Mercurial > hg > xemacs-beta
annotate lib-src/ootags.c @ 5856:27876789edc5
Be more careful about saving a non-nil value for START in message-stack.
lisp/ChangeLog addition:
2015-03-12 Aidan Kehoe <kehoea@parhasard.net>
* simple.el (append-message): Be more careful about saving a
non-nil value for START in message-stack.
author | Aidan Kehoe <kehoea@parhasard.net> |
---|---|
date | Thu, 12 Mar 2015 00:59:27 +0000 |
parents | b3824b7f5627 |
children |
rev | line source |
---|---|
428 | 1 /* Tags file maker to go with GNU Emacs |
2 Copyright (C) 1984, 87, 88, 89, 93, 94, 95 | |
3 Free Software Foundation, Inc. and Ken Arnold | |
4 | |
5 This file is not considered part of GNU Emacs. | |
6 | |
5406
061f4f90f874
Convert lib-src/ to GPLv3.
Mike Sperber <sperber@deinprogramm.de>
parents:
4932
diff
changeset
|
7 This program is free software: you can redistribute it and/or modify it |
061f4f90f874
Convert lib-src/ to GPLv3.
Mike Sperber <sperber@deinprogramm.de>
parents:
4932
diff
changeset
|
8 under the terms of the GNU General Public License as published by the |
061f4f90f874
Convert lib-src/ to GPLv3.
Mike Sperber <sperber@deinprogramm.de>
parents:
4932
diff
changeset
|
9 Free Software Foundation, either version 3 of the License, or (at your |
061f4f90f874
Convert lib-src/ to GPLv3.
Mike Sperber <sperber@deinprogramm.de>
parents:
4932
diff
changeset
|
10 option) any later version. |
061f4f90f874
Convert lib-src/ to GPLv3.
Mike Sperber <sperber@deinprogramm.de>
parents:
4932
diff
changeset
|
11 |
061f4f90f874
Convert lib-src/ to GPLv3.
Mike Sperber <sperber@deinprogramm.de>
parents:
4932
diff
changeset
|
12 This program is distributed in the hope that it will be useful, but WITHOUT |
061f4f90f874
Convert lib-src/ to GPLv3.
Mike Sperber <sperber@deinprogramm.de>
parents:
4932
diff
changeset
|
13 ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or |
061f4f90f874
Convert lib-src/ to GPLv3.
Mike Sperber <sperber@deinprogramm.de>
parents:
4932
diff
changeset
|
14 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License |
061f4f90f874
Convert lib-src/ to GPLv3.
Mike Sperber <sperber@deinprogramm.de>
parents:
4932
diff
changeset
|
15 for more details. |
428 | 16 |
17 You should have received a copy of the GNU General Public License | |
5406
061f4f90f874
Convert lib-src/ to GPLv3.
Mike Sperber <sperber@deinprogramm.de>
parents:
4932
diff
changeset
|
18 along with this program. If not, see <http://www.gnu.org/licenses/>. */ |
428 | 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 * Francesco Potorti` reorganised C and C++ based on work by Joe Wells. | |
28 * Regexp tags by Tom Tromey. | |
29 * | |
30 * Francesco Potorti` (F.Potorti@cnuce.cnr.it) is the current maintainer. | |
31 */ | |
32 | |
33 char pot_etags_version[] = "@(#) pot revision number is 12.28"; | |
34 | |
35 /* Prototyping magic snarfed from gmalloc.c */ | |
36 #if defined (__cplusplus) || defined (__STDC__) | |
37 #undef PP | |
38 #define PP(args) args | |
39 #undef __ptr_t | |
40 #define __ptr_t void * | |
41 #else /* Not C++ or ANSI C. */ | |
42 #undef PP | |
43 #define PP(args) () | |
44 #undef const | |
45 #define const | |
46 #undef __ptr_t | |
47 #define __ptr_t char * | |
48 #endif /* C++ or ANSI C. */ | |
49 | |
50 #ifdef HAVE_CONFIG_H | |
51 # include <config.h> | |
52 /* On some systems, Emacs defines static as nothing for the sake | |
53 of unexec. We don't want that here since we don't use unexec. */ | |
54 # undef static | |
55 # define ETAGS_REGEXPS /* use the regexp features */ | |
56 # define LONG_OPTIONS /* accept long options */ | |
57 #endif /* HAVE_CONFIG_H */ | |
58 | |
59 #define TRUE 1 | |
60 #define FALSE 0 | |
61 | |
62 #ifndef DEBUG | |
63 # define DEBUG FALSE | |
64 #endif | |
65 | |
442 | 66 #ifdef WIN32_NATIVE |
428 | 67 # include <stdlib.h> |
68 # include <fcntl.h> | |
69 # include <string.h> | |
70 # include <io.h> | |
71 # define MAXPATHLEN _MAX_PATH | |
442 | 72 # ifndef HAVE_CONFIG_H |
428 | 73 # define HAVE_GETCWD |
74 # endif /* not HAVE_CONFIG_H */ | |
442 | 75 #endif /* WIN32_NATIVE */ |
76 | |
77 #if !defined (WIN32_NATIVE) && defined (STDC_HEADERS) | |
428 | 78 #include <stdlib.h> |
79 #include <string.h> | |
80 #endif | |
81 | |
82 #ifdef HAVE_UNISTD_H | |
83 # include <unistd.h> | |
84 #else | |
85 # ifdef HAVE_GETCWD | |
86 extern char *getcwd (); | |
87 # endif | |
88 #endif /* HAVE_UNISTD_H */ | |
89 | |
90 #include <stdio.h> | |
91 #include <ctype.h> | |
92 #include <errno.h> | |
93 #include <sys/types.h> | |
94 #include <sys/stat.h> | |
95 | |
96 #if !defined (S_ISREG) && defined (S_IFREG) | |
97 # define S_ISREG(m) (((m) & S_IFMT) == S_IFREG) | |
98 #endif | |
99 | |
100 #ifdef LONG_OPTIONS | |
101 # include <getopt.h> | |
102 #else | |
103 # define getopt_long(argc,argv,optstr,lopts,lind) getopt (argc, argv, optstr) | |
104 extern char *optarg; | |
105 extern int optind, opterr; | |
106 #endif /* LONG_OPTIONS */ | |
107 | |
108 #ifdef ETAGS_REGEXPS | |
109 # include <regex.h> | |
110 #endif /* ETAGS_REGEXPS */ | |
111 | |
112 /* Define CTAGS to make the program "ctags" compatible with the usual one. | |
113 Leave it undefined to make the program "etags", which makes emacs-style | |
114 tag tables and tags typedefs, #defines and struct/union/enum by default. */ | |
115 #ifdef CTAGS | |
116 # undef CTAGS | |
117 # define CTAGS TRUE | |
118 #else | |
119 # define CTAGS FALSE | |
120 #endif | |
121 | |
122 /* Exit codes for success and failure. */ | |
123 #ifdef VMS | |
124 # define GOOD 1 | |
125 # define BAD 0 | |
126 #else | |
127 # define GOOD 0 | |
128 # define BAD 1 | |
129 #endif | |
130 | |
131 /* C extensions. */ | |
132 #define C_PLPL 0x00001 /* C++ */ | |
133 #define C_STAR 0x00003 /* C* */ | |
134 #define C_JAVA 0x00005 /* JAVA */ | |
135 #define YACC 0x10000 /* yacc file */ | |
136 | |
137 #define streq(s,t) ((DEBUG && (s) == NULL && (t) == NULL \ | |
138 && (abort (), 1)) || !strcmp (s, t)) | |
139 #define strneq(s,t,n) ((DEBUG && (s) == NULL && (t) == NULL \ | |
140 && (abort (), 1)) || !strncmp (s, t, n)) | |
141 | |
142 #define lowcase(c) tolower ((char)c) | |
143 | |
144 #define CHARS 256 /* 2^sizeof(char) */ | |
145 #define CHAR(x) ((unsigned int)x & (CHARS - 1)) | |
146 #define iswhite(c) (_wht[CHAR(c)]) /* c is white */ | |
147 #define notinname(c) (_nin[CHAR(c)]) /* c is not in a name */ | |
148 #define begtoken(c) (_btk[CHAR(c)]) /* c can start token */ | |
149 #define intoken(c) (_itk[CHAR(c)]) /* c can be in token */ | |
150 #define endtoken(c) (_etk[CHAR(c)]) /* c ends tokens */ | |
151 | |
152 /*#ifdef INFODOCK*/ | |
153 /*#undef OO_BROWSER*/ | |
154 /* Due to the way this file is constructed, this unfortunately doesn't */ | |
155 /* work except for documentation purposes. -slb */ | |
156 #define OO_BROWSER 1 | |
157 /*#endif*/ | |
158 | |
159 #ifdef OO_BROWSER | |
160 #define set_construct(construct) \ | |
161 if (!oo_browser_construct) oo_browser_construct = construct | |
438 | 162 void oo_browser_clear_all_globals(void); |
163 void oo_browser_clear_some_globals(void); | |
164 void oo_browser_check_and_clear_structtype(void); | |
428 | 165 #endif |
166 | |
167 /* | |
168 * xnew, xrnew -- allocate, reallocate storage | |
169 * | |
170 * SYNOPSIS: Type *xnew (int n, Type); | |
171 * Type *xrnew (OldPointer, int n, Type); | |
172 */ | |
173 #ifdef chkmalloc | |
174 # include "chkmalloc.h" | |
175 # define xnew(n,Type) ((Type *) trace_malloc (__FILE__, __LINE__, \ | |
176 (n) * sizeof (Type))) | |
177 # define xrnew(op,n,Type) ((Type *) trace_realloc (__FILE__, __LINE__, \ | |
178 (op), (n) * sizeof (Type))) | |
179 #else | |
180 # define xnew(n,Type) ((Type *) xmalloc ((n) * sizeof (Type))) | |
181 # define xrnew(op,n,Type) ((Type *) xrealloc ((op), (n) * sizeof (Type))) | |
182 #endif | |
183 | |
184 typedef int bool; | |
185 | |
438 | 186 typedef void Lang_function (FILE *); |
428 | 187 |
188 typedef struct | |
189 { | |
190 char *name; | |
191 Lang_function *function; | |
192 char **suffixes; | |
193 char **interpreters; | |
194 } language; | |
195 | |
196 typedef struct node_st | |
197 { /* sorting structure */ | |
198 char *name; /* function or type name */ | |
199 #ifdef OO_BROWSER | |
200 short int construct; /* Construct type for the OO-Browser */ | |
201 #endif | |
202 char *file; /* file name */ | |
203 bool is_func; /* use pattern or line no */ | |
204 bool been_warned; /* set if noticed dup */ | |
205 int lno; /* line number tag is on */ | |
206 long cno; /* character number line starts on */ | |
207 char *pat; /* search pattern */ | |
208 struct node_st *left, *right; /* left and right sons */ | |
209 } node; | |
210 | |
211 #ifdef OO_BROWSER | |
212 /* If you add to this array, you must add a corresponding entry to the | |
213 following enum. */ | |
214 static char *oo_browser_default_classes[] = | |
215 /* Lack of square brackets around some of these entries are intentional. */ | |
216 {"null", "class", "method", "[constant]", "[enumeration]", "[enum_label]", | |
217 "extern", "[function]", "[macro]", "objc", "[structure]", "[type]", | |
218 "[union]", "[variable]"}; | |
219 | |
220 /* If you add to this enum, you must add a corresponding entry to the | |
221 preceding array. */ | |
222 enum oo_browser_constructs {C_NULL, C_CLASS, C_METHOD, C_CONSTANT, C_ENUMERATION, | |
223 C_ENUM_LABEL, C_EXTERN, C_FUNCTION, C_MACRO, | |
224 C_OBJC, C_STRUCTURE, C_TYPE, C_UNION, C_VARIABLE}; | |
225 | |
226 enum oo_browser_constructs oo_browser_construct = C_NULL; | |
227 #endif | |
228 | |
229 /* | |
230 * A `linebuffer' is a structure which holds a line of text. | |
231 * `readline_internal' reads a line from a stream into a linebuffer | |
232 * and works regardless of the length of the line. | |
233 * SIZE is the size of BUFFER, LEN is the length of the string in | |
234 * BUFFER after readline reads it. | |
235 */ | |
236 typedef struct | |
237 { | |
238 long size; | |
239 int len; | |
240 char *buffer; | |
241 } linebuffer; | |
242 | |
243 extern char *getenv PP ((const char *envvar)); | |
244 | |
245 /* Many compilers barf on this: | |
246 Lang_function Asm_labels; | |
247 so let's write it this way */ | |
248 void Asm_labels PP ((FILE *inf)); | |
249 void C_entries PP ((int c_ext, FILE *inf)); | |
250 void default_C_entries PP ((FILE *inf)); | |
251 void plain_C_entries PP ((FILE *inf)); | |
252 void Cjava_entries PP ((FILE *inf)); | |
253 void Cplusplus_entries PP ((FILE *inf)); | |
254 void Yacc_entries PP ((FILE *inf)); | |
255 void Cobol_paragraphs PP ((FILE *inf)); | |
256 void Cstar_entries PP ((FILE *inf)); | |
257 void Erlang_functions PP ((FILE *inf)); | |
258 void Fortran_functions PP ((FILE *inf)); | |
259 void Lisp_functions PP ((FILE *inf)); | |
260 void Pascal_functions PP ((FILE *inf)); | |
261 void Perl_functions PP ((FILE *inf)); | |
262 void Postscript_functions PP ((FILE *inf)); | |
263 void Prolog_functions PP ((FILE *inf)); | |
264 void Python_functions PP ((FILE *inf)); | |
265 void Scheme_functions PP ((FILE *inf)); | |
266 void TeX_functions PP ((FILE *inf)); | |
267 void just_read_file PP ((FILE *inf)); | |
268 | |
269 void print_language_names PP ((void)); | |
270 void print_version PP ((void)); | |
271 void print_help PP ((void)); | |
272 | |
273 language *get_language_from_name PP ((char *name)); | |
274 language *get_language_from_interpreter PP ((char *interpreter)); | |
275 language *get_language_from_suffix PP ((char *suffix)); | |
276 int total_size_of_entries PP ((node *np)); | |
277 long readline PP ((linebuffer *lbp, FILE *stream)); | |
278 long readline_internal PP ((linebuffer *lbp, FILE *stream)); | |
279 #ifdef ETAGS_REGEXPS | |
280 void analyse_regex PP ((char *regex_arg)); | |
281 void add_regex PP ((char *regexp_pattern, language *lang)); | |
282 void free_patterns PP ((void)); | |
283 #endif /* ETAGS_REGEXPS */ | |
284 void error PP ((const char *s1, const char *s2)); | |
285 void suggest_asking_for_help PP ((void)); | |
286 void fatal PP ((char *s1, char *s2)); | |
287 void pfatal PP ((char *s1)); | |
288 void add_node PP ((node *np, node **cur_node_p)); | |
289 | |
290 void init PP ((void)); | |
291 void initbuffer PP ((linebuffer *lbp)); | |
292 void find_entries PP ((char *file, FILE *inf)); | |
293 void free_tree PP ((node *np)); | |
294 void pfnote PP ((char *name, bool is_func, char *linestart, int linelen, int lno, long cno)); | |
295 void new_pfnote PP ((char *name, int namelen, bool is_func, char *linestart, int linelen, int lno, long cno)); | |
296 void process_file PP ((char *file)); | |
297 void put_entries PP ((node *np)); | |
298 void takeprec PP ((void)); | |
299 | |
300 char *concat PP ((char *s1, char *s2, char *s3)); | |
301 char *skip_spaces PP ((char *cp)); | |
302 char *skip_non_spaces PP ((char *cp)); | |
303 char *savenstr PP ((char *cp, int len)); | |
304 char *savestr PP ((char *cp)); | |
305 char *etags_strchr PP ((char *sp, int c)); | |
306 char *etags_strrchr PP ((char *sp, int c)); | |
307 char *etags_getcwd PP ((void)); | |
308 char *relative_filename PP ((char *file, char *dir)); | |
309 char *absolute_filename PP ((char *file, char *dir)); | |
310 char *absolute_dirname PP ((char *file, char *dir)); | |
311 bool filename_is_absolute PP ((char *fn)); | |
312 void canonicalize_filename PP ((char *fn)); | |
313 void grow_linebuffer PP ((linebuffer *lbp, int toksize)); | |
314 long *xmalloc PP ((unsigned int size)); | |
315 long *xrealloc PP ((char *ptr, unsigned int size)); | |
316 | |
317 | |
318 char searchar = '/'; /* use /.../ searches */ | |
319 | |
320 char *tagfile; /* output file */ | |
321 char *progname; /* name this program was invoked with */ | |
322 char *cwd; /* current working directory */ | |
323 char *tagfiledir; /* directory of tagfile */ | |
324 FILE *tagf; /* ioptr for tags file */ | |
325 | |
326 char *curfile; /* current input file name */ | |
327 language *curlang; /* current language */ | |
328 | |
329 int lineno; /* line number of current line */ | |
330 long charno; /* current character number */ | |
331 long linecharno; /* charno of start of current line */ | |
332 char *dbp; /* pointer to start of current tag */ | |
333 node *head; /* the head of the binary tree of tags */ | |
334 | |
335 linebuffer lb; /* the current line */ | |
336 linebuffer token_name; /* used by C_entries as a temporary area */ | |
337 struct | |
338 { | |
339 long linepos; | |
340 linebuffer lb; /* used by C_entries instead of lb */ | |
341 } lbs[2]; | |
342 | |
343 /* boolean "functions" (see init) */ | |
344 bool _wht[CHARS], _nin[CHARS], _itk[CHARS], _btk[CHARS], _etk[CHARS]; | |
345 char | |
346 /* white chars */ | |
347 *white = " \f\t\n\r", | |
348 /* not in a name */ | |
349 *nonam = " \f\t\n\r(=,[;", | |
350 /* token ending chars */ | |
351 *endtk = " \t\n\r\"'#()[]{}=-+%*/&|^~!<>;,.:?", | |
352 /* token starting chars */ | |
353 *begtk = "ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz$~@", | |
354 /* valid in-token chars */ | |
355 *midtk = "ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz$0123456789"; | |
356 | |
357 bool append_to_tagfile; /* -a: append to tags */ | |
358 /* The following four default to TRUE for etags, but to FALSE for ctags. */ | |
359 bool typedefs; /* -t: create tags for C typedefs */ | |
360 bool typedefs_and_cplusplus; /* -T: create tags for C typedefs, level */ | |
361 /* 0 struct/enum/union decls, and C++ */ | |
362 /* member functions. */ | |
363 bool constantypedefs; /* -d: create tags for C #define, enum */ | |
364 /* constants and variables. */ | |
365 /* -D: opposite of -d. Default under ctags. */ | |
366 bool globals; /* create tags for global variables */ | |
367 bool members; /* create tags for C member variables */ | |
368 bool update; /* -u: update tags */ | |
369 bool vgrind_style; /* -v: create vgrind style index output */ | |
370 bool no_warnings; /* -w: suppress warnings */ | |
371 bool cxref_style; /* -x: create cxref style output */ | |
372 bool cplusplus; /* .[hc] means C++, not C */ | |
373 bool noindentypedefs; /* -I: ignore indentation in C */ | |
374 #ifdef OO_BROWSER | |
375 bool oo_browser_format; /* -O: OO-Browser tags format */ | |
376 #endif | |
377 | |
378 #ifdef LONG_OPTIONS | |
379 struct option longopts[] = | |
380 { | |
381 { "append", no_argument, NULL, 'a' }, | |
382 { "backward-search", no_argument, NULL, 'B' }, | |
383 { "c++", no_argument, NULL, 'C' }, | |
384 { "cxref", no_argument, NULL, 'x' }, | |
385 { "defines", no_argument, NULL, 'd' }, | |
386 { "no-defines", no_argument, NULL, 'D' }, | |
387 { "globals", no_argument, &globals, TRUE }, | |
388 { "no-globals", no_argument, &globals, FALSE }, | |
389 { "help", no_argument, NULL, 'h' }, | |
390 { "help", no_argument, NULL, 'H' }, | |
391 { "ignore-indentation", no_argument, NULL, 'I' }, | |
392 { "include", required_argument, NULL, 'i' }, | |
393 { "language", required_argument, NULL, 'l' }, | |
394 { "members", no_argument, &members, TRUE }, | |
395 { "no-members", no_argument, &members, FALSE }, | |
396 { "no-warn", no_argument, NULL, 'w' }, | |
397 { "output", required_argument, NULL, 'o' }, | |
398 #ifdef OO_BROWSER | |
399 { "oo-browser", no_argument, NULL, 'O' }, | |
400 #endif | |
401 #ifdef ETAGS_REGEXPS | |
402 { "regex", required_argument, NULL, 'r' }, | |
403 { "no-regex", no_argument, NULL, 'R' }, | |
404 #endif /* ETAGS_REGEXPS */ | |
405 { "typedefs", no_argument, NULL, 't' }, | |
406 { "typedefs-and-c++", no_argument, NULL, 'T' }, | |
407 { "update", no_argument, NULL, 'u' }, | |
408 { "version", no_argument, NULL, 'V' }, | |
409 { "vgrind", no_argument, NULL, 'v' }, | |
410 { 0 } | |
411 }; | |
412 #endif /* LONG_OPTIONS */ | |
413 | |
414 #ifdef ETAGS_REGEXPS | |
415 /* Structure defining a regular expression. Elements are | |
416 the compiled pattern, and the name string. */ | |
417 typedef struct pattern | |
418 { | |
419 struct pattern *p_next; | |
420 language *language; | |
421 char *regex; | |
422 struct re_pattern_buffer *pattern; | |
423 struct re_registers regs; | |
424 char *name_pattern; | |
425 bool error_signaled; | |
426 } pattern; | |
427 | |
428 /* Array of all regexps. */ | |
429 pattern *p_head = NULL; | |
430 #endif /* ETAGS_REGEXPS */ | |
431 | |
432 /* | |
433 * Language stuff. | |
434 */ | |
435 | |
436 /* Non-NULL if language fixed. */ | |
437 language *forced_lang = NULL; | |
438 | |
439 /* Assembly code */ | |
440 char *Asm_suffixes [] = { "a", /* Unix assembler */ | |
441 "asm", /* Microcontroller assembly */ | |
442 "def", /* BSO/Tasking definition includes */ | |
443 "inc", /* Microcontroller include files */ | |
444 "ins", /* Microcontroller include files */ | |
445 "s", "sa", /* Unix assembler */ | |
446 "src", /* BSO/Tasking C compiler output */ | |
447 NULL | |
448 }; | |
449 | |
450 /* Note that .c and .h can be considered C++, if the --c++ flag was | |
451 given. That is why default_C_entries is called here. */ | |
452 char *default_C_suffixes [] = | |
453 { "c", "h", NULL }; | |
454 | |
455 char *Cplusplus_suffixes [] = | |
456 { "C", "H", "c++", "cc", "cpp", "cxx", "h++", "hh", "hpp", "hxx", | |
457 "M", /* Objective C++ */ | |
458 "pdb", /* Postscript with C syntax */ | |
459 NULL }; | |
460 | |
461 char *Cjava_suffixes [] = | |
462 { "java", NULL }; | |
463 | |
464 char *Cobol_suffixes [] = | |
465 { "COB", "cob", NULL }; | |
466 | |
467 char *Cstar_suffixes [] = | |
468 { "cs", "hs", NULL }; | |
469 | |
470 char *Erlang_suffixes [] = | |
471 { "erl", "hrl", NULL }; | |
472 | |
473 char *Fortran_suffixes [] = | |
474 { "F", "f", "f90", "for", NULL }; | |
475 | |
476 char *Lisp_suffixes [] = | |
477 { "cl", "clisp", "el", "l", "lisp", "lsp", "ml", NULL }; | |
478 | |
479 char *Pascal_suffixes [] = | |
480 { "p", "pas", NULL }; | |
481 | |
482 char *Perl_suffixes [] = | |
483 { "pl", "pm", NULL }; | |
484 char *Perl_interpreters [] = | |
485 { "perl", "@PERL@", NULL }; | |
486 | |
487 char *plain_C_suffixes [] = | |
488 { "pc", /* Pro*C file */ | |
489 "m", /* Objective C file */ | |
490 "lm", /* Objective lex file */ | |
491 NULL }; | |
492 | |
493 char *Postscript_suffixes [] = | |
494 { "ps", NULL }; | |
495 | |
496 char *Prolog_suffixes [] = | |
497 { "prolog", NULL }; | |
498 | |
499 char *Python_suffixes [] = | |
500 { "py", NULL }; | |
501 | |
502 /* Can't do the `SCM' or `scm' prefix with a version number. */ | |
503 char *Scheme_suffixes [] = | |
504 { "SCM", "SM", "oak", "sch", "scheme", "scm", "sm", "ss", "t", NULL }; | |
505 | |
506 char *TeX_suffixes [] = | |
