
|
1716 /* We're not on a preprocessor line. */
|
|
1717 break;
|
|
1718 case dsharpseen:
|
|
1719 if (toktype == st_C_define)
|
|
1720 {
|
|
1721 definedef = ddefineseen;
|
|
1722 gotone = FALSE;
|
|
1723 }
|
|
1724 else
|
|
1725 {
|
|
1726 definedef = dignorerest;
|
|
1727 gotone = TRUE;
|
|
1728 }
|
|
1729 goto badone;
|
|
1730 case ddefineseen:
|
|
1731 /*
|
|
1732 * Make a tag for any macro.
|
|
1733 * This will flub up if there is a newline immediately following
|
|
1734 * the macro name.
|
|
1735 */
|
|
1736 *is_func = (c == '(');
|
|
1737 definedef = dignorerest;
|
|
1738 gotone = TRUE;
|
|
1739 if (!*is_func && !constantypedefs)
|
|
1740 goto badone;
|
|
1741 goto goodone;
|
|
1742 case dignorerest:
|
|
1743 goto badone;
|
|
1744 default:
|
|
1745 error ("internal error: definedef value");
|
|
1746 }
|
|
1747
|
|
1748 /*
|
|
1749 * Skip whitespace and comments after the token. This loop should
|
|
1750 * also skip C++ comments.
|
|
1751 */
|
|
1752 while (1)
|
|
1753 {
|
|
1754 /* At whitespace => skip it. */
|
|
1755 if (iswhite (c))
|
|
1756 {
|
|
1757 c = *lp++;
|
|
1758 }
|
|
1759 /* At a comment => skip to end of comment. */
|
|
1760 else if (c == '/' && *lp == '*')
|
|
1761 {
|
|
1762 /* If we find a comment, skip it. */
|
|
1763 while (!(c == '*' && *lp == '/'))
|
|
1764 {
|
|
1765 c = *lp++;
|
|
1766 if (c == 0)
|
|
1767 {
|
401
|
1768 lp--;
|
|
1769 break;
|
240
|
1770 }
|
|
1771 }
|
|
1772 if (c == '*' && *lp == '/')
|
|
1773 {
|
|
1774 lp++; /* lp now points past the '/' */
|
|
1775 c = *lp++; /* c is now the --whatever-- after the '/' */
|
|
1776 }
|
|
1777 }
|
|
1778 else
|
|
1779 break;
|
|
1780
|
|
1781 /* If we arrived at eof or eol, decide which one it is.
|
|
1782 If it's eol, advance to the next line. */
|
|
1783
|
|
1784 if (c == 0)
|
|
1785 {
|
401
|
1786 lp--;
|
|
1787 break;
|
240
|
1788 }
|
|
1789 }
|
|
1790
|
|
1791 /*
|
|
1792 * If you have custom token types, or when configuration files can
|
|
1793 * define custom token types, this switch will be larger.
|
|
1794 */
|
|
1795 switch (toktype)
|
|
1796 {
|
|
1797 case st_C_typedef:
|
|
1798 if (typedefs)
|
|
1799 {
|
|
1800 tydef = begin;
|
|
1801 goto badone;
|
|
1802 }
|
|
1803 break;
|
|
1804 case st_C_typespec:
|
|
1805 if (tydef == begin || tydef == end)
|
|
1806 {
|
|
1807 tydef = end;
|
|
1808 goto badone;
|
|
1809 }
|
|
1810 break;
|
|
1811 }
|
|
1812
|
|
1813 /*
|
|
1814 * This structdef business is currently only invoked when level==0.
|
|
1815 * It should be recursively invoked whatever the level, and a stack of
|
|
1816 * states kept, to allow for definitions of structs within structs.
|
|
1817 *
|
|
1818 * This structdef business is NOT invoked when we are ctags and the
|
|
1819 * file is plain C. This is because a struct tag may have the same
|
|
1820 * name as another tag, and this loses with ctags.
|
|
1821 *
|
|
1822 * This if statement deals with the tydef state machine as follows: if
|
|
1823 * tydef==begin and token is struct/union/class/enum, goto badone.
|
|
1824 * All the other code here is for the structdef state machine.
|
|
1825 */
|
|
1826 switch (toktype)
|
|
1827 {
|
|
1828 case st_C_struct:
|
|
1829 case st_C_enum:
|
|
1830 if (tydef == begin || (typedefs_and_cplusplus && level == 0 && structdef == snone))
|
|
1831 {
|
|
1832 structdef = skeyseen;
|
|
1833 structkey = tokse;
|
|
1834 }
|
|
1835 goto badone;
|
|
1836 }
|
|
1837
|
|
1838 if (structdef == skeyseen)
|
|
1839 {
|
|
1840 /* If next char is '{' or (for C++) ':', found a structure tag. */
|
|
1841 if (c == '{' || (c_ext && c == ':'))
|
|
1842 {
|
|
1843 /*
|
|
1844 * We should do this slightly differently for straight C:
|
|
1845 * instead of defining `tag', as we now do, we should define
|
|
1846 * `struct tag'. (Do this only if the find-tag defaulting is
|
|
1847 * done on a sophisticated per-mode basis, so that if the user
|
|
1848 * says meta-. anywhere in `struct foo', the default comes out
|
|
1849 * `struct foo', not `struct' or `foo'.) This will require
|
|
1850 * remembering which keyword (struct/union/class/enum) we saw, as a
|
|
1851 * Stab_entry* -- this will also make it possible to merge the
|
|
1852 * skeyseen and senumseen states, if we want.
