| 1 | /* Tags file maker to go with GNU Emacs -*- coding: utf-8 -*- |
| 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, 1987-1989, 1993-1995, 1998-2013 Free Software |
| 32 | Foundation, Inc. |
| 33 | |
| 34 | This file is not considered part of GNU Emacs. |
| 35 | |
| 36 | This program is free software: you can redistribute it and/or modify |
| 37 | it under the terms of the GNU General Public License as published by |
| 38 | the Free Software Foundation, either version 3 of the License, or |
| 39 | (at your option) any later version. |
| 40 | |
| 41 | This program is distributed in the hope that it will be useful, |
| 42 | but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 43 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 44 | GNU General Public License for more details. |
| 45 | |
| 46 | You should have received a copy of the GNU General Public License |
| 47 | along with this program. If not, see <http://www.gnu.org/licenses/>. */ |
| 48 | |
| 49 | |
| 50 | /* NB To comply with the above BSD license, copyright information is |
| 51 | reproduced in etc/ETAGS.README. That file should be updated when the |
| 52 | above notices are. |
| 53 | |
| 54 | To the best of our knowledge, this code was originally based on the |
| 55 | ctags.c distributed with BSD4.2, which was copyrighted by the |
| 56 | University of California, as described above. */ |
| 57 | |
| 58 | |
| 59 | /* |
| 60 | * Authors: |
| 61 | * 1983 Ctags originally by Ken Arnold. |
| 62 | * 1984 Fortran added by Jim Kleckner. |
| 63 | * 1984 Ed Pelegri-Llopart added C typedefs. |
| 64 | * 1985 Emacs TAGS format by Richard Stallman. |
| 65 | * 1989 Sam Kendall added C++. |
| 66 | * 1992 Joseph B. Wells improved C and C++ parsing. |
| 67 | * 1993 Francesco Potortì reorganized C and C++. |
| 68 | * 1994 Line-by-line regexp tags by Tom Tromey. |
| 69 | * 2001 Nested classes by Francesco Potortì (concept by Mykola Dzyuba). |
| 70 | * 2002 #line directives by Francesco Potortì. |
| 71 | * |
| 72 | * Francesco Potortì <pot@gnu.org> has maintained and improved it since 1993. |
| 73 | */ |
| 74 | |
| 75 | /* |
| 76 | * If you want to add support for a new language, start by looking at the LUA |
| 77 | * language, which is the simplest. Alternatively, consider distributing etags |
| 78 | * together with a configuration file containing regexp definitions for etags. |
| 79 | */ |
| 80 | |
| 81 | char pot_etags_version[] = "@(#) pot revision number is 17.38.1.4"; |
| 82 | |
| 83 | #define TRUE 1 |
| 84 | #define FALSE 0 |
| 85 | |
| 86 | #ifdef DEBUG |
| 87 | # undef DEBUG |
| 88 | # define DEBUG TRUE |
| 89 | #else |
| 90 | # define DEBUG FALSE |
| 91 | # define NDEBUG /* disable assert */ |
| 92 | #endif |
| 93 | |
| 94 | #include <config.h> |
| 95 | |
| 96 | #ifndef _GNU_SOURCE |
| 97 | # define _GNU_SOURCE 1 /* enables some compiler checks on GNU */ |
| 98 | #endif |
| 99 | |
| 100 | /* WIN32_NATIVE is for XEmacs. |
| 101 | MSDOS, WINDOWSNT, DOS_NT are for Emacs. */ |
| 102 | #ifdef WIN32_NATIVE |
| 103 | # undef MSDOS |
| 104 | # undef WINDOWSNT |
| 105 | # define WINDOWSNT |
| 106 | #endif /* WIN32_NATIVE */ |
| 107 | |
| 108 | #ifdef MSDOS |
| 109 | # undef MSDOS |
| 110 | # define MSDOS TRUE |
| 111 | # include <fcntl.h> |
| 112 | # include <sys/param.h> |
| 113 | # include <io.h> |
| 114 | #else |
| 115 | # define MSDOS FALSE |
| 116 | #endif /* MSDOS */ |
| 117 | |
| 118 | #ifdef WINDOWSNT |
| 119 | # include <fcntl.h> |
| 120 | # include <direct.h> |
| 121 | # include <io.h> |
| 122 | # define MAXPATHLEN _MAX_PATH |
| 123 | # undef HAVE_NTGUI |
| 124 | # undef DOS_NT |
| 125 | # define DOS_NT |
| 126 | #endif /* WINDOWSNT */ |
| 127 | |
| 128 | #include <unistd.h> |
| 129 | #include <stdarg.h> |
| 130 | #include <stdlib.h> |
| 131 | #include <string.h> |
| 132 | #include <stdio.h> |
| 133 | #include <ctype.h> |
| 134 | #include <errno.h> |
| 135 | #include <sys/types.h> |
| 136 | #include <sys/stat.h> |
| 137 | #include <c-strcase.h> |
| 138 | |
| 139 | #include <assert.h> |
| 140 | #ifdef NDEBUG |
| 141 | # undef assert /* some systems have a buggy assert.h */ |
| 142 | # define assert(x) ((void) 0) |
| 143 | #endif |
| 144 | |
| 145 | #include <getopt.h> |
| 146 | #include <regex.h> |
| 147 | |
| 148 | /* Define CTAGS to make the program "ctags" compatible with the usual one. |
| 149 | Leave it undefined to make the program "etags", which makes emacs-style |
| 150 | tag tables and tags typedefs, #defines and struct/union/enum by default. */ |
| 151 | #ifdef CTAGS |
| 152 | # undef CTAGS |
| 153 | # define CTAGS TRUE |
| 154 | #else |
| 155 | # define CTAGS FALSE |
| 156 | #endif |
| 157 | |
| 158 | #define streq(s,t) (assert ((s)!=NULL || (t)!=NULL), !strcmp (s, t)) |
| 159 | #define strcaseeq(s,t) (assert ((s)!=NULL && (t)!=NULL), !c_strcasecmp (s, t)) |
| 160 | #define strneq(s,t,n) (assert ((s)!=NULL || (t)!=NULL), !strncmp (s, t, n)) |
| 161 | #define strncaseeq(s,t,n) (assert ((s)!=NULL && (t)!=NULL), !c_strncasecmp (s, t, n)) |
| 162 | |
| 163 | #define CHARS 256 /* 2^sizeof(char) */ |
| 164 | #define CHAR(x) ((unsigned int)(x) & (CHARS - 1)) |
| 165 | #define iswhite(c) (_wht[CHAR (c)]) /* c is white (see white) */ |
| 166 | #define notinname(c) (_nin[CHAR (c)]) /* c is not in a name (see nonam) */ |
| 167 | #define begtoken(c) (_btk[CHAR (c)]) /* c can start token (see begtk) */ |
| 168 | #define intoken(c) (_itk[CHAR (c)]) /* c can be in token (see midtk) */ |
| 169 | #define endtoken(c) (_etk[CHAR (c)]) /* c ends tokens (see endtk) */ |
| 170 | |
| 171 | #define ISALNUM(c) isalnum (CHAR (c)) |
| 172 | #define ISALPHA(c) isalpha (CHAR (c)) |
| 173 | #define ISDIGIT(c) isdigit (CHAR (c)) |
| 174 | #define ISLOWER(c) islower (CHAR (c)) |
| 175 | |
| 176 | #define lowcase(c) tolower (CHAR (c)) |
| 177 | |
| 178 | |
| 179 | /* |
| 180 | * xnew, xrnew -- allocate, reallocate storage |
| 181 | * |
| 182 | * SYNOPSIS: Type *xnew (int n, Type); |
| 183 | * void xrnew (OldPointer, int n, Type); |
| 184 | */ |
| 185 | #if DEBUG |
| 186 | # include "chkmalloc.h" |
| 187 | # define xnew(n,Type) ((Type *) trace_malloc (__FILE__, __LINE__, \ |
| 188 | (n) * sizeof (Type))) |
| 189 | # define xrnew(op,n,Type) ((op) = (Type *) trace_realloc (__FILE__, __LINE__, \ |
| 190 | (char *) (op), (n) * sizeof (Type))) |
| 191 | #else |
| 192 | # define xnew(n,Type) ((Type *) xmalloc ((n) * sizeof (Type))) |
| 193 | # define xrnew(op,n,Type) ((op) = (Type *) xrealloc ( \ |
| 194 | (char *) (op), (n) * sizeof (Type))) |
| 195 | #endif |
| 196 | |
| 197 | #define bool int |
| 198 | |
| 199 | typedef void Lang_function (FILE *); |
| 200 | |
| 201 | typedef struct |
| 202 | { |
| 203 | const char *suffix; /* file name suffix for this compressor */ |
| 204 | const char *command; /* takes one arg and decompresses to stdout */ |
| 205 | } compressor; |
| 206 | |
| 207 | typedef struct |
| 208 | { |
| 209 | const char *name; /* language name */ |
| 210 | const char *help; /* detailed help for the language */ |
| 211 | Lang_function *function; /* parse function */ |
| 212 | const char **suffixes; /* name suffixes of this language's files */ |
| 213 | const char **filenames; /* names of this language's files */ |
| 214 | const char **interpreters; /* interpreters for this language */ |
| 215 | bool metasource; /* source used to generate other sources */ |
| 216 | } language; |
| 217 | |
| 218 | typedef struct fdesc |
| 219 | { |
| 220 | struct fdesc *next; /* for the linked list */ |
| 221 | char *infname; /* uncompressed input file name */ |
| 222 | char *infabsname; /* absolute uncompressed input file name */ |
| 223 | char *infabsdir; /* absolute dir of input file */ |
| 224 | char *taggedfname; /* file name to write in tagfile */ |
| 225 | language *lang; /* language of file */ |
| 226 | char *prop; /* file properties to write in tagfile */ |
| 227 | bool usecharno; /* etags tags shall contain char number */ |
| 228 | bool written; /* entry written in the tags file */ |
| 229 | } fdesc; |
| 230 | |
| 231 | typedef struct node_st |
| 232 | { /* sorting structure */ |
| 233 | struct node_st *left, *right; /* left and right sons */ |
| 234 | fdesc *fdp; /* description of file to whom tag belongs */ |
| 235 | char *name; /* tag name */ |
| 236 | char *regex; /* search regexp */ |
| 237 | bool valid; /* write this tag on the tag file */ |
| 238 | bool is_func; /* function tag: use regexp in CTAGS mode */ |
| 239 | bool been_warned; /* warning already given for duplicated tag */ |
| 240 | int lno; /* line number tag is on */ |
| 241 | long cno; /* character number line starts on */ |
| 242 | } node; |
| 243 | |
| 244 | /* |
| 245 | * A `linebuffer' is a structure which holds a line of text. |
| 246 | * `readline_internal' reads a line from a stream into a linebuffer |
| 247 | * and works regardless of the length of the line. |
| 248 | * SIZE is the size of BUFFER, LEN is the length of the string in |
| 249 | * BUFFER after readline reads it. |
| 250 | */ |
| 251 | typedef struct |
| 252 | { |
| 253 | long size; |
| 254 | int len; |
| 255 | char *buffer; |
| 256 | } linebuffer; |
| 257 | |
| 258 | /* Used to support mixing of --lang and file names. */ |
| 259 | typedef struct |
| 260 | { |
| 261 | enum { |
| 262 | at_language, /* a language specification */ |
| 263 | at_regexp, /* a regular expression */ |
| 264 | at_filename, /* a file name */ |
| 265 | at_stdin, /* read from stdin here */ |
| 266 | at_end /* stop parsing the list */ |
| 267 | } arg_type; /* argument type */ |
| 268 | language *lang; /* language associated with the argument */ |
| 269 | char *what; /* the argument itself */ |
| 270 | } argument; |
| 271 | |
| 272 | /* Structure defining a regular expression. */ |
| 273 | typedef struct regexp |
| 274 | { |
| 275 | struct regexp *p_next; /* pointer to next in list */ |
| 276 | language *lang; /* if set, use only for this language */ |
| 277 | char *pattern; /* the regexp pattern */ |
| 278 | char *name; /* tag name */ |
| 279 | struct re_pattern_buffer *pat; /* the compiled pattern */ |
| 280 | struct re_registers regs; /* re registers */ |
| 281 | bool error_signaled; /* already signaled for this regexp */ |
| 282 | bool force_explicit_name; /* do not allow implicit tag name */ |
| 283 | bool ignore_case; /* ignore case when matching */ |
| 284 | bool multi_line; /* do a multi-line match on the whole file */ |
| 285 | } regexp; |
| 286 | |
| 287 | |
| 288 | /* Many compilers barf on this: |
| 289 | Lang_function Ada_funcs; |
| 290 | so let's write it this way */ |
| 291 | static void Ada_funcs (FILE *); |
| 292 | static void Asm_labels (FILE *); |
| 293 | static void C_entries (int c_ext, FILE *); |
| 294 | static void default_C_entries (FILE *); |
| 295 | static void plain_C_entries (FILE *); |
| 296 | static void Cjava_entries (FILE *); |
| 297 | static void Cobol_paragraphs (FILE *); |
| 298 | static void Cplusplus_entries (FILE *); |
| 299 | static void Cstar_entries (FILE *); |
| 300 | static void Erlang_functions (FILE *); |
| 301 | static void Forth_words (FILE *); |
| 302 | static void Fortran_functions (FILE *); |
| 303 | static void HTML_labels (FILE *); |
| 304 | static void Lisp_functions (FILE *); |
| 305 | static void Lua_functions (FILE *); |
| 306 | static void Makefile_targets (FILE *); |
| 307 | static void Pascal_functions (FILE *); |
| 308 | static void Perl_functions (FILE *); |
| 309 | static void PHP_functions (FILE *); |
| 310 | static void PS_functions (FILE *); |
| 311 | static void Prolog_functions (FILE *); |
| 312 | static void Python_functions (FILE *); |
| 313 | static void Scheme_functions (FILE *); |
| 314 | static void TeX_commands (FILE *); |
| 315 | static void Texinfo_nodes (FILE *); |
| 316 | static void Yacc_entries (FILE *); |
| 317 | static void just_read_file (FILE *); |
| 318 | |
| 319 | static language *get_language_from_langname (const char *); |
| 320 | static void readline (linebuffer *, FILE *); |
| 321 | static long readline_internal (linebuffer *, FILE *); |
| 322 | static bool nocase_tail (const char *); |
| 323 | static void get_tag (char *, char **); |
| 324 | |
| 325 | static void analyse_regex (char *); |
| 326 | static void free_regexps (void); |
| 327 | static void regex_tag_multiline (void); |
| 328 | static void error (const char *, ...) ATTRIBUTE_FORMAT_PRINTF (1, 2); |
| 329 | static _Noreturn void suggest_asking_for_help (void); |
| 330 | _Noreturn void fatal (const char *, const char *); |
| 331 | static _Noreturn void pfatal (const char *); |
| 332 | static void add_node (node *, node **); |
| 333 | |
| 334 | static void init (void); |
| 335 | static void process_file_name (char *, language *); |
| 336 | static void process_file (FILE *, char *, language *); |
| 337 | static void find_entries (FILE *); |
| 338 | static void free_tree (node *); |
| 339 | static void free_fdesc (fdesc *); |
| 340 | static void pfnote (char *, bool, char *, int, int, long); |
| 341 | static void invalidate_nodes (fdesc *, node **); |
| 342 | static void put_entries (node *); |
| 343 | |
| 344 | static char *concat (const char *, const char *, const char *); |
| 345 | static char *skip_spaces (char *); |
| 346 | static char *skip_non_spaces (char *); |
| 347 | static char *skip_name (char *); |
| 348 | static char *savenstr (const char *, int); |
| 349 | static char *savestr (const char *); |
| 350 | static char *etags_strchr (const char *, int); |
| 351 | static char *etags_strrchr (const char *, int); |
| 352 | static char *etags_getcwd (void); |
| 353 | static char *relative_filename (char *, char *); |
| 354 | static char *absolute_filename (char *, char *); |
| 355 | static char *absolute_dirname (char *, char *); |
| 356 | static bool filename_is_absolute (char *f); |
| 357 | static void canonicalize_filename (char *); |
| 358 | static void linebuffer_init (linebuffer *); |
| 359 | static void linebuffer_setlen (linebuffer *, int); |
| 360 | static void *xmalloc (size_t); |
| 361 | static void *xrealloc (char *, size_t); |
| 362 | |
| 363 | \f |
| 364 | static char searchar = '/'; /* use /.../ searches */ |
| 365 | |
| 366 | static char *tagfile; /* output file */ |
| 367 | static char *progname; /* name this program was invoked with */ |