507 { "TeX", "bib", "clo", "cls", "ltx", "sty", "tex", NULL }; | |
508 | |
509 char *Yacc_suffixes [] = | |
510 { "y", "ym", NULL }; /* .ym is Objective yacc file */ | |
511 | |
512 /* | |
513 * Table of languages. | |
514 * | |
515 * It is ok for a given function to be listed under more than one | |
516 * name. I just didn't. | |
517 */ | |
518 | |
519 language lang_names [] = | |
520 { | |
521 { "asm", Asm_labels, Asm_suffixes, NULL }, | |
522 { "c", default_C_entries, default_C_suffixes, NULL }, | |
523 { "c++", Cplusplus_entries, Cplusplus_suffixes, NULL }, | |
524 { "c*", Cstar_entries, Cstar_suffixes, NULL }, | |
525 { "cobol", Cobol_paragraphs, Cobol_suffixes, NULL }, | |
526 { "erlang", Erlang_functions, Erlang_suffixes, NULL }, | |
527 { "fortran", Fortran_functions, Fortran_suffixes, NULL }, | |
528 { "java", Cjava_entries, Cjava_suffixes, NULL }, | |
529 { "lisp", Lisp_functions, Lisp_suffixes, NULL }, | |
530 { "pascal", Pascal_functions, Pascal_suffixes, NULL }, | |
531 { "perl", Perl_functions, Perl_suffixes, Perl_interpreters }, | |
532 { "postscript", Postscript_functions, Postscript_suffixes, NULL }, | |
533 { "proc", plain_C_entries, plain_C_suffixes, NULL }, | |
534 { "prolog", Prolog_functions, Prolog_suffixes, NULL }, | |
535 { "python", Python_functions, Python_suffixes, NULL }, | |
536 { "scheme", Scheme_functions, Scheme_suffixes, NULL }, | |
537 { "tex", TeX_functions, TeX_suffixes, NULL }, | |
538 { "yacc", Yacc_entries, Yacc_suffixes, NULL }, | |
539 { "auto", NULL }, /* default guessing scheme */ | |
540 { "none", just_read_file }, /* regexp matching only */ | |
541 { NULL, NULL } /* end of list */ | |
542 }; | |
543 | |
544 | |
545 void | |
546 print_language_names () | |
547 { | |
548 language *lang; | |
549 char **ext; | |
550 | |
551 puts ("\nThese are the currently supported languages, along with the\n\ | |
552 default file name suffixes:"); | |
553 for (lang = lang_names; lang->name != NULL; lang++) | |
554 { | |
555 printf ("\t%s\t", lang->name); | |
556 if (lang->suffixes != NULL) | |
557 for (ext = lang->suffixes; *ext != NULL; ext++) | |
558 printf (" .%s", *ext); | |
559 puts (""); | |
560 } | |
561 puts ("Where `auto' means use default language for files based on file\n\ | |
562 name suffix, and `none' means only do regexp processing on files.\n\ | |
563 If no language is specified and no matching suffix is found,\n\ | |
564 the first line of the file is read for a sharp-bang (#!) sequence\n\ | |
565 followed by the name of an interpreter. If no such sequence is found,\n\ | |
566 Fortran is tried first; if no tags are found, C is tried next."); | |
567 } | |
568 | |
569 #ifndef VERSION | |
570 # define VERSION "20" | |
571 #endif | |
572 void | |
573 print_version () | |
574 { | |
575 printf ("%s (GNU Emacs %s)\n", (CTAGS) ? "ctags" : "etags", VERSION); | |
576 puts ("Copyright (C) 1996 Free Software Foundation, Inc. and Ken Arnold"); | |
577 puts ("This program is distributed under the same terms as Emacs"); | |
578 | |
579 exit (GOOD); | |
580 } | |
581 | |
582 void | |
583 print_help () | |
584 { | |
585 printf ("Usage: %s [options] [[regex-option ...] file-name] ...\n\ | |
586 \n\ | |
587 These are the options accepted by %s.\n", progname, progname); | |
588 #ifdef LONG_OPTIONS | |
589 puts ("You may use unambiguous abbreviations for the long option names."); | |
590 #else | |
591 puts ("Long option names do not work with this executable, as it is not\n\ | |
592 linked with GNU getopt."); | |
593 #endif /* LONG_OPTIONS */ | |
594 puts ("A - as file name means read names from stdin (one per line)."); | |
595 if (!CTAGS) | |
596 printf (" Absolute names are stored in the output file as they are.\n\ | |
597 Relative ones are stored relative to the output file's directory."); | |
598 puts ("\n"); | |
599 | |
600 puts ("-a, --append\n\ | |
601 Append tag entries to existing tags file."); | |
602 | |
603 if (CTAGS) | |
604 puts ("-B, --backward-search\n\ | |
605 Write the search commands for the tag entries using '?', the\n\ | |
606 backward-search command instead of '/', the forward-search command."); | |
607 | |
608 puts ("-C, --c++\n\ | |
609 Treat files whose name suffix defaults to C language as C++ files."); | |
610 | |
611 if (CTAGS) | |
612 puts ("-d, --defines\n\ | |
613 Create tag entries for C #define constants and enum constants, too."); | |
614 else | |
615 puts ("-D, --no-defines\n\ | |
616 Don't create tag entries for C #define constants and enum constants.\n\ | |
617 This makes the tags file smaller."); | |
618 | |
619 if (!CTAGS) | |
620 { | |
621 puts ("-i FILE, --include=FILE\n\ | |
622 Include a note in tag file indicating that, when searching for\n\ | |
623 a tag, one should also consult the tags file FILE after\n\ | |
624 checking the current file."); | |
625 puts ("-l LANG, --language=LANG\n\ | |
626 Force the following files to be considered as written in the\n\ | |
627 named language up to the next --language=LANG option."); | |
628 } | |
629 | |
630 if (CTAGS) | |
631 puts ("--globals\n\ | |
632 Create tag entries for global variables in some languages."); | |
633 else | |
634 puts ("--no-globals\n\ | |
635 Do not create tag entries for global variables in some\n\ | |
636 languages. This makes the tags file smaller."); | |
637 puts ("--members\n\ | |
638 Create tag entries for member variables in C and derived languages."); | |
639 | |
640 #ifdef ETAGS_REGEXPS | |
641 puts ("-r /REGEXP/, --regex=/REGEXP/ or --regex=@regexfile\n\ | |
642 Make a tag for each line matching pattern REGEXP in the\n\ | |
643 following files. regexfile is a file containing one REGEXP\n\ | |
644 per line. REGEXP is anchored (as if preceded by ^).\n\ | |
645 The form /REGEXP/NAME/ creates a named tag. For example Tcl\n\ | |
646 named tags can be created with:\n\ | |
647 --regex=/proc[ \\t]+\\([^ \\t]+\\)/\\1/."); | |
648 puts ("-R, --no-regex\n\ | |
649 Don't create tags from regexps for the following files."); | |
650 #endif /* ETAGS_REGEXPS */ | |
651 puts ("-o FILE, --output=FILE\n\ | |
652 Write the tags to FILE."); | |
653 #ifdef OO_BROWSER | |
654 puts ("-O, --oo-browser\n\ | |
655 Generate a specialized tags format used only by the Altrasoft OO-Browser."); | |
656 #endif | |
657 puts ("-I, --ignore-indentation\n\ | |
658 Don't rely on indentation quite as much as normal. Currently,\n\ | |
659 this means not to assume that a closing brace in the first\n\ | |
660 column is the final brace of a function or structure\n\ | |
661 definition in C and C++."); | |
662 | |
663 if (CTAGS) | |
664 { | |
665 puts ("-t, --typedefs\n\ | |
666 Generate tag entries for C typedefs."); | |
667 puts ("-T, --typedefs-and-c++\n\ | |
668 Generate tag entries for C typedefs, C struct/enum/union tags,\n\ | |
669 and C++ member functions."); | |
670 puts ("-u, --update\n\ | |
671 Update the tag entries for the given files, leaving tag\n\ | |
672 entries for other files in place. Currently, this is\n\ | |
673 implemented by deleting the existing entries for the given\n\ | |
674 files and then rewriting the new entries at the end of the\n\ | |
675 tags file. It is often faster to simply rebuild the entire\n\ | |
676 tag file than to use this."); | |
677 puts ("-v, --vgrind\n\ | |
678 Generates an index of items intended for human consumption,\n\ | |
679 similar to the output of vgrind. The index is sorted, and\n\ | |
680 gives the page number of each item."); | |
681 puts ("-w, --no-warn\n\ | |
682 Suppress warning messages about entries defined in multiple\n\ | |
683 files."); | |
684 puts ("-x, --cxref\n\ | |
685 Like --vgrind, but in the style of cxref, rather than vgrind.\n\ | |
686 The output uses line numbers instead of page numbers, but\n\ | |
687 beyond that the differences are cosmetic; try both to see\n\ | |
688 which you like."); | |
689 } | |
690 | |
691 puts ("-V, --version\n\ | |
692 Print the version of the program.\n\ | |
693 -h, --help\n\ | |
694 Print this help message."); | |
695 | |
696 print_language_names (); | |
697 | |
698 puts (""); | |
699 puts ("Report bugs to bug-gnu-emacs@prep.ai.mit.edu"); | |
700 | |
701 exit (GOOD); | |
702 } | |
703 | |
704 | |
705 enum argument_type | |
706 { | |
707 at_language, | |
708 at_regexp, | |
709 at_filename | |
710 }; | |
711 | |
712 /* This structure helps us allow mixing of --lang and file names. */ | |
713 typedef struct | |
714 { | |
715 enum argument_type arg_type; | |
716 char *what; | |
717 language *lang; | |
718 } argument; | |
719 | |
720 #ifdef VMS /* VMS specific functions */ | |
721 | |
722 #define EOS '\0' | |
723 | |
724 /* This is a BUG! ANY arbitrary limit is a BUG! | |
725 Won't someone please fix this? */ | |
726 #define MAX_FILE_SPEC_LEN 255 | |
727 typedef struct { | |
728 short curlen; | |
729 char body[MAX_FILE_SPEC_LEN + 1]; | |
730 } vspec; | |
731 | |
732 /* | |
733 v1.05 nmm 26-Jun-86 fn_exp - expand specification of list of file names | |
734 returning in each successive call the next file name matching the input | |
735 spec. The function expects that each in_spec passed | |
736 to it will be processed to completion; in particular, up to and | |
737 including the call following that in which the last matching name | |
738 is returned, the function ignores the value of in_spec, and will | |
739 only start processing a new spec with the following call. | |
740 If an error occurs, on return out_spec contains the value | |
741 of in_spec when the error occurred. | |
742 | |
743 With each successive file name returned in out_spec, the | |
744 function's return value is one. When there are no more matching | |
745 names the function returns zero. If on the first call no file | |
746 matches in_spec, or there is any other error, -1 is returned. | |
747 */ | |
748 | |
749 #include <rmsdef.h> | |
750 #include <descrip.h> | |
751 #define OUTSIZE MAX_FILE_SPEC_LEN | |
752 short | |
753 fn_exp (out, in) | |
754 vspec *out; | |
755 char *in; | |
756 { | |
757 static long context = 0; | |
758 static struct dsc$descriptor_s o; | |
759 static struct dsc$descriptor_s i; | |
760 static bool pass1 = TRUE; | |
761 long status; | |
762 short retval; | |
763 | |
764 if (pass1) | |
765 { | |
766 pass1 = FALSE; | |
767 o.dsc$a_pointer = (char *) out; | |
768 o.dsc$w_length = (short)OUTSIZE; | |
769 i.dsc$a_pointer = in; | |
770 i.dsc$w_length = (short)strlen(in); | |
771 i.dsc$b_dtype = DSC$K_DTYPE_T; | |
772 i.dsc$b_class = DSC$K_CLASS_S; | |
773 o.dsc$b_dtype = DSC$K_DTYPE_VT; | |
774 o.dsc$b_class = DSC$K_CLASS_VS; | |
775 } | |
776 if ((status = lib$find_file(&i, &o, &context, 0, 0)) == RMS$_NORMAL) | |
777 { | |
778 out->body[out->curlen] = EOS; | |
779 return 1; | |
780 } | |
781 else if (status == RMS$_NMF) | |
782 retval = 0; | |
783 else | |
784 { | |
785 strcpy(out->body, in); | |
786 retval = -1; | |
787 } | |
788 lib$find_file_end(&context); | |
789 pass1 = TRUE; | |
790 return retval; | |
791 } | |
792 | |
793 /* | |
794 v1.01 nmm 19-Aug-85 gfnames - return in successive calls the | |
795 name of each file specified by the provided arg expanding wildcards. | |
796 */ | |
797 char * | |
798 gfnames (arg, p_error) | |
799 char *arg; | |
800 bool *p_error; | |
801 { | |
802 static vspec filename = {MAX_FILE_SPEC_LEN, "\0"}; | |
803 | |
804 switch (fn_exp (&filename, arg)) | |
805 { | |
806 case 1: | |
807 *p_error = FALSE; | |
808 return filename.body; | |
809 case 0: | |
810 *p_error = FALSE; | |
811 return NULL; | |
812 default: | |
813 *p_error = TRUE; | |
814 return filename.body; | |
815 } | |
816 } | |
817 | |
818 #ifndef OLD /* Newer versions of VMS do provide `system'. */ | |
819 system (cmd) | |
820 char *cmd; | |
821 { | |
822 error ("%s", "system() function not implemented under VMS"); | |
823 } | |
824 #endif | |
825 | |
826 #define VERSION_DELIM ';' | |
827 char *massage_name (s) | |
828 char *s; | |
829 { | |
830 char *start = s; | |
831 | |
832 for ( ; *s; s++) | |
833 if (*s == VERSION_DELIM) | |
834 { | |
835 *s = EOS; | |
836 break; | |
837 } | |
838 else | |
839 *s = lowcase (*s); | |
840 return start; | |
841 } | |
842 #endif /* VMS */ | |
843 | |
844 | |
845 int | |
438 | 846 main (int argc, char *argv[]) |
428 | 847 { |
848 int i; | |
849 unsigned int nincluded_files; | |
850 char **included_files; | |
851 char *this_file; | |
852 argument *argbuffer; | |
853 int current_arg, file_count; | |
854 linebuffer filename_lb; | |
855 #ifdef VMS | |
856 bool got_err; | |
857 #endif | |
858 | |
442 | 859 #ifdef WIN32_NATIVE |
428 | 860 _fmode = O_BINARY; /* all of files are treated as binary files */ |
442 | 861 #endif /* WIN32_NATIVE */ |
428 | 862 |
863 progname = argv[0]; | |
864 nincluded_files = 0; | |
865 included_files = xnew (argc, char *); | |
866 current_arg = 0; | |
867 file_count = 0; | |
868 | |
869 /* Allocate enough no matter what happens. Overkill, but each one | |
870 is small. */ | |
871 argbuffer = xnew (argc, argument); | |
872 | |
873 #ifdef ETAGS_REGEXPS | |
874 /* Set syntax for regular expression routines. */ | |
875 re_set_syntax (RE_SYNTAX_EMACS | RE_INTERVALS); | |
876 #endif /* ETAGS_REGEXPS */ | |
877 | |
878 /* | |
879 * If etags, always find typedefs and structure tags. Why not? | |
880 * Also default is to find macro constants, enum constants and | |
881 * global variables. | |
882 */ | |
883 if (!CTAGS) | |
884 { | |
885 typedefs = typedefs_and_cplusplus = constantypedefs = TRUE; | |
886 globals = TRUE; | |
887 members = FALSE; | |
888 } | |
889 | |
890 while (1) | |
891 { | |
892 int opt; | |
893 char *optstring; | |
894 | |
895 #ifdef ETAGS_REGEXPS | |
896 #ifndef OO_BROWSER | |
897 optstring = "-aCdDf:Il:o:r:RStTi:BuvxwVhH"; | |
898 #else | |
899 optstring = "-aCdDf:Il:o:r:RStTi:BOuvxwVhH"; | |
900 #endif | |
901 #else | |
902 #ifndef OO_BROWSER | |
903 optstring = "-aCdDf:Il:o:StTi:BuvxwVhH"; | |
904 #else | |
905 optstring = "-aCdDf:Il:o:StTi:BOuvxwVhH"; | |
906 #endif | |
907 #endif /* ETAGS_REGEXPS */ | |
908 | |
909 #ifndef LONG_OPTIONS | |
910 optstring = optstring + 1; | |
911 #endif /* LONG_OPTIONS */ | |
912 | |
913 opt = getopt_long (argc, argv, optstring, longopts, 0); | |
914 if (opt == EOF) | |
915 break; | |
916 | |
917 switch (opt) | |
918 { | |
919 case 0: | |
920 /* If getopt returns 0, then it has already processed a | |
921 long-named option. We should do nothing. */ | |
922 break; | |
923 | |
924 case 1: | |
925 /* This means that a file name has been seen. Record it. */ | |
926 argbuffer[current_arg].arg_type = at_filename; | |
927 argbuffer[current_arg].what = optarg; | |
928 ++current_arg; | |
929 ++file_count; | |
930 break; | |
931 | |
932 /* Common options. */ | |
933 case 'a': append_to_tagfile = TRUE; break; | |
934 case 'C': cplusplus = TRUE; break; | |
935 case 'd': constantypedefs = TRUE; break; | |
936 case 'D': constantypedefs = FALSE; break; | |
937 case 'f': /* for compatibility with old makefiles */ | |
938 case 'o': | |
939 if (tagfile) | |
940 { | |
941 /* convert char to string, to call error with */ | |
942 char buf[2]; | |
943 sprintf (buf, "%c", opt); | |
944 error ("-%s option may only be given once.", buf); | |
945 suggest_asking_for_help (); | |
946 } | |
947 tagfile = optarg; | |
948 break; | |
949 #ifdef OO_BROWSER | |
950 case 'O': | |
951 oo_browser_format = TRUE; | |
952 break; | |
953 #endif | |
954 case 'I': | |
955 case 'S': /* for backward compatibility */ | |
956 noindentypedefs = TRUE; | |
957 break; | |
958 case 'l': | |
959 { | |
960 language *lang = get_language_from_name (optarg); | |
961 if (lang != NULL) | |
962 { | |
963 argbuffer[current_arg].lang = lang; | |
964 argbuffer[current_arg].arg_type = at_language; | |
965 ++current_arg; | |
966 } | |
967 } | |
968 break; | |
969 #ifdef ETAGS_REGEXPS | |
970 case 'r': | |
971 argbuffer[current_arg].arg_type = at_regexp; | |
972 argbuffer[current_arg].what = optarg; | |
973 ++current_arg; | |
974 break; | |
975 case 'R': | |
976 argbuffer[current_arg].arg_type = at_regexp; | |
977 argbuffer[current_arg].what = NULL; | |
978 ++current_arg; | |
979 break; | |
980 #endif /* ETAGS_REGEXPS */ | |
981 case 'V': | |
982 print_version (); | |
983 break; | |
984 case 'h': | |
985 case 'H': | |
986 print_help (); | |
987 break; | |
988 case 't': | |
989 typedefs = TRUE; | |
990 break; | |
991 case 'T': | |
992 typedefs = typedefs_and_cplusplus = TRUE; | |
993 break; | |
994 #if (!CTAGS) | |
995 /* Etags options */ | |
996 case 'i': | |
997 included_files[nincluded_files++] = optarg; | |
998 break; | |
999 #else /* CTAGS */ | |
1000 /* Ctags options. */ | |
1001 case 'B': searchar = '?'; break; | |
1002 case 'u': update = TRUE; break; | |
1003 case 'v': vgrind_style = TRUE; /*FALLTHRU*/ | |
1004 case 'x': cxref_style = TRUE; break; | |
1005 case 'w': no_warnings = TRUE; break; | |
1006 #endif /* CTAGS */ | |
1007 default: | |
1008 suggest_asking_for_help (); | |
1009 } | |
1010 } | |
1011 | |
1012 for (; optind < argc; ++optind) | |
1013 { | |
1014 argbuffer[current_arg].arg_type = at_filename; | |
1015 argbuffer[current_arg].what = argv[optind]; | |
1016 ++current_arg; | |
1017 ++file_count; | |
1018 } | |
1019 | |
1020 if (nincluded_files == 0 && file_count == 0) | |
1021 { | |
1022 error ("no input files specified.", 0); | |
1023 suggest_asking_for_help (); | |
1024 } | |
1025 | |
1026 if (tagfile == NULL) | |
1027 tagfile = CTAGS ? "tags" : "TAGS"; | |
1028 cwd = etags_getcwd (); /* the current working directory */ | |
1029 if (cwd[strlen (cwd) - 1] != '/') | |
1030 { | |
1031 char *oldcwd = cwd; | |
1032 cwd = concat (oldcwd, "/", ""); | |
1033 free (oldcwd); | |
1034 } | |
1035 if (streq (tagfile, "-")) | |
1036 tagfiledir = cwd; | |
1037 else | |
1038 tagfiledir = absolute_dirname (tagfile, cwd); | |
1039 | |
1040 init (); /* set up boolean "functions" */ | |
1041 | |
1042 initbuffer (&lb); | |
1043 initbuffer (&token_name); | |
1044 initbuffer (&lbs[0].lb); | |
1045 initbuffer (&lbs[1].lb); | |
1046 initbuffer (&filename_lb); | |
1047 | |
1048 if (!CTAGS) | |
1049 { | |
1050 if (streq (tagfile, "-")) | |
1051 { | |
1052 tagf = stdout; | |
442 | 1053 #ifdef WIN32_NATIVE |
428 | 1054 /* Switch redirected `stdout' to binary mode (setting `_fmode' |
1055 doesn't take effect until after `stdout' is already open). */ | |
1056 if (!isatty (fileno (stdout))) | |
1057 setmode (fileno (stdout), O_BINARY); | |
442 | 1058 #endif /* WIN32_NATIVE */ |
428 | 1059 } |
1060 else | |
1061 tagf = fopen (tagfile, append_to_tagfile ? "a" : "w"); | |
1062 if (tagf == NULL) | |
1063 pfatal (tagfile); | |
1064 } | |
1065 | |
1066 /* | |
1067 * Loop through files finding functions. | |
1068 */ | |
1069 for (i = 0; i < current_arg; ++i) | |
1070 { | |
1071 switch (argbuffer[i].arg_type) | |
1072 { | |
1073 case at_language: | |
1074 forced_lang = argbuffer[i].lang; | |
1075 break; | |
1076 #ifdef ETAGS_REGEXPS | |
1077 case at_regexp: | |
1078 analyse_regex (argbuffer[i].what); | |
1079 break; | |
1080 #endif | |
1081 case at_filename: | |
1082 #ifdef VMS | |
1083 while ((this_file = gfnames (argbuffer[i].what, &got_err)) != NULL) | |
1084 { | |
1085 if (got_err) | |
1086 { | |
1087 error ("can't find file %s\n", this_file); | |
1088 argc--, argv++; | |
1089 } | |
1090 else | |
1091 { | |
1092 this_file = massage_name (this_file); | |
1093 } | |
1094 #else | |
1095 this_file = argbuffer[i].what; | |
1096 #endif | |
1097 #ifdef OO_BROWSER | |
1098 oo_browser_clear_all_globals(); | |
1099 #endif | |
1100 /* Input file named "-" means read file names from stdin | |
1101 (one per line) and use them. */ | |
1102 if (streq (this_file, "-")) | |
1103 while (readline_internal (&filename_lb, stdin) > 0) | |
1104 #ifdef OO_BROWSER | |
1105 { | |
1106 oo_browser_clear_some_globals(); | |
1107 #endif | |
1108 process_file (filename_lb.buffer); | |
1109 #ifdef OO_BROWSER | |
1110 } | |
1111 #endif | |
1112 else | |
1113 process_file (this_file); | |
1114 #ifdef VMS | |
1115 } | |
1116 #endif | |
1117 break; | |
1118 } | |
1119 } | |
1120 | |
1121 #ifdef ETAGS_REGEXPS | |
1122 free_patterns (); | |
1123 #endif /* ETAGS_REGEXPS */ | |
1124 | |
1125 if (!CTAGS) | |
1126 { | |
1127 while (nincluded_files-- > 0) | |
1128 fprintf (tagf, "\f\n%s,include\n", *included_files++); | |
1129 | |
1130 fclose (tagf); | |
1131 exit (GOOD); | |
1132 } | |
1133 | |
1134 /* If CTAGS, we are here. process_file did not write the tags yet, | |
1135 because we want them ordered. Let's do it now. */ | |
1136 if (cxref_style) | |
1137 { | |
1138 put_entries (head); | |
1139 exit (GOOD); | |