|
|
1853 */
|
|
1854 if (stab_type (structkey) == st_C_struct)
|
|
1855 {
|
|
1856 (void) strncpy (structtag, tokp->p, tokp->len);
|
|
1857 structtag[tokp->len] = '\0'; /* for struct/union/class */
|
|
1858 structdef = stagseen;
|
|
1859 }
|
|
1860 else
|
|
1861 {
|
|
1862 structtag[0] = '\0'; /* for enum */
|
|
1863 }
|
|
1864 *is_func = FALSE; /* not a function */
|
|
1865 goto goodone;
|
|
1866 }
|
|
1867 else
|
|
1868 {
|
|
1869 /* Not a definition: reset structdef */
|
|
1870 structdef = snone;
|
|
1871 (void) strcpy (structtag, "<error 3>");
|
|
1872 }
|
|
1873 /* Now what? And how does/should this stuff interact with tydef?? */
|
|
1874 /* Also maybe reset lp to *lpp for benefit of the function finding code. */
|
|
1875 }
|
|
1876 if (tydef == begin)
|
|
1877 {
|
|
1878 tydef = end;
|
|
1879 goto badone;
|
|
1880 }
|
|
1881 if (tydef == end)
|
|
1882 {
|
|
1883 *is_func = 0;
|
|
1884 goto goodone;
|
|
1885 }
|
|
1886 /* Detect GNUmacs's function-defining macros. */
|
621
|
1887 if (definedef == dnone)
|
240
|
1888 {
|
621
|
1889 if (strneq (tokp->p, "DEF", 3))
|
|
1890 {
|
|
1891 next_token_is_func = TRUE;
|
|
1892 goto badone;
|
|
1893 }
|
|
1894 else if (strneq (tokp->p, "EXFUN", 5))
|
|
1895 {
|
|
1896 next_token_is_func = FALSE;
|
|
1897 goto badone;
|
|
1898 }
|
240
|
1899 }
|
|
1900 if (next_token_is_func)
|
|
1901 {
|
|
1902 next_token_is_func = FALSE;
|
|
1903 goto goodone;
|
|
1904 }
|
|
1905 if (c != '(')
|
|
1906 goto badone;
|
|
1907 firsttok = FALSE;
|
|
1908 while ((c = *lp++) != ')')
|
|
1909 {
|
|
1910 if (c == 0)
|
|
1911 {
|
401
|
1912 lp--;
|
|
1913 break;
|
240
|
1914 }
|
|
1915 /*
|
|
1916 * This line used to confuse ctags:
|
|
1917 * int (*oldhup)();
|
|
1918 * This fixes it. A nonwhite char before the first
|
|
1919 * token, other than a / (in case of a comment in there)
|
|
1920 * makes this not a declaration.
|
|
1921 */
|
|
1922 if (begtoken (c) || c == '/')
|
|
1923 firsttok++;
|
|
1924 else if (!iswhite (c) && !firsttok)
|
|
1925 goto badone;
|
|
1926 }
|
|
1927 while (iswhite (c = *lp++))
|
|
1928 {
|
|
1929 if (c == 0)
|
|
1930 {
|
401
|
1931 lp--;
|
|
1932 break;
|
240
|
1933 }
|
|
1934 }
|
|
1935 if (!isgood (c))
|
|
1936 goto badone;
|
|
1937
|
|
1938 goodone:
|
|
1939 *lpp = lp - 1;
|
|
1940 return TRUE;
|
|
1941
|
|
1942 badone:
|
|
1943 *lpp = lp - 1;
|
|
1944 return FALSE;
|
|
1945 }
|
|
1946
|
|
1947 void
|
|
1948 getline (atcookie)
|
|
1949 long atcookie;
|
|
1950 {
|
|
1951 long saveftell = ftell (inf);
|
|
1952
|
|
1953 (void) fseek (inf, atcookie, 0);
|
|
1954 (void) readline (&lb1, inf);
|
|
1955 (void) fseek (inf, saveftell, 0);
|
|
1956 }
|
|
1957
|
|
1958 /* Fortran parsing */
|
|
1959
|
|
1960 char *dbp;
|
|
1961 int pfcnt;
|
|
1962
|
|
1963 int
|
|
1964 PF_funcs (fi)
|
|
1965 FILE *fi;
|
|
1966 {
|
|
1967 lineno = 0;
|
|
1968 charno = 0;
|
|
1969 pfcnt = 0;
|
|
1970
|
|
1971 while (!feof (fi))
|
|
1972 {
|
|
1973 lineno++;
|
|
1974 linecharno = charno;
|
|
1975 charno += readline (&lb, fi);
|
|
1976 dbp = lb.buffer;
|
|
1977 if (*dbp == '%')
|
|
1978 dbp++; /* Ratfor escape to fortran */
|
|
1979 while (isspace (*dbp))
|
|
1980 dbp++;
|
|
1981 if (*dbp == 0)
|
|
1982 continue;
|
|
1983 switch (*dbp | ' ')
|
|
1984 {
|
|
1985 case 'i':
|
|
1986 if (tail ("integer"))
|
|
1987 takeprec ();
|
|
1988 break;
|
|
1989 case 'r':
|
|
1990 if (tail ("real"))
|
|
1991 takeprec ();
|
|
1992 break;
|
|
1993 case 'l':
|
|
1994 if (tail ("logical"))
|
|
1995 takeprec ();
|
|
1996 break;
|
|
1997 case 'c':
|
|
1998 if (tail ("complex") || tail ("character"))
|
|
1999 takeprec ();
|
|
2000 break;
|
|
2001 case 'd':
|
|
2002 if (tail ("double"))
|
|
2003 {
|
|
2004 while (isspace (*dbp))
|
|
2005 dbp++;
|
|
2006 if (*dbp == 0)
|
|
2007 continue;
|
|
2008 if (tail ("precision"))
|
|
2009 break;
|
|
2010 continue;
|
|
2011 }
|
|
2012 break;
|
|
2013 }
|
|
2014 while (isspace (*dbp))
|
|
2015 dbp++;
|
|
2016 if (*dbp == 0)
|
|
2017 continue;
|
|
2018 switch (*dbp | ' ')
|
|
2019 {
|
|
2020 case 'f':
|
|
2021 if (tail ("function"))
|
|
2022 getit ();
|
|
2023 continue;
|
|
2024 case 's':
|
|
2025 if (tail ("subroutine"))
|
|
2026 getit ();
|
|
2027 continue;
|
|
2028 case 'p':
|
|
2029 if (tail ("program"))
|
|
2030 {
|
|
2031 getit ();
|
|
2032 continue;
|
|
2033 }
|
|
2034 if (tail ("procedure"))
|
|
2035 getit ();
|
|
2036 continue;
|
|
2037 }
|
|
2038 }
|
|
2039 return (pfcnt);
|
|
2040 }
|
|
2041
|
|
2042 logical
|
|
2043 tail (cp)
|
|
2044 char *cp;
|
|
2045 {
|
|
2046 register int len = 0;
|
|
2047
|
|
2048 while (*cp && (*cp & ~' ') == ((*(dbp + len)) & ~' '))
|
|
2049 cp++, len++;
|
|
2050 if (*cp == 0)
|
|
2051 {
|
|
2052 dbp += len;
|
|
2053 return (1);
|
|
2054 }
|
|