| 368 | static char *cwd; /* current working directory */ |
| 369 | static char *tagfiledir; /* directory of tagfile */ |
| 370 | static FILE *tagf; /* ioptr for tags file */ |
| 371 | static ptrdiff_t whatlen_max; /* maximum length of any 'what' member */ |
| 372 | |
| 373 | static fdesc *fdhead; /* head of file description list */ |
| 374 | static fdesc *curfdp; /* current file description */ |
| 375 | static int lineno; /* line number of current line */ |
| 376 | static long charno; /* current character number */ |
| 377 | static long linecharno; /* charno of start of current line */ |
| 378 | static char *dbp; /* pointer to start of current tag */ |
| 379 | |
| 380 | static const int invalidcharno = -1; |
| 381 | |
| 382 | static node *nodehead; /* the head of the binary tree of tags */ |
| 383 | static node *last_node; /* the last node created */ |
| 384 | |
| 385 | static linebuffer lb; /* the current line */ |
| 386 | static linebuffer filebuf; /* a buffer containing the whole file */ |
| 387 | static linebuffer token_name; /* a buffer containing a tag name */ |
| 388 | |
| 389 | /* boolean "functions" (see init) */ |
| 390 | static bool _wht[CHARS], _nin[CHARS], _itk[CHARS], _btk[CHARS], _etk[CHARS]; |
| 391 | static const char |
| 392 | /* white chars */ |
| 393 | *white = " \f\t\n\r\v", |
| 394 | /* not in a name */ |
| 395 | *nonam = " \f\t\n\r()=,;", /* look at make_tag before modifying! */ |
| 396 | /* token ending chars */ |
| 397 | *endtk = " \t\n\r\"'#()[]{}=-+%*/&|^~!<>;,.:?", |
| 398 | /* token starting chars */ |
| 399 | *begtk = "ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz$~@", |
| 400 | /* valid in-token chars */ |
| 401 | *midtk = "ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz$0123456789"; |
| 402 | |
| 403 | static bool append_to_tagfile; /* -a: append to tags */ |
| 404 | /* The next five default to TRUE in C and derived languages. */ |
| 405 | static bool typedefs; /* -t: create tags for C and Ada typedefs */ |
| 406 | static bool typedefs_or_cplusplus; /* -T: create tags for C typedefs, level */ |
| 407 | /* 0 struct/enum/union decls, and C++ */ |
| 408 | /* member functions. */ |
| 409 | static bool constantypedefs; /* -d: create tags for C #define, enum */ |
| 410 | /* constants and variables. */ |
| 411 | /* -D: opposite of -d. Default under ctags. */ |
| 412 | static bool globals; /* create tags for global variables */ |
| 413 | static bool members; /* create tags for C member variables */ |
| 414 | static bool declarations; /* --declarations: tag them and extern in C&Co*/ |
| 415 | static bool no_line_directive; /* ignore #line directives (undocumented) */ |
| 416 | static bool no_duplicates; /* no duplicate tags for ctags (undocumented) */ |
| 417 | static bool update; /* -u: update tags */ |
| 418 | static bool vgrind_style; /* -v: create vgrind style index output */ |
| 419 | static bool no_warnings; /* -w: suppress warnings (undocumented) */ |
| 420 | static bool cxref_style; /* -x: create cxref style output */ |
| 421 | static bool cplusplus; /* .[hc] means C++, not C (undocumented) */ |
| 422 | static bool ignoreindent; /* -I: ignore indentation in C */ |
| 423 | static bool packages_only; /* --packages-only: in Ada, only tag packages*/ |
| 424 | |
| 425 | /* STDIN is defined in LynxOS system headers */ |
| 426 | #ifdef STDIN |
| 427 | # undef STDIN |
| 428 | #endif |
| 429 | |
| 430 | #define STDIN 0x1001 /* returned by getopt_long on --parse-stdin */ |
| 431 | static bool parsing_stdin; /* --parse-stdin used */ |
| 432 | |
| 433 | static regexp *p_head; /* list of all regexps */ |
| 434 | static bool need_filebuf; /* some regexes are multi-line */ |
| 435 | |
| 436 | static struct option longopts[] = |
| 437 | { |
| 438 | { "append", no_argument, NULL, 'a' }, |
| 439 | { "packages-only", no_argument, &packages_only, TRUE }, |
| 440 | { "c++", no_argument, NULL, 'C' }, |
| 441 | { "declarations", no_argument, &declarations, TRUE }, |
| 442 | { "no-line-directive", no_argument, &no_line_directive, TRUE }, |
| 443 | { "no-duplicates", no_argument, &no_duplicates, TRUE }, |
| 444 | { "help", no_argument, NULL, 'h' }, |
| 445 | { "help", no_argument, NULL, 'H' }, |
| 446 | { "ignore-indentation", no_argument, NULL, 'I' }, |
| 447 | { "language", required_argument, NULL, 'l' }, |
| 448 | { "members", no_argument, &members, TRUE }, |
| 449 | { "no-members", no_argument, &members, FALSE }, |
| 450 | { "output", required_argument, NULL, 'o' }, |
| 451 | { "regex", required_argument, NULL, 'r' }, |
| 452 | { "no-regex", no_argument, NULL, 'R' }, |
| 453 | { "ignore-case-regex", required_argument, NULL, 'c' }, |
| 454 | { "parse-stdin", required_argument, NULL, STDIN }, |
| 455 | { "version", no_argument, NULL, 'V' }, |
| 456 | |
| 457 | #if CTAGS /* Ctags options */ |
| 458 | { "backward-search", no_argument, NULL, 'B' }, |
| 459 | { "cxref", no_argument, NULL, 'x' }, |
| 460 | { "defines", no_argument, NULL, 'd' }, |
| 461 | { "globals", no_argument, &globals, TRUE }, |
| 462 | { "typedefs", no_argument, NULL, 't' }, |
| 463 | { "typedefs-and-c++", no_argument, NULL, 'T' }, |
| 464 | { "update", no_argument, NULL, 'u' }, |
| 465 | { "vgrind", no_argument, NULL, 'v' }, |
| 466 | { "no-warn", no_argument, NULL, 'w' }, |
| 467 | |
| 468 | #else /* Etags options */ |
| 469 | { "no-defines", no_argument, NULL, 'D' }, |
| 470 | { "no-globals", no_argument, &globals, FALSE }, |
| 471 | { "include", required_argument, NULL, 'i' }, |
| 472 | #endif |
| 473 | { NULL } |
| 474 | }; |
| 475 | |
| 476 | static compressor compressors[] = |
| 477 | { |
| 478 | { "z", "gzip -d -c"}, |
| 479 | { "Z", "gzip -d -c"}, |
| 480 | { "gz", "gzip -d -c"}, |
| 481 | { "GZ", "gzip -d -c"}, |
| 482 | { "bz2", "bzip2 -d -c" }, |
| 483 | { "xz", "xz -d -c" }, |
| 484 | { NULL } |
| 485 | }; |
| 486 | |
| 487 | /* |
| 488 | * Language stuff. |
| 489 | */ |
| 490 | |
| 491 | /* Ada code */ |
| 492 | static const char *Ada_suffixes [] = |
| 493 | { "ads", "adb", "ada", NULL }; |
| 494 | static const char Ada_help [] = |
| 495 | "In Ada code, functions, procedures, packages, tasks and types are\n\ |
| 496 | tags. Use the `--packages-only' option to create tags for\n\ |
| 497 | packages only.\n\ |
| 498 | Ada tag names have suffixes indicating the type of entity:\n\ |
| 499 | Entity type: Qualifier:\n\ |
| 500 | ------------ ----------\n\ |
| 501 | function /f\n\ |
| 502 | procedure /p\n\ |
| 503 | package spec /s\n\ |
| 504 | package body /b\n\ |
| 505 | type /t\n\ |
| 506 | task /k\n\ |
| 507 | Thus, `M-x find-tag <RET> bidule/b <RET>' will go directly to the\n\ |
| 508 | body of the package `bidule', while `M-x find-tag <RET> bidule <RET>'\n\ |
| 509 | will just search for any tag `bidule'."; |
| 510 | |
| 511 | /* Assembly code */ |
| 512 | static const char *Asm_suffixes [] = |
| 513 | { "a", /* Unix assembler */ |
| 514 | "asm", /* Microcontroller assembly */ |
| 515 | "def", /* BSO/Tasking definition includes */ |
| 516 | "inc", /* Microcontroller include files */ |
| 517 | "ins", /* Microcontroller include files */ |
| 518 | "s", "sa", /* Unix assembler */ |
| 519 | "S", /* cpp-processed Unix assembler */ |
| 520 | "src", /* BSO/Tasking C compiler output */ |
| 521 | NULL |
| 522 | }; |
| 523 | static const char Asm_help [] = |
| 524 | "In assembler code, labels appearing at the beginning of a line,\n\ |
| 525 | followed by a colon, are tags."; |
| 526 | |
| 527 | |
| 528 | /* Note that .c and .h can be considered C++, if the --c++ flag was |
| 529 | given, or if the `class' or `template' keywords are met inside the file. |
| 530 | That is why default_C_entries is called for these. */ |
| 531 | static const char *default_C_suffixes [] = |
| 532 | { "c", "h", NULL }; |
| 533 | #if CTAGS /* C help for Ctags */ |
| 534 | static const char default_C_help [] = |
| 535 | "In C code, any C function is a tag. Use -t to tag typedefs.\n\ |
| 536 | Use -T to tag definitions of `struct', `union' and `enum'.\n\ |
| 537 | Use -d to tag `#define' macro definitions and `enum' constants.\n\ |
| 538 | Use --globals to tag global variables.\n\ |
| 539 | You can tag function declarations and external variables by\n\ |
| 540 | using `--declarations', and struct members by using `--members'."; |
| 541 | #else /* C help for Etags */ |
| 542 | static const char default_C_help [] = |
| 543 | "In C code, any C function or typedef is a tag, and so are\n\ |
| 544 | definitions of `struct', `union' and `enum'. `#define' macro\n\ |
| 545 | definitions and `enum' constants are tags unless you specify\n\ |
| 546 | `--no-defines'. Global variables are tags unless you specify\n\ |
| 547 | `--no-globals' and so are struct members unless you specify\n\ |
| 548 | `--no-members'. Use of `--no-globals', `--no-defines' and\n\ |
| 549 | `--no-members' can make the tags table file much smaller.\n\ |
| 550 | You can tag function declarations and external variables by\n\ |
| 551 | using `--declarations'."; |
| 552 | #endif /* C help for Ctags and Etags */ |
| 553 | |
| 554 | static const char *Cplusplus_suffixes [] = |
| 555 | { "C", "c++", "cc", "cpp", "cxx", "H", "h++", "hh", "hpp", "hxx", |
| 556 | "M", /* Objective C++ */ |
| 557 | "pdb", /* PostScript with C syntax */ |
| 558 | NULL }; |
| 559 | static const char Cplusplus_help [] = |
| 560 | "In C++ code, all the tag constructs of C code are tagged. (Use\n\ |
| 561 | --help --lang=c --lang=c++ for full help.)\n\ |
| 562 | In addition to C tags, member functions are also recognized. Member\n\ |
| 563 | variables are recognized unless you use the `--no-members' option.\n\ |
| 564 | Tags for variables and functions in classes are named `CLASS::VARIABLE'\n\ |
| 565 | and `CLASS::FUNCTION'. `operator' definitions have tag names like\n\ |
| 566 | `operator+'."; |
| 567 | |
| 568 | static const char *Cjava_suffixes [] = |
| 569 | { "java", NULL }; |
| 570 | static char Cjava_help [] = |
| 571 | "In Java code, all the tags constructs of C and C++ code are\n\ |
| 572 | tagged. (Use --help --lang=c --lang=c++ --lang=java for full help.)"; |
| 573 | |
| 574 | |
| 575 | static const char *Cobol_suffixes [] = |
| 576 | { "COB", "cob", NULL }; |
| 577 | static char Cobol_help [] = |
| 578 | "In Cobol code, tags are paragraph names; that is, any word\n\ |
| 579 | starting in column 8 and followed by a period."; |
| 580 | |
| 581 | static const char *Cstar_suffixes [] = |
| 582 | { "cs", "hs", NULL }; |
| 583 | |
| 584 | static const char *Erlang_suffixes [] = |
| 585 | { "erl", "hrl", NULL }; |
| 586 | static const char Erlang_help [] = |
| 587 | "In Erlang code, the tags are the functions, records and macros\n\ |
| 588 | defined in the file."; |
| 589 | |
| 590 | const char *Forth_suffixes [] = |
| 591 | { "fth", "tok", NULL }; |
| 592 | static const char Forth_help [] = |
| 593 | "In Forth code, tags are words defined by `:',\n\ |
| 594 | constant, code, create, defer, value, variable, buffer:, field."; |
| 595 | |
| 596 | static const char *Fortran_suffixes [] = |
| 597 | { "F", "f", "f90", "for", NULL }; |
| 598 | static const char Fortran_help [] = |
| 599 | "In Fortran code, functions, subroutines and block data are tags."; |
| 600 | |
| 601 | static const char *HTML_suffixes [] = |
| 602 | { "htm", "html", "shtml", NULL }; |
| 603 | static const char HTML_help [] = |
| 604 | "In HTML input files, the tags are the `title' and the `h1', `h2',\n\ |
| 605 | `h3' headers. Also, tags are `name=' in anchors and all\n\ |
| 606 | occurrences of `id='."; |
| 607 | |
| 608 | static const char *Lisp_suffixes [] = |
| 609 | { "cl", "clisp", "el", "l", "lisp", "LSP", "lsp", "ml", NULL }; |
| 610 | static const char Lisp_help [] = |
| 611 | "In Lisp code, any function defined with `defun', any variable\n\ |
| 612 | defined with `defvar' or `defconst', and in general the first\n\ |
| 613 | argument of any expression that starts with `(def' in column zero\n\ |
| 614 | is a tag.\n\ |
| 615 | The `--declarations' option tags \"(defvar foo)\" constructs too."; |
| 616 | |
| 617 | static const char *Lua_suffixes [] = |
| 618 | { "lua", "LUA", NULL }; |
| 619 | static const char Lua_help [] = |
| 620 | "In Lua scripts, all functions are tags."; |
| 621 | |
| 622 | static const char *Makefile_filenames [] = |
| 623 | { "Makefile", "makefile", "GNUMakefile", "Makefile.in", "Makefile.am", NULL}; |
| 624 | static const char Makefile_help [] = |
| 625 | "In makefiles, targets are tags; additionally, variables are tags\n\ |
| 626 | unless you specify `--no-globals'."; |
| 627 | |
| 628 | static const char *Objc_suffixes [] = |
| 629 | { "lm", /* Objective lex file */ |
| 630 | "m", /* Objective C file */ |
| 631 | NULL }; |
| 632 | static const char Objc_help [] = |
| 633 | "In Objective C code, tags include Objective C definitions for classes,\n\ |
| 634 | class categories, methods and protocols. Tags for variables and\n\ |
| 635 | functions in classes are named `CLASS::VARIABLE' and `CLASS::FUNCTION'.\n\ |
| 636 | (Use --help --lang=c --lang=objc --lang=java for full help.)"; |
| 637 | |
| 638 | static const char *Pascal_suffixes [] = |
| 639 | { "p", "pas", NULL }; |
| 640 | static const char Pascal_help [] = |
| 641 | "In Pascal code, the tags are the functions and procedures defined\n\ |
| 642 | in the file."; |
| 643 | /* " // this is for working around an Emacs highlighting bug... */ |
| 644 | |
| 645 | static const char *Perl_suffixes [] = |
| 646 | { "pl", "pm", NULL }; |
| 647 | static const char *Perl_interpreters [] = |
| 648 | { "perl", "@PERL@", NULL }; |
| 649 | static const char Perl_help [] = |
| 650 | "In Perl code, the tags are the packages, subroutines and variables\n\ |
| 651 | defined by the `package', `sub', `my' and `local' keywords. Use\n\ |
| 652 | `--globals' if you want to tag global variables. Tags for\n\ |
| 653 | subroutines are named `PACKAGE::SUB'. The name for subroutines\n\ |
| 654 | defined in the default package is `main::SUB'."; |
| 655 | |
| 656 | static const char *PHP_suffixes [] = |
| 657 | { "php", "php3", "php4", NULL }; |
| 658 | static const char PHP_help [] = |
| 659 | "In PHP code, tags are functions, classes and defines. Unless you use\n\ |