1140 } | |
1141 | |
1142 if (update) | |
1143 { | |
1144 char cmd[BUFSIZ]; | |
1145 for (i = 0; i < current_arg; ++i) | |
1146 { | |
1147 if (argbuffer[i].arg_type != at_filename) | |
1148 continue; | |
1149 sprintf (cmd, | |
1150 "mv %s OTAGS;fgrep -v '\t%s\t' OTAGS >%s;rm OTAGS", | |
1151 tagfile, argbuffer[i].what, tagfile); | |
1152 if (system (cmd) != GOOD) | |
1153 fatal ("failed to execute shell command", (char *)NULL); | |
1154 } | |
1155 append_to_tagfile = TRUE; | |
1156 } | |
1157 | |
1158 tagf = fopen (tagfile, append_to_tagfile ? "a" : "w"); | |
1159 if (tagf == NULL) | |
1160 pfatal (tagfile); | |
1161 put_entries (head); | |
1162 fclose (tagf); | |
1163 | |
1164 if (update) | |
1165 { | |
1166 char cmd[BUFSIZ]; | |
1167 sprintf (cmd, "sort %s -o %s", tagfile, tagfile); | |
1168 exit (system (cmd)); | |
1169 } | |
1170 return GOOD; | |
1171 } | |
1172 | |
1173 | |
1174 /* | |
1175 * Return a language given the name. | |
1176 */ | |
1177 language * | |
1178 get_language_from_name (name) | |
1179 char *name; | |
1180 { | |
1181 language *lang; | |
1182 | |
1183 if (name == NULL) | |
1184 error ("empty language name", (char *)NULL); | |
1185 else | |
1186 { | |
1187 for (lang = lang_names; lang->name != NULL; lang++) | |
1188 if (streq (name, lang->name)) | |
1189 return lang; | |
1190 error ("unknown language \"%s\"", name); | |
1191 } | |
1192 | |
1193 return NULL; | |
1194 } | |
1195 | |
1196 | |
1197 /* | |
1198 * Return a language given the interpreter name. | |
1199 */ | |
1200 language * | |
1201 get_language_from_interpreter (interpreter) | |
1202 char *interpreter; | |
1203 { | |
1204 language *lang; | |
1205 char **iname; | |
1206 | |
1207 if (interpreter == NULL) | |
1208 return NULL; | |
1209 for (lang = lang_names; lang->name != NULL; lang++) | |
1210 if (lang->interpreters != NULL) | |
1211 for (iname = lang->interpreters; *iname != NULL; iname++) | |
1212 if (streq (*iname, interpreter)) | |
1213 return lang; | |
1214 | |
1215 return NULL; | |
1216 } | |
1217 | |
1218 | |
1219 | |
1220 /* | |
1221 * Return a language given the file suffix. | |
1222 */ | |
1223 language * | |
1224 get_language_from_suffix (suffix) | |
1225 char *suffix; | |
1226 { | |
1227 language *lang; | |
1228 char **ext; | |
1229 | |
1230 if (suffix == NULL) | |
1231 return NULL; | |
1232 for (lang = lang_names; lang->name != NULL; lang++) | |
1233 if (lang->suffixes != NULL) | |
1234 for (ext = lang->suffixes; *ext != NULL; ext++) | |
1235 if (streq (*ext, suffix)) | |
1236 return lang; | |
1237 | |
1238 return NULL; | |
1239 } | |
1240 | |
1241 | |
1242 /* | |
1243 * This routine is called on each file argument. | |
1244 */ | |
1245 void | |
1246 process_file (file) | |
1247 char *file; | |
1248 { | |
1249 struct stat stat_buf; | |
1250 FILE *inf; | |
1251 | |
1252 canonicalize_filename (file); | |
1253 if (stat (file, &stat_buf) == 0 && !S_ISREG (stat_buf.st_mode)) | |
1254 { | |
1255 error ("skipping %s: it is not a regular file.", file); | |
1256 return; | |
1257 } | |
1258 if (streq (file, tagfile) && !streq (tagfile, "-")) | |
1259 { | |
1260 error ("skipping inclusion of %s in self.", file); | |
1261 return; | |
1262 } | |
1263 inf = fopen (file, "r"); | |
1264 if (inf == NULL) | |
1265 { | |
1266 perror (file); | |
1267 return; | |
1268 } | |
1269 | |
1270 find_entries (file, inf); | |
1271 | |
1272 if (!CTAGS) | |
1273 { | |
1274 char *filename; | |
1275 | |
1276 if (filename_is_absolute (file)) | |
1277 { | |
1278 /* file is an absolute file name. Canonicalise it. */ | |
1279 filename = absolute_filename (file, cwd); | |
1280 } | |
1281 else | |
1282 { | |
1283 /* file is a file name relative to cwd. Make it relative | |
1284 to the directory of the tags file. */ | |
1285 filename = relative_filename (file, tagfiledir); | |
1286 } | |
1287 #ifdef OO_BROWSER | |
1288 if (oo_browser_format) | |
1289 fprintf (tagf, "\f\n%s\n", filename); | |
1290 else | |
1291 #endif | |
1292 fprintf (tagf, "\f\n%s,%d\n", filename, total_size_of_entries (head)); | |
1293 free (filename); | |
1294 put_entries (head); | |
1295 free_tree (head); | |
1296 head = NULL; | |
1297 } | |
1298 } | |
1299 | |
1300 /* | |
1301 * This routine sets up the boolean pseudo-functions which work | |
1302 * by setting boolean flags dependent upon the corresponding character. | |
1303 * Every char which is NOT in that string is not a white char. Therefore, | |
1304 * all of the array "_wht" is set to FALSE, and then the elements | |
1305 * subscripted by the chars in "white" are set to TRUE. Thus "_wht" | |
1306 * of a char is TRUE if it is the string "white", else FALSE. | |
1307 */ | |
1308 void | |
1309 init () | |
1310 { | |
1311 register char *sp; | |
1312 register int i; | |
1313 | |
1314 for (i = 0; i < CHARS; i++) | |
1315 iswhite(i) = notinname(i) = begtoken(i) = intoken(i) = endtoken(i) = FALSE; | |
1316 for (sp = white; *sp != '\0'; sp++) iswhite (*sp) = TRUE; | |
1317 for (sp = nonam; *sp != '\0'; sp++) notinname (*sp) = TRUE; | |
1318 for (sp = begtk; *sp != '\0'; sp++) begtoken (*sp) = TRUE; | |
1319 for (sp = midtk; *sp != '\0'; sp++) intoken (*sp) = TRUE; | |
1320 for (sp = endtk; *sp != '\0'; sp++) endtoken (*sp) = TRUE; | |
1321 iswhite('\0') = iswhite('\n'); | |
1322 notinname('\0') = notinname('\n'); | |
1323 begtoken('\0') = begtoken('\n'); | |
1324 intoken('\0') = intoken('\n'); | |
1325 endtoken('\0') = endtoken('\n'); | |
1326 } | |
1327 | |
1328 /* | |
1329 * This routine opens the specified file and calls the function | |
1330 * which finds the function and type definitions. | |
1331 */ | |
1332 node *last_node = NULL; | |
1333 | |
1334 void | |
1335 find_entries (file, inf) | |
1336 char *file; | |
1337 FILE *inf; | |
1338 { | |
1339 char *cp; | |
1340 language *lang; | |
1341 node *old_last_node; | |
1342 | |
1343 curfile = savestr (file); | |
1344 | |
1345 /* If user specified a language, use it. */ | |
1346 lang = forced_lang; | |
1347 if (lang != NULL && lang->function != NULL) | |
1348 { | |
1349 curlang = lang; | |
1350 lang->function (inf); | |
1351 free (curfile); | |
1352 fclose (inf); | |
1353 return; | |
1354 } | |
1355 | |
1356 cp = etags_strrchr (file, '.'); | |
1357 if (cp != NULL) | |
1358 { | |
1359 cp += 1; | |
1360 lang = get_language_from_suffix (cp); | |
1361 if (lang != NULL && lang->function != NULL) | |
1362 { | |
1363 curlang = lang; | |
1364 lang->function (inf); | |
1365 free (curfile); | |
1366 fclose (inf); | |
1367 return; | |
1368 } | |
1369 } | |
1370 | |
1371 /* Look for sharp-bang as the first two characters. */ | |
1372 if (readline_internal (&lb, inf) > 0 | |
1373 && lb.len >= 2 | |
1374 && lb.buffer[0] == '#' | |
1375 && lb.buffer[1] == '!') | |
1376 { | |
1377 char *lp; | |
1378 | |
1379 /* Set lp to point at the first char after the last slash in the | |
1380 line or, if no slashes, at the first nonblank. Then set cp to | |
1381 the first successive blank and terminate the string. */ | |
1382 lp = etags_strrchr (lb.buffer+2, '/'); | |
1383 if (lp != NULL) | |
1384 lp += 1; | |
1385 else | |
1386 lp = skip_spaces (lb.buffer + 2); | |
1387 cp = skip_non_spaces (lp); | |
1388 *cp = '\0'; | |
1389 | |
1390 if (strlen (lp) > 0) | |
1391 { | |
1392 lang = get_language_from_interpreter (lp); | |
1393 if (lang != NULL && lang->function != NULL) | |
1394 { | |
1395 curlang = lang; | |
1396 lang->function (inf); | |
1397 fclose (inf); | |
1398 free (curfile); | |
1399 return; | |
1400 } | |
1401 } | |
1402 } | |
1403 rewind (inf); | |
1404 | |
1405 /* Try Fortran. */ | |
1406 old_last_node = last_node; | |
1407 curlang = get_language_from_name ("fortran"); | |
1408 Fortran_functions (inf); | |
1409 | |
1410 /* No Fortran entries found. Try C. */ | |
1411 if (old_last_node == last_node) | |
1412 { | |
1413 rewind (inf); | |
1414 curlang = get_language_from_name (cplusplus ? "c++" : "c"); | |
1415 default_C_entries (inf); | |
1416 } | |
1417 free (curfile); | |
1418 fclose (inf); | |
1419 return; | |
1420 } | |
1421 | |
1422 /* Record a tag. */ | |
1423 void | |
1424 pfnote (name, is_func, linestart, linelen, lno, cno) | |
1425 char *name; /* tag name, or NULL if unnamed */ | |
1426 bool is_func; /* tag is a function */ | |
1427 char *linestart; /* start of the line where tag is */ | |
1428 int linelen; /* length of the line where tag is */ | |
1429 int lno; /* line number */ | |
1430 long cno; /* character number */ | |
1431 { | |
1432 register node *np; | |
1433 | |
1434 if (CTAGS && name == NULL) | |
1435 return; | |
1436 | |
1437 np = xnew (1, node); | |
1438 | |
1439 /* If ctags mode, change name "main" to M<thisfilename>. */ | |
1440 if (CTAGS && !cxref_style && streq (name, "main")) | |
1441 { | |
1442 register char *fp = etags_strrchr (curfile, '/'); | |
1443 np->name = concat ("M", fp == 0 ? curfile : fp + 1, ""); | |
1444 fp = etags_strrchr (np->name, '.'); | |
1445 if (fp && fp[1] != '\0' && fp[2] == '\0') | |
1446 fp[0] = 0; | |
1447 } | |
1448 else | |
1449 np->name = name; | |
1450 np->been_warned = FALSE; | |
1451 np->file = curfile; | |
1452 np->is_func = is_func; | |
1453 np->lno = lno; | |
1454 /* Our char numbers are 0-base, because of C language tradition? | |
1455 ctags compatibility? old versions compatibility? I don't know. | |
1456 Anyway, since emacs's are 1-base we expect etags.el to take care | |
1457 of the difference. If we wanted to have 1-based numbers, we would | |
1458 uncomment the +1 below. */ | |
1459 np->cno = cno /* + 1 */ ; | |
1460 np->left = np->right = NULL; | |
1461 if (CTAGS && !cxref_style) | |
1462 { | |
1463 if (strlen (linestart) < 50) | |
1464 np->pat = concat (linestart, "$", ""); | |
1465 else | |
1466 np->pat = savenstr (linestart, 50); | |
1467 } | |
1468 else | |
1469 np->pat = savenstr (linestart, linelen); | |
1470 | |
1471 #ifdef OO_BROWSER | |
1472 if (oo_browser_format) | |
1473 np->construct = oo_browser_construct; | |
1474 oo_browser_construct = C_NULL; | |
1475 oo_browser_check_and_clear_structtype(); | |
1476 #endif | |
1477 | |
1478 add_node (np, &head); | |
1479 } | |
1480 | |
1481 /* Date: Wed, 22 Jan 1997 02:56:31 -0500 [last amended 18 Sep 1997] | |
1482 * From: Sam Kendall <kendall@mv.mv.com> | |
1483 * Subject: Proposal for firming up the TAGS format specification | |
1484 * To: F.Potorti@cnuce.cnr.it | |
1485 * | |
1486 * pfnote should emit the optimized form [unnamed tag] only if: | |
1487 * 1. name does not contain any of the characters " \t\r\n(),;"; | |
1488 * 2. linestart contains name as either a rightmost, or rightmost but | |
1489 * one character, substring; | |
1490 * 3. the character, if any, immediately before name in linestart must | |
1491 * be one of the characters " \t(),;"; | |
1492 * 4. the character, if any, immediately after name in linestart must | |
1493 * also be one of the characters " \t(),;". | |
1494 * | |
1495 * The real implementation uses the notinname() macro, which recognises | |
1496 * characters slightly different form " \t\r\n(),;". See the variable | |
1497 * `nonam'. | |
1498 */ | |
1499 #define traditional_tag_style TRUE | |
1500 void | |
1501 new_pfnote (name, namelen, is_func, linestart, linelen, lno, cno) | |
1502 char *name; /* tag name, or NULL if unnamed */ | |
1503 int namelen; /* tag length */ | |
1504 bool is_func; /* tag is a function */ | |
1505 char *linestart; /* start of the line where tag is */ | |
1506 int linelen; /* length of the line where tag is */ | |
1507 int lno; /* line number */ | |
1508 long cno; /* character number */ | |
1509 { | |
1510 register char *cp; | |
1511 bool named; | |
1512 | |
1513 named = TRUE; | |
1514 if (!CTAGS) | |
1515 { | |
1516 for (cp = name; !notinname (*cp); cp++) | |
1517 continue; | |
1518 if (*cp == '\0') /* rule #1 */ | |
1519 { | |
1520 cp = linestart + linelen - namelen; | |
1521 if (notinname (linestart[linelen-1])) | |
1522 cp -= 1; /* rule #4 */ | |
1523 #ifdef OO_BROWSER | |
1524 if (!oo_browser_format | |
1525 && cp >= linestart /* rule #2 */ | |
1526 #else | |
1527 if (cp >= linestart /* rule #2 */ | |
1528 #endif | |
1529 && (cp == linestart | |
1530 || notinname (cp[-1])) /* rule #3 */ | |
1531 && strneq (name, cp, namelen)) /* rule #2 */ | |
1532 named = FALSE; /* use unnamed tag */ | |
1533 } | |
1534 } | |
1535 | |
1536 if (named) | |
1537 name = savenstr (name, namelen); | |
1538 else | |
1539 name = NULL; | |
1540 pfnote (name, is_func, linestart, linelen, lno, cno); | |
1541 } | |
1542 | |
1543 /* | |
1544 * free_tree () | |
1545 * recurse on left children, iterate on right children. | |
1546 */ | |
1547 void | |
1548 free_tree (np) | |
1549 register node *np; | |
1550 { | |
1551 while (np) | |
1552 { | |
1553 register node *node_right = np->right; | |
1554 free_tree (np->left); | |
1555 if (np->name != NULL) | |
1556 free (np->name); | |
1557 free (np->pat); | |
1558 free (np); | |
1559 np = node_right; | |
1560 } | |
1561 } | |
1562 | |
1563 /* | |
1564 * add_node () | |
1565 * Adds a node to the tree of nodes. In etags mode, we don't keep | |
1566 * it sorted; we just keep a linear list. In ctags mode, maintain | |
1567 * an ordered tree, with no attempt at balancing. | |
1568 * | |
1569 * add_node is the only function allowed to add nodes, so it can | |
1570 * maintain state. | |
1571 */ | |
1572 void | |
1573 add_node (np, cur_node_p) | |
1574 node *np, **cur_node_p; | |
1575 { | |
1576 register int dif; | |
1577 register node *cur_node = *cur_node_p; | |
1578 | |
1579 if (cur_node == NULL) | |
1580 { | |
1581 *cur_node_p = np; | |
1582 last_node = np; | |
1583 return; | |
1584 } | |
1585 | |
1586 if (!CTAGS) | |
1587 { | |
1588 /* Etags Mode */ | |
1589 if (last_node == NULL) | |
1590 fatal ("internal error in add_node", (char *)NULL); | |
1591 last_node->right = np; | |
1592 last_node = np; | |
1593 } | |
1594 else | |
1595 { | |
1596 /* Ctags Mode */ | |
1597 dif = strcmp (np->name, cur_node->name); | |
1598 | |
1599 /* | |
1600 * If this tag name matches an existing one, then | |
1601 * do not add the node, but maybe print a warning. | |
1602 */ | |
1603 if (!dif) | |
1604 { | |
1605 if (streq (np->file, cur_node->file)) | |
1606 { | |
1607 if (!no_warnings) | |
1608 { | |
1609 fprintf (stderr, "Duplicate entry in file %s, line %d: %s\n", | |
1610 np->file, lineno, np->name); | |
1611 fprintf (stderr, "Second entry ignored\n"); | |
1612 } | |
1613 } | |
1614 else if (!cur_node->been_warned && !no_warnings) | |
1615 { | |
1616 fprintf | |
1617 (stderr, | |
1618 "Duplicate entry in files %s and %s: %s (Warning only)\n", | |
1619 np->file, cur_node->file, np->name); | |
1620 cur_node->been_warned = TRUE; | |
1621 } | |
1622 return; | |
1623 } | |
1624 | |
1625 /* Actually add the node */ | |
1626 add_node (np, dif < 0 ? &cur_node->left : &cur_node->right); | |
1627 } | |
1628 } | |
1629 | |
1630 #ifdef OO_BROWSER | |
1631 /* Default class name for the current OO-Browser tag. */ | |
1632 static char *oo_browser_class; | |
1633 /* Prefix character to use in OO-Browser listings for the current tag. */ | |
1634 static char oo_browser_prefix; | |
1635 #endif | |
1636 | |
1637 void | |
438 | 1638 put_entries (node *np) |
428 | 1639 { |
1640 register char *sp; | |
1641 | |
1642 if (np == NULL) | |
1643 return; | |
1644 | |
1645 /* Output subentries that precede this one */ | |
1646 put_entries (np->left); | |
1647 | |
1648 /* Output this entry */ | |
1649 | |
1650 if (!CTAGS) | |
1651 { | |
1652 #ifdef OO_BROWSER | |
1653 if (oo_browser_format) | |
1654 { | |
1655 /* Omit C++ `class' and `method' entries as well as Objective-C | |
1656 entries from this OO-Browser tags file since the browser handles | |
1657 them independently of this file. Omit `extern' variable declarations | |
1658 as they are unused by the OO-Browser. */ | |
1659 if (np->construct != C_CLASS | |
1660 && np->construct != C_METHOD | |
1661 && np->construct != C_EXTERN | |
1662 && np->construct != C_OBJC) | |
1663 { | |
1664 oo_browser_class = oo_browser_default_classes[np->construct]; | |
1665 switch (np->construct) | |
1666 { | |
1667 case C_CONSTANT: | |
1668 case C_ENUMERATION: | |
1669 case C_ENUM_LABEL: | |
1670 case C_STRUCTURE: | |
1671 case C_TYPE: | |
1672 case C_UNION: | |
1673 case C_VARIABLE: | |
1674 oo_browser_prefix = '='; | |
1675 break; | |
1676 case C_FUNCTION: | |
1677 case C_MACRO: | |
1678 oo_browser_prefix = '-'; | |
1679 break; | |
1680 } | |
1681 if (np->name != NULL) | |
1682 fprintf (tagf, "%s@%c %s@%s\n", | |
1683 oo_browser_class, oo_browser_prefix, | |
1684 np->name, np->pat); | |
1685 else | |
1686 fprintf (tagf, "%s@%c ???@%s\n", | |
1687 oo_browser_class, oo_browser_prefix, np->pat); | |
1688 } | |
1689 } | |
1690 else | |
1691 { | |
1692 #endif | |
1693 if (np->name != NULL) | |
1694 fprintf (tagf, "%s\177%s\001%d,%ld\n", | |
1695 np->pat, np->name, np->lno, np->cno); | |
1696 else | |
1697 fprintf (tagf, "%s\177%d,%ld\n", | |
1698 np->pat, np->lno, np->cno); | |
1699 #ifdef OO_BROWSER | |
1700 } | |
1701 #endif | |
1702 } | |
1703 else | |
1704 { | |
1705 if (np->name == NULL) | |
1706 error ("internal error: NULL name in ctags mode.", (char *)NULL); | |
1707 | |
1708 if (cxref_style) | |
1709 { | |
1710 if (vgrind_style) | |
1711 fprintf (stdout, "%s %s %d\n", | |
1712 np->name, np->file, (np->lno + 63) / 64); | |
1713 else | |
1714 fprintf (stdout, "%-16s %3d %-16s %s\n", | |
1715 np->name, np->lno, np->file, np->pat); | |
1716 } | |
1717 else | |
1718 { | |
1719 fprintf (tagf, "%s\t%s\t", np->name, np->file); | |
1720 | |
1721 if (np->is_func) | |
1722 { /* a function */ | |
1723 putc (searchar, tagf); | |
1724 putc ('^', tagf); | |
1725 | |
1726 for (sp = np->pat; *sp; sp++) | |
1727 { | |
1728 if (*sp == '\\' || *sp == searchar) | |
1729 putc ('\\', tagf); | |
1730 putc (*sp, tagf); | |
1731 } | |
1732 putc (searchar, tagf); | |
1733 } | |
1734 else | |
1735 { /* a typedef; text pattern inadequate */ | |
1736 fprintf (tagf, "%d", np->lno); | |
1737 } | |
1738 putc ('\n', tagf); | |
1739 } | |
1740 } | |
1741 | |
1742 /* Output subentries that follow this one */ | |
1743 put_entries (np->right); | |
1744 } | |
1745 | |
1746 /* Length of a number's decimal representation. */ | |
1747 int number_len PP ((long num)); | |
1748 int | |
1749 number_len (num) | |
1750 long num; | |
1751 { | |
1752 int len = 1; | |
1753 while ((num /= 10) > 0) | |
1754 len += 1; | |
1755 return len; | |
1756 } | |
1757 | |
1758 /* | |
1759 * Return total number of characters that put_entries will output for | |
1760 * the nodes in the subtree of the specified node. Works only if | |
1761 * we are not ctags, but called only in that case. This count | |
1762 * is irrelevant with the new tags.el, but is still supplied for | |
1763 * backward compatibility. | |
1764 */ | |
1765 int | |
1766 total_size_of_entries (np) | |
1767 register node *np; | |
1768 { | |
1769 register int total; | |
1770 | |
1771 if (np == NULL) | |
1772 return 0; | |
1773 | |
1774 for (total = 0; np != NULL; np = np->right) | |
1775 { | |
1776 /* Count left subentries. */ | |
1777 total += total_size_of_entries (np->left); | |
1778 | |
1779 /* Count this entry */ | |
1780 total += strlen (np->pat) + 1; | |
1781 total += number_len ((long) np->lno) + 1 + number_len (np->cno) + 1; | |
1782 if (np->name != NULL) | |
1783 total += 1 + strlen (np->name); /* \001name */ | |
1784 } | |
1785 | |
1786 return total; | |
1787 } | |
1788 | |
1789 /* | |
1790 * The C symbol tables. | |
1791 */ | |
1792 enum sym_type | |
1793 { | |
1794 st_none, | |
1795 st_C_objprot, st_C_objimpl, st_C_objend, | |
1796 st_C_gnumacro, | |
1797 st_C_ignore, | |
1798 st_C_javastruct, | |
1799 st_C_struct, st_C_enum, st_C_define, st_C_typedef, st_C_typespec, | |
1800 st_C_const | |
1801 #ifdef OO_BROWSER | |
1802 , st_C_union, st_C_class, st_C_extern, st_C_inline | |
1803 #endif | |
1804 }; | |
1805 | |
1806 /* Feed stuff between (but not including) %[ and %] lines to: | |
1807 gperf -c -k 1,3 -o -p -r -t | |
1808 %[ | |
1809 struct C_stab_entry { char *name; int c_ext; enum sym_type type; } | |
1810 %% | |
1811 @interface, 0, st_C_objprot | |
1812 @protocol, 0, st_C_objprot | |
1813 @implementation,0, st_C_objimpl | |
1814 @end, 0, st_C_objend | |
1815 import, C_JAVA, st_C_ignore | |
1816 package, C_JAVA, st_C_ignore | |
1817 friend, C_PLPL, st_C_ignore | |
1818 extends, C_JAVA, st_C_javastruct | |
1819 implements, C_JAVA, st_C_javastruct | |
1820 interface, C_JAVA, st_C_struct | |
1821 class, C_PLPL, st_C_class | |
1822 namespace, C_PLPL, st_C_struct | |
