Mercurial > hg > xemacs-beta
annotate lib-src/etags.c @ 5287:cd167465bf69
More permission consistency.
| author | Stephen J. Turnbull <stephen@xemacs.org> |
|---|---|
| date | Mon, 14 Jun 2010 15:03:08 +0900 |
| parents | 0f5bee973a7b |
| children | 061f4f90f874 |
| rev | line source |
|---|---|
| 458 | 1 /* Tags file maker to go with GNU Emacs -*- coding: latin-1 -*- |
| 3972 | 2 |
| 3 Copyright (C) 1984 The Regents of the University of California | |
| 4 | |
| 5 Redistribution and use in source and binary forms, with or without | |
| 6 modification, are permitted provided that the following conditions are | |
| 7 met: | |
| 8 1. Redistributions of source code must retain the above copyright | |
| 9 notice, this list of conditions and the following disclaimer. | |
| 10 2. Redistributions in binary form must reproduce the above copyright | |
| 11 notice, this list of conditions and the following disclaimer in the | |
| 12 documentation and/or other materials provided with the | |
| 13 distribution. | |
| 14 3. Neither the name of the University nor the names of its | |
| 15 contributors may be used to endorse or promote products derived | |
| 16 from this software without specific prior written permission. | |
| 17 | |
| 18 THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' | |
| 19 AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, | |
| 20 THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR | |
| 21 PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS | |
| 22 BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR | |
| 23 CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF | |
| 24 SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR | |
| 25 BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, | |
| 26 WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE | |
| 27 OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN | |
| 28 IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. | |
| 29 | |
| 30 | |
| 31 Copyright (C) 1984, 1987, 1988, 1989, 1993, 1994, 1995, 1998, 1999, | |
| 32 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007 | |
| 33 Free Software Foundation, Inc. | |
| 34 | |
| 35 This file is not considered part of GNU Emacs. | |
| 36 | |
| 37 This program is free software; you can redistribute it and/or modify | |
| 38 it under the terms of the GNU General Public License as published by | |
| 39 the Free Software Foundation; either version 2 of the License, or | |
| 40 (at your option) any later version. | |
| 41 | |
| 42 This program is distributed in the hope that it will be useful, | |
| 43 but WITHOUT ANY WARRANTY; without even the implied warranty of | |
| 44 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
| 45 GNU General Public License for more details. | |
| 46 | |
| 47 You should have received a copy of the GNU General Public License | |
| 48 along with this program; if not, write to the Free Software Foundation, | |
| 49 Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. */ | |
| 50 | |
| 51 | |
| 52 /* NB To comply with the above BSD license, copyright information is | |
| 53 reproduced in etc/ETAGS.README. That file should be updated when the | |
| 54 above notices are. | |
| 55 | |
| 56 To the best of our knowledge, this code was originally based on the | |
| 57 ctags.c distributed with BSD4.2, which was copyrighted by the | |
| 58 University of California, as described above. */ | |
| 59 | |
| 428 | 60 |
| 61 /* | |
| 62 * Authors: | |
| 3972 | 63 * 1983 Ctags originally by Ken Arnold. |
| 64 * 1984 Fortran added by Jim Kleckner. | |
| 65 * 1984 Ed Pelegri-Llopart added C typedefs. | |
| 66 * 1985 Emacs TAGS format by Richard Stallman. | |
| 458 | 67 * 1989 Sam Kendall added C++. |
| 2225 | 68 * 1992 Joseph B. Wells improved C and C++ parsing. |
| 69 * 1993 Francesco Potort́ reorganised C and C++. | |
| 70 * 1994 Line-by-line regexp tags by Tom Tromey. | |
| 71 * 2001 Nested classes by Francesco Potort́ (concept by Mykola Dzyuba). | |
| 72 * 2002 #line directives by Francesco Potort́. | |
| 428 | 73 * |
| 2225 | 74 * Francesco Potort́ <pot@gnu.org> has maintained and improved it since 1993. |
| 428 | 75 */ |
| 76 | |
| 2325 | 77 /* |
| 78 * If you want to add support for a new language, start by looking at the LUA | |
| 79 * language, which is the simplest. Alternatively, consider shipping a | |
| 80 * configuration file containing regexp definitions for etags. | |
| 81 */ | |
| 82 | |
| 3993 | 83 char pot_etags_version[] = "@(#) pot revision number is 17.33"; |
| 428 | 84 |
| 85 #define TRUE 1 | |
| 86 #define FALSE 0 | |
| 87 | |
| 458 | 88 #ifdef DEBUG |
| 89 # undef DEBUG | |
| 90 # define DEBUG TRUE | |
| 91 #else | |
| 92 # define DEBUG FALSE | |
| 93 # define NDEBUG /* disable assert */ | |
| 428 | 94 #endif |
| 95 | |
| 96 #ifdef HAVE_CONFIG_H | |
| 97 # include <config.h> | |
| 98 /* On some systems, Emacs defines static as nothing for the sake | |
| 99 of unexec. We don't want that here since we don't use unexec. */ | |
| 100 # undef static | |
| 3524 | 101 # ifndef PTR /* for XEmacs */ |
| 2225 | 102 # define PTR void * |
| 715 | 103 # endif |
| 3524 | 104 # ifndef __P /* for XEmacs */ |
| 2225 | 105 # define __P(args) args |
| 709 | 106 # endif |
| 2325 | 107 #else /* no config.h */ |
| 709 | 108 # if defined(__STDC__) && (__STDC__ || defined(__SUNPRO_C)) |
| 109 # define __P(args) args /* use prototypes */ | |
| 110 # define PTR void * /* for generic pointers */ | |
| 2325 | 111 # else /* not standard C */ |
| 709 | 112 # define __P(args) () /* no prototypes */ |
| 113 # define const /* remove const for old compilers' sake */ | |
| 114 # define PTR long * /* don't use void* */ | |
| 531 | 115 # endif |
| 116 #endif /* !HAVE_CONFIG_H */ | |
| 428 | 117 |
| 432 | 118 #ifndef _GNU_SOURCE |
| 119 # define _GNU_SOURCE 1 /* enables some compiler checks on GNU */ | |
| 120 #endif | |
| 121 | |
| 3524 | 122 /* WIN32_NATIVE is for XEmacs. |
| 458 | 123 MSDOS, WINDOWSNT, DOS_NT are for Emacs. */ |
| 442 | 124 #ifdef WIN32_NATIVE |
| 458 | 125 # undef MSDOS |
| 126 # undef WINDOWSNT | |
| 127 # define WINDOWSNT | |
| 128 #endif /* WIN32_NATIVE */ | |
| 129 | |
| 130 #ifdef MSDOS | |
| 131 # undef MSDOS | |
| 132 # define MSDOS TRUE | |
| 133 # include <fcntl.h> | |
| 134 # include <sys/param.h> | |
| 135 # include <io.h> | |
| 136 # ifndef HAVE_CONFIG_H | |
| 137 # define DOS_NT | |
| 138 # include <sys/config.h> | |
| 139 # endif | |
| 140 #else | |
| 141 # define MSDOS FALSE | |
| 142 #endif /* MSDOS */ | |
| 143 | |
| 144 #ifdef WINDOWSNT | |
| 428 | 145 # include <stdlib.h> |
| 146 # include <fcntl.h> | |
| 147 # include <string.h> | |
| 442 | 148 # include <direct.h> |
| 428 | 149 # include <io.h> |
| 150 # define MAXPATHLEN _MAX_PATH | |
| 458 | 151 # undef HAVE_NTGUI |
| 152 # undef DOS_NT | |
| 153 # define DOS_NT | |
| 428 | 154 # ifndef HAVE_GETCWD |
| 155 # define HAVE_GETCWD | |
| 156 # endif /* undef HAVE_GETCWD */ | |
| 2325 | 157 #else /* not WINDOWSNT */ |
| 442 | 158 # ifdef STDC_HEADERS |
| 159 # include <stdlib.h> | |
| 160 # include <string.h> | |
| 2325 | 161 # else /* no standard C headers */ |
| 3972 | 162 extern char *getenv (); |
| 163 extern char *strcpy (); | |
| 164 extern char *strncpy (); | |
| 165 extern char *strcat (); | |
| 166 extern char *strncat (); | |
| 167 extern unsigned long strlen (); | |
| 168 extern PTR malloc (); | |
| 169 extern PTR realloc (); | |
| 2325 | 170 # ifdef VMS |
| 171 # define EXIT_SUCCESS 1 | |
| 172 # define EXIT_FAILURE 0 | |
| 173 # else /* no VMS */ | |
| 174 # define EXIT_SUCCESS 0 | |
| 175 # define EXIT_FAILURE 1 | |
| 176 # endif | |
| 442 | 177 # endif |
| 458 | 178 #endif /* !WINDOWSNT */ |
| 428 | 179 |
| 180 #ifdef HAVE_UNISTD_H | |
| 181 # include <unistd.h> | |
| 182 #else | |
| 458 | 183 # if defined (HAVE_GETCWD) && !defined (WINDOWSNT) |
| 442 | 184 extern char *getcwd (char *buf, size_t size); |
| 428 | 185 # endif |
| 186 #endif /* HAVE_UNISTD_H */ | |
| 187 | |
| 188 #include <stdio.h> | |
| 189 #include <ctype.h> | |
| 190 #include <errno.h> | |
| 442 | 191 #ifndef errno |
| 192 extern int errno; | |
| 193 #endif | |
| 428 | 194 #include <sys/types.h> |
| 195 #include <sys/stat.h> | |
| 196 | |
| 458 | 197 #include <assert.h> |
| 198 #ifdef NDEBUG | |
| 199 # undef assert /* some systems have a buggy assert.h */ | |
| 200 # define assert(x) ((void) 0) | |
| 201 #endif | |
| 202 | |
| 428 | 203 #if !defined (S_ISREG) && defined (S_IFREG) |
| 204 # define S_ISREG(m) (((m) & S_IFMT) == S_IFREG) | |
| 205 #endif | |
| 206 | |
| 3517 | 207 #ifdef NO_LONG_OPTIONS /* define this if you don't have GNU getopt */ |
| 208 # define NO_LONG_OPTIONS TRUE | |
| 428 | 209 # define getopt_long(argc,argv,optstr,lopts,lind) getopt (argc, argv, optstr) |
| 210 extern char *optarg; | |
| 211 extern int optind, opterr; | |
| 3517 | 212 #else |
| 213 # define NO_LONG_OPTIONS FALSE | |
| 214 # include <getopt.h> | |
| 215 #endif /* NO_LONG_OPTIONS */ | |
| 216 | |
| 217 #ifndef HAVE_CONFIG_H /* this is a standalone compilation */ | |
| 218 # ifdef __CYGWIN__ /* compiling on Cygwin */ | |
| 709 | 219 !!! NOTICE !!! |
| 220 the regex.h distributed with Cygwin is not compatible with etags, alas! | |
| 221 If you want regular expression support, you should delete this notice and | |
| 222 arrange to use the GNU regex.h and regex.c. | |
| 223 # endif | |
| 3517 | 224 #endif |
| 225 #include <regex.h> | |
| 428 | 226 |
| 227 /* Define CTAGS to make the program "ctags" compatible with the usual one. | |
| 228 Leave it undefined to make the program "etags", which makes emacs-style | |
| 229 tag tables and tags typedefs, #defines and struct/union/enum by default. */ | |
| 230 #ifdef CTAGS | |
| 231 # undef CTAGS | |
| 232 # define CTAGS TRUE | |
| 233 #else | |
| 234 # define CTAGS FALSE | |
| 235 #endif | |
| 236 | |
| 458 | 237 #define streq(s,t) (assert((s)!=NULL || (t)!=NULL), !strcmp (s, t)) |
| 2225 | 238 #define strcaseeq(s,t) (assert((s)!=NULL && (t)!=NULL), !etags_strcasecmp (s, t)) |
| 458 | 239 #define strneq(s,t,n) (assert((s)!=NULL || (t)!=NULL), !strncmp (s, t, n)) |
| 2225 | 240 #define strncaseeq(s,t,n) (assert((s)!=NULL && (t)!=NULL), !etags_strncasecmp (s, t, n)) |
| 428 | 241 |
| 242 #define CHARS 256 /* 2^sizeof(char) */ | |
| 458 | 243 #define CHAR(x) ((unsigned int)(x) & (CHARS - 1)) |
| 2225 | 244 #define iswhite(c) (_wht[CHAR(c)]) /* c is white (see white) */ |
| 245 #define notinname(c) (_nin[CHAR(c)]) /* c is not in a name (see nonam) */ | |
| 246 #define begtoken(c) (_btk[CHAR(c)]) /* c can start token (see begtk) */ | |
| 247 #define intoken(c) (_itk[CHAR(c)]) /* c can be in token (see midtk) */ | |
| 248 #define endtoken(c) (_etk[CHAR(c)]) /* c ends tokens (see endtk) */ | |
| 428 | 249 |
| 458 | 250 #define ISALNUM(c) isalnum (CHAR(c)) |
| 251 #define ISALPHA(c) isalpha (CHAR(c)) | |
| 252 #define ISDIGIT(c) isdigit (CHAR(c)) | |
| 253 #define ISLOWER(c) islower (CHAR(c)) | |
| 254 | |
| 255 #define lowcase(c) tolower (CHAR(c)) | |
| 256 #define upcase(c) toupper (CHAR(c)) | |
| 257 | |
| 428 | 258 |
| 259 /* | |
| 260 * xnew, xrnew -- allocate, reallocate storage | |
| 261 * | |
| 262 * SYNOPSIS: Type *xnew (int n, Type); | |
| 458 | 263 * void xrnew (OldPointer, int n, Type); |
| 428 | 264 */ |
| 458 | 265 #if DEBUG |
| 428 | 266 # include "chkmalloc.h" |
| 267 # define xnew(n,Type) ((Type *) trace_malloc (__FILE__, __LINE__, \ | |
| 268 (n) * sizeof (Type))) | |
| 458 | 269 # define xrnew(op,n,Type) ((op) = (Type *) trace_realloc (__FILE__, __LINE__, \ |
| 270 (char *) (op), (n) * sizeof (Type))) | |
| 428 | 271 #else |
| 272 # define xnew(n,Type) ((Type *) xmalloc ((n) * sizeof (Type))) | |
| 458 | 273 # define xrnew(op,n,Type) ((op) = (Type *) xrealloc ( \ |
| 274 (char *) (op), (n) * sizeof (Type))) | |
| 428 | 275 #endif |
| 276 | |
| 709 | 277 #define bool int |
| 278 | |
| 279 typedef void Lang_function __P((FILE *)); | |
| 428 | 280 |
| 281 typedef struct | |
| 282 { | |
| 2225 | 283 char *suffix; /* file name suffix for this compressor */ |
| 284 char *command; /* takes one arg and decompresses to stdout */ | |
| 428 | 285 } compressor; |
| 286 | |
| 287 typedef struct | |
| 288 { | |
| 2225 | 289 char *name; /* language name */ |
| 290 char *help; /* detailed help for the language */ | |
| 291 Lang_function *function; /* parse function */ | |
| 292 char **suffixes; /* name suffixes of this language's files */ | |
| 293 char **filenames; /* names of this language's files */ | |
| 294 char **interpreters; /* interpreters for this language */ | |
| 295 bool metasource; /* source used to generate other sources */ | |
| 428 | 296 } language; |
| 297 | |
| 2225 | 298 typedef struct fdesc |
| 299 { | |
| 300 struct fdesc *next; /* for the linked list */ | |
| 301 char *infname; /* uncompressed input file name */ | |
| 302 char *infabsname; /* absolute uncompressed input file name */ | |
| 303 char *infabsdir; /* absolute dir of input file */ | |
| 304 char *taggedfname; /* file name to write in tagfile */ | |
| 305 language *lang; /* language of file */ | |
| 306 char *prop; /* file properties to write in tagfile */ | |
| 307 bool usecharno; /* etags tags shall contain char number */ | |
| 308 bool written; /* entry written in the tags file */ | |
| 309 } fdesc; | |
| 310 | |
| 428 | 311 typedef struct node_st |
| 2225 | 312 { /* sorting structure */ |
| 313 struct node_st *left, *right; /* left and right sons */ | |
| 314 fdesc *fdp; /* description of file to whom tag belongs */ | |
| 315 char *name; /* tag name */ | |
| 316 char *regex; /* search regexp */ | |
| 317 bool valid; /* write this tag on the tag file */ | |
| 318 bool is_func; /* function tag: use regexp in CTAGS mode */ | |
| 319 bool been_warned; /* warning already given for duplicated tag */ | |
| 320 int lno; /* line number tag is on */ | |
| 428 | 321 long cno; /* character number line starts on */ |
| 322 } node; | |
| 323 | |
| 324 /* | |
| 325 * A `linebuffer' is a structure which holds a line of text. | |
| 326 * `readline_internal' reads a line from a stream into a linebuffer | |
| 327 * and works regardless of the length of the line. | |
| 328 * SIZE is the size of BUFFER, LEN is the length of the string in | |
| 329 * BUFFER after readline reads it. | |
| 330 */ | |
| 331 typedef struct | |
| 332 { | |
| 333 long size; | |
| 334 int len; | |
| 335 char *buffer; | |
| 336 } linebuffer; | |
| 337 | |
| 2225 | 338 /* Used to support mixing of --lang and file names. */ |
| 339 typedef struct | |
| 340 { | |
| 341 enum { | |
| 342 at_language, /* a language specification */ | |
| 343 at_regexp, /* a regular expression */ | |
| 344 at_filename, /* a file name */ | |
| 345 at_stdin, /* read from stdin here */ | |
| 346 at_end /* stop parsing the list */ | |
| 347 } arg_type; /* argument type */ | |
| 348 language *lang; /* language associated with the argument */ | |
| 349 char *what; /* the argument itself */ | |
| 350 } argument; | |
| 351 | |
| 352 /* Structure defining a regular expression. */ | |
| 353 typedef struct regexp | |
| 354 { | |
| 355 struct regexp *p_next; /* pointer to next in list */ | |
| 356 language *lang; /* if set, use only for this language */ | |
| 357 char *pattern; /* the regexp pattern */ | |
| 358 char *name; /* tag name */ | |
| 359 struct re_pattern_buffer *pat; /* the compiled pattern */ | |
| 360 struct re_registers regs; /* re registers */ | |
| 361 bool error_signaled; /* already signaled for this regexp */ | |
| 362 bool force_explicit_name; /* do not allow implict tag name */ | |
| 363 bool ignore_case; /* ignore case when matching */ | |
| 364 bool multi_line; /* do a multi-line match on the whole file */ | |
| 365 } regexp; | |
| 366 | |
| 367 | |
| 428 | 368 /* Many compilers barf on this: |
| 369 Lang_function Ada_funcs; | |
| 370 so let's write it this way */ | |
| 709 | 371 static void Ada_funcs __P((FILE *)); |
| 372 static void Asm_labels __P((FILE *)); | |
| 373 static void C_entries __P((int c_ext, FILE *)); | |
| 374 static void default_C_entries __P((FILE *)); | |
| 375 static void plain_C_entries __P((FILE *)); | |
| 376 static void Cjava_entries __P((FILE *)); | |
| 377 static void Cobol_paragraphs __P((FILE *)); | |
| 378 static void Cplusplus_entries __P((FILE *)); | |
| 379 static void Cstar_entries __P((FILE *)); | |
| 380 static void Erlang_functions __P((FILE *)); | |
| 2554 | 381 static void Forth_words __P((FILE *)); |
| 709 | 382 static void Fortran_functions __P((FILE *)); |
| 2225 | 383 static void HTML_labels __P((FILE *)); |
| 709 | 384 static void Lisp_functions __P((FILE *)); |
| 2325 | 385 static void Lua_functions __P((FILE *)); |
| 709 | 386 static void Makefile_targets __P((FILE *)); |
| 387 static void Pascal_functions __P((FILE *)); | |
| 388 static void Perl_functions __P((FILE *)); | |
| 389 static void PHP_functions __P((FILE *)); | |
| 2225 | 390 static void PS_functions __P((FILE *)); |
| 709 | 391 static void Prolog_functions __P((FILE *)); |
| 392 static void Python_functions __P((FILE *)); | |
| 393 static void Scheme_functions __P((FILE *)); | |
| 394 static void TeX_commands __P((FILE *)); | |
| 395 static void Texinfo_nodes __P((FILE *)); | |
| 2225 | 396 static void Yacc_entries __P((FILE *)); |
| 709 | 397 static void just_read_file __P((FILE *)); |
| 398 | |
| 399 static void print_language_names __P((void)); | |
| 400 static void print_version __P((void)); | |
| 2225 | 401 static void print_help __P((argument *)); |
| 709 | 402 int main __P((int, char **)); |
| 403 | |
| 404 static compressor *get_compressor_from_suffix __P((char *, char **)); | |
| 405 static language *get_language_from_langname __P((const char *)); | |
| 406 static language *get_language_from_interpreter __P((char *)); | |
| 2225 | 407 static language *get_language_from_filename __P((char *, bool)); |
| 408 static void readline __P((linebuffer *, FILE *)); | |
| 709 | 409 static long readline_internal __P((linebuffer *, FILE *)); |
| 2225 | 410 static bool nocase_tail __P((char *)); |
| 411 static void get_tag __P((char *, char **)); | |
| 428 | 412 |
| 2225 | 413 static void analyse_regex __P((char *)); |
| 414 static void free_regexps __P((void)); | |
| 415 static void regex_tag_multiline __P((void)); | |
| 709 | 416 static void error __P((const char *, const char *)); |
| 417 static void suggest_asking_for_help __P((void)); | |
| 418 void fatal __P((char *, char *)); | |
| 419 static void pfatal __P((char *)); | |
| 420 static void add_node __P((node *, node **)); | |
| 421 | |
| 422 static void init __P((void)); | |
| 2225 | 423 static void process_file_name __P((char *, language *)); |
| 424 static void process_file __P((FILE *, char *, language *)); | |
| 425 static void find_entries __P((FILE *)); | |
| 709 | 426 static void free_tree __P((node *)); |
| 2225 | 427 static void free_fdesc __P((fdesc *)); |
| 709 | 428 static void pfnote __P((char *, bool, char *, int, int, long)); |
| 2225 | 429 static void make_tag __P((char *, int, bool, char *, int, int, long)); |
| 430 static void invalidate_nodes __P((fdesc *, node **)); | |
| 709 | 431 static void put_entries __P((node *)); |
| 432 | |
| 433 static char *concat __P((char *, char *, char *)); | |
| 434 static char *skip_spaces __P((char *)); | |
| 435 static char *skip_non_spaces __P((char *)); | |
| 436 static char *savenstr __P((char *, int)); | |
| 437 static char *savestr __P((char *)); | |
| 438 static char *etags_strchr __P((const char *, int)); | |
| 439 static char *etags_strrchr __P((const char *, int)); | |
| 2225 | 440 static int etags_strcasecmp __P((const char *, const char *)); |
| 441 static int etags_strncasecmp __P((const char *, const char *, int)); | |
| 709 | 442 static char *etags_getcwd __P((void)); |
| 443 static char *relative_filename __P((char *, char *)); | |
| 444 static char *absolute_filename __P((char *, char *)); | |
| 445 static char *absolute_dirname __P((char *, char *)); | |
| 446 static bool filename_is_absolute __P((char *f)); | |
| 447 static void canonicalize_filename __P((char *)); | |
| 2225 | 448 static void linebuffer_init __P((linebuffer *)); |
| 709 | 449 static void linebuffer_setlen __P((linebuffer *, int)); |
| 2225 | 450 static PTR xmalloc __P((unsigned int)); |
| 451 static PTR xrealloc __P((char *, unsigned int)); | |
| 428 | 452 |
| 453 | |
| 2225 | 454 static char searchar = '/'; /* use /.../ searches */ |
| 455 | |
| 456 static char *tagfile; /* output file */ | |
| 457 static char *progname; /* name this program was invoked with */ | |
| 458 static char *cwd; /* current working directory */ | |
| 459 static char *tagfiledir; /* directory of tagfile */ | |
| 460 static FILE *tagf; /* ioptr for tags file */ | |
| 461 | |
| 462 static fdesc *fdhead; /* head of file description list */ | |
| 463 static fdesc *curfdp; /* current file description */ | |
| 464 static int lineno; /* line number of current line */ | |
| 465 static long charno; /* current character number */ | |
| 466 static long linecharno; /* charno of start of current line */ | |
| 467 static char *dbp; /* pointer to start of current tag */ | |
| 468 | |
| 469 static const int invalidcharno = -1; | |
| 470 | |
| 471 static node *nodehead; /* the head of the binary tree of tags */ | |
| 472 static node *last_node; /* the last node created */ | |
| 473 | |
| 474 static linebuffer lb; /* the current line */ | |
| 475 static linebuffer filebuf; /* a buffer containing the whole file */ | |
| 476 static linebuffer token_name; /* a buffer containing a tag name */ | |
| 428 | 477 |
| 478 /* boolean "functions" (see init) */ | |
| 2225 | 479 static bool _wht[CHARS], _nin[CHARS], _itk[CHARS], _btk[CHARS], _etk[CHARS]; |
| 480 static char | |
| 428 | 481 /* white chars */ |
| 442 | 482 *white = " \f\t\n\r\v", |
| 428 | 483 /* not in a name */ |
| 2225 | 484 *nonam = " \f\t\n\r()=,;", /* look at make_tag before modifying! */ |
| 428 | 485 /* token ending chars */ |
| 486 *endtk = " \t\n\r\"'#()[]{}=-+%*/&|^~!<>;,.:?", | |
| 487 /* token starting chars */ | |
| 488 *begtk = "ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz$~@", | |
| 489 /* valid in-token chars */ | |
| 490 *midtk = "ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz$0123456789"; | |
| 491 | |
| 2225 | 492 static bool append_to_tagfile; /* -a: append to tags */ |
| 3972 | 493 /* The next five default to TRUE for etags, but to FALSE for ctags. */ |
| 2225 | 494 static bool typedefs; /* -t: create tags for C and Ada typedefs */ |
| 495 static bool typedefs_or_cplusplus; /* -T: create tags for C typedefs, level */ | |
| 428 | 496 /* 0 struct/enum/union decls, and C++ */ |
| 497 /* member functions. */ | |
| 2225 | 498 static bool constantypedefs; /* -d: create tags for C #define, enum */ |
| 428 | 499 /* constants and variables. */ |
| 500 /* -D: opposite of -d. Default under ctags. */ | |
| 2225 | 501 static bool globals; /* create tags for global variables */ |
| 3972 | 502 static bool members; /* create tags for C member variables */ |
| 2225 | 503 static bool declarations; /* --declarations: tag them and extern in C&Co*/ |
| 504 static bool no_line_directive; /* ignore #line directives (undocumented) */ | |
| 3876 | 505 static bool no_duplicates; /* no duplicate tags for ctags (undocumented) */ |
| 2225 | 506 static bool update; /* -u: update tags */ |
| 507 static bool vgrind_style; /* -v: create vgrind style index output */ | |
| 3876 | 508 static bool no_warnings; /* -w: suppress warnings (undocumented) */ |
| 2225 | 509 static bool cxref_style; /* -x: create cxref style output */ |
| 510 static bool cplusplus; /* .[hc] means C++, not C */ | |
| 511 static bool ignoreindent; /* -I: ignore indentation in C */ | |
| 512 static bool packages_only; /* --packages-only: in Ada, only tag packages*/ | |
| 513 | |
| 3517 | 514 /* STDIN is defined in LynxOS system headers */ |
| 515 #ifdef STDIN | |
| 516 # undef STDIN | |
| 517 #endif | |
| 518 | |
| 2225 | 519 #define STDIN 0x1001 /* returned by getopt_long on --parse-stdin */ |
| 520 static bool parsing_stdin; /* --parse-stdin used */ | |
| 521 | |
| 522 static regexp *p_head; /* list of all regexps */ | |
| 523 static bool need_filebuf; /* some regexes are multi-line */ | |
| 3517 | 524 |
| 2225 | 525 static struct option longopts[] = |
| 428 | 526 { |
| 3876 | 527 { "append", no_argument, NULL, 'a' }, |
| 528 { "packages-only", no_argument, &packages_only, TRUE }, | |
| 529 { "c++", no_argument, NULL, 'C' }, | |
| 530 { "declarations", no_argument, &declarations, TRUE }, | |
| 531 { "no-line-directive", no_argument, &no_line_directive, TRUE }, | |
| 532 { "no-duplicates", no_argument, &no_duplicates, TRUE }, | |
| 533 { "help", no_argument, NULL, 'h' }, | |
| 534 { "help", no_argument, NULL, 'H' }, | |
| 535 { "ignore-indentation", no_argument, NULL, 'I' }, | |
| 536 { "language", required_argument, NULL, 'l' }, | |
| 537 { "members", no_argument, &members, TRUE }, | |
| 538 { "no-members", no_argument, &members, FALSE }, | |
| 539 { "output", required_argument, NULL, 'o' }, | |
| 540 { "regex", required_argument, NULL, 'r' }, | |
| 541 { "no-regex", no_argument, NULL, 'R' }, | |
| 542 { "ignore-case-regex", required_argument, NULL, 'c' }, | |
| 2225 | 543 { "parse-stdin", required_argument, NULL, STDIN }, |
| 3876 | 544 { "version", no_argument, NULL, 'V' }, |
| 2225 | 545 |
| 3090 | 546 #if CTAGS /* Ctags options */ |
| 3876 | 547 { "backward-search", no_argument, NULL, 'B' }, |
| 548 { "cxref", no_argument, NULL, 'x' }, | |
| 549 { "defines", no_argument, NULL, 'd' }, | |
| 550 { "globals", no_argument, &globals, TRUE }, | |
| 551 { "typedefs", no_argument, NULL, 't' }, | |
| 552 { "typedefs-and-c++", no_argument, NULL, 'T' }, | |
| 553 { "update", no_argument, NULL, 'u' }, | |
| 554 { "vgrind", no_argument, NULL, 'v' }, | |
| 555 { "no-warn", no_argument, NULL, 'w' }, | |
| 2225 | 556 |
| 3090 | 557 #else /* Etags options */ |
| 3876 | 558 { "no-defines", no_argument, NULL, 'D' }, |
| 559 { "no-globals", no_argument, &globals, FALSE }, | |
| 560 { "include", required_argument, NULL, 'i' }, | |
| 2225 | 561 #endif |
| 428 | 562 { NULL } |
| 563 }; | |
| 564 | |
| 2225 | 565 static compressor compressors[] = |
| 428 | 566 { |
| 567 { "z", "gzip -d -c"}, | |
| 568 { "Z", "gzip -d -c"}, | |
| 569 { "gz", "gzip -d -c"}, | |
| 570 { "GZ", "gzip -d -c"}, | |
| 571 { "bz2", "bzip2 -d -c" }, | |
| 572 { NULL } | |
| 573 }; | |
| 574 | |
| 575 /* | |
| 576 * Language stuff. | |
| 577 */ | |
| 578 | |
| 579 /* Ada code */ | |
| 2225 | 580 static char *Ada_suffixes [] = |
| 428 | 581 { "ads", "adb", "ada", NULL }; |
| 2225 | 582 static char Ada_help [] = |
| 583 "In Ada code, functions, procedures, packages, tasks and types are\n\ | |
| 584 tags. Use the `--packages-only' option to create tags for\n\ | |
| 585 packages only.\n\ | |
| 586 Ada tag names have suffixes indicating the type of entity:\n\ | |
| 587 Entity type: Qualifier:\n\ | |
| 588 ------------ ----------\n\ | |
| 589 function /f\n\ | |
| 590 procedure /p\n\ | |
| 591 package spec /s\n\ | |
| 592 package body /b\n\ | |
| 593 type /t\n\ | |
| 594 task /k\n\ | |
| 595 Thus, `M-x find-tag <RET> bidule/b <RET>' will go directly to the\n\ | |
| 596 body of the package `bidule', while `M-x find-tag <RET> bidule <RET>'\n\ | |
| 597 will just search for any tag `bidule'."; | |
| 428 | 598 |
| 599 /* Assembly code */ | |
| 2225 | 600 static char *Asm_suffixes [] = |
| 601 { "a", /* Unix assembler */ | |
| 602 "asm", /* Microcontroller assembly */ | |
| 603 "def", /* BSO/Tasking definition includes */ | |
| 604 "inc", /* Microcontroller include files */ | |
| 605 "ins", /* Microcontroller include files */ | |
| 606 "s", "sa", /* Unix assembler */ | |
| 607 "S", /* cpp-processed Unix assembler */ | |
| 608 "src", /* BSO/Tasking C compiler output */ | |
| 609 NULL | |
| 610 }; | |
| 611 static char Asm_help [] = | |
| 612 "In assembler code, labels appearing at the beginning of a line,\n\ | |
| 613 followed by a colon, are tags."; | |
| 614 | |
| 428 | 615 |
| 616 /* Note that .c and .h can be considered C++, if the --c++ flag was | |
| 2225 | 617 given, or if the `class' or `template' keyowrds are met inside the file. |
| 458 | 618 That is why default_C_entries is called for these. */ |
| 2225 | 619 static char *default_C_suffixes [] = |
| 428 | 620 { "c", "h", NULL }; |
| 2225 | 621 static char default_C_help [] = |
| 622 "In C code, any C function or typedef is a tag, and so are\n\ | |
| 623 definitions of `struct', `union' and `enum'. `#define' macro\n\ | |
| 624 definitions and `enum' constants are tags unless you specify\n\ | |
| 625 `--no-defines'. Global variables are tags unless you specify\n\ | |
| 3972 | 626 `--no-globals' and so are struct members unless you specify\n\ |
| 627 `--no-members'. Use of `--no-globals', `--no-defines' and\n\ | |
| 628 `--no-members' can make the tags table file much smaller.\n\ | |
| 2225 | 629 You can tag function declarations and external variables by\n\ |
| 3972 | 630 using `--declarations'."; |
| 2225 | 631 |
| 632 static char *Cplusplus_suffixes [] = | |
| 458 | 633 { "C", "c++", "cc", "cpp", "cxx", "H", "h++", "hh", "hpp", "hxx", |
| 428 | 634 "M", /* Objective C++ */ |
| 635 "pdb", /* Postscript with C syntax */ | |
| 636 NULL }; | |
| 2225 | 637 static char Cplusplus_help [] = |
| 638 "In C++ code, all the tag constructs of C code are tagged. (Use\n\ | |
| 639 --help --lang=c --lang=c++ for full help.)\n\ | |
| 3972 | 640 In addition to C tags, member functions are also recognized. Member\n\ |
| 641 variables are recognized unless you use the `--no-members' option.\n\ | |
| 2225 | 642 Tags for variables and functions in classes are named `CLASS::VARIABLE'\n\ |
| 643 and `CLASS::FUNCTION'. `operator' definitions have tag names like\n\ | |
| 644 `operator+'."; | |
| 645 | |
| 646 static char *Cjava_suffixes [] = | |
| 428 | 647 { "java", NULL }; |
| 2225 | 648 static char Cjava_help [] = |
| 649 "In Java code, all the tags constructs of C and C++ code are\n\ | |
| 650 tagged. (Use --help --lang=c --lang=c++ --lang=java for full help.)"; | |
| 651 | |
| 652 | |
| 653 static char *Cobol_suffixes [] = | |
| 428 | 654 { "COB", "cob", NULL }; |
| 2225 | 655 static char Cobol_help [] = |
| 656 "In Cobol code, tags are paragraph names; that is, any word\n\ | |
| 657 starting in column 8 and followed by a period."; | |
| 658 | |
| 659 static char *Cstar_suffixes [] = | |