| 660 | the `--no-members' option, vars are tags too."; |
| 661 | |
| 662 | static const char *plain_C_suffixes [] = |
| 663 | { "pc", /* Pro*C file */ |
| 664 | NULL }; |
| 665 | |
| 666 | static const char *PS_suffixes [] = |
| 667 | { "ps", "psw", NULL }; /* .psw is for PSWrap */ |
| 668 | static const char PS_help [] = |
| 669 | "In PostScript code, the tags are the functions."; |
| 670 | |
| 671 | static const char *Prolog_suffixes [] = |
| 672 | { "prolog", NULL }; |
| 673 | static const char Prolog_help [] = |
| 674 | "In Prolog code, tags are predicates and rules at the beginning of\n\ |
| 675 | line."; |
| 676 | |
| 677 | static const char *Python_suffixes [] = |
| 678 | { "py", NULL }; |
| 679 | static const char Python_help [] = |
| 680 | "In Python code, `def' or `class' at the beginning of a line\n\ |
| 681 | generate a tag."; |
| 682 | |
| 683 | /* Can't do the `SCM' or `scm' prefix with a version number. */ |
| 684 | static const char *Scheme_suffixes [] = |
| 685 | { "oak", "sch", "scheme", "SCM", "scm", "SM", "sm", "ss", "t", NULL }; |
| 686 | static const char Scheme_help [] = |
| 687 | "In Scheme code, tags include anything defined with `def' or with a\n\ |
| 688 | construct whose name starts with `def'. They also include\n\ |
| 689 | variables set with `set!' at top level in the file."; |
| 690 | |
| 691 | static const char *TeX_suffixes [] = |
| 692 | { "bib", "clo", "cls", "ltx", "sty", "TeX", "tex", NULL }; |
| 693 | static const char TeX_help [] = |
| 694 | "In LaTeX text, the argument of any of the commands `\\chapter',\n\ |
| 695 | `\\section', `\\subsection', `\\subsubsection', `\\eqno', `\\label',\n\ |
| 696 | `\\ref', `\\cite', `\\bibitem', `\\part', `\\appendix', `\\entry',\n\ |
| 697 | `\\index', `\\def', `\\newcommand', `\\renewcommand',\n\ |
| 698 | `\\newenvironment' or `\\renewenvironment' is a tag.\n\ |
| 699 | \n\ |
| 700 | Other commands can be specified by setting the environment variable\n\ |
| 701 | `TEXTAGS' to a colon-separated list like, for example,\n\ |
| 702 | TEXTAGS=\"mycommand:myothercommand\"."; |
| 703 | |
| 704 | |
| 705 | static const char *Texinfo_suffixes [] = |
| 706 | { "texi", "texinfo", "txi", NULL }; |
| 707 | static const char Texinfo_help [] = |
| 708 | "for texinfo files, lines starting with @node are tagged."; |
| 709 | |
| 710 | static const char *Yacc_suffixes [] = |
| 711 | { "y", "y++", "ym", "yxx", "yy", NULL }; /* .ym is Objective yacc file */ |
| 712 | static const char Yacc_help [] = |
| 713 | "In Bison or Yacc input files, each rule defines as a tag the\n\ |
| 714 | nonterminal it constructs. The portions of the file that contain\n\ |
| 715 | C code are parsed as C code (use --help --lang=c --lang=yacc\n\ |
| 716 | for full help)."; |
| 717 | |
| 718 | static const char auto_help [] = |
| 719 | "`auto' is not a real language, it indicates to use\n\ |
| 720 | a default language for files base on file name suffix and file contents."; |
| 721 | |
| 722 | static const char none_help [] = |
| 723 | "`none' is not a real language, it indicates to only do\n\ |
| 724 | regexp processing on files."; |
| 725 | |
| 726 | static const char no_lang_help [] = |
| 727 | "No detailed help available for this language."; |
| 728 | |
| 729 | |
| 730 | /* |
| 731 | * Table of languages. |
| 732 | * |
| 733 | * It is ok for a given function to be listed under more than one |
| 734 | * name. I just didn't. |
| 735 | */ |
| 736 | |
| 737 | static language lang_names [] = |
| 738 | { |
| 739 | { "ada", Ada_help, Ada_funcs, Ada_suffixes }, |
| 740 | { "asm", Asm_help, Asm_labels, Asm_suffixes }, |
| 741 | { "c", default_C_help, default_C_entries, default_C_suffixes }, |
| 742 | { "c++", Cplusplus_help, Cplusplus_entries, Cplusplus_suffixes }, |
| 743 | { "c*", no_lang_help, Cstar_entries, Cstar_suffixes }, |
| 744 | { "cobol", Cobol_help, Cobol_paragraphs, Cobol_suffixes }, |
| 745 | { "erlang", Erlang_help, Erlang_functions, Erlang_suffixes }, |
| 746 | { "forth", Forth_help, Forth_words, Forth_suffixes }, |
| 747 | { "fortran", Fortran_help, Fortran_functions, Fortran_suffixes }, |
| 748 | { "html", HTML_help, HTML_labels, HTML_suffixes }, |
| 749 | { "java", Cjava_help, Cjava_entries, Cjava_suffixes }, |
| 750 | { "lisp", Lisp_help, Lisp_functions, Lisp_suffixes }, |
| 751 | { "lua", Lua_help, Lua_functions, Lua_suffixes }, |
| 752 | { "makefile", Makefile_help,Makefile_targets,NULL,Makefile_filenames}, |
| 753 | { "objc", Objc_help, plain_C_entries, Objc_suffixes }, |
| 754 | { "pascal", Pascal_help, Pascal_functions, Pascal_suffixes }, |
| 755 | { "perl",Perl_help,Perl_functions,Perl_suffixes,NULL,Perl_interpreters}, |
| 756 | { "php", PHP_help, PHP_functions, PHP_suffixes }, |
| 757 | { "postscript",PS_help, PS_functions, PS_suffixes }, |
| 758 | { "proc", no_lang_help, plain_C_entries, plain_C_suffixes }, |
| 759 | { "prolog", Prolog_help, Prolog_functions, Prolog_suffixes }, |
| 760 | { "python", Python_help, Python_functions, Python_suffixes }, |
| 761 | { "scheme", Scheme_help, Scheme_functions, Scheme_suffixes }, |
| 762 | { "tex", TeX_help, TeX_commands, TeX_suffixes }, |
| 763 | { "texinfo", Texinfo_help, Texinfo_nodes, Texinfo_suffixes }, |
| 764 | { "yacc", Yacc_help,Yacc_entries,Yacc_suffixes,NULL,NULL,TRUE}, |
| 765 | { "auto", auto_help }, /* default guessing scheme */ |
| 766 | { "none", none_help, just_read_file }, /* regexp matching only */ |
| 767 | { NULL } /* end of list */ |
| 768 | }; |
| 769 | |
| 770 | \f |
| 771 | static void |
| 772 | print_language_names (void) |
| 773 | { |
| 774 | language *lang; |
| 775 | const char **name, **ext; |
| 776 | |
| 777 | puts ("\nThese are the currently supported languages, along with the\n\ |
| 778 | default file names and dot suffixes:"); |
| 779 | for (lang = lang_names; lang->name != NULL; lang++) |
| 780 | { |
| 781 | printf (" %-*s", 10, lang->name); |
| 782 | if (lang->filenames != NULL) |
| 783 | for (name = lang->filenames; *name != NULL; name++) |
| 784 | printf (" %s", *name); |
| 785 | if (lang->suffixes != NULL) |
| 786 | for (ext = lang->suffixes; *ext != NULL; ext++) |
| 787 | printf (" .%s", *ext); |
| 788 | puts (""); |
| 789 | } |
| 790 | puts ("where `auto' means use default language for files based on file\n\ |
| 791 | name suffix, and `none' means only do regexp processing on files.\n\ |
| 792 | If no language is specified and no matching suffix is found,\n\ |
| 793 | the first line of the file is read for a sharp-bang (#!) sequence\n\ |
| 794 | followed by the name of an interpreter. If no such sequence is found,\n\ |
| 795 | Fortran is tried first; if no tags are found, C is tried next.\n\ |
| 796 | When parsing any C file, a \"class\" or \"template\" keyword\n\ |
| 797 | switches to C++."); |
| 798 | puts ("Compressed files are supported using gzip, bzip2, and xz.\n\ |
| 799 | \n\ |
| 800 | For detailed help on a given language use, for example,\n\ |
| 801 | etags --help --lang=ada."); |
| 802 | } |
| 803 | |
| 804 | #ifndef EMACS_NAME |
| 805 | # define EMACS_NAME "standalone" |
| 806 | #endif |
| 807 | #ifndef VERSION |
| 808 | # define VERSION "17.38.1.4" |
| 809 | #endif |
| 810 | static _Noreturn void |
| 811 | print_version (void) |
| 812 | { |
| 813 | char emacs_copyright[] = COPYRIGHT; |
| 814 | |
| 815 | printf ("%s (%s %s)\n", (CTAGS) ? "ctags" : "etags", EMACS_NAME, VERSION); |
| 816 | puts (emacs_copyright); |
| 817 | puts ("This program is distributed under the terms in ETAGS.README"); |
| 818 | |
| 819 | exit (EXIT_SUCCESS); |
| 820 | } |
| 821 | |
| 822 | #ifndef PRINT_UNDOCUMENTED_OPTIONS_HELP |
| 823 | # define PRINT_UNDOCUMENTED_OPTIONS_HELP FALSE |
| 824 | #endif |
| 825 | |
| 826 | static _Noreturn void |
| 827 | print_help (argument *argbuffer) |
| 828 | { |
| 829 | bool help_for_lang = FALSE; |
| 830 | |
| 831 | for (; argbuffer->arg_type != at_end; argbuffer++) |
| 832 | if (argbuffer->arg_type == at_language) |
| 833 | { |
| 834 | if (help_for_lang) |
| 835 | puts (""); |
| 836 | puts (argbuffer->lang->help); |
| 837 | help_for_lang = TRUE; |
| 838 | } |
| 839 | |
| 840 | if (help_for_lang) |
| 841 | exit (EXIT_SUCCESS); |
| 842 | |
| 843 | printf ("Usage: %s [options] [[regex-option ...] file-name] ...\n\ |
| 844 | \n\ |
| 845 | These are the options accepted by %s.\n", progname, progname); |
| 846 | puts ("You may use unambiguous abbreviations for the long option names."); |
| 847 | puts (" A - as file name means read names from stdin (one per line).\n\ |
| 848 | Absolute names are stored in the output file as they are.\n\ |
| 849 | Relative ones are stored relative to the output file's directory.\n"); |
| 850 | |
| 851 | puts ("-a, --append\n\ |
| 852 | Append tag entries to existing tags file."); |
| 853 | |
| 854 | puts ("--packages-only\n\ |
| 855 | For Ada files, only generate tags for packages."); |
| 856 | |
| 857 | if (CTAGS) |
| 858 | puts ("-B, --backward-search\n\ |
| 859 | Write the search commands for the tag entries using '?', the\n\ |
| 860 | backward-search command instead of '/', the forward-search command."); |
| 861 | |
| 862 | /* This option is mostly obsolete, because etags can now automatically |
| 863 | detect C++. Retained for backward compatibility and for debugging and |
| 864 | experimentation. In principle, we could want to tag as C++ even |
| 865 | before any "class" or "template" keyword. |
| 866 | puts ("-C, --c++\n\ |
| 867 | Treat files whose name suffix defaults to C language as C++ files."); |
| 868 | */ |
| 869 | |
| 870 | puts ("--declarations\n\ |
| 871 | In C and derived languages, create tags for function declarations,"); |
| 872 | if (CTAGS) |
| 873 | puts ("\tand create tags for extern variables if --globals is used."); |
| 874 | else |
| 875 | puts |
| 876 | ("\tand create tags for extern variables unless --no-globals is used."); |
| 877 | |
| 878 | if (CTAGS) |
| 879 | puts ("-d, --defines\n\ |
| 880 | Create tag entries for C #define constants and enum constants, too."); |
| 881 | else |
| 882 | puts ("-D, --no-defines\n\ |
| 883 | Don't create tag entries for C #define constants and enum constants.\n\ |
| 884 | This makes the tags file smaller."); |
| 885 | |
| 886 | if (!CTAGS) |
| 887 | puts ("-i FILE, --include=FILE\n\ |
| 888 | Include a note in tag file indicating that, when searching for\n\ |
| 889 | a tag, one should also consult the tags file FILE after\n\ |
| 890 | checking the current file."); |
| 891 | |
| 892 | puts ("-l LANG, --language=LANG\n\ |
| 893 | Force the following files to be considered as written in the\n\ |
| 894 | named language up to the next --language=LANG option."); |
| 895 | |
| 896 | if (CTAGS) |
| 897 | puts ("--globals\n\ |
| 898 | Create tag entries for global variables in some languages."); |
| 899 | else |
| 900 | puts ("--no-globals\n\ |
| 901 | Do not create tag entries for global variables in some\n\ |
| 902 | languages. This makes the tags file smaller."); |
| 903 | |
| 904 | if (PRINT_UNDOCUMENTED_OPTIONS_HELP) |
| 905 | puts ("--no-line-directive\n\ |
| 906 | Ignore #line preprocessor directives in C and derived languages."); |
| 907 | |
| 908 | if (CTAGS) |
| 909 | puts ("--members\n\ |
| 910 | Create tag entries for members of structures in some languages."); |
| 911 | else |
| 912 | puts ("--no-members\n\ |
| 913 | Do not create tag entries for members of structures\n\ |
| 914 | in some languages."); |
| 915 | |
| 916 | puts ("-r REGEXP, --regex=REGEXP or --regex=@regexfile\n\ |
| 917 | Make a tag for each line matching a regular expression pattern\n\ |
| 918 | in the following files. {LANGUAGE}REGEXP uses REGEXP for LANGUAGE\n\ |
| 919 | files only. REGEXFILE is a file containing one REGEXP per line.\n\ |
| 920 | REGEXP takes the form /TAGREGEXP/TAGNAME/MODS, where TAGNAME/ is\n\ |
| 921 | optional. The TAGREGEXP pattern is anchored (as if preceded by ^)."); |
| 922 | puts (" If TAGNAME/ is present, the tags created are named.\n\ |
| 923 | For example Tcl named tags can be created with:\n\ |
| 924 | --regex=\"/proc[ \\t]+\\([^ \\t]+\\)/\\1/.\".\n\ |
| 925 | MODS are optional one-letter modifiers: `i' means to ignore case,\n\ |
| 926 | `m' means to allow multi-line matches, `s' implies `m' and\n\ |
| 927 | causes dot to match any character, including newline."); |
| 928 | |
| 929 | puts ("-R, --no-regex\n\ |
| 930 | Don't create tags from regexps for the following files."); |
| 931 | |
| 932 | puts ("-I, --ignore-indentation\n\ |
| 933 | In C and C++ do not assume that a closing brace in the first\n\ |
| 934 | column is the final brace of a function or structure definition."); |
| 935 | |
| 936 | puts ("-o FILE, --output=FILE\n\ |
| 937 | Write the tags to FILE."); |
| 938 | |
| 939 | puts ("--parse-stdin=NAME\n\ |
| 940 | Read from standard input and record tags as belonging to file NAME."); |
| 941 | |
| 942 | if (CTAGS) |
| 943 | { |
| 944 | puts ("-t, --typedefs\n\ |
| 945 | Generate tag entries for C and Ada typedefs."); |
| 946 | puts ("-T, --typedefs-and-c++\n\ |
| 947 | Generate tag entries for C typedefs, C struct/enum/union tags,\n\ |
| 948 | and C++ member functions."); |
| 949 | } |
| 950 | |
| 951 | if (CTAGS) |
| 952 | puts ("-u, --update\n\ |
| 953 | Update the tag entries for the given files, leaving tag\n\ |
| 954 | entries for other files in place. Currently, this is\n\ |
| 955 | implemented by deleting the existing entries for the given\n\ |
| 956 | files and then rewriting the new entries at the end of the\n\ |
| 957 | tags file. It is often faster to simply rebuild the entire\n\ |
| 958 | tag file than to use this."); |
| 959 | |
| 960 | if (CTAGS) |
| 961 | { |
| 962 | puts ("-v, --vgrind\n\ |
| 963 | Print on the standard output an index of items intended for\n\ |
| 964 | human consumption, similar to the output of vgrind. The index\n\ |
| 965 | is sorted, and gives the page number of each item."); |
| 966 | |
| 967 | if (PRINT_UNDOCUMENTED_OPTIONS_HELP) |
| 968 | puts ("-w, --no-duplicates\n\ |
| 969 | Do not create duplicate tag entries, for compatibility with\n\ |
| 970 | traditional ctags."); |
| 971 | |
| 972 | if (PRINT_UNDOCUMENTED_OPTIONS_HELP) |
| 973 | puts ("-w, --no-warn\n\ |
| 974 | Suppress warning messages about duplicate tag entries."); |
| 975 | |