1823 domain, C_STAR, st_C_struct | |
1824 union, 0, st_C_union | |
1825 struct, 0, st_C_struct | |
1826 enum, 0, st_C_enum | |
1827 typedef, 0, st_C_typedef | |
1828 define, 0, st_C_define | |
1829 inline, 0, st_C_inline | |
1830 bool, C_PLPL, st_C_typespec | |
1831 long, 0, st_C_typespec | |
1832 short, 0, st_C_typespec | |
1833 int, 0, st_C_typespec | |
1834 char, 0, st_C_typespec | |
1835 float, 0, st_C_typespec | |
1836 double, 0, st_C_typespec | |
1837 signed, 0, st_C_typespec | |
1838 unsigned, 0, st_C_typespec | |
1839 auto, 0, st_C_typespec | |
1840 void, 0, st_C_typespec | |
1841 extern, 0, st_C_extern | |
1842 static, 0, st_C_typespec | |
1843 const, 0, st_C_const | |
1844 volatile, 0, st_C_typespec | |
1845 explicit, C_PLPL, st_C_typespec | |
1846 mutable, C_PLPL, st_C_typespec | |
1847 typename, C_PLPL, st_C_typespec | |
1848 # DEFUN used in emacs, the next three used in glibc (SYSCALL only for mach). | |
1849 DEFUN, 0, st_C_gnumacro | |
1850 SYSCALL, 0, st_C_gnumacro | |
1851 ENTRY, 0, st_C_gnumacro | |
1852 PSEUDO, 0, st_C_gnumacro | |
1853 # These are defined inside C functions, so currently they are not met. | |
1854 # EXFUN used in glibc, DEFVAR_* in emacs. | |
1855 #EXFUN, 0, st_C_gnumacro | |
1856 #DEFVAR_, 0, st_C_gnumacro | |
1857 %] | |
1858 and replace lines between %< and %> with its output. */ | |
1859 /*%<*/ | |
1860 /* C code produced by gperf version 2.5 (GNU C++ version) */ | |
1861 /* Command-line: gperf -c -k 1,3 -o -p -r -t */ | |
1862 struct C_stab_entry { char *name; int c_ext; enum sym_type type; }; | |
1863 | |
1864 #define TOTAL_KEYWORDS 41 | |
1865 #define MIN_WORD_LENGTH 3 | |
1866 #define MAX_WORD_LENGTH 15 | |
1867 #define MIN_HASH_VALUE 13 | |
1868 #define MAX_HASH_VALUE 129 | |
1869 /* maximum key range = 117, duplicates = 0 */ | |
1870 | |
1871 static unsigned int | |
438 | 1872 hash (char *str, unsigned int len) |
428 | 1873 { |
1874 static unsigned char asso_values[] = | |
1875 { | |
1876 130, 130, 130, 130, 130, 130, 130, 130, 130, 130, | |
1877 130, 130, 130, 130, 130, 130, 130, 130, 130, 130, | |
1878 130, 130, 130, 130, 130, 130, 130, 130, 130, 130, | |
1879 130, 130, 130, 130, 130, 130, 130, 130, 130, 130, | |
1880 130, 130, 130, 130, 130, 130, 130, 130, 130, 130, | |
1881 130, 130, 130, 130, 130, 130, 130, 130, 130, 130, | |
1882 130, 130, 130, 130, 13, 130, 130, 130, 33, 32, | |
1883 47, 130, 130, 130, 130, 130, 130, 130, 130, 130, | |
1884 5, 130, 130, 20, 32, 130, 130, 130, 130, 130, | |
1885 130, 130, 130, 130, 130, 130, 130, 47, 55, 8, | |
1886 15, 33, 61, 38, 130, 60, 130, 130, 2, 9, | |
1887 10, 62, 59, 130, 28, 27, 50, 19, 3, 130, | |
1888 130, 130, 130, 130, 130, 130, 130, 130, | |
1889 }; | |
438 | 1890 return len + |
1891 asso_values[(unsigned char) str[2]] + | |
1892 asso_values[(unsigned char) str[0]]; | |
428 | 1893 } |
1894 | |
438 | 1895 static struct C_stab_entry * |
1896 in_word_set (char *str, unsigned int len) | |
428 | 1897 { |
1898 static struct C_stab_entry wordlist[] = | |
1899 { | |
1900 {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, | |
1901 {"",}, {"",}, {"",}, {"",}, | |
1902 {"volatile", 0, st_C_typespec}, | |
1903 {"",}, {"",}, | |
1904 {"long", 0, st_C_typespec}, | |
1905 {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, | |
1906 {"const", 0, st_C_const}, | |
1907 {"",}, {"",}, {"",}, | |
1908 {"@end", 0, st_C_objend}, | |
1909 {"namespace", C_PLPL, st_C_struct}, | |
1910 {"",}, | |
1911 {"domain", C_STAR, st_C_struct}, | |
1912 {"",}, {"",}, | |
1913 {"@interface", 0, st_C_objprot}, | |
1914 {"",}, {"",}, {"",}, | |
1915 {"@implementation", 0, st_C_objimpl}, | |
1916 {"",}, {"",}, | |
1917 {"double", 0, st_C_typespec}, | |
1918 {"",}, {"",}, | |
1919 {"PSEUDO", 0, st_C_gnumacro}, | |
1920 {"",}, {"",}, {"",}, | |
1921 {"SYSCALL", 0, st_C_gnumacro}, | |
1922 {"",}, {"",}, | |
1923 {"@protocol", 0, st_C_objprot}, | |
1924 {"",}, {"",}, {"",}, | |
1925 {"unsigned", 0, st_C_typespec}, | |
1926 {"",}, | |
1927 {"enum", 0, st_C_enum}, | |
1928 {"",}, {"",}, | |
1929 {"char", 0, st_C_typespec}, | |
1930 {"class", C_PLPL, st_C_class}, | |
1931 {"struct", 0, st_C_struct}, | |
1932 {"",}, {"",}, {"",}, {"",}, | |
1933 {"mutable", C_PLPL, st_C_typespec}, | |
1934 {"void", 0, st_C_typespec}, | |
1935 {"inline", 0, st_C_inline}, | |
1936 {"ENTRY", 0, st_C_gnumacro}, | |
1937 {"",}, | |
1938 {"signed", 0, st_C_typespec}, | |
1939 {"",}, {"",}, | |
1940 {"package", C_JAVA, st_C_ignore}, | |
1941 {"",}, {"",}, {"",}, {"",}, {"",}, | |
1942 {"static", 0, st_C_typespec}, | |
1943 {"",}, | |
1944 {"define", 0, st_C_define}, | |
1945 {"",}, | |
1946 {"union", 0, st_C_union}, | |
1947 {"DEFUN", 0, st_C_gnumacro}, | |
1948 {"",}, {"",}, {"",}, | |
1949 {"extern", 0, st_C_extern}, | |
1950 {"extends", C_JAVA, st_C_javastruct}, | |
1951 {"",}, {"",}, {"",}, | |
1952 {"short", 0, st_C_typespec}, | |
1953 {"",}, {"",}, {"",}, {"",}, {"",}, | |
1954 {"explicit", C_PLPL, st_C_typespec}, | |
1955 {"auto", 0, st_C_typespec}, | |
1956 {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, | |
1957 {"",}, {"",}, | |
1958 {"int", 0, st_C_typespec}, | |
1959 {"",}, {"",}, | |
1960 {"typedef", 0, st_C_typedef}, | |
1961 {"typename", C_PLPL, st_C_typespec}, | |
1962 {"",}, | |
1963 {"interface", C_JAVA, st_C_struct}, | |
1964 {"",}, | |
1965 {"bool", C_PLPL, st_C_typespec}, | |
1966 {"",}, {"",}, {"",}, | |
1967 {"import", C_JAVA, st_C_ignore}, | |
1968 {"",}, | |
1969 {"friend", C_PLPL, st_C_ignore}, | |
1970 {"float", 0, st_C_typespec}, | |
1971 {"implements", C_JAVA, st_C_javastruct}, | |
1972 }; | |
1973 | |
1974 if (len <= MAX_WORD_LENGTH && len >= MIN_WORD_LENGTH) | |
1975 { | |
1976 register int key = hash (str, len); | |
1977 | |
1978 if (key <= MAX_HASH_VALUE && key >= 0) | |
1979 { | |
1980 register char *s = wordlist[key].name; | |
1981 | |
1982 if (*s == *str && !strncmp (str + 1, s + 1, len - 1)) | |
1983 return &wordlist[key]; | |
1984 } | |
1985 } | |
1986 return 0; | |
1987 } | |
1988 /*%>*/ | |
1989 | |
1990 enum sym_type C_symtype PP ((char *str, int len, int c_ext)); | |
1991 enum sym_type | |
1992 C_symtype (str, len, c_ext) | |
1993 char *str; | |
1994 int len; | |
1995 int c_ext; | |
1996 { | |
1997 register struct C_stab_entry *se = in_word_set (str, len); | |
1998 | |
1999 if (se == NULL || (se->c_ext && !(c_ext & se->c_ext))) | |
2000 return st_none; | |
2001 return se->type; | |
2002 } | |
2003 | |
2004 /* | |
2005 * C functions and variables are recognized using a simple | |
2006 * finite automaton. fvdef is its state variable. | |
2007 */ | |
2008 enum | |
2009 { | |
2010 fvnone, /* nothing seen */ | |
2011 fvnameseen, /* function or variable name seen */ | |
2012 fstartlist, /* func: just after open parenthesis */ | |
2013 finlist, /* func: in parameter list */ | |
2014 flistseen, /* func: after parameter list */ | |
2015 fignore, /* func: before open brace */ | |
2016 vignore /* var-like: ignore until ';' */ | |
2017 } fvdef; | |
2018 | |
2019 | |
2020 /* | |
2021 * typedefs are recognized using a simple finite automaton. | |
2022 * typdef is its state variable. | |
2023 */ | |
2024 enum | |
2025 { | |
2026 tnone, /* nothing seen */ | |
2027 ttypedseen, /* typedef keyword seen */ | |
2028 tinbody, /* inside typedef body */ | |
2029 tend, /* just before typedef tag */ | |
2030 tignore /* junk after typedef tag */ | |
2031 } typdef; | |
2032 | |
2033 | |
2034 /* | |
2035 * struct-like structures (enum, struct and union) are recognized | |
2036 * using another simple finite automaton. `structdef' is its state | |
2037 * variable. | |
2038 */ | |
2039 enum | |
2040 { | |
2041 snone, /* nothing seen yet */ | |
2042 skeyseen, /* struct-like keyword seen */ | |
2043 stagseen, /* struct-like tag seen */ | |
2044 scolonseen, /* colon seen after struct-like tag */ | |
2045 sinbody /* in struct body: recognize member func defs*/ | |
2046 } structdef; | |
2047 | |
2048 /* | |
2049 * When structdef is stagseen, scolonseen, or sinbody, structtag is the | |
2050 * struct tag, and structtype is the type of the preceding struct-like | |
2051 * keyword. | |
2052 */ | |
2053 char *structtag = "<uninited>"; | |
2054 enum sym_type structtype; | |
2055 | |
2056 #ifdef OO_BROWSER | |
2057 void | |
438 | 2058 oo_browser_check_and_clear_structtype(void) |
428 | 2059 { |
2060 /* Allow for multiple enum_label tags. */ | |
2061 if (structtype != st_C_enum) | |
2062 structtype = st_none; | |
2063 } | |
2064 #endif | |
2065 | |
2066 /* | |
2067 * When objdef is different from onone, objtag is the name of the class. | |
2068 */ | |
2069 char *objtag = "<uninited>"; | |
2070 | |
2071 /* | |
2072 * Yet another little state machine to deal with preprocessor lines. | |
2073 */ | |
2074 enum | |
2075 { | |
2076 dnone, /* nothing seen */ | |
2077 dsharpseen, /* '#' seen as first char on line */ | |
2078 ddefineseen, /* '#' and 'define' seen */ | |
2079 dignorerest /* ignore rest of line */ | |
2080 } definedef; | |
2081 | |
2082 /* | |
2083 * State machine for Objective C protocols and implementations. | |
2084 * Tom R.Hageman <tom@basil.icce.rug.nl> | |
2085 */ | |
2086 enum | |
2087 { | |
2088 onone, /* nothing seen */ | |
2089 oprotocol, /* @interface or @protocol seen */ | |
2090 oimplementation, /* @implementations seen */ | |
2091 otagseen, /* class name seen */ | |
2092 oparenseen, /* parenthesis before category seen */ | |
2093 ocatseen, /* category name seen */ | |
2094 oinbody, /* in @implementation body */ | |
2095 omethodsign, /* in @implementation body, after +/- */ | |
2096 omethodtag, /* after method name */ | |
2097 omethodcolon, /* after method colon */ | |
2098 omethodparm, /* after method parameter */ | |
2099 oignore /* wait for @end */ | |
2100 } objdef; | |
2101 | |
2102 /* | |
2103 * Use this structure to keep info about the token read, and how it | |
2104 * should be tagged. Used by the make_C_tag function to build a tag. | |
2105 */ | |
2106 typedef struct | |
2107 { | |
2108 bool valid; | |
2109 char *str; | |
2110 bool named; | |
2111 int linelen; | |
2112 int lineno; | |
2113 long linepos; | |
2114 char *buffer; | |
2115 } token; | |
2116 | |
2117 token tok; /* latest token read */ | |
2118 | |
2119 /* | |
2120 * Set this to TRUE, and the next token considered is called a function. | |
2121 * Used only for GNU emacs's function-defining macros. | |
2122 */ | |
2123 bool next_token_is_func; | |
2124 | |
2125 /* | |
2126 * TRUE in the rules part of a yacc file, FALSE outside (parse as C). | |
2127 */ | |
2128 bool yacc_rules; | |
2129 | |
2130 /* | |
2131 * methodlen is the length of the method name stored in token_name. | |
2132 */ | |
2133 int methodlen; | |
2134 | |
2135 #ifdef OO_BROWSER | |
2136 void | |
438 | 2137 oo_browser_clear_all_globals(void) |
428 | 2138 { |
2139 /* Initialize globals so there is no carry over between files. */ | |
2140 oo_browser_construct = C_NULL; | |
2141 fvdef = fvnone; typdef = tnone; structdef = snone; | |
2142 definedef = dnone; objdef = onone; | |
2143 structtype = st_none; | |
2144 next_token_is_func = yacc_rules = FALSE; | |
2145 } | |
2146 | |
2147 void | |
438 | 2148 oo_browser_clear_some_globals(void) |
428 | 2149 { |
2150 oo_browser_construct = C_NULL; | |
2151 structtype = st_none; | |
2152 } | |
2153 #endif | |
2154 | |
2155 /* | |
2156 * consider_token () | |
2157 * checks to see if the current token is at the start of a | |
2158 * function or variable, or corresponds to a typedef, or | |
2159 * is a struct/union/enum tag, or #define, or an enum constant. | |
2160 * | |
2161 * *IS_FUNC gets TRUE iff the token is a function or #define macro | |
2162 * with args. C_EXT is which language we are looking at. | |
2163 * | |
2164 * In the future we will need some way to adjust where the end of | |
2165 * the token is; for instance, implementing the C++ keyword | |
2166 * `operator' properly will adjust the end of the token to be after | |
2167 * whatever follows `operator'. | |
2168 * | |
2169 * Globals | |
2170 * fvdef IN OUT | |
2171 * structdef IN OUT | |
2172 * definedef IN OUT | |
2173 * typdef IN OUT | |
2174 * objdef IN OUT | |
2175 * next_token_is_func IN OUT | |
2176 */ | |
2177 bool consider_token PP ((char *str, int len, int c, int c_ext, | |
2178 int cblev, int parlev, bool *is_func_or_var)); | |
2179 bool | |
2180 consider_token (str, len, c, c_ext, cblev, parlev, is_func_or_var) | |
2181 register char *str; /* IN: token pointer */ | |
2182 register int len; /* IN: token length */ | |
2183 register int c; /* IN: first char after the token */ | |
2184 int c_ext; /* IN: C extensions mask */ | |
2185 int cblev; /* IN: curly brace level */ | |
2186 int parlev; /* IN: parenthesis level */ | |
2187 bool *is_func_or_var; /* OUT: function or variable found */ | |
2188 { | |
2189 enum sym_type toktype = C_symtype (str, len, c_ext); | |
2190 | |
2191 #ifdef OO_BROWSER | |
2192 switch (toktype) | |
2193 { | |
2194 case st_C_struct: | |
2195 set_construct(C_STRUCTURE); | |
2196 break; | |
2197 case st_C_union: | |
2198 set_construct(C_UNION); | |
2199 break; | |
2200 case st_C_class: | |
2201 set_construct(C_CLASS); | |
2202 break; | |
2203 case st_C_enum: | |
2204 set_construct(C_ENUMERATION); | |
2205 break; | |
2206 case st_C_typedef: | |
2207 set_construct(C_TYPE); | |
2208 break; | |
2209 case st_C_extern: | |
2210 set_construct(C_EXTERN); | |
2211 break; | |
2212 case st_C_inline: | |
2213 set_construct(C_FUNCTION); | |
2214 break; | |
2215 } | |
2216 #endif | |
2217 | |
2218 /* | |
2219 * Advance the definedef state machine. | |
2220 */ | |
2221 switch (definedef) | |
2222 { | |
2223 case dnone: | |
2224 /* We're not on a preprocessor line. */ | |
2225 break; | |
2226 case dsharpseen: | |
2227 if (toktype == st_C_define) | |
2228 { | |
2229 definedef = ddefineseen; | |
2230 } | |
2231 else | |
2232 { | |
2233 definedef = dignorerest; | |
2234 } | |
2235 return FALSE; | |
2236 case ddefineseen: | |
2237 /* | |
2238 * Make a tag for any macro, unless it is a constant | |
2239 * and constantypedefs is FALSE. | |
2240 */ | |
2241 definedef = dignorerest; | |
2242 #ifndef OO_BROWSER | |
2243 *is_func_or_var = (c == '('); | |
2244 #else | |
2245 { | |
2246 char *p = str + len * sizeof(char); | |
2247 | |
2248 if (*p == '(') | |
2249 /* This must be a macro since there is no | |
2250 whitespace between the opening parenthesis | |
2251 and the definition name. */ | |
2252 *is_func_or_var = TRUE; | |
2253 else | |
2254 { | |
2255 *is_func_or_var = FALSE; | |
2256 | |
2257 /* Handle possible whitespace between macro tag and opening | |
2258 parenthesis and ensure this is an actual macro. | |
2259 -- Bob Weiner, Altrasoft, 11/19/1997 */ | |
2260 while (*p && isspace(*p)) p++; | |
2261 if (*p) c = *p; | |
2262 | |
2263 /* Skip over nested parentheses. */ | |
2264 if (c == '(') | |
2265 { | |
2266 short depth = 1; | |
2267 | |
2268 while (*++p && depth > 0 && *p != '\n') | |
2269 { | |
2270 switch (*p) | |
2271 { | |
2272 case '(': | |
2273 depth++; break; | |
2274 case ')': | |
2275 depth--; break; | |
2276 } | |
2277 } | |
2278 | |
2279 /* If this is a macro, we have just passed | |
2280 the arguments and there will be more on | |
2281 the line before the NULL character that marks | |
2282 the end of the line token. */ | |
2283 while (*p == ' ' || *p == '\t') p++; | |
2284 if (*p) *is_func_or_var = TRUE; | |
2285 } | |
2286 } | |
2287 } | |
2288 | |
2289 set_construct((*is_func_or_var) ? C_MACRO : C_CONSTANT); | |
2290 #endif | |
2291 if (!*is_func_or_var && !constantypedefs) | |
2292 return FALSE; | |
2293 else | |
2294 return TRUE; | |
2295 case dignorerest: | |
2296 return FALSE; | |
2297 default: | |
2298 error ("internal error: definedef value.", (char *)NULL); | |
2299 } | |
2300 | |
2301 /* | |
2302 * Now typedefs | |
2303 */ | |
2304 switch (typdef) | |
2305 { | |
2306 case tnone: | |
2307 if (toktype == st_C_typedef) | |
2308 { | |
2309 if (typedefs) | |
2310 typdef = ttypedseen; | |
2311 fvdef = fvnone; | |
2312 return FALSE; | |
2313 } | |
2314 break; | |
2315 case ttypedseen: | |
2316 switch (toktype) | |
2317 { | |
2318 case st_C_const: | |
2319 set_construct(C_CONSTANT); | |
2320 /* fall through */ | |
2321 case st_none: | |
2322 case st_C_typespec: | |
2323 #ifdef OO_BROWSER | |
2324 case st_C_extern: | |
2325 #endif | |
2326 typdef = tend; | |
2327 break; | |
2328 case st_C_struct: | |
2329 case st_C_enum: | |
2330 #ifdef OO_BROWSER | |
2331 case st_C_union: | |
2332 case st_C_class: | |
2333 #endif | |
2334 break; | |
2335 } | |
2336 /* Do not return here, so the structdef stuff has a chance. */ | |
2337 break; | |
2338 case tend: | |
2339 switch (toktype) | |
2340 { | |
2341 case st_C_const: | |
2342 set_construct(C_CONSTANT); | |
2343 /* fall through */ | |
2344 case st_C_typespec: | |
2345 case st_C_struct: | |
2346 case st_C_enum: | |
2347 #ifdef OO_BROWSER | |
2348 case st_C_extern: | |
2349 case st_C_union: | |
2350 case st_C_class: | |
2351 #endif | |
2352 return FALSE; | |
2353 } | |
2354 return TRUE; | |
2355 } | |
2356 | |
2357 /* | |
2358 * This structdef business is currently only invoked when cblev==0. | |
2359 * It should be recursively invoked whatever the curly brace level, | |
2360 * and a stack of states kept, to allow for definitions of structs | |
2361 * within structs. | |
2362 * | |
2363 * This structdef business is NOT invoked when we are ctags and the | |
2364 * file is plain C. This is because a struct tag may have the same | |
2365 * name as another tag, and this loses with ctags. | |
2366 */ | |
2367 switch (toktype) | |
2368 { | |
2369 case st_C_javastruct: | |
2370 if (structdef == stagseen) | |
2371 structdef = scolonseen; | |
2372 return FALSE; | |
2373 case st_C_struct: | |
2374 case st_C_enum: | |
2375 #ifdef OO_BROWSER | |
2376 case st_C_union: | |
2377 case st_C_class: | |
2378 case st_C_extern: | |
2379 #endif | |
2380 if (typdef == ttypedseen | |
2381 || (typedefs_and_cplusplus && cblev == 0 && structdef == snone)) | |
2382 { | |
2383 structdef = skeyseen; | |
2384 structtype = toktype; | |
2385 } | |
2386 return FALSE; | |
2387 } | |
2388 | |
2389 if (structdef == skeyseen) | |
2390 { | |
2391 /* Save the tag for struct/union/class, for functions and variables | |
2392 that may be defined inside. */ | |
2393 #ifndef OO_BROWSER | |
2394 if (structtype == st_C_struct) | |
2395 #else | |
2396 if (structtype == st_C_struct | |
2397 || structtype == st_C_union | |
2398 || structtype == st_C_class) | |
2399 #endif | |
2400 structtag = savenstr (str, len); | |
2401 else | |
2402 structtag = "<enum>"; | |
2403 structdef = stagseen; | |
2404 return TRUE; | |
2405 } | |
2406 | |
2407 /* Avoid entering fvdef stuff if typdef is going on. */ | |
2408 if (typdef != tnone) | |
2409 { | |
2410 definedef = dnone; | |
2411 return FALSE; | |
2412 } | |
2413 | |
2414 /* Detect GNU macros. | |
2415 | |
2416 DEFUN note for writers of emacs C code: | |
2417 The DEFUN macro, used in emacs C source code, has a first arg | |
2418 that is a string (the lisp function name), and a second arg that | |
2419 is a C function name. Since etags skips strings, the second arg | |
2420 is tagged. This is unfortunate, as it would be better to tag the | |
2421 first arg. The simplest way to deal with this problem would be | |
2422 to name the tag with a name built from the function name, by | |
2423 removing the initial 'F' character and substituting '-' for '_'. | |
2424 Anyway, this assumes that the conventions of naming lisp | |
2425 functions will never change. Currently, this method is not | |
2426 implemented, so writers of emacs code are recommended to put the | |
2427 first two args of a DEFUN on the same line. */ | |
2428 if (definedef == dnone && toktype == st_C_gnumacro) | |
2429 { | |
2430 next_token_is_func = TRUE; | |
2431 return FALSE; | |
2432 } | |
2433 if (next_token_is_func) | |
2434 { | |
2435 next_token_is_func = FALSE; | |
2436 fvdef = fignore; | |
2437 *is_func_or_var = TRUE; | |
2438 return TRUE; | |
2439 } | |
2440 | |
2441 /* Detect Objective C constructs. */ | |
2442 switch (objdef) | |
2443 { | |
2444 case onone: | |
2445 switch (toktype) | |
2446 { | |
2447 case st_C_objprot: | |
2448 #ifdef OO_BROWSER | |
2449 set_construct(C_OBJC); | |
2450 #endif | |
2451 objdef = oprotocol; | |
2452 return FALSE; | |
2453 case st_C_objimpl: | |
2454 #ifdef OO_BROWSER | |
2455 set_construct(C_OBJC); | |
2456 #endif | |
2457 objdef = oimplementation; | |
2458 return FALSE; | |
2459 } | |
2460 break; | |