| 428 | 660 { "cs", "hs", NULL }; |
| 661 | |
| 2225 | 662 static char *Erlang_suffixes [] = |
| 428 | 663 { "erl", "hrl", NULL }; |
| 2225 | 664 static char Erlang_help [] = |
| 665 "In Erlang code, the tags are the functions, records and macros\n\ | |
| 666 defined in the file."; | |
| 667 | |
| 2554 | 668 char *Forth_suffixes [] = |
| 669 { "fth", "tok", NULL }; | |
| 670 static char Forth_help [] = | |
| 671 "In Forth code, tags are words defined by `:',\n\ | |
| 672 constant, code, create, defer, value, variable, buffer:, field."; | |
| 673 | |
| 2225 | 674 static char *Fortran_suffixes [] = |
| 428 | 675 { "F", "f", "f90", "for", NULL }; |
| 2225 | 676 static char Fortran_help [] = |
| 677 "In Fortran code, functions, subroutines and block data are tags."; | |
| 678 | |
| 679 static char *HTML_suffixes [] = | |
| 680 { "htm", "html", "shtml", NULL }; | |
| 681 static char HTML_help [] = | |
| 682 "In HTML input files, the tags are the `title' and the `h1', `h2',\n\ | |
| 683 `h3' headers. Also, tags are `name=' in anchors and all\n\ | |
| 684 occurrences of `id='."; | |
| 685 | |
| 686 static char *Lisp_suffixes [] = | |
| 458 | 687 { "cl", "clisp", "el", "l", "lisp", "LSP", "lsp", "ml", NULL }; |
| 2225 | 688 static char Lisp_help [] = |
| 689 "In Lisp code, any function defined with `defun', any variable\n\ | |
| 690 defined with `defvar' or `defconst', and in general the first\n\ | |
| 691 argument of any expression that starts with `(def' in column zero\n\ | |
| 692 is a tag."; | |
| 693 | |
| 2325 | 694 static char *Lua_suffixes [] = |
| 695 { "lua", "LUA", NULL }; | |
| 696 static char Lua_help [] = | |
| 697 "In Lua scripts, all functions are tags."; | |
| 698 | |
| 2225 | 699 static char *Makefile_filenames [] = |
| 458 | 700 { "Makefile", "makefile", "GNUMakefile", "Makefile.in", "Makefile.am", NULL}; |
| 2225 | 701 static char Makefile_help [] = |
| 702 "In makefiles, targets are tags; additionally, variables are tags\n\ | |
| 703 unless you specify `--no-globals'."; | |
| 704 | |
| 705 static char *Objc_suffixes [] = | |
| 458 | 706 { "lm", /* Objective lex file */ |
| 428 | 707 "m", /* Objective C file */ |
| 708 NULL }; | |
| 2225 | 709 static char Objc_help [] = |
| 710 "In Objective C code, tags include Objective C definitions for classes,\n\ | |
| 711 class categories, methods and protocols. Tags for variables and\n\ | |
| 3517 | 712 functions in classes are named `CLASS::VARIABLE' and `CLASS::FUNCTION'.\n\ |
| 713 (Use --help --lang=c --lang=objc --lang=java for full help.)"; | |
| 2225 | 714 |
| 715 static char *Pascal_suffixes [] = | |
| 716 { "p", "pas", NULL }; | |
| 717 static char Pascal_help [] = | |
| 718 "In Pascal code, the tags are the functions and procedures defined\n\ | |
| 719 in the file."; | |
| 3517 | 720 /* " // this is for working around an Emacs highlighting bug... */ |
| 2225 | 721 |
| 722 static char *Perl_suffixes [] = | |
| 723 { "pl", "pm", NULL }; | |
| 724 static char *Perl_interpreters [] = | |
| 725 { "perl", "@PERL@", NULL }; | |
| 726 static char Perl_help [] = | |
| 727 "In Perl code, the tags are the packages, subroutines and variables\n\ | |
| 728 defined by the `package', `sub', `my' and `local' keywords. Use\n\ | |
| 729 `--globals' if you want to tag global variables. Tags for\n\ | |
| 730 subroutines are named `PACKAGE::SUB'. The name for subroutines\n\ | |
| 731 defined in the default package is `main::SUB'."; | |
| 732 | |
| 733 static char *PHP_suffixes [] = | |
| 734 { "php", "php3", "php4", NULL }; | |
| 735 static char PHP_help [] = | |
| 3972 | 736 "In PHP code, tags are functions, classes and defines. Unless you use\n\ |
| 737 the `--no-members' option, vars are tags too."; | |
| 2225 | 738 |
| 739 static char *plain_C_suffixes [] = | |
| 740 { "pc", /* Pro*C file */ | |
| 741 NULL }; | |
| 742 | |
| 743 static char *PS_suffixes [] = | |
| 428 | 744 { "ps", "psw", NULL }; /* .psw is for PSWrap */ |
| 2225 | 745 static char PS_help [] = |
| 746 "In PostScript code, the tags are the functions."; | |
| 747 | |
| 748 static char *Prolog_suffixes [] = | |
| 428 | 749 { "prolog", NULL }; |
| 2225 | 750 static char Prolog_help [] = |
| 751 "In Prolog code, tags are predicates and rules at the beginning of\n\ | |
| 752 line."; | |
| 753 | |
| 754 static char *Python_suffixes [] = | |
| 428 | 755 { "py", NULL }; |
| 2225 | 756 static char Python_help [] = |
| 757 "In Python code, `def' or `class' at the beginning of a line\n\ | |
| 758 generate a tag."; | |
| 428 | 759 |
| 760 /* Can't do the `SCM' or `scm' prefix with a version number. */ | |
| 2225 | 761 static char *Scheme_suffixes [] = |
| 458 | 762 { "oak", "sch", "scheme", "SCM", "scm", "SM", "sm", "ss", "t", NULL }; |
| 2225 | 763 static char Scheme_help [] = |
| 764 "In Scheme code, tags include anything defined with `def' or with a\n\ | |
| 765 construct whose name starts with `def'. They also include\n\ | |
| 766 variables set with `set!' at top level in the file."; | |
| 767 | |
| 768 static char *TeX_suffixes [] = | |
| 458 | 769 { "bib", "clo", "cls", "ltx", "sty", "TeX", "tex", NULL }; |
| 2225 | 770 static char TeX_help [] = |
| 771 "In LaTeX text, the argument of any of the commands `\\chapter',\n\ | |
| 772 `\\section', `\\subsection', `\\subsubsection', `\\eqno', `\\label',\n\ | |
| 773 `\\ref', `\\cite', `\\bibitem', `\\part', `\\appendix', `\\entry',\n\ | |
| 774 `\\index', `\\def', `\\newcommand', `\\renewcommand',\n\ | |
| 775 `\\newenvironment' or `\\renewenvironment' is a tag.\n\ | |
| 776 \n\ | |
| 777 Other commands can be specified by setting the environment variable\n\ | |
| 778 `TEXTAGS' to a colon-separated list like, for example,\n\ | |
| 779 TEXTAGS=\"mycommand:myothercommand\"."; | |
| 780 | |
| 781 | |
| 782 static char *Texinfo_suffixes [] = | |
| 458 | 783 { "texi", "texinfo", "txi", NULL }; |
| 2225 | 784 static char Texinfo_help [] = |
| 785 "for texinfo files, lines starting with @node are tagged."; | |
| 786 | |
| 787 static char *Yacc_suffixes [] = | |
| 458 | 788 { "y", "y++", "ym", "yxx", "yy", NULL }; /* .ym is Objective yacc file */ |
| 2225 | 789 static char Yacc_help [] = |
| 790 "In Bison or Yacc input files, each rule defines as a tag the\n\ | |
| 791 nonterminal it constructs. The portions of the file that contain\n\ | |
| 792 C code are parsed as C code (use --help --lang=c --lang=yacc\n\ | |
| 793 for full help)."; | |
| 794 | |
| 795 static char auto_help [] = | |
| 796 "`auto' is not a real language, it indicates to use\n\ | |
| 797 a default language for files base on file name suffix and file contents."; | |
| 798 | |
| 799 static char none_help [] = | |
| 800 "`none' is not a real language, it indicates to only do\n\ | |
| 801 regexp processing on files."; | |
| 802 | |
| 803 static char no_lang_help [] = | |
| 804 "No detailed help available for this language."; | |
| 805 | |
| 428 | 806 |
| 807 /* | |
| 808 * Table of languages. | |
| 809 * | |
| 810 * It is ok for a given function to be listed under more than one | |
| 811 * name. I just didn't. | |
| 812 */ | |
| 813 | |
| 2225 | 814 static language lang_names [] = |
| 428 | 815 { |
| 2225 | 816 { "ada", Ada_help, Ada_funcs, Ada_suffixes }, |
| 817 { "asm", Asm_help, Asm_labels, Asm_suffixes }, | |
| 818 { "c", default_C_help, default_C_entries, default_C_suffixes }, | |
| 819 { "c++", Cplusplus_help, Cplusplus_entries, Cplusplus_suffixes }, | |
| 820 { "c*", no_lang_help, Cstar_entries, Cstar_suffixes }, | |
| 821 { "cobol", Cobol_help, Cobol_paragraphs, Cobol_suffixes }, | |
| 822 { "erlang", Erlang_help, Erlang_functions, Erlang_suffixes }, | |
| 2554 | 823 { "forth", Forth_help, Forth_words, Forth_suffixes }, |
| 2225 | 824 { "fortran", Fortran_help, Fortran_functions, Fortran_suffixes }, |
| 825 { "html", HTML_help, HTML_labels, HTML_suffixes }, | |
| 826 { "java", Cjava_help, Cjava_entries, Cjava_suffixes }, | |
| 827 { "lisp", Lisp_help, Lisp_functions, Lisp_suffixes }, | |
| 2325 | 828 { "lua", Lua_help, Lua_functions, Lua_suffixes }, |
| 2225 | 829 { "makefile", Makefile_help,Makefile_targets,NULL,Makefile_filenames}, |
| 830 { "objc", Objc_help, plain_C_entries, Objc_suffixes }, | |
| 831 { "pascal", Pascal_help, Pascal_functions, Pascal_suffixes }, | |
| 832 { "perl",Perl_help,Perl_functions,Perl_suffixes,NULL,Perl_interpreters}, | |
| 833 { "php", PHP_help, PHP_functions, PHP_suffixes }, | |
| 834 { "postscript",PS_help, PS_functions, PS_suffixes }, | |
| 835 { "proc", no_lang_help, plain_C_entries, plain_C_suffixes }, | |
| 836 { "prolog", Prolog_help, Prolog_functions, Prolog_suffixes }, | |
| 837 { "python", Python_help, Python_functions, Python_suffixes }, | |
| 838 { "scheme", Scheme_help, Scheme_functions, Scheme_suffixes }, | |
| 839 { "tex", TeX_help, TeX_commands, TeX_suffixes }, | |
| 840 { "texinfo", Texinfo_help, Texinfo_nodes, Texinfo_suffixes }, | |
| 841 { "yacc", Yacc_help,Yacc_entries,Yacc_suffixes,NULL,NULL,TRUE}, | |
| 842 { "auto", auto_help }, /* default guessing scheme */ | |
| 843 { "none", none_help, just_read_file }, /* regexp matching only */ | |
| 844 { NULL } /* end of list */ | |
| 428 | 845 }; |
| 458 | 846 |
| 428 | 847 |
| 848 static void | |
| 849 print_language_names () | |
| 850 { | |
| 851 language *lang; | |
| 458 | 852 char **name, **ext; |
| 428 | 853 |
| 854 puts ("\nThese are the currently supported languages, along with the\n\ | |
| 458 | 855 default file names and dot suffixes:"); |
| 428 | 856 for (lang = lang_names; lang->name != NULL; lang++) |
| 857 { | |
| 458 | 858 printf (" %-*s", 10, lang->name); |
| 859 if (lang->filenames != NULL) | |
| 860 for (name = lang->filenames; *name != NULL; name++) | |
| 861 printf (" %s", *name); | |
| 428 | 862 if (lang->suffixes != NULL) |
| 863 for (ext = lang->suffixes; *ext != NULL; ext++) | |
| 864 printf (" .%s", *ext); | |
| 865 puts (""); | |
| 866 } | |
| 2225 | 867 puts ("where `auto' means use default language for files based on file\n\ |
| 428 | 868 name suffix, and `none' means only do regexp processing on files.\n\ |
| 869 If no language is specified and no matching suffix is found,\n\ | |
| 870 the first line of the file is read for a sharp-bang (#!) sequence\n\ | |
| 871 followed by the name of an interpreter. If no such sequence is found,\n\ | |
| 872 Fortran is tried first; if no tags are found, C is tried next.\n\ | |
| 2225 | 873 When parsing any C file, a \"class\" or \"template\" keyword\n\ |
| 874 switches to C++."); | |
| 875 puts ("Compressed files are supported using gzip and bzip2.\n\ | |
| 876 \n\ | |
| 877 For detailed help on a given language use, for example,\n\ | |
| 878 etags --help --lang=ada."); | |
| 428 | 879 } |
| 880 | |
| 442 | 881 #ifndef EMACS_NAME |
| 2225 | 882 # define EMACS_NAME "standalone" |
| 428 | 883 #endif |
| 3972 | 884 #ifndef VERSION |
| 3993 | 885 # define VERSION "17.33" |
| 428 | 886 #endif |
| 887 static void | |
| 888 print_version () | |
| 889 { | |
| 3972 | 890 printf ("%s (%s %s)\n", (CTAGS) ? "ctags" : "etags", EMACS_NAME, VERSION); |
| 891 puts ("Copyright (C) 2007 Free Software Foundation, Inc."); | |
| 892 puts ("This program is distributed under the terms in ETAGS.README"); | |
| 428 | 893 |
| 2225 | 894 exit (EXIT_SUCCESS); |
| 428 | 895 } |
| 896 | |
| 3993 | 897 #ifndef PRINT_UNDOCUMENTED_OPTIONS_HELP |
| 898 # define PRINT_UNDOCUMENTED_OPTIONS_HELP FALSE | |
| 899 #endif | |
| 900 | |
| 428 | 901 static void |
| 2225 | 902 print_help (argbuffer) |
| 903 argument *argbuffer; | |
| 428 | 904 { |
| 2225 | 905 bool help_for_lang = FALSE; |
| 906 | |
| 907 for (; argbuffer->arg_type != at_end; argbuffer++) | |
| 908 if (argbuffer->arg_type == at_language) | |
| 909 { | |
| 910 if (help_for_lang) | |
| 911 puts (""); | |
| 912 puts (argbuffer->lang->help); | |
| 913 help_for_lang = TRUE; | |
| 914 } | |
| 915 | |
| 916 if (help_for_lang) | |
| 917 exit (EXIT_SUCCESS); | |
| 918 | |
| 428 | 919 printf ("Usage: %s [options] [[regex-option ...] file-name] ...\n\ |
| 920 \n\ | |
| 921 These are the options accepted by %s.\n", progname, progname); | |
| 3517 | 922 if (NO_LONG_OPTIONS) |
| 923 puts ("WARNING: long option names do not work with this executable,\n\ | |
| 924 as it is not linked with GNU getopt."); | |
| 2325 | 925 else |
| 3517 | 926 puts ("You may use unambiguous abbreviations for the long option names."); |
| 2225 | 927 puts (" A - as file name means read names from stdin (one per line).\n\ |
| 928 Absolute names are stored in the output file as they are.\n\ | |
| 929 Relative ones are stored relative to the output file's directory.\n"); | |
| 930 | |
| 3090 | 931 puts ("-a, --append\n\ |
| 428 | 932 Append tag entries to existing tags file."); |
| 933 | |
| 934 puts ("--packages-only\n\ | |
| 2225 | 935 For Ada files, only generate tags for packages."); |
| 428 | 936 |
| 937 if (CTAGS) | |
| 938 puts ("-B, --backward-search\n\ | |
| 939 Write the search commands for the tag entries using '?', the\n\ | |
| 940 backward-search command instead of '/', the forward-search command."); | |
| 941 | |
| 458 | 942 /* This option is mostly obsolete, because etags can now automatically |
| 943 detect C++. Retained for backward compatibility and for debugging and | |
| 944 experimentation. In principle, we could want to tag as C++ even | |
| 2225 | 945 before any "class" or "template" keyword. |
| 428 | 946 puts ("-C, --c++\n\ |
| 947 Treat files whose name suffix defaults to C language as C++ files."); | |
| 458 | 948 */ |
| 428 | 949 |
| 950 puts ("--declarations\n\ | |
| 951 In C and derived languages, create tags for function declarations,"); | |
| 952 if (CTAGS) | |
| 953 puts ("\tand create tags for extern variables if --globals is used."); | |
| 954 else | |
| 955 puts | |
| 956 ("\tand create tags for extern variables unless --no-globals is used."); | |
| 957 | |
| 958 if (CTAGS) | |
| 959 puts ("-d, --defines\n\ | |
| 960 Create tag entries for C #define constants and enum constants, too."); | |
| 961 else | |
| 962 puts ("-D, --no-defines\n\ | |
| 963 Don't create tag entries for C #define constants and enum constants.\n\ | |
| 964 This makes the tags file smaller."); | |
| 965 | |
| 966 if (!CTAGS) | |
| 2225 | 967 puts ("-i FILE, --include=FILE\n\ |
| 428 | 968 Include a note in tag file indicating that, when searching for\n\ |
| 969 a tag, one should also consult the tags file FILE after\n\ | |
| 970 checking the current file."); | |
| 2225 | 971 |
| 972 puts ("-l LANG, --language=LANG\n\ | |
| 428 | 973 Force the following files to be considered as written in the\n\ |
| 974 named language up to the next --language=LANG option."); | |
| 975 | |
| 976 if (CTAGS) | |
| 977 puts ("--globals\n\ | |
| 978 Create tag entries for global variables in some languages."); | |
| 979 else | |
| 980 puts ("--no-globals\n\ | |
| 981 Do not create tag entries for global variables in some\n\ | |
| 982 languages. This makes the tags file smaller."); | |
| 3993 | 983 |
| 984 if (PRINT_UNDOCUMENTED_OPTIONS_HELP) | |
| 985 puts ("--no-line-directive\n\ | |
| 986 Ignore #line preprocessor directives in C and derived languages."); | |
| 987 | |
| 3972 | 988 if (CTAGS) |
| 989 puts ("--members\n\ | |
| 2225 | 990 Create tag entries for members of structures in some languages."); |
| 3972 | 991 else |
| 992 puts ("--no-members\n\ | |
| 993 Do not create tag entries for members of structures\n\ | |
| 994 in some languages."); | |
| 428 | 995 |
| 2225 | 996 puts ("-r REGEXP, --regex=REGEXP or --regex=@regexfile\n\ |
| 997 Make a tag for each line matching a regular expression pattern\n\ | |
| 998 in the following files. {LANGUAGE}REGEXP uses REGEXP for LANGUAGE\n\ | |
| 999 files only. REGEXFILE is a file containing one REGEXP per line.\n\ | |
| 1000 REGEXP takes the form /TAGREGEXP/TAGNAME/MODS, where TAGNAME/ is\n\ | |
| 1001 optional. The TAGREGEXP pattern is anchored (as if preceded by ^)."); | |
| 1002 puts (" If TAGNAME/ is present, the tags created are named.\n\ | |
| 428 | 1003 For example Tcl named tags can be created with:\n\ |
| 2225 | 1004 --regex=\"/proc[ \\t]+\\([^ \\t]+\\)/\\1/.\".\n\ |
| 1005 MODS are optional one-letter modifiers: `i' means to ignore case,\n\ | |
| 1006 `m' means to allow multi-line matches, `s' implies `m' and\n\ | |
| 1007 causes dot to match any character, including newline."); | |
| 3993 | 1008 |
| 428 | 1009 puts ("-R, --no-regex\n\ |
| 1010 Don't create tags from regexps for the following files."); | |
| 3993 | 1011 |
| 2225 | 1012 puts ("-I, --ignore-indentation\n\ |
| 1013 In C and C++ do not assume that a closing brace in the first\n\ | |
| 1014 column is the final brace of a function or structure definition."); | |
| 3993 | 1015 |
| 428 | 1016 puts ("-o FILE, --output=FILE\n\ |
| 1017 Write the tags to FILE."); | |
| 3993 | 1018 |
| 2225 | 1019 puts ("--parse-stdin=NAME\n\ |
| 1020 Read from standard input and record tags as belonging to file NAME."); | |
| 428 | 1021 |
| 1022 if (CTAGS) | |
| 1023 { | |
| 1024 puts ("-t, --typedefs\n\ | |
| 1025 Generate tag entries for C and Ada typedefs."); | |
| 1026 puts ("-T, --typedefs-and-c++\n\ | |
| 1027 Generate tag entries for C typedefs, C struct/enum/union tags,\n\ | |
| 1028 and C++ member functions."); | |
| 2225 | 1029 } |
| 1030 | |
| 1031 if (CTAGS) | |
| 1032 puts ("-u, --update\n\ | |
| 428 | 1033 Update the tag entries for the given files, leaving tag\n\ |
| 1034 entries for other files in place. Currently, this is\n\ | |
| 1035 implemented by deleting the existing entries for the given\n\ | |
| 1036 files and then rewriting the new entries at the end of the\n\ | |
| 1037 tags file. It is often faster to simply rebuild the entire\n\ | |
| 1038 tag file than to use this."); | |
| 2225 | 1039 |
| 1040 if (CTAGS) | |
| 1041 { | |
| 428 | 1042 puts ("-v, --vgrind\n\ |
| 3090 | 1043 Print on the standard output an index of items intended for\n\ |
| 1044 human consumption, similar to the output of vgrind. The index\n\ | |
| 1045 is sorted, and gives the page number of each item."); | |
| 3993 | 1046 |
| 1047 if (PRINT_UNDOCUMENTED_OPTIONS_HELP) | |
| 1048 puts ("-w, --no-duplicates\n\ | |
| 3876 | 1049 Do not create duplicate tag entries, for compatibility with\n\ |
| 1050 traditional ctags."); | |
| 3993 | 1051 |
| 1052 if (PRINT_UNDOCUMENTED_OPTIONS_HELP) | |
| 1053 puts ("-w, --no-warn\n\ | |
| 3876 | 1054 Suppress warning messages about duplicate tag entries."); |
| 3993 | 1055 |
| 428 | 1056 puts ("-x, --cxref\n\ |
| 1057 Like --vgrind, but in the style of cxref, rather than vgrind.\n\ | |
| 1058 The output uses line numbers instead of page numbers, but\n\ | |
| 1059 beyond that the differences are cosmetic; try both to see\n\ | |
| 1060 which you like."); | |
| 1061 } | |
| 1062 | |
| 1063 puts ("-V, --version\n\ | |
| 1064 Print the version of the program.\n\ | |
| 1065 -h, --help\n\ | |
| 2225 | 1066 Print this help message.\n\ |
| 1067 Followed by one or more `--language' options prints detailed\n\ | |
| 1068 help about tag generation for the specified languages."); | |
| 428 | 1069 |
| 1070 print_language_names (); | |
| 1071 | |
| 1072 puts (""); | |
| 1073 puts ("Report bugs to bug-gnu-emacs@gnu.org"); | |
| 1074 | |
| 2225 | 1075 exit (EXIT_SUCCESS); |
| 428 | 1076 } |
| 1077 | |
| 1078 | |
| 1079 #ifdef VMS /* VMS specific functions */ | |
| 1080 | |
| 1081 #define EOS '\0' | |
| 1082 | |
| 1083 /* This is a BUG! ANY arbitrary limit is a BUG! | |
| 1084 Won't someone please fix this? */ | |
| 1085 #define MAX_FILE_SPEC_LEN 255 | |
| 1086 typedef struct { | |
| 1087 short curlen; | |
| 1088 char body[MAX_FILE_SPEC_LEN + 1]; | |
| 1089 } vspec; | |
| 1090 | |
| 1091 /* | |
| 1092 v1.05 nmm 26-Jun-86 fn_exp - expand specification of list of file names | |
| 1093 returning in each successive call the next file name matching the input | |
| 1094 spec. The function expects that each in_spec passed | |
| 1095 to it will be processed to completion; in particular, up to and | |
| 1096 including the call following that in which the last matching name | |
| 1097 is returned, the function ignores the value of in_spec, and will | |
| 1098 only start processing a new spec with the following call. | |
| 1099 If an error occurs, on return out_spec contains the value | |
| 1100 of in_spec when the error occurred. | |
| 1101 | |
| 1102 With each successive file name returned in out_spec, the | |
| 1103 function's return value is one. When there are no more matching | |
| 1104 names the function returns zero. If on the first call no file | |
| 1105 matches in_spec, or there is any other error, -1 is returned. | |
| 1106 */ | |
| 1107 | |
| 1108 #include <rmsdef.h> | |
| 1109 #include <descrip.h> | |
| 1110 #define OUTSIZE MAX_FILE_SPEC_LEN | |
| 442 | 1111 static short |
| 428 | 1112 fn_exp (out, in) |
| 1113 vspec *out; | |
| 1114 char *in; | |
| 1115 { | |
| 1116 static long context = 0; | |
| 1117 static struct dsc$descriptor_s o; | |
| 1118 static struct dsc$descriptor_s i; | |
| 1119 static bool pass1 = TRUE; | |
| 1120 long status; | |
| 1121 short retval; | |
| 1122 | |
| 1123 if (pass1) | |
| 1124 { | |
| 1125 pass1 = FALSE; | |
| 1126 o.dsc$a_pointer = (char *) out; | |
| 1127 o.dsc$w_length = (short)OUTSIZE; | |
| 1128 i.dsc$a_pointer = in; | |
| 1129 i.dsc$w_length = (short)strlen(in); | |
| 1130 i.dsc$b_dtype = DSC$K_DTYPE_T; | |
| 1131 i.dsc$b_class = DSC$K_CLASS_S; | |
| 1132 o.dsc$b_dtype = DSC$K_DTYPE_VT; | |
| 1133 o.dsc$b_class = DSC$K_CLASS_VS; | |
| 1134 } | |
| 1135 if ((status = lib$find_file(&i, &o, &context, 0, 0)) == RMS$_NORMAL) | |
| 1136 { | |
| 1137 out->body[out->curlen] = EOS; | |
| 1138 return 1; | |
| 1139 } | |
| 1140 else if (status == RMS$_NMF) | |
| 1141 retval = 0; | |
| 1142 else | |
| 1143 { | |
| 1144 strcpy(out->body, in); | |
| 1145 retval = -1; | |
| 1146 } | |
| 1147 lib$find_file_end(&context); | |
| 1148 pass1 = TRUE; | |
| 1149 return retval; | |
| 1150 } | |
| 1151 | |
| 1152 /* | |
| 1153 v1.01 nmm 19-Aug-85 gfnames - return in successive calls the | |
| 1154 name of each file specified by the provided arg expanding wildcards. | |
| 1155 */ | |
| 442 | 1156 static char * |
| 428 | 1157 gfnames (arg, p_error) |
| 1158 char *arg; | |
| 1159 bool *p_error; | |
| 1160 { | |
| 1161 static vspec filename = {MAX_FILE_SPEC_LEN, "\0"}; | |
| 1162 | |
| 1163 switch (fn_exp (&filename, arg)) | |
| 1164 { | |
| 1165 case 1: | |
| 1166 *p_error = FALSE; | |
| 1167 return filename.body; | |
| 1168 case 0: | |
| 1169 *p_error = FALSE; | |
| 1170 return NULL; | |
| 1171 default: | |
| 1172 *p_error = TRUE; | |
| 1173 return filename.body; | |
| 1174 } | |
| 1175 } | |
| 1176 | |
| 1177 #ifndef OLD /* Newer versions of VMS do provide `system'. */ | |
| 1178 system (cmd) | |
| 1179 char *cmd; | |
| 1180 { | |
| 1181 error ("%s", "system() function not implemented under VMS"); | |
| 1182 } | |
| 1183 #endif | |
| 1184 | |
| 1185 #define VERSION_DELIM ';' | |
| 1186 char *massage_name (s) | |
| 1187 char *s; | |
| 1188 { | |
| 1189 char *start = s; | |
| 1190 | |
| 1191 for ( ; *s; s++) | |
| 1192 if (*s == VERSION_DELIM) | |
| 1193 { | |
| 1194 *s = EOS; | |
| 1195 break; | |
| 1196 } | |
| 1197 else | |
| 1198 *s = lowcase (*s); | |
| 1199 return start; | |
| 1200 } | |
| 1201 #endif /* VMS */ | |
| 1202 | |
| 1203 | |
| 1204 int | |
| 1205 main (argc, argv) | |
| 1206 int argc; | |
| 1207 char *argv[]; | |
| 1208 { | |
| 1209 int i; | |
| 1210 unsigned int nincluded_files; | |
| 1211 char **included_files; | |
| 1212 argument *argbuffer; | |
| 1213 int current_arg, file_count; | |
| 1214 linebuffer filename_lb; | |
| 2225 | 1215 bool help_asked = FALSE; |
| 428 | 1216 #ifdef VMS |
| 1217 bool got_err; | |
| 1218 #endif | |
| 2225 | 1219 char *optstring; |
| 1220 int opt; | |
| 1221 | |
| 428 | 1222 |
| 458 | 1223 #ifdef DOS_NT |
| 428 | 1224 _fmode = O_BINARY; /* all of files are treated as binary files */ |
| 458 | 1225 #endif /* DOS_NT */ |
| 428 | 1226 |
| 1227 progname = argv[0]; | |
| 1228 nincluded_files = 0; | |
| 1229 included_files = xnew (argc, char *); | |
| 1230 current_arg = 0; | |
| 1231 file_count = 0; | |
| 1232 | |
| 1233 /* Allocate enough no matter what happens. Overkill, but each one | |
| 1234 is small. */ | |
| 1235 argbuffer = xnew (argc, argument); | |
| 1236 | |
| 1237 /* | |
| 1238 * If etags, always find typedefs and structure tags. Why not? | |
| 3972 | 1239 * Also default to find macro constants, enum constants, struct |
| 1240 * members and global variables. | |
| 428 | 1241 */ |
| 1242 if (!CTAGS) | |
| 1243 { | |
| 458 | 1244 typedefs = typedefs_or_cplusplus = constantypedefs = TRUE; |
| 3972 | 1245 globals = members = TRUE; |
| 428 | 1246 } |
| 1247 | |
| 2554 | 1248 /* When the optstring begins with a '-' getopt_long does not rearrange the |
| 1249 non-options arguments to be at the end, but leaves them alone. */ | |
| 3517 | 1250 optstring = concat (NO_LONG_OPTIONS ? "" : "-", |
| 1251 "ac:Cf:Il:o:r:RSVhH", | |
| 3090 | 1252 (CTAGS) ? "BxdtTuvw" : "Di:"); |
| 1253 | |
| 1254 while ((opt = getopt_long (argc, argv, optstring, longopts, NULL)) != EOF) | |
| 2225 | 1255 switch (opt) |
| 1256 { | |
| 1257 case 0: | |
| 1258 /* If getopt returns 0, then it has already processed a | |
| 1259 long-named option. We should do nothing. */ | |
| 428 | 1260 break; |
| 1261 | |
| 2225 | 1262 case 1: |
| 1263 /* This means that a file name has been seen. Record it. */ | |
| 1264 argbuffer[current_arg].arg_type = at_filename; | |
| 1265 argbuffer[current_arg].what = optarg; | |
| 1266 ++current_arg; | |
| 1267 ++file_count; | |
| 1268 break; | |
| 1269 | |
| 1270 case STDIN: | |
| 1271 /* Parse standard input. Idea by Vivek <vivek@etla.org>. */ | |
| 1272 argbuffer[current_arg].arg_type = at_stdin; | |
| 1273 argbuffer[current_arg].what = optarg; | |
| 1274 ++current_arg; | |
| 1275 ++file_count; | |
| 1276 if (parsing_stdin) | |
| 1277 fatal ("cannot parse standard input more than once", (char *)NULL); | |
| 1278 parsing_stdin = TRUE; | |
| 1279 break; | |
| 1280 | |
| 1281 /* Common options. */ | |
| 3090 | 1282 case 'a': append_to_tagfile = TRUE; break; |
| 2225 | 1283 case 'C': cplusplus = TRUE; break; |
| 1284 case 'f': /* for compatibility with old makefiles */ | |
| 1285 case 'o': | |
| 1286 if (tagfile) | |
| 428 | 1287 { |
| 2225 | 1288 error ("-o option may only be given once.", (char *)NULL); |
| 1289 suggest_asking_for_help (); | |
| 1290 /* NOTREACHED */ | |
| 428 | 1291 } |
| 2225 | 1292 tagfile = optarg; |
| 1293 break; | |
| 1294 case 'I': | |
| 1295 case 'S': /* for backward compatibility */ | |
| 1296 ignoreindent = TRUE; | |
| 1297 break; | |
| 1298 case 'l': | |
| 1299 { | |
| 1300 language *lang = get_language_from_langname (optarg); | |
| 1301 if (lang != NULL) | |
| 1302 { | |
| 1303 argbuffer[current_arg].lang = lang; | |
| 1304 argbuffer[current_arg].arg_type = at_language; | |
| 1305 ++current_arg; | |
| 1306 } | |
| 428 | 1307 } |
| 2225 | 1308 break; |
| 1309 case 'c': | |
| 1310 /* Backward compatibility: support obsolete --ignore-case-regexp. */ | |
| 1311 optarg = concat (optarg, "i", ""); /* memory leak here */ | |
| 1312 /* FALLTHRU */ | |
| 1313 case 'r': | |
| 1314 argbuffer[current_arg].arg_type = at_regexp; | |
| 1315 argbuffer[current_arg].what = optarg; | |
| 1316 ++current_arg; | |
| 1317 break; | |
| 1318 case 'R': | |
| 1319 argbuffer[current_arg].arg_type = at_regexp; | |
| 1320 argbuffer[current_arg].what = NULL; | |
| 1321 ++current_arg; | |
| 1322 break; | |
| 1323 case 'V': | |
| 1324 print_version (); | |
| 1325 break; | |
| 1326 case 'h': | |
| 1327 case 'H': | |
| 1328 help_asked = TRUE; | |
| 1329 break; | |
| 1330 | |
| 1331 /* Etags options */ | |
| 1332 case 'D': constantypedefs = FALSE; break; | |
| 1333 case 'i': included_files[nincluded_files++] = optarg; break; | |
| 1334 | |
| 1335 /* Ctags options. */ | |
| 1336 case 'B': searchar = '?'; break; | |
| 1337 case 'd': constantypedefs = TRUE; break; | |
| 1338 case 't': typedefs = TRUE; break; | |
| 1339 case 'T': typedefs = typedefs_or_cplusplus = TRUE; break; | |
| 1340 case 'u': update = TRUE; break; | |
| 1341 case 'v': vgrind_style = TRUE; /*FALLTHRU*/ | |
| 1342 case 'x': cxref_style = TRUE; break; | |
| 1343 case 'w': no_warnings = TRUE; break; | |
| 1344 default: | |
| 1345 suggest_asking_for_help (); | |
| 1346 /* NOTREACHED */ | |
| 1347 } | |
| 1348 | |
| 2554 | 1349 /* No more options. Store the rest of arguments. */ |
| 2225 | 1350 for (; optind < argc; optind++) |
| 428 | 1351 { |
| 1352 argbuffer[current_arg].arg_type = at_filename; | |
| 1353 argbuffer[current_arg].what = argv[optind]; | |
| 1354 ++current_arg; | |
| 1355 ++file_count; | |
| 1356 } | |
| 1357 | |
| 2225 | 1358 argbuffer[current_arg].arg_type = at_end; |
| 1359 | |
| 1360 if (help_asked) | |
| 1361 print_help (argbuffer); | |
| 1362 /* NOTREACHED */ | |
| 1363 | |
| 428 | 1364 if (nincluded_files == 0 && file_count == 0) |
| 1365 { | |
| 1366 error ("no input files specified.", (char *)NULL); | |
| 1367 suggest_asking_for_help (); | |
| 2225 | 1368 /* NOTREACHED */ |
| 428 | 1369 } |
| 1370 | |
| 1371 if (tagfile == NULL) | |
| 1372 tagfile = CTAGS ? "tags" : "TAGS"; | |
| 1373 cwd = etags_getcwd (); /* the current working directory */ | |
| 1374 if (cwd[strlen (cwd) - 1] != '/') | |
| 1375 { | |
| 1376 char *oldcwd = cwd; | |
| 1377 cwd = concat (oldcwd, "/", ""); | |
| 1378 free (oldcwd); | |
| 1379 } | |
| 2325 | 1380 /* Relative file names are made relative to the current directory. */ |
| 1381 if (streq (tagfile, "-") | |
| 1382 || strneq (tagfile, "/dev/", 5)) | |
| 428 | 1383 tagfiledir = cwd; |
| 1384 else | |
| 1385 tagfiledir = absolute_dirname (tagfile, cwd); | |
| 1386 | |
| 1387 init (); /* set up boolean "functions" */ | |
| 1388 | |
| 2225 | 1389 linebuffer_init (&lb); |
| 1390 linebuffer_init (&filename_lb); | |
| 1391 linebuffer_init (&filebuf); | |
| 1392 linebuffer_init (&token_name); | |
| 428 | 1393 |
| 1394 if (!CTAGS) | |
| 1395 { | |
| 1396 if (streq (tagfile, "-")) | |
| 1397 { | |
| 1398 tagf = stdout; | |
| 458 | 1399 #ifdef DOS_NT |
| 428 | 1400 /* Switch redirected `stdout' to binary mode (setting `_fmode' |
| 1401 doesn't take effect until after `stdout' is already open). */ | |
| 1402 if (!isatty (fileno (stdout))) | |
| 1403 setmode (fileno (stdout), O_BINARY); | |
| 458 | 1404 #endif /* DOS_NT */ |
| 428 | 1405 } |
| 1406 else | |
| 1407 tagf = fopen (tagfile, append_to_tagfile ? "a" : "w"); | |
| 1408 if (tagf == NULL) | |