2055 return (0);
|
|
2056 }
|
|
2057
|
|
2058 void
|
|
2059 takeprec ()
|
|
2060 {
|
|
2061 while (isspace (*dbp))
|
|
2062 dbp++;
|
|
2063 if (*dbp != '*')
|
|
2064 return;
|
|
2065 dbp++;
|
|
2066 while (isspace (*dbp))
|
|
2067 dbp++;
|
|
2068 if (!isdigit (*dbp))
|
|
2069 {
|
|
2070 --dbp; /* force failure */
|
|
2071 return;
|
|
2072 }
|
|
2073 do
|
|
2074 dbp++;
|
|
2075 while (isdigit (*dbp));
|
|
2076 }
|
|
2077
|
|
2078 void
|
|
2079 getit ()
|
|
2080 {
|
|
2081 register char *cp;
|
|
2082 char c;
|
|
2083 char nambuf[BUFSIZ];
|
|
2084
|
|
2085 while (isspace (*dbp))
|
|
2086 dbp++;
|
|
2087 if (*dbp == 0 || (!isalpha (*dbp)) && (*dbp != '_') && (*dbp != '$'))
|
|
2088 return;
|
|
2089 for (cp = dbp + 1; *cp && (isalpha (*cp) || isdigit (*cp)
|
|
2090 || (*cp == '_') || (*cp == '$')); cp++)
|
|
2091 continue;
|
|
2092 c = cp[0];
|
|
2093 cp[0] = 0;
|
|
2094 (void) strcpy (nambuf, dbp);
|
|
2095 cp[0] = c;
|
|
2096 pfnote (nambuf, TRUE, FALSE, lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
|
|
2097 pfcnt++;
|
|
2098 }
|
|
2099
|
|
2100 /* Handle a file of assembler code. */
|
|
2101
|
|
2102 void
|
|
2103 Asm_funcs (fi)
|
|
2104 FILE *fi;
|
|
2105 {
|
|
2106 int i;
|
|
2107 register char c;
|
|
2108
|
|
2109 lineno = 0;
|
|
2110 charno = 0;
|
|
2111 pfcnt = 0;
|
|
2112
|
|
2113 while (!feof (fi))
|
|
2114 {
|
|
2115 lineno++;
|
|
2116 linecharno = charno;
|
|
2117 charno += readline (&lb, fi);
|
|
2118 dbp = lb.buffer;
|
|
2119
|
|
2120 for (i = 0; ((c = dbp[i]) && !isspace (c)) && (c != ':'); i++)
|
|
2121 ;
|
|
2122
|
|
2123 if ((i > 0) && (c == ':'))
|
|
2124 getit ();
|
|
2125 }
|
|
2126 }
|
|
2127
|
|
2128 /* Added by Mosur Mohan, 4/22/88 */
|
|
2129 /* Pascal parsing */
|
|
2130
|
|
2131 #define GET_NEW_LINE \
|
|
2132 { \
|
|
2133 linecharno = charno; lineno++; \
|
|
2134 charno += 1 + readline (&lb, inf); \
|
|
2135 dbp = lb.buffer; \
|
|
2136 }
|
|
2137
|
|
2138 /* Locates tags for procedures & functions.
|
|
2139 * Doesn't do any type- or var-definitions.
|
|
2140 * It does look for the keyword "extern" or "forward"
|
|
2141 * immediately following the procedure statement;
|
|
2142 * if found, the tag is skipped.
|
|
2143 */
|
|
2144
|
|
2145 void
|
|
2146 PAS_funcs (fi)
|
|
2147 FILE *fi;
|
|
2148 {
|
|
2149 struct linebuffer tline; /* mostly copied from C_entries */
|
|
2150 long save_lcno;
|
|
2151 int save_lineno;
|
|
2152 char c, *cp;
|
|
2153 char nambuf[BUFSIZ];
|
|
2154
|
|
2155 logical /* each of these flags is TRUE iff: */
|
|
2156 incomm1, /* point is inside {..} comment */
|
|
2157 incomm2, /* point is inside (*..*) comment */
|
|
2158 inquote, /* point is inside '..' string */
|
|
2159 get_tagname, /* point is after PROCEDURE/FUNCTION */
|
|
2160 /* keyword, so next item = potential tag */
|
|
2161 found_tag, /* point is after a potential tag */
|
|
2162 inparms, /* point is within parameter-list */
|
|
2163 verify_tag; /* point has passed the parm-list, so the */
|
|
2164 /* next token will determine whether */
|
|
2165 /* this is a FORWARD/EXTERN to be */
|
|
2166 /* ignored, or whether it is a real tag */
|
|
2167
|
|
2168 lineno = 0;
|
|
2169 charno = 0;
|
|
2170 dbp = lb.buffer;
|
|
2171 *dbp = 0;
|
|
2172 initbuffer (&tline);
|
|
2173
|
|
2174 incomm1 = incomm2 = inquote = FALSE;
|
|
2175 found_tag = FALSE; /* have a proc name; check if extern */
|
|
2176 get_tagname = FALSE; /* have found "procedure" keyword */
|
|
2177 inparms = FALSE; /* found '(' after "proc" */
|
|
2178 verify_tag = FALSE; /* check if "extern" is ahead */
|
|
2179
|
|
2180 /* long main loop to get next char */
|
|
2181 while (!feof (fi))
|
|
2182 {
|
|
2183 c = *dbp++;
|
|
2184 if (c == 0) /* if end of line */
|
|
2185 {
|
|
2186 GET_NEW_LINE;
|
|
2187 if (*dbp == 0)
|
|
2188 continue;
|
|
2189 if (!((found_tag && verify_tag) ||
|
|
2190 get_tagname))
|
|
2191 c = *dbp++; /* only if don't need *dbp pointing */
|
|
2192 /* to the beginning of the name of */
|
|
2193 /* the procedure or function */
|
|
2194 }
|
|
2195 if (incomm1) /* within { - } comments */
|
|
2196 {
|
|
2197 if (c == '}')
|
|
2198 incomm1 = FALSE;
|
|
2199 continue;
|
|
2200 }
|
|
2201 else if (incomm2) /* within (* - *) comments */
|
|
2202 {
|
|
2203 if (c == '*')
|
|
2204 {
|
|
2205 while ((c = *dbp++) == '*')
|
|
2206 continue;
|
|
2207 if (c == 0)
|
|
2208 GET_NEW_LINE;
|
|
2209 if (c == ')')
|
|
2210 incomm2 = FALSE;
|
|
2211 }
|
|
2212 continue;
|
|
2213 }
|
|
2214 else if (inquote)
|
|
2215 {
|
|
2216 if (c == '\'')
|
|
2217 inquote = FALSE;
|
|
2218 continue;
|
|
2219 }
|
|
2220 else
|
|
2221 switch (c)
|
|
2222 {
|
|
2223 case '\'':
|
|
2224 inquote = TRUE; /* found first quote */
|
|
2225 continue;
|
|