| 976 | puts ("-x, --cxref\n\ |
| 977 | Like --vgrind, but in the style of cxref, rather than vgrind.\n\ |
| 978 | The output uses line numbers instead of page numbers, but\n\ |
| 979 | beyond that the differences are cosmetic; try both to see\n\ |
| 980 | which you like."); |
| 981 | } |
| 982 | |
| 983 | puts ("-V, --version\n\ |
| 984 | Print the version of the program.\n\ |
| 985 | -h, --help\n\ |
| 986 | Print this help message.\n\ |
| 987 | Followed by one or more `--language' options prints detailed\n\ |
| 988 | help about tag generation for the specified languages."); |
| 989 | |
| 990 | print_language_names (); |
| 991 | |
| 992 | puts (""); |
| 993 | puts ("Report bugs to bug-gnu-emacs@gnu.org"); |
| 994 | |
| 995 | exit (EXIT_SUCCESS); |
| 996 | } |
| 997 | |
| 998 | \f |
| 999 | int |
| 1000 | main (int argc, char **argv) |
| 1001 | { |
| 1002 | int i; |
| 1003 | unsigned int nincluded_files; |
| 1004 | char **included_files; |
| 1005 | argument *argbuffer; |
| 1006 | int current_arg, file_count; |
| 1007 | linebuffer filename_lb; |
| 1008 | bool help_asked = FALSE; |
| 1009 | ptrdiff_t len; |
| 1010 | char *optstring; |
| 1011 | int opt; |
| 1012 | |
| 1013 | |
| 1014 | #ifdef DOS_NT |
| 1015 | _fmode = O_BINARY; /* all of files are treated as binary files */ |
| 1016 | #endif /* DOS_NT */ |
| 1017 | |
| 1018 | progname = argv[0]; |
| 1019 | nincluded_files = 0; |
| 1020 | included_files = xnew (argc, char *); |
| 1021 | current_arg = 0; |
| 1022 | file_count = 0; |
| 1023 | |
| 1024 | /* Allocate enough no matter what happens. Overkill, but each one |
| 1025 | is small. */ |
| 1026 | argbuffer = xnew (argc, argument); |
| 1027 | |
| 1028 | /* |
| 1029 | * Always find typedefs and structure tags. |
| 1030 | * Also default to find macro constants, enum constants, struct |
| 1031 | * members and global variables. Do it for both etags and ctags. |
| 1032 | */ |
| 1033 | typedefs = typedefs_or_cplusplus = constantypedefs = TRUE; |
| 1034 | globals = members = TRUE; |
| 1035 | |
| 1036 | /* When the optstring begins with a '-' getopt_long does not rearrange the |
| 1037 | non-options arguments to be at the end, but leaves them alone. */ |
| 1038 | optstring = concat ("-ac:Cf:Il:o:r:RSVhH", |
| 1039 | (CTAGS) ? "BxdtTuvw" : "Di:", |
| 1040 | ""); |
| 1041 | |
| 1042 | while ((opt = getopt_long (argc, argv, optstring, longopts, NULL)) != EOF) |
| 1043 | switch (opt) |
| 1044 | { |
| 1045 | case 0: |
| 1046 | /* If getopt returns 0, then it has already processed a |
| 1047 | long-named option. We should do nothing. */ |
| 1048 | break; |
| 1049 | |
| 1050 | case 1: |
| 1051 | /* This means that a file name has been seen. Record it. */ |
| 1052 | argbuffer[current_arg].arg_type = at_filename; |
| 1053 | argbuffer[current_arg].what = optarg; |
| 1054 | len = strlen (optarg); |
| 1055 | if (whatlen_max < len) |
| 1056 | whatlen_max = len; |
| 1057 | ++current_arg; |
| 1058 | ++file_count; |
| 1059 | break; |
| 1060 | |
| 1061 | case STDIN: |
| 1062 | /* Parse standard input. Idea by Vivek <vivek@etla.org>. */ |
| 1063 | argbuffer[current_arg].arg_type = at_stdin; |
| 1064 | argbuffer[current_arg].what = optarg; |
| 1065 | len = strlen (optarg); |
| 1066 | if (whatlen_max < len) |
| 1067 | whatlen_max = len; |
| 1068 | ++current_arg; |
| 1069 | ++file_count; |
| 1070 | if (parsing_stdin) |
| 1071 | fatal ("cannot parse standard input more than once", (char *)NULL); |
| 1072 | parsing_stdin = TRUE; |
| 1073 | break; |
| 1074 | |
| 1075 | /* Common options. */ |
| 1076 | case 'a': append_to_tagfile = TRUE; break; |
| 1077 | case 'C': cplusplus = TRUE; break; |
| 1078 | case 'f': /* for compatibility with old makefiles */ |
| 1079 | case 'o': |
| 1080 | if (tagfile) |
| 1081 | { |
| 1082 | error ("-o option may only be given once."); |
| 1083 | suggest_asking_for_help (); |
| 1084 | /* NOTREACHED */ |
| 1085 | } |
| 1086 | tagfile = optarg; |
| 1087 | break; |
| 1088 | case 'I': |
| 1089 | case 'S': /* for backward compatibility */ |
| 1090 | ignoreindent = TRUE; |
| 1091 | break; |
| 1092 | case 'l': |
| 1093 | { |
| 1094 | language *lang = get_language_from_langname (optarg); |
| 1095 | if (lang != NULL) |
| 1096 | { |
| 1097 | argbuffer[current_arg].lang = lang; |
| 1098 | argbuffer[current_arg].arg_type = at_language; |
| 1099 | ++current_arg; |
| 1100 | } |
| 1101 | } |
| 1102 | break; |
| 1103 | case 'c': |
| 1104 | /* Backward compatibility: support obsolete --ignore-case-regexp. */ |
| 1105 | optarg = concat (optarg, "i", ""); /* memory leak here */ |
| 1106 | /* FALLTHRU */ |
| 1107 | case 'r': |
| 1108 | argbuffer[current_arg].arg_type = at_regexp; |
| 1109 | argbuffer[current_arg].what = optarg; |
| 1110 | len = strlen (optarg); |
| 1111 | if (whatlen_max < len) |
| 1112 | whatlen_max = len; |
| 1113 | ++current_arg; |
| 1114 | break; |
| 1115 | case 'R': |
| 1116 | argbuffer[current_arg].arg_type = at_regexp; |
| 1117 | argbuffer[current_arg].what = NULL; |
| 1118 | ++current_arg; |
| 1119 | break; |
| 1120 | case 'V': |
| 1121 | print_version (); |
| 1122 | break; |
| 1123 | case 'h': |
| 1124 | case 'H': |
| 1125 | help_asked = TRUE; |
| 1126 | break; |
| 1127 | |
| 1128 | /* Etags options */ |
| 1129 | case 'D': constantypedefs = FALSE; break; |
| 1130 | case 'i': included_files[nincluded_files++] = optarg; break; |
| 1131 | |
| 1132 | /* Ctags options. */ |
| 1133 | case 'B': searchar = '?'; break; |
| 1134 | case 'd': constantypedefs = TRUE; break; |
| 1135 | case 't': typedefs = TRUE; break; |
| 1136 | case 'T': typedefs = typedefs_or_cplusplus = TRUE; break; |
| 1137 | case 'u': update = TRUE; break; |
| 1138 | case 'v': vgrind_style = TRUE; /*FALLTHRU*/ |
| 1139 | case 'x': cxref_style = TRUE; break; |
| 1140 | case 'w': no_warnings = TRUE; break; |
| 1141 | default: |
| 1142 | suggest_asking_for_help (); |
| 1143 | /* NOTREACHED */ |
| 1144 | } |
| 1145 | |
| 1146 | /* No more options. Store the rest of arguments. */ |
| 1147 | for (; optind < argc; optind++) |
| 1148 | { |
| 1149 | argbuffer[current_arg].arg_type = at_filename; |
| 1150 | argbuffer[current_arg].what = argv[optind]; |
| 1151 | len = strlen (argv[optind]); |
| 1152 | if (whatlen_max < len) |
| 1153 | whatlen_max = len; |
| 1154 | ++current_arg; |
| 1155 | ++file_count; |
| 1156 | } |
| 1157 | |
| 1158 | argbuffer[current_arg].arg_type = at_end; |
| 1159 | |
| 1160 | if (help_asked) |
| 1161 | print_help (argbuffer); |
| 1162 | /* NOTREACHED */ |
| 1163 | |
| 1164 | if (nincluded_files == 0 && file_count == 0) |
| 1165 | { |
| 1166 | error ("no input files specified."); |
| 1167 | suggest_asking_for_help (); |
| 1168 | /* NOTREACHED */ |
| 1169 | } |
| 1170 | |
| 1171 | if (tagfile == NULL) |
| 1172 | tagfile = savestr (CTAGS ? "tags" : "TAGS"); |
| 1173 | cwd = etags_getcwd (); /* the current working directory */ |
| 1174 | if (cwd[strlen (cwd) - 1] != '/') |
| 1175 | { |
| 1176 | char *oldcwd = cwd; |
| 1177 | cwd = concat (oldcwd, "/", ""); |
| 1178 | free (oldcwd); |
| 1179 | } |
| 1180 | |
| 1181 | /* Compute base directory for relative file names. */ |
| 1182 | if (streq (tagfile, "-") |
| 1183 | || strneq (tagfile, "/dev/", 5)) |
| 1184 | tagfiledir = cwd; /* relative file names are relative to cwd */ |
| 1185 | else |
| 1186 | { |
| 1187 | canonicalize_filename (tagfile); |
| 1188 | tagfiledir = absolute_dirname (tagfile, cwd); |
| 1189 | } |
| 1190 | |
| 1191 | init (); /* set up boolean "functions" */ |
| 1192 | |
| 1193 | linebuffer_init (&lb); |
| 1194 | linebuffer_init (&filename_lb); |
| 1195 | linebuffer_init (&filebuf); |
| 1196 | linebuffer_init (&token_name); |
| 1197 | |
| 1198 | if (!CTAGS) |
| 1199 | { |
| 1200 | if (streq (tagfile, "-")) |
| 1201 | { |
| 1202 | tagf = stdout; |
| 1203 | #ifdef DOS_NT |
| 1204 | /* Switch redirected `stdout' to binary mode (setting `_fmode' |
| 1205 | doesn't take effect until after `stdout' is already open). */ |
| 1206 | if (!isatty (fileno (stdout))) |
| 1207 | setmode (fileno (stdout), O_BINARY); |
| 1208 | #endif /* DOS_NT */ |
| 1209 | } |
| 1210 | else |
| 1211 | tagf = fopen (tagfile, append_to_tagfile ? "a" : "w"); |
| 1212 | if (tagf == NULL) |
| 1213 | pfatal (tagfile); |
| 1214 | } |
| 1215 | |
| 1216 | /* |
| 1217 | * Loop through files finding functions. |
| 1218 | */ |
| 1219 | for (i = 0; i < current_arg; i++) |
| 1220 | { |
| 1221 | static language *lang; /* non-NULL if language is forced */ |
| 1222 | char *this_file; |
| 1223 | |
| 1224 | switch (argbuffer[i].arg_type) |
| 1225 | { |
| 1226 | case at_language: |
| 1227 | lang = argbuffer[i].lang; |
| 1228 | break; |
| 1229 | case at_regexp: |
| 1230 | analyse_regex (argbuffer[i].what); |
| 1231 | break; |
| 1232 | case at_filename: |
| 1233 | this_file = argbuffer[i].what; |
| 1234 | /* Input file named "-" means read file names from stdin |
| 1235 | (one per line) and use them. */ |
| 1236 | if (streq (this_file, "-")) |
| 1237 | { |
| 1238 | if (parsing_stdin) |
| 1239 | fatal ("cannot parse standard input AND read file names from it", |
| 1240 | (char *)NULL); |
| 1241 | while (readline_internal (&filename_lb, stdin) > 0) |
| 1242 | process_file_name (filename_lb.buffer, lang); |
| 1243 | } |
| 1244 | else |
| 1245 | process_file_name (this_file, lang); |
| 1246 | break; |
| 1247 | case at_stdin: |
| 1248 | this_file = argbuffer[i].what; |
| 1249 | process_file (stdin, this_file, lang); |
| 1250 | break; |
| 1251 | } |
| 1252 | } |
| 1253 | |
| 1254 | free_regexps (); |
| 1255 | free (lb.buffer); |
| 1256 | free (filebuf.buffer); |
| 1257 | free (token_name.buffer); |
| 1258 | |
| 1259 | if (!CTAGS || cxref_style) |
| 1260 | { |
| 1261 | /* Write the remaining tags to tagf (ETAGS) or stdout (CXREF). */ |
| 1262 | put_entries (nodehead); |
| 1263 | free_tree (nodehead); |
| 1264 | nodehead = NULL; |
| 1265 | if (!CTAGS) |
| 1266 | { |
| 1267 | fdesc *fdp; |
| 1268 | |
| 1269 | /* Output file entries that have no tags. */ |
| 1270 | for (fdp = fdhead; fdp != NULL; fdp = fdp->next) |
| 1271 | if (!fdp->written) |
| 1272 | fprintf (tagf, "\f\n%s,0\n", fdp->taggedfname); |
| 1273 | |
| 1274 | while (nincluded_files-- > 0) |
| 1275 | fprintf (tagf, "\f\n%s,include\n", *included_files++); |
| 1276 | |
| 1277 | if (fclose (tagf) == EOF) |
| 1278 | pfatal (tagfile); |
| 1279 | } |
| 1280 | |
| 1281 | exit (EXIT_SUCCESS); |
| 1282 | } |
| 1283 | |
| 1284 | /* From here on, we are in (CTAGS && !cxref_style) */ |
| 1285 | if (update) |
| 1286 | { |
| 1287 | char *cmd = |
| 1288 | xmalloc (strlen (tagfile) + whatlen_max + |
| 1289 | sizeof "mv..OTAGS;fgrep -v '\t\t' OTAGS >;rm OTAGS"); |
| 1290 | for (i = 0; i < current_arg; ++i) |
| 1291 | { |
| 1292 | switch (argbuffer[i].arg_type) |
| 1293 | { |
| 1294 | case at_filename: |
| 1295 | case at_stdin: |
| 1296 | break; |
| 1297 | default: |
| 1298 | continue; /* the for loop */ |
| 1299 | } |
| 1300 | strcpy (cmd, "mv "); |
| 1301 | strcat (cmd, tagfile); |
| 1302 | strcat (cmd, " OTAGS;fgrep -v '\t"); |
| 1303 | strcat (cmd, argbuffer[i].what); |
| 1304 | strcat (cmd, "\t' OTAGS >"); |
| 1305 | strcat (cmd, tagfile); |
| 1306 | strcat (cmd, ";rm OTAGS"); |
| 1307 | if (system (cmd) != EXIT_SUCCESS) |
| 1308 | fatal ("failed to execute shell command", (char *)NULL); |
| 1309 | } |
| 1310 | free (cmd); |
| 1311 | append_to_tagfile = TRUE; |
| 1312 | } |
| 1313 | |
| 1314 | tagf = fopen (tagfile, append_to_tagfile ? "a" : "w"); |
| 1315 | if (tagf == NULL) |
| 1316 | pfatal (tagfile); |
| 1317 | put_entries (nodehead); /* write all the tags (CTAGS) */ |
| 1318 | free_tree (nodehead); |
| 1319 | nodehead = NULL; |
| 1320 | if (fclose (tagf) == EOF) |
| 1321 | pfatal (tagfile); |
| 1322 | |
| 1323 | if (CTAGS) |
| 1324 | if (append_to_tagfile || update) |
| 1325 | { |
| 1326 | char *cmd = xmalloc (2 * strlen (tagfile) + sizeof "sort -u -o.."); |
| 1327 | /* Maybe these should be used: |
| 1328 | setenv ("LC_COLLATE", "C", 1); |
| 1329 | setenv ("LC_ALL", "C", 1); */ |
| 1330 | strcpy (cmd, "sort -u -o "); |
| 1331 | strcat (cmd, tagfile); |
| 1332 | strcat (cmd, " "); |
| 1333 | strcat (cmd, tagfile); |
| 1334 | exit (system (cmd)); |
| 1335 | } |
| 1336 | return EXIT_SUCCESS; |
| 1337 | } |
| 1338 | |
| 1339 | |
| 1340 | /* |
| 1341 | * Return a compressor given the file name. If EXTPTR is non-zero, |
| 1342 | * return a pointer into FILE where the compressor-specific |
| 1343 | * extension begins. If no compressor is found, NULL is returned |
| 1344 | * and EXTPTR is not significant. |
| 1345 | * Idea by Vladimir Alexiev <vladimir@cs.ualberta.ca> (1998) |
| 1346 | */ |
| 1347 | static compressor * |
| 1348 | get_compressor_from_suffix (char *file, char **extptr) |
| 1349 | { |
| 1350 | compressor *compr; |
| 1351 | char *slash, *suffix; |
| 1352 | |
| 1353 | /* File has been processed by canonicalize_filename, |
| 1354 | so we don't need to consider backslashes on DOS_NT. */ |
| 1355 | slash = etags_strrchr (file, '/'); |
| 1356 | suffix = etags_strrchr (file, '.'); |
| 1357 | if (suffix == NULL || suffix < slash) |
| 1358 | return NULL; |
| 1359 | if (extptr != NULL) |
| 1360 | *extptr = suffix; |
| 1361 | suffix += 1; |
| 1362 | /* Let those poor souls who live with DOS 8+3 file name limits get |
| 1363 | some solace by treating foo.cgz as if it were foo.c.gz, etc. |
| 1364 | Only the first do loop is run if not MSDOS */ |
| 1365 | do |
| 1366 | { |
| 1367 | for (compr = compressors; compr->suffix != NULL; compr++) |
| 1368 | if (streq (compr->suffix, suffix)) |
| 1369 | return compr; |
| 1370 | if (!MSDOS) |
| 1371 | break; /* do it only once: not really a loop */ |
| 1372 | if (extptr != NULL) |
| 1373 | *extptr = ++suffix; |
| 1374 | } while (*suffix != '\0'); |
| 1375 | return NULL; |
| 1376 | } |
| 1377 | |
| 1378 | |
| 1379 | |
| 1380 | /* |
| 1381 | * Return a language given the name. |
| 1382 | */ |
| 1383 | static language * |
| 1384 | get_language_from_langname (const char *name) |
| 1385 | { |
| 1386 | language *lang; |
| 1387 | |
| 1388 | if (name == NULL) |
| 1389 | error ("empty language name"); |
| 1390 | else |
| 1391 | { |
| 1392 | for (lang = lang_names; lang->name != NULL; lang++) |
| 1393 | if (streq (name, lang->name)) |
| 1394 | return lang; |