| 1409 pfatal (tagfile); | |
| 1410 } | |
| 1411 | |
| 1412 /* | |
| 1413 * Loop through files finding functions. | |
| 1414 */ | |
| 2225 | 1415 for (i = 0; i < current_arg; i++) |
| 428 | 1416 { |
| 2225 | 1417 static language *lang; /* non-NULL if language is forced */ |
| 1418 char *this_file; | |
| 1419 | |
| 428 | 1420 switch (argbuffer[i].arg_type) |
| 1421 { | |
| 1422 case at_language: | |
| 2225 | 1423 lang = argbuffer[i].lang; |
| 428 | 1424 break; |
| 1425 case at_regexp: | |
| 2225 | 1426 analyse_regex (argbuffer[i].what); |
| 428 | 1427 break; |
| 1428 case at_filename: | |
| 1429 #ifdef VMS | |
| 1430 while ((this_file = gfnames (argbuffer[i].what, &got_err)) != NULL) | |
| 1431 { | |
| 1432 if (got_err) | |
| 1433 { | |
| 1434 error ("can't find file %s\n", this_file); | |
| 1435 argc--, argv++; | |
| 1436 } | |
| 1437 else | |
| 1438 { | |
| 1439 this_file = massage_name (this_file); | |
| 1440 } | |
| 1441 #else | |
| 1442 this_file = argbuffer[i].what; | |
| 1443 #endif | |
| 1444 /* Input file named "-" means read file names from stdin | |
| 1445 (one per line) and use them. */ | |
| 1446 if (streq (this_file, "-")) | |
| 2225 | 1447 { |
| 1448 if (parsing_stdin) | |
| 1449 fatal ("cannot parse standard input AND read file names from it", | |
| 1450 (char *)NULL); | |
| 1451 while (readline_internal (&filename_lb, stdin) > 0) | |
| 1452 process_file_name (filename_lb.buffer, lang); | |
| 1453 } | |
| 428 | 1454 else |
| 2225 | 1455 process_file_name (this_file, lang); |
| 428 | 1456 #ifdef VMS |
| 1457 } | |
| 1458 #endif | |
| 1459 break; | |
| 2225 | 1460 case at_stdin: |
| 1461 this_file = argbuffer[i].what; | |
| 1462 process_file (stdin, this_file, lang); | |
| 1463 break; | |
| 428 | 1464 } |
| 1465 } | |
| 1466 | |
| 2225 | 1467 free_regexps (); |
| 1468 free (lb.buffer); | |
| 1469 free (filebuf.buffer); | |
| 1470 free (token_name.buffer); | |
| 1471 | |
| 1472 if (!CTAGS || cxref_style) | |
| 428 | 1473 { |
| 3090 | 1474 /* Write the remaining tags to tagf (ETAGS) or stdout (CXREF). */ |
| 1475 put_entries (nodehead); | |
| 2225 | 1476 free_tree (nodehead); |
| 1477 nodehead = NULL; | |
| 1478 if (!CTAGS) | |
| 1479 { | |
| 1480 fdesc *fdp; | |
| 1481 | |
| 1482 /* Output file entries that have no tags. */ | |
| 1483 for (fdp = fdhead; fdp != NULL; fdp = fdp->next) | |
| 1484 if (!fdp->written) | |
| 1485 fprintf (tagf, "\f\n%s,0\n", fdp->taggedfname); | |
| 1486 | |
| 1487 while (nincluded_files-- > 0) | |
| 1488 fprintf (tagf, "\f\n%s,include\n", *included_files++); | |
| 3090 | 1489 |
| 1490 if (fclose (tagf) == EOF) | |
| 1491 pfatal (tagfile); | |
| 2225 | 1492 } |
| 1493 | |
| 1494 exit (EXIT_SUCCESS); | |
| 428 | 1495 } |
| 1496 | |
| 1497 if (update) | |
| 1498 { | |
| 1499 char cmd[BUFSIZ]; | |
| 1500 for (i = 0; i < current_arg; ++i) | |
| 1501 { | |
| 2225 | 1502 switch (argbuffer[i].arg_type) |
| 1503 { | |
| 1504 case at_filename: | |
| 1505 case at_stdin: | |
| 1506 break; | |
| 1507 default: | |
| 1508 continue; /* the for loop */ | |
| 1509 } | |
| 428 | 1510 sprintf (cmd, |
| 1511 "mv %s OTAGS;fgrep -v '\t%s\t' OTAGS >%s;rm OTAGS", | |
| 1512 tagfile, argbuffer[i].what, tagfile); | |
| 2225 | 1513 if (system (cmd) != EXIT_SUCCESS) |
| 428 | 1514 fatal ("failed to execute shell command", (char *)NULL); |
| 1515 } | |
| 1516 append_to_tagfile = TRUE; | |
| 1517 } | |
| 1518 | |
| 1519 tagf = fopen (tagfile, append_to_tagfile ? "a" : "w"); | |
| 1520 if (tagf == NULL) | |
| 1521 pfatal (tagfile); | |
| 2225 | 1522 put_entries (nodehead); /* write all the tags (CTAGS) */ |
| 1523 free_tree (nodehead); | |
| 1524 nodehead = NULL; | |
| 1525 if (fclose (tagf) == EOF) | |
| 1526 pfatal (tagfile); | |
| 428 | 1527 |
| 3090 | 1528 if (CTAGS) |
| 1529 if (append_to_tagfile || update) | |
| 1530 { | |
| 3876 | 1531 char cmd[2*BUFSIZ+20]; |
| 1532 /* Maybe these should be used: | |
| 1533 setenv ("LC_COLLATE", "C", 1); | |
| 1534 setenv ("LC_ALL", "C", 1); */ | |
| 1535 sprintf (cmd, "sort -u -o %.*s %.*s", BUFSIZ, tagfile, BUFSIZ, tagfile); | |
| 3090 | 1536 exit (system (cmd)); |
| 1537 } | |
| 2225 | 1538 return EXIT_SUCCESS; |
| 428 | 1539 } |
| 1540 | |
| 1541 | |
| 1542 /* | |
| 1543 * Return a compressor given the file name. If EXTPTR is non-zero, | |
| 1544 * return a pointer into FILE where the compressor-specific | |
| 1545 * extension begins. If no compressor is found, NULL is returned | |
| 1546 * and EXTPTR is not significant. | |
| 458 | 1547 * Idea by Vladimir Alexiev <vladimir@cs.ualberta.ca> (1998) |
| 428 | 1548 */ |
| 442 | 1549 static compressor * |
| 428 | 1550 get_compressor_from_suffix (file, extptr) |
| 1551 char *file; | |
| 1552 char **extptr; | |
| 1553 { | |
| 1554 compressor *compr; | |
| 1555 char *slash, *suffix; | |
| 1556 | |
| 1557 /* This relies on FN to be after canonicalize_filename, | |
| 458 | 1558 so we don't need to consider backslashes on DOS_NT. */ |
| 428 | 1559 slash = etags_strrchr (file, '/'); |
| 1560 suffix = etags_strrchr (file, '.'); | |
| 1561 if (suffix == NULL || suffix < slash) | |
| 1562 return NULL; | |
| 1563 if (extptr != NULL) | |
| 1564 *extptr = suffix; | |
| 1565 suffix += 1; | |
| 1566 /* Let those poor souls who live with DOS 8+3 file name limits get | |
| 1567 some solace by treating foo.cgz as if it were foo.c.gz, etc. | |
| 458 | 1568 Only the first do loop is run if not MSDOS */ |
| 428 | 1569 do |
| 1570 { | |
| 1571 for (compr = compressors; compr->suffix != NULL; compr++) | |
| 1572 if (streq (compr->suffix, suffix)) | |
| 1573 return compr; | |
| 458 | 1574 if (!MSDOS) |
| 442 | 1575 break; /* do it only once: not really a loop */ |
| 428 | 1576 if (extptr != NULL) |
| 1577 *extptr = ++suffix; | |
| 1578 } while (*suffix != '\0'); | |
| 1579 return NULL; | |
| 1580 } | |
| 1581 | |
| 1582 | |
| 1583 | |
| 1584 /* | |
| 1585 * Return a language given the name. | |
| 1586 */ | |
| 442 | 1587 static language * |
| 458 | 1588 get_language_from_langname (name) |
| 709 | 1589 const char *name; |
| 428 | 1590 { |
| 1591 language *lang; | |
| 1592 | |
| 1593 if (name == NULL) | |
| 1594 error ("empty language name", (char *)NULL); | |
| 1595 else | |
| 1596 { | |
| 1597 for (lang = lang_names; lang->name != NULL; lang++) | |
| 1598 if (streq (name, lang->name)) | |
| 1599 return lang; | |
| 1600 error ("unknown language \"%s\"", name); | |
| 1601 } | |
| 1602 | |
| 1603 return NULL; | |
| 1604 } | |
| 1605 | |
| 1606 | |
| 1607 /* | |
| 1608 * Return a language given the interpreter name. | |
| 1609 */ | |
| 442 | 1610 static language * |
| 428 | 1611 get_language_from_interpreter (interpreter) |
| 1612 char *interpreter; | |
| 1613 { | |
| 1614 language *lang; | |
| 1615 char **iname; | |
| 1616 | |
| 1617 if (interpreter == NULL) | |
| 1618 return NULL; | |
| 1619 for (lang = lang_names; lang->name != NULL; lang++) | |
| 1620 if (lang->interpreters != NULL) | |
| 1621 for (iname = lang->interpreters; *iname != NULL; iname++) | |
| 1622 if (streq (*iname, interpreter)) | |
| 1623 return lang; | |
| 1624 | |
| 1625 return NULL; | |
| 1626 } | |
| 1627 | |
| 1628 | |
| 1629 | |
| 1630 /* | |
| 1631 * Return a language given the file name. | |
| 1632 */ | |
| 442 | 1633 static language * |
| 2225 | 1634 get_language_from_filename (file, case_sensitive) |
| 428 | 1635 char *file; |
| 2225 | 1636 bool case_sensitive; |
| 428 | 1637 { |
| 1638 language *lang; | |
| 458 | 1639 char **name, **ext, *suffix; |
| 1640 | |
| 1641 /* Try whole file name first. */ | |
| 1642 for (lang = lang_names; lang->name != NULL; lang++) | |
| 1643 if (lang->filenames != NULL) | |
| 1644 for (name = lang->filenames; *name != NULL; name++) | |
| 2225 | 1645 if ((case_sensitive) |
| 1646 ? streq (*name, file) | |
| 1647 : strcaseeq (*name, file)) | |
| 458 | 1648 return lang; |
| 1649 | |
| 1650 /* If not found, try suffix after last dot. */ | |
| 428 | 1651 suffix = etags_strrchr (file, '.'); |
| 1652 if (suffix == NULL) | |
| 1653 return NULL; | |
| 1654 suffix += 1; | |
| 1655 for (lang = lang_names; lang->name != NULL; lang++) | |
| 1656 if (lang->suffixes != NULL) | |
| 1657 for (ext = lang->suffixes; *ext != NULL; ext++) | |
| 2225 | 1658 if ((case_sensitive) |
| 1659 ? streq (*ext, suffix) | |
| 1660 : strcaseeq (*ext, suffix)) | |
| 428 | 1661 return lang; |
| 1662 return NULL; | |
| 1663 } | |
| 1664 | |
| 2225 | 1665 |
| 428 | 1666 /* |
| 1667 * This routine is called on each file argument. | |
| 1668 */ | |
| 442 | 1669 static void |
| 2225 | 1670 process_file_name (file, lang) |
| 428 | 1671 char *file; |
| 2225 | 1672 language *lang; |
| 428 | 1673 { |
| 1674 struct stat stat_buf; | |
| 1675 FILE *inf; | |
| 2225 | 1676 fdesc *fdp; |
| 428 | 1677 compressor *compr; |
| 1678 char *compressed_name, *uncompressed_name; | |
| 1679 char *ext, *real_name; | |
| 2225 | 1680 int retval; |
| 428 | 1681 |
| 1682 canonicalize_filename (file); | |
| 1683 if (streq (file, tagfile) && !streq (tagfile, "-")) | |
| 1684 { | |
| 1685 error ("skipping inclusion of %s in self.", file); | |
| 1686 return; | |
| 1687 } | |
| 1688 if ((compr = get_compressor_from_suffix (file, &ext)) == NULL) | |
| 1689 { | |
| 1690 compressed_name = NULL; | |
| 1691 real_name = uncompressed_name = savestr (file); | |
| 1692 } | |
| 1693 else | |
| 1694 { | |
| 1695 real_name = compressed_name = savestr (file); | |
| 1696 uncompressed_name = savenstr (file, ext - file); | |
| 1697 } | |
| 1698 | |
| 2225 | 1699 /* If the canonicalized uncompressed name |
| 1700 has already been dealt with, skip it silently. */ | |
| 1701 for (fdp = fdhead; fdp != NULL; fdp = fdp->next) | |
| 428 | 1702 { |
| 2225 | 1703 assert (fdp->infname != NULL); |
| 1704 if (streq (uncompressed_name, fdp->infname)) | |
| 1705 goto cleanup; | |
| 1706 } | |
| 428 | 1707 |
| 1708 if (stat (real_name, &stat_buf) != 0) | |
| 1709 { | |
| 1710 /* Reset real_name and try with a different name. */ | |
| 1711 real_name = NULL; | |
| 1712 if (compressed_name != NULL) /* try with the given suffix */ | |
| 1713 { | |
| 1714 if (stat (uncompressed_name, &stat_buf) == 0) | |
| 1715 real_name = uncompressed_name; | |
| 1716 } | |
| 1717 else /* try all possible suffixes */ | |
| 1718 { | |
| 1719 for (compr = compressors; compr->suffix != NULL; compr++) | |
| 1720 { | |
| 1721 compressed_name = concat (file, ".", compr->suffix); | |
| 1722 if (stat (compressed_name, &stat_buf) != 0) | |
| 1723 { | |
| 458 | 1724 if (MSDOS) |
| 1725 { | |
| 1726 char *suf = compressed_name + strlen (file); | |
| 1727 size_t suflen = strlen (compr->suffix) + 1; | |
| 1728 for ( ; suf[1]; suf++, suflen--) | |
| 1729 { | |
| 1730 memmove (suf, suf + 1, suflen); | |
| 1731 if (stat (compressed_name, &stat_buf) == 0) | |
| 1732 { | |
| 1733 real_name = compressed_name; | |
| 1734 break; | |
| 1735 } | |
| 1736 } | |
| 1737 if (real_name != NULL) | |
| 1738 break; | |
| 1739 } /* MSDOS */ | |
| 428 | 1740 free (compressed_name); |
| 1741 compressed_name = NULL; | |
| 1742 } | |
| 1743 else | |
| 1744 { | |
| 1745 real_name = compressed_name; | |
| 1746 break; | |
| 1747 } | |
| 1748 } | |
| 1749 } | |
| 1750 if (real_name == NULL) | |
| 1751 { | |
| 1752 perror (file); | |
| 2225 | 1753 goto cleanup; |
| 428 | 1754 } |
| 1755 } /* try with a different name */ | |
| 1756 | |
| 1757 if (!S_ISREG (stat_buf.st_mode)) | |
| 1758 { | |
| 1759 error ("skipping %s: it is not a regular file.", real_name); | |
| 2225 | 1760 goto cleanup; |
| 428 | 1761 } |
| 1762 if (real_name == compressed_name) | |
| 1763 { | |
| 1764 char *cmd = concat (compr->command, " ", real_name); | |
| 458 | 1765 inf = (FILE *) popen (cmd, "r"); |
| 428 | 1766 free (cmd); |
| 1767 } | |
| 1768 else | |
| 1769 inf = fopen (real_name, "r"); | |
| 1770 if (inf == NULL) | |
| 1771 { | |
| 1772 perror (real_name); | |
| 2225 | 1773 goto cleanup; |
| 428 | 1774 } |
| 1775 | |
| 2225 | 1776 process_file (inf, uncompressed_name, lang); |
| 428 | 1777 |
| 1778 if (real_name == compressed_name) | |
| 2225 | 1779 retval = pclose (inf); |
| 428 | 1780 else |
| 2225 | 1781 retval = fclose (inf); |
| 1782 if (retval < 0) | |
| 1783 pfatal (file); | |
| 1784 | |
| 1785 cleanup: | |
| 1786 if (compressed_name) free (compressed_name); | |
| 1787 if (uncompressed_name) free (uncompressed_name); | |
| 1788 last_node = NULL; | |
| 1789 curfdp = NULL; | |
| 1790 return; | |
| 1791 } | |
| 1792 | |
| 1793 static void | |
| 1794 process_file (fh, fn, lang) | |
| 1795 FILE *fh; | |
| 1796 char *fn; | |
| 1797 language *lang; | |
| 1798 { | |
| 1799 static const fdesc emptyfdesc; | |
| 1800 fdesc *fdp; | |
| 1801 | |
| 1802 /* Create a new input file description entry. */ | |
| 1803 fdp = xnew (1, fdesc); | |
| 1804 *fdp = emptyfdesc; | |
| 1805 fdp->next = fdhead; | |
| 1806 fdp->infname = savestr (fn); | |
| 1807 fdp->lang = lang; | |
| 1808 fdp->infabsname = absolute_filename (fn, cwd); | |
| 1809 fdp->infabsdir = absolute_dirname (fn, cwd); | |
| 1810 if (filename_is_absolute (fn)) | |
| 428 | 1811 { |
| 2225 | 1812 /* An absolute file name. Canonicalize it. */ |
| 1813 fdp->taggedfname = absolute_filename (fn, NULL); | |
| 1814 } | |
| 1815 else | |
| 1816 { | |
| 1817 /* A file name relative to cwd. Make it relative | |
| 1818 to the directory of the tags file. */ | |
| 1819 fdp->taggedfname = relative_filename (fn, tagfiledir); | |
| 1820 } | |
| 1821 fdp->usecharno = TRUE; /* use char position when making tags */ | |
| 1822 fdp->prop = NULL; | |
| 1823 fdp->written = FALSE; /* not written on tags file yet */ | |
| 1824 | |
| 1825 fdhead = fdp; | |
| 1826 curfdp = fdhead; /* the current file description */ | |
| 1827 | |
| 1828 find_entries (fh); | |
| 1829 | |
| 1830 /* If not Ctags, and if this is not metasource and if it contained no #line | |
| 1831 directives, we can write the tags and free all nodes pointing to | |
| 1832 curfdp. */ | |
| 1833 if (!CTAGS | |
| 1834 && curfdp->usecharno /* no #line directives in this file */ | |
| 1835 && !curfdp->lang->metasource) | |
| 1836 { | |
| 1837 node *np, *prev; | |
| 1838 | |
| 1839 /* Look for the head of the sublist relative to this file. See add_node | |
| 1840 for the structure of the node tree. */ | |
| 1841 prev = NULL; | |
| 1842 for (np = nodehead; np != NULL; prev = np, np = np->left) | |
| 1843 if (np->fdp == curfdp) | |
| 1844 break; | |
| 1845 | |
| 1846 /* If we generated tags for this file, write and delete them. */ | |
| 1847 if (np != NULL) | |
| 428 | 1848 { |
| 2225 | 1849 /* This is the head of the last sublist, if any. The following |
| 1850 instructions depend on this being true. */ | |
| 1851 assert (np->left == NULL); | |
| 1852 | |
| 1853 assert (fdhead == curfdp); | |
| 1854 assert (last_node->fdp == curfdp); | |
| 1855 put_entries (np); /* write tags for file curfdp->taggedfname */ | |
| 1856 free_tree (np); /* remove the written nodes */ | |
| 1857 if (prev == NULL) | |
| 1858 nodehead = NULL; /* no nodes left */ | |
| 1859 else | |
| 1860 prev->left = NULL; /* delete the pointer to the sublist */ | |
| 428 | 1861 } |
| 1862 } | |
| 1863 } | |
| 1864 | |
| 1865 /* | |
| 1866 * This routine sets up the boolean pseudo-functions which work | |
| 1867 * by setting boolean flags dependent upon the corresponding character. | |
| 1868 * Every char which is NOT in that string is not a white char. Therefore, | |
| 1869 * all of the array "_wht" is set to FALSE, and then the elements | |
| 1870 * subscripted by the chars in "white" are set to TRUE. Thus "_wht" | |
| 1871 * of a char is TRUE if it is the string "white", else FALSE. | |
| 1872 */ | |
| 442 | 1873 static void |
| 428 | 1874 init () |
| 1875 { | |
| 1876 register char *sp; | |
| 1877 register int i; | |
| 1878 | |
| 1879 for (i = 0; i < CHARS; i++) | |
| 1880 iswhite(i) = notinname(i) = begtoken(i) = intoken(i) = endtoken(i) = FALSE; | |
| 1881 for (sp = white; *sp != '\0'; sp++) iswhite (*sp) = TRUE; | |
| 1882 for (sp = nonam; *sp != '\0'; sp++) notinname (*sp) = TRUE; | |
| 442 | 1883 notinname('\0') = notinname('\n'); |
| 428 | 1884 for (sp = begtk; *sp != '\0'; sp++) begtoken (*sp) = TRUE; |
| 442 | 1885 begtoken('\0') = begtoken('\n'); |
| 428 | 1886 for (sp = midtk; *sp != '\0'; sp++) intoken (*sp) = TRUE; |
| 442 | 1887 intoken('\0') = intoken('\n'); |
| 428 | 1888 for (sp = endtk; *sp != '\0'; sp++) endtoken (*sp) = TRUE; |
| 1889 endtoken('\0') = endtoken('\n'); | |
| 1890 } | |
| 1891 | |
| 1892 /* | |
| 1893 * This routine opens the specified file and calls the function | |
| 1894 * which finds the function and type definitions. | |
| 1895 */ | |
| 442 | 1896 static void |
| 2225 | 1897 find_entries (inf) |
| 428 | 1898 FILE *inf; |
| 1899 { | |
| 1900 char *cp; | |
| 2225 | 1901 language *lang = curfdp->lang; |
| 1902 Lang_function *parser = NULL; | |
| 428 | 1903 |
| 1904 /* If user specified a language, use it. */ | |
| 1905 if (lang != NULL && lang->function != NULL) | |
| 1906 { | |
| 2225 | 1907 parser = lang->function; |
| 428 | 1908 } |
| 1909 | |
| 2225 | 1910 /* Else try to guess the language given the file name. */ |
| 1911 if (parser == NULL) | |
| 428 | 1912 { |
| 2225 | 1913 lang = get_language_from_filename (curfdp->infname, TRUE); |
| 1914 if (lang != NULL && lang->function != NULL) | |
| 1915 { | |
| 1916 curfdp->lang = lang; | |
| 1917 parser = lang->function; | |
| 1918 } | |
| 428 | 1919 } |
| 1920 | |
| 2225 | 1921 /* Else look for sharp-bang as the first two characters. */ |
| 1922 if (parser == NULL | |
| 1923 && readline_internal (&lb, inf) > 0 | |
| 428 | 1924 && lb.len >= 2 |
| 1925 && lb.buffer[0] == '#' | |
| 1926 && lb.buffer[1] == '!') | |
| 1927 { | |
| 1928 char *lp; | |
| 1929 | |
| 1930 /* Set lp to point at the first char after the last slash in the | |
| 1931 line or, if no slashes, at the first nonblank. Then set cp to | |
| 1932 the first successive blank and terminate the string. */ | |
| 1933 lp = etags_strrchr (lb.buffer+2, '/'); | |
| 1934 if (lp != NULL) | |
| 1935 lp += 1; | |
| 1936 else | |
| 1937 lp = skip_spaces (lb.buffer + 2); | |
| 1938 cp = skip_non_spaces (lp); | |
| 1939 *cp = '\0'; | |
| 1940 | |
| 1941 if (strlen (lp) > 0) | |
| 1942 { | |
| 1943 lang = get_language_from_interpreter (lp); | |
| 1944 if (lang != NULL && lang->function != NULL) | |
| 1945 { | |
| 2225 | 1946 curfdp->lang = lang; |
| 1947 parser = lang->function; | |
| 428 | 1948 } |
| 1949 } | |
| 1950 } | |
| 2225 | 1951 |
| 428 | 1952 /* We rewind here, even if inf may be a pipe. We fail if the |
| 1953 length of the first line is longer than the pipe block size, | |
| 1954 which is unlikely. */ | |
| 1955 rewind (inf); | |
| 1956 | |
| 2225 | 1957 /* Else try to guess the language given the case insensitive file name. */ |
| 1958 if (parser == NULL) | |
| 1959 { | |
| 1960 lang = get_language_from_filename (curfdp->infname, FALSE); | |
| 1961 if (lang != NULL && lang->function != NULL) | |
| 1962 { | |
| 1963 curfdp->lang = lang; | |
| 1964 parser = lang->function; | |
| 1965 } | |
| 1966 } | |
| 1967 | |
| 1968 /* Else try Fortran or C. */ | |
| 1969 if (parser == NULL) | |
| 428 | 1970 { |
| 2225 | 1971 node *old_last_node = last_node; |
| 1972 | |
| 1973 curfdp->lang = get_language_from_langname ("fortran"); | |
| 1974 find_entries (inf); | |
| 1975 | |
| 1976 if (old_last_node == last_node) | |
| 1977 /* No Fortran entries found. Try C. */ | |
| 1978 { | |
| 1979 /* We do not tag if rewind fails. | |
| 1980 Only the file name will be recorded in the tags file. */ | |
| 1981 rewind (inf); | |
| 1982 curfdp->lang = get_language_from_langname (cplusplus ? "c++" : "c"); | |
| 1983 find_entries (inf); | |
| 1984 } | |
| 1985 return; | |
| 428 | 1986 } |
| 2225 | 1987 |
| 1988 if (!no_line_directive | |
| 1989 && curfdp->lang != NULL && curfdp->lang->metasource) | |
| 1990 /* It may be that this is a bingo.y file, and we already parsed a bingo.c | |
| 1991 file, or anyway we parsed a file that is automatically generated from | |
| 1992 this one. If this is the case, the bingo.c file contained #line | |
| 1993 directives that generated tags pointing to this file. Let's delete | |
| 1994 them all before parsing this file, which is the real source. */ | |
| 1995 { | |
| 1996 fdesc **fdpp = &fdhead; | |
| 1997 while (*fdpp != NULL) | |
| 1998 if (*fdpp != curfdp | |
| 1999 && streq ((*fdpp)->taggedfname, curfdp->taggedfname)) | |
| 2000 /* We found one of those! We must delete both the file description | |
| 2001 and all tags referring to it. */ | |
| 2002 { | |
| 2003 fdesc *badfdp = *fdpp; | |
| 2004 | |
| 2005 /* Delete the tags referring to badfdp->taggedfname | |
| 2006 that were obtained from badfdp->infname. */ | |
| 2007 invalidate_nodes (badfdp, &nodehead); | |
| 2008 | |
| 2009 *fdpp = badfdp->next; /* remove the bad description from the list */ | |
| 2010 free_fdesc (badfdp); | |
| 2011 } | |
| 2012 else | |
| 2013 fdpp = &(*fdpp)->next; /* advance the list pointer */ | |
| 2014 } | |
| 2015 | |
| 2016 assert (parser != NULL); | |
| 2017 | |
| 2018 /* Generic initialisations before reading from file. */ | |
| 2019 linebuffer_setlen (&filebuf, 0); /* reset the file buffer */ | |
| 2020 | |
| 2021 /* Generic initialisations before parsing file with readline. */ | |
| 2022 lineno = 0; /* reset global line number */ | |
| 2023 charno = 0; /* reset global char number */ | |
| 2024 linecharno = 0; /* reset global char number of line start */ | |
| 2025 | |
| 2026 parser (inf); | |
| 2027 | |
| 2028 regex_tag_multiline (); | |
| 428 | 2029 } |
| 458 | 2030 |
| 428 | 2031 |
| 2225 | 2032 /* |
| 2033 * Check whether an implicitly named tag should be created, | |
| 2034 * then call `pfnote'. | |
| 2035 * NAME is a string that is internally copied by this function. | |
| 2036 * | |
| 2037 * TAGS format specification | |
| 2038 * Idea by Sam Kendall <kendall@mv.mv.com> (1997) | |
| 2039 * The following is explained in some more detail in etc/ETAGS.EBNF. | |
| 2040 * | |
| 2041 * make_tag creates tags with "implicit tag names" (unnamed tags) | |
| 2042 * if the following are all true, assuming NONAM=" \f\t\n\r()=,;": | |
| 2043 * 1. NAME does not contain any of the characters in NONAM; | |
| 2044 * 2. LINESTART contains name as either a rightmost, or rightmost but | |
| 2045 * one character, substring; | |
| 2046 * 3. the character, if any, immediately before NAME in LINESTART must | |
| 2047 * be a character in NONAM; | |
| 2048 * 4. the character, if any, immediately after NAME in LINESTART must | |
| 2049 * also be a character in NONAM. | |
| 2050 * | |
| 2051 * The implementation uses the notinname() macro, which recognises the | |
| 2052 * characters stored in the string `nonam'. | |
| 2053 * etags.el needs to use the same characters that are in NONAM. | |
| 2054 */ | |
| 2055 static void | |
| 2056 make_tag (name, namelen, is_func, linestart, linelen, lno, cno) | |
| 2057 char *name; /* tag name, or NULL if unnamed */ | |
| 2058 int namelen; /* tag length */ | |
| 2059 bool is_func; /* tag is a function */ | |
| 2060 char *linestart; /* start of the line where tag is */ | |
| 2061 int linelen; /* length of the line where tag is */ | |
| 2062 int lno; /* line number */ | |
| 2063 long cno; /* character number */ | |
| 2064 { | |
| 2065 bool named = (name != NULL && namelen > 0); | |
| 2066 | |
| 2067 if (!CTAGS && named) /* maybe set named to false */ | |
| 2068 /* Let's try to make an implicit tag name, that is, create an unnamed tag | |
| 2069 such that etags.el can guess a name from it. */ | |
| 2070 { | |
| 2071 int i; | |
| 2072 register char *cp = name; | |
| 2073 | |
| 2074 for (i = 0; i < namelen; i++) | |
| 2075 if (notinname (*cp++)) | |
| 2076 break; | |
| 2077 if (i == namelen) /* rule #1 */ | |
| 2078 { | |
| 2079 cp = linestart + linelen - namelen; | |
| 2080 if (notinname (linestart[linelen-1])) | |
| 2081 cp -= 1; /* rule #4 */ | |
| 2082 if (cp >= linestart /* rule #2 */ | |
| 2083 && (cp == linestart | |
| 2084 || notinname (cp[-1])) /* rule #3 */ | |
| 2085 && strneq (name, cp, namelen)) /* rule #2 */ | |
| 2086 named = FALSE; /* use implicit tag name */ | |
| 2087 } | |
| 2088 } | |
| 2089 | |
| 2090 if (named) | |
| 2091 name = savenstr (name, namelen); | |
| 2092 else | |
| 2093 name = NULL; | |
| 2094 pfnote (name, is_func, linestart, linelen, lno, cno); | |
| 2095 } | |
| 2096 | |
| 428 | 2097 /* Record a tag. */ |
| 442 | 2098 static void |
| 428 | 2099 pfnote (name, is_func, linestart, linelen, lno, cno) |
| 2100 char *name; /* tag name, or NULL if unnamed */ | |
| 2101 bool is_func; /* tag is a function */ | |
| 2102 char *linestart; /* start of the line where tag is */ | |
| 2103 int linelen; /* length of the line where tag is */ | |
| 2104 int lno; /* line number */ | |
| 2105 long cno; /* character number */ | |
| 2106 { | |
| 2107 register node *np; | |
| 2108 | |
| 2225 | 2109 assert (name == NULL || name[0] != '\0'); |
| 428 | 2110 if (CTAGS && name == NULL) |
| 2111 return; | |
| 2112 | |
| 2113 np = xnew (1, node); | |
| 2114 | |
| 2115 /* If ctags mode, change name "main" to M<thisfilename>. */ | |
| 2116 if (CTAGS && !cxref_style && streq (name, "main")) | |
| 2117 { | |
| 2225 | 2118 register char *fp = etags_strrchr (curfdp->taggedfname, '/'); |
| 2119 np->name = concat ("M", fp == NULL ? curfdp->taggedfname : fp + 1, ""); | |
| 428 | 2120 fp = etags_strrchr (np->name, '.'); |
| 2121 if (fp != NULL && fp[1] != '\0' && fp[2] == '\0') | |
| 2122 fp[0] = '\0'; | |
| 2123 } | |
| 2124 else | |
| 2125 np->name = name; | |
| 2225 | 2126 np->valid = TRUE; |
| 428 | 2127 np->been_warned = FALSE; |
| 2225 | 2128 np->fdp = curfdp; |
| 428 | 2129 np->is_func = is_func; |
| 2130 np->lno = lno; | |
| 2225 | 2131 if (np->fdp->usecharno) |
| 2132 /* Our char numbers are 0-base, because of C language tradition? | |
| 2133 ctags compatibility? old versions compatibility? I don't know. | |
| 2134 Anyway, since emacs's are 1-base we expect etags.el to take care | |
| 2135 of the difference. If we wanted to have 1-based numbers, we would | |
| 2136 uncomment the +1 below. */ | |
| 2137 np->cno = cno /* + 1 */ ; | |
| 2138 else | |
| 2139 np->cno = invalidcharno; | |
| 428 | 2140 np->left = np->right = NULL; |
| 2141 if (CTAGS && !cxref_style) | |
| 2142 { | |
| 2143 if (strlen (linestart) < 50) | |
| 2225 | 2144 np->regex = concat (linestart, "$", ""); |
| 428 | 2145 else |
| 2225 | 2146 np->regex = savenstr (linestart, 50); |
| 428 | 2147 } |
| 2148 else | |
| 2225 | 2149 np->regex = savenstr (linestart, linelen); |
| 2150 | |
| 2151 add_node (np, &nodehead); | |
| 428 | 2152 } |
| 2153 | |
| 2154 /* | |
| 2155 * free_tree () | |
| 2156 * recurse on left children, iterate on right children. | |
| 2157 */ | |
| 442 | 2158 static void |
| 428 | 2159 free_tree (np) |
| 2160 register node *np; | |
| 2161 { | |
| 2162 while (np) | |
| 2163 { | |
| 2164 register node *node_right = np->right; | |
| 2165 free_tree (np->left); | |
| 2166 if (np->name != NULL) | |
| 2167 free (np->name); | |
| 2225 | 2168 free (np->regex); |
| 428 | 2169 free (np); |
| 2170 np = node_right; | |
| 2171 } | |
| 2172 } | |
| 2173 | |
| 2174 /* | |
| 2225 | 2175 * free_fdesc () |
| 2176 * delete a file description | |
| 2177 */ | |
| 2178 static void | |
| 2179 free_fdesc (fdp) | |
| 2180 register fdesc *fdp; | |
| 2181 { | |
| 2182 if (fdp->infname != NULL) free (fdp->infname); | |
| 2183 if (fdp->infabsname != NULL) free (fdp->infabsname); | |
| 2184 if (fdp->infabsdir != NULL) free (fdp->infabsdir); | |
| 2185 if (fdp->taggedfname != NULL) free (fdp->taggedfname); | |
| 2186 if (fdp->prop != NULL) free (fdp->prop); | |
| 2187 free (fdp); | |
| 2188 } | |
| 2189 | |
| 2190 /* | |
| 428 | 2191 * add_node () |
| 2225 | 2192 * Adds a node to the tree of nodes. In etags mode, sort by file |
| 2193 * name. In ctags mode, sort by tag name. Make no attempt at | |
| 2194 * balancing. | |
| 428 | 2195 * |
| 2196 * add_node is the only function allowed to add nodes, so it can | |
| 2197 * maintain state. | |
| 2198 */ | |
| 442 | 2199 static void |
| 428 | 2200 add_node (np, cur_node_p) |
| 2201 node *np, **cur_node_p; | |
| 2202 { | |
| 2203 register int dif; | |
| 2204 register node *cur_node = *cur_node_p; | |
| 2205 | |
| 2206 if (cur_node == NULL) | |
| 2207 { | |
| 2208 *cur_node_p = np; | |
| 2209 last_node = np; | |
| 2210 return; | |
| 2211 } | |
| 2212 | |
| 2213 if (!CTAGS) | |
| 2225 | 2214 /* Etags Mode */ |
| 428 | 2215 { |
| 2225 | 2216 /* For each file name, tags are in a linked sublist on the right |
| 2217 pointer. The first tags of different files are a linked list | |
| 2218 on the left pointer. last_node points to the end of the last | |
| 2219 used sublist. */ | |
| 2220 if (last_node != NULL && last_node->fdp == np->fdp) | |
| 2221 { | |
| 2222 /* Let's use the same sublist as the last added node. */ | |
| 2223 assert (last_node->right == NULL); | |
| 2224 last_node->right = np; | |
| 2225 last_node = np; | |
| 2226 } | |
| 2227 else if (cur_node->fdp == np->fdp) | |
| 2228 { | |
| 2229 /* Scanning the list we found the head of a sublist which is | |
| 2230 good for us. Let's scan this sublist. */ | |
| 2231 add_node (np, &cur_node->right); | |
| 2232 } | |
| 2233 else | |
| 2234 /* The head of this sublist is not good for us. Let's try the | |
| 2235 next one. */ | |
| 2236 add_node (np, &cur_node->left); | |
| 2237 } /* if ETAGS mode */ | |
| 2238 | |
| 428 | 2239 else |
| 2240 { | |
| 2241 /* Ctags Mode */ | |
| 2242 dif = strcmp (np->name, cur_node->name); | |
| 2243 | |
| 2244 /* | |
| 2245 * If this tag name matches an existing one, then | |
| 2246 * do not add the node, but maybe print a warning. | |
| 2247 */ | |
| 3876 | 2248 if (no_duplicates && !dif) |
| 428 | 2249 { |
| 2225 | 2250 if (np->fdp == cur_node->fdp) |
| 428 | 2251 { |
| 2252 if (!no_warnings) | |
| 2253 { | |
| 2254 fprintf (stderr, "Duplicate entry in file %s, line %d: %s\n", | |
| 2225 | 2255 np->fdp->infname, lineno, np->name); |
| 428 | 2256 fprintf (stderr, "Second entry ignored\n"); |
| 2257 } | |
| 2258 } | |
| 2259 else if (!cur_node->been_warned && !no_warnings) | |
| 2260 { | |
| 2261 fprintf | |
| 2262 (stderr, | |
| 2263 "Duplicate entry in files %s and %s: %s (Warning only)\n", | |
| 2225 | 2264 np->fdp->infname, cur_node->fdp->infname, np->name); |
| 428 | 2265 cur_node->been_warned = TRUE; |
| 2266 } | |
| 2267 return; | |
| 2268 } | |
| 2269 | |
| 2270 /* Actually add the node */ | |
| 2271 add_node (np, dif < 0 ? &cur_node->left : &cur_node->right); | |