2226 case '{': /* found open-{-comment */
|
|
2227 incomm1 = TRUE;
|
|
2228 continue;
|
|
2229 case '(':
|
|
2230 if (*dbp == '*') /* found open-(*-comment */
|
|
2231 {
|
|
2232 incomm2 = TRUE;
|
|
2233 dbp++;
|
|
2234 }
|
|
2235 else if (found_tag) /* found '(' after tag, i.e., parm-list */
|
|
2236 inparms = TRUE;
|
|
2237 continue;
|
|
2238 case ')': /* end of parms list */
|
|
2239 if (inparms)
|
|
2240 inparms = FALSE;
|
|
2241 continue;
|
|
2242 case ';':
|
|
2243 if ((found_tag) && (!inparms)) /* end of proc or fn stmt */
|
|
2244 {
|
|
2245 verify_tag = TRUE;
|
|
2246 break;
|
|
2247 }
|
|
2248 continue;
|
|
2249 }
|
|
2250 if ((found_tag) && (verify_tag) && (*dbp != ' '))
|
|
2251 {
|
|
2252 /* check if this is an "extern" declaration */
|
|
2253 if (*dbp == 0)
|
|
2254 continue;
|
|
2255 if ((*dbp == 'e') || (*dbp == 'E'))
|
|
2256 {
|
|
2257 if (tail ("extern")) /* superfluous, really! */
|
|
2258 {
|
|
2259 found_tag = FALSE;
|
|
2260 verify_tag = FALSE;
|
|
2261 }
|
|
2262 }
|
|
2263 else if ((*dbp == 'f') || (*dbp == 'F'))
|
|
2264 {
|
|
2265 if (tail ("forward")) /* check for forward reference */
|
|
2266 {
|
|
2267 found_tag = FALSE;
|
|
2268 verify_tag = FALSE;
|
|
2269 }
|
|
2270 }
|
|
2271 if ((found_tag) && (verify_tag)) /* not external proc, so make tag */
|
|
2272 {
|
|
2273 found_tag = FALSE;
|
|
2274 verify_tag = FALSE;
|
|
2275 pfnote (nambuf, TRUE, FALSE,
|
|
2276 tline.buffer, cp - tline.buffer + 1,
|
|
2277 save_lineno, save_lcno);
|
|
2278 continue;
|
|
2279 }
|
|
2280 }
|
|
2281 if (get_tagname) /* grab name of proc or fn */
|
|
2282 {
|
|
2283 if (*dbp == 0)
|
|
2284 continue;
|
|
2285
|
|
2286 /* save all values for later tagging */
|
|
2287 tline.size = lb.size;
|
|
2288 strcpy (tline.buffer, lb.buffer);
|
|
2289 save_lineno = lineno;
|
|
2290 save_lcno = linecharno;
|
|
2291
|
|
2292 /* grab block name */
|
|
2293 for (cp = dbp + 1; *cp && (!endtoken (*cp)); cp++)
|
|
2294 continue;
|
|
2295 c = cp[0];
|
|
2296 cp[0] = 0;
|
|
2297 strcpy (nambuf, dbp);
|
|
2298 cp[0] = c;
|
|
2299 dbp = cp; /* restore dbp to e-o-token */
|
|
2300 get_tagname = FALSE;
|
|
2301 found_tag = TRUE;
|
|
2302 continue;
|
|
2303
|
|
2304 /* and proceed to check for "extern" */
|
|
2305 }
|
|
2306 if ((!incomm1) && (!incomm2) && (!inquote) &&
|
|
2307 (!found_tag) && (!get_tagname))
|
|
2308 {
|
|
2309 /* check for proc/fn keywords */
|
|
2310 switch (c | ' ')
|
|
2311 {
|
|
2312 case 'p':
|
|
2313 if (tail ("rocedure")) /* c = 'p', dbp has advanced */
|
|
2314 get_tagname = TRUE;
|
|
2315 continue;
|
|
2316 case 'f':
|
|
2317 if (tail ("unction"))
|
|
2318 get_tagname = TRUE;
|
|
2319 continue;
|
|
2320 }
|
|
2321 }
|
|
2322 } /* while not e-o-f */
|
|
2323 }
|
|
2324
|
|
2325 /*
|
|
2326 * lisp tag functions
|
|
2327 * just look for (def or (DEF
|
|
2328 */
|
|
2329
|
|
2330 void
|
|
2331 L_funcs (fi)
|
|
2332 FILE *fi;
|
|
2333 {
|
|
2334 lineno = 0;
|
|
2335 charno = 0;
|
|
2336 pfcnt = 0;
|
|
2337
|
|
2338 while (!feof (fi))
|
|
2339 {
|
|
2340 lineno++;
|
|
2341 linecharno = charno;
|
|
2342 charno += readline (&lb, fi);
|
|
2343 dbp = lb.buffer;
|
|
2344 if (dbp[0] == '(')
|
|
2345 {
|
|
2346 if (L_isdef (dbp))
|
|
2347 {
|
|
2348 while (!isspace (*dbp))
|
|
2349 dbp++;
|
|
2350 while (isspace (*dbp))
|
|
2351 dbp++;
|
|
2352 L_getit ();
|
|
2353 }
|
|
2354 else
|
|
2355 {
|
|
2356 /* Check for (foo::defmumble name-defined ... */
|
|
2357 while (*dbp && *dbp != ':' && !isspace (*dbp)
|
|
2358 && *dbp != '(' && *dbp != ')')
|
|
2359 dbp++;
|
|
2360 if (*dbp == ':')
|
|
2361 {
|
|
2362 while (*dbp == ':')
|
|
2363 dbp++;
|
|
2364
|
|
2365 if (L_isdef (dbp))
|
|
2366 {
|
|
2367 while (!isspace (*dbp))
|
|
2368 dbp++;
|
|
2369 while (isspace (*dbp))
|
|
2370 dbp++;
|
|
2371 L_getit ();
|
|
2372 }
|
|
2373 }
|
|
2374 }
|
|
2375 }
|
|
2376 }
|
|
2377 }
|
|
2378
|
|
2379 int
|
|
2380 L_isdef (dbp)
|
|
2381 char *dbp;
|
|
2382 {
|
|
2383 return ((dbp[1] == 'D' || dbp[1] == 'd') &&
|
|
2384 (dbp[2] == 'E' || dbp[2] == 'e') &&
|
|
2385 (dbp[3] == 'F' || dbp[3] == 'f'));
|
|
2386 }
|
|
2387
|
|
2388 void
|
|
2389 L_getit ()
|
|
2390 {
|
|
2391 register char *cp;
|
|
2392 char c;
|
|
2393 char nambuf[BUFSIZ];
|
|
2394
|
|
2395 if (*dbp == 0)
|
|
2396 return;
|
|
2397 for (cp = dbp + 1; *cp && *cp != '(' && *cp != ' '; cp++)
|
|
2398 continue;
|
|
2399 c = cp[0];
|
|
2400 cp[0] = 0;
|
|
2401 (void) strcpy (nambuf, dbp);
|
|
2402 cp[0] = c;
|
|
2403 pfnote (nambuf, TRUE, FALSE, lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
|
|
2404 pfcnt++;
|
|
2405 }
|
|
2406
|
|
2407 /*
|
|
2408 * Scheme tag functions
|
|
2409 * look for (def... xyzzy
|
|
2410 * look for (def... (xyzzy
|
|
2411 * look for (def ... ((...(xyzzy ....