2461 case oimplementation: | |
2462 /* Save the class tag for functions or variables defined inside. */ | |
2463 objtag = savenstr (str, len); | |
2464 objdef = oinbody; | |
2465 return FALSE; | |
2466 case oprotocol: | |
2467 /* Save the class tag for categories. */ | |
2468 objtag = savenstr (str, len); | |
2469 objdef = otagseen; | |
2470 *is_func_or_var = TRUE; | |
2471 return TRUE; | |
2472 case oparenseen: | |
2473 objdef = ocatseen; | |
2474 *is_func_or_var = TRUE; | |
2475 return TRUE; | |
2476 case oinbody: | |
2477 break; | |
2478 case omethodsign: | |
2479 if (parlev == 0) | |
2480 { | |
2481 objdef = omethodtag; | |
2482 methodlen = len; | |
2483 grow_linebuffer (&token_name, methodlen + 1); | |
2484 strncpy (token_name.buffer, str, len); | |
2485 token_name.buffer[methodlen] = '\0'; | |
2486 token_name.len = methodlen; | |
2487 return TRUE; | |
2488 } | |
2489 return FALSE; | |
2490 case omethodcolon: | |
2491 if (parlev == 0) | |
2492 objdef = omethodparm; | |
2493 return FALSE; | |
2494 case omethodparm: | |
2495 if (parlev == 0) | |
2496 { | |
2497 objdef = omethodtag; | |
2498 methodlen += len; | |
2499 grow_linebuffer (&token_name, methodlen + 1); | |
2500 strncat (token_name.buffer, str, len); | |
2501 token_name.len = methodlen; | |
2502 return TRUE; | |
2503 } | |
2504 return FALSE; | |
2505 case oignore: | |
2506 if (toktype == st_C_objend) | |
2507 { | |
2508 /* Memory leakage here: the string pointed by objtag is | |
2509 never released, because many tests would be needed to | |
2510 avoid breaking on incorrect input code. The amount of | |
2511 memory leaked here is the sum of the lengths of the | |
2512 class tags. | |
2513 free (objtag); */ | |
2514 objdef = onone; | |
2515 } | |
2516 return FALSE; | |
2517 } | |
2518 | |
2519 /* A function, variable or enum constant? */ | |
2520 switch (toktype) | |
2521 { | |
2522 case st_C_const: | |
2523 set_construct(C_CONSTANT); | |
2524 /* fall through */ | |
2525 case st_C_typespec: | |
2526 #ifdef OO_BROWSER | |
2527 case st_C_extern: | |
2528 #endif | |
2529 if (fvdef != finlist && fvdef != fignore && fvdef != vignore) | |
2530 fvdef = fvnone; /* should be useless */ | |
2531 return FALSE; | |
2532 case st_C_ignore: | |
2533 fvdef = vignore; | |
2534 return FALSE; | |
2535 case st_none: | |
2536 if (constantypedefs && structdef == sinbody && structtype == st_C_enum) | |
2537 #ifdef OO_BROWSER | |
2538 { | |
2539 oo_browser_construct = C_ENUM_LABEL; | |
2540 #endif | |
2541 return TRUE; | |
2542 #ifdef OO_BROWSER | |
2543 } | |
2544 #endif | |
2545 if (fvdef == fvnone) | |
2546 { | |
2547 fvdef = fvnameseen; /* function or variable */ | |
2548 *is_func_or_var = TRUE; | |
2549 return TRUE; | |
2550 } | |
2551 } | |
2552 | |
2553 return FALSE; | |
2554 } | |
2555 | |
2556 /* | |
2557 * C_entries () | |
2558 * This routine finds functions, variables, typedefs, | |
2559 * #define's, enum constants and struct/union/enum definitions in | |
2560 * #C syntax and adds them to the list. | |
2561 */ | |
2562 #define current_lb_is_new (newndx == curndx) | |
2563 #define switch_line_buffers() (curndx = 1 - curndx) | |
2564 | |
2565 #define curlb (lbs[curndx].lb) | |
2566 #define othlb (lbs[1-curndx].lb) | |
2567 #define newlb (lbs[newndx].lb) | |
2568 #define curlinepos (lbs[curndx].linepos) | |
2569 #define othlinepos (lbs[1-curndx].linepos) | |
2570 #define newlinepos (lbs[newndx].linepos) | |
2571 | |
2572 #define CNL_SAVE_DEFINEDEF() \ | |
2573 do { \ | |
2574 curlinepos = charno; \ | |
2575 lineno++; \ | |
2576 linecharno = charno; \ | |
2577 charno += readline (&curlb, inf); \ | |
2578 lp = curlb.buffer; \ | |
2579 quotednl = FALSE; \ | |
2580 newndx = curndx; \ | |
2581 } while (0) | |
2582 | |
2583 #define CNL() \ | |
2584 do { \ | |
2585 CNL_SAVE_DEFINEDEF(); \ | |
2586 if (savetok.valid) \ | |
2587 { \ | |
2588 tok = savetok; \ | |
2589 savetok.valid = FALSE; \ | |
2590 } \ | |
2591 definedef = dnone; \ | |
2592 } while (0) | |
2593 | |
2594 | |
2595 void make_C_tag PP ((bool isfun)); | |
2596 void | |
2597 make_C_tag (isfun) | |
2598 bool isfun; | |
2599 { | |
2600 /* This function should never be called when tok.valid is FALSE, but | |
2601 we must protect against invalid input or internal errors. */ | |
2602 if (tok.valid) | |
2603 { | |
2604 if (traditional_tag_style) | |
2605 { | |
2606 /* This was the original code. Now we call new_pfnote instead, | |
2607 which uses the new method for naming tags (see new_pfnote). */ | |
2608 char *name = NULL; | |
2609 | |
2610 if (CTAGS || tok.named) | |
2611 name = savestr (token_name.buffer); | |
2612 pfnote (name, isfun, | |
2613 tok.buffer, tok.linelen, tok.lineno, tok.linepos); | |
2614 } | |
2615 else | |
2616 new_pfnote (token_name.buffer, token_name.len, isfun, | |
2617 tok.buffer, tok.linelen, tok.lineno, tok.linepos); | |
2618 tok.valid = FALSE; | |
2619 } | |
2620 else if (DEBUG) | |
2621 abort (); | |
2622 } | |
2623 | |
2624 | |
2625 void | |
2626 C_entries (c_ext, inf) | |
2627 int c_ext; /* extension of C */ | |
2628 FILE *inf; /* input file */ | |
2629 { | |
2630 register char c; /* latest char read; '\0' for end of line */ | |
2631 register char *lp; /* pointer one beyond the character `c' */ | |
2632 int curndx, newndx; /* indices for current and new lb */ | |
2633 register int tokoff; /* offset in line of start of current token */ | |
2634 register int toklen; /* length of current token */ | |
2635 char *qualifier; /* string used to qualify names */ | |
2636 int qlen; /* length of qualifier */ | |
2637 int cblev; /* current curly brace level */ | |
2638 int parlev; /* current parenthesis level */ | |
2639 bool incomm, inquote, inchar, quotednl, midtoken; | |
2640 bool cplpl, cjava; | |
4932 | 2641 token savetok = {0}; /* token saved during preprocessor handling */ |
428 | 2642 |
2643 | |
2644 tokoff = toklen = 0; /* keep compiler quiet */ | |
2645 curndx = newndx = 0; | |
2646 lineno = 0; | |
2647 charno = 0; | |
2648 lp = curlb.buffer; | |
2649 *lp = 0; | |
2650 | |
2651 fvdef = fvnone; typdef = tnone; structdef = snone; | |
2652 definedef = dnone; objdef = onone; | |
2653 next_token_is_func = yacc_rules = FALSE; | |
2654 midtoken = inquote = inchar = incomm = quotednl = FALSE; | |
2655 tok.valid = savetok.valid = FALSE; | |
2656 cblev = 0; | |
2657 parlev = 0; | |
2658 cplpl = (c_ext & C_PLPL) == C_PLPL; | |
2659 cjava = (c_ext & C_JAVA) == C_JAVA; | |
2660 if (cjava) | |
2661 { qualifier = "."; qlen = 1; } | |
2662 else | |
2663 { qualifier = "::"; qlen = 2; } | |
2664 | |
2665 while (!feof (inf)) | |
2666 { | |
2667 c = *lp++; | |
2668 if (c == '\\') | |
2669 { | |
2670 /* If we're at the end of the line, the next character is a | |
2671 '\0'; don't skip it, because it's the thing that tells us | |
2672 to read the next line. */ | |
2673 if (*lp == '\0') | |
2674 { | |
2675 quotednl = TRUE; | |
2676 continue; | |
2677 } | |
2678 lp++; | |
2679 c = ' '; | |
2680 } | |
2681 else if (incomm) | |
2682 { | |
2683 switch (c) | |
2684 { | |
2685 case '*': | |
2686 if (*lp == '/') | |
2687 { | |
2688 c = *lp++; | |
2689 incomm = FALSE; | |
2690 } | |
2691 break; | |
2692 case '\0': | |
2693 /* Newlines inside comments do not end macro definitions in | |
2694 traditional cpp. */ | |
2695 CNL_SAVE_DEFINEDEF (); | |
2696 break; | |
2697 } | |
2698 continue; | |
2699 } | |
2700 else if (inquote) | |
2701 { | |
2702 switch (c) | |
2703 { | |
2704 case '"': | |
2705 inquote = FALSE; | |
2706 break; | |
2707 case '\0': | |
2708 /* Newlines inside strings do not end macro definitions | |
2709 in traditional cpp, even though compilers don't | |
2710 usually accept them. */ | |
2711 CNL_SAVE_DEFINEDEF (); | |
2712 break; | |
2713 } | |
2714 continue; | |
2715 } | |
2716 else if (inchar) | |
2717 { | |
2718 switch (c) | |
2719 { | |
2720 case '\0': | |
2721 /* Hmmm, something went wrong. */ | |
2722 CNL (); | |
2723 /* FALLTHRU */ | |
2724 case '\'': | |
2725 inchar = FALSE; | |
2726 break; | |
2727 } | |
2728 continue; | |
2729 } | |
2730 else | |
2731 switch (c) | |
2732 { | |
2733 case '"': | |
2734 inquote = TRUE; | |
2735 if (fvdef != finlist && fvdef != fignore && fvdef !=vignore) | |
2736 fvdef = fvnone; | |
2737 continue; | |
2738 case '\'': | |
2739 inchar = TRUE; | |
2740 if (fvdef != finlist && fvdef != fignore && fvdef !=vignore) | |
2741 fvdef = fvnone; | |
2742 continue; | |
2743 case '/': | |
2744 if (*lp == '*') | |
2745 { | |
2746 lp++; | |
2747 incomm = TRUE; | |
2748 continue; | |
2749 } | |
2750 else if (/* cplpl && */ *lp == '/') | |
2751 { | |
2752 c = '\0'; | |
2753 break; | |
2754 } | |
2755 else | |
2756 break; | |
2757 case '%': | |
2758 if ((c_ext & YACC) && *lp == '%') | |
2759 { | |
2760 /* entering or exiting rules section in yacc file */ | |
2761 lp++; | |
2762 definedef = dnone; fvdef = fvnone; | |
2763 typdef = tnone; structdef = snone; | |
2764 next_token_is_func = FALSE; | |
2765 midtoken = inquote = inchar = incomm = quotednl = FALSE; | |
2766 cblev = 0; | |
2767 yacc_rules = !yacc_rules; | |
2768 continue; | |
2769 } | |
2770 else | |
2771 break; | |
2772 case '#': | |
2773 if (definedef == dnone) | |
2774 { | |
2775 char *cp; | |
2776 bool cpptoken = TRUE; | |
2777 | |
2778 /* Look back on this line. If all blanks, or nonblanks | |
2779 followed by an end of comment, this is a preprocessor | |
2780 token. */ | |
2781 for (cp = newlb.buffer; cp < lp-1; cp++) | |
2782 if (!iswhite (*cp)) | |
2783 { | |
2784 if (*cp == '*' && *(cp+1) == '/') | |
2785 { | |
2786 cp++; | |
2787 cpptoken = TRUE; | |
2788 } | |
2789 else | |
2790 cpptoken = FALSE; | |
2791 } | |
2792 if (cpptoken) | |
2793 definedef = dsharpseen; | |
2794 } /* if (definedef == dnone) */ | |
2795 | |
2796 continue; | |
2797 } /* switch (c) */ | |
2798 | |
2799 | |
2800 /* Consider token only if some complicated conditions are satisfied. */ | |
2801 if ((definedef != dnone | |
2802 || (cblev == 0 && structdef != scolonseen) | |
2803 || (cblev == 1 && cplpl && structdef == sinbody) | |
2804 || (structdef == sinbody && structtype == st_C_enum)) | |
2805 && typdef != tignore | |
2806 && definedef != dignorerest | |
2807 && fvdef != finlist) | |
2808 { | |
2809 if (midtoken) | |
2810 { | |
2811 if (endtoken (c)) | |
2812 { | |
2813 if (c == ':' && cplpl && *lp == ':' && begtoken(*(lp + 1))) | |
2814 { | |
2815 /* | |
2816 * This handles :: in the middle, but not at the | |
2817 * beginning of an identifier. | |
2818 */ | |
2819 lp += 2; | |
2820 toklen += 3; | |
2821 #ifdef OO_BROWSER | |
2822 set_construct(C_METHOD); | |
2823 #endif | |
2824 } | |
2825 else | |
2826 { | |
2827 bool funorvar = FALSE; | |
2828 | |
2829 if (yacc_rules | |
2830 || consider_token (newlb.buffer + tokoff, toklen, c, | |
2831 c_ext, cblev, parlev, &funorvar)) | |
2832 { | |
2833 tok.named = FALSE; | |
2834 if (structdef == sinbody | |
2835 && definedef == dnone | |
2836 && funorvar) | |
2837 /* function or var defined in C++ class body */ | |
2838 { | |
2839 int len = strlen (structtag) + qlen + toklen; | |
2840 grow_linebuffer (&token_name, len + 1); | |
2841 strcpy (token_name.buffer, structtag); | |
2842 strcat (token_name.buffer, qualifier); | |
2843 strncat (token_name.buffer, | |
2844 newlb.buffer + tokoff, toklen); | |
2845 token_name.len = len; | |
2846 tok.named = TRUE; | |
2847 #ifdef OO_BROWSER | |
2848 oo_browser_construct = C_METHOD; | |
2849 #endif | |
2850 } | |
2851 else if (objdef == ocatseen) | |
2852 /* Objective C category */ | |
2853 { | |
2854 int len = strlen (objtag) + 2 + toklen; | |
2855 grow_linebuffer (&token_name, len + 1); | |
2856 strcpy (token_name.buffer, objtag); | |
2857 strcat (token_name.buffer, "("); | |
2858 strncat (token_name.buffer, | |
2859 newlb.buffer + tokoff, toklen); | |
2860 strcat (token_name.buffer, ")"); | |
2861 token_name.len = len; | |
2862 tok.named = TRUE; | |
2863 #ifdef OO_BROWSER | |
2864 oo_browser_construct = C_OBJC; | |
2865 #endif | |
2866 } | |
2867 else if (objdef == omethodtag | |
2868 || objdef == omethodparm) | |
2869 /* Objective C method */ | |
2870 { | |
2871 tok.named = TRUE; | |
2872 #ifdef OO_BROWSER | |
2873 oo_browser_construct = C_OBJC; | |
2874 #endif | |
2875 } | |
2876 else | |
2877 { | |
2878 grow_linebuffer (&token_name, toklen + 1); | |
2879 strncpy (token_name.buffer, | |
2880 newlb.buffer + tokoff, toklen); | |
2881 token_name.buffer[toklen] = '\0'; | |
2882 token_name.len = toklen; | |
2883 /* Name macros. */ | |
2884 tok.named | |
2885 = (structdef == stagseen | |
2886 || typdef == tend | |
2887 #ifdef OO_BROWSER | |
2888 /* Also name #define constants, | |
2889 enumerations and enum_labels. | |
2890 Conditionalize `funorvar' reference | |
2891 here or #defines will appear without | |
2892 their #names. | |
2893 -- Bob Weiner, Altrasoft, 4/25/1998 */ | |
2894 || ((oo_browser_format || funorvar) | |
2895 && definedef == dignorerest) | |
2896 || (oo_browser_format | |
2897 && (oo_browser_construct == C_ENUMERATION | |
2898 || oo_browser_construct == C_ENUM_LABEL)) | |
2899 #else | |
2900 || (funorvar | |
2901 && definedef == dignorerest) | |
2902 #endif | |
2903 ); | |
2904 } | |
2905 tok.lineno = lineno; | |
2906 tok.linelen = tokoff + toklen + 1; | |
2907 tok.buffer = newlb.buffer; | |
2908 tok.linepos = newlinepos; | |
2909 tok.valid = TRUE; | |
2910 | |
2911 if (definedef == dnone | |
2912 && (fvdef == fvnameseen | |
2913 || structdef == stagseen | |
2914 || typdef == tend | |
2915 || objdef != onone)) | |
2916 { | |
2917 if (current_lb_is_new) | |
2918 switch_line_buffers (); | |
2919 } | |
2920 else | |
2921 make_C_tag (funorvar); | |
2922 } | |
2923 midtoken = FALSE; | |
2924 } | |
2925 } /* if (endtoken (c)) */ | |
2926 else if (intoken (c)) | |
2927 { | |
2928 toklen++; | |
2929 continue; | |
2930 } | |
2931 } /* if (midtoken) */ | |
2932 else if (begtoken (c)) | |
2933 { | |
2934 switch (definedef) | |
2935 { | |
2936 case dnone: | |
2937 switch (fvdef) | |
2938 { | |
2939 case fstartlist: | |
2940 fvdef = finlist; | |
2941 continue; | |
2942 case flistseen: | |
2943 #ifdef OO_BROWSER | |
2944 set_construct(C_MACRO); | |
2945 #endif | |
2946 make_C_tag (TRUE); /* a function */ | |
2947 fvdef = fignore; | |
2948 break; | |
2949 case fvnameseen: | |
2950 fvdef = fvnone; | |
2951 break; | |
2952 } | |
2953 if (structdef == stagseen && !cjava) | |
2954 structdef = snone; | |
2955 break; | |
2956 case dsharpseen: | |
2957 savetok = tok; | |
2958 } | |
2959 if (!yacc_rules || lp == newlb.buffer + 1) | |
2960 { | |
2961 tokoff = lp - 1 - newlb.buffer; | |
2962 toklen = 1; | |
2963 midtoken = TRUE; | |
2964 } | |
2965 continue; | |
2966 } /* if (begtoken) */ | |
2967 } /* if must look at token */ | |
2968 | |
2969 | |
2970 /* Detect end of line, colon, comma, semicolon and various braces | |
2971 after having handled a token.*/ | |
2972 switch (c) | |
2973 { | |
2974 case ':': | |
2975 if (definedef != dnone) | |
2976 break; | |
2977 switch (objdef) | |
2978 { | |
2979 case otagseen: | |
2980 objdef = oignore; | |
2981 make_C_tag (TRUE); /* an Objective C class */ | |
2982 break; | |
2983 case omethodtag: | |
2984 case omethodparm: | |
2985 objdef = omethodcolon; | |
2986 methodlen += 1; | |
2987 grow_linebuffer (&token_name, methodlen + 1); | |
2988 strcat (token_name.buffer, ":"); | |
2989 token_name.len = methodlen; | |
2990 break; | |
2991 } | |
2992 if (structdef == stagseen) | |
2993 structdef = scolonseen; | |
2994 else | |
2995 switch (fvdef) | |
2996 { | |
2997 case fvnameseen: | |
2998 if (yacc_rules) | |
2999 { | |
3000 make_C_tag (FALSE); /* a yacc function */ | |
3001 fvdef = fignore; | |
3002 } | |
3003 break; | |
3004 case fstartlist: | |
3005 fvdef = fvnone; | |
3006 break; | |
3007 } | |
3008 break; | |
3009 case ';': | |
3010 if (definedef != dnone) | |
3011 break; | |
3012 if (cblev == 0) | |
3013 switch (typdef) | |
3014 { | |
3015 case tend: | |
3016 #ifdef OO_BROWSER | |
3017 set_construct(C_TYPE); | |
3018 #endif | |
3019 make_C_tag (FALSE); /* a typedef */ | |
3020 /* FALLTHRU */ | |
3021 default: | |
3022 typdef = tnone; | |
3023 } | |
3024 switch (fvdef) | |
3025 { | |
3026 case fignore: | |
3027 break; | |
3028 case fvnameseen: | |
3029 if ((globals && cblev == 0) || (members && cblev == 1)) | |
3030 #ifndef OO_BROWSER | |
3031 make_C_tag (FALSE); /* a variable */ | |
3032 #else | |
3033 /* if (constantypedefs && structdef == snone)*/ | |
3034 { | |
3035 tok.named = TRUE; | |
3036 switch (structtype) | |
3037 { | |
3038 case st_C_enum: | |
3039 set_construct(C_ENUMERATION); | |
3040 break; | |
3041 case st_C_class: | |
3042 set_construct(C_CLASS); | |
3043 break; | |
3044 default: | |
3045 set_construct(C_VARIABLE); | |
3046 break; | |
3047 } | |
3048 make_C_tag (FALSE); | |
3049 /* Force reset of st_C_enum structtype value. */ | |
3050 structtype = st_none; | |
3051 } | |
3052 #endif | |
3053 /* FALLTHRU */ | |
3054 default: | |
3055 fvdef = fvnone; | |
3056 /* The following instruction invalidates the token. | |
3057 Probably the token should be invalidated in all | |
3058 other cases where some state machine is reset. */ | |
3059 tok.valid = FALSE; | |
3060 } | |
3061 if (structdef == stagseen) | |
3062 structdef = snone; | |
3063 break; | |
3064 case ',': | |
3065 if (definedef != dnone) | |
3066 break; | |
3067 switch (objdef) | |
3068 { | |
3069 case omethodtag: | |
3070 case omethodparm: | |
3071 make_C_tag (TRUE); /* an Objective C method */ | |
3072 objdef = oinbody; | |
3073 break; | |
3074 } | |
3075 switch (fvdef) | |
3076 { | |
3077 case finlist: | |
3078 case fignore: | |
3079 case vignore: | |
3080 break; | |
3081 case fvnameseen: | |
3082 if ((globals && cblev == 0) || (members && cblev == 1)) | |
3083 make_C_tag (FALSE); /* a variable */ | |
3084 break; | |
3085 default: | |
3086 fvdef = fvnone; | |
3087 } | |
3088 if (structdef == stagseen) | |
3089 structdef = snone; | |
3090 break; | |
3091 case '[': | |
3092 if (definedef != dnone) | |
3093 break; | |
3094 if (cblev == 0 && typdef == tend) | |
3095 { | |
3096 #ifdef OO_BROWSER | |
3097 set_construct(C_TYPE); | |
3098 #endif | |
3099 typdef = tignore; | |
3100 make_C_tag (FALSE); /* a typedef */ | |
3101 break; | |
3102 } | |
3103 switch (fvdef) | |
3104 { | |
3105 case finlist: | |
3106 case fignore: | |
3107 case vignore: | |
3108 break; | |
3109 case fvnameseen: | |
3110 #ifndef OO_BROWSER | |
3111 if ((globals && cblev == 0) || (members && cblev == 1)) | |
3112 make_C_tag (FALSE); /* a variable */ | |
3113 #else | |
3114 if (constantypedefs && structdef == snone) | |
3115 { | |
3116 tok.named = TRUE; | |
3117 switch (structtype) | |
3118 { | |
3119 case st_C_enum: | |
3120 set_construct(C_ENUMERATION); | |
3121 break; | |
3122 case st_C_class: | |
3123 set_construct(C_CLASS); | |
3124 break; | |
3125 default: | |
3126 set_construct(C_VARIABLE); | |
3127 break; | |
3128 } | |
3129 make_C_tag (FALSE); | |
3130 /* Force reset of st_C_enum structtype value. */ | |
3131 structtype = st_none; | |
3132 } | |
3133 #endif | |
3134 /* FALLTHRU */ | |
3135 default: | |
3136 fvdef = fvnone; | |
3137 } | |
3138 if (structdef == stagseen) | |
3139 structdef = snone; | |
3140 break; | |
3141 case '(': | |
3142 if (definedef != dnone) | |
3143 break; | |
3144 if (objdef == otagseen && parlev == 0) | |
3145 objdef = oparenseen; | |
3146 switch (fvdef) | |
3147 { | |
3148 case fvnone: | |
3149 switch (typdef) | |
3150 { | |
3151 case ttypedseen: | |
3152 case tend: | |
3153 if (tok.valid && *lp != '*') | |
3154 { | |
3155 /* This handles constructs like: | |
3156 typedef void OperatorFun (int fun); */ | |
3157 typdef = tignore; | |
3158 #ifdef OO_BROWSER | |
3159 set_construct(C_TYPE); | |
3160 #endif | |
3161 make_C_tag (FALSE); | |
3162 } | |
3163 break; | |
3164 } /* switch (typdef) */ | |
3165 break; | |
3166 case fvnameseen: | |
3167 fvdef = fstartlist; | |
3168 break; | |
3169 case flistseen: | |
3170 fvdef = finlist; | |
3171 break; | |
3172 } | |
3173 parlev++; | |
3174 break; | |
3175 case ')': | |
3176 if (definedef != dnone) | |
3177 break; | |
3178 if (objdef == ocatseen && parlev == 1) | |
3179 { | |
3180 make_C_tag (TRUE); /* an Objective C category */ | |
3181 objdef = oignore; | |
3182 } | |
3183 if (--parlev == 0) | |
3184 { | |
3185 switch (fvdef) | |
3186 { | |
3187 case fstartlist: | |
3188 case finlist: | |
3189 fvdef = flistseen; | |
3190 break; | |
3191 } | |
3192 if (cblev == 0 && typdef == tend) | |
3193 { | |
3194 #ifdef OO_BROWSER | |
3195 set_construct(C_TYPE); | |
3196 #endif | |
3197 typdef = tignore; | |