| 2225 | 2272 } /* if CTAGS mode */ |
| 2273 } | |
| 2274 | |
| 2275 /* | |
| 2276 * invalidate_nodes () | |
| 2277 * Scan the node tree and invalidate all nodes pointing to the | |
| 2278 * given file description (CTAGS case) or free them (ETAGS case). | |
| 2279 */ | |
| 2280 static void | |
| 2281 invalidate_nodes (badfdp, npp) | |
| 2282 fdesc *badfdp; | |
| 2283 node **npp; | |
| 2284 { | |
| 2285 node *np = *npp; | |
| 2286 | |
| 2287 if (np == NULL) | |
| 2288 return; | |
| 2289 | |
| 2290 if (CTAGS) | |
| 2291 { | |
| 2292 if (np->left != NULL) | |
| 2293 invalidate_nodes (badfdp, &np->left); | |
| 2294 if (np->fdp == badfdp) | |
| 2295 np->valid = FALSE; | |
| 2296 if (np->right != NULL) | |
| 2297 invalidate_nodes (badfdp, &np->right); | |
| 2298 } | |
| 2299 else | |
| 2300 { | |
| 2301 assert (np->fdp != NULL); | |
| 2302 if (np->fdp == badfdp) | |
| 2303 { | |
| 2304 *npp = np->left; /* detach the sublist from the list */ | |
| 2305 np->left = NULL; /* isolate it */ | |
| 2306 free_tree (np); /* free it */ | |
| 2307 invalidate_nodes (badfdp, npp); | |
| 2308 } | |
| 2309 else | |
| 2310 invalidate_nodes (badfdp, &np->left); | |
| 428 | 2311 } |
| 2312 } | |
| 458 | 2313 |
| 428 | 2314 |
| 2225 | 2315 static int total_size_of_entries __P((node *)); |
| 2316 static int number_len __P((long)); | |
| 2317 | |
| 2318 /* Length of a non-negative number's decimal representation. */ | |
| 428 | 2319 static int |
| 2320 number_len (num) | |
| 2321 long num; | |
| 2322 { | |
| 2323 int len = 1; | |
| 2324 while ((num /= 10) > 0) | |
| 2325 len += 1; | |
| 2326 return len; | |
| 2327 } | |
| 2328 | |
| 2329 /* | |
| 2330 * Return total number of characters that put_entries will output for | |
| 2225 | 2331 * the nodes in the linked list at the right of the specified node. |
| 2332 * This count is irrelevant with etags.el since emacs 19.34 at least, | |
| 2333 * but is still supplied for backward compatibility. | |
| 428 | 2334 */ |
| 442 | 2335 static int |
| 428 | 2336 total_size_of_entries (np) |
| 2337 register node *np; | |
| 2338 { | |
| 2225 | 2339 register int total = 0; |
| 2340 | |
| 2341 for (; np != NULL; np = np->right) | |
| 2342 if (np->valid) | |
| 2343 { | |
| 2344 total += strlen (np->regex) + 1; /* pat\177 */ | |
| 2345 if (np->name != NULL) | |
| 2346 total += strlen (np->name) + 1; /* name\001 */ | |
| 2347 total += number_len ((long) np->lno) + 1; /* lno, */ | |
| 2348 if (np->cno != invalidcharno) /* cno */ | |
| 2349 total += number_len (np->cno); | |
| 2350 total += 1; /* newline */ | |
| 2351 } | |
| 2352 | |
| 2353 return total; | |
| 2354 } | |
| 2355 | |
| 2356 static void | |
| 2357 put_entries (np) | |
| 2358 register node *np; | |
| 2359 { | |
| 2360 register char *sp; | |
| 2361 static fdesc *fdp = NULL; | |
| 428 | 2362 |
| 2363 if (np == NULL) | |
| 2225 | 2364 return; |
| 2365 | |
| 2366 /* Output subentries that precede this one */ | |
| 2367 if (CTAGS) | |
| 2368 put_entries (np->left); | |
| 2369 | |
| 2370 /* Output this entry */ | |
| 2371 if (np->valid) | |
| 428 | 2372 { |
| 2225 | 2373 if (!CTAGS) |
| 2374 { | |
| 2375 /* Etags mode */ | |
| 2376 if (fdp != np->fdp) | |
| 2377 { | |
| 2378 fdp = np->fdp; | |
| 2379 fprintf (tagf, "\f\n%s,%d\n", | |
| 2380 fdp->taggedfname, total_size_of_entries (np)); | |
| 2381 fdp->written = TRUE; | |
| 2382 } | |
| 2383 fputs (np->regex, tagf); | |
| 2384 fputc ('\177', tagf); | |
| 2385 if (np->name != NULL) | |
| 2386 { | |
| 2387 fputs (np->name, tagf); | |
| 2388 fputc ('\001', tagf); | |
| 2389 } | |
| 2390 fprintf (tagf, "%d,", np->lno); | |
| 2391 if (np->cno != invalidcharno) | |
| 2392 fprintf (tagf, "%ld", np->cno); | |
| 2393 fputs ("\n", tagf); | |
| 2394 } | |
| 2395 else | |
| 2396 { | |
| 2397 /* Ctags mode */ | |
| 2398 if (np->name == NULL) | |
| 2399 error ("internal error: NULL name in ctags mode.", (char *)NULL); | |
| 2400 | |
| 2401 if (cxref_style) | |
| 2402 { | |
| 2403 if (vgrind_style) | |
| 2404 fprintf (stdout, "%s %s %d\n", | |
| 2405 np->name, np->fdp->taggedfname, (np->lno + 63) / 64); | |
| 2406 else | |
| 2407 fprintf (stdout, "%-16s %3d %-16s %s\n", | |
| 2408 np->name, np->lno, np->fdp->taggedfname, np->regex); | |
| 2409 } | |
| 2410 else | |
| 2411 { | |
| 2412 fprintf (tagf, "%s\t%s\t", np->name, np->fdp->taggedfname); | |
| 2413 | |
| 2414 if (np->is_func) | |
| 2415 { /* function or #define macro with args */ | |
| 2416 putc (searchar, tagf); | |
| 2417 putc ('^', tagf); | |
| 2418 | |
| 2419 for (sp = np->regex; *sp; sp++) | |
| 2420 { | |
| 2421 if (*sp == '\\' || *sp == searchar) | |
| 2422 putc ('\\', tagf); | |
| 2423 putc (*sp, tagf); | |
| 2424 } | |
| 2425 putc (searchar, tagf); | |
| 2426 } | |
| 2427 else | |
| 2428 { /* anything else; text pattern inadequate */ | |
| 2429 fprintf (tagf, "%d", np->lno); | |
| 2430 } | |
| 2431 putc ('\n', tagf); | |
| 2432 } | |
| 2433 } | |
| 2434 } /* if this node contains a valid tag */ | |
| 2435 | |
| 2436 /* Output subentries that follow this one */ | |
| 2437 put_entries (np->right); | |
| 2438 if (!CTAGS) | |
| 2439 put_entries (np->left); | |
| 428 | 2440 } |
| 458 | 2441 |
| 428 | 2442 |
| 458 | 2443 /* C extensions. */ |
| 2444 #define C_EXT 0x00fff /* C extensions */ | |
| 2445 #define C_PLAIN 0x00000 /* C */ | |
| 2446 #define C_PLPL 0x00001 /* C++ */ | |
| 2447 #define C_STAR 0x00003 /* C* */ | |
| 2448 #define C_JAVA 0x00005 /* JAVA */ | |
| 2449 #define C_AUTO 0x01000 /* C, but switch to C++ if `class' is met */ | |
| 2450 #define YACC 0x10000 /* yacc file */ | |
| 2451 | |
| 428 | 2452 /* |
| 2453 * The C symbol tables. | |
| 2454 */ | |
| 2455 enum sym_type | |
| 2456 { | |
| 2457 st_none, | |
| 2458 st_C_objprot, st_C_objimpl, st_C_objend, | |
| 2459 st_C_gnumacro, | |
| 2325 | 2460 st_C_ignore, st_C_attribute, |
| 428 | 2461 st_C_javastruct, |
| 2462 st_C_operator, | |
| 531 | 2463 st_C_class, st_C_template, |
| 2325 | 2464 st_C_struct, st_C_extern, st_C_enum, st_C_define, st_C_typedef |
| 428 | 2465 }; |
| 2466 | |
| 709 | 2467 static unsigned int hash __P((const char *, unsigned int)); |
| 2468 static struct C_stab_entry * in_word_set __P((const char *, unsigned int)); | |
| 2469 static enum sym_type C_symtype __P((char *, int, int)); | |
| 442 | 2470 |
| 428 | 2471 /* Feed stuff between (but not including) %[ and %] lines to: |
| 2325 | 2472 gperf -m 5 |
| 428 | 2473 %[ |
| 2325 | 2474 %compare-strncmp |
| 2475 %enum | |
| 2476 %struct-type | |
| 428 | 2477 struct C_stab_entry { char *name; int c_ext; enum sym_type type; } |
| 2478 %% | |
| 2325 | 2479 if, 0, st_C_ignore |
| 2480 for, 0, st_C_ignore | |
| 2481 while, 0, st_C_ignore | |
| 2482 switch, 0, st_C_ignore | |
| 2483 return, 0, st_C_ignore | |
| 2484 __attribute__, 0, st_C_attribute | |
| 2485 @interface, 0, st_C_objprot | |
| 2486 @protocol, 0, st_C_objprot | |
| 2487 @implementation,0, st_C_objimpl | |
| 2488 @end, 0, st_C_objend | |
| 3876 | 2489 import, (C_JAVA & ~C_PLPL), st_C_ignore |
| 2490 package, (C_JAVA & ~C_PLPL), st_C_ignore | |
| 2325 | 2491 friend, C_PLPL, st_C_ignore |
| 3876 | 2492 extends, (C_JAVA & ~C_PLPL), st_C_javastruct |
| 2493 implements, (C_JAVA & ~C_PLPL), st_C_javastruct | |
| 2494 interface, (C_JAVA & ~C_PLPL), st_C_struct | |
| 2325 | 2495 class, 0, st_C_class |
| 2496 namespace, C_PLPL, st_C_struct | |
| 2497 domain, C_STAR, st_C_struct | |
| 2498 union, 0, st_C_struct | |
| 2499 struct, 0, st_C_struct | |
| 2500 extern, 0, st_C_extern | |
| 2501 enum, 0, st_C_enum | |
| 2502 typedef, 0, st_C_typedef | |
| 2503 define, 0, st_C_define | |
| 3517 | 2504 undef, 0, st_C_define |
| 2325 | 2505 operator, C_PLPL, st_C_operator |
| 2506 template, 0, st_C_template | |
| 428 | 2507 # DEFUN used in emacs, the next three used in glibc (SYSCALL only for mach). |
| 2325 | 2508 DEFUN, 0, st_C_gnumacro |
| 2509 SYSCALL, 0, st_C_gnumacro | |
| 2510 ENTRY, 0, st_C_gnumacro | |
| 2511 PSEUDO, 0, st_C_gnumacro | |
| 428 | 2512 # These are defined inside C functions, so currently they are not met. |
| 2513 # EXFUN used in glibc, DEFVAR_* in emacs. | |
| 2325 | 2514 #EXFUN, 0, st_C_gnumacro |
| 2515 #DEFVAR_, 0, st_C_gnumacro | |
| 428 | 2516 %] |
| 2325 | 2517 and replace lines between %< and %> with its output, then: |
| 2518 - remove the #if characterset check | |
| 2519 - make in_word_set static and not inline. */ | |
| 428 | 2520 /*%<*/ |
| 2325 | 2521 /* C code produced by gperf version 3.0.1 */ |
| 2522 /* Command-line: gperf -m 5 */ | |
| 3517 | 2523 /* Computed positions: -k'2-3' */ |
| 2325 | 2524 |
| 428 | 2525 struct C_stab_entry { char *name; int c_ext; enum sym_type type; }; |
| 3517 | 2526 /* maximum key range = 33, duplicates = 0 */ |
| 428 | 2527 |
| 2528 #ifdef __GNUC__ | |
| 2529 __inline | |
| 2325 | 2530 #else |
| 2531 #ifdef __cplusplus | |
| 2532 inline | |
| 2533 #endif | |
| 428 | 2534 #endif |
| 2535 static unsigned int | |
| 2536 hash (str, len) | |
| 2537 register const char *str; | |
| 2538 register unsigned int len; | |
| 2539 { | |
| 2540 static unsigned char asso_values[] = | |
| 2541 { | |
| 3517 | 2542 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, |
| 2543 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, | |
| 2544 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, | |
| 2545 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, | |
| 2546 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, | |
| 2547 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, | |
| 2548 35, 35, 35, 35, 35, 35, 35, 35, 35, 15, | |
| 2549 14, 35, 35, 35, 35, 35, 35, 35, 14, 35, | |
| 2550 35, 35, 35, 12, 13, 35, 35, 35, 35, 12, | |
| 2551 35, 35, 35, 35, 35, 1, 35, 16, 35, 6, | |
| 2552 23, 0, 0, 35, 22, 0, 35, 35, 5, 0, | |
| 2553 0, 15, 1, 35, 6, 35, 8, 19, 35, 16, | |
| 2554 4, 5, 35, 35, 35, 35, 35, 35, 35, 35, | |
| 2555 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, | |
| 2556 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, | |
| 2557 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, | |
| 2558 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, | |
| 2559 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, | |
| 2560 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, | |
| 2561 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, | |
| 2562 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, | |
| 2563 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, | |
| 2564 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, | |
| 2565 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, | |
| 2566 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, | |
| 2567 35, 35, 35, 35, 35, 35 | |
| 428 | 2568 }; |
| 3517 | 2569 register int hval = len; |
| 2570 | |
| 2571 switch (hval) | |
| 2572 { | |
| 2573 default: | |
| 2574 hval += asso_values[(unsigned char)str[2]]; | |
| 2575 /*FALLTHROUGH*/ | |
| 2576 case 2: | |
| 2577 hval += asso_values[(unsigned char)str[1]]; | |
| 2578 break; | |
| 2579 } | |
| 2580 return hval; | |
| 428 | 2581 } |
| 2582 | |
| 2583 static struct C_stab_entry * | |
| 2584 in_word_set (str, len) | |
| 2585 register const char *str; | |
| 2586 register unsigned int len; | |
| 2587 { | |
| 2325 | 2588 enum |
| 2589 { | |
| 3517 | 2590 TOTAL_KEYWORDS = 32, |
| 2325 | 2591 MIN_WORD_LENGTH = 2, |
| 2592 MAX_WORD_LENGTH = 15, | |
| 3517 | 2593 MIN_HASH_VALUE = 2, |
| 2594 MAX_HASH_VALUE = 34 | |
| 2325 | 2595 }; |
| 2596 | |
| 428 | 2597 static struct C_stab_entry wordlist[] = |
| 2598 { | |
| 3517 | 2599 {""}, {""}, |
| 2325 | 2600 {"if", 0, st_C_ignore}, |
| 3517 | 2601 {""}, |
| 2325 | 2602 {"@end", 0, st_C_objend}, |
| 3517 | 2603 {"union", 0, st_C_struct}, |
| 2604 {"define", 0, st_C_define}, | |
| 3876 | 2605 {"import", (C_JAVA & ~C_PLPL), st_C_ignore}, |
| 3517 | 2606 {"template", 0, st_C_template}, |
| 2607 {"operator", C_PLPL, st_C_operator}, | |
| 2608 {"@interface", 0, st_C_objprot}, | |
| 3876 | 2609 {"implements", (C_JAVA & ~C_PLPL), st_C_javastruct}, |
| 3517 | 2610 {"friend", C_PLPL, st_C_ignore}, |
| 2611 {"typedef", 0, st_C_typedef}, | |
| 2612 {"return", 0, st_C_ignore}, | |
| 2613 {"@implementation",0, st_C_objimpl}, | |
| 2614 {"@protocol", 0, st_C_objprot}, | |
| 3876 | 2615 {"interface", (C_JAVA & ~C_PLPL), st_C_struct}, |
| 2325 | 2616 {"extern", 0, st_C_extern}, |
| 3876 | 2617 {"extends", (C_JAVA & ~C_PLPL), st_C_javastruct}, |
| 3517 | 2618 {"struct", 0, st_C_struct}, |
| 2619 {"domain", C_STAR, st_C_struct}, | |
| 2620 {"switch", 0, st_C_ignore}, | |
| 2621 {"enum", 0, st_C_enum}, | |
| 2325 | 2622 {"for", 0, st_C_ignore}, |
| 3517 | 2623 {"namespace", C_PLPL, st_C_struct}, |
| 2624 {"class", 0, st_C_class}, | |
| 2625 {"while", 0, st_C_ignore}, | |
| 2626 {"undef", 0, st_C_define}, | |
| 3876 | 2627 {"package", (C_JAVA & ~C_PLPL), st_C_ignore}, |
| 2325 | 2628 {"__attribute__", 0, st_C_attribute}, |
| 2629 {"SYSCALL", 0, st_C_gnumacro}, | |
| 3517 | 2630 {"ENTRY", 0, st_C_gnumacro}, |
| 2325 | 2631 {"PSEUDO", 0, st_C_gnumacro}, |
| 2632 {"DEFUN", 0, st_C_gnumacro} | |
| 428 | 2633 }; |
| 2634 | |
| 2635 if (len <= MAX_WORD_LENGTH && len >= MIN_WORD_LENGTH) | |
| 2636 { | |
| 2637 register int key = hash (str, len); | |
| 2638 | |
| 2639 if (key <= MAX_HASH_VALUE && key >= 0) | |
| 2640 { | |
| 2641 register const char *s = wordlist[key].name; | |
| 2642 | |
| 2325 | 2643 if (*str == *s && !strncmp (str + 1, s + 1, len - 1) && s[len] == '\0') |
| 428 | 2644 return &wordlist[key]; |
| 2645 } | |
| 2646 } | |
| 2647 return 0; | |
| 2648 } | |
| 2649 /*%>*/ | |
| 2650 | |
| 2651 static enum sym_type | |
| 2652 C_symtype (str, len, c_ext) | |
| 2653 char *str; | |
| 2654 int len; | |
| 2655 int c_ext; | |
| 2656 { | |
| 2657 register struct C_stab_entry *se = in_word_set (str, len); | |
| 2658 | |
| 2659 if (se == NULL || (se->c_ext && !(c_ext & se->c_ext))) | |
| 2660 return st_none; | |
| 2661 return se->type; | |
| 2662 } | |
| 458 | 2663 |
| 428 | 2664 |
| 458 | 2665 /* |
| 2325 | 2666 * Ignoring __attribute__ ((list)) |
| 2667 */ | |
| 2668 static bool inattribute; /* looking at an __attribute__ construct */ | |
| 2669 | |
| 2670 /* | |
| 458 | 2671 * C functions and variables are recognized using a simple |
| 2672 * finite automaton. fvdef is its state variable. | |
| 2673 */ | |
| 2225 | 2674 static enum |
| 428 | 2675 { |
| 2676 fvnone, /* nothing seen */ | |
| 458 | 2677 fdefunkey, /* Emacs DEFUN keyword seen */ |
| 2678 fdefunname, /* Emacs DEFUN name seen */ | |
| 428 | 2679 foperator, /* func: operator keyword seen (cplpl) */ |
| 2680 fvnameseen, /* function or variable name seen */ | |
| 2681 fstartlist, /* func: just after open parenthesis */ | |
| 2682 finlist, /* func: in parameter list */ | |
| 2683 flistseen, /* func: after parameter list */ | |
| 2684 fignore, /* func: before open brace */ | |
| 2685 vignore /* var-like: ignore until ';' */ | |
| 2686 } fvdef; | |
| 2687 | |
| 2225 | 2688 static bool fvextern; /* func or var: extern keyword seen; */ |
| 428 | 2689 |
| 458 | 2690 /* |
| 2691 * typedefs are recognized using a simple finite automaton. | |
| 2692 * typdef is its state variable. | |
| 2693 */ | |
| 2225 | 2694 static enum |
| 428 | 2695 { |
| 2696 tnone, /* nothing seen */ | |
| 2697 tkeyseen, /* typedef keyword seen */ | |
| 2698 ttypeseen, /* defined type seen */ | |
| 2699 tinbody, /* inside typedef body */ | |
| 2700 tend, /* just before typedef tag */ | |
| 2701 tignore /* junk after typedef tag */ | |
| 2702 } typdef; | |
| 2703 | |
| 458 | 2704 /* |
| 2705 * struct-like structures (enum, struct and union) are recognized | |
| 2706 * using another simple finite automaton. `structdef' is its state | |
| 2707 * variable. | |
| 2708 */ | |
| 2225 | 2709 static enum |
| 428 | 2710 { |
| 458 | 2711 snone, /* nothing seen yet, |
| 2325 | 2712 or in struct body if bracelev > 0 */ |
| 428 | 2713 skeyseen, /* struct-like keyword seen */ |
| 2714 stagseen, /* struct-like tag seen */ | |
| 458 | 2715 scolonseen /* colon seen after struct-like tag */ |
| 428 | 2716 } structdef; |
| 2717 | |
| 2718 /* | |
| 2719 * When objdef is different from onone, objtag is the name of the class. | |
| 2720 */ | |
| 2225 | 2721 static char *objtag = "<uninited>"; |
| 428 | 2722 |
| 2723 /* | |
| 2724 * Yet another little state machine to deal with preprocessor lines. | |
| 2725 */ | |
| 2225 | 2726 static enum |
| 428 | 2727 { |
| 2728 dnone, /* nothing seen */ | |
| 2729 dsharpseen, /* '#' seen as first char on line */ | |
| 2730 ddefineseen, /* '#' and 'define' seen */ | |
| 2731 dignorerest /* ignore rest of line */ | |
| 2732 } definedef; | |
| 2733 | |
| 2734 /* | |
| 2735 * State machine for Objective C protocols and implementations. | |
| 458 | 2736 * Idea by Tom R.Hageman <tom@basil.icce.rug.nl> (1995) |
| 428 | 2737 */ |
| 2225 | 2738 static enum |
| 428 | 2739 { |
| 2740 onone, /* nothing seen */ | |
| 2741 oprotocol, /* @interface or @protocol seen */ | |
| 2742 oimplementation, /* @implementations seen */ | |
| 2743 otagseen, /* class name seen */ | |
| 2744 oparenseen, /* parenthesis before category seen */ | |
| 2745 ocatseen, /* category name seen */ | |
| 2746 oinbody, /* in @implementation body */ | |
| 2747 omethodsign, /* in @implementation body, after +/- */ | |
| 2748 omethodtag, /* after method name */ | |
| 2749 omethodcolon, /* after method colon */ | |
| 2750 omethodparm, /* after method parameter */ | |
| 2751 oignore /* wait for @end */ | |
| 2752 } objdef; | |
| 2753 | |
| 2754 | |
| 2755 /* | |
| 2756 * Use this structure to keep info about the token read, and how it | |
| 2757 * should be tagged. Used by the make_C_tag function to build a tag. | |
| 2758 */ | |
| 2225 | 2759 static struct tok |
| 428 | 2760 { |
| 2225 | 2761 char *line; /* string containing the token */ |
| 2762 int offset; /* where the token starts in LINE */ | |
| 2763 int length; /* token length */ | |
| 2764 /* | |
| 2765 The previous members can be used to pass strings around for generic | |
| 2766 purposes. The following ones specifically refer to creating tags. In this | |
| 2767 case the token contained here is the pattern that will be used to create a | |
| 2768 tag. | |
| 2769 */ | |
| 2770 bool valid; /* do not create a tag; the token should be | |
| 2771 invalidated whenever a state machine is | |
| 2772 reset prematurely */ | |
| 2773 bool named; /* create a named tag */ | |
| 2774 int lineno; /* source line number of tag */ | |
| 2775 long linepos; /* source char number of tag */ | |
| 458 | 2776 } token; /* latest token read */ |
| 428 | 2777 |
| 2778 /* | |
| 458 | 2779 * Variables and functions for dealing with nested structures. |
| 2780 * Idea by Mykola Dzyuba <mdzyuba@yahoo.com> (2001) | |
| 428 | 2781 */ |
| 709 | 2782 static void pushclass_above __P((int, char *, int)); |
| 2783 static void popclass_above __P((int)); | |
| 2784 static void write_classname __P((linebuffer *, char *qualifier)); | |
| 458 | 2785 |
| 2225 | 2786 static struct { |
| 458 | 2787 char **cname; /* nested class names */ |
| 2325 | 2788 int *bracelev; /* nested class brace level */ |
| 458 | 2789 int nl; /* class nesting level (elements used) */ |
| 2790 int size; /* length of the array */ | |
| 2791 } cstack; /* stack for nested declaration tags */ | |
| 2792 /* Current struct nesting depth (namespace, class, struct, union, enum). */ | |
| 2793 #define nestlev (cstack.nl) | |
| 2225 | 2794 /* After struct keyword or in struct body, not inside a nested function. */ |
| 458 | 2795 #define instruct (structdef == snone && nestlev > 0 \ |
| 2325 | 2796 && bracelev == cstack.bracelev[nestlev-1] + 1) |
| 458 | 2797 |
| 2798 static void | |
| 2325 | 2799 pushclass_above (bracelev, str, len) |
| 2800 int bracelev; | |
| 458 | 2801 char *str; |
| 2802 int len; | |
| 2803 { | |
| 2804 int nl; | |
| 2805 | |
| 2325 | 2806 popclass_above (bracelev); |
| 458 | 2807 nl = cstack.nl; |
| 2808 if (nl >= cstack.size) | |
| 2809 { | |
| 2810 int size = cstack.size *= 2; | |
| 2811 xrnew (cstack.cname, size, char *); | |
| 2325 | 2812 xrnew (cstack.bracelev, size, int); |
| 458 | 2813 } |
| 2325 | 2814 assert (nl == 0 || cstack.bracelev[nl-1] < bracelev); |
| 458 | 2815 cstack.cname[nl] = (str == NULL) ? NULL : savenstr (str, len); |
| 2325 | 2816 cstack.bracelev[nl] = bracelev; |
| 458 | 2817 cstack.nl = nl + 1; |
| 2818 } | |
| 2819 | |
| 2820 static void | |
| 2325 | 2821 popclass_above (bracelev) |
| 2822 int bracelev; | |
| 458 | 2823 { |
| 2824 int nl; | |
| 2825 | |
| 2826 for (nl = cstack.nl - 1; | |
| 2325 | 2827 nl >= 0 && cstack.bracelev[nl] >= bracelev; |
| 458 | 2828 nl--) |
| 2829 { | |
| 2830 if (cstack.cname[nl] != NULL) | |
| 2831 free (cstack.cname[nl]); | |
| 2832 cstack.nl = nl; | |
| 2833 } | |
| 2834 } | |
| 2835 | |
| 2836 static void | |
| 2837 write_classname (cn, qualifier) | |
| 2838 linebuffer *cn; | |
| 2839 char *qualifier; | |
| 2840 { | |
| 2841 int i, len; | |
| 2842 int qlen = strlen (qualifier); | |
| 2843 | |
| 2844 if (cstack.nl == 0 || cstack.cname[0] == NULL) | |
| 2845 { | |
| 2846 len = 0; | |
| 2847 cn->len = 0; | |
| 2848 cn->buffer[0] = '\0'; | |
| 2849 } | |
| 2850 else | |
| 2851 { | |
| 2852 len = strlen (cstack.cname[0]); | |
| 2853 linebuffer_setlen (cn, len); | |
| 2854 strcpy (cn->buffer, cstack.cname[0]); | |
| 2855 } | |
| 2856 for (i = 1; i < cstack.nl; i++) | |
| 2857 { | |
| 2858 char *s; | |
| 2859 int slen; | |
| 2860 | |
| 2861 s = cstack.cname[i]; | |
| 2862 if (s == NULL) | |
| 2863 continue; | |
| 2864 slen = strlen (s); | |
| 2865 len += slen + qlen; | |
| 2866 linebuffer_setlen (cn, len); | |
| 2867 strncat (cn->buffer, qualifier, qlen); | |
| 2868 strncat (cn->buffer, s, slen); | |
| 2869 } | |
| 2870 } | |
| 2871 | |
| 2872 | |
| 709 | 2873 static bool consider_token __P((char *, int, int, int *, int, int, bool *)); |
| 2874 static void make_C_tag __P((bool)); | |
| 442 | 2875 |
| 428 | 2876 /* |
| 2877 * consider_token () | |
| 2878 * checks to see if the current token is at the start of a | |
| 2879 * function or variable, or corresponds to a typedef, or | |
| 2880 * is a struct/union/enum tag, or #define, or an enum constant. | |
| 2881 * | |
| 2882 * *IS_FUNC gets TRUE iff the token is a function or #define macro | |
| 458 | 2883 * with args. C_EXTP points to which language we are looking at. |
| 428 | 2884 * |
| 2885 * Globals | |
| 2886 * fvdef IN OUT | |
| 2887 * structdef IN OUT | |
| 2888 * definedef IN OUT | |
| 2889 * typdef IN OUT | |
| 2890 * objdef IN OUT | |
| 2891 */ | |
| 2892 | |
| 2893 static bool | |
| 2325 | 2894 consider_token (str, len, c, c_extp, bracelev, parlev, is_func_or_var) |
| 428 | 2895 register char *str; /* IN: token pointer */ |
| 2896 register int len; /* IN: token length */ | |
| 442 | 2897 register int c; /* IN: first char after the token */ |
| 458 | 2898 int *c_extp; /* IN, OUT: C extensions mask */ |
| 2325 | 2899 int bracelev; /* IN: brace level */ |
| 428 | 2900 int parlev; /* IN: parenthesis level */ |
| 2901 bool *is_func_or_var; /* OUT: function or variable found */ | |
| 2902 { | |
| 2325 | 2903 /* When structdef is stagseen, scolonseen, or snone with bracelev > 0, |
| 458 | 2904 structtype is the type of the preceding struct-like keyword, and |
| 2325 | 2905 structbracelev is the brace level where it has been seen. */ |
| 458 | 2906 static enum sym_type structtype; |
| 2325 | 2907 static int structbracelev; |
| 458 | 2908 static enum sym_type toktype; |
| 2909 | |
| 2910 | |
| 2911 toktype = C_symtype (str, len, *c_extp); | |
| 428 | 2912 |
| 2913 /* | |
| 2325 | 2914 * Skip __attribute__ |
| 428 | 2915 */ |
| 2325 | 2916 if (toktype == st_C_attribute) |
| 428 | 2917 { |
| 2325 | 2918 inattribute = TRUE; |
| 428 | 2919 return FALSE; |
| 2325 | 2920 } |
| 2921 | |
| 2922 /* | |
| 2923 * Advance the definedef state machine. | |
| 2924 */ | |
| 2925 switch (definedef) | |
| 2926 { | |
| 2927 case dnone: | |
| 2928 /* We're not on a preprocessor line. */ | |
| 2929 if (toktype == st_C_gnumacro) | |
| 2930 { | |
| 2931 fvdef = fdefunkey; | |
| 2932 return FALSE; | |
| 2933 } | |
| 2934 break; | |
| 2935 case dsharpseen: | |
| 2936 if (toktype == st_C_define) | |
| 2937 { | |
| 2938 definedef = ddefineseen; | |
| 2939 } | |
| 2940 else | |
| 2941 { | |
| 2942 definedef = dignorerest; | |
| 2943 } | |
| 2944 return FALSE; | |
| 2945 case ddefineseen: | |
| 2946 /* | |
| 2947 * Make a tag for any macro, unless it is a constant | |
| 2948 * and constantypedefs is FALSE. | |
| 2949 */ | |
| 2950 definedef = dignorerest; | |
| 2951 *is_func_or_var = (c == '('); | |
| 2952 if (!*is_func_or_var && !constantypedefs) | |
| 2953 return FALSE; | |
| 2954 else | |
| 2955 return TRUE; | |
| 2956 case dignorerest: | |
| 2957 return FALSE; | |
| 2958 default: | |
| 2959 error ("internal error: definedef value.", (char *)NULL); | |
| 2960 } | |
| 2961 | |
| 2962 /* | |
| 2963 * Now typedefs | |
| 2964 */ | |
| 2965 switch (typdef) | |
| 2966 { | |
| 2967 case tnone: | |
| 2968 if (toktype == st_C_typedef) | |
| 2969 { | |
| 2970 if (typedefs) | |
| 2971 typdef = tkeyseen; | |
| 2972 fvextern = FALSE; | |
| 2973 fvdef = fvnone; | |
| 2974 return FALSE; | |
| 2975 } | |
| 2976 break; | |
| 2977 case tkeyseen: | |
| 2978 switch (toktype) | |
| 2979 { | |
| 2980 case st_none: | |
| 2981 case st_C_class: | |
| 2982 case st_C_struct: | |
| 2983 case st_C_enum: | |
| 2984 typdef = ttypeseen; | |
| 2985 } | |
| 2986 break; | |
| 2987 case ttypeseen: | |
| 2988 if (structdef == snone && fvdef == fvnone) | |
| 2989 { | |
| 2990 fvdef = fvnameseen; | |
| 2991 return TRUE; | |
| 2992 } | |
| 2993 break; | |
| 2994 case tend: | |
| 2995 switch (toktype) | |
| 2996 { | |
| 2997 case st_C_class: | |
| 2998 case st_C_struct: | |
| 2999 case st_C_enum: | |
| 3000 return FALSE; | |
| 3001 } | |
| 3002 return TRUE; | |
| 3003 } | |
| 3004 | |
| 3005 /* | |
| 3006 * This structdef business is NOT invoked when we are ctags and the | |
| 3007 * file is plain C. This is because a struct tag may have the same | |
| 3008 * name as another tag, and this loses with ctags. | |
| 3009 */ | |
| 3010 switch (toktype) | |
| 3011 { | |
| 3012 case st_C_javastruct: | |
| 3013 if (structdef == stagseen) | |
| 3014 structdef = scolonseen; | |
| 3015 return FALSE; | |
| 3016 case st_C_template: | |
| 3017 case st_C_class: | |
| 3018 if ((*c_extp & C_AUTO) /* automatic detection of C++ language */ | |
| 3019 && bracelev == 0 | |
| 3020 && definedef == dnone && structdef == snone | |
| 3021 && typdef == tnone && fvdef == fvnone) | |
| 3022 *c_extp = (*c_extp | C_PLPL) & ~C_AUTO; | |
| 3023 if (toktype == st_C_template) | |
| 3024 break; | |
| 3025 /* FALLTHRU */ | |
| 3026 case st_C_struct: | |
| 3027 case st_C_enum: | |
| 3028 if (parlev == 0 | |
| 3029 && fvdef != vignore | |
| 3030 && (typdef == tkeyseen | |
| 3031 || (typedefs_or_cplusplus && structdef == snone))) | |
| 3032 { | |
| 3033 structdef = skeyseen; | |
| 3034 structtype = toktype; | |
| 3035 structbracelev = bracelev; | |
| 3036 if (fvdef == fvnameseen) | |
| 3037 fvdef = fvnone; | |
| 3038 } | |
| 3039 return FALSE; | |
| 3040 } | |
| 3041 | |
| 3042 if (structdef == skeyseen) | |
| 3043 { | |
| 3044 structdef = stagseen; | |
| 3045 return TRUE; | |
| 3046 } | |
| 3047 | |
| 3048 if (typdef != tnone) | |
| 3049 definedef = dnone; | |
| 3050 | |
| 3051 /* Detect Objective C constructs. */ | |
| 3052 switch (objdef) | |
| 3053 { | |
| 3054 case onone: | |
| 3055 switch (toktype) | |
| 3056 { | |
| 3057 case st_C_objprot: | |
| 3058 objdef = oprotocol; | |
| 3059 return FALSE; | |
| 3060 case st_C_objimpl: | |
| 3061 objdef = oimplementation; | |
| 3062 return FALSE; | |
| 3063 } | |
| 3064 break; | |
| 3065 case oimplementation: | |
| 3066 /* Save the class tag for functions or variables defined inside. */ | |
| 3067 objtag = savenstr (str, len); | |
| 3068 objdef = oinbody; | |
| 3069 return FALSE; | |
| 3070 case oprotocol: | |
| 3071 /* Save the class tag for categories. */ | |
| 3072 objtag = savenstr (str, len); | |
| 3073 objdef = otagseen; | |
| 3074 *is_func_or_var = TRUE; | |
| 3075 return TRUE; | |
| 3076 case oparenseen: | |
| 3077 objdef = ocatseen; | |
| 3078 *is_func_or_var = TRUE; | |
| 3079 return TRUE; | |
| 3080 case oinbody: | |
| 3081 break; | |
| 3082 case omethodsign: | |
| 3083 if (parlev == 0) | |
| 3084 { | |
| 3085 fvdef = fvnone; | |
| 3086 objdef = omethodtag; | |
| 3087 linebuffer_setlen (&token_name, len); | |
| 3088 strncpy (token_name.buffer, str, len); | |
| 3089 token_name.buffer[len] = '\0'; | |
| 3090 return TRUE; | |
| 3091 } | |
| 3092 return FALSE; | |
| 3093 case omethodcolon: | |
| 3094 if (parlev == 0) | |
| 3095 objdef = omethodparm; | |
| 3096 return FALSE; | |
| 3097 case omethodparm: | |
| 3098 if (parlev == 0) | |
| 3099 { | |
| 3100 fvdef = fvnone; | |
| 3101 objdef = omethodtag; | |
| 3102 linebuffer_setlen (&token_name, token_name.len + len); | |
| 3103 strncat (token_name.buffer, str, len); | |
| 3104 return TRUE; | |
| 3105 } | |
| 3106 return FALSE; | |
| 3107 case oignore: | |
| 3108 if (toktype == st_C_objend) | |
| 3109 { | |
| 3110 /* Memory leakage here: the string pointed by objtag is | |
| 3111 never released, because many tests would be needed to | |
| 3112 avoid breaking on incorrect input code. The amount of | |
| 3113 memory leaked here is the sum of the lengths of the | |
| 3114 class tags. | |
| 3115 free (objtag); */ | |
| 3116 objdef = onone; | |
| 3117 } | |
| 3118 return FALSE; | |
| 3119 } | |
| 3120 | |
| 3121 /* A function, variable or enum constant? */ | |
| 3122 switch (toktype) | |
| 3123 { | |
| 3124 case st_C_extern: | |
| 3125 fvextern = TRUE; | |
| 3126 switch (fvdef) | |
| 3127 { | |
| 3128 case finlist: | |
| 3129 case flistseen: | |
| 3130 case fignore: | |
| 3131 case vignore: | |
| 3132 break; | |
| 3133 default: | |
| 3134 fvdef = fvnone; | |
| 3135 } | |
| 3136 return FALSE; | |
| 3137 case st_C_ignore: | |
| 3138 fvextern = FALSE; | |
| 3139 fvdef = vignore; | |
| 3140 return FALSE; | |
| 3141 case st_C_operator: | |
| 3142 fvdef = foperator; | |
| 3143 *is_func_or_var = TRUE; | |
| 3144 return TRUE; | |
| 3145 case st_none: | |
| 3146 if (constantypedefs | |
| 3147 && structdef == snone | |
| 3148 && structtype == st_C_enum && bracelev > structbracelev) | |
| 3149 return TRUE; /* enum constant */ | |
| 3150 switch (fvdef) | |
| 3151 { | |
| 3152 case fdefunkey: | |
| 3153 if (bracelev > 0) | |
| 3154 break; | |
| 3155 fvdef = fdefunname; /* GNU macro */ | |
| 3156 *is_func_or_var = TRUE; | |