| 1395 | error ("unknown language \"%s\"", name); |
| 1396 | } |
| 1397 | |
| 1398 | return NULL; |
| 1399 | } |
| 1400 | |
| 1401 | |
| 1402 | /* |
| 1403 | * Return a language given the interpreter name. |
| 1404 | */ |
| 1405 | static language * |
| 1406 | get_language_from_interpreter (char *interpreter) |
| 1407 | { |
| 1408 | language *lang; |
| 1409 | const char **iname; |
| 1410 | |
| 1411 | if (interpreter == NULL) |
| 1412 | return NULL; |
| 1413 | for (lang = lang_names; lang->name != NULL; lang++) |
| 1414 | if (lang->interpreters != NULL) |
| 1415 | for (iname = lang->interpreters; *iname != NULL; iname++) |
| 1416 | if (streq (*iname, interpreter)) |
| 1417 | return lang; |
| 1418 | |
| 1419 | return NULL; |
| 1420 | } |
| 1421 | |
| 1422 | |
| 1423 | |
| 1424 | /* |
| 1425 | * Return a language given the file name. |
| 1426 | */ |
| 1427 | static language * |
| 1428 | get_language_from_filename (char *file, int case_sensitive) |
| 1429 | { |
| 1430 | language *lang; |
| 1431 | const char **name, **ext, *suffix; |
| 1432 | |
| 1433 | /* Try whole file name first. */ |
| 1434 | for (lang = lang_names; lang->name != NULL; lang++) |
| 1435 | if (lang->filenames != NULL) |
| 1436 | for (name = lang->filenames; *name != NULL; name++) |
| 1437 | if ((case_sensitive) |
| 1438 | ? streq (*name, file) |
| 1439 | : strcaseeq (*name, file)) |
| 1440 | return lang; |
| 1441 | |
| 1442 | /* If not found, try suffix after last dot. */ |
| 1443 | suffix = etags_strrchr (file, '.'); |
| 1444 | if (suffix == NULL) |
| 1445 | return NULL; |
| 1446 | suffix += 1; |
| 1447 | for (lang = lang_names; lang->name != NULL; lang++) |
| 1448 | if (lang->suffixes != NULL) |
| 1449 | for (ext = lang->suffixes; *ext != NULL; ext++) |
| 1450 | if ((case_sensitive) |
| 1451 | ? streq (*ext, suffix) |
| 1452 | : strcaseeq (*ext, suffix)) |
| 1453 | return lang; |
| 1454 | return NULL; |
| 1455 | } |
| 1456 | |
| 1457 | \f |
| 1458 | /* |
| 1459 | * This routine is called on each file argument. |
| 1460 | */ |
| 1461 | static void |
| 1462 | process_file_name (char *file, language *lang) |
| 1463 | { |
| 1464 | struct stat stat_buf; |
| 1465 | FILE *inf; |
| 1466 | fdesc *fdp; |
| 1467 | compressor *compr; |
| 1468 | char *compressed_name, *uncompressed_name; |
| 1469 | char *ext, *real_name; |
| 1470 | int retval; |
| 1471 | |
| 1472 | canonicalize_filename (file); |
| 1473 | if (streq (file, tagfile) && !streq (tagfile, "-")) |
| 1474 | { |
| 1475 | error ("skipping inclusion of %s in self.", file); |
| 1476 | return; |
| 1477 | } |
| 1478 | if ((compr = get_compressor_from_suffix (file, &ext)) == NULL) |
| 1479 | { |
| 1480 | compressed_name = NULL; |
| 1481 | real_name = uncompressed_name = savestr (file); |
| 1482 | } |
| 1483 | else |
| 1484 | { |
| 1485 | real_name = compressed_name = savestr (file); |
| 1486 | uncompressed_name = savenstr (file, ext - file); |
| 1487 | } |
| 1488 | |
| 1489 | /* If the canonicalized uncompressed name |
| 1490 | has already been dealt with, skip it silently. */ |
| 1491 | for (fdp = fdhead; fdp != NULL; fdp = fdp->next) |
| 1492 | { |
| 1493 | assert (fdp->infname != NULL); |
| 1494 | if (streq (uncompressed_name, fdp->infname)) |
| 1495 | goto cleanup; |
| 1496 | } |
| 1497 | |
| 1498 | if (stat (real_name, &stat_buf) != 0) |
| 1499 | { |
| 1500 | /* Reset real_name and try with a different name. */ |
| 1501 | real_name = NULL; |
| 1502 | if (compressed_name != NULL) /* try with the given suffix */ |
| 1503 | { |
| 1504 | if (stat (uncompressed_name, &stat_buf) == 0) |
| 1505 | real_name = uncompressed_name; |
| 1506 | } |
| 1507 | else /* try all possible suffixes */ |
| 1508 | { |
| 1509 | for (compr = compressors; compr->suffix != NULL; compr++) |
| 1510 | { |
| 1511 | compressed_name = concat (file, ".", compr->suffix); |
| 1512 | if (stat (compressed_name, &stat_buf) != 0) |
| 1513 | { |
| 1514 | if (MSDOS) |
| 1515 | { |
| 1516 | char *suf = compressed_name + strlen (file); |
| 1517 | size_t suflen = strlen (compr->suffix) + 1; |
| 1518 | for ( ; suf[1]; suf++, suflen--) |
| 1519 | { |
| 1520 | memmove (suf, suf + 1, suflen); |
| 1521 | if (stat (compressed_name, &stat_buf) == 0) |
| 1522 | { |
| 1523 | real_name = compressed_name; |
| 1524 | break; |
| 1525 | } |
| 1526 | } |
| 1527 | if (real_name != NULL) |
| 1528 | break; |
| 1529 | } /* MSDOS */ |
| 1530 | free (compressed_name); |
| 1531 | compressed_name = NULL; |
| 1532 | } |
| 1533 | else |
| 1534 | { |
| 1535 | real_name = compressed_name; |
| 1536 | break; |
| 1537 | } |
| 1538 | } |
| 1539 | } |
| 1540 | if (real_name == NULL) |
| 1541 | { |
| 1542 | perror (file); |
| 1543 | goto cleanup; |
| 1544 | } |
| 1545 | } /* try with a different name */ |
| 1546 | |
| 1547 | if (!S_ISREG (stat_buf.st_mode)) |
| 1548 | { |
| 1549 | error ("skipping %s: it is not a regular file.", real_name); |
| 1550 | goto cleanup; |
| 1551 | } |
| 1552 | if (real_name == compressed_name) |
| 1553 | { |
| 1554 | char *cmd = concat (compr->command, " ", real_name); |
| 1555 | inf = (FILE *) popen (cmd, "r"); |
| 1556 | free (cmd); |
| 1557 | } |
| 1558 | else |
| 1559 | inf = fopen (real_name, "r"); |
| 1560 | if (inf == NULL) |
| 1561 | { |
| 1562 | perror (real_name); |
| 1563 | goto cleanup; |
| 1564 | } |
| 1565 | |
| 1566 | process_file (inf, uncompressed_name, lang); |
| 1567 | |
| 1568 | if (real_name == compressed_name) |
| 1569 | retval = pclose (inf); |
| 1570 | else |
| 1571 | retval = fclose (inf); |
| 1572 | if (retval < 0) |
| 1573 | pfatal (file); |
| 1574 | |
| 1575 | cleanup: |
| 1576 | free (compressed_name); |
| 1577 | free (uncompressed_name); |
| 1578 | last_node = NULL; |
| 1579 | curfdp = NULL; |
| 1580 | return; |
| 1581 | } |
| 1582 | |
| 1583 | static void |
| 1584 | process_file (FILE *fh, char *fn, language *lang) |
| 1585 | { |
| 1586 | static const fdesc emptyfdesc; |
| 1587 | fdesc *fdp; |
| 1588 | |
| 1589 | /* Create a new input file description entry. */ |
| 1590 | fdp = xnew (1, fdesc); |
| 1591 | *fdp = emptyfdesc; |
| 1592 | fdp->next = fdhead; |
| 1593 | fdp->infname = savestr (fn); |
| 1594 | fdp->lang = lang; |
| 1595 | fdp->infabsname = absolute_filename (fn, cwd); |
| 1596 | fdp->infabsdir = absolute_dirname (fn, cwd); |
| 1597 | if (filename_is_absolute (fn)) |
| 1598 | { |
| 1599 | /* An absolute file name. Canonicalize it. */ |
| 1600 | fdp->taggedfname = absolute_filename (fn, NULL); |
| 1601 | } |
| 1602 | else |
| 1603 | { |
| 1604 | /* A file name relative to cwd. Make it relative |
| 1605 | to the directory of the tags file. */ |
| 1606 | fdp->taggedfname = relative_filename (fn, tagfiledir); |
| 1607 | } |
| 1608 | fdp->usecharno = TRUE; /* use char position when making tags */ |
| 1609 | fdp->prop = NULL; |
| 1610 | fdp->written = FALSE; /* not written on tags file yet */ |
| 1611 | |
| 1612 | fdhead = fdp; |
| 1613 | curfdp = fdhead; /* the current file description */ |
| 1614 | |
| 1615 | find_entries (fh); |
| 1616 | |
| 1617 | /* If not Ctags, and if this is not metasource and if it contained no #line |
| 1618 | directives, we can write the tags and free all nodes pointing to |
| 1619 | curfdp. */ |
| 1620 | if (!CTAGS |
| 1621 | && curfdp->usecharno /* no #line directives in this file */ |
| 1622 | && !curfdp->lang->metasource) |
| 1623 | { |
| 1624 | node *np, *prev; |
| 1625 | |
| 1626 | /* Look for the head of the sublist relative to this file. See add_node |
| 1627 | for the structure of the node tree. */ |
| 1628 | prev = NULL; |
| 1629 | for (np = nodehead; np != NULL; prev = np, np = np->left) |
| 1630 | if (np->fdp == curfdp) |
| 1631 | break; |
| 1632 | |
| 1633 | /* If we generated tags for this file, write and delete them. */ |
| 1634 | if (np != NULL) |
| 1635 | { |
| 1636 | /* This is the head of the last sublist, if any. The following |
| 1637 | instructions depend on this being true. */ |
| 1638 | assert (np->left == NULL); |
| 1639 | |
| 1640 | assert (fdhead == curfdp); |
| 1641 | assert (last_node->fdp == curfdp); |
| 1642 | put_entries (np); /* write tags for file curfdp->taggedfname */ |
| 1643 | free_tree (np); /* remove the written nodes */ |
| 1644 | if (prev == NULL) |
| 1645 | nodehead = NULL; /* no nodes left */ |
| 1646 | else |
| 1647 | prev->left = NULL; /* delete the pointer to the sublist */ |
| 1648 | } |
| 1649 | } |
| 1650 | } |
| 1651 | |
| 1652 | /* |
| 1653 | * This routine sets up the boolean pseudo-functions which work |
| 1654 | * by setting boolean flags dependent upon the corresponding character. |
| 1655 | * Every char which is NOT in that string is not a white char. Therefore, |
| 1656 | * all of the array "_wht" is set to FALSE, and then the elements |
| 1657 | * subscripted by the chars in "white" are set to TRUE. Thus "_wht" |
| 1658 | * of a char is TRUE if it is the string "white", else FALSE. |
| 1659 | */ |
| 1660 | static void |
| 1661 | init (void) |
| 1662 | { |
| 1663 | register const char *sp; |
| 1664 | register int i; |
| 1665 | |
| 1666 | for (i = 0; i < CHARS; i++) |
| 1667 | iswhite (i) = notinname (i) = begtoken (i) = intoken (i) = endtoken (i) = FALSE; |
| 1668 | for (sp = white; *sp != '\0'; sp++) iswhite (*sp) = TRUE; |
| 1669 | for (sp = nonam; *sp != '\0'; sp++) notinname (*sp) = TRUE; |
| 1670 | notinname ('\0') = notinname ('\n'); |
| 1671 | for (sp = begtk; *sp != '\0'; sp++) begtoken (*sp) = TRUE; |
| 1672 | begtoken ('\0') = begtoken ('\n'); |
| 1673 | for (sp = midtk; *sp != '\0'; sp++) intoken (*sp) = TRUE; |
| 1674 | intoken ('\0') = intoken ('\n'); |
| 1675 | for (sp = endtk; *sp != '\0'; sp++) endtoken (*sp) = TRUE; |
| 1676 | endtoken ('\0') = endtoken ('\n'); |
| 1677 | } |
| 1678 | |
| 1679 | /* |
| 1680 | * This routine opens the specified file and calls the function |
| 1681 | * which finds the function and type definitions. |
| 1682 | */ |
| 1683 | static void |
| 1684 | find_entries (FILE *inf) |
| 1685 | { |
| 1686 | char *cp; |
| 1687 | language *lang = curfdp->lang; |
| 1688 | Lang_function *parser = NULL; |
| 1689 | |
| 1690 | /* If user specified a language, use it. */ |
| 1691 | if (lang != NULL && lang->function != NULL) |
| 1692 | { |
| 1693 | parser = lang->function; |
| 1694 | } |
| 1695 | |
| 1696 | /* Else try to guess the language given the file name. */ |
| 1697 | if (parser == NULL) |
| 1698 | { |
| 1699 | lang = get_language_from_filename (curfdp->infname, TRUE); |
| 1700 | if (lang != NULL && lang->function != NULL) |
| 1701 | { |
| 1702 | curfdp->lang = lang; |
| 1703 | parser = lang->function; |
| 1704 | } |
| 1705 | } |
| 1706 | |
| 1707 | /* Else look for sharp-bang as the first two characters. */ |
| 1708 | if (parser == NULL |
| 1709 | && readline_internal (&lb, inf) > 0 |
| 1710 | && lb.len >= 2 |
| 1711 | && lb.buffer[0] == '#' |
| 1712 | && lb.buffer[1] == '!') |
| 1713 | { |
| 1714 | char *lp; |
| 1715 | |
| 1716 | /* Set lp to point at the first char after the last slash in the |
| 1717 | line or, if no slashes, at the first nonblank. Then set cp to |
| 1718 | the first successive blank and terminate the string. */ |
| 1719 | lp = etags_strrchr (lb.buffer+2, '/'); |
| 1720 | if (lp != NULL) |
| 1721 | lp += 1; |
| 1722 | else |
| 1723 | lp = skip_spaces (lb.buffer + 2); |
| 1724 | cp = skip_non_spaces (lp); |
| 1725 | *cp = '\0'; |
| 1726 | |
| 1727 | if (strlen (lp) > 0) |
| 1728 | { |
| 1729 | lang = get_language_from_interpreter (lp); |
| 1730 | if (lang != NULL && lang->function != NULL) |
| 1731 | { |
| 1732 | curfdp->lang = lang; |
| 1733 | parser = lang->function; |
| 1734 | } |
| 1735 | } |
| 1736 | } |
| 1737 | |
| 1738 | /* We rewind here, even if inf may be a pipe. We fail if the |
| 1739 | length of the first line is longer than the pipe block size, |
| 1740 | which is unlikely. */ |
| 1741 | rewind (inf); |
| 1742 | |
| 1743 | /* Else try to guess the language given the case insensitive file name. */ |
| 1744 | if (parser == NULL) |
| 1745 | { |
| 1746 | lang = get_language_from_filename (curfdp->infname, FALSE); |
| 1747 | if (lang != NULL && lang->function != NULL) |
| 1748 | { |
| 1749 | curfdp->lang = lang; |
| 1750 | parser = lang->function; |
| 1751 | } |
| 1752 | } |
| 1753 | |
| 1754 | /* Else try Fortran or C. */ |
| 1755 | if (parser == NULL) |
| 1756 | { |
| 1757 | node *old_last_node = last_node; |
| 1758 | |
| 1759 | curfdp->lang = get_language_from_langname ("fortran"); |
| 1760 | find_entries (inf); |
| 1761 | |
| 1762 | if (old_last_node == last_node) |
| 1763 | /* No Fortran entries found. Try C. */ |
| 1764 | { |
| 1765 | /* We do not tag if rewind fails. |
| 1766 | Only the file name will be recorded in the tags file. */ |
| 1767 | rewind (inf); |
| 1768 | curfdp->lang = get_language_from_langname (cplusplus ? "c++" : "c"); |
| 1769 | find_entries (inf); |
| 1770 | } |
| 1771 | return; |
| 1772 | } |
| 1773 | |
| 1774 | if (!no_line_directive |
| 1775 | && curfdp->lang != NULL && curfdp->lang->metasource) |
| 1776 | /* It may be that this is a bingo.y file, and we already parsed a bingo.c |
| 1777 | file, or anyway we parsed a file that is automatically generated from |
| 1778 | this one. If this is the case, the bingo.c file contained #line |
| 1779 | directives that generated tags pointing to this file. Let's delete |
| 1780 | them all before parsing this file, which is the real source. */ |
| 1781 | { |
| 1782 | fdesc **fdpp = &fdhead; |
| 1783 | while (*fdpp != NULL) |
| 1784 | if (*fdpp != curfdp |
| 1785 | && streq ((*fdpp)->taggedfname, curfdp->taggedfname)) |
| 1786 | /* We found one of those! We must delete both the file description |
| 1787 | and all tags referring to it. */ |
| 1788 | { |
| 1789 | fdesc *badfdp = *fdpp; |
| 1790 | |
| 1791 | /* Delete the tags referring to badfdp->taggedfname |
| 1792 | that were obtained from badfdp->infname. */ |
| 1793 | invalidate_nodes (badfdp, &nodehead); |
| 1794 | |
| 1795 | *fdpp = badfdp->next; /* remove the bad description from the list */ |
| 1796 | free_fdesc (badfdp); |
| 1797 | } |
| 1798 | else |
| 1799 | fdpp = &(*fdpp)->next; /* advance the list pointer */ |
| 1800 | } |
| 1801 | |
| 1802 | assert (parser != NULL); |
| 1803 | |
| 1804 | /* Generic initializations before reading from file. */ |