|
|
2412 * look for (set! xyzzy
|
|
2413 */
|
|
2414
|
|
2415 static void get_scheme ();
|
|
2416
|
|
2417 void
|
|
2418 Scheme_funcs (fi)
|
|
2419 FILE *fi;
|
|
2420 {
|
|
2421 lineno = 0;
|
|
2422 charno = 0;
|
|
2423 pfcnt = 0;
|
|
2424
|
|
2425 while (!feof (fi))
|
|
2426 {
|
|
2427 lineno++;
|
|
2428 linecharno = charno;
|
|
2429 charno += readline (&lb, fi);
|
|
2430 dbp = lb.buffer;
|
|
2431 if (dbp[0] == '(' &&
|
|
2432 (dbp[1] == 'D' || dbp[1] == 'd') &&
|
|
2433 (dbp[2] == 'E' || dbp[2] == 'e') &&
|
|
2434 (dbp[3] == 'F' || dbp[3] == 'f'))
|
|
2435 {
|
|
2436 while (!isspace (*dbp))
|
|
2437 dbp++;
|
|
2438 /* Skip over open parens and white space */
|
|
2439 while (*dbp && (isspace (*dbp) || *dbp == '('))
|
|
2440 dbp++;
|
|
2441 get_scheme ();
|
|
2442 }
|
|
2443 if (dbp[0] == '(' &&
|
|
2444 (dbp[1] == 'S' || dbp[1] == 's') &&
|
|
2445 (dbp[2] == 'E' || dbp[2] == 'e') &&
|
|
2446 (dbp[3] == 'T' || dbp[3] == 't') &&
|
|
2447 (dbp[4] == '!' || dbp[4] == '!') &&
|
|
2448 (isspace (dbp[5])))
|
|
2449 {
|
|
2450 while (!isspace (*dbp))
|
|
2451 dbp++;
|
|
2452 /* Skip over white space */
|
|
2453 while (isspace (*dbp))
|
|
2454 dbp++;
|
|
2455 get_scheme ();
|
|
2456 }
|
|
2457 }
|
|
2458 }
|
|
2459
|
|
2460 static void
|
|
2461 get_scheme ()
|
|
2462 {
|
|
2463 register char *cp;
|
|
2464 char c;
|
|
2465 char nambuf[BUFSIZ];
|
|
2466
|
|
2467 if (*dbp == 0)
|
|
2468 return;
|
|
2469 /* Go till you get to white space or a syntactic break */
|
|
2470 for (cp = dbp + 1; *cp && *cp != '(' && *cp != ')' && !isspace (*cp); cp++)
|
|
2471 continue;
|
|
2472 /* Null terminate the string there. */
|
|
2473 c = cp[0];
|
|
2474 cp[0] = 0;
|
|
2475 /* Copy the string */
|
|
2476 strcpy (nambuf, dbp);
|
|
2477 /* Unterminate the string */
|
|
2478 cp[0] = c;
|
|
2479 /* Announce the change */
|
|
2480 pfnote (nambuf, TRUE, FALSE, lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
|
|
2481 pfcnt++;
|
|
2482 }
|
|
2483
|
|
2484 /* Find tags in TeX and LaTeX input files. */
|
|
2485
|
|
2486 /* TEX_toktab is a table of TeX control sequences that define tags.
|
|
2487 Each TEX_tabent records one such control sequence.
|
|
2488 CONVERT THIS TO USE THE Stab TYPE!! */
|
|
2489
|
|
2490 struct TEX_tabent
|
|
2491 {
|
|
2492 char *name;
|
|
2493 int len;
|
|
2494 };
|
|
2495
|
|
2496 struct TEX_tabent *TEX_toktab = NULL; /* Table with tag tokens */
|
|
2497
|
|
2498 /* Default set of control sequences to put into TEX_toktab.
|
|
2499 The value of environment var TEXTAGS is prepended to this. */
|
|
2500
|
|
2501 static char *TEX_defenv =
|
|
2502 ":chapter:section:subsection:subsubsection:eqno:label:ref:cite:bibitem:typeout";
|
|
2503
|
|
2504 void TEX_mode ();
|
|
2505 struct TEX_tabent *TEX_decode_env ();
|
|
2506 void TEX_getit ();
|
|
2507 int TEX_Token ();
|
|
2508
|
|
2509 static char TEX_esc = '\\';
|
|
2510 static char TEX_opgrp = '{';
|
|
2511 static char TEX_clgrp = '}';
|
|
2512
|
|
2513 /*
|
|
2514 * TeX/LaTeX scanning loop.