| 3157 return TRUE; | |
| 3158 case fvnone: | |
| 3159 switch (typdef) | |
| 3160 { | |
| 3161 case ttypeseen: | |
| 3162 return FALSE; | |
| 3163 case tnone: | |
| 3164 if ((strneq (str, "asm", 3) && endtoken (str[3])) | |
| 3165 || (strneq (str, "__asm__", 7) && endtoken (str[7]))) | |
| 3166 { | |
| 3167 fvdef = vignore; | |
| 3168 return FALSE; | |
| 3169 } | |
| 3170 break; | |
| 3171 } | |
| 3172 /* FALLTHRU */ | |
| 3173 case fvnameseen: | |
| 2225 | 3174 if (len >= 10 && strneq (str+len-10, "::operator", 10)) |
| 458 | 3175 { |
| 2225 | 3176 if (*c_extp & C_AUTO) /* automatic detection of C++ */ |
| 3177 *c_extp = (*c_extp | C_PLPL) & ~C_AUTO; | |
| 458 | 3178 fvdef = foperator; |
| 3179 *is_func_or_var = TRUE; | |
| 3180 return TRUE; | |
| 3181 } | |
| 2325 | 3182 if (bracelev > 0 && !instruct) |
| 458 | 3183 break; |
| 428 | 3184 fvdef = fvnameseen; /* function or variable */ |
| 3185 *is_func_or_var = TRUE; | |
| 3186 return TRUE; | |
| 3187 } | |
| 3188 break; | |
| 3189 } | |
| 3190 | |
| 3191 return FALSE; | |
| 3192 } | |
| 3193 | |
| 458 | 3194 |
| 428 | 3195 /* |
| 458 | 3196 * C_entries often keeps pointers to tokens or lines which are older than |
| 3197 * the line currently read. By keeping two line buffers, and switching | |
| 3198 * them at end of line, it is possible to use those pointers. | |
| 428 | 3199 */ |
| 2225 | 3200 static struct |
| 458 | 3201 { |
| 3202 long linepos; | |
| 3203 linebuffer lb; | |
| 3204 } lbs[2]; | |
| 3205 | |
| 428 | 3206 #define current_lb_is_new (newndx == curndx) |
| 3207 #define switch_line_buffers() (curndx = 1 - curndx) | |
| 3208 | |
| 3209 #define curlb (lbs[curndx].lb) | |
| 3210 #define newlb (lbs[newndx].lb) | |
| 3211 #define curlinepos (lbs[curndx].linepos) | |
| 3212 #define newlinepos (lbs[newndx].linepos) | |
| 3213 | |
| 2225 | 3214 #define plainc ((c_ext & C_EXT) == C_PLAIN) |
| 3215 #define cplpl (c_ext & C_PLPL) | |
| 3216 #define cjava ((c_ext & C_JAVA) == C_JAVA) | |
| 3217 | |
| 428 | 3218 #define CNL_SAVE_DEFINEDEF() \ |
| 3219 do { \ | |
| 3220 curlinepos = charno; \ | |
| 2225 | 3221 readline (&curlb, inf); \ |
| 428 | 3222 lp = curlb.buffer; \ |
| 3223 quotednl = FALSE; \ | |
| 3224 newndx = curndx; \ | |
| 3225 } while (0) | |
| 3226 | |
| 3227 #define CNL() \ | |
| 3228 do { \ | |
| 3229 CNL_SAVE_DEFINEDEF(); \ | |
| 458 | 3230 if (savetoken.valid) \ |
| 428 | 3231 { \ |
| 458 | 3232 token = savetoken; \ |
| 3233 savetoken.valid = FALSE; \ | |
| 428 | 3234 } \ |
| 3235 definedef = dnone; \ | |
| 3236 } while (0) | |
| 3237 | |
| 3238 | |
| 3239 static void | |
| 3240 make_C_tag (isfun) | |
| 3241 bool isfun; | |
| 3242 { | |
| 3876 | 3243 /* This function is never called when token.valid is FALSE, but |
| 428 | 3244 we must protect against invalid input or internal errors. */ |
| 2225 | 3245 if (!DEBUG && !token.valid) |
| 3246 return; | |
| 3247 | |
| 3248 if (token.valid) | |
| 3249 make_tag (token_name.buffer, token_name.len, isfun, token.line, | |
| 3250 token.offset+token.length+1, token.lineno, token.linepos); | |
| 3251 else /* this case is optimised away if !DEBUG */ | |
| 3252 make_tag (concat ("INVALID TOKEN:-->", token_name.buffer, ""), | |
| 3253 token_name.len + 17, isfun, token.line, | |
| 3254 token.offset+token.length+1, token.lineno, token.linepos); | |
| 3255 | |
| 3256 token.valid = FALSE; | |
| 428 | 3257 } |
| 3258 | |
| 3259 | |
| 458 | 3260 /* |
| 3261 * C_entries () | |
| 3262 * This routine finds functions, variables, typedefs, | |
| 3263 * #define's, enum constants and struct/union/enum definitions in | |
| 3264 * C syntax and adds them to the list. | |
| 3265 */ | |
| 442 | 3266 static void |
| 428 | 3267 C_entries (c_ext, inf) |
| 3268 int c_ext; /* extension of C */ | |
| 3269 FILE *inf; /* input file */ | |
| 3270 { | |
| 3271 register char c; /* latest char read; '\0' for end of line */ | |
| 3272 register char *lp; /* pointer one beyond the character `c' */ | |
| 3273 int curndx, newndx; /* indices for current and new lb */ | |
| 3274 register int tokoff; /* offset in line of start of current token */ | |
| 3275 register int toklen; /* length of current token */ | |
| 3276 char *qualifier; /* string used to qualify names */ | |
| 3277 int qlen; /* length of qualifier */ | |
| 2325 | 3278 int bracelev; /* current brace level */ |
| 3279 int bracketlev; /* current bracket level */ | |
| 428 | 3280 int parlev; /* current parenthesis level */ |
| 2325 | 3281 int attrparlev; /* __attribute__ parenthesis level */ |
| 3282 int templatelev; /* current template level */ | |
| 3283 int typdefbracelev; /* bracelev where a typedef struct body begun */ | |
| 428 | 3284 bool incomm, inquote, inchar, quotednl, midtoken; |
| 458 | 3285 bool yacc_rules; /* in the rules part of a yacc file */ |
| 3876 | 3286 struct tok savetoken = {0}; /* token saved during preprocessor handling */ |
| 458 | 3287 |
| 3288 | |
| 2225 | 3289 linebuffer_init (&lbs[0].lb); |
| 3290 linebuffer_init (&lbs[1].lb); | |
| 458 | 3291 if (cstack.size == 0) |
| 3292 { | |
| 3293 cstack.size = (DEBUG) ? 1 : 4; | |
| 3294 cstack.nl = 0; | |
| 3295 cstack.cname = xnew (cstack.size, char *); | |
| 2325 | 3296 cstack.bracelev = xnew (cstack.size, int); |
| 458 | 3297 } |
| 428 | 3298 |
| 2325 | 3299 tokoff = toklen = typdefbracelev = 0; /* keep compiler quiet */ |
| 428 | 3300 curndx = newndx = 0; |
| 3301 lp = curlb.buffer; | |
| 3302 *lp = 0; | |
| 3303 | |
| 3304 fvdef = fvnone; fvextern = FALSE; typdef = tnone; | |
| 3305 structdef = snone; definedef = dnone; objdef = onone; | |
| 458 | 3306 yacc_rules = FALSE; |
| 428 | 3307 midtoken = inquote = inchar = incomm = quotednl = FALSE; |
| 458 | 3308 token.valid = savetoken.valid = FALSE; |
| 2325 | 3309 bracelev = bracketlev = parlev = attrparlev = templatelev = 0; |
| 428 | 3310 if (cjava) |
| 3311 { qualifier = "."; qlen = 1; } | |
| 3312 else | |
| 3313 { qualifier = "::"; qlen = 2; } | |
| 3314 | |
| 458 | 3315 |
| 428 | 3316 while (!feof (inf)) |
| 3317 { | |
| 3318 c = *lp++; | |
| 3319 if (c == '\\') | |
| 3320 { | |
| 2325 | 3321 /* If we are at the end of the line, the next character is a |
| 3322 '\0'; do not skip it, because it is what tells us | |
| 428 | 3323 to read the next line. */ |
| 3324 if (*lp == '\0') | |
| 3325 { | |
| 3326 quotednl = TRUE; | |
| 3327 continue; | |
| 3328 } | |
| 3329 lp++; | |
| 3330 c = ' '; | |
| 3331 } | |
| 3332 else if (incomm) | |
| 3333 { | |
| 3334 switch (c) | |
| 3335 { | |
| 3336 case '*': | |
| 3337 if (*lp == '/') | |
| 3338 { | |
| 3339 c = *lp++; | |
| 3340 incomm = FALSE; | |
| 3341 } | |
| 3342 break; | |
| 3343 case '\0': | |
| 3344 /* Newlines inside comments do not end macro definitions in | |
| 3345 traditional cpp. */ | |
| 3346 CNL_SAVE_DEFINEDEF (); | |
| 3347 break; | |
| 3348 } | |
| 3349 continue; | |
| 3350 } | |
| 3351 else if (inquote) | |
| 3352 { | |
| 3353 switch (c) | |
| 3354 { | |
| 3355 case '"': | |
| 3356 inquote = FALSE; | |
| 3357 break; | |
| 3358 case '\0': | |
| 3359 /* Newlines inside strings do not end macro definitions | |
| 3360 in traditional cpp, even though compilers don't | |
| 3361 usually accept them. */ | |
| 3362 CNL_SAVE_DEFINEDEF (); | |
| 3363 break; | |
| 3364 } | |
| 3365 continue; | |
| 3366 } | |
| 3367 else if (inchar) | |
| 3368 { | |
| 3369 switch (c) | |
| 3370 { | |
| 3371 case '\0': | |
| 3372 /* Hmmm, something went wrong. */ | |
| 3373 CNL (); | |
| 3374 /* FALLTHRU */ | |
| 3375 case '\'': | |
| 3376 inchar = FALSE; | |
| 3377 break; | |
| 3378 } | |
| 3379 continue; | |
| 3380 } | |
| 2325 | 3381 else if (bracketlev > 0) |
| 3382 { | |
| 3383 switch (c) | |
| 3384 { | |
| 3385 case ']': | |
| 3386 if (--bracketlev > 0) | |
| 428 | 3387 continue; |
| 2325 | 3388 break; |
| 3389 case '\0': | |
| 3390 CNL_SAVE_DEFINEDEF (); | |
| 3391 break; | |
| 3392 } | |
| 3393 continue; | |
| 3394 } | |
| 3395 else switch (c) | |
| 3396 { | |
| 3397 case '"': | |
| 3398 inquote = TRUE; | |
| 3399 if (inattribute) | |
| 3400 break; | |
| 3401 switch (fvdef) | |
| 3402 { | |
| 3403 case fdefunkey: | |
| 3404 case fstartlist: | |
| 3405 case finlist: | |
| 3406 case fignore: | |
| 3407 case vignore: | |
| 428 | 3408 break; |
| 2325 | 3409 default: |
| 3410 fvextern = FALSE; | |
| 3411 fvdef = fvnone; | |
| 3412 } | |
| 3413 continue; | |
| 3414 case '\'': | |
| 3415 inchar = TRUE; | |
| 3416 if (inattribute) | |
| 3417 break; | |
| 3418 if (fvdef != finlist && fvdef != fignore && fvdef !=vignore) | |
| 3419 { | |
| 3420 fvextern = FALSE; | |
| 3421 fvdef = fvnone; | |
| 3422 } | |
| 3423 continue; | |
| 3424 case '/': | |
| 3425 if (*lp == '*') | |
| 3426 { | |
| 3972 | 3427 incomm = TRUE; |
| 2325 | 3428 lp++; |
| 3972 | 3429 c = ' '; |
| 2325 | 3430 } |
| 3431 else if (/* cplpl && */ *lp == '/') | |
| 3432 { | |
| 3433 c = '\0'; | |
| 3434 } | |
| 3972 | 3435 break; |
| 2325 | 3436 case '%': |
| 3437 if ((c_ext & YACC) && *lp == '%') | |
| 3438 { | |
| 3439 /* Entering or exiting rules section in yacc file. */ | |
| 3440 lp++; | |
| 3441 definedef = dnone; fvdef = fvnone; fvextern = FALSE; | |
| 3442 typdef = tnone; structdef = snone; | |
| 3443 midtoken = inquote = inchar = incomm = quotednl = FALSE; | |
| 3444 bracelev = 0; | |
| 3445 yacc_rules = !yacc_rules; | |
| 3446 continue; | |
| 3447 } | |
| 3448 else | |
| 3449 break; | |
| 3450 case '#': | |
| 3451 if (definedef == dnone) | |
| 3452 { | |
| 3453 char *cp; | |
| 3454 bool cpptoken = TRUE; | |
| 3455 | |
| 3456 /* Look back on this line. If all blanks, or nonblanks | |
| 3457 followed by an end of comment, this is a preprocessor | |
| 3458 token. */ | |
| 3459 for (cp = newlb.buffer; cp < lp-1; cp++) | |
| 3460 if (!iswhite (*cp)) | |
| 3461 { | |
| 3462 if (*cp == '*' && *(cp+1) == '/') | |
| 3463 { | |
| 3464 cp++; | |
| 3465 cpptoken = TRUE; | |
| 3466 } | |
| 3467 else | |
| 3468 cpptoken = FALSE; | |
| 3469 } | |
| 3470 if (cpptoken) | |
| 3471 definedef = dsharpseen; | |
| 3472 } /* if (definedef == dnone) */ | |
| 3473 continue; | |
| 3474 case '[': | |
| 3475 bracketlev++; | |
| 428 | 3476 continue; |
| 2325 | 3477 } /* switch (c) */ |
| 428 | 3478 |
| 3479 | |
| 458 | 3480 /* Consider token only if some involved conditions are satisfied. */ |
| 3481 if (typdef != tignore | |
| 428 | 3482 && definedef != dignorerest |
| 458 | 3483 && fvdef != finlist |
| 2325 | 3484 && templatelev == 0 |
| 458 | 3485 && (definedef != dnone |
| 2325 | 3486 || structdef != scolonseen) |
| 3487 && !inattribute) | |
| 428 | 3488 { |
| 3489 if (midtoken) | |
| 3490 { | |
| 3491 if (endtoken (c)) | |
| 3492 { | |
| 2225 | 3493 if (c == ':' && *lp == ':' && begtoken (lp[1])) |
| 3494 /* This handles :: in the middle, | |
| 3495 but not at the beginning of an identifier. | |
| 3496 Also, space-separated :: is not recognised. */ | |
| 428 | 3497 { |
| 2225 | 3498 if (c_ext & C_AUTO) /* automatic detection of C++ */ |
| 3499 c_ext = (c_ext | C_PLPL) & ~C_AUTO; | |
| 428 | 3500 lp += 2; |
| 3501 toklen += 2; | |
| 3502 c = lp[-1]; | |
| 458 | 3503 goto still_in_token; |
| 428 | 3504 } |
| 3505 else | |
| 3506 { | |
| 458 | 3507 bool funorvar = FALSE; |
| 3508 | |
| 428 | 3509 if (yacc_rules |
| 3510 || consider_token (newlb.buffer + tokoff, toklen, c, | |
| 2325 | 3511 &c_ext, bracelev, parlev, |
| 3512 &funorvar)) | |
| 428 | 3513 { |
| 3514 if (fvdef == foperator) | |
| 3515 { | |
| 3516 char *oldlp = lp; | |
| 3517 lp = skip_spaces (lp-1); | |
| 3518 if (*lp != '\0') | |
| 3519 lp += 1; | |
| 3520 while (*lp != '\0' | |
| 442 | 3521 && !iswhite (*lp) && *lp != '(') |
| 428 | 3522 lp += 1; |
| 3523 c = *lp++; | |
| 3524 toklen += lp - oldlp; | |
| 3525 } | |
| 458 | 3526 token.named = FALSE; |
| 2225 | 3527 if (!plainc |
| 458 | 3528 && nestlev > 0 && definedef == dnone) |
| 3529 /* in struct body */ | |
| 428 | 3530 { |
| 458 | 3531 write_classname (&token_name, qualifier); |
| 3532 linebuffer_setlen (&token_name, | |
| 3533 token_name.len+qlen+toklen); | |
| 428 | 3534 strcat (token_name.buffer, qualifier); |
| 3535 strncat (token_name.buffer, | |
| 3536 newlb.buffer + tokoff, toklen); | |
| 458 | 3537 token.named = TRUE; |
| 428 | 3538 } |
| 3539 else if (objdef == ocatseen) | |
| 3540 /* Objective C category */ | |
| 3541 { | |
| 3542 int len = strlen (objtag) + 2 + toklen; | |
| 458 | 3543 linebuffer_setlen (&token_name, len); |
| 428 | 3544 strcpy (token_name.buffer, objtag); |
| 3545 strcat (token_name.buffer, "("); | |
| 3546 strncat (token_name.buffer, | |
| 3547 newlb.buffer + tokoff, toklen); | |
| 3548 strcat (token_name.buffer, ")"); | |
| 458 | 3549 token.named = TRUE; |
| 428 | 3550 } |
| 3551 else if (objdef == omethodtag | |
| 3552 || objdef == omethodparm) | |
| 3553 /* Objective C method */ | |
| 3554 { | |
| 458 | 3555 token.named = TRUE; |
| 3556 } | |
| 3557 else if (fvdef == fdefunname) | |
| 3558 /* GNU DEFUN and similar macros */ | |
| 3559 { | |
| 3560 bool defun = (newlb.buffer[tokoff] == 'F'); | |
| 3561 int off = tokoff; | |
| 3562 int len = toklen; | |
| 3563 | |
| 3564 /* Rewrite the tag so that emacs lisp DEFUNs | |
| 3565 can be found by their elisp name */ | |
| 3566 if (defun) | |
| 3567 { | |
| 3568 off += 1; | |
| 3569 len -= 1; | |
| 3570 } | |
| 3571 linebuffer_setlen (&token_name, len); | |
| 3572 strncpy (token_name.buffer, | |
| 3573 newlb.buffer + off, len); | |
| 3574 token_name.buffer[len] = '\0'; | |
| 3575 if (defun) | |
| 3576 while (--len >= 0) | |
| 3577 if (token_name.buffer[len] == '_') | |
| 3578 token_name.buffer[len] = '-'; | |
| 3579 token.named = defun; | |
| 428 | 3580 } |
| 3581 else | |
| 3582 { | |
| 458 | 3583 linebuffer_setlen (&token_name, toklen); |
| 428 | 3584 strncpy (token_name.buffer, |
| 3585 newlb.buffer + tokoff, toklen); | |
| 3586 token_name.buffer[toklen] = '\0'; | |
| 3587 /* Name macros and members. */ | |
| 458 | 3588 token.named = (structdef == stagseen |
| 3589 || typdef == ttypeseen | |
| 3590 || typdef == tend | |
| 3591 || (funorvar | |
| 3592 && definedef == dignorerest) | |
| 3593 || (funorvar | |
| 3594 && definedef == dnone | |
| 3595 && structdef == snone | |
| 2325 | 3596 && bracelev > 0)); |
| 428 | 3597 } |
| 458 | 3598 token.lineno = lineno; |
| 3599 token.offset = tokoff; | |
| 3600 token.length = toklen; | |
| 3601 token.line = newlb.buffer; | |
| 3602 token.linepos = newlinepos; | |
| 3603 token.valid = TRUE; | |
| 428 | 3604 |
| 3605 if (definedef == dnone | |
| 3606 && (fvdef == fvnameseen | |
| 3607 || fvdef == foperator | |
| 3608 || structdef == stagseen | |
| 3609 || typdef == tend | |
| 458 | 3610 || typdef == ttypeseen |
| 428 | 3611 || objdef != onone)) |
| 3612 { | |
| 3613 if (current_lb_is_new) | |
| 3614 switch_line_buffers (); | |
| 3615 } | |
| 458 | 3616 else if (definedef != dnone |
| 3617 || fvdef == fdefunname | |
| 3618 || instruct) | |
| 428 | 3619 make_C_tag (funorvar); |
| 3620 } | |
| 2325 | 3621 else /* not yacc and consider_token failed */ |
| 3622 { | |
| 3623 if (inattribute && fvdef == fignore) | |
| 3624 { | |
| 3625 /* We have just met __attribute__ after a | |
| 3626 function parameter list: do not tag the | |
| 3627 function again. */ | |
| 3628 fvdef = fvnone; | |
| 3629 } | |
| 3630 } | |
| 428 | 3631 midtoken = FALSE; |
| 3632 } | |
| 3633 } /* if (endtoken (c)) */ | |
| 3634 else if (intoken (c)) | |
| 458 | 3635 still_in_token: |
| 428 | 3636 { |
| 3637 toklen++; | |
| 3638 continue; | |
| 3639 } | |
| 3640 } /* if (midtoken) */ | |
| 3641 else if (begtoken (c)) | |
| 3642 { | |
| 3643 switch (definedef) | |
| 3644 { | |
| 3645 case dnone: | |
| 3646 switch (fvdef) | |
| 3647 { | |
| 3648 case fstartlist: | |
| 2325 | 3649 /* This prevents tagging fb in |
| 3650 void (__attribute__((noreturn)) *fb) (void); | |
| 3651 Fixing this is not easy and not very important. */ | |
| 428 | 3652 fvdef = finlist; |
| 3653 continue; | |
| 3654 case flistseen: | |
| 2225 | 3655 if (plainc || declarations) |
| 3656 { | |
| 3657 make_C_tag (TRUE); /* a function */ | |
| 3658 fvdef = fignore; | |
| 3659 } | |
| 428 | 3660 break; |
| 3661 } | |
| 3662 if (structdef == stagseen && !cjava) | |
| 458 | 3663 { |
| 2325 | 3664 popclass_above (bracelev); |
| 458 | 3665 structdef = snone; |
| 3666 } | |
| 428 | 3667 break; |
| 3668 case dsharpseen: | |
| 458 | 3669 savetoken = token; |
| 2225 | 3670 break; |
| 428 | 3671 } |
| 3672 if (!yacc_rules || lp == newlb.buffer + 1) | |
| 3673 { | |
| 3674 tokoff = lp - 1 - newlb.buffer; | |
| 3675 toklen = 1; | |
| 3676 midtoken = TRUE; | |
| 3677 } | |
| 3678 continue; | |
| 3679 } /* if (begtoken) */ | |
| 3680 } /* if must look at token */ | |
| 3681 | |
| 3682 | |
| 3683 /* Detect end of line, colon, comma, semicolon and various braces | |
| 3684 after having handled a token.*/ | |
| 3685 switch (c) | |
| 3686 { | |
| 3687 case ':': | |
| 2325 | 3688 if (inattribute) |
| 3689 break; | |
| 458 | 3690 if (yacc_rules && token.offset == 0 && token.valid) |
| 3691 { | |
| 3692 make_C_tag (FALSE); /* a yacc function */ | |
| 3693 break; | |
| 3694 } | |
| 428 | 3695 if (definedef != dnone) |
| 3696 break; | |
| 3697 switch (objdef) | |
| 3698 { | |
| 3699 case otagseen: | |
| 3700 objdef = oignore; | |
| 3701 make_C_tag (TRUE); /* an Objective C class */ | |
| 3702 break; | |
| 3703 case omethodtag: | |
| 3704 case omethodparm: | |
| 3705 objdef = omethodcolon; | |
| 458 | 3706 linebuffer_setlen (&token_name, token_name.len + 1); |
| 428 | 3707 strcat (token_name.buffer, ":"); |
| 3708 break; | |
| 3709 } | |
| 3710 if (structdef == stagseen) | |
| 2225 | 3711 { |
| 3712 structdef = scolonseen; | |
| 3713 break; | |
| 3714 } | |
| 3715 /* Should be useless, but may be work as a safety net. */ | |
| 3716 if (cplpl && fvdef == flistseen) | |
| 3717 { | |
| 3718 make_C_tag (TRUE); /* a function */ | |
| 3719 fvdef = fignore; | |
| 3720 break; | |
| 3721 } | |
| 428 | 3722 break; |
| 3723 case ';': | |
| 2325 | 3724 if (definedef != dnone || inattribute) |
| 428 | 3725 break; |
| 458 | 3726 switch (typdef) |
| 428 | 3727 { |
| 458 | 3728 case tend: |
| 3729 case ttypeseen: | |
| 3730 make_C_tag (FALSE); /* a typedef */ | |
| 3731 typdef = tnone; | |
| 3732 fvdef = fvnone; | |
| 428 | 3733 break; |
| 458 | 3734 case tnone: |
| 3735 case tinbody: | |
| 3736 case tignore: | |
| 3737 switch (fvdef) | |
| 3738 { | |
| 3739 case fignore: | |
| 2225 | 3740 if (typdef == tignore || cplpl) |
| 458 | 3741 fvdef = fvnone; |
| 3742 break; | |
| 3743 case fvnameseen: | |
| 2325 | 3744 if ((globals && bracelev == 0 && (!fvextern || declarations)) |
| 458 | 3745 || (members && instruct)) |
| 3746 make_C_tag (FALSE); /* a variable */ | |
| 3747 fvextern = FALSE; | |
| 3748 fvdef = fvnone; | |
| 3749 token.valid = FALSE; | |
| 3750 break; | |
| 3751 case flistseen: | |
| 2325 | 3752 if ((declarations |
| 3753 && (cplpl || !instruct) | |
| 3754 && (typdef == tnone || (typdef != tignore && instruct))) | |
| 3755 || (members | |
| 3756 && plainc && instruct)) | |
| 3757 make_C_tag (TRUE); /* a function */ | |
| 458 | 3758 /* FALLTHRU */ |
| 3759 default: | |
| 3760 fvextern = FALSE; | |
| 3761 fvdef = fvnone; | |
| 3762 if (declarations | |
| 2225 | 3763 && cplpl && structdef == stagseen) |
| 458 | 3764 make_C_tag (FALSE); /* forward declaration */ |
| 3765 else | |
| 3766 token.valid = FALSE; | |
| 3767 } /* switch (fvdef) */ | |
| 428 | 3768 /* FALLTHRU */ |
| 3769 default: | |
| 458 | 3770 if (!instruct) |
| 3771 typdef = tnone; | |
| 428 | 3772 } |
| 3773 if (structdef == stagseen) | |
| 3774 structdef = snone; | |
| 3775 break; | |
| 3776 case ',': | |
| 2325 | 3777 if (definedef != dnone || inattribute) |
| 428 | 3778 break; |
| 3779 switch (objdef) | |
| 3780 { | |
| 3781 case omethodtag: | |
| 3782 case omethodparm: | |
| 3783 make_C_tag (TRUE); /* an Objective C method */ | |
| 3784 objdef = oinbody; | |
| 3785 break; | |
| 3786 } | |
| 3787 switch (fvdef) | |
| 3788 { | |
| 458 | 3789 case fdefunkey: |
| 428 | 3790 case foperator: |
| 458 | 3791 case fstartlist: |
| 428 | 3792 case finlist: |
| 3793 case fignore: | |
| 3794 case vignore: | |
| 3795 break; | |
| 458 | 3796 case fdefunname: |
| 3797 fvdef = fignore; | |
| 3798 break; | |
| 2325 | 3799 case fvnameseen: |
| 3800 if (parlev == 0 | |
| 3801 && ((globals | |
| 3802 && bracelev == 0 | |
| 3803 && templatelev == 0 | |
| 3804 && (!fvextern || declarations)) | |
| 3805 || (members && instruct))) | |
| 3806 make_C_tag (FALSE); /* a variable */ | |
| 458 | 3807 break; |
| 2325 | 3808 case flistseen: |
| 458 | 3809 if ((declarations && typdef == tnone && !instruct) |
| 3810 || (members && typdef != tignore && instruct)) | |
| 3811 { | |
| 2325 | 3812 make_C_tag (TRUE); /* a function */ |
| 458 | 3813 fvdef = fvnameseen; |
| 3814 } | |
| 3815 else if (!declarations) | |
| 3816 fvdef = fvnone; | |
| 3817 token.valid = FALSE; | |
| 428 | 3818 break; |
| 3819 default: | |
| 3820 fvdef = fvnone; | |
| 3821 } | |
| 3822 if (structdef == stagseen) | |
| 3823 structdef = snone; | |
| 3824 break; | |
| 2325 | 3825 case ']': |
| 3826 if (definedef != dnone || inattribute) | |
| 428 | 3827 break; |
| 458 | 3828 if (structdef == stagseen) |
| 3829 structdef = snone; | |
| 3830 switch (typdef) | |
| 428 | 3831 { |
| 458 | 3832 case ttypeseen: |
| 3833 case tend: | |
| 428 | 3834 typdef = tignore; |
| 3835 make_C_tag (FALSE); /* a typedef */ | |
| 3836 break; | |
| 458 | 3837 case tnone: |
| 3838 case tinbody: | |
| 3839 switch (fvdef) | |
| 3840 { | |
| 3841 case foperator: | |
| 3842 case finlist: | |
| 3843 case fignore: | |
| 3844 case vignore: | |
| 3845 break; | |
| 3846 case fvnameseen: | |
| 2325 | 3847 if ((members && bracelev == 1) |
| 3848 || (globals && bracelev == 0 | |
| 458 | 3849 && (!fvextern || declarations))) |
| 3850 make_C_tag (FALSE); /* a variable */ | |
| 3851 /* FALLTHRU */ | |
| 3852 default: | |
| 3853 fvdef = fvnone; | |
| 3854 } | |
| 428 | 3855 break; |
| 3856 } | |
| 3857 break; | |
| 3858 case '(': | |
| 2325 | 3859 if (inattribute) |
| 3860 { | |
| 3861 attrparlev++; | |
| 3862 break; | |
| 3863 } | |
| 428 | 3864 if (definedef != dnone) |
| 3865 break; | |
| 3866 if (objdef == otagseen && parlev == 0) | |
| 3867 objdef = oparenseen; | |
| 3868 switch (fvdef) | |
| 3869 { | |
| 3870 case fvnameseen: | |
| 3871 if (typdef == ttypeseen | |
| 3872 && *lp != '*' | |
| 458 | 3873 && !instruct) |
| 428 | 3874 { |
| 3875 /* This handles constructs like: | |
| 3876 typedef void OperatorFun (int fun); */ | |
| 3877 make_C_tag (FALSE); | |
| 3878 typdef = tignore; | |
| 458 | 3879 fvdef = fignore; |
| 3880 break; | |
| 428 | 3881 } |
| 3882 /* FALLTHRU */ | |
| 3883 case foperator: | |
| 3884 fvdef = fstartlist; | |
| 3885 break; | |
| 3886 case flistseen: | |
| 3887 fvdef = finlist; | |
| 3888 break; | |
| 3889 } | |
| 3890 parlev++; | |
| 3891 break; | |
| 3892 case ')': | |
| 2325 | 3893 if (inattribute) |
| 3894 { | |
| 3895 if (--attrparlev == 0) | |
| 3896 inattribute = FALSE; | |
| 3897 break; | |
| 3898 } | |
| 428 | 3899 if (definedef != dnone) |
| 3900 break; | |
| 3901 if (objdef == ocatseen && parlev == 1) | |
| 3902 { | |
| 3903 make_C_tag (TRUE); /* an Objective C category */ | |
| 3904 objdef = oignore; | |
| 3905 } | |
| 3906 if (--parlev == 0) | |
| 3907 { | |
| 3908 switch (fvdef) | |
| 3909 { | |
| 3910 case fstartlist: | |
| 3911 case finlist: | |
| 3912 fvdef = flistseen; | |
| 3913 break; | |
| 3914 } | |
| 458 | 3915 if (!instruct |
| 3916 && (typdef == tend | |
| 3917 || typdef == ttypeseen)) | |
| 428 | 3918 { |
| 3919 typdef = tignore; | |
| 3920 make_C_tag (FALSE); /* a typedef */ | |
| 3921 } | |
| 3922 } | |
| 3923 else if (parlev < 0) /* can happen due to ill-conceived #if's. */ | |
| 3924 parlev = 0; | |
| 3925 break; | |
| 3926 case '{': | |
| 3927 if (definedef != dnone) | |
| 3928 break; | |
| 3929 if (typdef == ttypeseen) | |
| 3930 { | |
| 531 | 3931 /* Whenever typdef is set to tinbody (currently only |
| 2325 | 3932 here), typdefbracelev should be set to bracelev. */ |
| 531 | 3933 typdef = tinbody; |
| 2325 | 3934 typdefbracelev = bracelev; |
| 428 | 3935 } |
| 3936 switch (fvdef) | |
| 3937 { | |
| 3938 case flistseen: | |
| 458 | 3939 make_C_tag (TRUE); /* a function */ |
| 428 | 3940 /* FALLTHRU */ |
| 3941 case fignore: | |
| 3942 fvdef = fvnone; | |
| 3943 break; | |
| 3944 case fvnone: | |
| 3945 switch (objdef) | |
| 3946 { | |
| 3947 case otagseen: | |
| 3948 make_C_tag (TRUE); /* an Objective C class */ | |
| 3949 objdef = oignore; | |
| 3950 break; | |
| 3951 case omethodtag: | |
| 3952 case omethodparm: | |
| 3953 make_C_tag (TRUE); /* an Objective C method */ | |
| 3954 objdef = oinbody; | |
| 3955 break; | |
| 3956 default: | |
| 3957 /* Neutralize `extern "C" {' grot. */ | |
| 2325 | 3958 if (bracelev == 0 && structdef == snone && nestlev == 0 |
| 458 | 3959 && typdef == tnone) |
| 2325 | 3960 bracelev = -1; |
| 428 | 3961 } |
| 2225 | 3962 break; |
| 428 | 3963 } |
| 458 | 3964 switch (structdef) |
| 3965 { | |
| 3966 case skeyseen: /* unnamed struct */ | |
| 2325 | 3967 pushclass_above (bracelev, NULL, 0); |
| 458 | 3968 structdef = snone; |
| 3969 break; | |
| 3970 case stagseen: /* named struct or enum */ | |
| 3971 case scolonseen: /* a class */ | |
| 2325 | 3972 pushclass_above (bracelev,token.line+token.offset, token.length); |
| 458 | 3973 structdef = snone; |
| 3974 make_C_tag (FALSE); /* a struct or enum */ | |
| 3975 break; | |
| 3976 } | |
| 2325 | 3977 bracelev++; |
| 428 | 3978 break; |
| 3979 case '*': | |
| 3980 if (definedef != dnone) | |
| 3981 break; | |
| 3982 if (fvdef == fstartlist) | |
| 2225 | 3983 { |
| 3984 fvdef = fvnone; /* avoid tagging `foo' in `foo (*bar()) ()' */ | |
| 3985 token.valid = FALSE; | |
| 3986 } | |
| 428 | 3987 break; |
| 3988 case '}': | |
| 3989 if (definedef != dnone) | |
| 3990 break; | |
| 2225 | 3991 if (!ignoreindent && lp == newlb.buffer + 1) |
| 428 | 3992 { |
| 2325 | 3993 if (bracelev != 0) |
| 2225 | 3994 token.valid = FALSE; |
| 2325 | 3995 bracelev = 0; /* reset brace level if first column */ |
| 428 | 3996 parlev = 0; /* also reset paren level, just in case... */ |
| 3997 } | |
| 2325 | 3998 else if (bracelev > 0) |
| 3999 bracelev--; | |
| 2225 | 4000 else |
| 4001 token.valid = FALSE; /* something gone amiss, token unreliable */ | |
| 2325 | 4002 popclass_above (bracelev); |
| 458 | 4003 structdef = snone; |
| 2325 | 4004 /* Only if typdef == tinbody is typdefbracelev significant. */ |
| 4005 if (typdef == tinbody && bracelev <= typdefbracelev) | |
| 428 | 4006 { |
| 2325 | 4007 assert (bracelev == typdefbracelev); |
| 458 | 4008 typdef = tend; |
| 428 | 4009 } |
| 4010 break; | |
| 4011 case '=': | |
| 4012 if (definedef != dnone) | |
| 4013 break; | |
| 4014 switch (fvdef) | |
| 4015 { | |
| 4016 case foperator: | |
| 4017 case finlist: | |
| 4018 case fignore: | |
| 4019 case vignore: | |
| 4020 break; | |
| 4021 case fvnameseen: | |
| 2325 | 4022 if ((members && bracelev == 1) |
| 4023 || (globals && bracelev == 0 && (!fvextern || declarations))) | |
| 428 | 4024 make_C_tag (FALSE); /* a variable */ |
| 4025 /* FALLTHRU */ | |
| 4026 default: | |
| 4027 fvdef = vignore; | |
| 4028 } | |
| 4029 break; | |
| 458 | 4030 case '<': |
| 2325 | 4031 if (cplpl |
| 4032 && (structdef == stagseen || fvdef == fvnameseen)) | |
| 458 | 4033 { |
| 2325 | 4034 templatelev++; |
| 458 | 4035 break; |
| 4036 } | |
| 4037 goto resetfvdef; | |
| 4038 case '>': | |
| 2325 | 4039 if (templatelev > 0) |
| 458 | 4040 { |
| 2325 | 4041 templatelev--; |
| 458 | 4042 break; |
| 4043 } | |
| 4044 goto resetfvdef; | |
| 428 | 4045 case '+': |
| 4046 case '-': | |
| 2325 | 4047 if (objdef == oinbody && bracelev == 0) |
| 428 | 4048 { |
| 4049 objdef = omethodsign; | |
| 4050 break; | |
| 4051 } | |
| 4052 /* FALLTHRU */ | |
| 458 | 4053 resetfvdef: |
| 2325 | 4054 case '#': case '~': case '&': case '%': case '/': |
| 4055 case '|': case '^': case '!': case '.': case '?': | |
| 428 | 4056 if (definedef != dnone) |
| 4057 break; | |
| 4058 /* These surely cannot follow a function tag in C. */ | |
| 4059 switch (fvdef) | |
| 4060 { | |
| 4061 case foperator: | |
| 4062 case finlist: | |
| 4063 case fignore: | |
| 4064 case vignore: | |
| 4065 break; | |
| 4066 default: | |
| 4067 fvdef = fvnone; | |
| 4068 } | |
| 4069 break; | |
| 4070 case '\0': | |
| 4071 if (objdef == otagseen) | |
| 4072 { | |
| 4073 make_C_tag (TRUE); /* an Objective C class */ | |
| 4074 objdef = oignore; | |
| 4075 } | |
| 4076 /* If a macro spans multiple lines don't reset its state. */ | |
| 4077 if (quotednl) | |
| 4078 CNL_SAVE_DEFINEDEF (); | |
| 4079 else | |
| 4080 CNL (); | |
| 4081 break; | |
| 4082 } /* switch (c) */ | |
| 4083 | |
| 4084 } /* while not eof */ | |
| 458 | 4085 |
| 4086 free (lbs[0].lb.buffer); | |
| 4087 free (lbs[1].lb.buffer); | |
| 428 | 4088 } |
| 4089 | |
| 4090 /* | |
| 4091 * Process either a C++ file or a C file depending on the setting | |
| 4092 * of a global flag. | |
| 4093 */ | |
| 442 | 4094 static void |
| 428 | 4095 default_C_entries (inf) |
| 4096 FILE *inf; | |
| 4097 { | |
| 458 | 4098 C_entries (cplusplus ? C_PLPL : C_AUTO, inf); |
| 428 | 4099 } |
| 4100 | |
| 458 | 4101 /* Always do plain C. */ |
| 442 | 4102 static void |
| 428 | 4103 plain_C_entries (inf) |
| 4104 FILE *inf; | |
| 4105 { | |
| 4106 C_entries (0, inf); | |
| 4107 } | |
| 4108 | |
| 4109 /* Always do C++. */ | |
| 442 | 4110 static void |
| 428 | 4111 Cplusplus_entries (inf) |
| 4112 FILE *inf; | |
| 4113 { | |
| 4114 C_entries (C_PLPL, inf); | |
| 4115 } | |
| 4116 | |
| 4117 /* Always do Java. */ | |
| 442 | 4118 static void |
| 428 | 4119 Cjava_entries (inf) |
| 4120 FILE *inf; | |
| 4121 { | |
| 4122 C_entries (C_JAVA, inf); | |
| 4123 } | |
| 4124 | |
| 4125 /* Always do C*. */ | |
| 442 | 4126 static void |
| 428 | 4127 Cstar_entries (inf) |
| 4128 FILE *inf; | |
| 4129 { | |
| 4130 C_entries (C_STAR, inf); | |
| 4131 } | |
| 4132 | |
| 4133 /* Always do Yacc. */ | |
| 442 | 4134 static void |
| 428 | 4135 Yacc_entries (inf) |
| 4136 FILE *inf; | |
| 4137 { | |
| 4138 C_entries (YACC, inf); | |
| 4139 } | |
| 458 | 4140 |
| 428 | 4141 |
| 709 | 4142 /* Useful macros. */ |
| 428 | 4143 #define LOOP_ON_INPUT_LINES(file_pointer, line_buffer, char_pointer) \ |
| 2225 | 4144 for (; /* loop initialization */ \ |
| 428 | 4145 !feof (file_pointer) /* loop test */ \ |
| 2225 | 4146 && /* instructions at start of loop */ \ |
| 4147 (readline (&line_buffer, file_pointer), \ | |
| 4148 char_pointer = line_buffer.buffer, \ | |