| 1805 | linebuffer_setlen (&filebuf, 0); /* reset the file buffer */ |
| 1806 | |
| 1807 | /* Generic initializations before parsing file with readline. */ |
| 1808 | lineno = 0; /* reset global line number */ |
| 1809 | charno = 0; /* reset global char number */ |
| 1810 | linecharno = 0; /* reset global char number of line start */ |
| 1811 | |
| 1812 | parser (inf); |
| 1813 | |
| 1814 | regex_tag_multiline (); |
| 1815 | } |
| 1816 | |
| 1817 | \f |
| 1818 | /* |
| 1819 | * Check whether an implicitly named tag should be created, |
| 1820 | * then call `pfnote'. |
| 1821 | * NAME is a string that is internally copied by this function. |
| 1822 | * |
| 1823 | * TAGS format specification |
| 1824 | * Idea by Sam Kendall <kendall@mv.mv.com> (1997) |
| 1825 | * The following is explained in some more detail in etc/ETAGS.EBNF. |
| 1826 | * |
| 1827 | * make_tag creates tags with "implicit tag names" (unnamed tags) |
| 1828 | * if the following are all true, assuming NONAM=" \f\t\n\r()=,;": |
| 1829 | * 1. NAME does not contain any of the characters in NONAM; |
| 1830 | * 2. LINESTART contains name as either a rightmost, or rightmost but |
| 1831 | * one character, substring; |
| 1832 | * 3. the character, if any, immediately before NAME in LINESTART must |
| 1833 | * be a character in NONAM; |
| 1834 | * 4. the character, if any, immediately after NAME in LINESTART must |
| 1835 | * also be a character in NONAM. |
| 1836 | * |
| 1837 | * The implementation uses the notinname() macro, which recognizes the |
| 1838 | * characters stored in the string `nonam'. |
| 1839 | * etags.el needs to use the same characters that are in NONAM. |
| 1840 | */ |
| 1841 | static void |
| 1842 | make_tag (const char *name, /* tag name, or NULL if unnamed */ |
| 1843 | int namelen, /* tag length */ |
| 1844 | int is_func, /* tag is a function */ |
| 1845 | char *linestart, /* start of the line where tag is */ |
| 1846 | int linelen, /* length of the line where tag is */ |
| 1847 | int lno, /* line number */ |
| 1848 | long int cno) /* character number */ |
| 1849 | { |
| 1850 | bool named = (name != NULL && namelen > 0); |
| 1851 | char *nname = NULL; |
| 1852 | |
| 1853 | if (!CTAGS && named) /* maybe set named to false */ |
| 1854 | /* Let's try to make an implicit tag name, that is, create an unnamed tag |
| 1855 | such that etags.el can guess a name from it. */ |
| 1856 | { |
| 1857 | int i; |
| 1858 | register const char *cp = name; |
| 1859 | |
| 1860 | for (i = 0; i < namelen; i++) |
| 1861 | if (notinname (*cp++)) |
| 1862 | break; |
| 1863 | if (i == namelen) /* rule #1 */ |
| 1864 | { |
| 1865 | cp = linestart + linelen - namelen; |
| 1866 | if (notinname (linestart[linelen-1])) |
| 1867 | cp -= 1; /* rule #4 */ |
| 1868 | if (cp >= linestart /* rule #2 */ |
| 1869 | && (cp == linestart |
| 1870 | || notinname (cp[-1])) /* rule #3 */ |
| 1871 | && strneq (name, cp, namelen)) /* rule #2 */ |
| 1872 | named = FALSE; /* use implicit tag name */ |
| 1873 | } |
| 1874 | } |
| 1875 | |
| 1876 | if (named) |
| 1877 | nname = savenstr (name, namelen); |
| 1878 | |
| 1879 | pfnote (nname, is_func, linestart, linelen, lno, cno); |
| 1880 | } |
| 1881 | |
| 1882 | /* Record a tag. */ |
| 1883 | static void |
| 1884 | pfnote (char *name, int is_func, char *linestart, int linelen, int lno, long int cno) |
| 1885 | /* tag name, or NULL if unnamed */ |
| 1886 | /* tag is a function */ |
| 1887 | /* start of the line where tag is */ |
| 1888 | /* length of the line where tag is */ |
| 1889 | /* line number */ |
| 1890 | /* character number */ |
| 1891 | { |
| 1892 | register node *np; |
| 1893 | |
| 1894 | assert (name == NULL || name[0] != '\0'); |
| 1895 | if (CTAGS && name == NULL) |
| 1896 | return; |
| 1897 | |
| 1898 | np = xnew (1, node); |
| 1899 | |
| 1900 | /* If ctags mode, change name "main" to M<thisfilename>. */ |
| 1901 | if (CTAGS && !cxref_style && streq (name, "main")) |
| 1902 | { |
| 1903 | register char *fp = etags_strrchr (curfdp->taggedfname, '/'); |
| 1904 | np->name = concat ("M", fp == NULL ? curfdp->taggedfname : fp + 1, ""); |
| 1905 | fp = etags_strrchr (np->name, '.'); |
| 1906 | if (fp != NULL && fp[1] != '\0' && fp[2] == '\0') |
| 1907 | fp[0] = '\0'; |
| 1908 | } |
| 1909 | else |
| 1910 | np->name = name; |
| 1911 | np->valid = TRUE; |
| 1912 | np->been_warned = FALSE; |
| 1913 | np->fdp = curfdp; |
| 1914 | np->is_func = is_func; |
| 1915 | np->lno = lno; |
| 1916 | if (np->fdp->usecharno) |
| 1917 | /* Our char numbers are 0-base, because of C language tradition? |
| 1918 | ctags compatibility? old versions compatibility? I don't know. |
| 1919 | Anyway, since emacs's are 1-base we expect etags.el to take care |
| 1920 | of the difference. If we wanted to have 1-based numbers, we would |
| 1921 | uncomment the +1 below. */ |
| 1922 | np->cno = cno /* + 1 */ ; |
| 1923 | else |
| 1924 | np->cno = invalidcharno; |
| 1925 | np->left = np->right = NULL; |
| 1926 | if (CTAGS && !cxref_style) |
| 1927 | { |
| 1928 | if (strlen (linestart) < 50) |
| 1929 | np->regex = concat (linestart, "$", ""); |
| 1930 | else |
| 1931 | np->regex = savenstr (linestart, 50); |
| 1932 | } |
| 1933 | else |
| 1934 | np->regex = savenstr (linestart, linelen); |
| 1935 | |
| 1936 | add_node (np, &nodehead); |
| 1937 | } |
| 1938 | |
| 1939 | /* |
| 1940 | * free_tree () |
| 1941 | * recurse on left children, iterate on right children. |
| 1942 | */ |
| 1943 | static void |
| 1944 | free_tree (register node *np) |
| 1945 | { |
| 1946 | while (np) |
| 1947 | { |
| 1948 | register node *node_right = np->right; |
| 1949 | free_tree (np->left); |
| 1950 | free (np->name); |
| 1951 | free (np->regex); |
| 1952 | free (np); |
| 1953 | np = node_right; |
| 1954 | } |
| 1955 | } |
| 1956 | |
| 1957 | /* |
| 1958 | * free_fdesc () |
| 1959 | * delete a file description |
| 1960 | */ |
| 1961 | static void |
| 1962 | free_fdesc (register fdesc *fdp) |
| 1963 | { |
| 1964 | free (fdp->infname); |
| 1965 | free (fdp->infabsname); |
| 1966 | free (fdp->infabsdir); |
| 1967 | free (fdp->taggedfname); |
| 1968 | free (fdp->prop); |
| 1969 | free (fdp); |
| 1970 | } |
| 1971 | |
| 1972 | /* |
| 1973 | * add_node () |
| 1974 | * Adds a node to the tree of nodes. In etags mode, sort by file |
| 1975 | * name. In ctags mode, sort by tag name. Make no attempt at |
| 1976 | * balancing. |
| 1977 | * |
| 1978 | * add_node is the only function allowed to add nodes, so it can |
| 1979 | * maintain state. |
| 1980 | */ |
| 1981 | static void |
| 1982 | add_node (node *np, node **cur_node_p) |
| 1983 | { |
| 1984 | register int dif; |
| 1985 | register node *cur_node = *cur_node_p; |
| 1986 | |
| 1987 | if (cur_node == NULL) |
| 1988 | { |
| 1989 | *cur_node_p = np; |
| 1990 | last_node = np; |
| 1991 | return; |
| 1992 | } |
| 1993 | |
| 1994 | if (!CTAGS) |
| 1995 | /* Etags Mode */ |
| 1996 | { |
| 1997 | /* For each file name, tags are in a linked sublist on the right |
| 1998 | pointer. The first tags of different files are a linked list |
| 1999 | on the left pointer. last_node points to the end of the last |
| 2000 | used sublist. */ |
| 2001 | if (last_node != NULL && last_node->fdp == np->fdp) |
| 2002 | { |
| 2003 | /* Let's use the same sublist as the last added node. */ |
| 2004 | assert (last_node->right == NULL); |
| 2005 | last_node->right = np; |
| 2006 | last_node = np; |
| 2007 | } |
| 2008 | else if (cur_node->fdp == np->fdp) |
| 2009 | { |
| 2010 | /* Scanning the list we found the head of a sublist which is |
| 2011 | good for us. Let's scan this sublist. */ |
| 2012 | add_node (np, &cur_node->right); |
| 2013 | } |
| 2014 | else |
| 2015 | /* The head of this sublist is not good for us. Let's try the |
| 2016 | next one. */ |
| 2017 | add_node (np, &cur_node->left); |
| 2018 | } /* if ETAGS mode */ |
| 2019 | |
| 2020 | else |
| 2021 | { |
| 2022 | /* Ctags Mode */ |
| 2023 | dif = strcmp (np->name, cur_node->name); |
| 2024 | |
| 2025 | /* |
| 2026 | * If this tag name matches an existing one, then |
| 2027 | * do not add the node, but maybe print a warning. |
| 2028 | */ |
| 2029 | if (no_duplicates && !dif) |
| 2030 | { |
| 2031 | if (np->fdp == cur_node->fdp) |
| 2032 | { |
| 2033 | if (!no_warnings) |
| 2034 | { |
| 2035 | fprintf (stderr, "Duplicate entry in file %s, line %d: %s\n", |
| 2036 | np->fdp->infname, lineno, np->name); |
| 2037 | fprintf (stderr, "Second entry ignored\n"); |
| 2038 | } |
| 2039 | } |
| 2040 | else if (!cur_node->been_warned && !no_warnings) |
| 2041 | { |
| 2042 | fprintf |
| 2043 | (stderr, |
| 2044 | "Duplicate entry in files %s and %s: %s (Warning only)\n", |
| 2045 | np->fdp->infname, cur_node->fdp->infname, np->name); |
| 2046 | cur_node->been_warned = TRUE; |
| 2047 | } |
| 2048 | return; |
| 2049 | } |
| 2050 | |
| 2051 | /* Actually add the node */ |
| 2052 | add_node (np, dif < 0 ? &cur_node->left : &cur_node->right); |
| 2053 | } /* if CTAGS mode */ |
| 2054 | } |
| 2055 | |
| 2056 | /* |
| 2057 | * invalidate_nodes () |
| 2058 | * Scan the node tree and invalidate all nodes pointing to the |
| 2059 | * given file description (CTAGS case) or free them (ETAGS case). |
| 2060 | */ |
| 2061 | static void |
| 2062 | invalidate_nodes (fdesc *badfdp, node **npp) |
| 2063 | { |
| 2064 | node *np = *npp; |
| 2065 | |
| 2066 | if (np == NULL) |
| 2067 | return; |
| 2068 | |
| 2069 | if (CTAGS) |
| 2070 | { |
| 2071 | if (np->left != NULL) |
| 2072 | invalidate_nodes (badfdp, &np->left); |
| 2073 | if (np->fdp == badfdp) |
| 2074 | np->valid = FALSE; |
| 2075 | if (np->right != NULL) |
| 2076 | invalidate_nodes (badfdp, &np->right); |
| 2077 | } |
| 2078 | else |
| 2079 | { |
| 2080 | assert (np->fdp != NULL); |
| 2081 | if (np->fdp == badfdp) |
| 2082 | { |
| 2083 | *npp = np->left; /* detach the sublist from the list */ |
| 2084 | np->left = NULL; /* isolate it */ |
| 2085 | free_tree (np); /* free it */ |
| 2086 | invalidate_nodes (badfdp, npp); |
| 2087 | } |
| 2088 | else |
| 2089 | invalidate_nodes (badfdp, &np->left); |
| 2090 | } |
| 2091 | } |
| 2092 | |
| 2093 | \f |
| 2094 | static int total_size_of_entries (node *); |
| 2095 | static int number_len (long) ATTRIBUTE_CONST; |
| 2096 | |
| 2097 | /* Length of a non-negative number's decimal representation. */ |
| 2098 | static int |
| 2099 | number_len (long int num) |
| 2100 | { |
| 2101 | int len = 1; |
| 2102 | while ((num /= 10) > 0) |
| 2103 | len += 1; |
| 2104 | return len; |
| 2105 | } |
| 2106 | |
| 2107 | /* |
| 2108 | * Return total number of characters that put_entries will output for |
| 2109 | * the nodes in the linked list at the right of the specified node. |
| 2110 | * This count is irrelevant with etags.el since emacs 19.34 at least, |
| 2111 | * but is still supplied for backward compatibility. |
| 2112 | */ |
| 2113 | static int |
| 2114 | total_size_of_entries (register node *np) |
| 2115 | { |
| 2116 | register int total = 0; |
| 2117 | |
| 2118 | for (; np != NULL; np = np->right) |
| 2119 | if (np->valid) |
| 2120 | { |
| 2121 | total += strlen (np->regex) + 1; /* pat\177 */ |
| 2122 | if (np->name != NULL) |
| 2123 | total += strlen (np->name) + 1; /* name\001 */ |
| 2124 | total += number_len ((long) np->lno) + 1; /* lno, */ |
| 2125 | if (np->cno != invalidcharno) /* cno */ |
| 2126 | total += number_len (np->cno); |
| 2127 | total += 1; /* newline */ |
| 2128 | } |
| 2129 | |
| 2130 | return total; |
| 2131 | } |
| 2132 | |
| 2133 | static void |
| 2134 | put_entries (register node *np) |
| 2135 | { |
| 2136 | register char *sp; |
| 2137 | static fdesc *fdp = NULL; |
| 2138 | |
| 2139 | if (np == NULL) |
| 2140 | return; |
| 2141 | |
| 2142 | /* Output subentries that precede this one */ |
| 2143 | if (CTAGS) |
| 2144 | put_entries (np->left); |
| 2145 | |
| 2146 | /* Output this entry */ |
| 2147 | if (np->valid) |
| 2148 | { |
| 2149 | if (!CTAGS) |
| 2150 | { |
| 2151 | /* Etags mode */ |
| 2152 | if (fdp != np->fdp) |
| 2153 | { |
| 2154 | fdp = np->fdp; |
| 2155 | fprintf (tagf, "\f\n%s,%d\n", |
| 2156 | fdp->taggedfname, total_size_of_entries (np)); |
| 2157 | fdp->written = TRUE; |
| 2158 | } |
| 2159 | fputs (np->regex, tagf); |
| 2160 | fputc ('\177', tagf); |
| 2161 | if (np->name != NULL) |
| 2162 | { |
| 2163 | fputs (np->name, tagf); |
| 2164 | fputc ('\001', tagf); |
| 2165 | } |
| 2166 | fprintf (tagf, "%d,", np->lno); |
| 2167 | if (np->cno != invalidcharno) |
| 2168 | fprintf (tagf, "%ld", np->cno); |
| 2169 | fputs ("\n", tagf); |
| 2170 | } |
| 2171 | else |
| 2172 | { |
| 2173 | /* Ctags mode */ |
| 2174 | if (np->name == NULL) |
| 2175 | error ("internal error: NULL name in ctags mode."); |
| 2176 | |
| 2177 | if (cxref_style) |
| 2178 | { |
| 2179 | if (vgrind_style) |
| 2180 | fprintf (stdout, "%s %s %d\n", |
| 2181 | np->name, np->fdp->taggedfname, (np->lno + 63) / 64); |
| 2182 | else |
| 2183 | fprintf (stdout, "%-16s %3d %-16s %s\n", |
| 2184 | np->name, np->lno, np->fdp->taggedfname, np->regex); |
| 2185 | } |
| 2186 | else |
| 2187 | { |
| 2188 | fprintf (tagf, "%s\t%s\t", np->name, np->fdp->taggedfname); |
| 2189 | |
| 2190 | if (np->is_func) |
| 2191 | { /* function or #define macro with args */ |
| 2192 | putc (searchar, tagf); |
| 2193 | putc ('^', tagf); |
| 2194 | |
| 2195 | for (sp = np->regex; *sp; sp++) |
| 2196 | { |
| 2197 | if (*sp == '\\' || *sp == searchar) |
| 2198 | putc ('\\', tagf); |
| 2199 | putc (*sp, tagf); |
| 2200 | } |
| 2201 | putc (searchar, tagf); |
| 2202 | } |
| 2203 | else |
| 2204 | { /* anything else; text pattern inadequate */ |
| 2205 | fprintf (tagf, "%d", np->lno); |
| 2206 | } |
| 2207 | putc ('\n', tagf); |
| 2208 | } |
| 2209 | } |
| 2210 | } /* if this node contains a valid tag */ |
| 2211 | |
| 2212 | /* Output subentries that follow this one */ |
| 2213 | put_entries (np->right); |
| 2214 | if (!CTAGS) |
| 2215 | put_entries (np->left); |
| 2216 | } |
| 2217 | |
| 2218 | \f |
| 2219 | /* C extensions. */ |
| 2220 | #define C_EXT 0x00fff /* C extensions */ |
| 2221 | #define C_PLAIN 0x00000 /* C */ |
| 2222 | #define C_PLPL 0x00001 /* C++ */ |
| 2223 | #define C_STAR 0x00003 /* C* */ |
| 2224 | #define C_JAVA 0x00005 /* JAVA */ |
| 2225 | #define C_AUTO 0x01000 /* C, but switch to C++ if `class' is met */ |