|
|
2515 */
|
|
2516
|
|
2517 void
|
|
2518 TEX_funcs (fi)
|
|
2519 FILE *fi;
|
|
2520 {
|
|
2521 char *lasthit;
|
|
2522
|
|
2523 lineno = 0;
|
|
2524 charno = 0;
|
|
2525 pfcnt = 0;
|
|
2526
|
|
2527 /* Select either \ or ! as escape character. */
|
|
2528 TEX_mode (fi);
|
|
2529
|
|
2530 /* Initialize token table once from environment. */
|
|
2531 if (!TEX_toktab)
|
|
2532 TEX_toktab = TEX_decode_env ("TEXTAGS", TEX_defenv);
|
|
2533
|
|
2534 while (!feof (fi))
|
|
2535 {
|
|
2536 lineno++;
|
|
2537 linecharno = charno;
|
|
2538 charno += readline (&lb, fi);
|
|
2539 dbp = lb.buffer;
|
|
2540 lasthit = dbp;
|
|
2541
|
|
2542 while (!feof (fi))
|
|
2543 { /* Scan each line in file */
|
|
2544 lineno++;
|
|
2545 linecharno = charno;
|
|
2546 charno += readline (&lb, fi);
|
|
2547 dbp = lb.buffer;
|
|
2548 lasthit = dbp;
|
|
2549 while (dbp = index (dbp, TEX_esc)) /* Look at each escape in line */
|
|
2550 {
|
|
2551 register int i;
|
|
2552
|
|
2553 if (!*(++dbp))
|
|
2554 break;
|
|
2555 linecharno += dbp - lasthit;
|
|
2556 lasthit = dbp;
|
|
2557 i = TEX_Token (lasthit);
|
|
2558 if (0 <= i)
|
|
2559 {
|
|
2560 TEX_getit (lasthit, TEX_toktab[i].len);
|
|
2561 break; /* We only save a line once */
|
|
2562 }
|
|
2563 }
|
|
2564 }
|
|
2565 }
|
|
2566 }
|
|
2567
|
|
2568 #define TEX_LESC '\\'
|
|
2569 #define TEX_SESC '!'
|
|
2570 #define TEX_cmt '%'
|
|
2571
|
|
2572 /* Figure out whether TeX's escapechar is '\\' or '!' and set grouping */
|
|
2573 /* chars accordingly. */
|
|
2574
|
|
2575 void
|
|
2576 TEX_mode (f)
|
|
2577 FILE *f;
|
|
2578 {
|
|
2579 int c;
|
|
2580
|
|
2581 while ((c = getc (f)) != EOF)
|
|
2582 {
|
|
2583 /* Skip to next line if we hit the TeX comment char. */
|
|
2584 if (c == TEX_cmt)
|
|
2585 while (c != '\n')
|
|
2586 c = getc (f);
|
|
2587 else if (c == TEX_LESC || c == TEX_SESC )
|
|
2588 break;
|
|
2589 }
|
|
2590
|
|
2591 if (c == TEX_LESC)
|
|
2592 {
|
|
2593 TEX_esc = TEX_LESC;
|
|
2594 TEX_opgrp = '{';
|
|
2595 TEX_clgrp = '}';
|
|
2596 }
|
|
2597 else
|
|
2598 {
|
|
2599 TEX_esc = TEX_SESC;
|
|
2600 TEX_opgrp = '<';
|
|
2601 TEX_clgrp = '>';
|
|
2602 }
|
|
2603 rewind (f);
|
|
2604 }
|
|
2605
|
|
2606 /* Read environment and prepend it to the default string. */
|
|
2607 /* Build token table. */
|
|
2608
|
|
2609 struct TEX_tabent *
|
|
2610 TEX_decode_env (evarname, defenv)
|
|
2611 char *evarname;
|
|
2612 char *defenv;
|
|
2613 {
|
|
2614 register char *env, *p;
|
|
2615 extern char *savenstr (), *index ();
|
|
2616
|
|
2617 struct TEX_tabent *tab;
|
|
2618 int size, i;
|
|
2619
|
|
2620 /* Append default string to environment. */
|
|
2621 env = getenv (evarname);
|
|
2622 if (!env)
|
|
2623 env = defenv;
|
|
2624 else
|
|
2625 env = concat (env, defenv, "");
|
|
2626
|
|
2627 /* Allocate a token table */
|
|
2628 for (size = 1, p = env; p;)
|
|
2629 if ((p = index (p, ':')) && *(++p))
|
|
2630 size++;
|
|
2631 tab = xnew (size, struct TEX_tabent);
|
|
2632
|
|
2633 /* Unpack environment string into token table. Be careful about */
|
|
2634 /* zero-length strings (leading ':', "::" and trailing ':') */
|
|
2635 for (i = 0; *env;)
|
|
2636 {
|
|
2637 p = index (env, ':');
|
|
2638 if (!p) /* End of environment string. */
|
|
2639 p = env + strlen (env);
|
|
2640 if (p - env > 0)
|
|
2641 { /* Only non-zero strings. */
|
|
2642 tab[i].name = savenstr (env, p - env);
|
|
2643 tab[i].len = strlen (tab[i].name);
|
|
2644 i++;
|
|
2645 }
|
|
2646 if (*p)
|
|
2647 env = p + 1;
|
|
2648 else
|
|
2649 {
|
|
2650 tab[i].name = NULL; /* Mark end of table. */
|
|
2651 tab[i].len = 0;
|
|
2652 break;
|
|
2653 }
|
|
2654 }
|
|
2655 return tab;
|
|
2656 }
|
|
2657
|
|
2658 /* Record a tag defined by a TeX command of length LEN and starting at NAME.
|
|
2659 The name being defined actually starts at (NAME + LEN + 1).
|
|
2660 But we seem to include the TeX command in the tag name. */
|
|
2661
|
|
2662 void
|
|
2663 TEX_getit (name, len)
|
|
2664 char *name;
|
|
2665 int len;
|
|
2666 {
|
|
2667 char *p = name + len;
|
|
2668 char nambuf[BUFSIZ];
|
|
2669
|
|
2670 if (*name == 0)
|
|
2671 return;
|
|
2672
|
|
2673 /* Let tag name extend to next group close (or end of line) */
|
|
2674 while (*p && *p != TEX_clgrp)
|
|
2675 p++;
|
|
2676 (void) strncpy (nambuf, name, p - name);
|
|
2677 nambuf[p - name] = 0;
|
|
2678
|
|
2679 pfnote (nambuf, TRUE, FALSE, lb.buffer, strlen (lb.buffer), lineno, linecharno);
|
|
2680 pfcnt++;
|
|
2681 }
|
|
2682
|
|
2683 /* If the text at CP matches one of the tag-defining TeX command names,
|
|
2684 return the index of that command in TEX_toktab.