| 428 | 4149 TRUE); \ |
| 4150 ) | |
| 2554 | 4151 |
| 4152 #define LOOKING_AT(cp, kw) /* kw is the keyword, a literal string */ \ | |
| 4153 ((assert("" kw), TRUE) /* syntax error if not a literal string */ \ | |
| 4154 && strneq ((cp), kw, sizeof(kw)-1) /* cp points at kw */ \ | |
| 4155 && notinname ((cp)[sizeof(kw)-1]) /* end of kw */ \ | |
| 4156 && ((cp) = skip_spaces((cp)+sizeof(kw)-1))) /* skip spaces */ | |
| 4157 | |
| 4158 /* Similar to LOOKING_AT but does not use notinname, does not skip */ | |
| 4159 #define LOOKING_AT_NOCASE(cp, kw) /* the keyword is a literal string */ \ | |
| 4160 ((assert("" kw), TRUE) /* syntax error if not a literal string */ \ | |
| 4161 && strncaseeq ((cp), kw, sizeof(kw)-1) /* cp points at kw */ \ | |
| 4162 && ((cp) += sizeof(kw)-1)) /* skip spaces */ | |
| 428 | 4163 |
| 4164 /* | |
| 4165 * Read a file, but do no processing. This is used to do regexp | |
| 4166 * matching on files that have no language defined. | |
| 4167 */ | |
| 442 | 4168 static void |
| 428 | 4169 just_read_file (inf) |
| 4170 FILE *inf; | |
| 4171 { | |
| 4172 register char *dummy; | |
| 4173 | |
| 4174 LOOP_ON_INPUT_LINES (inf, lb, dummy) | |
| 4175 continue; | |
| 4176 } | |
| 458 | 4177 |
| 428 | 4178 |
| 4179 /* Fortran parsing */ | |
| 4180 | |
| 2225 | 4181 static void F_takeprec __P((void)); |
| 4182 static void F_getit __P((FILE *)); | |
| 428 | 4183 |
| 442 | 4184 static void |
| 2225 | 4185 F_takeprec () |
| 428 | 4186 { |
| 4187 dbp = skip_spaces (dbp); | |
| 4188 if (*dbp != '*') | |
| 4189 return; | |
| 4190 dbp++; | |
| 4191 dbp = skip_spaces (dbp); | |
| 4192 if (strneq (dbp, "(*)", 3)) | |
| 4193 { | |
| 4194 dbp += 3; | |
| 4195 return; | |
| 4196 } | |
| 458 | 4197 if (!ISDIGIT (*dbp)) |
| 428 | 4198 { |
| 4199 --dbp; /* force failure */ | |
| 4200 return; | |
| 4201 } | |
| 4202 do | |
| 4203 dbp++; | |
| 458 | 4204 while (ISDIGIT (*dbp)); |
| 428 | 4205 } |
| 4206 | |
| 4207 static void | |
| 2225 | 4208 F_getit (inf) |
| 428 | 4209 FILE *inf; |
| 4210 { | |
| 4211 register char *cp; | |
| 4212 | |
| 4213 dbp = skip_spaces (dbp); | |
| 4214 if (*dbp == '\0') | |
| 4215 { | |
| 2225 | 4216 readline (&lb, inf); |
| 428 | 4217 dbp = lb.buffer; |
| 4218 if (dbp[5] != '&') | |
| 4219 return; | |
| 4220 dbp += 6; | |
| 4221 dbp = skip_spaces (dbp); | |
| 4222 } | |
| 458 | 4223 if (!ISALPHA (*dbp) && *dbp != '_' && *dbp != '$') |
| 428 | 4224 return; |
| 4225 for (cp = dbp + 1; *cp != '\0' && intoken (*cp); cp++) | |
| 4226 continue; | |
| 2225 | 4227 make_tag (dbp, cp-dbp, TRUE, |
| 4228 lb.buffer, cp - lb.buffer + 1, lineno, linecharno); | |
| 428 | 4229 } |
| 4230 | |
| 4231 | |
| 442 | 4232 static void |
| 428 | 4233 Fortran_functions (inf) |
| 4234 FILE *inf; | |
| 4235 { | |
| 4236 LOOP_ON_INPUT_LINES (inf, lb, dbp) | |
| 4237 { | |
| 4238 if (*dbp == '%') | |
| 4239 dbp++; /* Ratfor escape to fortran */ | |
| 4240 dbp = skip_spaces (dbp); | |
| 4241 if (*dbp == '\0') | |
| 4242 continue; | |
| 4243 switch (lowcase (*dbp)) | |
| 4244 { | |
| 4245 case 'i': | |
| 2225 | 4246 if (nocase_tail ("integer")) |
| 4247 F_takeprec (); | |
| 428 | 4248 break; |
| 4249 case 'r': | |
| 2225 | 4250 if (nocase_tail ("real")) |
| 4251 F_takeprec (); | |
| 428 | 4252 break; |
| 4253 case 'l': | |
| 2225 | 4254 if (nocase_tail ("logical")) |
| 4255 F_takeprec (); | |
| 428 | 4256 break; |
| 4257 case 'c': | |
| 2225 | 4258 if (nocase_tail ("complex") || nocase_tail ("character")) |
| 4259 F_takeprec (); | |
| 428 | 4260 break; |
| 4261 case 'd': | |
| 2225 | 4262 if (nocase_tail ("double")) |
| 428 | 4263 { |
| 4264 dbp = skip_spaces (dbp); | |
| 4265 if (*dbp == '\0') | |
| 4266 continue; | |
| 2225 | 4267 if (nocase_tail ("precision")) |
| 428 | 4268 break; |
| 4269 continue; | |
| 4270 } | |
| 4271 break; | |
| 4272 } | |
| 4273 dbp = skip_spaces (dbp); | |
| 4274 if (*dbp == '\0') | |
| 4275 continue; | |
| 4276 switch (lowcase (*dbp)) | |
| 4277 { | |
| 4278 case 'f': | |
| 2225 | 4279 if (nocase_tail ("function")) |
| 4280 F_getit (inf); | |
| 428 | 4281 continue; |
| 4282 case 's': | |
| 2225 | 4283 if (nocase_tail ("subroutine")) |
| 4284 F_getit (inf); | |
| 428 | 4285 continue; |
| 4286 case 'e': | |
| 2225 | 4287 if (nocase_tail ("entry")) |
| 4288 F_getit (inf); | |
| 428 | 4289 continue; |
| 4290 case 'b': | |
| 2225 | 4291 if (nocase_tail ("blockdata") || nocase_tail ("block data")) |
| 428 | 4292 { |
| 4293 dbp = skip_spaces (dbp); | |
| 4294 if (*dbp == '\0') /* assume un-named */ | |
| 2225 | 4295 make_tag ("blockdata", 9, TRUE, |
| 4296 lb.buffer, dbp - lb.buffer, lineno, linecharno); | |
| 428 | 4297 else |
| 2225 | 4298 F_getit (inf); /* look for name */ |
| 428 | 4299 } |
| 4300 continue; | |
| 4301 } | |
| 4302 } | |
| 4303 } | |
| 458 | 4304 |
| 428 | 4305 |
| 4306 /* | |
| 4307 * Ada parsing | |
| 2225 | 4308 * Original code by |
| 4309 * Philippe Waroquiers (1998) | |
| 428 | 4310 */ |
| 442 | 4311 |
| 2225 | 4312 static void Ada_getit __P((FILE *, char *)); |
| 442 | 4313 |
| 428 | 4314 /* Once we are positioned after an "interesting" keyword, let's get |
| 4315 the real tag value necessary. */ | |
| 4316 static void | |
| 2225 | 4317 Ada_getit (inf, name_qualifier) |
| 428 | 4318 FILE *inf; |
| 4319 char *name_qualifier; | |
| 4320 { | |
| 4321 register char *cp; | |
| 4322 char *name; | |
| 4323 char c; | |
| 4324 | |
| 4325 while (!feof (inf)) | |
| 4326 { | |
| 4327 dbp = skip_spaces (dbp); | |
| 4328 if (*dbp == '\0' | |
| 4329 || (dbp[0] == '-' && dbp[1] == '-')) | |
| 4330 { | |
| 2225 | 4331 readline (&lb, inf); |
| 428 | 4332 dbp = lb.buffer; |
| 4333 } | |
| 2225 | 4334 switch (lowcase(*dbp)) |
| 428 | 4335 { |
| 4336 case 'b': | |
| 2225 | 4337 if (nocase_tail ("body")) |
| 428 | 4338 { |
| 4339 /* Skipping body of procedure body or package body or .... | |
| 4340 resetting qualifier to body instead of spec. */ | |
| 4341 name_qualifier = "/b"; | |
| 4342 continue; | |
| 4343 } | |
| 4344 break; | |
| 4345 case 't': | |
| 4346 /* Skipping type of task type or protected type ... */ | |
| 2225 | 4347 if (nocase_tail ("type")) |
| 428 | 4348 continue; |
| 4349 break; | |
| 4350 } | |
| 4351 if (*dbp == '"') | |
| 4352 { | |
| 4353 dbp += 1; | |
| 4354 for (cp = dbp; *cp != '\0' && *cp != '"'; cp++) | |
| 4355 continue; | |
| 4356 } | |
| 4357 else | |
| 4358 { | |
| 4359 dbp = skip_spaces (dbp); | |
| 4360 for (cp = dbp; | |
| 4361 (*cp != '\0' | |
| 458 | 4362 && (ISALPHA (*cp) || ISDIGIT (*cp) || *cp == '_' || *cp == '.')); |
| 428 | 4363 cp++) |
| 4364 continue; | |
| 4365 if (cp == dbp) | |
| 4366 return; | |
| 4367 } | |
| 4368 c = *cp; | |
| 4369 *cp = '\0'; | |
| 4370 name = concat (dbp, name_qualifier, ""); | |
| 4371 *cp = c; | |
| 2225 | 4372 make_tag (name, strlen (name), TRUE, |
| 4373 lb.buffer, cp - lb.buffer + 1, lineno, linecharno); | |
| 4374 free (name); | |
| 428 | 4375 if (c == '"') |
| 4376 dbp = cp + 1; | |
| 4377 return; | |
| 4378 } | |
| 4379 } | |
| 4380 | |
| 442 | 4381 static void |
| 428 | 4382 Ada_funcs (inf) |
| 4383 FILE *inf; | |
| 4384 { | |
| 4385 bool inquote = FALSE; | |
| 2225 | 4386 bool skip_till_semicolumn = FALSE; |
| 428 | 4387 |
| 4388 LOOP_ON_INPUT_LINES (inf, lb, dbp) | |
| 4389 { | |
| 4390 while (*dbp != '\0') | |
| 4391 { | |
| 4392 /* Skip a string i.e. "abcd". */ | |
| 4393 if (inquote || (*dbp == '"')) | |
| 4394 { | |
| 4395 dbp = etags_strchr ((inquote) ? dbp : dbp+1, '"'); | |
| 4396 if (dbp != NULL) | |
| 4397 { | |
| 4398 inquote = FALSE; | |
| 4399 dbp += 1; | |
| 4400 continue; /* advance char */ | |
| 4401 } | |
| 4402 else | |
| 4403 { | |
| 4404 inquote = TRUE; | |
| 4405 break; /* advance line */ | |
| 4406 } | |
| 4407 } | |
| 4408 | |
| 4409 /* Skip comments. */ | |
| 4410 if (dbp[0] == '-' && dbp[1] == '-') | |
| 4411 break; /* advance line */ | |
| 4412 | |
| 4413 /* Skip character enclosed in single quote i.e. 'a' | |
| 4414 and skip single quote starting an attribute i.e. 'Image. */ | |
| 4415 if (*dbp == '\'') | |
| 4416 { | |
| 4417 dbp++ ; | |
| 4418 if (*dbp != '\0') | |
| 4419 dbp++; | |
| 4420 continue; | |
| 4421 } | |
| 4422 | |
| 2225 | 4423 if (skip_till_semicolumn) |
| 4424 { | |
| 4425 if (*dbp == ';') | |
| 4426 skip_till_semicolumn = FALSE; | |
| 4427 dbp++; | |
| 4428 continue; /* advance char */ | |
| 4429 } | |
| 4430 | |
| 428 | 4431 /* Search for beginning of a token. */ |
| 4432 if (!begtoken (*dbp)) | |
| 4433 { | |
| 4434 dbp++; | |
| 4435 continue; /* advance char */ | |
| 4436 } | |
| 4437 | |
| 4438 /* We are at the beginning of a token. */ | |
| 2225 | 4439 switch (lowcase(*dbp)) |
| 428 | 4440 { |
| 4441 case 'f': | |
| 2225 | 4442 if (!packages_only && nocase_tail ("function")) |
| 4443 Ada_getit (inf, "/f"); | |
| 428 | 4444 else |
| 4445 break; /* from switch */ | |
| 4446 continue; /* advance char */ | |
| 4447 case 'p': | |
| 2225 | 4448 if (!packages_only && nocase_tail ("procedure")) |
| 4449 Ada_getit (inf, "/p"); | |
| 4450 else if (nocase_tail ("package")) | |
| 4451 Ada_getit (inf, "/s"); | |
| 4452 else if (nocase_tail ("protected")) /* protected type */ | |
| 4453 Ada_getit (inf, "/t"); | |
| 428 | 4454 else |
| 4455 break; /* from switch */ | |
| 4456 continue; /* advance char */ | |
| 2225 | 4457 |
| 4458 case 'u': | |
| 4459 if (typedefs && !packages_only && nocase_tail ("use")) | |
| 4460 { | |
| 4461 /* when tagging types, avoid tagging use type Pack.Typename; | |
| 4462 for this, we will skip everything till a ; */ | |
| 4463 skip_till_semicolumn = TRUE; | |
| 4464 continue; /* advance char */ | |
| 4465 } | |
| 4466 | |
| 428 | 4467 case 't': |
| 2225 | 4468 if (!packages_only && nocase_tail ("task")) |
| 4469 Ada_getit (inf, "/k"); | |
| 4470 else if (typedefs && !packages_only && nocase_tail ("type")) | |
| 428 | 4471 { |
| 2225 | 4472 Ada_getit (inf, "/t"); |
| 428 | 4473 while (*dbp != '\0') |
| 4474 dbp += 1; | |
| 4475 } | |
| 4476 else | |
| 4477 break; /* from switch */ | |
| 4478 continue; /* advance char */ | |
| 4479 } | |
| 4480 | |
| 4481 /* Look for the end of the token. */ | |
| 4482 while (!endtoken (*dbp)) | |
| 4483 dbp++; | |
| 4484 | |
| 4485 } /* advance char */ | |
| 4486 } /* advance line */ | |
| 4487 } | |
| 458 | 4488 |
| 428 | 4489 |
| 4490 /* | |
| 4491 * Unix and microcontroller assembly tag handling | |
| 2225 | 4492 * Labels: /^[a-zA-Z_.$][a-zA_Z0-9_.$]*[: ^I^J]/ |
| 4493 * Idea by Bob Weiner, Motorola Inc. (1994) | |
| 428 | 4494 */ |
| 442 | 4495 static void |
| 428 | 4496 Asm_labels (inf) |
| 4497 FILE *inf; | |
| 4498 { | |
| 4499 register char *cp; | |
| 4500 | |
| 4501 LOOP_ON_INPUT_LINES (inf, lb, cp) | |
| 4502 { | |
| 4503 /* If first char is alphabetic or one of [_.$], test for colon | |
| 4504 following identifier. */ | |
| 458 | 4505 if (ISALPHA (*cp) || *cp == '_' || *cp == '.' || *cp == '$') |
| 428 | 4506 { |
| 4507 /* Read past label. */ | |
| 4508 cp++; | |
| 458 | 4509 while (ISALNUM (*cp) || *cp == '_' || *cp == '.' || *cp == '$') |
| 428 | 4510 cp++; |
| 442 | 4511 if (*cp == ':' || iswhite (*cp)) |
| 2225 | 4512 /* Found end of label, so copy it and add it to the table. */ |
| 4513 make_tag (lb.buffer, cp - lb.buffer, TRUE, | |
| 428 | 4514 lb.buffer, cp - lb.buffer + 1, lineno, linecharno); |
| 4515 } | |
| 4516 } | |
| 4517 } | |
| 458 | 4518 |
| 428 | 4519 |
| 4520 /* | |
| 458 | 4521 * Perl support |
| 2225 | 4522 * Perl sub names: /^sub[ \t\n]+[^ \t\n{]+/ |
| 428 | 4523 * Perl variable names: /^(my|local).../ |
| 2225 | 4524 * Original code by Bart Robinson <lomew@cs.utah.edu> (1995) |
| 4525 * Additions by Michael Ernst <mernst@alum.mit.edu> (1997) | |
| 4526 * Ideas by Kai Großjohann <Kai.Grossjohann@CS.Uni-Dortmund.DE> (2001) | |
| 428 | 4527 */ |
| 442 | 4528 static void |
| 428 | 4529 Perl_functions (inf) |
| 4530 FILE *inf; | |
| 4531 { | |
| 2225 | 4532 char *package = savestr ("main"); /* current package name */ |
| 428 | 4533 register char *cp; |
| 4534 | |
| 4535 LOOP_ON_INPUT_LINES (inf, lb, cp) | |
| 4536 { | |
| 2225 | 4537 skip_spaces(cp); |
| 4538 | |
| 4539 if (LOOKING_AT (cp, "package")) | |
| 4540 { | |
| 4541 free (package); | |
| 4542 get_tag (cp, &package); | |
| 4543 } | |
| 4544 else if (LOOKING_AT (cp, "sub")) | |
| 428 | 4545 { |
| 2225 | 4546 char *pos; |
| 4547 char *sp = cp; | |
| 4548 | |
| 4549 while (!notinname (*cp)) | |
| 4550 cp++; | |
| 4551 if (cp == sp) | |
| 4552 continue; /* nothing found */ | |
| 4553 if ((pos = etags_strchr (sp, ':')) != NULL | |
| 4554 && pos < cp && pos[1] == ':') | |
| 4555 /* The name is already qualified. */ | |
| 4556 make_tag (sp, cp - sp, TRUE, | |
| 4557 lb.buffer, cp - lb.buffer + 1, lineno, linecharno); | |
| 4558 else | |
| 4559 /* Qualify it. */ | |
| 4560 { | |
| 4561 char savechar, *name; | |
| 4562 | |
| 4563 savechar = *cp; | |
| 4564 *cp = '\0'; | |
| 4565 name = concat (package, "::", sp); | |
| 4566 *cp = savechar; | |
| 4567 make_tag (name, strlen(name), TRUE, | |
| 4568 lb.buffer, cp - lb.buffer + 1, lineno, linecharno); | |
| 4569 free (name); | |
| 4570 } | |
| 428 | 4571 } |
| 2225 | 4572 else if (globals) /* only if we are tagging global vars */ |
| 428 | 4573 { |
| 2225 | 4574 /* Skip a qualifier, if any. */ |
| 4575 bool qual = LOOKING_AT (cp, "my") || LOOKING_AT (cp, "local"); | |
| 428 | 4576 /* After "my" or "local", but before any following paren or space. */ |
| 2225 | 4577 char *varstart = cp; |
| 4578 | |
| 4579 if (qual /* should this be removed? If yes, how? */ | |
| 4580 && (*cp == '$' || *cp == '@' || *cp == '%')) | |
| 428 | 4581 { |
| 2225 | 4582 varstart += 1; |
| 4583 do | |
| 428 | 4584 cp++; |
| 2225 | 4585 while (ISALNUM (*cp) || *cp == '_'); |
| 428 | 4586 } |
| 2225 | 4587 else if (qual) |
| 428 | 4588 { |
| 4589 /* Should be examining a variable list at this point; | |
| 4590 could insist on seeing an open parenthesis. */ | |
| 4591 while (*cp != '\0' && *cp != ';' && *cp != '=' && *cp != ')') | |
| 4592 cp++; | |
| 4593 } | |
| 2225 | 4594 else |
| 4595 continue; | |
| 4596 | |
| 4597 make_tag (varstart, cp - varstart, FALSE, | |
| 4598 lb.buffer, cp - lb.buffer + 1, lineno, linecharno); | |
| 428 | 4599 } |
| 4600 } | |
| 3517 | 4601 free (package); |
| 428 | 4602 } |
| 458 | 4603 |
| 709 | 4604 |
| 428 | 4605 /* |
| 458 | 4606 * Python support |
| 2225 | 4607 * Look for /^[\t]*def[ \t\n]+[^ \t\n(:]+/ or /^class[ \t\n]+[^ \t\n(:]+/ |
| 4608 * Idea by Eric S. Raymond <esr@thyrsus.com> (1997) | |
| 4609 * More ideas by seb bacon <seb@jamkit.com> (2002) | |
| 428 | 4610 */ |
| 442 | 4611 static void |
| 428 | 4612 Python_functions (inf) |
| 4613 FILE *inf; | |
| 4614 { | |
| 4615 register char *cp; | |
| 4616 | |
| 4617 LOOP_ON_INPUT_LINES (inf, lb, cp) | |
| 2225 | 4618 { |
| 4619 cp = skip_spaces (cp); | |
| 4620 if (LOOKING_AT (cp, "def") || LOOKING_AT (cp, "class")) | |
| 4621 { | |
| 4622 char *name = cp; | |
| 4623 while (!notinname (*cp) && *cp != ':') | |
| 4624 cp++; | |
| 4625 make_tag (name, cp - name, TRUE, | |
| 4626 lb.buffer, cp - lb.buffer + 1, lineno, linecharno); | |
| 4627 } | |
| 4628 } | |
| 709 | 4629 } |
| 4630 | |
| 4631 | |
| 4632 /* | |
| 4633 * PHP support | |
| 4634 * Look for: | |
| 4635 * - /^[ \t]*function[ \t\n]+[^ \t\n(]+/ | |
| 4636 * - /^[ \t]*class[ \t\n]+[^ \t\n]+/ | |
| 4637 * - /^[ \t]*define\(\"[^\"]+/ | |
| 4638 * Only with --members: | |
| 4639 * - /^[ \t]*var[ \t\n]+\$[^ \t\n=;]/ | |
| 2225 | 4640 * Idea by Diez B. Roggisch (2001) |
| 709 | 4641 */ |
| 4642 static void | |
| 4643 PHP_functions (inf) | |
| 4644 FILE *inf; | |
| 4645 { | |
| 2225 | 4646 register char *cp, *name; |
| 709 | 4647 bool search_identifier = FALSE; |
| 4648 | |
| 4649 LOOP_ON_INPUT_LINES (inf, lb, cp) | |
| 428 | 4650 { |
| 709 | 4651 cp = skip_spaces (cp); |
| 2225 | 4652 name = cp; |
| 709 | 4653 if (search_identifier |
| 4654 && *cp != '\0') | |
| 428 | 4655 { |
| 2225 | 4656 while (!notinname (*cp)) |
| 428 | 4657 cp++; |
| 2225 | 4658 make_tag (name, cp - name, TRUE, |
| 4659 lb.buffer, cp - lb.buffer + 1, lineno, linecharno); | |
| 709 | 4660 search_identifier = FALSE; |
| 428 | 4661 } |
| 709 | 4662 else if (LOOKING_AT (cp, "function")) |
| 4663 { | |
| 4664 if(*cp == '&') | |
| 4665 cp = skip_spaces (cp+1); | |
| 4666 if(*cp != '\0') | |
| 4667 { | |
| 2225 | 4668 name = cp; |
| 4669 while (!notinname (*cp)) | |
| 709 | 4670 cp++; |
| 2225 | 4671 make_tag (name, cp - name, TRUE, |
| 4672 lb.buffer, cp - lb.buffer + 1, lineno, linecharno); | |
| 709 | 4673 } |
| 4674 else | |
| 4675 search_identifier = TRUE; | |
| 4676 } | |
| 4677 else if (LOOKING_AT (cp, "class")) | |
| 428 | 4678 { |
| 709 | 4679 if (*cp != '\0') |
| 4680 { | |
| 2225 | 4681 name = cp; |
| 709 | 4682 while (*cp != '\0' && !iswhite (*cp)) |
| 4683 cp++; | |
| 2225 | 4684 make_tag (name, cp - name, FALSE, |
| 4685 lb.buffer, cp - lb.buffer + 1, lineno, linecharno); | |
| 709 | 4686 } |
| 4687 else | |
| 4688 search_identifier = TRUE; | |
| 4689 } | |
| 4690 else if (strneq (cp, "define", 6) | |
| 4691 && (cp = skip_spaces (cp+6)) | |
| 4692 && *cp++ == '(' | |
| 4693 && (*cp == '"' || *cp == '\'')) | |
| 4694 { | |
| 4695 char quote = *cp++; | |
| 2225 | 4696 name = cp; |
| 709 | 4697 while (*cp != quote && *cp != '\0') |
| 428 | 4698 cp++; |
| 2225 | 4699 make_tag (name, cp - name, FALSE, |
| 4700 lb.buffer, cp - lb.buffer + 1, lineno, linecharno); | |
| 709 | 4701 } |
| 4702 else if (members | |
| 4703 && LOOKING_AT (cp, "var") | |
| 4704 && *cp == '$') | |
| 4705 { | |
| 2225 | 4706 name = cp; |
| 4707 while (!notinname(*cp)) | |
| 709 | 4708 cp++; |
| 2225 | 4709 make_tag (name, cp - name, FALSE, |
| 4710 lb.buffer, cp - lb.buffer + 1, lineno, linecharno); | |
| 428 | 4711 } |
| 4712 } | |
| 4713 } | |
| 458 | 4714 |
| 428 | 4715 |
| 2225 | 4716 /* |
| 428 | 4717 * Cobol tag functions |
| 4718 * We could look for anything that could be a paragraph name. | |
| 4719 * i.e. anything that starts in column 8 is one word and ends in a full stop. | |
| 2225 | 4720 * Idea by Corny de Souza (1993) |
| 428 | 4721 */ |
| 442 | 4722 static void |
| 428 | 4723 Cobol_paragraphs (inf) |
| 4724 FILE *inf; | |
| 4725 { | |
| 4726 register char *bp, *ep; | |
| 4727 | |
| 4728 LOOP_ON_INPUT_LINES (inf, lb, bp) | |
| 4729 { | |
| 4730 if (lb.len < 9) | |
| 4731 continue; | |
| 4732 bp += 8; | |
| 4733 | |
| 4734 /* If eoln, compiler option or comment ignore whole line. */ | |
| 458 | 4735 if (bp[-1] != ' ' || !ISALNUM (bp[0])) |
| 428 | 4736 continue; |
| 4737 | |
| 458 | 4738 for (ep = bp; ISALNUM (*ep) || *ep == '-'; ep++) |
| 428 | 4739 continue; |
| 4740 if (*ep++ == '.') | |
| 2225 | 4741 make_tag (bp, ep - bp, TRUE, |
| 4742 lb.buffer, ep - lb.buffer + 1, lineno, linecharno); | |
| 428 | 4743 } |
| 4744 } | |
| 458 | 4745 |
| 4746 | |
| 4747 /* | |
| 4748 * Makefile support | |
| 2225 | 4749 * Ideas by Assar Westerlund <assar@sics.se> (2001) |
| 458 | 4750 */ |
| 4751 static void | |
| 4752 Makefile_targets (inf) | |
| 4753 FILE *inf; | |
| 4754 { | |
| 4755 register char *bp; | |
| 4756 | |
| 4757 LOOP_ON_INPUT_LINES (inf, lb, bp) | |
| 4758 { | |
| 4759 if (*bp == '\t' || *bp == '#') | |
| 4760 continue; | |
| 4761 while (*bp != '\0' && *bp != '=' && *bp != ':') | |
| 4762 bp++; | |
| 2225 | 4763 if (*bp == ':' || (globals && *bp == '=')) |
| 3876 | 4764 { |
| 4765 /* We should detect if there is more than one tag, but we do not. | |
| 4766 We just skip initial and final spaces. */ | |
| 4767 char * namestart = skip_spaces (lb.buffer); | |
| 4768 while (--bp > namestart) | |
| 4769 if (!notinname (*bp)) | |
| 4770 break; | |
| 4771 make_tag (namestart, bp - namestart + 1, TRUE, | |
| 4772 lb.buffer, bp - lb.buffer + 2, lineno, linecharno); | |
| 4773 } | |
| 458 | 4774 } |
| 4775 } | |
| 4776 | |
| 428 | 4777 |
| 4778 /* | |
| 2225 | 4779 * Pascal parsing |
| 4780 * Original code by Mosur K. Mohan (1989) | |
| 4781 * | |
| 428 | 4782 * Locates tags for procedures & functions. Doesn't do any type- or |
| 4783 * var-definitions. It does look for the keyword "extern" or | |
| 4784 * "forward" immediately following the procedure statement; if found, | |
| 4785 * the tag is skipped. | |
| 4786 */ | |
| 442 | 4787 static void |
| 428 | 4788 Pascal_functions (inf) |
| 4789 FILE *inf; | |
| 4790 { | |
| 4791 linebuffer tline; /* mostly copied from C_entries */ | |
| 4792 long save_lcno; | |
| 2225 | 4793 int save_lineno, namelen, taglen; |
| 4794 char c, *name; | |
| 428 | 4795 |
| 4796 bool /* each of these flags is TRUE iff: */ | |
| 4797 incomment, /* point is inside a comment */ | |
| 4798 inquote, /* point is inside '..' string */ | |
| 4799 get_tagname, /* point is after PROCEDURE/FUNCTION | |
| 4800 keyword, so next item = potential tag */ | |
| 4801 found_tag, /* point is after a potential tag */ | |
| 4802 inparms, /* point is within parameter-list */ | |
| 4803 verify_tag; /* point has passed the parm-list, so the | |
| 4804 next token will determine whether this | |
| 4805 is a FORWARD/EXTERN to be ignored, or | |
| 4806 whether it is a real tag */ | |
| 4807 | |
| 2225 | 4808 save_lcno = save_lineno = namelen = taglen = 0; /* keep compiler quiet */ |
| 4809 name = NULL; /* keep compiler quiet */ | |
| 428 | 4810 dbp = lb.buffer; |
| 4811 *dbp = '\0'; | |
| 2225 | 4812 linebuffer_init (&tline); |
| 428 | 4813 |
| 4814 incomment = inquote = FALSE; | |
| 4815 found_tag = FALSE; /* have a proc name; check if extern */ | |
| 2225 | 4816 get_tagname = FALSE; /* found "procedure" keyword */ |
| 428 | 4817 inparms = FALSE; /* found '(' after "proc" */ |
| 4818 verify_tag = FALSE; /* check if "extern" is ahead */ | |
| 4819 | |
| 4820 | |
| 4821 while (!feof (inf)) /* long main loop to get next char */ | |
| 4822 { | |
| 4823 c = *dbp++; | |
| 4824 if (c == '\0') /* if end of line */ | |
| 4825 { | |
| 2225 | 4826 readline (&lb, inf); |
| 428 | 4827 dbp = lb.buffer; |
| 4828 if (*dbp == '\0') | |
| 4829 continue; | |
| 4830 if (!((found_tag && verify_tag) | |
| 4831 || get_tagname)) | |
| 4832 c = *dbp++; /* only if don't need *dbp pointing | |
| 4833 to the beginning of the name of | |
| 4834 the procedure or function */ | |
| 4835 } | |
| 4836 if (incomment) | |
| 4837 { | |
| 4838 if (c == '}') /* within { } comments */ | |
| 4839 incomment = FALSE; | |
| 4840 else if (c == '*' && *dbp == ')') /* within (* *) comments */ | |
| 4841 { | |
| 4842 dbp++; | |
| 4843 incomment = FALSE; | |
| 4844 } | |
| 4845 continue; | |
| 4846 } | |
| 4847 else if (inquote) | |
| 4848 { | |
| 4849 if (c == '\'') | |
| 4850 inquote = FALSE; | |
| 4851 continue; | |
| 4852 } | |
| 4853 else | |
| 4854 switch (c) | |
| 4855 { | |
| 4856 case '\'': | |
| 4857 inquote = TRUE; /* found first quote */ | |
| 4858 continue; | |
| 4859 case '{': /* found open { comment */ | |
| 4860 incomment = TRUE; | |
| 4861 continue; | |
| 4862 case '(': | |
| 4863 if (*dbp == '*') /* found open (* comment */ | |
| 4864 { | |
| 4865 incomment = TRUE; | |
| 4866 dbp++; | |
| 4867 } | |
| 4868 else if (found_tag) /* found '(' after tag, i.e., parm-list */ | |
| 4869 inparms = TRUE; | |
| 4870 continue; | |
| 4871 case ')': /* end of parms list */ | |
| 4872 if (inparms) | |
| 4873 inparms = FALSE; | |
| 4874 continue; | |
| 4875 case ';': | |
| 4876 if (found_tag && !inparms) /* end of proc or fn stmt */ | |
| 4877 { | |
| 4878 verify_tag = TRUE; | |
| 4879 break; | |
| 4880 } | |
| 4881 continue; | |
| 4882 } | |
| 4883 if (found_tag && verify_tag && (*dbp != ' ')) | |
| 4884 { | |
| 2225 | 4885 /* Check if this is an "extern" declaration. */ |
| 428 | 4886 if (*dbp == '\0') |
| 4887 continue; | |
| 4888 if (lowcase (*dbp == 'e')) | |
| 4889 { | |
| 2225 | 4890 if (nocase_tail ("extern")) /* superfluous, really! */ |
| 428 | 4891 { |
| 4892 found_tag = FALSE; | |
| 4893 verify_tag = FALSE; | |
| 4894 } | |
| 4895 } | |
| 4896 else if (lowcase (*dbp) == 'f') | |
| 4897 { | |
| 2225 | 4898 if (nocase_tail ("forward")) /* check for forward reference */ |
| 428 | 4899 { |
| 4900 found_tag = FALSE; | |
| 4901 verify_tag = FALSE; | |
| 4902 } | |
| 4903 } | |
| 4904 if (found_tag && verify_tag) /* not external proc, so make tag */ | |
| 4905 { | |
| 4906 found_tag = FALSE; | |
| 4907 verify_tag = FALSE; | |
| 2225 | 4908 make_tag (name, namelen, TRUE, |
| 4909 tline.buffer, taglen, save_lineno, save_lcno); | |
| 428 | 4910 continue; |
| 4911 } | |
| 4912 } | |
| 4913 if (get_tagname) /* grab name of proc or fn */ | |
| 4914 { | |
| 2225 | 4915 char *cp; |
| 4916 | |
| 428 | 4917 if (*dbp == '\0') |
| 4918 continue; | |
| 4919 | |
| 2225 | 4920 /* Find block name. */ |
| 4921 for (cp = dbp + 1; *cp != '\0' && !endtoken (*cp); cp++) | |
| 4922 continue; | |
| 4923 | |
| 4924 /* Save all values for later tagging. */ | |
| 458 | 4925 linebuffer_setlen (&tline, lb.len); |
| 428 | 4926 strcpy (tline.buffer, lb.buffer); |
| 4927 save_lineno = lineno; | |
| 4928 save_lcno = linecharno; | |
| 2225 | 4929 name = tline.buffer + (dbp - lb.buffer); |
| 4930 namelen = cp - dbp; | |
| 4931 taglen = cp - lb.buffer + 1; | |
| 4932 | |
| 428 | 4933 dbp = cp; /* set dbp to e-o-token */ |
| 4934 get_tagname = FALSE; | |
| 4935 found_tag = TRUE; | |
| 4936 continue; | |
| 4937 | |
| 2225 | 4938 /* And proceed to check for "extern". */ |
| 428 | 4939 } |
| 4940 else if (!incomment && !inquote && !found_tag) | |
| 4941 { | |
| 2225 | 4942 /* Check for proc/fn keywords. */ |
| 428 | 4943 switch (lowcase (c)) |
| 4944 { | |
| 4945 case 'p': | |
| 2225 | 4946 if (nocase_tail ("rocedure")) /* c = 'p', dbp has advanced */ |
| 428 | 4947 get_tagname = TRUE; |
| 4948 continue; | |
| 4949 case 'f': | |
| 2225 | 4950 if (nocase_tail ("unction")) |
| 428 | 4951 get_tagname = TRUE; |
| 4952 continue; | |
| 4953 } | |
| 4954 } | |
| 2225 | 4955 } /* while not eof */ |
| 428 | 4956 |
| 4957 free (tline.buffer); | |
| 4958 } | |
| 458 | 4959 |
| 428 | 4960 |
| 4961 /* | |
| 458 | 4962 * Lisp tag functions |
| 428 | 4963 * look for (def or (DEF, quote or QUOTE |
| 4964 */ | |
| 442 | 4965 |
| 709 | 4966 static void L_getit __P((void)); |
| 442 | 4967 |
| 428 | 4968 static void |
| 4969 L_getit () | |
| 4970 { | |
| 4971 if (*dbp == '\'') /* Skip prefix quote */ | |
| 4972 dbp++; | |
| 4973 else if (*dbp == '(') | |
| 4974 { | |
| 2225 | 4975 dbp++; |
| 4976 /* Try to skip "(quote " */ | |
| 4977 if (!LOOKING_AT (dbp, "quote") && !LOOKING_AT (dbp, "QUOTE")) | |
| 4978 /* Ok, then skip "(" before name in (defstruct (foo)) */ | |
| 4979 dbp = skip_spaces (dbp); | |
| 428 | 4980 } |
| 2225 | 4981 get_tag (dbp, NULL); |
| 428 | 4982 } |
| 4983 | |
| 442 | 4984 static void |
| 428 | 4985 Lisp_functions (inf) |
| 4986 FILE *inf; | |
| 4987 { | |
| 4988 LOOP_ON_INPUT_LINES (inf, lb, dbp) | |
| 4989 { | |
| 2225 | 4990 if (dbp[0] != '(') |
| 4991 continue; | |
| 4992 | |
| 4993 if (strneq (dbp+1, "def", 3) || strneq (dbp+1, "DEF", 3)) | |
| 428 | 4994 { |
| 2225 | 4995 dbp = skip_non_spaces (dbp); |
| 4996 dbp = skip_spaces (dbp); | |
| 4997 L_getit (); | |
| 4998 } | |
| 4999 else | |
| 5000 { | |
| 5001 /* Check for (foo::defmumble name-defined ... */ | |
| 5002 do | |
| 5003 dbp++; | |
| 5004 while (!notinname (*dbp) && *dbp != ':'); | |
| 5005 if (*dbp == ':') | |
| 428 | 5006 { |
| 5007 do | |
| 5008 dbp++; | |
| 2225 | 5009 while (*dbp == ':'); |
| 5010 | |
| 5011 if (strneq (dbp, "def", 3) || strneq (dbp, "DEF", 3)) | |
| 428 | 5012 { |
| 2225 | 5013 dbp = skip_non_spaces (dbp); |
| 5014 dbp = skip_spaces (dbp); | |
| 5015 L_getit (); | |
| 428 | 5016 } |
| 5017 } | |
| 5018 } | |
| 5019 } | |
| 5020 } | |
| 458 | 5021 |
| 428 | 5022 |
| 5023 /* | |
| 2325 | 5024 * Lua script language parsing |
| 5025 * Original code by David A. Capello <dacap@users.sourceforge.net> (2004) | |
| 5026 * | |
| 5027 * "function" and "local function" are tags if they start at column 1. | |
| 5028 */ | |
| 5029 static void | |
| 5030 Lua_functions (inf) | |
| 5031 FILE *inf; | |
| 5032 { | |
| 5033 register char *bp; | |
| 5034 | |
| 5035 LOOP_ON_INPUT_LINES (inf, lb, bp) | |
| 5036 { | |
| 5037 if (bp[0] != 'f' && bp[0] != 'l') | |
| 5038 continue; | |
| 5039 | |
| 3876 | 5040 (void)LOOKING_AT (bp, "local"); /* skip possible "local" */ |
| 2325 | 5041 |
| 5042 if (LOOKING_AT (bp, "function")) | |
| 5043 get_tag (bp, NULL); | |
| 5044 } | |
| 5045 } | |
| 5046 | |
| 5047 | |
| 5048 /* | |
| 2554 | 5049 * Postscript tags |
| 428 | 5050 * Just look for lines where the first character is '/' |
| 5051 * Also look at "defineps" for PSWrap | |
| 2225 | 5052 * Ideas by: |
| 5053 * Richard Mlynarik <mly@adoc.xerox.com> (1997) | |
| 5054 * Masatake Yamato <masata-y@is.aist-nara.ac.jp> (1999) | |
| 428 | 5055 */ |
| 442 | 5056 static void |
| 2225 | 5057 PS_functions (inf) |
| 428 | 5058 FILE *inf; |
| 5059 { | |
| 5060 register char *bp, *ep; | |
| 5061 | |
| 5062 LOOP_ON_INPUT_LINES (inf, lb, bp) | |
| 5063 { | |
| 5064 if (bp[0] == '/') | |
| 5065 { | |
| 5066 for (ep = bp+1; | |
| 5067 *ep != '\0' && *ep != ' ' && *ep != '{'; | |
| 5068 ep++) | |
| 5069 continue; | |
| 2225 | 5070 make_tag (bp, ep - bp, TRUE, |
| 5071 lb.buffer, ep - lb.buffer + 1, lineno, linecharno); | |
| 428 | 5072 } |
| 2225 | 5073 else if (LOOKING_AT (bp, "defineps")) |
| 5074 get_tag (bp, NULL); | |
| 428 | 5075 } |
| 5076 } | |
| 5077 | |
| 5078 | |
| 5079 /* | |
| 2554 | 5080 * Forth tags |
| 5081 * Ignore anything after \ followed by space or in ( ) | |
| 5082 * Look for words defined by : | |
| 5083 * Look for constant, code, create, defer, value, and variable | |
| 5084 * OBP extensions: Look for buffer:, field, | |
| 5085 * Ideas by Eduardo Horvath <eeh@netbsd.org> (2004) | |
| 5086 */ | |
| 5087 static void | |
| 5088 Forth_words (inf) | |
| 5089 FILE *inf; | |
| 5090 { | |
| 5091 register char *bp; | |
| 5092 | |
| 5093 LOOP_ON_INPUT_LINES (inf, lb, bp) | |
| 5094 while ((bp = skip_spaces (bp))[0] != '\0') | |
| 5095 if (bp[0] == '\\' && iswhite(bp[1])) | |
| 5096 break; /* read next line */ | |
| 5097 else if (bp[0] == '(' && iswhite(bp[1])) | |
| 5098 do /* skip to ) or eol */ | |
| 5099 bp++; | |
| 5100 while (*bp != ')' && *bp != '\0'); | |
| 5101 else if ((bp[0] == ':' && iswhite(bp[1]) && bp++) | |
| 5102 || LOOKING_AT_NOCASE (bp, "constant") | |
| 5103 || LOOKING_AT_NOCASE (bp, "code") | |
| 5104 || LOOKING_AT_NOCASE (bp, "create") | |
| 5105 || LOOKING_AT_NOCASE (bp, "defer") | |
| 5106 || LOOKING_AT_NOCASE (bp, "value") | |
| 5107 || LOOKING_AT_NOCASE (bp, "variable") | |
| 5108 || LOOKING_AT_NOCASE (bp, "buffer:") | |
| 5109 || LOOKING_AT_NOCASE (bp, "field")) | |