| 2226 | #define YACC 0x10000 /* yacc file */ |
| 2227 | |
| 2228 | /* |
| 2229 | * The C symbol tables. |
| 2230 | */ |
| 2231 | enum sym_type |
| 2232 | { |
| 2233 | st_none, |
| 2234 | st_C_objprot, st_C_objimpl, st_C_objend, |
| 2235 | st_C_gnumacro, |
| 2236 | st_C_ignore, st_C_attribute, |
| 2237 | st_C_javastruct, |
| 2238 | st_C_operator, |
| 2239 | st_C_class, st_C_template, |
| 2240 | st_C_struct, st_C_extern, st_C_enum, st_C_define, st_C_typedef |
| 2241 | }; |
| 2242 | |
| 2243 | /* Feed stuff between (but not including) %[ and %] lines to: |
| 2244 | gperf -m 5 |
| 2245 | %[ |
| 2246 | %compare-strncmp |
| 2247 | %enum |
| 2248 | %struct-type |
| 2249 | struct C_stab_entry { char *name; int c_ext; enum sym_type type; } |
| 2250 | %% |
| 2251 | if, 0, st_C_ignore |
| 2252 | for, 0, st_C_ignore |
| 2253 | while, 0, st_C_ignore |
| 2254 | switch, 0, st_C_ignore |
| 2255 | return, 0, st_C_ignore |
| 2256 | __attribute__, 0, st_C_attribute |
| 2257 | GTY, 0, st_C_attribute |
| 2258 | @interface, 0, st_C_objprot |
| 2259 | @protocol, 0, st_C_objprot |
| 2260 | @implementation,0, st_C_objimpl |
| 2261 | @end, 0, st_C_objend |
| 2262 | import, (C_JAVA & ~C_PLPL), st_C_ignore |
| 2263 | package, (C_JAVA & ~C_PLPL), st_C_ignore |
| 2264 | friend, C_PLPL, st_C_ignore |
| 2265 | extends, (C_JAVA & ~C_PLPL), st_C_javastruct |
| 2266 | implements, (C_JAVA & ~C_PLPL), st_C_javastruct |
| 2267 | interface, (C_JAVA & ~C_PLPL), st_C_struct |
| 2268 | class, 0, st_C_class |
| 2269 | namespace, C_PLPL, st_C_struct |
| 2270 | domain, C_STAR, st_C_struct |
| 2271 | union, 0, st_C_struct |
| 2272 | struct, 0, st_C_struct |
| 2273 | extern, 0, st_C_extern |
| 2274 | enum, 0, st_C_enum |
| 2275 | typedef, 0, st_C_typedef |
| 2276 | define, 0, st_C_define |
| 2277 | undef, 0, st_C_define |
| 2278 | operator, C_PLPL, st_C_operator |
| 2279 | template, 0, st_C_template |
| 2280 | # DEFUN used in emacs, the next three used in glibc (SYSCALL only for mach). |
| 2281 | DEFUN, 0, st_C_gnumacro |
| 2282 | SYSCALL, 0, st_C_gnumacro |
| 2283 | ENTRY, 0, st_C_gnumacro |
| 2284 | PSEUDO, 0, st_C_gnumacro |
| 2285 | # These are defined inside C functions, so currently they are not met. |
| 2286 | # EXFUN used in glibc, DEFVAR_* in emacs. |
| 2287 | #EXFUN, 0, st_C_gnumacro |
| 2288 | #DEFVAR_, 0, st_C_gnumacro |
| 2289 | %] |
| 2290 | and replace lines between %< and %> with its output, then: |
| 2291 | - remove the #if characterset check |
| 2292 | - make in_word_set static and not inline. */ |
| 2293 | /*%<*/ |
| 2294 | /* C code produced by gperf version 3.0.1 */ |
| 2295 | /* Command-line: gperf -m 5 */ |
| 2296 | /* Computed positions: -k'2-3' */ |
| 2297 | |
| 2298 | struct C_stab_entry { const char *name; int c_ext; enum sym_type type; }; |
| 2299 | /* maximum key range = 33, duplicates = 0 */ |
| 2300 | |
| 2301 | static int |
| 2302 | hash (const char *str, int len) |
| 2303 | { |
| 2304 | static char const asso_values[] = |
| 2305 | { |
| 2306 | 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, |
| 2307 | 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, |
| 2308 | 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, |
| 2309 | 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, |
| 2310 | 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, |
| 2311 | 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, |
| 2312 | 35, 35, 35, 35, 35, 35, 35, 35, 35, 3, |
| 2313 | 26, 35, 35, 35, 35, 35, 35, 35, 27, 35, |
| 2314 | 35, 35, 35, 24, 0, 35, 35, 35, 35, 0, |
| 2315 | 35, 35, 35, 35, 35, 1, 35, 16, 35, 6, |
| 2316 | 23, 0, 0, 35, 22, 0, 35, 35, 5, 0, |
| 2317 | 0, 15, 1, 35, 6, 35, 8, 19, 35, 16, |
| 2318 | 4, 5, 35, 35, 35, 35, 35, 35, 35, 35, |
| 2319 | 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, |
| 2320 | 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, |
| 2321 | 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, |
| 2322 | 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, |
| 2323 | 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, |
| 2324 | 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, |
| 2325 | 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, |
| 2326 | 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, |
| 2327 | 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, |
| 2328 | 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, |
| 2329 | 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, |
| 2330 | 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, |
| 2331 | 35, 35, 35, 35, 35, 35 |
| 2332 | }; |
| 2333 | int hval = len; |
| 2334 | |
| 2335 | switch (hval) |
| 2336 | { |
| 2337 | default: |
| 2338 | hval += asso_values[(unsigned char) str[2]]; |
| 2339 | /*FALLTHROUGH*/ |
| 2340 | case 2: |
| 2341 | hval += asso_values[(unsigned char) str[1]]; |
| 2342 | break; |
| 2343 | } |
| 2344 | return hval; |
| 2345 | } |
| 2346 | |
| 2347 | static struct C_stab_entry * |
| 2348 | in_word_set (register const char *str, register unsigned int len) |
| 2349 | { |
| 2350 | enum |
| 2351 | { |
| 2352 | TOTAL_KEYWORDS = 33, |
| 2353 | MIN_WORD_LENGTH = 2, |
| 2354 | MAX_WORD_LENGTH = 15, |
| 2355 | MIN_HASH_VALUE = 2, |
| 2356 | MAX_HASH_VALUE = 34 |
| 2357 | }; |
| 2358 | |
| 2359 | static struct C_stab_entry wordlist[] = |
| 2360 | { |
| 2361 | {""}, {""}, |
| 2362 | {"if", 0, st_C_ignore}, |
| 2363 | {"GTY", 0, st_C_attribute}, |
| 2364 | {"@end", 0, st_C_objend}, |
| 2365 | {"union", 0, st_C_struct}, |
| 2366 | {"define", 0, st_C_define}, |
| 2367 | {"import", (C_JAVA & ~C_PLPL), st_C_ignore}, |
| 2368 | {"template", 0, st_C_template}, |
| 2369 | {"operator", C_PLPL, st_C_operator}, |
| 2370 | {"@interface", 0, st_C_objprot}, |
| 2371 | {"implements", (C_JAVA & ~C_PLPL), st_C_javastruct}, |
| 2372 | {"friend", C_PLPL, st_C_ignore}, |
| 2373 | {"typedef", 0, st_C_typedef}, |
| 2374 | {"return", 0, st_C_ignore}, |
| 2375 | {"@implementation",0, st_C_objimpl}, |
| 2376 | {"@protocol", 0, st_C_objprot}, |
| 2377 | {"interface", (C_JAVA & ~C_PLPL), st_C_struct}, |
| 2378 | {"extern", 0, st_C_extern}, |
| 2379 | {"extends", (C_JAVA & ~C_PLPL), st_C_javastruct}, |
| 2380 | {"struct", 0, st_C_struct}, |
| 2381 | {"domain", C_STAR, st_C_struct}, |
| 2382 | {"switch", 0, st_C_ignore}, |
| 2383 | {"enum", 0, st_C_enum}, |
| 2384 | {"for", 0, st_C_ignore}, |
| 2385 | {"namespace", C_PLPL, st_C_struct}, |
| 2386 | {"class", 0, st_C_class}, |
| 2387 | {"while", 0, st_C_ignore}, |
| 2388 | {"undef", 0, st_C_define}, |
| 2389 | {"package", (C_JAVA & ~C_PLPL), st_C_ignore}, |
| 2390 | {"__attribute__", 0, st_C_attribute}, |
| 2391 | {"SYSCALL", 0, st_C_gnumacro}, |
| 2392 | {"ENTRY", 0, st_C_gnumacro}, |
| 2393 | {"PSEUDO", 0, st_C_gnumacro}, |
| 2394 | {"DEFUN", 0, st_C_gnumacro} |
| 2395 | }; |
| 2396 | |
| 2397 | if (len <= MAX_WORD_LENGTH && len >= MIN_WORD_LENGTH) |
| 2398 | { |
| 2399 | int key = hash (str, len); |
| 2400 | |
| 2401 | if (key <= MAX_HASH_VALUE && key >= 0) |
| 2402 | { |
| 2403 | const char *s = wordlist[key].name; |
| 2404 | |
| 2405 | if (*str == *s && !strncmp (str + 1, s + 1, len - 1) && s[len] == '\0') |
| 2406 | return &wordlist[key]; |
| 2407 | } |
| 2408 | } |
| 2409 | return 0; |
| 2410 | } |
| 2411 | /*%>*/ |
| 2412 | |
| 2413 | static enum sym_type |
| 2414 | C_symtype (char *str, int len, int c_ext) |
| 2415 | { |
| 2416 | register struct C_stab_entry *se = in_word_set (str, len); |
| 2417 | |
| 2418 | if (se == NULL || (se->c_ext && !(c_ext & se->c_ext))) |
| 2419 | return st_none; |
| 2420 | return se->type; |
| 2421 | } |
| 2422 | |
| 2423 | \f |
| 2424 | /* |
| 2425 | * Ignoring __attribute__ ((list)) |
| 2426 | */ |
| 2427 | static bool inattribute; /* looking at an __attribute__ construct */ |
| 2428 | |
| 2429 | /* |
| 2430 | * C functions and variables are recognized using a simple |
| 2431 | * finite automaton. fvdef is its state variable. |
| 2432 | */ |
| 2433 | static enum |
| 2434 | { |
| 2435 | fvnone, /* nothing seen */ |
| 2436 | fdefunkey, /* Emacs DEFUN keyword seen */ |
| 2437 | fdefunname, /* Emacs DEFUN name seen */ |
| 2438 | foperator, /* func: operator keyword seen (cplpl) */ |
| 2439 | fvnameseen, /* function or variable name seen */ |
| 2440 | fstartlist, /* func: just after open parenthesis */ |
| 2441 | finlist, /* func: in parameter list */ |
| 2442 | flistseen, /* func: after parameter list */ |
| 2443 | fignore, /* func: before open brace */ |
| 2444 | vignore /* var-like: ignore until ';' */ |
| 2445 | } fvdef; |
| 2446 | |
| 2447 | static bool fvextern; /* func or var: extern keyword seen; */ |
| 2448 | |
| 2449 | /* |
| 2450 | * typedefs are recognized using a simple finite automaton. |
| 2451 | * typdef is its state variable. |
| 2452 | */ |
| 2453 | static enum |
| 2454 | { |
| 2455 | tnone, /* nothing seen */ |
| 2456 | tkeyseen, /* typedef keyword seen */ |
| 2457 | ttypeseen, /* defined type seen */ |
| 2458 | tinbody, /* inside typedef body */ |
| 2459 | tend, /* just before typedef tag */ |
| 2460 | tignore /* junk after typedef tag */ |
| 2461 | } typdef; |
| 2462 | |
| 2463 | /* |
| 2464 | * struct-like structures (enum, struct and union) are recognized |
| 2465 | * using another simple finite automaton. `structdef' is its state |
| 2466 | * variable. |
| 2467 | */ |
| 2468 | static enum |
| 2469 | { |
| 2470 | snone, /* nothing seen yet, |
| 2471 | or in struct body if bracelev > 0 */ |
| 2472 | skeyseen, /* struct-like keyword seen */ |
| 2473 | stagseen, /* struct-like tag seen */ |
| 2474 | scolonseen /* colon seen after struct-like tag */ |
| 2475 | } structdef; |
| 2476 | |
| 2477 | /* |
| 2478 | * When objdef is different from onone, objtag is the name of the class. |
| 2479 | */ |
| 2480 | static const char *objtag = "<uninited>"; |
| 2481 | |
| 2482 | /* |
| 2483 | * Yet another little state machine to deal with preprocessor lines. |
| 2484 | */ |
| 2485 | static enum |
| 2486 | { |
| 2487 | dnone, /* nothing seen */ |
| 2488 | dsharpseen, /* '#' seen as first char on line */ |
| 2489 | ddefineseen, /* '#' and 'define' seen */ |
| 2490 | dignorerest /* ignore rest of line */ |
| 2491 | } definedef; |
| 2492 | |
| 2493 | /* |
| 2494 | * State machine for Objective C protocols and implementations. |
| 2495 | * Idea by Tom R.Hageman <tom@basil.icce.rug.nl> (1995) |
| 2496 | */ |
| 2497 | static enum |
| 2498 | { |
| 2499 | onone, /* nothing seen */ |
| 2500 | oprotocol, /* @interface or @protocol seen */ |
| 2501 | oimplementation, /* @implementations seen */ |
| 2502 | otagseen, /* class name seen */ |
| 2503 | oparenseen, /* parenthesis before category seen */ |
| 2504 | ocatseen, /* category name seen */ |
| 2505 | oinbody, /* in @implementation body */ |
| 2506 | omethodsign, /* in @implementation body, after +/- */ |
| 2507 | omethodtag, /* after method name */ |
| 2508 | omethodcolon, /* after method colon */ |
| 2509 | omethodparm, /* after method parameter */ |
| 2510 | oignore /* wait for @end */ |
| 2511 | } objdef; |
| 2512 | |
| 2513 | |
| 2514 | /* |
| 2515 | * Use this structure to keep info about the token read, and how it |
| 2516 | * should be tagged. Used by the make_C_tag function to build a tag. |
| 2517 | */ |
| 2518 | static struct tok |
| 2519 | { |
| 2520 | char *line; /* string containing the token */ |
| 2521 | int offset; /* where the token starts in LINE */ |
| 2522 | int length; /* token length */ |
| 2523 | /* |
| 2524 | The previous members can be used to pass strings around for generic |
| 2525 | purposes. The following ones specifically refer to creating tags. In this |
| 2526 | case the token contained here is the pattern that will be used to create a |
| 2527 | tag. |
| 2528 | */ |
| 2529 | bool valid; /* do not create a tag; the token should be |
| 2530 | invalidated whenever a state machine is |
| 2531 | reset prematurely */ |
| 2532 | bool named; /* create a named tag */ |
| 2533 | int lineno; /* source line number of tag */ |
| 2534 | long linepos; /* source char number of tag */ |
| 2535 | } token; /* latest token read */ |
| 2536 | |
| 2537 | /* |
| 2538 | * Variables and functions for dealing with nested structures. |
| 2539 | * Idea by Mykola Dzyuba <mdzyuba@yahoo.com> (2001) |
| 2540 | */ |
| 2541 | static void pushclass_above (int, char *, int); |
| 2542 | static void popclass_above (int); |
| 2543 | static void write_classname (linebuffer *, const char *qualifier); |
| 2544 | |
| 2545 | static struct { |
| 2546 | char **cname; /* nested class names */ |
| 2547 | int *bracelev; /* nested class brace level */ |
| 2548 | int nl; /* class nesting level (elements used) */ |
| 2549 | int size; /* length of the array */ |
| 2550 | } cstack; /* stack for nested declaration tags */ |
| 2551 | /* Current struct nesting depth (namespace, class, struct, union, enum). */ |
| 2552 | #define nestlev (cstack.nl) |
| 2553 | /* After struct keyword or in struct body, not inside a nested function. */ |
| 2554 | #define instruct (structdef == snone && nestlev > 0 \ |
| 2555 | && bracelev == cstack.bracelev[nestlev-1] + 1) |
| 2556 | |
| 2557 | static void |
| 2558 | pushclass_above (int bracelev, char *str, int len) |
| 2559 | { |
| 2560 | int nl; |
| 2561 | |
| 2562 | popclass_above (bracelev); |
| 2563 | nl = cstack.nl; |
| 2564 | if (nl >= cstack.size) |
| 2565 | { |
| 2566 | int size = cstack.size *= 2; |
| 2567 | xrnew (cstack.cname, size, char *); |
| 2568 | xrnew (cstack.bracelev, size, int); |
| 2569 | } |
| 2570 | assert (nl == 0 || cstack.bracelev[nl-1] < bracelev); |
| 2571 | cstack.cname[nl] = (str == NULL) ? NULL : savenstr (str, len); |
| 2572 | cstack.bracelev[nl] = bracelev; |
| 2573 | cstack.nl = nl + 1; |
| 2574 | } |
| 2575 | |
| 2576 | static void |
| 2577 | popclass_above (int bracelev) |
| 2578 | { |
| 2579 | int nl; |
| 2580 | |
| 2581 | for (nl = cstack.nl - 1; |
| 2582 | nl >= 0 && cstack.bracelev[nl] >= bracelev; |
| 2583 | nl--) |
| 2584 | { |
| 2585 | free (cstack.cname[nl]); |
| 2586 | cstack.nl = nl; |
| 2587 | } |
| 2588 | } |
| 2589 | |
| 2590 | static void |
| 2591 | write_classname (linebuffer *cn, const char *qualifier) |
| 2592 | { |
| 2593 | int i, len; |
| 2594 | int qlen = strlen (qualifier); |
| 2595 | |