3198 make_C_tag (FALSE); /* a typedef */ | |
3199 } | |
3200 } | |
3201 else if (parlev < 0) /* can happen due to ill-conceived #if's. */ | |
3202 parlev = 0; | |
3203 break; | |
3204 case '{': | |
3205 if (definedef != dnone) | |
3206 break; | |
3207 if (typdef == ttypedseen) | |
3208 typdef = tinbody; | |
3209 switch (structdef) | |
3210 { | |
3211 case skeyseen: /* unnamed struct */ | |
3212 structdef = sinbody; | |
3213 structtag = "_anonymous_"; | |
3214 break; | |
3215 case stagseen: | |
3216 case scolonseen: /* named struct */ | |
3217 structdef = sinbody; | |
3218 make_C_tag (FALSE); /* a struct */ | |
3219 break; | |
3220 } | |
3221 switch (fvdef) | |
3222 { | |
3223 case flistseen: | |
3224 #ifdef OO_BROWSER | |
3225 set_construct(C_FUNCTION); | |
3226 /* Ensure function name is recorded. | |
3227 -- Bob Weiner, Altrasoft */ | |
3228 tok.named = TRUE; | |
3229 #endif | |
3230 make_C_tag (TRUE); /* a function */ | |
3231 /* FALLTHRU */ | |
3232 case fignore: | |
3233 fvdef = fvnone; | |
3234 break; | |
3235 case fvnone: | |
3236 switch (objdef) | |
3237 { | |
3238 case otagseen: | |
3239 make_C_tag (TRUE); /* an Objective C class */ | |
3240 objdef = oignore; | |
3241 break; | |
3242 case omethodtag: | |
3243 case omethodparm: | |
3244 make_C_tag (TRUE); /* an Objective C method */ | |
3245 objdef = oinbody; | |
3246 break; | |
3247 default: | |
3248 /* Neutralize `extern "C" {' grot. */ | |
3249 if (cblev == 0 && structdef == snone && typdef == tnone) | |
3250 cblev = -1; | |
3251 } | |
3252 } | |
3253 cblev++; | |
3254 break; | |
3255 case '*': | |
3256 if (definedef != dnone) | |
3257 break; | |
3258 if (fvdef == fstartlist) | |
3259 fvdef = fvnone; /* avoid tagging `foo' in `foo (*bar()) ()' */ | |
3260 break; | |
3261 case '}': | |
3262 if (definedef != dnone) | |
3263 break; | |
3264 if (!noindentypedefs && lp == newlb.buffer + 1) | |
3265 { | |
3266 cblev = 0; /* reset curly brace level if first column */ | |
3267 parlev = 0; /* also reset paren level, just in case... */ | |
3268 } | |
3269 else if (cblev > 0) | |
3270 cblev--; | |
3271 if (cblev == 0) | |
3272 { | |
3273 if (typdef == tinbody) | |
3274 typdef = tend; | |
3275 /* Memory leakage here: the string pointed by structtag is | |
3276 never released, because I fear to miss something and | |
3277 break things while freeing the area. The amount of | |
3278 memory leaked here is the sum of the lengths of the | |
3279 struct tags. | |
3280 if (structdef == sinbody) | |
3281 free (structtag); */ | |
3282 | |
3283 structdef = snone; | |
3284 structtag = "<error>"; | |
3285 #ifdef OO_BROWSER | |
3286 /* Next line added to avoid any state carryover between | |
3287 functions. -- Bob Weiner, Altrasoft, 11/19/1997 */ | |
3288 fvdef = fvnone; oo_browser_construct = C_NULL; | |
3289 #endif | |
3290 } | |
3291 break; | |
3292 case '=': | |
3293 if (definedef != dnone) | |
3294 break; | |
3295 #ifdef OO_BROWSER | |
3296 { | |
3297 int is_method = 0; | |
3298 #endif | |
3299 switch (fvdef) | |
3300 { | |
3301 case finlist: | |
3302 case fignore: | |
3303 case vignore: | |
3304 break; | |
3305 case fvnameseen: | |
3306 if ((globals && cblev == 0) || (members && cblev == 1)) | |
3307 #ifndef OO_BROWSER | |
3308 make_C_tag (FALSE); /* a variable */ | |
3309 #else | |
3310 { | |
3311 tok.named = TRUE; | |
3312 switch (structtype) | |
3313 { | |
3314 case st_C_enum: | |
3315 set_construct(C_ENUMERATION); | |
3316 break; | |
3317 case st_C_class: | |
3318 set_construct(C_CLASS); | |
3319 break; | |
3320 default: | |
3321 /* a global variable */ | |
3322 set_construct(C_VARIABLE); | |
3323 break; | |
3324 } | |
3325 | |
3326 /* ootags categorizes each tag found whereas etags doesn't. | |
3327 Set the is_method flag if this tag has been marked as | |
3328 such by an earlier section of code. | |
3329 -- Steve Baur, Altrasoft, 5/7/1998 */ | |
3330 is_method = (oo_browser_construct == C_METHOD); | |
3331 | |
3332 make_C_tag (FALSE); | |
3333 /* Force reset of st_C_enum structtype value. */ | |
3334 structtype = st_none; | |
3335 } | |
3336 #endif | |
3337 /* FALLTHRU */ | |
3338 default: | |
3339 #ifdef OO_BROWSER | |
3340 fvdef = is_method ? fignore : vignore; | |
3341 #else | |
3342 fvdef = vignore; | |
3343 #endif | |
3344 } | |
3345 #ifdef OO_BROWSER | |
3346 } | |
3347 #endif | |
3348 break; | |
3349 case '+': | |
3350 case '-': | |
3351 if (objdef == oinbody && cblev == 0) | |
3352 { | |
3353 objdef = omethodsign; | |
3354 break; | |
3355 } | |
3356 /* FALLTHRU */ | |
3357 case '#': case '~': case '&': case '%': case '/': case '|': | |
3358 case '^': case '!': case '<': case '>': case '.': case '?': case ']': | |
3359 if (definedef != dnone) | |
3360 break; | |
3361 #ifdef OO_BROWSER | |
3362 if (!cplpl) | |
3363 { | |
3364 #endif | |
3365 /* The above characters cannot follow a function tag in C, so | |
3366 unmark this as a function entry. For C++, these characters | |
3367 may follow an `operator' function construct, so skip the | |
3368 unmarking conditional below. | |
3369 -- Steve Baur, Altrasoft, 5/7/1998 */ | |
3370 if (fvdef != finlist && fvdef != fignore && fvdef != vignore) | |
3371 fvdef = fvnone; | |
3372 #ifdef OO_BROWSER | |
3373 } | |
3374 #endif | |
3375 break; | |
3376 case '\0': | |
3377 if (objdef == otagseen) | |
3378 { | |
3379 make_C_tag (TRUE); /* an Objective C class */ | |
3380 objdef = oignore; | |
3381 } | |
3382 /* If a macro spans multiple lines don't reset its state. */ | |
3383 if (quotednl) | |
3384 CNL_SAVE_DEFINEDEF (); | |
3385 else | |
3386 CNL (); | |
3387 break; | |
3388 } /* switch (c) */ | |
3389 | |
3390 } /* while not eof */ | |
3391 } | |
3392 | |
3393 /* | |
3394 * Process either a C++ file or a C file depending on the setting | |
3395 * of a global flag. | |
3396 */ | |
3397 void | |
3398 default_C_entries (inf) | |
3399 FILE *inf; | |
3400 { | |
3401 C_entries (cplusplus ? C_PLPL : 0, inf); | |
3402 } | |
3403 | |
3404 /* Always do plain ANSI C. */ | |
3405 void | |
3406 plain_C_entries (inf) | |
3407 FILE *inf; | |
3408 { | |
3409 C_entries (0, inf); | |
3410 } | |
3411 | |
3412 /* Always do C++. */ | |
3413 void | |
3414 Cplusplus_entries (inf) | |
3415 FILE *inf; | |
3416 { | |
3417 C_entries (C_PLPL, inf); | |
3418 } | |
3419 | |
3420 /* Always do Java. */ | |
3421 void | |
3422 Cjava_entries (inf) | |
3423 FILE *inf; | |
3424 { | |
3425 C_entries (C_JAVA, inf); | |
3426 } | |
3427 | |
3428 /* Always do C*. */ | |
3429 void | |
3430 Cstar_entries (inf) | |
3431 FILE *inf; | |
3432 { | |
3433 C_entries (C_STAR, inf); | |
3434 } | |
3435 | |
3436 /* Always do Yacc. */ | |
3437 void | |
3438 Yacc_entries (inf) | |
3439 FILE *inf; | |
3440 { | |
3441 C_entries (YACC, inf); | |
3442 } | |
3443 | |
3444 /* A useful macro. */ | |
3445 #define LOOP_ON_INPUT_LINES(file_pointer, line_buffer, char_pointer) \ | |
3446 for (lineno = charno = 0; /* loop initialization */ \ | |
3447 !feof (file_pointer) /* loop test */ \ | |
3448 && (lineno++, /* instructions at start of loop */ \ | |
3449 linecharno = charno, \ | |
3450 charno += readline (&line_buffer, file_pointer), \ | |
3451 char_pointer = lb.buffer, \ | |
3452 TRUE); \ | |
3453 ) | |
3454 | |
3455 | |
3456 /* | |
3457 * Read a file, but do no processing. This is used to do regexp | |
3458 * matching on files that have no language defined. | |
3459 */ | |
3460 void | |
3461 just_read_file (inf) | |
3462 FILE *inf; | |
3463 { | |
3464 register char *dummy; | |
3465 | |
3466 LOOP_ON_INPUT_LINES (inf, lb, dummy) | |
3467 continue; | |
3468 } | |
3469 | |
3470 /* Fortran parsing */ | |
3471 | |
3472 bool tail PP ((char *cp)); | |
3473 bool | |
3474 tail (cp) | |
3475 char *cp; | |
3476 { | |
3477 register int len = 0; | |
3478 | |
3479 while (*cp && lowcase(*cp) == lowcase(dbp[len])) | |
3480 cp++, len++; | |
3481 if (*cp == '\0' && !intoken(dbp[len])) | |
3482 { | |
3483 dbp += len; | |
3484 return TRUE; | |
3485 } | |
3486 return FALSE; | |
3487 } | |
3488 | |
3489 void | |
3490 takeprec () | |
3491 { | |
3492 dbp = skip_spaces (dbp); | |
3493 if (*dbp != '*') | |
3494 return; | |
3495 dbp++; | |
3496 dbp = skip_spaces (dbp); | |
3497 if (strneq (dbp, "(*)", 3)) | |
3498 { | |
3499 dbp += 3; | |
3500 return; | |
3501 } | |
3502 if (!isdigit (*dbp)) | |
3503 { | |
3504 --dbp; /* force failure */ | |
3505 return; | |
3506 } | |
3507 do | |
3508 dbp++; | |
3509 while (isdigit (*dbp)); | |
3510 } | |
3511 | |
3512 void getit PP ((FILE *inf)); | |
3513 void | |
3514 getit (inf) | |
3515 FILE *inf; | |
3516 { | |
3517 register char *cp; | |
3518 | |
3519 dbp = skip_spaces (dbp); | |
3520 if (*dbp == '\0') | |
3521 { | |
3522 lineno++; | |
3523 linecharno = charno; | |
3524 charno += readline (&lb, inf); | |
3525 dbp = lb.buffer; | |
3526 if (dbp[5] != '&') | |
3527 return; | |
3528 dbp += 6; | |
3529 dbp = skip_spaces (dbp); | |
3530 } | |
3531 if (!isalpha (*dbp) | |
3532 && *dbp != '_' | |
3533 && *dbp != '$') | |
3534 return; | |
3535 for (cp = dbp + 1; *cp && intoken (*cp); cp++) | |
3536 continue; | |
3537 pfnote ((CTAGS) ? savenstr (dbp, cp-dbp) : NULL, TRUE, | |
3538 lb.buffer, cp - lb.buffer + 1, lineno, linecharno); | |
3539 } | |
3540 | |
3541 | |
3542 void | |
3543 Fortran_functions (inf) | |
3544 FILE *inf; | |
3545 { | |
3546 LOOP_ON_INPUT_LINES (inf, lb, dbp) | |
3547 { | |
3548 if (*dbp == '%') | |
3549 dbp++; /* Ratfor escape to fortran */ | |
3550 dbp = skip_spaces (dbp); | |
3551 if (*dbp == '\0') | |
3552 continue; | |
3553 switch (lowcase (*dbp)) | |
3554 { | |
3555 case 'i': | |
3556 if (tail ("integer")) | |
3557 takeprec (); | |
3558 break; | |
3559 case 'r': | |
3560 if (tail ("real")) | |
3561 takeprec (); | |
3562 break; | |
3563 case 'l': | |
3564 if (tail ("logical")) | |
3565 takeprec (); | |
3566 break; | |
3567 case 'c': | |
3568 if (tail ("complex") || tail ("character")) | |
3569 takeprec (); | |
3570 break; | |
3571 case 'd': | |
3572 if (tail ("double")) | |
3573 { | |
3574 dbp = skip_spaces (dbp); | |
3575 if (*dbp == '\0') | |
3576 continue; | |
3577 if (tail ("precision")) | |
3578 break; | |
3579 continue; | |
3580 } | |
3581 break; | |
3582 } | |
3583 dbp = skip_spaces (dbp); | |
3584 if (*dbp == '\0') | |
3585 continue; | |
3586 switch (lowcase (*dbp)) | |
3587 { | |
3588 case 'f': | |
3589 if (tail ("function")) | |
3590 getit (inf); | |
3591 continue; | |
3592 case 's': | |
3593 if (tail ("subroutine")) | |
3594 getit (inf); | |
3595 continue; | |
3596 case 'e': | |
3597 if (tail ("entry")) | |
3598 getit (inf); | |
3599 continue; | |
3600 case 'p': | |
3601 if (tail ("program")) | |
3602 { | |
3603 getit (inf); | |
3604 continue; | |
3605 } | |
3606 if (tail ("procedure")) | |
3607 getit (inf); | |
3608 continue; | |
3609 } | |
3610 } | |
3611 } | |
3612 | |
3613 /* | |
3614 * Bob Weiner, Motorola Inc., 4/3/94 | |
3615 * Unix and microcontroller assembly tag handling | |
3616 * look for '^[a-zA-Z_.$][a-zA_Z0-9_.$]*[: ^I^J]' | |
3617 */ | |
3618 void | |
438 | 3619 Asm_labels (FILE *inf) |
428 | 3620 { |
3621 register char *cp; | |
3622 | |
3623 LOOP_ON_INPUT_LINES (inf, lb, cp) | |
3624 { | |
3625 /* If first char is alphabetic or one of [_.$], test for colon | |
3626 following identifier. */ | |
3627 if (isalpha (*cp) || *cp == '_' || *cp == '.' || *cp == '$') | |
3628 { | |
3629 /* Read past label. */ | |
3630 cp++; | |
3631 while (isalnum (*cp) || *cp == '_' || *cp == '.' || *cp == '$') | |
3632 cp++; | |
3633 if (*cp == ':' || isspace (*cp)) | |
3634 { | |
3635 /* Found end of label, so copy it and add it to the table. */ | |
3636 pfnote ((CTAGS) ? savenstr(lb.buffer, cp-lb.buffer) : NULL, TRUE, | |
3637 lb.buffer, cp - lb.buffer + 1, lineno, linecharno); | |
3638 } | |
3639 } | |
3640 } | |
3641 } | |
3642 | |
3643 /* | |
3644 * Perl support by Bart Robinson <lomew@cs.utah.edu> | |
3645 * enhanced by Michael Ernst <mernst@alum.mit.edu> | |
3646 * Perl sub names: look for /^sub[ \t\n]+[^ \t\n{]+/ | |
3647 * Perl variable names: /^(my|local).../ | |
3648 */ | |
3649 void | |
3650 Perl_functions (inf) | |
3651 FILE *inf; | |
3652 { | |
3653 register char *cp; | |
3654 | |
3655 LOOP_ON_INPUT_LINES (inf, lb, cp) | |
3656 { | |
3657 if (*cp++ == 's' | |
3658 && *cp++ == 'u' | |
3659 && *cp++ == 'b' && isspace (*cp++)) | |
3660 { | |
3661 cp = skip_spaces (cp); | |
3662 if (*cp != '\0') | |
3663 { | |
3664 while (*cp != '\0' | |
3665 && !isspace (*cp) && *cp != '{' && *cp != '(') | |
3666 cp++; | |
3667 pfnote ((CTAGS) ? savenstr(lb.buffer, cp-lb.buffer) : NULL, TRUE, | |
3668 lb.buffer, cp - lb.buffer + 1, lineno, linecharno); | |
3669 } | |
3670 } | |
3671 else if (globals /* only if tagging global vars is enabled */ | |
3672 && ((cp = lb.buffer, | |
3673 *cp++ == 'm' | |
3674 && *cp++ == 'y') | |
3675 || (cp = lb.buffer, | |
3676 *cp++ == 'l' | |
3677 && *cp++ == 'o' | |
3678 && *cp++ == 'c' | |
3679 && *cp++ == 'a' | |
3680 && *cp++ == 'l')) | |
3681 && (*cp == '(' || isspace (*cp))) | |
3682 { | |
3683 /* After "my" or "local", but before any following paren or space. */ | |
3684 char *varname = NULL; | |
3685 | |
3686 cp = skip_spaces (cp); | |
3687 if (*cp == '$' || *cp == '@' || *cp == '%') | |
3688 { | |
3689 char* varstart = ++cp; | |
3690 while (isalnum (*cp) || *cp == '_') | |
3691 cp++; | |
3692 varname = savenstr (varstart, cp-varstart); | |
3693 } | |
3694 else | |
3695 { | |
3696 /* Should be examining a variable list at this point; | |
3697 could insist on seeing an open parenthesis. */ | |
3698 while (*cp != '\0' && *cp != ';' && *cp != '=' && *cp != ')') | |
3699 cp++; | |
3700 } | |
3701 | |
3702 /* Perhaps I should back cp up one character, so the TAGS table | |
3703 doesn't mention (and so depend upon) the following char. */ | |
3704 pfnote ((CTAGS) ? savenstr (lb.buffer, cp-lb.buffer) : varname, | |
3705 FALSE, lb.buffer, cp - lb.buffer + 1, lineno, linecharno); | |
3706 } | |
3707 } | |
3708 } | |
3709 | |
3710 /* | |
3711 * Python support by Eric S. Raymond <esr@thyrsus.com> | |
3712 * Look for /^def[ \t\n]+[^ \t\n(:]+/ or /^class[ \t\n]+[^ \t\n(:]+/ | |
3713 */ | |
3714 void | |
3715 Python_functions (inf) | |
3716 FILE *inf; | |
3717 { | |
3718 register char *cp; | |
3719 | |
3720 LOOP_ON_INPUT_LINES (inf, lb, cp) | |
3721 { | |
3722 if (*cp++ == 'd' | |
3723 && *cp++ == 'e' | |
3724 && *cp++ == 'f' && isspace (*cp++)) | |
3725 { | |
3726 cp = skip_spaces (cp); | |
3727 while (*cp != '\0' && !isspace (*cp) && *cp != '(' && *cp != ':') | |
3728 cp++; | |
3729 pfnote ((char *) NULL, TRUE, | |
3730 lb.buffer, cp - lb.buffer + 1, lineno, linecharno); | |
3731 } | |
3732 | |
3733 cp = lb.buffer; | |
3734 if (*cp++ == 'c' | |
3735 && *cp++ == 'l' | |
3736 && *cp++ == 'a' | |
3737 && *cp++ == 's' | |
3738 && *cp++ == 's' && isspace (*cp++)) | |
3739 { | |
3740 cp = skip_spaces (cp); | |
3741 while (*cp != '\0' && !isspace (*cp) && *cp != '(' && *cp != ':') | |
3742 cp++; | |
3743 pfnote ((char *) NULL, TRUE, | |
3744 lb.buffer, cp - lb.buffer + 1, lineno, linecharno); | |
3745 } | |
3746 } | |
3747 } | |
3748 | |
3749 /* Idea by Corny de Souza | |
3750 * Cobol tag functions | |
3751 * We could look for anything that could be a paragraph name. | |
3752 * i.e. anything that starts in column 8 is one word and ends in a full stop. | |
3753 */ | |
3754 void | |
3755 Cobol_paragraphs (inf) | |
3756 FILE *inf; | |
3757 { | |
3758 register char *bp, *ep; | |
3759 | |
3760 LOOP_ON_INPUT_LINES (inf, lb, bp) | |
3761 { | |
3762 if (lb.len < 9) | |
3763 continue; | |
3764 bp += 8; | |
3765 | |
3766 /* If eoln, compiler option or comment ignore whole line. */ | |
3767 if (bp[-1] != ' ' || !isalnum (bp[0])) | |
3768 continue; | |
3769 | |
3770 for (ep = bp; isalnum (*ep) || *ep == '-'; ep++) | |
3771 continue; | |
3772 if (*ep++ == '.') | |
3773 pfnote ((CTAGS) ? savenstr (bp, ep-bp) : NULL, TRUE, | |
3774 lb.buffer, ep - lb.buffer + 1, lineno, linecharno); | |
3775 } | |
3776 } | |
3777 | |
3778 /* Added by Mosur Mohan, 4/22/88 */ | |
3779 /* Pascal parsing */ | |
3780 | |
3781 /* | |
3782 * Locates tags for procedures & functions. Doesn't do any type- or | |
3783 * var-definitions. It does look for the keyword "extern" or | |
3784 * "forward" immediately following the procedure statement; if found, | |
3785 * the tag is skipped. | |
3786 */ | |
3787 void | |
3788 Pascal_functions (inf) | |
3789 FILE *inf; | |
3790 { | |
3791 linebuffer tline; /* mostly copied from C_entries */ | |
3792 long save_lcno; | |
3793 int save_lineno, save_len; | |
3794 char c, *cp, *namebuf; | |
3795 | |
3796 bool /* each of these flags is TRUE iff: */ | |
3797 incomment, /* point is inside a comment */ | |
3798 inquote, /* point is inside '..' string */ | |
3799 get_tagname, /* point is after PROCEDURE/FUNCTION | |
3800 keyword, so next item = potential tag */ | |
3801 found_tag, /* point is after a potential tag */ | |
3802 inparms, /* point is within parameter-list */ | |
3803 verify_tag; /* point has passed the parm-list, so the | |
3804 next token will determine whether this | |
3805 is a FORWARD/EXTERN to be ignored, or | |
3806 whether it is a real tag */ | |
3807 | |
3808 save_lcno = save_lineno = save_len = 0; /* keep compiler quiet */ | |
3809 namebuf = NULL; /* keep compiler quiet */ | |
3810 lineno = 0; | |
3811 charno = 0; | |
3812 dbp = lb.buffer; | |
3813 *dbp = '\0'; | |
3814 initbuffer (&tline); | |
3815 | |
3816 incomment = inquote = FALSE; | |
3817 found_tag = FALSE; /* have a proc name; check if extern */ | |
3818 get_tagname = FALSE; /* have found "procedure" keyword */ | |
3819 inparms = FALSE; /* found '(' after "proc" */ | |
3820 verify_tag = FALSE; /* check if "extern" is ahead */ | |
3821 | |
3822 | |
3823 while (!feof (inf)) /* long main loop to get next char */ | |
3824 { | |
3825 c = *dbp++; | |
3826 if (c == '\0') /* if end of line */ | |
3827 { | |
3828 lineno++; | |
3829 linecharno = charno; | |
3830 charno += readline (&lb, inf); | |
3831 dbp = lb.buffer; | |
3832 if (*dbp == '\0') | |
3833 continue; | |
3834 if (!((found_tag && verify_tag) | |
3835 || get_tagname)) | |
3836 c = *dbp++; /* only if don't need *dbp pointing | |
3837 to the beginning of the name of | |
3838 the procedure or function */ | |
3839 } | |
3840 if (incomment) | |
3841 { | |
3842 if (c == '}') /* within { } comments */ | |
3843 incomment = FALSE; | |
3844 else if (c == '*' && *dbp == ')') /* within (* *) comments */ | |
3845 { | |
3846 dbp++; | |
3847 incomment = FALSE; | |
3848 } | |
3849 continue; | |
3850 } | |
3851 else if (inquote) | |
3852 { | |
3853 if (c == '\'') | |
3854 inquote = FALSE; | |
3855 continue; | |
3856 } | |
3857 else | |
3858 switch (c) | |
3859 { | |
3860 case '\'': | |
3861 inquote = TRUE; /* found first quote */ | |
3862 continue; | |
3863 case '{': /* found open { comment */ | |
3864 incomment = TRUE; | |
3865 continue; | |
3866 case '(': | |
3867 if (*dbp == '*') /* found open (* comment */ | |
3868 { | |
3869 incomment = TRUE; | |
3870 dbp++; | |
3871 } | |
3872 else if (found_tag) /* found '(' after tag, i.e., parm-list */ | |
3873 inparms = TRUE; | |
3874 continue; | |
3875 case ')': /* end of parms list */ | |
3876 if (inparms) | |
3877 inparms = FALSE; | |
3878 continue; | |
3879 case ';': | |
3880 if (found_tag && !inparms) /* end of proc or fn stmt */ | |
3881 { | |
3882 verify_tag = TRUE; | |
3883 break; | |
3884 } | |
3885 continue; | |
3886 } | |
3887 if (found_tag && verify_tag && (*dbp != ' ')) | |
3888 { | |
3889 /* check if this is an "extern" declaration */ | |
3890 if (*dbp == '\0') | |
3891 continue; | |
3892 if (lowcase (*dbp == 'e')) | |
3893 { | |
3894 if (tail ("extern")) /* superfluous, really! */ | |
3895 { | |
3896 found_tag = FALSE; | |
3897 verify_tag = FALSE; | |
3898 } | |
3899 } | |
3900 else if (lowcase (*dbp) == 'f') | |
3901 { | |
3902 if (tail ("forward")) /* check for forward reference */ | |
3903 { | |
3904 found_tag = FALSE; | |
3905 verify_tag = FALSE; | |
3906 } | |
3907 } | |
3908 if (found_tag && verify_tag) /* not external proc, so make tag */ | |
3909 { | |
3910 found_tag = FALSE; | |
3911 verify_tag = FALSE; | |
3912 pfnote (namebuf, TRUE, | |
3913 tline.buffer, save_len, save_lineno, save_lcno); | |
3914 continue; | |
3915 } | |
3916 } | |
3917 if (get_tagname) /* grab name of proc or fn */ | |
3918 { | |
3919 if (*dbp == '\0') | |
3920 continue; | |
3921 | |
3922 /* save all values for later tagging */ | |
3923 grow_linebuffer (&tline, lb.len + 1); | |
3924 strcpy (tline.buffer, lb.buffer); | |
3925 save_lineno = lineno; | |