| 5110 get_tag (skip_spaces (bp), NULL); /* Yay! A definition! */ | |
| 5111 else | |
| 5112 bp = skip_non_spaces (bp); | |
| 5113 } | |
| 5114 | |
| 5115 | |
| 5116 /* | |
| 428 | 5117 * Scheme tag functions |
| 5118 * look for (def... xyzzy | |
| 2225 | 5119 * (def... (xyzzy |
| 5120 * (def ... ((...(xyzzy .... | |
| 5121 * (set! xyzzy | |
| 5122 * Original code by Ken Haase (1985?) | |
| 428 | 5123 */ |
| 442 | 5124 static void |
| 428 | 5125 Scheme_functions (inf) |
| 5126 FILE *inf; | |
| 5127 { | |
| 5128 register char *bp; | |
| 5129 | |
| 5130 LOOP_ON_INPUT_LINES (inf, lb, bp) | |
| 5131 { | |
| 2225 | 5132 if (strneq (bp, "(def", 4) || strneq (bp, "(DEF", 4)) |
| 428 | 5133 { |
| 2225 | 5134 bp = skip_non_spaces (bp+4); |
| 428 | 5135 /* Skip over open parens and white space */ |
| 2225 | 5136 while (notinname (*bp)) |
| 428 | 5137 bp++; |
| 2225 | 5138 get_tag (bp, NULL); |
| 428 | 5139 } |
| 709 | 5140 if (LOOKING_AT (bp, "(SET!") || LOOKING_AT (bp, "(set!")) |
| 2225 | 5141 get_tag (bp, NULL); |
| 428 | 5142 } |
| 5143 } | |
| 458 | 5144 |
| 428 | 5145 |
| 5146 /* Find tags in TeX and LaTeX input files. */ | |
| 5147 | |
| 5148 /* TEX_toktab is a table of TeX control sequences that define tags. | |
| 2225 | 5149 * Each entry records one such control sequence. |
| 5150 * | |
| 5151 * Original code from who knows whom. | |
| 5152 * Ideas by: | |
| 5153 * Stefan Monnier (2002) | |
| 5154 */ | |
| 5155 | |
| 5156 static linebuffer *TEX_toktab = NULL; /* Table with tag tokens */ | |
| 428 | 5157 |
| 5158 /* Default set of control sequences to put into TEX_toktab. | |
| 5159 The value of environment var TEXTAGS is prepended to this. */ | |
| 2225 | 5160 static char *TEX_defenv = "\ |
| 428 | 5161 :chapter:section:subsection:subsubsection:eqno:label:ref:cite:bibitem\ |
| 2225 | 5162 :part:appendix:entry:index:def\ |
| 5163 :newcommand:renewcommand:newenvironment:renewenvironment"; | |
| 428 | 5164 |
| 709 | 5165 static void TEX_mode __P((FILE *)); |
| 2225 | 5166 static void TEX_decode_env __P((char *, char *)); |
| 5167 | |
| 5168 static char TEX_esc = '\\'; | |
| 5169 static char TEX_opgrp = '{'; | |
| 5170 static char TEX_clgrp = '}'; | |
| 428 | 5171 |
| 5172 /* | |
| 5173 * TeX/LaTeX scanning loop. | |
| 5174 */ | |
| 442 | 5175 static void |
| 458 | 5176 TeX_commands (inf) |
| 428 | 5177 FILE *inf; |
| 5178 { | |
| 2225 | 5179 char *cp; |
| 5180 linebuffer *key; | |
| 428 | 5181 |
| 5182 /* Select either \ or ! as escape character. */ | |
| 5183 TEX_mode (inf); | |
| 5184 | |
| 5185 /* Initialize token table once from environment. */ | |
| 2225 | 5186 if (TEX_toktab == NULL) |
| 5187 TEX_decode_env ("TEXTAGS", TEX_defenv); | |
| 428 | 5188 |
| 5189 LOOP_ON_INPUT_LINES (inf, lb, cp) | |
| 5190 { | |
| 2225 | 5191 /* Look at each TEX keyword in line. */ |
| 5192 for (;;) | |
| 428 | 5193 { |
| 2225 | 5194 /* Look for a TEX escape. */ |
| 5195 while (*cp++ != TEX_esc) | |
| 5196 if (cp[-1] == '\0' || cp[-1] == '%') | |
| 5197 goto tex_next_line; | |
| 5198 | |
| 5199 for (key = TEX_toktab; key->buffer != NULL; key++) | |
| 5200 if (strneq (cp, key->buffer, key->len)) | |
| 5201 { | |
| 5202 register char *p; | |
| 5203 int namelen, linelen; | |
| 5204 bool opgrp = FALSE; | |
| 5205 | |
| 5206 cp = skip_spaces (cp + key->len); | |
| 5207 if (*cp == TEX_opgrp) | |
| 5208 { | |
| 5209 opgrp = TRUE; | |
| 5210 cp++; | |
| 5211 } | |
| 5212 for (p = cp; | |
| 5213 (!iswhite (*p) && *p != '#' && | |
| 5214 *p != TEX_opgrp && *p != TEX_clgrp); | |
| 5215 p++) | |
| 5216 continue; | |
| 5217 namelen = p - cp; | |
| 5218 linelen = lb.len; | |
| 5219 if (!opgrp || *p == TEX_clgrp) | |
| 5220 { | |
| 5221 while (*p != '\0' && *p != TEX_opgrp && *p != TEX_clgrp) | |
| 3876 | 5222 p++; |
| 2225 | 5223 linelen = p - lb.buffer + 1; |
| 5224 } | |
| 5225 make_tag (cp, namelen, TRUE, | |
| 5226 lb.buffer, linelen, lineno, linecharno); | |
| 5227 goto tex_next_line; /* We only tag a line once */ | |
| 5228 } | |
| 428 | 5229 } |
| 2225 | 5230 tex_next_line: |
| 5231 ; | |
| 428 | 5232 } |
| 5233 } | |
| 5234 | |
| 5235 #define TEX_LESC '\\' | |
| 5236 #define TEX_SESC '!' | |
| 5237 | |
| 5238 /* Figure out whether TeX's escapechar is '\\' or '!' and set grouping | |
| 5239 chars accordingly. */ | |
| 442 | 5240 static void |
| 428 | 5241 TEX_mode (inf) |
| 5242 FILE *inf; | |
| 5243 { | |
| 5244 int c; | |
| 5245 | |
| 5246 while ((c = getc (inf)) != EOF) | |
| 5247 { | |
| 5248 /* Skip to next line if we hit the TeX comment char. */ | |
| 2225 | 5249 if (c == '%') |
| 3876 | 5250 while (c != '\n' && c != EOF) |
| 428 | 5251 c = getc (inf); |
| 5252 else if (c == TEX_LESC || c == TEX_SESC ) | |
| 5253 break; | |
| 5254 } | |
| 5255 | |
| 5256 if (c == TEX_LESC) | |
| 5257 { | |
| 5258 TEX_esc = TEX_LESC; | |
| 5259 TEX_opgrp = '{'; | |
| 5260 TEX_clgrp = '}'; | |
| 5261 } | |
| 5262 else | |
| 5263 { | |
| 5264 TEX_esc = TEX_SESC; | |
| 5265 TEX_opgrp = '<'; | |
| 5266 TEX_clgrp = '>'; | |
| 5267 } | |
| 5268 /* If the input file is compressed, inf is a pipe, and rewind may fail. | |
| 5269 No attempt is made to correct the situation. */ | |
| 5270 rewind (inf); | |
| 5271 } | |
| 5272 | |
| 5273 /* Read environment and prepend it to the default string. | |
| 5274 Build token table. */ | |
| 2225 | 5275 static void |
| 428 | 5276 TEX_decode_env (evarname, defenv) |
| 5277 char *evarname; | |
| 5278 char *defenv; | |
| 5279 { | |
| 5280 register char *env, *p; | |
| 2225 | 5281 int i, len; |
| 428 | 5282 |
| 5283 /* Append default string to environment. */ | |
| 5284 env = getenv (evarname); | |
| 5285 if (!env) | |
| 5286 env = defenv; | |
| 5287 else | |
| 5288 { | |
| 5289 char *oldenv = env; | |
| 5290 env = concat (oldenv, defenv, ""); | |
| 5291 } | |
| 5292 | |
| 5293 /* Allocate a token table */ | |
| 2225 | 5294 for (len = 1, p = env; p;) |
| 428 | 5295 if ((p = etags_strchr (p, ':')) && *++p != '\0') |
| 2225 | 5296 len++; |
| 5297 TEX_toktab = xnew (len, linebuffer); | |
| 428 | 5298 |
| 5299 /* Unpack environment string into token table. Be careful about */ | |
| 5300 /* zero-length strings (leading ':', "::" and trailing ':') */ | |
| 2225 | 5301 for (i = 0; *env != '\0';) |
| 428 | 5302 { |
| 5303 p = etags_strchr (env, ':'); | |
| 5304 if (!p) /* End of environment string. */ | |
| 5305 p = env + strlen (env); | |
| 5306 if (p - env > 0) | |
| 5307 { /* Only non-zero strings. */ | |
| 2225 | 5308 TEX_toktab[i].buffer = savenstr (env, p - env); |
| 5309 TEX_toktab[i].len = p - env; | |
| 428 | 5310 i++; |
| 5311 } | |
| 5312 if (*p) | |
| 5313 env = p + 1; | |
| 5314 else | |
| 5315 { | |
| 2225 | 5316 TEX_toktab[i].buffer = NULL; /* Mark end of table. */ |
| 5317 TEX_toktab[i].len = 0; | |
| 428 | 5318 break; |
| 5319 } | |
| 5320 } | |
| 5321 } | |
| 458 | 5322 |
| 5323 | |
| 5324 /* Texinfo support. Dave Love, Mar. 2000. */ | |
| 5325 static void | |
| 5326 Texinfo_nodes (inf) | |
| 5327 FILE * inf; | |
| 5328 { | |
| 5329 char *cp, *start; | |
| 5330 LOOP_ON_INPUT_LINES (inf, lb, cp) | |
| 709 | 5331 if (LOOKING_AT (cp, "@node")) |
| 5332 { | |
| 5333 start = cp; | |
| 5334 while (*cp != '\0' && *cp != ',') | |
| 5335 cp++; | |
| 2225 | 5336 make_tag (start, cp - start, TRUE, |
| 5337 lb.buffer, cp - lb.buffer + 1, lineno, linecharno); | |
| 5338 } | |
| 5339 } | |
| 5340 | |
| 5341 | |
| 5342 /* | |
| 5343 * HTML support. | |
| 5344 * Contents of <title>, <h1>, <h2>, <h3> are tags. | |
| 5345 * Contents of <a name=xxx> are tags with name xxx. | |
| 5346 * | |
| 5347 * Francesco Potort́, 2002. | |
| 5348 */ | |
| 5349 static void | |
| 5350 HTML_labels (inf) | |
| 5351 FILE * inf; | |
| 5352 { | |
| 5353 bool getnext = FALSE; /* next text outside of HTML tags is a tag */ | |
| 5354 bool skiptag = FALSE; /* skip to the end of the current HTML tag */ | |
| 5355 bool intag = FALSE; /* inside an html tag, looking for ID= */ | |
| 5356 bool inanchor = FALSE; /* when INTAG, is an anchor, look for NAME= */ | |
| 5357 char *end; | |
| 5358 | |
| 5359 | |
| 5360 linebuffer_setlen (&token_name, 0); /* no name in buffer */ | |
| 5361 | |
| 5362 LOOP_ON_INPUT_LINES (inf, lb, dbp) | |
| 5363 for (;;) /* loop on the same line */ | |
| 5364 { | |
| 5365 if (skiptag) /* skip HTML tag */ | |
| 5366 { | |
| 5367 while (*dbp != '\0' && *dbp != '>') | |
| 5368 dbp++; | |
| 5369 if (*dbp == '>') | |
| 5370 { | |
| 5371 dbp += 1; | |
| 5372 skiptag = FALSE; | |
| 5373 continue; /* look on the same line */ | |
| 5374 } | |
| 5375 break; /* go to next line */ | |
| 5376 } | |
| 5377 | |
| 5378 else if (intag) /* look for "name=" or "id=" */ | |
| 5379 { | |
| 5380 while (*dbp != '\0' && *dbp != '>' | |
| 5381 && lowcase (*dbp) != 'n' && lowcase (*dbp) != 'i') | |
| 5382 dbp++; | |
| 5383 if (*dbp == '\0') | |
| 5384 break; /* go to next line */ | |
| 5385 if (*dbp == '>') | |
| 5386 { | |
| 5387 dbp += 1; | |
| 5388 intag = FALSE; | |
| 5389 continue; /* look on the same line */ | |
| 5390 } | |
| 5391 if ((inanchor && LOOKING_AT_NOCASE (dbp, "name=")) | |
| 5392 || LOOKING_AT_NOCASE (dbp, "id=")) | |
| 5393 { | |
| 5394 bool quoted = (dbp[0] == '"'); | |
| 5395 | |
| 5396 if (quoted) | |
| 5397 for (end = ++dbp; *end != '\0' && *end != '"'; end++) | |
| 5398 continue; | |
| 5399 else | |
| 5400 for (end = dbp; *end != '\0' && intoken (*end); end++) | |
| 5401 continue; | |
| 5402 linebuffer_setlen (&token_name, end - dbp); | |
| 5403 strncpy (token_name.buffer, dbp, end - dbp); | |
| 5404 token_name.buffer[end - dbp] = '\0'; | |
| 5405 | |
| 5406 dbp = end; | |
| 5407 intag = FALSE; /* we found what we looked for */ | |
| 5408 skiptag = TRUE; /* skip to the end of the tag */ | |
| 5409 getnext = TRUE; /* then grab the text */ | |
| 5410 continue; /* look on the same line */ | |
| 5411 } | |
| 5412 dbp += 1; | |
| 5413 } | |
| 5414 | |
| 5415 else if (getnext) /* grab next tokens and tag them */ | |
| 5416 { | |
| 5417 dbp = skip_spaces (dbp); | |
| 5418 if (*dbp == '\0') | |
| 5419 break; /* go to next line */ | |
| 5420 if (*dbp == '<') | |
| 5421 { | |
| 5422 intag = TRUE; | |
| 5423 inanchor = (lowcase (dbp[1]) == 'a' && !intoken (dbp[2])); | |
| 5424 continue; /* look on the same line */ | |
| 5425 } | |
| 5426 | |
| 5427 for (end = dbp + 1; *end != '\0' && *end != '<'; end++) | |
| 5428 continue; | |
| 5429 make_tag (token_name.buffer, token_name.len, TRUE, | |
| 5430 dbp, end - dbp, lineno, linecharno); | |
| 5431 linebuffer_setlen (&token_name, 0); /* no name in buffer */ | |
| 5432 getnext = FALSE; | |
| 5433 break; /* go to next line */ | |
| 5434 } | |
| 5435 | |
| 5436 else /* look for an interesting HTML tag */ | |
| 5437 { | |
| 5438 while (*dbp != '\0' && *dbp != '<') | |
| 5439 dbp++; | |
| 5440 if (*dbp == '\0') | |
| 5441 break; /* go to next line */ | |
| 5442 intag = TRUE; | |
| 5443 if (lowcase (dbp[1]) == 'a' && !intoken (dbp[2])) | |
| 5444 { | |
| 5445 inanchor = TRUE; | |
| 5446 continue; /* look on the same line */ | |
| 5447 } | |
| 5448 else if (LOOKING_AT_NOCASE (dbp, "<title>") | |
| 5449 || LOOKING_AT_NOCASE (dbp, "<h1>") | |
| 5450 || LOOKING_AT_NOCASE (dbp, "<h2>") | |
| 5451 || LOOKING_AT_NOCASE (dbp, "<h3>")) | |
| 5452 { | |
| 5453 intag = FALSE; | |
| 5454 getnext = TRUE; | |
| 5455 continue; /* look on the same line */ | |
| 5456 } | |
| 5457 dbp += 1; | |
| 5458 } | |
| 709 | 5459 } |
| 458 | 5460 } |
| 5461 | |
| 428 | 5462 |
| 5463 /* | |
| 2225 | 5464 * Prolog support |
| 428 | 5465 * |
| 709 | 5466 * Assumes that the predicate or rule starts at column 0. |
| 5467 * Only the first clause of a predicate or rule is added. | |
| 2225 | 5468 * Original code by Sunichirou Sugou (1989) |
| 5469 * Rewritten by Anders Lindgren (1996) | |
| 428 | 5470 */ |
| 709 | 5471 static int prolog_pr __P((char *, char *)); |
| 5472 static void prolog_skip_comment __P((linebuffer *, FILE *)); | |
| 5473 static int prolog_atom __P((char *, int)); | |
| 442 | 5474 |
| 5475 static void | |
| 428 | 5476 Prolog_functions (inf) |
| 5477 FILE *inf; | |
| 5478 { | |
| 5479 char *cp, *last; | |
| 5480 int len; | |
| 5481 int allocated; | |
| 5482 | |
| 5483 allocated = 0; | |
| 5484 len = 0; | |
| 5485 last = NULL; | |
| 5486 | |
| 5487 LOOP_ON_INPUT_LINES (inf, lb, cp) | |
| 5488 { | |
| 5489 if (cp[0] == '\0') /* Empty line */ | |
| 5490 continue; | |
| 442 | 5491 else if (iswhite (cp[0])) /* Not a predicate */ |
| 428 | 5492 continue; |
| 5493 else if (cp[0] == '/' && cp[1] == '*') /* comment. */ | |
| 5494 prolog_skip_comment (&lb, inf); | |
| 709 | 5495 else if ((len = prolog_pr (cp, last)) > 0) |
| 428 | 5496 { |
| 709 | 5497 /* Predicate or rule. Store the function name so that we |
| 5498 only generate a tag for the first clause. */ | |
| 428 | 5499 if (last == NULL) |
| 5500 last = xnew(len + 1, char); | |
| 5501 else if (len + 1 > allocated) | |
| 458 | 5502 xrnew (last, len + 1, char); |
| 428 | 5503 allocated = len + 1; |
| 5504 strncpy (last, cp, len); | |
| 5505 last[len] = '\0'; | |
| 5506 } | |
| 5507 } | |
| 3517 | 5508 if (last != NULL) |
| 5509 free (last); | |
| 428 | 5510 } |
| 5511 | |
| 5512 | |
| 442 | 5513 static void |
| 428 | 5514 prolog_skip_comment (plb, inf) |
| 5515 linebuffer *plb; | |
| 5516 FILE *inf; | |
| 5517 { | |
| 5518 char *cp; | |
| 5519 | |
| 5520 do | |
| 5521 { | |
| 5522 for (cp = plb->buffer; *cp != '\0'; cp++) | |
| 5523 if (cp[0] == '*' && cp[1] == '/') | |
| 5524 return; | |
| 2225 | 5525 readline (plb, inf); |
| 428 | 5526 } |
| 5527 while (!feof(inf)); | |
| 5528 } | |
| 5529 | |
| 5530 /* | |
| 709 | 5531 * A predicate or rule definition is added if it matches: |
| 428 | 5532 * <beginning of line><Prolog Atom><whitespace>( |
| 709 | 5533 * or <beginning of line><Prolog Atom><whitespace>:- |
| 428 | 5534 * |
| 5535 * It is added to the tags database if it doesn't match the | |
| 5536 * name of the previous clause header. | |
| 5537 * | |
| 709 | 5538 * Return the size of the name of the predicate or rule, or 0 if no |
| 5539 * header was found. | |
| 428 | 5540 */ |
| 442 | 5541 static int |
| 709 | 5542 prolog_pr (s, last) |
| 428 | 5543 char *s; |
| 5544 char *last; /* Name of last clause. */ | |
| 5545 { | |
| 5546 int pos; | |
| 5547 int len; | |
| 5548 | |
| 5549 pos = prolog_atom (s, 0); | |
| 5550 if (pos < 1) | |
| 5551 return 0; | |
| 5552 | |
| 5553 len = pos; | |
| 5554 pos = skip_spaces (s + pos) - s; | |
| 5555 | |
| 709 | 5556 if ((s[pos] == '.' |
| 5557 || (s[pos] == '(' && (pos += 1)) | |
| 5558 || (s[pos] == ':' && s[pos + 1] == '-' && (pos += 2))) | |
| 5559 && (last == NULL /* save only the first clause */ | |
| 2325 | 5560 || len != (int)strlen (last) |
| 709 | 5561 || !strneq (s, last, len))) |
| 428 | 5562 { |
| 2225 | 5563 make_tag (s, len, TRUE, s, pos, lineno, linecharno); |
| 428 | 5564 return len; |
| 5565 } | |
| 709 | 5566 else |
| 5567 return 0; | |
| 428 | 5568 } |
| 5569 | |
| 5570 /* | |
| 5571 * Consume a Prolog atom. | |
| 5572 * Return the number of bytes consumed, or -1 if there was an error. | |
| 5573 * | |
| 5574 * A prolog atom, in this context, could be one of: | |
| 5575 * - An alphanumeric sequence, starting with a lower case letter. | |
| 5576 * - A quoted arbitrary string. Single quotes can escape themselves. | |
| 5577 * Backslash quotes everything. | |
| 5578 */ | |
| 442 | 5579 static int |
| 428 | 5580 prolog_atom (s, pos) |
| 5581 char *s; | |
| 5582 int pos; | |
| 5583 { | |
| 5584 int origpos; | |
| 5585 | |
| 5586 origpos = pos; | |
| 5587 | |
| 458 | 5588 if (ISLOWER(s[pos]) || (s[pos] == '_')) |
| 428 | 5589 { |
| 5590 /* The atom is unquoted. */ | |
| 5591 pos++; | |
| 458 | 5592 while (ISALNUM(s[pos]) || (s[pos] == '_')) |
| 428 | 5593 { |
| 5594 pos++; | |
| 5595 } | |
| 5596 return pos - origpos; | |
| 5597 } | |
| 5598 else if (s[pos] == '\'') | |
| 5599 { | |
| 5600 pos++; | |
| 5601 | |
| 2225 | 5602 for (;;) |
| 428 | 5603 { |
| 5604 if (s[pos] == '\'') | |
| 5605 { | |
| 5606 pos++; | |
| 5607 if (s[pos] != '\'') | |
| 5608 break; | |
| 5609 pos++; /* A double quote */ | |
| 5610 } | |
| 5611 else if (s[pos] == '\0') | |
| 5612 /* Multiline quoted atoms are ignored. */ | |
| 5613 return -1; | |
| 5614 else if (s[pos] == '\\') | |
| 5615 { | |
| 5616 if (s[pos+1] == '\0') | |
| 5617 return -1; | |
| 5618 pos += 2; | |
| 5619 } | |
| 5620 else | |
| 5621 pos++; | |
| 5622 } | |
| 5623 return pos - origpos; | |
| 5624 } | |
| 5625 else | |
| 5626 return -1; | |
| 5627 } | |
| 458 | 5628 |
| 428 | 5629 |
| 5630 /* | |
| 2225 | 5631 * Support for Erlang |
| 428 | 5632 * |
| 5633 * Generates tags for functions, defines, and records. | |
| 5634 * Assumes that Erlang functions start at column 0. | |
| 2225 | 5635 * Original code by Anders Lindgren (1996) |
| 428 | 5636 */ |
| 709 | 5637 static int erlang_func __P((char *, char *)); |
| 5638 static void erlang_attribute __P((char *)); | |
| 2225 | 5639 static int erlang_atom __P((char *)); |
| 442 | 5640 |
| 5641 static void | |
| 428 | 5642 Erlang_functions (inf) |
| 5643 FILE *inf; | |
| 5644 { | |
| 5645 char *cp, *last; | |
| 5646 int len; | |
| 5647 int allocated; | |
| 5648 | |
| 5649 allocated = 0; | |
| 5650 len = 0; | |
| 5651 last = NULL; | |
| 5652 | |
| 5653 LOOP_ON_INPUT_LINES (inf, lb, cp) | |
| 5654 { | |
| 5655 if (cp[0] == '\0') /* Empty line */ | |
| 5656 continue; | |
| 442 | 5657 else if (iswhite (cp[0])) /* Not function nor attribute */ |
| 428 | 5658 continue; |
| 5659 else if (cp[0] == '%') /* comment */ | |
| 5660 continue; | |
| 5661 else if (cp[0] == '"') /* Sometimes, strings start in column one */ | |
| 5662 continue; | |
| 5663 else if (cp[0] == '-') /* attribute, e.g. "-define" */ | |
| 5664 { | |
| 5665 erlang_attribute (cp); | |
| 3517 | 5666 if (last != NULL) |
| 5667 { | |
| 5668 free (last); | |
| 5669 last = NULL; | |
| 5670 } | |
| 428 | 5671 } |
| 5672 else if ((len = erlang_func (cp, last)) > 0) | |
| 5673 { | |
| 5674 /* | |
| 5675 * Function. Store the function name so that we only | |
| 5676 * generates a tag for the first clause. | |
| 5677 */ | |
| 5678 if (last == NULL) | |
| 5679 last = xnew (len + 1, char); | |
| 5680 else if (len + 1 > allocated) | |
| 458 | 5681 xrnew (last, len + 1, char); |
| 428 | 5682 allocated = len + 1; |
| 5683 strncpy (last, cp, len); | |
| 5684 last[len] = '\0'; | |
| 5685 } | |
| 5686 } | |
| 3517 | 5687 if (last != NULL) |
| 5688 free (last); | |
| 428 | 5689 } |
| 5690 | |
| 5691 | |
| 5692 /* | |
| 5693 * A function definition is added if it matches: | |
| 5694 * <beginning of line><Erlang Atom><whitespace>( | |
| 5695 * | |
| 5696 * It is added to the tags database if it doesn't match the | |
| 5697 * name of the previous clause header. | |
| 5698 * | |
| 5699 * Return the size of the name of the function, or 0 if no function | |
| 5700 * was found. | |
| 5701 */ | |
| 442 | 5702 static int |
| 428 | 5703 erlang_func (s, last) |
| 5704 char *s; | |
| 5705 char *last; /* Name of last clause. */ | |
| 5706 { | |
| 5707 int pos; | |
| 5708 int len; | |
| 5709 | |
| 2225 | 5710 pos = erlang_atom (s); |
| 428 | 5711 if (pos < 1) |
| 5712 return 0; | |
| 5713 | |
| 5714 len = pos; | |
| 5715 pos = skip_spaces (s + pos) - s; | |
| 5716 | |
| 5717 /* Save only the first clause. */ | |
| 5718 if (s[pos++] == '(' | |
| 5719 && (last == NULL | |
| 5720 || len != (int)strlen (last) | |
| 5721 || !strneq (s, last, len))) | |
| 5722 { | |
| 2225 | 5723 make_tag (s, len, TRUE, s, pos, lineno, linecharno); |
| 428 | 5724 return len; |
| 5725 } | |
| 5726 | |
| 5727 return 0; | |
| 5728 } | |
| 5729 | |
| 5730 | |
| 5731 /* | |
| 5732 * Handle attributes. Currently, tags are generated for defines | |
| 5733 * and records. | |
| 5734 * | |
| 5735 * They are on the form: | |
| 5736 * -define(foo, bar). | |
| 5737 * -define(Foo(M, N), M+N). | |
| 5738 * -record(graph, {vtab = notable, cyclic = true}). | |
| 5739 */ | |
| 442 | 5740 static void |
| 428 | 5741 erlang_attribute (s) |
| 5742 char *s; | |
| 5743 { | |
| 2225 | 5744 char *cp = s; |
| 5745 | |
| 5746 if ((LOOKING_AT (cp, "-define") || LOOKING_AT (cp, "-record")) | |
| 5747 && *cp++ == '(') | |
| 428 | 5748 { |
| 2225 | 5749 int len = erlang_atom (skip_spaces (cp)); |
| 5750 if (len > 0) | |
| 5751 make_tag (cp, len, TRUE, s, cp + len - s, lineno, linecharno); | |
| 428 | 5752 } |
| 5753 return; | |
| 5754 } | |
| 5755 | |
| 5756 | |
| 5757 /* | |
| 5758 * Consume an Erlang atom (or variable). | |
| 5759 * Return the number of bytes consumed, or -1 if there was an error. | |
| 5760 */ | |
| 442 | 5761 static int |
| 2225 | 5762 erlang_atom (s) |
| 428 | 5763 char *s; |
| 5764 { | |
| 2225 | 5765 int pos = 0; |
| 428 | 5766 |
| 458 | 5767 if (ISALPHA (s[pos]) || s[pos] == '_') |
| 428 | 5768 { |
| 5769 /* The atom is unquoted. */ | |
| 2225 | 5770 do |
| 428 | 5771 pos++; |
| 2225 | 5772 while (ISALNUM (s[pos]) || s[pos] == '_'); |
| 428 | 5773 } |
| 5774 else if (s[pos] == '\'') | |
| 5775 { | |
| 2225 | 5776 for (pos++; s[pos] != '\''; pos++) |
| 5777 if (s[pos] == '\0' /* multiline quoted atoms are ignored */ | |
| 5778 || (s[pos] == '\\' && s[++pos] == '\0')) | |
| 5779 return 0; | |
| 428 | 5780 pos++; |
| 5781 } | |
| 2225 | 5782 |
| 5783 return pos; | |
| 428 | 5784 } |
| 458 | 5785 |
| 428 | 5786 |
| 709 | 5787 static char *scan_separators __P((char *)); |
| 2225 | 5788 static void add_regex __P((char *, language *)); |
| 709 | 5789 static char *substitute __P((char *, char *, struct re_registers *)); |
| 442 | 5790 |
| 2225 | 5791 /* |
| 5792 * Take a string like "/blah/" and turn it into "blah", verifying | |
| 5793 * that the first and last characters are the same, and handling | |
| 5794 * quoted separator characters. Actually, stops on the occurrence of | |
| 5795 * an unquoted separator. Also process \t, \n, etc. and turn into | |
| 5796 * appropriate characters. Works in place. Null terminates name string. | |
| 5797 * Returns pointer to terminating separator, or NULL for | |
| 5798 * unterminated regexps. | |
| 5799 */ | |
| 428 | 5800 static char * |
| 5801 scan_separators (name) | |
| 5802 char *name; | |
| 5803 { | |
| 5804 char sep = name[0]; | |
| 5805 char *copyto = name; | |
| 5806 bool quoted = FALSE; | |
| 5807 | |
| 5808 for (++name; *name != '\0'; ++name) | |
| 5809 { | |
| 5810 if (quoted) | |
| 5811 { | |
| 2225 | 5812 switch (*name) |
| 428 | 5813 { |
| 2225 | 5814 case 'a': *copyto++ = '\007'; break; /* BEL (bell) */ |
| 5815 case 'b': *copyto++ = '\b'; break; /* BS (back space) */ | |
| 5816 case 'd': *copyto++ = 0177; break; /* DEL (delete) */ | |
| 5817 case 'e': *copyto++ = 033; break; /* ESC (delete) */ | |
| 5818 case 'f': *copyto++ = '\f'; break; /* FF (form feed) */ | |
| 5819 case 'n': *copyto++ = '\n'; break; /* NL (new line) */ | |
| 5820 case 'r': *copyto++ = '\r'; break; /* CR (carriage return) */ | |
| 5821 case 't': *copyto++ = '\t'; break; /* TAB (horizontal tab) */ | |
| 5822 case 'v': *copyto++ = '\v'; break; /* VT (vertical tab) */ | |
| 5823 default: | |
| 5824 if (*name == sep) | |
| 5825 *copyto++ = sep; | |
| 5826 else | |
| 5827 { | |
| 5828 /* Something else is quoted, so preserve the quote. */ | |
| 5829 *copyto++ = '\\'; | |
| 5830 *copyto++ = *name; | |
| 5831 } | |
| 5832 break; | |
| 428 | 5833 } |
| 5834 quoted = FALSE; | |
| 5835 } | |
| 5836 else if (*name == '\\') | |
| 5837 quoted = TRUE; | |
| 5838 else if (*name == sep) | |
| 5839 break; | |
| 5840 else | |
| 5841 *copyto++ = *name; | |
| 5842 } | |
| 2225 | 5843 if (*name != sep) |
| 5844 name = NULL; /* signal unterminated regexp */ | |
| 428 | 5845 |
| 5846 /* Terminate copied string. */ | |
| 5847 *copyto = '\0'; | |
| 5848 return name; | |
| 5849 } | |
| 5850 | |
| 5851 /* Look at the argument of --regex or --no-regex and do the right | |
| 5852 thing. Same for each line of a regexp file. */ | |
| 442 | 5853 static void |
| 2225 | 5854 analyse_regex (regex_arg) |
| 428 | 5855 char *regex_arg; |
| 5856 { | |
| 5857 if (regex_arg == NULL) | |
| 709 | 5858 { |
| 2225 | 5859 free_regexps (); /* --no-regex: remove existing regexps */ |
| 709 | 5860 return; |
| 5861 } | |
| 428 | 5862 |
| 5863 /* A real --regexp option or a line in a regexp file. */ | |
| 5864 switch (regex_arg[0]) | |
| 5865 { | |
| 5866 /* Comments in regexp file or null arg to --regex. */ | |
| 5867 case '\0': | |
| 5868 case ' ': | |
| 5869 case '\t': | |
| 5870 break; | |
| 5871 | |
| 5872 /* Read a regex file. This is recursive and may result in a | |
| 5873 loop, which will stop when the file descriptors are exhausted. */ | |
| 5874 case '@': | |
| 5875 { | |
| 5876 FILE *regexfp; | |
| 5877 linebuffer regexbuf; | |
| 5878 char *regexfile = regex_arg + 1; | |
| 5879 | |
| 5880 /* regexfile is a file containing regexps, one per line. */ | |
| 5881 regexfp = fopen (regexfile, "r"); | |
| 5882 if (regexfp == NULL) | |
| 5883 { | |
| 5884 pfatal (regexfile); | |
| 5885 return; | |
| 5886 } | |
| 2225 | 5887 linebuffer_init (®exbuf); |
| 428 | 5888 while (readline_internal (®exbuf, regexfp) > 0) |
| 2225 | 5889 analyse_regex (regexbuf.buffer); |
| 428 | 5890 free (regexbuf.buffer); |
| 5891 fclose (regexfp); | |
| 5892 } | |
| 5893 break; | |
| 5894 | |
| 5895 /* Regexp to be used for a specific language only. */ | |
| 5896 case '{': | |
| 5897 { | |
| 5898 language *lang; | |
| 5899 char *lang_name = regex_arg + 1; | |
| 5900 char *cp; | |
| 5901 | |
| 5902 for (cp = lang_name; *cp != '}'; cp++) | |
| 5903 if (*cp == '\0') | |
| 5904 { | |
| 5905 error ("unterminated language name in regex: %s", regex_arg); | |
| 5906 return; | |
| 5907 } | |
| 2225 | 5908 *cp++ = '\0'; |
| 458 | 5909 lang = get_language_from_langname (lang_name); |
| 428 | 5910 if (lang == NULL) |
| 5911 return; | |
| 2225 | 5912 add_regex (cp, lang); |
| 428 | 5913 } |
| 5914 break; | |
| 5915 | |
| 5916 /* Regexp to be used for any language. */ | |
| 5917 default: | |
| 2225 | 5918 add_regex (regex_arg, NULL); |
| 428 | 5919 break; |
| 5920 } | |
| 5921 } | |
| 5922 | |
| 2225 | 5923 /* Separate the regexp pattern, compile it, |
| 5924 and care for optional name and modifiers. */ | |
| 442 | 5925 static void |
| 2225 | 5926 add_regex (regexp_pattern, lang) |
| 428 | 5927 char *regexp_pattern; |
| 5928 language *lang; | |
| 5929 { | |
| 531 | 5930 static struct re_pattern_buffer zeropattern; |
| 2225 | 5931 char sep, *pat, *name, *modifiers; |
| 428 | 5932 const char *err; |
| 5933 struct re_pattern_buffer *patbuf; | |
| 2225 | 5934 regexp *rp; |
| 5935 bool | |
| 5936 force_explicit_name = TRUE, /* do not use implicit tag names */ | |
| 5937 ignore_case = FALSE, /* case is significant */ | |
| 5938 multi_line = FALSE, /* matches are done one line at a time */ | |
| 5939 single_line = FALSE; /* dot does not match newline */ | |
| 5940 | |
| 5941 | |
| 5942 if (strlen(regexp_pattern) < 3) | |
| 5943 { | |
| 5944 error ("null regexp", (char *)NULL); | |
| 5945 return; | |
| 5946 } | |
| 5947 sep = regexp_pattern[0]; | |
| 5948 name = scan_separators (regexp_pattern); | |
| 5949 if (name == NULL) | |
| 428 | 5950 { |
| 5951 error ("%s: unterminated regexp", regexp_pattern); | |
| 5952 return; | |
| 5953 } | |
| 2225 | 5954 if (name[1] == sep) |
| 428 | 5955 { |
| 2225 | 5956 error ("null name for regexp \"%s\"", regexp_pattern); |
| 428 | 5957 return; |
| 5958 } | |
| 2225 | 5959 modifiers = scan_separators (name); |
| 5960 if (modifiers == NULL) /* no terminating separator --> no name */ | |
| 5961 { | |
| 5962 modifiers = name; | |
| 5963 name = ""; | |
| 5964 } | |
| 5965 else | |
| 5966 modifiers += 1; /* skip separator */ | |
| 5967 | |
| 5968 /* Parse regex modifiers. */ | |
| 5969 for (; modifiers[0] != '\0'; modifiers++) | |
| 5970 switch (modifiers[0]) | |
| 5971 { | |
| 5972 case 'N': | |
| 5973 if (modifiers == name) | |
| 5974 error ("forcing explicit tag name but no name, ignoring", NULL); | |
| 5975 force_explicit_name = TRUE; | |
| 5976 break; | |
| 5977 case 'i': | |
| 5978 ignore_case = TRUE; | |
| 5979 break; | |
| 5980 case 's': | |
| 5981 single_line = TRUE; | |
| 5982 /* FALLTHRU */ | |
| 5983 case 'm': | |
| 5984 multi_line = TRUE; | |
| 5985 need_filebuf = TRUE; | |
| 5986 break; | |
| 5987 default: | |
| 5988 { | |
| 5989 char wrongmod [2]; | |
| 5990 wrongmod[0] = modifiers[0]; | |
| 5991 wrongmod[1] = '\0'; | |
| 5992 error ("invalid regexp modifier `%s', ignoring", wrongmod); | |
| 5993 } | |
| 5994 break; | |
| 5995 } | |
| 428 | 5996 |
| 5997 patbuf = xnew (1, struct re_pattern_buffer); | |
| 531 | 5998 *patbuf = zeropattern; |
| 5999 if (ignore_case) | |
| 2225 | 6000 { |
| 6001 static char lc_trans[CHARS]; | |
| 6002 int i; | |
| 6003 for (i = 0; i < CHARS; i++) | |
| 6004 lc_trans[i] = lowcase (i); | |
| 6005 patbuf->translate = lc_trans; /* translation table to fold case */ | |
| 6006 } | |
| 6007 | |
| 6008 if (multi_line) | |
| 6009 pat = concat ("^", regexp_pattern, ""); /* anchor to beginning of line */ | |
| 6010 else | |
| 6011 pat = regexp_pattern; | |
| 6012 | |
| 6013 if (single_line) | |
| 6014 re_set_syntax (RE_SYNTAX_EMACS | RE_DOT_NEWLINE); | |
| 6015 else | |
| 6016 re_set_syntax (RE_SYNTAX_EMACS); | |
| 6017 | |
| 6018 err = re_compile_pattern (pat, strlen (regexp_pattern), patbuf); | |
| 6019 if (multi_line) | |
| 6020 free (pat); | |
| 428 | 6021 if (err != NULL) |
| 6022 { | |
| 6023 error ("%s while compiling pattern", err); | |
| 6024 return; | |
| 6025 } | |
| 6026 | |
| 2225 | 6027 rp = p_head; |
| 6028 p_head = xnew (1, regexp); | |
| 6029 p_head->pattern = savestr (regexp_pattern); | |
| 6030 p_head->p_next = rp; | |
| 709 | 6031 p_head->lang = lang; |
| 6032 p_head->pat = patbuf; | |
| 2225 | 6033 p_head->name = savestr (name); |
| 428 | 6034 p_head->error_signaled = FALSE; |
| 2225 | 6035 p_head->force_explicit_name = force_explicit_name; |
| 6036 p_head->ignore_case = ignore_case; | |
| 6037 p_head->multi_line = multi_line; | |
| 428 | 6038 } |
| 6039 | |
| 6040 /* | |
| 6041 * Do the substitutions indicated by the regular expression and | |
| 6042 * arguments. | |
| 6043 */ | |
| 6044 static char * | |
| 6045 substitute (in, out, regs) | |
| 6046 char *in, *out; | |
| 6047 struct re_registers *regs; | |
| 6048 { | |
| 6049 char *result, *t; | |
| 6050 int size, dig, diglen; | |
| 6051 | |
| 6052 result = NULL; | |
| 6053 size = strlen (out); | |
| 6054 | |
| 6055 /* Pass 1: figure out how much to allocate by finding all \N strings. */ | |
| 6056 if (out[size - 1] == '\\') | |
| 6057 fatal ("pattern error in \"%s\"", out); | |