|
|
2685 Otherwise return -1. */
|
|
2686
|
|
2687 /* Keep the capital `T' in `Token' for dumb truncating compilers
|
|
2688 (this distinguishes it from `TEX_toktab' */
|
|
2689 int
|
|
2690 TEX_Token (cp)
|
|
2691 char *cp;
|
|
2692 {
|
|
2693 int i;
|
|
2694
|
|
2695 for (i = 0; TEX_toktab[i].len > 0; i++)
|
|
2696 if (strncmp (TEX_toktab[i].name, cp, TEX_toktab[i].len) == 0)
|
|
2697 return i;
|
|
2698 return -1;
|
|
2699 }
|
|
2700
|
|
2701 /* Support for Prolog. */
|
|
2702
|
|
2703 /* whole head (not only functor, but also arguments)
|
|
2704 is gotten in compound term. */
|
|
2705
|
|
2706 void
|
|
2707 prolog_getit (s, lineno, linecharno)
|
|
2708 char *s;
|
|
2709 int lineno;
|
|
2710 long linecharno;
|
|
2711 {
|
|
2712 char nambuf[BUFSIZ], *save_s, tmpc;
|
|
2713 int insquote, npar;
|
|
2714
|
|
2715 save_s = s;
|
|
2716 insquote = FALSE;
|
|
2717 npar = 0;
|
|
2718 while (1)
|
|
2719 {
|
|
2720 if (*s == '\0') /* syntax error. */
|
|
2721 return;
|
|
2722 else if (insquote && *s == '\'' && *(s + 1) == '\'')
|
|
2723 s += 2;
|
|
2724 else if (*s == '\'')
|
|
2725 {
|
|
2726 insquote = !insquote;
|
|
2727 s++;
|
|
2728 }
|
|
2729 else if (!insquote && *s == '(')
|
|
2730 {
|
|
2731 npar++;
|
|
2732 s++;
|
|
2733 }
|
|
2734 else if (!insquote && *s == ')')
|
|
2735 {
|
|
2736 npar--;
|
|
2737 s++;
|
|
2738 if (npar == 0)
|
|
2739 break;
|
|
2740 else if (npar < 0) /* syntax error. */
|
|
2741 return;
|
|
2742 }
|
|
2743 else if (!insquote && *s == '.' && (isspace (*(s + 1)) || *(s + 1) == '\0'))
|
|
2744 { /* fullstop. */
|
|
2745 if (npar != 0) /* syntax error. */
|
|
2746 return;
|
|
2747 s++;
|
|
2748 break;
|
|
2749 }
|
|
2750 else
|
|
2751 s++;
|
|
2752 }
|
|
2753 tmpc = *s;
|
|
2754 *s = '\0';
|
|
2755 strcpy (nambuf, save_s);
|
|
2756 *s = tmpc;
|
|
2757 pfnote (nambuf, TRUE, save_s, strlen (nambuf), lineno, linecharno);
|
|
2758 }
|
|
2759
|
|
2760 /* It is assumed that prolog predicate starts from column 0. */
|
|
2761
|
|
2762 void
|
|
2763 prolog_funcs (fi)
|
|
2764 FILE *fi;
|
|
2765 {
|
|
2766 void skip_comment (), prolog_getit ();
|
|
2767
|
|
2768 lineno = linecharno = charno = 0;
|
|
2769 while (!feof (fi))
|
|
2770 {
|
|
2771 lineno++;
|
|
2772 linecharno += charno;
|
|
2773 charno = readline (&lb, fi) + 1; /* 1 for newline. */
|
|
2774 dbp = lb.buffer;
|
|
2775 if (isspace (dbp[0])) /* not predicate header. */
|
|
2776 continue;
|
|
2777 else if (dbp[0] == '%') /* comment. */
|
|
2778 continue;
|
|
2779 else if (dbp[0] == '/' && dbp[1] == '*') /* comment. */
|
|
2780 skip_comment (&lb, fi, &lineno, &linecharno);
|
|
2781 else /* found. */
|
|
2782 prolog_getit (dbp, lineno, linecharno);
|
|
2783 }
|
|
2784 }
|
|
2785
|
|
2786 void
|
|
2787 skip_comment (plb, fi, plineno, plinecharno)
|
|
2788 struct linebuffer *plb;
|
|
2789 FILE *fi;
|
|
2790 int *plineno; /* result */
|
|
2791 long *plinecharno; /* result */
|
|
2792 {
|
|
2793 while (!substr ("*/", plb->buffer))
|
|
2794 {
|
|
2795 (*plineno)++;
|
|
2796 *plinecharno += readline (plb, fi) + 1;
|
|
2797 } /* 1 for newline. */
|
|
2798 }
|
|
2799
|
|
2800 /* Return TRUE if 'sub' exists somewhere in 's'. */
|
|
2801
|
|
2802 int
|
|
2803 substr (sub, s)
|
|
2804 char *sub;
|
|
2805 char *s;
|
|
2806 {
|
|
2807 while (*s && (s = index (s, *sub)))
|
|
2808 if (prestr (sub, s))
|
|
2809 return (TRUE);
|
|
2810 else
|
|
2811 s++;
|
|
2812 return (FALSE);
|
|
2813 }
|
|
2814
|
|
2815 /* Return TRUE if 'pre' is prefix of string 's'. */
|
|
2816
|
|
2817 int
|
|
2818 prestr (pre, s)
|
|
2819 char *pre;
|
|
2820 char *s;
|
|
2821 {
|
|
2822 if (*pre == '\0')
|
|
2823 return (TRUE);
|
|
2824 else if (*pre == *s)
|
|
2825 return (prestr (pre + 1, s + 1));
|
|
2826 else
|
|
2827 return (FALSE);
|
|
2828 }
|
|
2829
|
|
2830 /* Initialize a linebuffer for use */
|
|
2831
|
|
2832 void
|
|
2833 initbuffer (linebuffer)
|
|
2834 struct linebuffer *linebuffer;
|
|
2835 {
|
|
2836 linebuffer->size = 200;
|
|
2837 linebuffer->buffer = xnew (200, char);
|
|
2838 }
|
|
2839
|
|
2840 /*
|
|
2841 * Read a line of text from `stream' into `linebuffer'.