3926 save_lcno = linecharno; | |
3927 | |
3928 /* grab block name */ | |
3929 for (cp = dbp + 1; *cp != '\0' && !endtoken (*cp); cp++) | |
3930 continue; | |
3931 namebuf = (CTAGS) ? savenstr (dbp, cp-dbp) : NULL; | |
3932 dbp = cp; /* set dbp to e-o-token */ | |
3933 save_len = dbp - lb.buffer + 1; | |
3934 get_tagname = FALSE; | |
3935 found_tag = TRUE; | |
3936 continue; | |
3937 | |
3938 /* and proceed to check for "extern" */ | |
3939 } | |
3940 else if (!incomment && !inquote && !found_tag) | |
3941 { | |
3942 /* check for proc/fn keywords */ | |
3943 switch (lowcase (c)) | |
3944 { | |
3945 case 'p': | |
3946 if (tail ("rocedure")) /* c = 'p', dbp has advanced */ | |
3947 get_tagname = TRUE; | |
3948 continue; | |
3949 case 'f': | |
3950 if (tail ("unction")) | |
3951 get_tagname = TRUE; | |
3952 continue; | |
3953 } | |
3954 } | |
3955 } /* while not eof */ | |
3956 | |
3957 free (tline.buffer); | |
3958 } | |
3959 | |
3960 /* | |
3961 * lisp tag functions | |
3962 * look for (def or (DEF, quote or QUOTE | |
3963 */ | |
3964 int L_isdef PP ((char *strp)); | |
3965 int | |
3966 L_isdef (strp) | |
3967 register char *strp; | |
3968 { | |
3969 return ((strp[1] == 'd' || strp[1] == 'D') | |
3970 && (strp[2] == 'e' || strp[2] == 'E') | |
3971 && (strp[3] == 'f' || strp[3] == 'F')); | |
3972 } | |
3973 int L_isquote PP ((char *strp)); | |
3974 int | |
3975 L_isquote (strp) | |
3976 register char *strp; | |
3977 { | |
3978 return ((*++strp == 'q' || *strp == 'Q') | |
3979 && (*++strp == 'u' || *strp == 'U') | |
3980 && (*++strp == 'o' || *strp == 'O') | |
3981 && (*++strp == 't' || *strp == 'T') | |
3982 && (*++strp == 'e' || *strp == 'E') | |
3983 && isspace (*++strp)); | |
3984 } | |
3985 | |
3986 void L_getit PP ((void)); | |
3987 void | |
3988 L_getit () | |
3989 { | |
3990 register char *cp; | |
3991 | |
3992 if (*dbp == '\'') /* Skip prefix quote */ | |
3993 dbp++; | |
3994 else if (*dbp == '(') | |
3995 { | |
3996 if (L_isquote (dbp)) | |
3997 dbp += 7; /* Skip "(quote " */ | |
3998 else | |
3999 dbp += 1; /* Skip "(" before name in (defstruct (foo)) */ | |
4000 dbp = skip_spaces (dbp); | |
4001 } | |
4002 | |
4003 for (cp = dbp /*+1*/; | |
4004 *cp != '\0' && *cp != '(' && *cp != ' ' && *cp != ')'; | |
4005 cp++) | |
4006 continue; | |
4007 if (cp == dbp) | |
4008 return; | |
4009 | |
4010 pfnote ((CTAGS) ? savenstr (dbp, cp-dbp) : NULL, TRUE, | |
4011 lb.buffer, cp - lb.buffer + 1, lineno, linecharno); | |
4012 } | |
4013 | |
4014 void | |
4015 Lisp_functions (inf) | |
4016 FILE *inf; | |
4017 { | |
4018 LOOP_ON_INPUT_LINES (inf, lb, dbp) | |
4019 { | |
4020 if (dbp[0] == '(') | |
4021 { | |
4022 if (L_isdef (dbp)) | |
4023 { | |
4024 dbp = skip_non_spaces (dbp); | |
4025 dbp = skip_spaces (dbp); | |
4026 L_getit (); | |
4027 } | |
4028 else | |
4029 { | |
4030 /* Check for (foo::defmumble name-defined ... */ | |
4031 do | |
4032 dbp++; | |
4033 while (*dbp != '\0' && !isspace (*dbp) | |
4034 && *dbp != ':' && *dbp != '(' && *dbp != ')'); | |
4035 if (*dbp == ':') | |
4036 { | |
4037 do | |
4038 dbp++; | |
4039 while (*dbp == ':'); | |
4040 | |
4041 if (L_isdef (dbp - 1)) | |
4042 { | |
4043 dbp = skip_non_spaces (dbp); | |
4044 dbp = skip_spaces (dbp); | |
4045 L_getit (); | |
4046 } | |
4047 } | |
4048 } | |
4049 } | |
4050 } | |
4051 } | |
4052 | |
4053 /* | |
4054 * Postscript tag functions | |
4055 * Just look for lines where the first character is '/' | |
4056 * Richard Mlynarik <mly@adoc.xerox.com> | |
4057 */ | |
4058 void | |
4059 Postscript_functions (inf) | |
4060 FILE *inf; | |
4061 { | |
4062 register char *bp, *ep; | |
4063 | |
4064 LOOP_ON_INPUT_LINES (inf, lb, bp) | |
4065 { | |
4066 if (bp[0] == '/') | |
4067 { | |
4068 for (ep = bp+1; | |
4069 *ep != '\0' && *ep != ' ' && *ep != '{'; | |
4070 ep++) | |
4071 continue; | |
4072 pfnote ((CTAGS) ? savenstr (bp, ep-bp) : NULL, TRUE, | |
4073 lb.buffer, ep - lb.buffer + 1, lineno, linecharno); | |
4074 } | |
4075 } | |
4076 } | |
4077 | |
4078 | |
4079 /* | |
4080 * Scheme tag functions | |
4081 * look for (def... xyzzy | |
4082 * look for (def... (xyzzy | |
4083 * look for (def ... ((...(xyzzy .... | |
4084 * look for (set! xyzzy | |
4085 */ | |
4086 | |
4087 void get_scheme PP ((void)); | |
4088 | |
4089 void | |
4090 Scheme_functions (inf) | |
4091 FILE *inf; | |
4092 { | |
4093 LOOP_ON_INPUT_LINES (inf, lb, dbp) | |
4094 { | |
4095 if (dbp[0] == '(' | |
4096 && (dbp[1] == 'D' || dbp[1] == 'd') | |
4097 && (dbp[2] == 'E' || dbp[2] == 'e') | |
4098 && (dbp[3] == 'F' || dbp[3] == 'f')) | |
4099 { | |
4100 dbp = skip_non_spaces (dbp); | |
4101 /* Skip over open parens and white space */ | |
4102 while (isspace (*dbp) || *dbp == '(') | |
4103 dbp++; | |
4104 get_scheme (); | |
4105 } | |
4106 if (dbp[0] == '(' | |
4107 && (dbp[1] == 'S' || dbp[1] == 's') | |
4108 && (dbp[2] == 'E' || dbp[2] == 'e') | |
4109 && (dbp[3] == 'T' || dbp[3] == 't') | |
4110 && (dbp[4] == '!' || dbp[4] == '!') | |
4111 && (isspace (dbp[5]))) | |
4112 { | |
4113 dbp = skip_non_spaces (dbp); | |
4114 dbp = skip_spaces (dbp); | |
4115 get_scheme (); | |
4116 } | |
4117 } | |
4118 } | |
4119 | |
4120 void | |
4121 get_scheme () | |
4122 { | |
4123 register char *cp; | |
4124 | |
4125 if (*dbp == '\0') | |
4126 return; | |
4127 /* Go till you get to white space or a syntactic break */ | |
4128 for (cp = dbp + 1; | |
4129 *cp != '\0' && *cp != '(' && *cp != ')' && !isspace (*cp); | |
4130 cp++) | |
4131 continue; | |
4132 pfnote ((CTAGS) ? savenstr (dbp, cp-dbp) : NULL, TRUE, | |
4133 lb.buffer, cp - lb.buffer + 1, lineno, linecharno); | |
4134 } | |
4135 | |
4136 /* Find tags in TeX and LaTeX input files. */ | |
4137 | |
4138 /* TEX_toktab is a table of TeX control sequences that define tags. | |
4139 Each TEX_tabent records one such control sequence. | |
4140 CONVERT THIS TO USE THE Stab TYPE!! */ | |
4141 struct TEX_tabent | |
4142 { | |
4143 char *name; | |
4144 int len; | |
4145 }; | |
4146 | |
4147 struct TEX_tabent *TEX_toktab = NULL; /* Table with tag tokens */ | |
4148 | |
4149 /* Default set of control sequences to put into TEX_toktab. | |
4150 The value of environment var TEXTAGS is prepended to this. */ | |
4151 | |
4152 char *TEX_defenv = "\ | |
4153 :chapter:section:subsection:subsubsection:eqno:label:ref:cite:bibitem\ | |
4154 :part:appendix:entry:index"; | |
4155 | |
4156 void TEX_mode PP ((FILE *inf)); | |
4157 struct TEX_tabent *TEX_decode_env PP ((char *evarname, char *defenv)); | |
4158 int TEX_Token PP ((char *cp)); | |
4159 | |
4160 char TEX_esc = '\\'; | |
4161 char TEX_opgrp = '{'; | |
4162 char TEX_clgrp = '}'; | |
4163 | |
4164 /* | |
4165 * TeX/LaTeX scanning loop. | |
4166 */ | |
4167 void | |
4168 TeX_functions (inf) | |
4169 FILE *inf; | |
4170 { | |
4171 char *cp, *lasthit; | |
4172 register int i; | |
4173 | |
4174 /* Select either \ or ! as escape character. */ | |
4175 TEX_mode (inf); | |
4176 | |
4177 /* Initialize token table once from environment. */ | |
4178 if (!TEX_toktab) | |
4179 TEX_toktab = TEX_decode_env ("TEXTAGS", TEX_defenv); | |
4180 | |
4181 LOOP_ON_INPUT_LINES (inf, lb, cp) | |
4182 { | |
4183 lasthit = cp; | |
4184 /* Look at each esc in line. */ | |
4185 while ((cp = etags_strchr (cp, TEX_esc)) != NULL) | |
4186 { | |
4187 if (*++cp == '\0') | |
4188 break; | |
4189 linecharno += cp - lasthit; | |
4190 lasthit = cp; | |
4191 i = TEX_Token (lasthit); | |
4192 if (i >= 0) | |
4193 { | |
4194 /* We seem to include the TeX command in the tag name. | |
4195 register char *p; | |
4196 for (p = lasthit + TEX_toktab[i].len; | |
4197 *p != '\0' && *p != TEX_clgrp; | |
4198 p++) | |
4199 continue; */ | |
4200 pfnote (/*savenstr (lasthit, p-lasthit)*/ (char *)NULL, TRUE, | |
4201 lb.buffer, lb.len, lineno, linecharno); | |
4202 break; /* We only tag a line once */ | |
4203 } | |
4204 } | |
4205 } | |
4206 } | |
4207 | |
4208 #define TEX_LESC '\\' | |
4209 #define TEX_SESC '!' | |
4210 #define TEX_cmt '%' | |
4211 | |
4212 /* Figure out whether TeX's escapechar is '\\' or '!' and set grouping | |
4213 chars accordingly. */ | |
4214 void | |
4215 TEX_mode (inf) | |
4216 FILE *inf; | |
4217 { | |
4218 int c; | |
4219 | |
4220 while ((c = getc (inf)) != EOF) | |
4221 { | |
4222 /* Skip to next line if we hit the TeX comment char. */ | |
4223 if (c == TEX_cmt) | |
4224 while (c != '\n') | |
4225 c = getc (inf); | |
4226 else if (c == TEX_LESC || c == TEX_SESC ) | |
4227 break; | |
4228 } | |
4229 | |
4230 if (c == TEX_LESC) | |
4231 { | |
4232 TEX_esc = TEX_LESC; | |
4233 TEX_opgrp = '{'; | |
4234 TEX_clgrp = '}'; | |
4235 } | |
4236 else | |
4237 { | |
4238 TEX_esc = TEX_SESC; | |
4239 TEX_opgrp = '<'; | |
4240 TEX_clgrp = '>'; | |
4241 } | |
4242 rewind (inf); | |
4243 } | |
4244 | |
4245 /* Read environment and prepend it to the default string. | |
4246 Build token table. */ | |
4247 struct TEX_tabent * | |
4248 TEX_decode_env (evarname, defenv) | |
4249 char *evarname; | |
4250 char *defenv; | |
4251 { | |
4252 register char *env, *p; | |
4253 | |
4254 struct TEX_tabent *tab; | |
4255 int size, i; | |
4256 | |
4257 /* Append default string to environment. */ | |
4258 env = getenv (evarname); | |
4259 if (!env) | |
4260 env = defenv; | |
4261 else | |
4262 { | |
4263 char *oldenv = env; | |
4264 env = concat (oldenv, defenv, ""); | |
4265 free (oldenv); | |
4266 } | |
4267 | |
4268 /* Allocate a token table */ | |
4269 for (size = 1, p = env; p;) | |
4270 if ((p = etags_strchr (p, ':')) && *++p != '\0') | |
4271 size++; | |
4272 /* Add 1 to leave room for null terminator. */ | |
4273 tab = xnew (size + 1, struct TEX_tabent); | |
4274 | |
4275 /* Unpack environment string into token table. Be careful about */ | |
4276 /* zero-length strings (leading ':', "::" and trailing ':') */ | |
4277 for (i = 0; *env;) | |
4278 { | |
4279 p = etags_strchr (env, ':'); | |
4280 if (!p) /* End of environment string. */ | |
4281 p = env + strlen (env); | |
4282 if (p - env > 0) | |
4283 { /* Only non-zero strings. */ | |
4284 tab[i].name = savenstr (env, p - env); | |
4285 tab[i].len = strlen (tab[i].name); | |
4286 i++; | |
4287 } | |
4288 if (*p) | |
4289 env = p + 1; | |
4290 else | |
4291 { | |
4292 tab[i].name = NULL; /* Mark end of table. */ | |
4293 tab[i].len = 0; | |
4294 break; | |
4295 } | |
4296 } | |
4297 return tab; | |
4298 } | |
4299 | |
4300 /* If the text at CP matches one of the tag-defining TeX command names, | |
4301 return the pointer to the first occurrence of that command in TEX_toktab. | |
4302 Otherwise return -1. | |
4303 Keep the capital `T' in `token' for dumb truncating compilers | |
4304 (this distinguishes it from `TEX_toktab' */ | |
4305 int | |
4306 TEX_Token (cp) | |
4307 char *cp; | |
4308 { | |
4309 int i; | |
4310 | |
4311 for (i = 0; TEX_toktab[i].len > 0; i++) | |
4312 if (strneq (TEX_toktab[i].name, cp, TEX_toktab[i].len)) | |
4313 return i; | |
4314 return -1; | |
4315 } | |
4316 | |
4317 /* | |
4318 * Prolog support (rewritten) by Anders Lindgren, Mar. 96 | |
4319 * | |
4320 * Assumes that the predicate starts at column 0. | |
4321 * Only the first clause of a predicate is added. | |
4322 */ | |
4323 int prolog_pred PP ((char *s, char *last)); | |
4324 void prolog_skip_comment PP ((linebuffer *plb, FILE *inf)); | |
4325 int prolog_atom PP ((char *s, int pos)); | |
4326 | |
4327 void | |
4328 Prolog_functions (inf) | |
4329 FILE *inf; | |
4330 { | |
4331 char *cp, *last; | |
4332 int len; | |
4333 int allocated; | |
4334 | |
4335 allocated = 0; | |
4336 len = 0; | |
4337 last = NULL; | |
4338 | |
4339 LOOP_ON_INPUT_LINES (inf, lb, cp) | |
4340 { | |
4341 if (cp[0] == '\0') /* Empty line */ | |
4342 continue; | |
4343 else if (isspace (cp[0])) /* Not a predicate */ | |
4344 continue; | |
4345 else if (cp[0] == '/' && cp[1] == '*') /* comment. */ | |
4346 prolog_skip_comment (&lb, inf); | |
4347 else if ((len = prolog_pred (cp, last)) > 0) | |
4348 { | |
4349 /* Predicate. Store the function name so that we only | |
4350 generate a tag for the first clause. */ | |
4351 if (last == NULL) | |
4352 last = xnew(len + 1, char); | |
4353 else if (len + 1 > allocated) | |
4354 last = xrnew (last, len + 1, char); | |
4355 allocated = len + 1; | |
4356 strncpy (last, cp, len); | |
4357 last[len] = '\0'; | |
4358 } | |
4359 } | |
4360 } | |
4361 | |
4362 | |
4363 void | |
4364 prolog_skip_comment (plb, inf) | |
4365 linebuffer *plb; | |
4366 FILE *inf; | |
4367 { | |
4368 char *cp; | |
4369 | |
4370 do | |
4371 { | |
4372 for (cp = plb->buffer; *cp != '\0'; cp++) | |
4373 if (cp[0] == '*' && cp[1] == '/') | |
4374 return; | |
4375 lineno++; | |
4376 linecharno += readline (plb, inf); | |
4377 } | |
4378 while (!feof(inf)); | |
4379 } | |
4380 | |
4381 /* | |
4382 * A predicate definition is added if it matches: | |
4383 * <beginning of line><Prolog Atom><whitespace>( | |
4384 * | |
4385 * It is added to the tags database if it doesn't match the | |
4386 * name of the previous clause header. | |
4387 * | |
4388 * Return the size of the name of the predicate, or 0 if no header | |
4389 * was found. | |
4390 */ | |
4391 int | |
4392 prolog_pred (s, last) | |
4393 char *s; | |
4394 char *last; /* Name of last clause. */ | |
4395 { | |
4396 int pos; | |
4397 int len; | |
4398 | |
4399 pos = prolog_atom (s, 0); | |
4400 if (pos < 1) | |
4401 return 0; | |
4402 | |
4403 len = pos; | |
4404 pos = skip_spaces (s + pos) - s; | |
4405 | |
4406 if ((s[pos] == '(') || (s[pos] == '.')) | |
4407 { | |
4408 if (s[pos] == '(') | |
4409 pos++; | |
4410 | |
4411 /* Save only the first clause. */ | |
4412 if (last == NULL | |
647 | 4413 || len != (int) strlen (last) |
428 | 4414 || !strneq (s, last, len)) |
4415 { | |
4416 pfnote ((CTAGS) ? savenstr (s, len) : NULL, TRUE, | |
4417 s, pos, lineno, linecharno); | |
4418 return len; | |
4419 } | |
4420 } | |
4421 return 0; | |
4422 } | |
4423 | |
4424 /* | |
4425 * Consume a Prolog atom. | |
4426 * Return the number of bytes consumed, or -1 if there was an error. | |
4427 * | |
4428 * A prolog atom, in this context, could be one of: | |
4429 * - An alphanumeric sequence, starting with a lower case letter. | |
4430 * - A quoted arbitrary string. Single quotes can escape themselves. | |
4431 * Backslash quotes everything. | |
4432 */ | |
4433 int | |
4434 prolog_atom (s, pos) | |
4435 char *s; | |
4436 int pos; | |
4437 { | |
4438 int origpos; | |
4439 | |
4440 origpos = pos; | |
4441 | |
4442 if (islower(s[pos]) || (s[pos] == '_')) | |
4443 { | |
4444 /* The atom is unquoted. */ | |
4445 pos++; | |
4446 while (isalnum(s[pos]) || (s[pos] == '_')) | |
4447 { | |
4448 pos++; | |
4449 } | |
4450 return pos - origpos; | |
4451 } | |
4452 else if (s[pos] == '\'') | |
4453 { | |
4454 pos++; | |
4455 | |
4456 while (1) | |
4457 { | |
4458 if (s[pos] == '\'') | |
4459 { | |
4460 pos++; | |
4461 if (s[pos] != '\'') | |
4462 break; | |
4463 pos++; /* A double quote */ | |
4464 } | |
4465 else if (s[pos] == '\0') | |
4466 /* Multiline quoted atoms are ignored. */ | |
4467 return -1; | |
4468 else if (s[pos] == '\\') | |
4469 { | |
4470 if (s[pos+1] == '\0') | |
4471 return -1; | |
4472 pos += 2; | |
4473 } | |
4474 else | |
4475 pos++; | |
4476 } | |
4477 return pos - origpos; | |
4478 } | |
4479 else | |
4480 return -1; | |
4481 } | |
4482 | |
4483 /* | |
4484 * Support for Erlang -- Anders Lindgren, Feb 1996. | |
4485 * | |
4486 * Generates tags for functions, defines, and records. | |
4487 * | |
4488 * Assumes that Erlang functions start at column 0. | |
4489 */ | |
4490 int erlang_func PP ((char *s, char *last)); | |
4491 void erlang_attribute PP ((char *s)); | |
4492 int erlang_atom PP ((char *s, int pos)); | |
4493 | |
4494 void | |
4495 Erlang_functions (inf) | |
4496 FILE *inf; | |
4497 { | |
4498 char *cp, *last; | |
4499 int len; | |
4500 int allocated; | |
4501 | |
4502 allocated = 0; | |
4503 len = 0; | |
4504 last = NULL; | |
4505 | |
4506 LOOP_ON_INPUT_LINES (inf, lb, cp) | |
4507 { | |
4508 if (cp[0] == '\0') /* Empty line */ | |
4509 continue; | |
4510 else if (isspace (cp[0])) /* Not function nor attribute */ | |
4511 continue; | |
4512 else if (cp[0] == '%') /* comment */ | |
4513 continue; | |
4514 else if (cp[0] == '"') /* Sometimes, strings start in column one */ | |
4515 continue; | |
4516 else if (cp[0] == '-') /* attribute, e.g. "-define" */ | |
4517 { | |
4518 erlang_attribute (cp); | |
4519 last = NULL; | |
4520 } | |
4521 else if ((len = erlang_func (cp, last)) > 0) | |
4522 { | |
4523 /* | |
4524 * Function. Store the function name so that we only | |
4525 * generates a tag for the first clause. | |
4526 */ | |
4527 if (last == NULL) | |
4528 last = xnew (len + 1, char); | |
4529 else if (len + 1 > allocated) | |
4530 last = xrnew (last, len + 1, char); | |
4531 allocated = len + 1; | |
4532 strncpy (last, cp, len); | |
4533 last[len] = '\0'; | |
4534 } | |
4535 } | |
4536 } | |
4537 | |
4538 | |
4539 /* | |
4540 * A function definition is added if it matches: | |
4541 * <beginning of line><Erlang Atom><whitespace>( | |
4542 * | |
4543 * It is added to the tags database if it doesn't match the | |
4544 * name of the previous clause header. | |
4545 * | |
4546 * Return the size of the name of the function, or 0 if no function | |
4547 * was found. | |
4548 */ | |
4549 int | |
4550 erlang_func (s, last) | |
4551 char *s; | |
4552 char *last; /* Name of last clause. */ | |
4553 { | |
4554 int pos; | |
4555 int len; | |
4556 | |
4557 pos = erlang_atom (s, 0); | |
4558 if (pos < 1) | |
4559 return 0; | |
4560 | |
4561 len = pos; | |
4562 pos = skip_spaces (s + pos) - s; | |
4563 | |
4564 /* Save only the first clause. */ | |
4565 if (s[pos++] == '(' | |
4566 && (last == NULL | |
647 | 4567 || len != (int) strlen (last) |
428 | 4568 || !strneq (s, last, len))) |
4569 { | |
4570 pfnote ((CTAGS) ? savenstr (s, len) : NULL, TRUE, | |
4571 s, pos, lineno, linecharno); | |
4572 return len; | |
4573 } | |
4574 | |
4575 return 0; | |
4576 } | |
4577 | |
4578 | |
4579 /* | |
4580 * Handle attributes. Currently, tags are generated for defines | |
4581 * and records. | |
4582 * | |
4583 * They are on the form: | |
4584 * -define(foo, bar). | |
4585 * -define(Foo(M, N), M+N). | |
4586 * -record(graph, {vtab = notable, cyclic = true}). | |
4587 */ | |
4588 void | |
4589 erlang_attribute (s) | |
4590 char *s; | |
4591 { | |
4592 int pos; | |
4593 int len; | |
4594 | |
4595 if (strneq (s, "-define", 7) || strneq (s, "-record", 7)) | |
4596 { | |
4597 pos = skip_spaces (s + 7) - s; | |
4598 if (s[pos++] == '(') | |
4599 { | |
4600 pos = skip_spaces (s + pos) - s; | |
4601 len = erlang_atom (s, pos); | |
4602 if (len != 0) | |
4603 pfnote ((CTAGS) ? savenstr (& s[pos], len) : NULL, TRUE, | |
4604 s, pos + len, lineno, linecharno); | |
4605 } | |
4606 } | |
4607 return; | |
4608 } | |
4609 | |
4610 | |
4611 /* | |
4612 * Consume an Erlang atom (or variable). | |
4613 * Return the number of bytes consumed, or -1 if there was an error. | |
4614 */ | |
4615 int | |
4616 erlang_atom (s, pos) | |
4617 char *s; | |
4618 int pos; | |
4619 { | |
4620 int origpos; | |
4621 | |
4622 origpos = pos; | |
4623 | |
4624 if (isalpha (s[pos]) || s[pos] == '_') | |
4625 { | |
4626 /* The atom is unquoted. */ | |
4627 pos++; | |
4628 while (isalnum (s[pos]) || s[pos] == '_') | |
4629 pos++; | |
4630 return pos - origpos; | |
4631 } | |
4632 else if (s[pos] == '\'') | |
4633 { | |
4634 pos++; | |
4635 | |
4636 while (1) | |
4637 { | |
4638 if (s[pos] == '\'') | |
4639 { | |
4640 pos++; | |
4641 break; | |
4642 } | |
4643 else if (s[pos] == '\0') | |
4644 /* Multiline quoted atoms are ignored. */ | |
4645 return -1; | |
4646 else if (s[pos] == '\\') | |
4647 { | |
4648 if (s[pos+1] == '\0') | |
4649 return -1; | |
4650 pos += 2; | |
4651 } | |
4652 else | |
4653 pos++; | |
4654 } | |
4655 return pos - origpos; | |
4656 } | |
4657 else | |
4658 return -1; | |
4659 } | |
4660 | |
4661 #ifdef ETAGS_REGEXPS | |
4662 | |
4663 /* Take a string like "/blah/" and turn it into "blah", making sure | |
4664 that the first and last characters are the same, and handling | |
4665 quoted separator characters. Actually, stops on the occurrence of | |
4666 an unquoted separator. Also turns "\t" into a Tab character. | |
4667 Returns pointer to terminating separator. Works in place. Null | |
4668 terminates name string. */ | |
4669 char * scan_separators PP ((char *name)); | |
4670 char * | |
4671 scan_separators (name) | |
4672 char *name; | |
4673 { | |
4674 char sep = name[0]; | |