| 6058 for (t = etags_strchr (out, '\\'); | |
| 6059 t != NULL; | |
| 6060 t = etags_strchr (t + 2, '\\')) | |
| 458 | 6061 if (ISDIGIT (t[1])) |
| 428 | 6062 { |
| 6063 dig = t[1] - '0'; | |
| 6064 diglen = regs->end[dig] - regs->start[dig]; | |
| 6065 size += diglen - 2; | |
| 6066 } | |
| 6067 else | |
| 6068 size -= 1; | |
| 6069 | |
| 6070 /* Allocate space and do the substitutions. */ | |
| 2225 | 6071 assert (size >= 0); |
| 428 | 6072 result = xnew (size + 1, char); |
| 6073 | |
| 6074 for (t = result; *out != '\0'; out++) | |
| 458 | 6075 if (*out == '\\' && ISDIGIT (*++out)) |
| 428 | 6076 { |
| 6077 dig = *out - '0'; | |
| 6078 diglen = regs->end[dig] - regs->start[dig]; | |
| 6079 strncpy (t, in + regs->start[dig], diglen); | |
| 6080 t += diglen; | |
| 6081 } | |
| 6082 else | |
| 6083 *t++ = *out; | |
| 6084 *t = '\0'; | |
| 6085 | |
| 2225 | 6086 assert (t <= result + size); |
| 6087 assert (t - result == (int)strlen (result)); | |
| 428 | 6088 |
| 6089 return result; | |
| 6090 } | |
| 6091 | |
| 2225 | 6092 /* Deallocate all regexps. */ |
| 442 | 6093 static void |
| 2225 | 6094 free_regexps () |
| 428 | 6095 { |
| 2225 | 6096 regexp *rp; |
| 428 | 6097 while (p_head != NULL) |
| 6098 { | |
| 2225 | 6099 rp = p_head->p_next; |
| 6100 free (p_head->pattern); | |
| 6101 free (p_head->name); | |
| 428 | 6102 free (p_head); |
| 2225 | 6103 p_head = rp; |
| 428 | 6104 } |
| 6105 return; | |
| 6106 } | |
| 2225 | 6107 |
| 6108 /* | |
| 6109 * Reads the whole file as a single string from `filebuf' and looks for | |
| 6110 * multi-line regular expressions, creating tags on matches. | |
| 6111 * readline already dealt with normal regexps. | |
| 6112 * | |
| 6113 * Idea by Ben Wing <ben@666.com> (2002). | |
| 6114 */ | |
| 6115 static void | |
| 6116 regex_tag_multiline () | |
| 6117 { | |
| 6118 char *buffer = filebuf.buffer; | |
| 6119 regexp *rp; | |
| 6120 char *name; | |
| 6121 | |
| 6122 for (rp = p_head; rp != NULL; rp = rp->p_next) | |
| 6123 { | |
| 6124 int match = 0; | |
| 6125 | |
| 6126 if (!rp->multi_line) | |
| 6127 continue; /* skip normal regexps */ | |
| 6128 | |
| 6129 /* Generic initialisations before parsing file from memory. */ | |
| 6130 lineno = 1; /* reset global line number */ | |
| 6131 charno = 0; /* reset global char number */ | |
| 6132 linecharno = 0; /* reset global char number of line start */ | |
| 6133 | |
| 6134 /* Only use generic regexps or those for the current language. */ | |
| 6135 if (rp->lang != NULL && rp->lang != curfdp->lang) | |
| 6136 continue; | |
| 6137 | |
| 6138 while (match >= 0 && match < filebuf.len) | |
| 6139 { | |
| 6140 match = re_search (rp->pat, buffer, filebuf.len, charno, | |
| 6141 filebuf.len - match, &rp->regs); | |
| 6142 switch (match) | |
| 6143 { | |
| 6144 case -2: | |
| 6145 /* Some error. */ | |
| 6146 if (!rp->error_signaled) | |
| 6147 { | |
| 6148 error ("regexp stack overflow while matching \"%s\"", | |
| 6149 rp->pattern); | |
| 6150 rp->error_signaled = TRUE; | |
| 6151 } | |
| 6152 break; | |
| 6153 case -1: | |
| 6154 /* No match. */ | |
| 6155 break; | |
| 6156 default: | |
| 6157 if (match == rp->regs.end[0]) | |
| 6158 { | |
| 6159 if (!rp->error_signaled) | |
| 6160 { | |
| 6161 error ("regexp matches the empty string: \"%s\"", | |
| 6162 rp->pattern); | |
| 6163 rp->error_signaled = TRUE; | |
| 6164 } | |
| 6165 match = -3; /* exit from while loop */ | |
| 6166 break; | |
| 6167 } | |
| 6168 | |
| 6169 /* Match occurred. Construct a tag. */ | |
| 6170 while (charno < rp->regs.end[0]) | |
| 6171 if (buffer[charno++] == '\n') | |
| 6172 lineno++, linecharno = charno; | |
| 6173 name = rp->name; | |
| 6174 if (name[0] == '\0') | |
| 6175 name = NULL; | |
| 6176 else /* make a named tag */ | |
| 6177 name = substitute (buffer, rp->name, &rp->regs); | |
| 6178 if (rp->force_explicit_name) | |
| 6179 /* Force explicit tag name, if a name is there. */ | |
| 6180 pfnote (name, TRUE, buffer + linecharno, | |
| 6181 charno - linecharno + 1, lineno, linecharno); | |
| 6182 else | |
| 6183 make_tag (name, strlen (name), TRUE, buffer + linecharno, | |
| 6184 charno - linecharno + 1, lineno, linecharno); | |
| 6185 break; | |
| 6186 } | |
| 6187 } | |
| 6188 } | |
| 6189 } | |
| 6190 | |
| 428 | 6191 |
| 2225 | 6192 static bool |
| 6193 nocase_tail (cp) | |
| 6194 char *cp; | |
| 428 | 6195 { |
| 2225 | 6196 register int len = 0; |
| 6197 | |
| 6198 while (*cp != '\0' && lowcase (*cp) == lowcase (dbp[len])) | |
| 6199 cp++, len++; | |
| 6200 if (*cp == '\0' && !intoken (dbp[len])) | |
| 6201 { | |
| 6202 dbp += len; | |
| 6203 return TRUE; | |
| 6204 } | |
| 6205 return FALSE; | |
| 428 | 6206 } |
| 6207 | |
| 442 | 6208 static void |
| 2225 | 6209 get_tag (bp, namepp) |
| 6210 register char *bp; | |
| 6211 char **namepp; | |
| 428 | 6212 { |
| 2225 | 6213 register char *cp = bp; |
| 6214 | |
| 6215 if (*bp != '\0') | |
| 6216 { | |
| 6217 /* Go till you get to white space or a syntactic break */ | |
| 6218 for (cp = bp + 1; !notinname (*cp); cp++) | |
| 6219 continue; | |
| 6220 make_tag (bp, cp - bp, TRUE, | |
| 6221 lb.buffer, cp - lb.buffer + 1, lineno, linecharno); | |
| 6222 } | |
| 6223 | |
| 6224 if (namepp != NULL) | |
| 6225 *namepp = savenstr (bp, cp - bp); | |
| 428 | 6226 } |
| 6227 | |
| 6228 /* | |
| 6229 * Read a line of text from `stream' into `lbp', excluding the | |
| 6230 * newline or CR-NL, if any. Return the number of characters read from | |
| 6231 * `stream', which is the length of the line including the newline. | |
| 6232 * | |
| 2225 | 6233 * On DOS or Windows we do not count the CR character, if any before the |
| 6234 * NL, in the returned length; this mirrors the behavior of Emacs on those | |
| 428 | 6235 * platforms (for text files, it translates CR-NL to NL as it reads in the |
| 6236 * file). | |
| 2225 | 6237 * |
| 6238 * If multi-line regular expressions are requested, each line read is | |
| 6239 * appended to `filebuf'. | |
| 428 | 6240 */ |
| 442 | 6241 static long |
| 428 | 6242 readline_internal (lbp, stream) |
| 6243 linebuffer *lbp; | |
| 6244 register FILE *stream; | |
| 6245 { | |
| 6246 char *buffer = lbp->buffer; | |
| 6247 register char *p = lbp->buffer; | |
| 6248 register char *pend; | |
| 6249 int chars_deleted; | |
| 6250 | |
| 6251 pend = p + lbp->size; /* Separate to avoid 386/IX compiler bug. */ | |
| 6252 | |
| 2225 | 6253 for (;;) |
| 428 | 6254 { |
| 6255 register int c = getc (stream); | |
| 6256 if (p == pend) | |
| 6257 { | |
| 6258 /* We're at the end of linebuffer: expand it. */ | |
| 6259 lbp->size *= 2; | |
| 458 | 6260 xrnew (buffer, lbp->size, char); |
| 428 | 6261 p += buffer - lbp->buffer; |
| 6262 pend = buffer + lbp->size; | |
| 6263 lbp->buffer = buffer; | |
| 6264 } | |
| 6265 if (c == EOF) | |
| 6266 { | |
| 6267 *p = '\0'; | |
| 6268 chars_deleted = 0; | |
| 6269 break; | |
| 6270 } | |
| 6271 if (c == '\n') | |
| 6272 { | |
| 6273 if (p > buffer && p[-1] == '\r') | |
| 6274 { | |
| 6275 p -= 1; | |
| 458 | 6276 #ifdef DOS_NT |
| 428 | 6277 /* Assume CRLF->LF translation will be performed by Emacs |
| 6278 when loading this file, so CRs won't appear in the buffer. | |
| 6279 It would be cleaner to compensate within Emacs; | |
| 6280 however, Emacs does not know how many CRs were deleted | |
| 6281 before any given point in the file. */ | |
| 6282 chars_deleted = 1; | |
| 6283 #else | |
| 6284 chars_deleted = 2; | |
| 6285 #endif | |
| 6286 } | |
| 6287 else | |
| 6288 { | |
| 6289 chars_deleted = 1; | |
| 6290 } | |
| 6291 *p = '\0'; | |
| 6292 break; | |
| 6293 } | |
| 6294 *p++ = c; | |
| 6295 } | |
| 6296 lbp->len = p - buffer; | |
| 6297 | |
| 2225 | 6298 if (need_filebuf /* we need filebuf for multi-line regexps */ |
| 6299 && chars_deleted > 0) /* not at EOF */ | |
| 6300 { | |
| 6301 while (filebuf.size <= filebuf.len + lbp->len + 1) /* +1 for \n */ | |
| 6302 { | |
| 6303 /* Expand filebuf. */ | |
| 6304 filebuf.size *= 2; | |
| 6305 xrnew (filebuf.buffer, filebuf.size, char); | |
| 6306 } | |
| 6307 strncpy (filebuf.buffer + filebuf.len, lbp->buffer, lbp->len); | |
| 6308 filebuf.len += lbp->len; | |
| 6309 filebuf.buffer[filebuf.len++] = '\n'; | |
| 6310 filebuf.buffer[filebuf.len] = '\0'; | |
| 6311 } | |
| 6312 | |
| 428 | 6313 return lbp->len + chars_deleted; |
| 6314 } | |
| 6315 | |
| 6316 /* | |
| 6317 * Like readline_internal, above, but in addition try to match the | |
| 2225 | 6318 * input line against relevant regular expressions and manage #line |
| 6319 * directives. | |
| 428 | 6320 */ |
| 2225 | 6321 static void |
| 428 | 6322 readline (lbp, stream) |
| 6323 linebuffer *lbp; | |
| 6324 FILE *stream; | |
| 6325 { | |
| 2225 | 6326 long result; |
| 6327 | |
| 6328 linecharno = charno; /* update global char number of line start */ | |
| 6329 result = readline_internal (lbp, stream); /* read line */ | |
| 6330 lineno += 1; /* increment global line number */ | |
| 6331 charno += result; /* increment global char number */ | |
| 6332 | |
| 6333 /* Honour #line directives. */ | |
| 6334 if (!no_line_directive) | |
| 6335 { | |
| 6336 static bool discard_until_line_directive; | |
| 6337 | |
| 6338 /* Check whether this is a #line directive. */ | |
| 6339 if (result > 12 && strneq (lbp->buffer, "#line ", 6)) | |
| 6340 { | |
| 3876 | 6341 unsigned int lno; |
| 6342 int start = 0; | |
| 6343 | |
| 6344 if (sscanf (lbp->buffer, "#line %u \"%n", &lno, &start) >= 1 | |
| 6345 && start > 0) /* double quote character found */ | |
| 2225 | 6346 { |
| 6347 char *endp = lbp->buffer + start; | |
| 6348 | |
| 6349 while ((endp = etags_strchr (endp, '"')) != NULL | |
| 6350 && endp[-1] == '\\') | |
| 6351 endp++; | |
| 6352 if (endp != NULL) | |
| 6353 /* Ok, this is a real #line directive. Let's deal with it. */ | |
| 6354 { | |
| 6355 char *taggedabsname; /* absolute name of original file */ | |
| 6356 char *taggedfname; /* name of original file as given */ | |
| 6357 char *name; /* temp var */ | |
| 6358 | |
| 6359 discard_until_line_directive = FALSE; /* found it */ | |
| 6360 name = lbp->buffer + start; | |
| 6361 *endp = '\0'; | |
| 6362 canonicalize_filename (name); /* for DOS */ | |
| 3876 | 6363 taggedabsname = absolute_filename (name, tagfiledir); |
| 2225 | 6364 if (filename_is_absolute (name) |
| 6365 || filename_is_absolute (curfdp->infname)) | |
| 6366 taggedfname = savestr (taggedabsname); | |
| 6367 else | |
| 6368 taggedfname = relative_filename (taggedabsname,tagfiledir); | |
| 6369 | |
| 6370 if (streq (curfdp->taggedfname, taggedfname)) | |
| 6371 /* The #line directive is only a line number change. We | |
| 6372 deal with this afterwards. */ | |
| 6373 free (taggedfname); | |
| 6374 else | |
| 6375 /* The tags following this #line directive should be | |
| 6376 attributed to taggedfname. In order to do this, set | |
| 6377 curfdp accordingly. */ | |
| 6378 { | |
| 6379 fdesc *fdp; /* file description pointer */ | |
| 6380 | |
| 6381 /* Go look for a file description already set up for the | |
| 6382 file indicated in the #line directive. If there is | |
| 6383 one, use it from now until the next #line | |
| 6384 directive. */ | |
| 6385 for (fdp = fdhead; fdp != NULL; fdp = fdp->next) | |
| 6386 if (streq (fdp->infname, curfdp->infname) | |
| 6387 && streq (fdp->taggedfname, taggedfname)) | |
| 6388 /* If we remove the second test above (after the &&) | |
| 6389 then all entries pertaining to the same file are | |
| 6390 coalesced in the tags file. If we use it, then | |
| 6391 entries pertaining to the same file but generated | |
| 6392 from different files (via #line directives) will | |
| 6393 go into separate sections in the tags file. These | |
| 6394 alternatives look equivalent. The first one | |
| 6395 destroys some apparently useless information. */ | |
| 6396 { | |
| 6397 curfdp = fdp; | |
| 6398 free (taggedfname); | |
| 6399 break; | |
| 6400 } | |
| 6401 /* Else, if we already tagged the real file, skip all | |
| 6402 input lines until the next #line directive. */ | |
| 6403 if (fdp == NULL) /* not found */ | |
| 6404 for (fdp = fdhead; fdp != NULL; fdp = fdp->next) | |
| 6405 if (streq (fdp->infabsname, taggedabsname)) | |
| 6406 { | |
| 6407 discard_until_line_directive = TRUE; | |
| 6408 free (taggedfname); | |
| 6409 break; | |
| 6410 } | |
| 6411 /* Else create a new file description and use that from | |
| 6412 now on, until the next #line directive. */ | |
| 6413 if (fdp == NULL) /* not found */ | |
| 6414 { | |
| 6415 fdp = fdhead; | |
| 6416 fdhead = xnew (1, fdesc); | |
| 6417 *fdhead = *curfdp; /* copy curr. file description */ | |
| 6418 fdhead->next = fdp; | |
| 6419 fdhead->infname = savestr (curfdp->infname); | |
| 6420 fdhead->infabsname = savestr (curfdp->infabsname); | |
| 6421 fdhead->infabsdir = savestr (curfdp->infabsdir); | |
| 6422 fdhead->taggedfname = taggedfname; | |
| 6423 fdhead->usecharno = FALSE; | |
| 6424 fdhead->prop = NULL; | |
| 6425 fdhead->written = FALSE; | |
| 6426 curfdp = fdhead; | |
| 6427 } | |
| 6428 } | |
| 6429 free (taggedabsname); | |
| 6430 lineno = lno - 1; | |
| 6431 readline (lbp, stream); | |
| 6432 return; | |
| 6433 } /* if a real #line directive */ | |
| 6434 } /* if #line is followed by a a number */ | |
| 6435 } /* if line begins with "#line " */ | |
| 6436 | |
| 6437 /* If we are here, no #line directive was found. */ | |
| 6438 if (discard_until_line_directive) | |
| 6439 { | |
| 6440 if (result > 0) | |
| 6441 { | |
| 6442 /* Do a tail recursion on ourselves, thus discarding the contents | |
| 6443 of the line buffer. */ | |
| 6444 readline (lbp, stream); | |
| 6445 return; | |
| 6446 } | |
| 6447 /* End of file. */ | |
| 6448 discard_until_line_directive = FALSE; | |
| 6449 return; | |
| 6450 } | |
| 6451 } /* if #line directives should be considered */ | |
| 6452 | |
| 6453 { | |
| 6454 int match; | |
| 6455 regexp *rp; | |
| 6456 char *name; | |
| 6457 | |
| 6458 /* Match against relevant regexps. */ | |
| 6459 if (lbp->len > 0) | |
| 6460 for (rp = p_head; rp != NULL; rp = rp->p_next) | |
| 6461 { | |
| 6462 /* Only use generic regexps or those for the current language. | |
| 6463 Also do not use multiline regexps, which is the job of | |
| 6464 regex_tag_multiline. */ | |
| 6465 if ((rp->lang != NULL && rp->lang != fdhead->lang) | |
| 6466 || rp->multi_line) | |
| 6467 continue; | |
| 6468 | |
| 6469 match = re_match (rp->pat, lbp->buffer, lbp->len, 0, &rp->regs); | |
| 6470 switch (match) | |
| 6471 { | |
| 6472 case -2: | |
| 6473 /* Some error. */ | |
| 6474 if (!rp->error_signaled) | |
| 6475 { | |
| 6476 error ("regexp stack overflow while matching \"%s\"", | |
| 6477 rp->pattern); | |
| 6478 rp->error_signaled = TRUE; | |
| 6479 } | |
| 6480 break; | |
| 6481 case -1: | |
| 6482 /* No match. */ | |
| 6483 break; | |
| 6484 case 0: | |
| 6485 /* Empty string matched. */ | |
| 6486 if (!rp->error_signaled) | |
| 6487 { | |
| 6488 error ("regexp matches the empty string: \"%s\"", rp->pattern); | |
| 6489 rp->error_signaled = TRUE; | |
| 6490 } | |
| 6491 break; | |
| 6492 default: | |
| 6493 /* Match occurred. Construct a tag. */ | |
| 6494 name = rp->name; | |
| 6495 if (name[0] == '\0') | |
| 6496 name = NULL; | |
| 6497 else /* make a named tag */ | |
| 6498 name = substitute (lbp->buffer, rp->name, &rp->regs); | |
| 6499 if (rp->force_explicit_name) | |
| 6500 /* Force explicit tag name, if a name is there. */ | |
| 6501 pfnote (name, TRUE, lbp->buffer, match, lineno, linecharno); | |
| 6502 else | |
| 6503 make_tag (name, strlen (name), TRUE, | |
| 6504 lbp->buffer, match, lineno, linecharno); | |
| 6505 break; | |
| 6506 } | |
| 6507 } | |
| 6508 } | |
| 428 | 6509 } |
| 458 | 6510 |
| 428 | 6511 |
| 6512 /* | |
| 6513 * Return a pointer to a space of size strlen(cp)+1 allocated | |
| 6514 * with xnew where the string CP has been copied. | |
| 6515 */ | |
| 442 | 6516 static char * |
| 428 | 6517 savestr (cp) |
| 6518 char *cp; | |
| 6519 { | |
| 6520 return savenstr (cp, strlen (cp)); | |
| 6521 } | |
| 6522 | |
| 6523 /* | |
| 6524 * Return a pointer to a space of size LEN+1 allocated with xnew where | |
| 6525 * the string CP has been copied for at most the first LEN characters. | |
| 6526 */ | |
| 442 | 6527 static char * |
| 428 | 6528 savenstr (cp, len) |
| 6529 char *cp; | |
| 6530 int len; | |
| 6531 { | |
| 6532 register char *dp; | |
| 6533 | |
| 6534 dp = xnew (len + 1, char); | |
| 6535 strncpy (dp, cp, len); | |
| 6536 dp[len] = '\0'; | |
| 6537 return dp; | |
| 6538 } | |
| 6539 | |
| 6540 /* | |
| 6541 * Return the ptr in sp at which the character c last | |
| 6542 * appears; NULL if not found | |
| 442 | 6543 * |
| 6544 * Identical to POSIX strrchr, included for portability. | |
| 428 | 6545 */ |
| 442 | 6546 static char * |
| 428 | 6547 etags_strrchr (sp, c) |
| 442 | 6548 register const char *sp; |
| 6549 register int c; | |
| 428 | 6550 { |
| 438 | 6551 register const char *r; |
| 428 | 6552 |
| 6553 r = NULL; | |
| 6554 do | |
| 6555 { | |
| 6556 if (*sp == c) | |
| 6557 r = sp; | |
| 6558 } while (*sp++); | |
| 442 | 6559 return (char *)r; |
| 428 | 6560 } |
| 6561 | |
| 6562 /* | |
| 6563 * Return the ptr in sp at which the character c first | |
| 6564 * appears; NULL if not found | |
| 442 | 6565 * |
| 6566 * Identical to POSIX strchr, included for portability. | |
| 428 | 6567 */ |
| 442 | 6568 static char * |
| 428 | 6569 etags_strchr (sp, c) |
| 442 | 6570 register const char *sp; |
| 6571 register int c; | |
| 428 | 6572 { |
| 6573 do | |
| 6574 { | |
| 6575 if (*sp == c) | |
| 442 | 6576 return (char *)sp; |
| 428 | 6577 } while (*sp++); |
| 6578 return NULL; | |
| 6579 } | |
| 6580 | |
| 2225 | 6581 /* |
| 6582 * Compare two strings, ignoring case for alphabetic characters. | |
| 6583 * | |
| 6584 * Same as BSD's strcasecmp, included for portability. | |
| 6585 */ | |
| 6586 static int | |
| 6587 etags_strcasecmp (s1, s2) | |
| 6588 register const char *s1; | |
| 6589 register const char *s2; | |
| 6590 { | |
| 6591 while (*s1 != '\0' | |
| 6592 && (ISALPHA (*s1) && ISALPHA (*s2) | |
| 6593 ? lowcase (*s1) == lowcase (*s2) | |
| 6594 : *s1 == *s2)) | |
| 6595 s1++, s2++; | |
| 6596 | |
| 6597 return (ISALPHA (*s1) && ISALPHA (*s2) | |
| 6598 ? lowcase (*s1) - lowcase (*s2) | |
| 6599 : *s1 - *s2); | |
| 6600 } | |
| 6601 | |
| 6602 /* | |
| 6603 * Compare two strings, ignoring case for alphabetic characters. | |
| 6604 * Stop after a given number of characters | |
| 6605 * | |
| 6606 * Same as BSD's strncasecmp, included for portability. | |
| 6607 */ | |
| 6608 static int | |
| 6609 etags_strncasecmp (s1, s2, n) | |
| 6610 register const char *s1; | |
| 6611 register const char *s2; | |
| 6612 register int n; | |
| 6613 { | |
| 6614 while (*s1 != '\0' && n-- > 0 | |
| 6615 && (ISALPHA (*s1) && ISALPHA (*s2) | |
| 6616 ? lowcase (*s1) == lowcase (*s2) | |
| 6617 : *s1 == *s2)) | |
| 6618 s1++, s2++; | |
| 6619 | |
| 6620 if (n < 0) | |
| 6621 return 0; | |
| 6622 else | |
| 6623 return (ISALPHA (*s1) && ISALPHA (*s2) | |
| 6624 ? lowcase (*s1) - lowcase (*s2) | |
| 6625 : *s1 - *s2); | |
| 6626 } | |
| 6627 | |
| 2554 | 6628 /* Skip spaces (end of string is not space), return new pointer. */ |
| 442 | 6629 static char * |
| 428 | 6630 skip_spaces (cp) |
| 6631 char *cp; | |
| 6632 { | |
| 442 | 6633 while (iswhite (*cp)) |
| 428 | 6634 cp++; |
| 6635 return cp; | |
| 6636 } | |
| 6637 | |
| 2554 | 6638 /* Skip non spaces, except end of string, return new pointer. */ |
| 442 | 6639 static char * |
| 428 | 6640 skip_non_spaces (cp) |
| 6641 char *cp; | |
| 6642 { | |
| 442 | 6643 while (*cp != '\0' && !iswhite (*cp)) |
| 428 | 6644 cp++; |
| 6645 return cp; | |
| 6646 } | |
| 6647 | |
| 6648 /* Print error message and exit. */ | |
| 458 | 6649 void |
| 428 | 6650 fatal (s1, s2) |
| 6651 char *s1, *s2; | |
| 6652 { | |
| 6653 error (s1, s2); | |
| 2225 | 6654 exit (EXIT_FAILURE); |
| 428 | 6655 } |
| 6656 | |
| 442 | 6657 static void |
| 428 | 6658 pfatal (s1) |
| 6659 char *s1; | |
| 6660 { | |
| 6661 perror (s1); | |
| 2225 | 6662 exit (EXIT_FAILURE); |
| 428 | 6663 } |
| 6664 | |
| 442 | 6665 static void |
| 428 | 6666 suggest_asking_for_help () |
| 6667 { | |
| 2325 | 6668 fprintf (stderr, "\tTry `%s %s' for a complete list of options.\n", |
| 3517 | 6669 progname, NO_LONG_OPTIONS ? "-h" : "--help"); |
| 2225 | 6670 exit (EXIT_FAILURE); |
| 428 | 6671 } |
| 6672 | |
| 6673 /* Print error message. `s1' is printf control string, `s2' is arg for it. */ | |
| 442 | 6674 static void |
| 428 | 6675 error (s1, s2) |
| 6676 const char *s1, *s2; | |
| 6677 { | |
| 6678 fprintf (stderr, "%s: ", progname); | |
| 6679 fprintf (stderr, s1, s2); | |
| 6680 fprintf (stderr, "\n"); | |
| 6681 } | |
| 6682 | |
| 6683 /* Return a newly-allocated string whose contents | |
| 6684 concatenate those of s1, s2, s3. */ | |
| 442 | 6685 static char * |
| 428 | 6686 concat (s1, s2, s3) |
| 6687 char *s1, *s2, *s3; | |
| 6688 { | |
| 6689 int len1 = strlen (s1), len2 = strlen (s2), len3 = strlen (s3); | |
| 6690 char *result = xnew (len1 + len2 + len3 + 1, char); | |
| 6691 | |
| 6692 strcpy (result, s1); | |
| 6693 strcpy (result + len1, s2); | |
| 6694 strcpy (result + len1 + len2, s3); | |
| 6695 result[len1 + len2 + len3] = '\0'; | |
| 6696 | |
| 6697 return result; | |
| 6698 } | |
| 458 | 6699 |
| 428 | 6700 |
| 6701 /* Does the same work as the system V getcwd, but does not need to | |
| 6702 guess the buffer size in advance. */ | |
| 442 | 6703 static char * |
| 428 | 6704 etags_getcwd () |
| 6705 { | |
| 6706 #ifdef HAVE_GETCWD | |
| 6707 int bufsize = 200; | |
| 6708 char *path = xnew (bufsize, char); | |
| 6709 | |
| 6710 while (getcwd (path, bufsize) == NULL) | |
| 6711 { | |
| 6712 if (errno != ERANGE) | |
| 6713 pfatal ("getcwd"); | |
| 6714 bufsize *= 2; | |
| 6715 free (path); | |
| 6716 path = xnew (bufsize, char); | |
| 6717 } | |
| 6718 | |
| 6719 canonicalize_filename (path); | |
| 6720 return path; | |
| 6721 | |
| 6722 #else /* not HAVE_GETCWD */ | |
| 458 | 6723 #if MSDOS |
| 6724 | |
| 6725 char *p, path[MAXPATHLEN + 1]; /* Fixed size is safe on MSDOS. */ | |
| 6726 | |
| 6727 getwd (path); | |
| 6728 | |
| 6729 for (p = path; *p != '\0'; p++) | |
| 6730 if (*p == '\\') | |
| 6731 *p = '/'; | |
| 6732 else | |
| 6733 *p = lowcase (*p); | |
| 6734 | |
| 6735 return strdup (path); | |
| 6736 #else /* not MSDOS */ | |
| 428 | 6737 linebuffer path; |
| 6738 FILE *pipe; | |
| 6739 | |
| 2225 | 6740 linebuffer_init (&path); |
| 428 | 6741 pipe = (FILE *) popen ("pwd 2>/dev/null", "r"); |
| 6742 if (pipe == NULL || readline_internal (&path, pipe) == 0) | |
| 6743 pfatal ("pwd"); | |
| 6744 pclose (pipe); | |
| 6745 | |
| 6746 return path.buffer; | |
| 458 | 6747 #endif /* not MSDOS */ |
| 428 | 6748 #endif /* not HAVE_GETCWD */ |
| 6749 } | |
| 6750 | |
| 6751 /* Return a newly allocated string containing the file name of FILE | |
| 6752 relative to the absolute directory DIR (which should end with a slash). */ | |
| 442 | 6753 static char * |
| 428 | 6754 relative_filename (file, dir) |
| 6755 char *file, *dir; | |
| 6756 { | |
| 6757 char *fp, *dp, *afn, *res; | |
| 6758 int i; | |
| 6759 | |
| 6760 /* Find the common root of file and dir (with a trailing slash). */ | |
| 6761 afn = absolute_filename (file, cwd); | |
| 6762 fp = afn; | |
| 6763 dp = dir; | |
| 6764 while (*fp++ == *dp++) | |
| 6765 continue; | |
| 6766 fp--, dp--; /* back to the first differing char */ | |
| 458 | 6767 #ifdef DOS_NT |
| 428 | 6768 if (fp == afn && afn[0] != '/') /* cannot build a relative name */ |
| 6769 return afn; | |
| 6770 #endif | |
| 6771 do /* look at the equal chars until '/' */ | |
| 6772 fp--, dp--; | |
| 6773 while (*fp != '/'); | |
| 6774 | |
| 6775 /* Build a sequence of "../" strings for the resulting relative file name. */ | |
| 6776 i = 0; | |
| 6777 while ((dp = etags_strchr (dp + 1, '/')) != NULL) | |
| 6778 i += 1; | |
| 6779 res = xnew (3*i + strlen (fp + 1) + 1, char); | |
| 6780 res[0] = '\0'; | |
| 6781 while (i-- > 0) | |
| 6782 strcat (res, "../"); | |
| 6783 | |
| 6784 /* Add the file name relative to the common root of file and dir. */ | |
| 6785 strcat (res, fp + 1); | |
| 6786 free (afn); | |
| 6787 | |
| 6788 return res; | |
| 6789 } | |
| 6790 | |
| 6791 /* Return a newly allocated string containing the absolute file name | |
| 6792 of FILE given DIR (which should end with a slash). */ | |
| 442 | 6793 static char * |
| 428 | 6794 absolute_filename (file, dir) |
| 6795 char *file, *dir; | |
| 6796 { | |
| 6797 char *slashp, *cp, *res; | |
| 6798 | |
| 6799 if (filename_is_absolute (file)) | |
| 6800 res = savestr (file); | |
| 458 | 6801 #ifdef DOS_NT |
| 428 | 6802 /* We don't support non-absolute file names with a drive |
| 6803 letter, like `d:NAME' (it's too much hassle). */ | |
| 6804 else if (file[1] == ':') | |
| 6805 fatal ("%s: relative file names with drive letters not supported", file); | |
| 6806 #endif | |
| 6807 else | |
| 6808 res = concat (dir, file, ""); | |
| 6809 | |
| 6810 /* Delete the "/dirname/.." and "/." substrings. */ | |
| 6811 slashp = etags_strchr (res, '/'); | |
| 6812 while (slashp != NULL && slashp[0] != '\0') | |
| 6813 { | |
| 6814 if (slashp[1] == '.') | |
| 6815 { | |
| 6816 if (slashp[2] == '.' | |
| 6817 && (slashp[3] == '/' || slashp[3] == '\0')) | |
| 6818 { | |
| 6819 cp = slashp; | |
| 6820 do | |
| 6821 cp--; | |
| 6822 while (cp >= res && !filename_is_absolute (cp)); | |
| 6823 if (cp < res) | |
| 6824 cp = slashp; /* the absolute name begins with "/.." */ | |
| 458 | 6825 #ifdef DOS_NT |
| 6826 /* Under MSDOS and NT we get `d:/NAME' as absolute | |
| 428 | 6827 file name, so the luser could say `d:/../NAME'. |
| 6828 We silently treat this as `d:/NAME'. */ | |
| 6829 else if (cp[0] != '/') | |
| 6830 cp = slashp; | |
| 6831 #endif | |
|
4768
0f5bee973a7b
Etags: use memmove instead of strcpy to move characters within a string. See
Jerry James <james@xemacs.org>
parents:
3993
diff
changeset
|
6832 memmove (cp, slashp + 3, strlen (slashp + 3) + 1); |
| 428 | 6833 slashp = cp; |
| 6834 continue; | |
| 6835 } | |
| 6836 else if (slashp[2] == '/' || slashp[2] == '\0') | |
| 6837 { | |
|
4768
0f5bee973a7b
Etags: use memmove instead of strcpy to move characters within a string. See
Jerry James <james@xemacs.org>
parents:
3993
diff
changeset
|
6838 memmove (slashp, slashp + 2, strlen (slashp + 2) + 1); |
| 428 | 6839 continue; |
| 6840 } | |
| 6841 } | |
| 6842 | |
| 6843 slashp = etags_strchr (slashp + 1, '/'); | |
| 6844 } | |
| 6845 | |
| 3517 | 6846 if (res[0] == '\0') /* just a safety net: should never happen */ |
| 6847 { | |
| 6848 free (res); | |
| 6849 return savestr ("/"); | |
| 6850 } | |
| 428 | 6851 else |
| 6852 return res; | |
| 6853 } | |
| 6854 | |
| 6855 /* Return a newly allocated string containing the absolute | |
| 6856 file name of dir where FILE resides given DIR (which should | |
| 6857 end with a slash). */ | |
| 442 | 6858 static char * |
| 428 | 6859 absolute_dirname (file, dir) |
| 6860 char *file, *dir; | |
| 6861 { | |
| 6862 char *slashp, *res; | |
| 6863 char save; | |
| 6864 | |
| 6865 canonicalize_filename (file); | |
| 6866 slashp = etags_strrchr (file, '/'); | |
| 6867 if (slashp == NULL) | |
| 6868 return savestr (dir); | |
| 6869 save = slashp[1]; | |
| 6870 slashp[1] = '\0'; | |
| 6871 res = absolute_filename (file, dir); | |
| 6872 slashp[1] = save; | |
| 6873 | |
| 6874 return res; | |
| 6875 } | |
| 6876 | |
| 6877 /* Whether the argument string is an absolute file name. The argument | |
| 6878 string must have been canonicalized with canonicalize_filename. */ | |
| 442 | 6879 static bool |
| 428 | 6880 filename_is_absolute (fn) |
| 6881 char *fn; | |
| 6882 { | |
| 6883 return (fn[0] == '/' | |
| 458 | 6884 #ifdef DOS_NT |
| 6885 || (ISALPHA(fn[0]) && fn[1] == ':' && fn[2] == '/') | |
| 428 | 6886 #endif |
| 6887 ); | |
| 6888 } | |
| 6889 | |
| 6890 /* Translate backslashes into slashes. Works in place. */ | |
| 442 | 6891 static void |
| 428 | 6892 canonicalize_filename (fn) |
| 6893 register char *fn; | |
| 6894 { | |
| 458 | 6895 #ifdef DOS_NT |
| 428 | 6896 /* Canonicalize drive letter case. */ |
| 458 | 6897 if (fn[0] != '\0' && fn[1] == ':' && ISLOWER (fn[0])) |
| 6898 fn[0] = upcase (fn[0]); | |
| 428 | 6899 /* Convert backslashes to slashes. */ |
| 6900 for (; *fn != '\0'; fn++) | |
| 6901 if (*fn == '\\') | |
| 6902 *fn = '/'; | |
| 6903 #else | |
| 6904 /* No action. */ | |
| 6905 fn = NULL; /* shut up the compiler */ | |
| 6906 #endif | |
| 6907 } | |
| 6908 | |
| 2225 | 6909 |
| 6910 /* Initialize a linebuffer for use */ | |
| 6911 static void | |
| 6912 linebuffer_init (lbp) | |
| 6913 linebuffer *lbp; | |
| 6914 { | |
| 6915 lbp->size = (DEBUG) ? 3 : 200; | |
| 6916 lbp->buffer = xnew (lbp->size, char); | |
| 6917 lbp->buffer[0] = '\0'; | |
| 6918 lbp->len = 0; | |
| 6919 } | |
| 6920 | |
| 458 | 6921 /* Set the minimum size of a string contained in a linebuffer. */ |
| 442 | 6922 static void |
| 458 | 6923 linebuffer_setlen (lbp, toksize) |
| 428 | 6924 linebuffer *lbp; |
| 6925 int toksize; | |
| 6926 { | |
| 458 | 6927 while (lbp->size <= toksize) |
| 6928 { | |
| 6929 lbp->size *= 2; | |
| 6930 xrnew (lbp->buffer, lbp->size, char); | |
| 6931 } | |
| 6932 lbp->len = toksize; | |
| 428 | 6933 } |
| 6934 | |
| 2225 | 6935 /* Like malloc but get fatal error if memory is exhausted. */ |
| 6936 static PTR | |
| 428 | 6937 xmalloc (size) |
| 6938 unsigned int size; | |
| 6939 { | |
| 709 | 6940 PTR result = (PTR) malloc (size); |
| 428 | 6941 if (result == NULL) |
| 6942 fatal ("virtual memory exhausted", (char *)NULL); | |
| 6943 return result; | |
| 6944 } | |
| 6945 | |
| 2225 | 6946 static PTR |
| 428 | 6947 xrealloc (ptr, size) |
| 6948 char *ptr; | |
| 6949 unsigned int size; | |
| 6950 { | |
| 709 | 6951 PTR result = (PTR) realloc (ptr, size); |
| 428 | 6952 if (result == NULL) |
| 6953 fatal ("virtual memory exhausted", (char *)NULL); | |
| 6954 return result; | |
| 6955 } | |
| 709 | 6956 |
| 6957 /* | |
| 6958 * Local Variables: | |
| 6959 * indent-tabs-mode: t | |
| 6960 * tab-width: 8 | |
| 2225 | 6961 * fill-column: 79 |
| 6962 * c-font-lock-extra-types: ("FILE" "bool" "language" "linebuffer" "fdesc" "node" "regexp") | |
| 3972 | 6963 * c-file-style: "gnu" |
| 709 | 6964 * End: |
| 6965 */ | |
| 2225 | 6966 |
| 6967 /* arch-tag: 8a9b748d-390c-4922-99db-2eeefa921051 | |
| 6968 (do not change this comment) */ | |
| 6969 | |
| 6970 /* etags.c ends here */ |