|
|
2842 * Return the number of characters read from `stream',
|
|
2843 * which is the length of the line including the newline, if any.
|
|
2844 */
|
|
2845 long
|
|
2846 readline (linebuffer, stream)
|
|
2847 struct linebuffer *linebuffer;
|
|
2848 register FILE *stream;
|
|
2849 {
|
|
2850 char *buffer = linebuffer->buffer;
|
|
2851 register char *p = linebuffer->buffer;
|
|
2852 register char *pend;
|
|
2853 int newline; /* 1 if ended with newline, 0 if ended with EOF */
|
|
2854
|
|
2855 pend = p + linebuffer->size; /* Separate to avoind 386/IX compiler bug. */
|
|
2856
|
|
2857 while (1)
|
|
2858 {
|
|
2859 register int c = getc (stream);
|
|
2860 if (p == pend)
|
|
2861 {
|
|
2862 linebuffer->size *= 2;
|
|
2863 buffer = (char *) xrealloc (buffer, linebuffer->size);
|
|
2864 p += buffer - linebuffer->buffer;
|
|
2865 pend = buffer + linebuffer->size;
|
|
2866 linebuffer->buffer = buffer;
|
|
2867 }
|
|
2868 if (c < 0 || c == '\n')
|
|
2869 {
|
|
2870 *p = 0;
|
|
2871 newline = (c == '\n' ? 1 : 0);
|
|
2872 break;
|
|
2873 }
|
|
2874 *p++ = c;
|
|
2875 }
|
|
2876
|
|
2877 return p - buffer + newline;
|
|
2878 }
|
|
2879
|
|
2880 char *
|
|
2881 savestr (cp)
|
|
2882 char *cp;
|
|
2883 {
|
|
2884 return savenstr (cp, strlen (cp));
|
|
2885 }
|
|
2886
|
|
2887 char *
|
|
2888 savenstr (cp, len)
|
|
2889 char *cp;
|
|
2890 int len;
|
|
2891 {
|
|
2892 register char *dp;
|
|
2893
|
|
2894 dp = xnew (len + 1, char);
|
|
2895 (void) strncpy (dp, cp, len);
|
|
2896 dp[len] = '\0';
|
|
2897 return dp;
|
|
2898 }
|
|
2899
|
|
2900 #ifdef notdef
|
|
2901 /*
|
|
2902 * Return the ptr in sp at which the character c last
|
|
2903 * appears; NULL if not found
|
|
2904 *
|
|
2905 * Identical to v7 rindex, included for portability.
|
|
2906 */
|
|
2907
|
|
2908 char *
|
|
2909 rindex (sp, c)
|
|
2910 register char *sp, c;
|
|
2911 {
|
|
2912 register char *r;
|
|
2913
|
|
2914 r = NULL;
|
|
2915 do
|
|
2916 {
|
|
2917 if (*sp == c)
|
|
2918 r = sp;
|
|
2919 } while (*sp++);
|
|
2920 return (r);
|
|
2921 }
|
|
2922
|
|
2923 /*
|
|
2924 * Return the ptr in sp at which the character c first
|
|
2925 * appears; NULL if not found
|
|
2926 *
|
|
2927 * Identical to v7 index, included for portability.
|
|
2928 */
|
|
2929
|
|
2930 char *
|
|
2931 index (sp, c)
|
|
2932 register char *sp, c;
|
|
2933 {
|
|
2934 do
|
|
2935 {
|
|
2936 if (*sp == c)
|
|
2937 return (sp);
|
|
2938 } while (*sp++);
|
|
2939 return (NULL);
|
|
2940 }
|
|
2941
|
|
2942 #endif /* notdef */
|
|
2943
|
|
2944 /* Print error message and exit. */
|
|
2945
|
|
2946 /* VARARGS1 */
|
|
2947 void
|
|
2948 fatal (s1, s2)
|
|
2949 char *s1, *s2;
|
|
2950 {
|
|
2951 error (s1, s2);
|
|
2952 exit (1);
|
|
2953 }
|
|
2954
|
|
2955 /* Print error message. `s1' is printf control string, `s2' is arg for it. */
|
|
2956
|
|
2957 /* VARARGS1 */
|
|
2958 void
|
|
2959 error (s1, s2)
|
|
2960 char *s1, *s2;
|
|
2961 {
|
|
2962 fprintf (stderr, "%s: ", progname);
|
|
2963 fprintf (stderr, s1, s2);
|
|
2964 fprintf (stderr, "\n");
|
|
2965 }
|
|
2966
|
|
2967 /* Return a newly-allocated string whose contents concatenate those of s1, s2, s3. */
|
|
2968
|
|
2969 char *
|
|
2970 concat (s1, s2, s3)
|
|
2971 char *s1, *s2, *s3;
|
|
2972 {
|
|
2973 int len1 = strlen (s1), len2 = strlen (s2), len3 = strlen (s3);
|
|
2974 char *result = xnew (len1 + len2 + len3 + 1, char);
|
|
2975
|
|
2976 (void) strcpy (result, s1);
|
|
2977 (void) strcpy (result + len1, s2);
|
|
2978 (void) strcpy (result + len1 + len2, s3);
|
|
2979 *(result + len1 + len2 + len3) = 0;
|
|
2980
|
|
2981 return result;
|
|
2982 }
|
|
2983
|
|
2984 /* Like malloc but get fatal error if memory is exhausted. */
|
|
2985
|
|
2986 char *
|
|
2987 xmalloc (size)
|
|
2988 int size;
|
|
2989 {
|
|
2990 char *result = malloc (size);
|
|
2991 if (!result)
|
|
2992 fatal ("virtual memory exhausted", 0);
|
|
2993 return result;
|
|
2994 }
|
|
2995
|
|
2996 char *
|
|
2997 xrealloc (ptr, size)
|
|
2998 char *ptr;
|
|
2999 int size;
|
|
3000 {
|
|
3001 char *result = realloc (ptr, size);
|
|
3002 if (!result)
|
|
3003 fatal ("virtual memory exhausted");
|
|
3004 return result;
|
|
3005 }
|