4675 char *copyto = name; | |
4676 bool quoted = FALSE; | |
4677 | |
4678 for (++name; *name != '\0'; ++name) | |
4679 { | |
4680 if (quoted) | |
4681 { | |
4682 if (*name == 't') | |
4683 *copyto++ = '\t'; | |
4684 else if (*name == sep) | |
4685 *copyto++ = sep; | |
4686 else | |
4687 { | |
4688 /* Something else is quoted, so preserve the quote. */ | |
4689 *copyto++ = '\\'; | |
4690 *copyto++ = *name; | |
4691 } | |
4692 quoted = FALSE; | |
4693 } | |
4694 else if (*name == '\\') | |
4695 quoted = TRUE; | |
4696 else if (*name == sep) | |
4697 break; | |
4698 else | |
4699 *copyto++ = *name; | |
4700 } | |
4701 | |
4702 /* Terminate copied string. */ | |
4703 *copyto = '\0'; | |
4704 return name; | |
4705 } | |
4706 | |
4707 /* Look at the argument of --regex or --no-regex and do the right | |
4708 thing. Same for each line of a regexp file. */ | |
4709 void | |
4710 analyse_regex (regex_arg) | |
4711 char *regex_arg; | |
4712 { | |
4713 if (regex_arg == NULL) | |
4714 free_patterns (); /* --no-regex: remove existing regexps */ | |
4715 | |
4716 /* A real --regexp option or a line in a regexp file. */ | |
4717 switch (regex_arg[0]) | |
4718 { | |
4719 /* Comments in regexp file or null arg to --regex. */ | |
4720 case '\0': | |
4721 case ' ': | |
4722 case '\t': | |
4723 break; | |
4724 | |
4725 /* Read a regex file. This is recursive and may result in a | |
4726 loop, which will stop when the file descriptors are exhausted. */ | |
4727 case '@': | |
4728 { | |
4729 FILE *regexfp; | |
4730 linebuffer regexbuf; | |
4731 char *regexfile = regex_arg + 1; | |
4732 | |
4733 /* regexfile is a file containing regexps, one per line. */ | |
4734 regexfp = fopen (regexfile, "r"); | |
4735 if (regexfp == NULL) | |
4736 { | |
4737 pfatal (regexfile); | |
4738 return; | |
4739 } | |
4740 initbuffer (®exbuf); | |
4741 while (readline_internal (®exbuf, regexfp) > 0) | |
4742 analyse_regex (regexbuf.buffer); | |
4743 free (regexbuf.buffer); | |
4744 fclose (regexfp); | |
4745 } | |
4746 break; | |
4747 | |
4748 /* Regexp to be used for a specific language only. */ | |
4749 case '{': | |
4750 { | |
4751 language *lang; | |
4752 char *lang_name = regex_arg + 1; | |
4753 char *cp; | |
4754 | |
4755 for (cp = lang_name; *cp != '}'; cp++) | |
4756 if (*cp == '\0') | |
4757 { | |
4758 error ("unterminated language name in regex: %s", regex_arg); | |
4759 return; | |
4760 } | |
4761 *cp = '\0'; | |
4762 lang = get_language_from_name (lang_name); | |
4763 if (lang == NULL) | |
4764 return; | |
4765 add_regex (cp + 1, lang); | |
4766 } | |
4767 break; | |
4768 | |
4769 /* Regexp to be used for any language. */ | |
4770 default: | |
4771 add_regex (regex_arg, NULL); | |
4772 break; | |
4773 } | |
4774 } | |
4775 | |
4776 /* Turn a name, which is an ed-style (but Emacs syntax) regular | |
4777 expression, into a real regular expression by compiling it. */ | |
4778 void | |
4779 add_regex (regexp_pattern, lang) | |
4780 char *regexp_pattern; | |
4781 language *lang; | |
4782 { | |
4783 char *name; | |
4784 const char *err; | |
4785 struct re_pattern_buffer *patbuf; | |
4786 pattern *pp; | |
4787 | |
4788 | |
4789 if (regexp_pattern[strlen(regexp_pattern)-1] != regexp_pattern[0]) | |
4790 { | |
4791 error ("%s: unterminated regexp", regexp_pattern); | |
4792 return; | |
4793 } | |
4794 name = scan_separators (regexp_pattern); | |
4795 if (regexp_pattern[0] == '\0') | |
4796 { | |
4797 error ("null regexp", (char *)NULL); | |
4798 return; | |
4799 } | |
4800 (void) scan_separators (name); | |
4801 | |
4802 patbuf = xnew (1, struct re_pattern_buffer); | |
4803 patbuf->translate = NULL; | |
4804 patbuf->fastmap = NULL; | |
4805 patbuf->buffer = NULL; | |
4806 patbuf->allocated = 0; | |
4807 | |
4808 err = re_compile_pattern (regexp_pattern, strlen (regexp_pattern), patbuf); | |
4809 if (err != NULL) | |
4810 { | |
4811 error ("%s while compiling pattern", err); | |
4812 return; | |
4813 } | |
4814 | |
4815 pp = p_head; | |
4816 p_head = xnew (1, pattern); | |
4817 p_head->regex = savestr (regexp_pattern); | |
4818 p_head->p_next = pp; | |
4819 p_head->language = lang; | |
4820 p_head->pattern = patbuf; | |
4821 p_head->name_pattern = savestr (name); | |
4822 p_head->error_signaled = FALSE; | |
4823 } | |
4824 | |
4825 /* | |
4826 * Do the substitutions indicated by the regular expression and | |
4827 * arguments. | |
4828 */ | |
4829 char * substitute PP ((char *in, char *out, struct re_registers *regs)); | |
4830 char * | |
4831 substitute (in, out, regs) | |
4832 char *in, *out; | |
4833 struct re_registers *regs; | |
4834 { | |
4835 char *result, *t; | |
4836 int size, dig, diglen; | |
4837 | |
4838 result = NULL; | |
4839 size = strlen (out); | |
4840 | |
4841 /* Pass 1: figure out how much to allocate by finding all \N strings. */ | |
4842 if (out[size - 1] == '\\') | |
4843 fatal ("pattern error in \"%s\"", out); | |
4844 for (t = etags_strchr (out, '\\'); | |
4845 t != NULL; | |
4846 t = etags_strchr (t + 2, '\\')) | |
4847 if (isdigit (t[1])) | |
4848 { | |
4849 dig = t[1] - '0'; | |
4850 diglen = regs->end[dig] - regs->start[dig]; | |
4851 size += diglen - 2; | |
4852 } | |
4853 else | |
4854 size -= 1; | |
4855 | |
4856 /* Allocate space and do the substitutions. */ | |
4857 result = xnew (size + 1, char); | |
4858 | |
4859 for (t = result; *out != '\0'; out++) | |
4860 if (*out == '\\' && isdigit (*++out)) | |
4861 { | |
4862 /* Using "dig2" satisfies my debugger. Bleah. */ | |
4863 dig = *out - '0'; | |
4864 diglen = regs->end[dig] - regs->start[dig]; | |
4865 strncpy (t, in + regs->start[dig], diglen); | |
4866 t += diglen; | |
4867 } | |
4868 else | |
4869 *t++ = *out; | |
4870 *t = '\0'; | |
4871 | |
5820
b3824b7f5627
Some changes to eliminate warnings with Apple clang version 1.7.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5406
diff
changeset
|
4872 if (DEBUG && (t > result + size || t - result != (int) (strlen (result)))) |
428 | 4873 abort (); |
4874 | |
4875 return result; | |
4876 } | |
4877 | |
4878 /* Deallocate all patterns. */ | |
4879 void | |
4880 free_patterns () | |
4881 { | |
4882 pattern *pp; | |
4883 while (p_head != NULL) | |
4884 { | |
4885 pp = p_head->p_next; | |
4886 free (p_head->regex); | |
4887 free (p_head->name_pattern); | |
4888 free (p_head); | |
4889 p_head = pp; | |
4890 } | |
4891 return; | |
4892 } | |
4893 | |
4894 #endif /* ETAGS_REGEXPS */ | |
4895 /* Initialize a linebuffer for use */ | |
4896 void | |
4897 initbuffer (lbp) | |
4898 linebuffer *lbp; | |
4899 { | |
4900 lbp->size = 200; | |
4901 lbp->buffer = xnew (200, char); | |
4902 } | |
4903 | |
4904 /* | |
4905 * Read a line of text from `stream' into `lbp', excluding the | |
4906 * newline or CR-NL, if any. Return the number of characters read from | |
4907 * `stream', which is the length of the line including the newline. | |
4908 * | |
4909 * On DOS or Windows we do not count the CR character, if any, before the | |
4910 * NL, in the returned length; this mirrors the behavior of emacs on those | |
4911 * platforms (for text files, it translates CR-NL to NL as it reads in the | |
4912 * file). | |
4913 */ | |
4914 long | |
4915 readline_internal (lbp, stream) | |
4916 linebuffer *lbp; | |
4917 register FILE *stream; | |
4918 { | |
4919 char *buffer = lbp->buffer; | |
4920 register char *p = lbp->buffer; | |
4921 register char *pend; | |
4922 int chars_deleted; | |
4923 | |
4924 pend = p + lbp->size; /* Separate to avoid 386/IX compiler bug. */ | |
4925 | |
4926 while (1) | |
4927 { | |
4928 register int c = getc (stream); | |
4929 if (p == pend) | |
4930 { | |
4931 /* We're at the end of linebuffer: expand it. */ | |
4932 lbp->size *= 2; | |
4933 buffer = xrnew (buffer, lbp->size, char); | |
4934 p += buffer - lbp->buffer; | |
4935 pend = buffer + lbp->size; | |
4936 lbp->buffer = buffer; | |
4937 } | |
4938 if (c == EOF) | |
4939 { | |
4940 *p = '\0'; | |
4941 chars_deleted = 0; | |
4942 break; | |
4943 } | |
4944 if (c == '\n') | |
4945 { | |
4946 if (p > buffer && p[-1] == '\r') | |
4947 { | |
4948 p -= 1; | |
442 | 4949 #ifdef WIN32_NATIVE |
428 | 4950 /* Assume CRLF->LF translation will be performed by Emacs |
4951 when loading this file, so CRs won't appear in the buffer. | |
4952 It would be cleaner to compensate within Emacs; | |
4953 however, Emacs does not know how many CRs were deleted | |
4954 before any given point in the file. */ | |
4955 chars_deleted = 1; | |
4956 #else | |
4957 chars_deleted = 2; | |
4958 #endif | |
4959 } | |
4960 else | |
4961 { | |
4962 chars_deleted = 1; | |
4963 } | |
4964 *p = '\0'; | |
4965 break; | |
4966 } | |
4967 *p++ = c; | |
4968 } | |
4969 lbp->len = p - buffer; | |
4970 | |
4971 return lbp->len + chars_deleted; | |
4972 } | |
4973 | |
4974 /* | |
4975 * Like readline_internal, above, but in addition try to match the | |
4976 * input line against relevant regular expressions. | |
4977 */ | |
4978 long | |
4979 readline (lbp, stream) | |
4980 linebuffer *lbp; | |
4981 FILE *stream; | |
4982 { | |
4983 /* Read new line. */ | |
4984 long result = readline_internal (lbp, stream); | |
4985 #ifdef ETAGS_REGEXPS | |
4986 int match; | |
4987 pattern *pp; | |
4988 | |
4989 /* Match against relevant patterns. */ | |
4990 if (lbp->len > 0) | |
4991 for (pp = p_head; pp != NULL; pp = pp->p_next) | |
4992 { | |
4993 /* Only use generic regexps or those for the current language. */ | |
4994 if (pp->language != NULL && pp->language != curlang) | |
4995 continue; | |
4996 | |
4997 match = re_match (pp->pattern, lbp->buffer, lbp->len, 0, &pp->regs); | |
4998 switch (match) | |
4999 { | |
5000 case -2: | |
5001 /* Some error. */ | |
5002 if (!pp->error_signaled) | |
5003 { | |
5004 error ("error while matching \"%s\"", pp->regex); | |
5005 pp->error_signaled = TRUE; | |
5006 } | |
5007 break; | |
5008 case -1: | |
5009 /* No match. */ | |
5010 break; | |
5011 default: | |
5012 /* Match occurred. Construct a tag. */ | |
5013 if (pp->name_pattern[0] != '\0') | |
5014 { | |
5015 /* Make a named tag. */ | |
5016 char *name = substitute (lbp->buffer, | |
5017 pp->name_pattern, &pp->regs); | |
5018 if (name != NULL) | |
5019 pfnote (name, TRUE, lbp->buffer, match, lineno, linecharno); | |
5020 } | |
5021 else | |
5022 { | |
5023 /* Make an unnamed tag. */ | |
5024 pfnote ((char *)NULL, TRUE, | |
5025 lbp->buffer, match, lineno, linecharno); | |
5026 } | |
5027 break; | |
5028 } | |
5029 } | |
5030 #endif /* ETAGS_REGEXPS */ | |
5031 | |
5032 return result; | |
5033 } | |
5034 | |
5035 /* | |
5036 * Return a pointer to a space of size strlen(cp)+1 allocated | |
5037 * with xnew where the string CP has been copied. | |
5038 */ | |
5039 char * | |
5040 savestr (cp) | |
5041 char *cp; | |
5042 { | |
5043 return savenstr (cp, strlen (cp)); | |
5044 } | |
5045 | |
5046 /* | |
5047 * Return a pointer to a space of size LEN+1 allocated with xnew where | |
5048 * the string CP has been copied for at most the first LEN characters. | |
5049 */ | |
5050 char * | |
5051 savenstr (cp, len) | |
5052 char *cp; | |
5053 int len; | |
5054 { | |
5055 register char *dp; | |
5056 | |
5057 dp = xnew (len + 1, char); | |
5058 strncpy (dp, cp, len); | |
5059 dp[len] = '\0'; | |
5060 return dp; | |
5061 } | |
5062 | |
5063 /* | |
5064 * Return the ptr in sp at which the character c last | |
5065 * appears; NULL if not found | |
5066 * | |
5067 * Identical to System V strrchr, included for portability. | |
5068 */ | |
5069 char * | |
5070 etags_strrchr (sp, c) | |
5071 register char *sp; | |
5072 register int c; | |
5073 { | |
5074 register char *r; | |
5075 | |
5076 r = NULL; | |
5077 do | |
5078 { | |
5079 if (*sp == c) | |
5080 r = sp; | |
5081 } while (*sp++); | |
5082 return r; | |
5083 } | |
5084 | |
5085 | |
5086 /* | |
5087 * Return the ptr in sp at which the character c first | |
5088 * appears; NULL if not found | |
5089 * | |
5090 * Identical to System V strchr, included for portability. | |
5091 */ | |
5092 char * | |
5093 etags_strchr (sp, c) | |
5094 register char *sp; | |
5095 register int c; | |
5096 { | |
5097 do | |
5098 { | |
5099 if (*sp == c) | |
5100 return sp; | |
5101 } while (*sp++); | |
5102 return NULL; | |
5103 } | |
5104 | |
5105 /* Skip spaces, return new pointer. */ | |
5106 char * | |
5107 skip_spaces (cp) | |
5108 char *cp; | |
5109 { | |
5110 while (isspace (*cp)) /* isspace('\0')==FALSE */ | |
5111 cp++; | |
5112 return cp; | |
5113 } | |
5114 | |
5115 /* Skip non spaces, return new pointer. */ | |
5116 char * | |
5117 skip_non_spaces (cp) | |
5118 char *cp; | |
5119 { | |
5120 while (!iswhite (*cp)) /* iswhite('\0')==TRUE */ | |
5121 cp++; | |
5122 return cp; | |
5123 } | |
5124 | |
5125 /* Print error message and exit. */ | |
5126 void | |
5127 fatal (s1, s2) | |
5128 char *s1, *s2; | |
5129 { | |
5130 error (s1, s2); | |
5131 exit (BAD); | |
5132 } | |
5133 | |
5134 void | |
5135 pfatal (s1) | |
5136 char *s1; | |
5137 { | |
5138 perror (s1); | |
5139 exit (BAD); | |
5140 } | |
5141 | |
5142 void | |
5143 suggest_asking_for_help () | |
5144 { | |
5145 fprintf (stderr, "\tTry `%s %s' for a complete list of options.\n", | |
5146 progname, | |
5147 #ifdef LONG_OPTIONS | |
5148 "--help" | |
5149 #else | |
5150 "-h" | |
5151 #endif | |
5152 ); | |
5153 exit (BAD); | |
5154 } | |
5155 | |
5156 /* Print error message. `s1' is printf control string, `s2' is arg for it. */ | |
5157 void | |
5158 error (s1, s2) | |
5159 const char *s1, *s2; | |
5160 { | |
5161 fprintf (stderr, "%s: ", progname); | |
5162 fprintf (stderr, s1, s2); | |
5163 fprintf (stderr, "\n"); | |
5164 } | |
5165 | |
5166 /* Return a newly-allocated string whose contents | |
5167 concatenate those of s1, s2, s3. */ | |
5168 char * | |
5169 concat (s1, s2, s3) | |
5170 char *s1, *s2, *s3; | |
5171 { | |
5172 int len1 = strlen (s1), len2 = strlen (s2), len3 = strlen (s3); | |
5173 char *result = xnew (len1 + len2 + len3 + 1, char); | |
5174 | |
5175 strcpy (result, s1); | |
5176 strcpy (result + len1, s2); | |
5177 strcpy (result + len1 + len2, s3); | |
5178 result[len1 + len2 + len3] = '\0'; | |
5179 | |
5180 return result; | |
5181 } | |
5182 | |
5183 /* Does the same work as the system V getcwd, but does not need to | |
5184 guess the buffer size in advance. */ | |
5185 char * | |
5186 etags_getcwd () | |
5187 { | |
5188 #ifdef HAVE_GETCWD | |
5189 int bufsize = 200; | |
5190 char *path = xnew (bufsize, char); | |
5191 | |
5192 while (getcwd (path, bufsize) == NULL) | |
5193 { | |
5194 if (errno != ERANGE) | |
5195 pfatal ("getcwd"); | |
5196 bufsize *= 2; | |
5197 free (path); | |
5198 path = xnew (bufsize, char); | |
5199 } | |
5200 | |
5201 canonicalize_filename (path); | |
5202 return path; | |
5203 | |
5204 #else /* not HAVE_GETCWD */ | |
5205 linebuffer path; | |
5206 FILE *pipe; | |
5207 | |
5208 initbuffer (&path); | |
5209 pipe = (FILE *) popen ("pwd 2>/dev/null", "r"); | |
5210 if (pipe == NULL || readline_internal (&path, pipe) == 0) | |
5211 pfatal ("pwd"); | |
5212 pclose (pipe); | |
5213 | |
5214 return path.buffer; | |
5215 #endif /* not HAVE_GETCWD */ | |
5216 } | |
5217 | |
5218 /* Return a newly allocated string containing the file name of FILE | |
5219 relative to the absolute directory DIR (which should end with a slash). */ | |
5220 char * | |
5221 relative_filename (file, dir) | |
5222 char *file, *dir; | |
5223 { | |
5224 char *fp, *dp, *afn, *res; | |
5225 int i; | |
5226 | |
5227 /* Find the common root of file and dir (with a trailing slash). */ | |
5228 afn = absolute_filename (file, cwd); | |
5229 fp = afn; | |
5230 dp = dir; | |
5231 while (*fp++ == *dp++) | |
5232 continue; | |
5233 fp--, dp--; /* back to the first differing char */ | |
5234 do /* look at the equal chars until '/' */ | |
5235 fp--, dp--; | |
5236 while (*fp != '/'); | |
5237 | |
5238 /* Build a sequence of "../" strings for the resulting relative file name. */ | |
5239 i = 0; | |
5240 while ((dp = etags_strchr (dp + 1, '/')) != NULL) | |
5241 i += 1; | |
5242 res = xnew (3*i + strlen (fp + 1) + 1, char); | |
5243 res[0] = '\0'; | |
5244 while (i-- > 0) | |
5245 strcat (res, "../"); | |
5246 | |
5247 /* Add the file name relative to the common root of file and dir. */ | |
5248 strcat (res, fp + 1); | |
5249 free (afn); | |
5250 | |
5251 return res; | |
5252 } | |
5253 | |
5254 /* Return a newly allocated string containing the absolute file name | |
5255 of FILE given DIR (which should end with a slash). */ | |
5256 char * | |
5257 absolute_filename (file, dir) | |
5258 char *file, *dir; | |
5259 { | |
5260 char *slashp, *cp, *res; | |
5261 | |
5262 if (filename_is_absolute (file)) | |
5263 res = savestr (file); | |
442 | 5264 #ifdef WIN32_NATIVE |
428 | 5265 /* We don't support non-absolute file names with a drive |
5266 letter, like `d:NAME' (it's too much hassle). */ | |
5267 else if (file[1] == ':') | |
5268 fatal ("%s: relative file names with drive letters not supported", file); | |
5269 #endif | |
5270 else | |
5271 res = concat (dir, file, ""); | |
5272 | |
5273 /* Delete the "/dirname/.." and "/." substrings. */ | |
5274 slashp = etags_strchr (res, '/'); | |
5275 while (slashp != NULL && slashp[0] != '\0') | |
5276 { | |
5277 if (slashp[1] == '.') | |
5278 { | |
5279 if (slashp[2] == '.' | |
5280 && (slashp[3] == '/' || slashp[3] == '\0')) | |
5281 { | |
5282 cp = slashp; | |
5283 do | |
5284 cp--; | |
5285 while (cp >= res && !filename_is_absolute (cp)); | |
5286 if (cp < res) | |
5287 cp = slashp; /* the absolute name begins with "/.." */ | |
442 | 5288 #ifdef WIN32_NATIVE |
5289 /* Under Windows we get `d:/NAME' as absolute | |
428 | 5290 file name, so the luser could say `d:/../NAME'. |
5291 We silently treat this as `d:/NAME'. */ | |
5292 else if (cp[0] != '/') | |
5293 cp = slashp; | |
5294 #endif | |
5295 strcpy (cp, slashp + 3); | |
5296 slashp = cp; | |
5297 continue; | |
5298 } | |
5299 else if (slashp[2] == '/' || slashp[2] == '\0') | |
5300 { | |
5301 strcpy (slashp, slashp + 2); | |
5302 continue; | |
5303 } | |
5304 } | |
5305 | |
5306 slashp = etags_strchr (slashp + 1, '/'); | |
5307 } | |
5308 | |
5309 if (res[0] == '\0') | |
5310 return savestr ("/"); | |
5311 else | |
5312 return res; | |
5313 } | |
5314 | |
5315 /* Return a newly allocated string containing the absolute | |
5316 file name of dir where FILE resides given DIR (which should | |
5317 end with a slash). */ | |
5318 char * | |
5319 absolute_dirname (file, dir) | |
5320 char *file, *dir; | |
5321 { | |
5322 char *slashp, *res; | |
5323 char save; | |
5324 | |
5325 canonicalize_filename (file); | |
5326 slashp = etags_strrchr (file, '/'); | |
5327 if (slashp == NULL) | |
5328 return savestr (dir); | |
5329 save = slashp[1]; | |
5330 slashp[1] = '\0'; | |
5331 res = absolute_filename (file, dir); | |
5332 slashp[1] = save; | |
5333 | |
5334 return res; | |
5335 } | |
5336 | |
5337 /* Whether the argument string is an absolute file name. The argument | |
5338 string must have been canonicalized with canonicalize_filename. */ | |
5339 bool | |
5340 filename_is_absolute (fn) | |
5341 char *fn; | |
5342 { | |
5343 return (fn[0] == '/' | |
442 | 5344 #ifdef WIN32_NATIVE |
428 | 5345 || (isalpha(fn[0]) && fn[1] == ':' && fn[2] == '/') |
5346 #endif | |
5347 ); | |
5348 } | |
5349 | |
5350 /* Translate backslashes into slashes. Works in place. */ | |
5351 void | |
5352 canonicalize_filename (fn) | |
5353 register char *fn; | |
5354 { | |
442 | 5355 #ifdef WIN32_NATIVE |
428 | 5356 for (; *fn != '\0'; fn++) |
5357 if (*fn == '\\') | |
5358 *fn = '/'; | |
5359 #else | |
5360 /* No action. */ | |
5361 #endif | |
5362 } | |
5363 | |
5364 /* Increase the size of a linebuffer. */ | |
5365 void | |
5366 grow_linebuffer (lbp, toksize) | |
5367 linebuffer *lbp; | |
5368 int toksize; | |
5369 { | |
5370 while (lbp->size < toksize) | |
5371 lbp->size *= 2; | |
5372 lbp->buffer = xrnew (lbp->buffer, lbp->size, char); | |
5373 } | |
5374 | |
5375 /* Like malloc but get fatal error if memory is exhausted. */ | |
5376 long * | |
5377 xmalloc (size) | |
5378 unsigned int size; | |
5379 { | |
5380 long *result = (long *) malloc (size); | |
5381 if (result == NULL) | |
5382 fatal ("virtual memory exhausted", (char *)NULL); | |
5383 return result; | |
5384 } | |
5385 | |
5386 long * | |
5387 xrealloc (ptr, size) | |
5388 char *ptr; | |
5389 unsigned int size; | |
5390 { | |
5391 long *result = (long *) realloc (ptr, size); | |
5392 if (result == NULL) | |
5393 fatal ("virtual memory exhausted", (char *)NULL); | |
5394 return result; | |
5395 } |