| 2596 | if (cstack.nl == 0 || cstack.cname[0] == NULL) |
| 2597 | { |
| 2598 | len = 0; |
| 2599 | cn->len = 0; |
| 2600 | cn->buffer[0] = '\0'; |
| 2601 | } |
| 2602 | else |
| 2603 | { |
| 2604 | len = strlen (cstack.cname[0]); |
| 2605 | linebuffer_setlen (cn, len); |
| 2606 | strcpy (cn->buffer, cstack.cname[0]); |
| 2607 | } |
| 2608 | for (i = 1; i < cstack.nl; i++) |
| 2609 | { |
| 2610 | char *s = cstack.cname[i]; |
| 2611 | if (s == NULL) |
| 2612 | continue; |
| 2613 | linebuffer_setlen (cn, len + qlen + strlen (s)); |
| 2614 | len += sprintf (cn->buffer + len, "%s%s", qualifier, s); |
| 2615 | } |
| 2616 | } |
| 2617 | |
| 2618 | \f |
| 2619 | static bool consider_token (char *, int, int, int *, int, int, bool *); |
| 2620 | static void make_C_tag (bool); |
| 2621 | |
| 2622 | /* |
| 2623 | * consider_token () |
| 2624 | * checks to see if the current token is at the start of a |
| 2625 | * function or variable, or corresponds to a typedef, or |
| 2626 | * is a struct/union/enum tag, or #define, or an enum constant. |
| 2627 | * |
| 2628 | * *IS_FUNC gets TRUE if the token is a function or #define macro |
| 2629 | * with args. C_EXTP points to which language we are looking at. |
| 2630 | * |
| 2631 | * Globals |
| 2632 | * fvdef IN OUT |
| 2633 | * structdef IN OUT |
| 2634 | * definedef IN OUT |
| 2635 | * typdef IN OUT |
| 2636 | * objdef IN OUT |
| 2637 | */ |
| 2638 | |
| 2639 | static bool |
| 2640 | consider_token (register char *str, register int len, register int c, int *c_extp, int bracelev, int parlev, int *is_func_or_var) |
| 2641 | /* IN: token pointer */ |
| 2642 | /* IN: token length */ |
| 2643 | /* IN: first char after the token */ |
| 2644 | /* IN, OUT: C extensions mask */ |
| 2645 | /* IN: brace level */ |
| 2646 | /* IN: parenthesis level */ |
| 2647 | /* OUT: function or variable found */ |
| 2648 | { |
| 2649 | /* When structdef is stagseen, scolonseen, or snone with bracelev > 0, |
| 2650 | structtype is the type of the preceding struct-like keyword, and |
| 2651 | structbracelev is the brace level where it has been seen. */ |
| 2652 | static enum sym_type structtype; |
| 2653 | static int structbracelev; |
| 2654 | static enum sym_type toktype; |
| 2655 | |
| 2656 | |
| 2657 | toktype = C_symtype (str, len, *c_extp); |
| 2658 | |
| 2659 | /* |
| 2660 | * Skip __attribute__ |
| 2661 | */ |
| 2662 | if (toktype == st_C_attribute) |
| 2663 | { |
| 2664 | inattribute = TRUE; |
| 2665 | return FALSE; |
| 2666 | } |
| 2667 | |
| 2668 | /* |
| 2669 | * Advance the definedef state machine. |
| 2670 | */ |
| 2671 | switch (definedef) |
| 2672 | { |
| 2673 | case dnone: |
| 2674 | /* We're not on a preprocessor line. */ |
| 2675 | if (toktype == st_C_gnumacro) |
| 2676 | { |
| 2677 | fvdef = fdefunkey; |
| 2678 | return FALSE; |
| 2679 | } |
| 2680 | break; |
| 2681 | case dsharpseen: |
| 2682 | if (toktype == st_C_define) |
| 2683 | { |
| 2684 | definedef = ddefineseen; |
| 2685 | } |
| 2686 | else |
| 2687 | { |
| 2688 | definedef = dignorerest; |
| 2689 | } |
| 2690 | return FALSE; |
| 2691 | case ddefineseen: |
| 2692 | /* |
| 2693 | * Make a tag for any macro, unless it is a constant |
| 2694 | * and constantypedefs is FALSE. |
| 2695 | */ |
| 2696 | definedef = dignorerest; |
| 2697 | *is_func_or_var = (c == '('); |
| 2698 | if (!*is_func_or_var && !constantypedefs) |
| 2699 | return FALSE; |
| 2700 | else |
| 2701 | return TRUE; |
| 2702 | case dignorerest: |
| 2703 | return FALSE; |
| 2704 | default: |
| 2705 | error ("internal error: definedef value."); |
| 2706 | } |
| 2707 | |
| 2708 | /* |
| 2709 | * Now typedefs |
| 2710 | */ |
| 2711 | switch (typdef) |
| 2712 | { |
| 2713 | case tnone: |
| 2714 | if (toktype == st_C_typedef) |
| 2715 | { |
| 2716 | if (typedefs) |
| 2717 | typdef = tkeyseen; |
| 2718 | fvextern = FALSE; |
| 2719 | fvdef = fvnone; |
| 2720 | return FALSE; |
| 2721 | } |
| 2722 | break; |
| 2723 | case tkeyseen: |
| 2724 | switch (toktype) |
| 2725 | { |
| 2726 | case st_none: |
| 2727 | case st_C_class: |
| 2728 | case st_C_struct: |
| 2729 | case st_C_enum: |
| 2730 | typdef = ttypeseen; |
| 2731 | } |
| 2732 | break; |
| 2733 | case ttypeseen: |
| 2734 | if (structdef == snone && fvdef == fvnone) |
| 2735 | { |
| 2736 | fvdef = fvnameseen; |
| 2737 | return TRUE; |
| 2738 | } |
| 2739 | break; |
| 2740 | case tend: |
| 2741 | switch (toktype) |
| 2742 | { |
| 2743 | case st_C_class: |
| 2744 | case st_C_struct: |
| 2745 | case st_C_enum: |
| 2746 | return FALSE; |
| 2747 | } |
| 2748 | return TRUE; |
| 2749 | } |
| 2750 | |
| 2751 | switch (toktype) |
| 2752 | { |
| 2753 | case st_C_javastruct: |
| 2754 | if (structdef == stagseen) |
| 2755 | structdef = scolonseen; |
| 2756 | return FALSE; |
| 2757 | case st_C_template: |
| 2758 | case st_C_class: |
| 2759 | if ((*c_extp & C_AUTO) /* automatic detection of C++ language */ |
| 2760 | && bracelev == 0 |
| 2761 | && definedef == dnone && structdef == snone |
| 2762 | && typdef == tnone && fvdef == fvnone) |
| 2763 | *c_extp = (*c_extp | C_PLPL) & ~C_AUTO; |
| 2764 | if (toktype == st_C_template) |
| 2765 | break; |
| 2766 | /* FALLTHRU */ |
| 2767 | case st_C_struct: |
| 2768 | case st_C_enum: |
| 2769 | if (parlev == 0 |
| 2770 | && fvdef != vignore |
| 2771 | && (typdef == tkeyseen |
| 2772 | || (typedefs_or_cplusplus && structdef == snone))) |
| 2773 | { |
| 2774 | structdef = skeyseen; |
| 2775 | structtype = toktype; |
| 2776 | structbracelev = bracelev; |
| 2777 | if (fvdef == fvnameseen) |
| 2778 | fvdef = fvnone; |
| 2779 | } |
| 2780 | return FALSE; |
| 2781 | } |
| 2782 | |
| 2783 | if (structdef == skeyseen) |
| 2784 | { |
| 2785 | structdef = stagseen; |
| 2786 | return TRUE; |
| 2787 | } |
| 2788 | |
| 2789 | if (typdef != tnone) |
| 2790 | definedef = dnone; |
| 2791 | |
| 2792 | /* Detect Objective C constructs. */ |
| 2793 | switch (objdef) |
| 2794 | { |
| 2795 | case onone: |
| 2796 | switch (toktype) |
| 2797 | { |
| 2798 | case st_C_objprot: |
| 2799 | objdef = oprotocol; |
| 2800 | return FALSE; |
| 2801 | case st_C_objimpl: |
| 2802 | objdef = oimplementation; |
| 2803 | return FALSE; |
| 2804 | } |
| 2805 | break; |
| 2806 | case oimplementation: |
| 2807 | /* Save the class tag for functions or variables defined inside. */ |
| 2808 | objtag = savenstr (str, len); |
| 2809 | objdef = oinbody; |
| 2810 | return FALSE; |
| 2811 | case oprotocol: |
| 2812 | /* Save the class tag for categories. */ |
| 2813 | objtag = savenstr (str, len); |
| 2814 | objdef = otagseen; |
| 2815 | *is_func_or_var = TRUE; |
| 2816 | return TRUE; |
| 2817 | case oparenseen: |
| 2818 | objdef = ocatseen; |
| 2819 | *is_func_or_var = TRUE; |
| 2820 | return TRUE; |
| 2821 | case oinbody: |
| 2822 | break; |
| 2823 | case omethodsign: |
| 2824 | if (parlev == 0) |
| 2825 | { |
| 2826 | fvdef = fvnone; |
| 2827 | objdef = omethodtag; |
| 2828 | linebuffer_setlen (&token_name, len); |
| 2829 | memcpy (token_name.buffer, str, len); |
| 2830 | token_name.buffer[len] = '\0'; |
| 2831 | return TRUE; |
| 2832 | } |
| 2833 | return FALSE; |
| 2834 | case omethodcolon: |
| 2835 | if (parlev == 0) |
| 2836 | objdef = omethodparm; |
| 2837 | return FALSE; |
| 2838 | case omethodparm: |
| 2839 | if (parlev == 0) |
| 2840 | { |
| 2841 | int oldlen = token_name.len; |
| 2842 | fvdef = fvnone; |
| 2843 | objdef = omethodtag; |
| 2844 | linebuffer_setlen (&token_name, oldlen + len); |
| 2845 | memcpy (token_name.buffer + oldlen, str, len); |
| 2846 | token_name.buffer[oldlen + len] = '\0'; |
| 2847 | return TRUE; |
| 2848 | } |
| 2849 | return FALSE; |
| 2850 | case oignore: |
| 2851 | if (toktype == st_C_objend) |
| 2852 | { |
| 2853 | /* Memory leakage here: the string pointed by objtag is |
| 2854 | never released, because many tests would be needed to |
| 2855 | avoid breaking on incorrect input code. The amount of |
| 2856 | memory leaked here is the sum of the lengths of the |
| 2857 | class tags. |
| 2858 | free (objtag); */ |
| 2859 | objdef = onone; |
| 2860 | } |
| 2861 | return FALSE; |
| 2862 | } |
| 2863 | |
| 2864 | /* A function, variable or enum constant? */ |
| 2865 | switch (toktype) |
| 2866 | { |
| 2867 | case st_C_extern: |
| 2868 | fvextern = TRUE; |
| 2869 | switch (fvdef) |
| 2870 | { |
| 2871 | case finlist: |
| 2872 | case flistseen: |
| 2873 | case fignore: |
| 2874 | case vignore: |
| 2875 | break; |
| 2876 | default: |
| 2877 | fvdef = fvnone; |
| 2878 | } |
| 2879 | return FALSE; |
| 2880 | case st_C_ignore: |
| 2881 | fvextern = FALSE; |
| 2882 | fvdef = vignore; |
| 2883 | return FALSE; |
| 2884 | case st_C_operator: |
| 2885 | fvdef = foperator; |
| 2886 | *is_func_or_var = TRUE; |
| 2887 | return TRUE; |
| 2888 | case st_none: |
| 2889 | if (constantypedefs |
| 2890 | && structdef == snone |
| 2891 | && structtype == st_C_enum && bracelev > structbracelev) |
| 2892 | return TRUE; /* enum constant */ |
| 2893 | switch (fvdef) |
| 2894 | { |
| 2895 | case fdefunkey: |
| 2896 | if (bracelev > 0) |
| 2897 | break; |
| 2898 | fvdef = fdefunname; /* GNU macro */ |
| 2899 | *is_func_or_var = TRUE; |
| 2900 | return TRUE; |
| 2901 | case fvnone: |
| 2902 | switch (typdef) |
| 2903 | { |
| 2904 | case ttypeseen: |
| 2905 | return FALSE; |
| 2906 | case tnone: |
| 2907 | if ((strneq (str, "asm", 3) && endtoken (str[3])) |
| 2908 | || (strneq (str, "__asm__", 7) && endtoken (str[7]))) |
| 2909 | { |
| 2910 | fvdef = vignore; |
| 2911 | return FALSE; |
| 2912 | } |
| 2913 | break; |
| 2914 | } |
| 2915 | /* FALLTHRU */ |
| 2916 | case fvnameseen: |
| 2917 | if (len >= 10 && strneq (str+len-10, "::operator", 10)) |
| 2918 | { |
| 2919 | if (*c_extp & C_AUTO) /* automatic detection of C++ */ |
| 2920 | *c_extp = (*c_extp | C_PLPL) & ~C_AUTO; |
| 2921 | fvdef = foperator; |
| 2922 | *is_func_or_var = TRUE; |
| 2923 | return TRUE; |
| 2924 | } |
| 2925 | if (bracelev > 0 && !instruct) |
| 2926 | break; |
| 2927 | fvdef = fvnameseen; /* function or variable */ |
| 2928 | *is_func_or_var = TRUE; |
| 2929 | return TRUE; |
| 2930 | } |
| 2931 | break; |
| 2932 | } |
| 2933 | |
| 2934 | return FALSE; |
| 2935 | } |
| 2936 | |
| 2937 | \f |
| 2938 | /* |
| 2939 | * C_entries often keeps pointers to tokens or lines which are older than |
| 2940 | * the line currently read. By keeping two line buffers, and switching |
| 2941 | * them at end of line, it is possible to use those pointers. |
| 2942 | */ |
| 2943 | static struct |
| 2944 | { |
| 2945 | long linepos; |
| 2946 | linebuffer lb; |
| 2947 | } lbs[2]; |
| 2948 | |
| 2949 | #define current_lb_is_new (newndx == curndx) |
| 2950 | #define switch_line_buffers() (curndx = 1 - curndx) |
| 2951 | |
| 2952 | #define curlb (lbs[curndx].lb) |
| 2953 | #define newlb (lbs[newndx].lb) |
| 2954 | #define curlinepos (lbs[curndx].linepos) |
| 2955 | #define newlinepos (lbs[newndx].linepos) |
| 2956 | |
| 2957 | #define plainc ((c_ext & C_EXT) == C_PLAIN) |
| 2958 | #define cplpl (c_ext & C_PLPL) |
| 2959 | #define cjava ((c_ext & C_JAVA) == C_JAVA) |
| 2960 | |
| 2961 | #define CNL_SAVE_DEFINEDEF() \ |
| 2962 | do { \ |
| 2963 | curlinepos = charno; \ |
| 2964 | readline (&curlb, inf); \ |
| 2965 | lp = curlb.buffer; \ |
| 2966 | quotednl = FALSE; \ |
| 2967 | newndx = curndx; \ |
| 2968 | } while (0) |
| 2969 | |
| 2970 | #define CNL() \ |
| 2971 | do { \ |
| 2972 | CNL_SAVE_DEFINEDEF(); \ |
| 2973 | if (savetoken.valid) \ |
| 2974 | { \ |
| 2975 | token = savetoken; \ |
| 2976 | savetoken.valid = FALSE; \ |
| 2977 | } \ |
| 2978 | definedef = dnone; \ |
| 2979 | } while (0) |
| 2980 | |
| 2981 | |
| 2982 | static void |
| 2983 | make_C_tag (int isfun) |
| 2984 | { |
| 2985 | /* This function is never called when token.valid is FALSE, but |
| 2986 | we must protect against invalid input or internal errors. */ |
| 2987 | if (token.valid) |
| 2988 | make_tag (token_name.buffer, token_name.len, isfun, token.line, |
| 2989 | token.offset+token.length+1, token.lineno, token.linepos); |
| 2990 | else if (DEBUG) |
| 2991 | { /* this branch is optimized away if !DEBUG */ |
| 2992 | make_tag (concat ("INVALID TOKEN:-->", token_name.buffer, ""), |
| 2993 | token_name.len + 17, isfun, token.line, |
| 2994 | token.offset+token.length+1, token.lineno, token.linepos); |
| 2995 | error ("INVALID TOKEN"); |
| 2996 | } |
| 2997 | |
| 2998 | token.valid = FALSE; |
| 2999 | } |
| 3000 | |
| 3001 | |
| 3002 | /* |
| 3003 | * C_entries () |
| 3004 | * This routine finds functions, variables, typedefs, |
| 3005 | * #define's, enum constants and struct/union/enum definitions in |
| 3006 | * C syntax and adds them to the list. |
| 3007 | */ |
| 3008 | static void |
| 3009 | C_entries (int c_ext, FILE *inf) |
| 3010 | /* extension of C */ |
| 3011 | /* input file */ |
| 3012 | { |
| 3013 | register char c; /* latest char read; '\0' for end of line */ |
| 3014 | register char *lp; /* pointer one beyond the character `c' */ |
| 3015 | int curndx, newndx; /* indices for current and new lb */ |
| 3016 | register int tokoff; /* offset in line of start of current token */ |
| 3017 | register int toklen; /* length of current token */ |
| 3018 | const char *qualifier; /* string used to qualify names */ |
| 3019 | int qlen; /* length of qualifier */ |
| 3020 | int bracelev; /* current brace level */ |
| 3021 | int bracketlev; /* current bracket level */ |
| 3022 | int parlev; /* current parenthesis level */ |
| 3023 | int attrparlev; /* __attribute__ parenthesis level */ |
| 3024 | int templatelev; /* current template level */ |
| 3025 | int typdefbracelev; /* bracelev where a typedef struct body begun */ |
| 3026 | bool incomm, inquote, inchar, quotednl, midtoken; |
| 3027 | bool yacc_rules; /* in the rules part of a yacc file */ |
| 3028 | struct tok savetoken = {0}; /* token saved during preprocessor handling */ |
| 3029 | |
| 3030 | |
| 3031 | linebuffer_init (&lbs[0].lb); |
| 3032 | linebuffer_init (&lbs[1].lb); |
| 3033 | if (cstack.size == 0) |
| 3034 | { |
| 3035 | cstack.size = (DEBUG) ? 1 : 4; |
| 3036 | cstack.nl = 0; |
| 3037 | cstack.cname = xnew (cstack.size, char *); |
| 3038 | cstack.bracelev = xnew (cstack.size, int); |
| 3039 | } |
| 3040 | |
| 3041 | tokoff = toklen = typdefbracelev = 0; /* keep compiler quiet */ |
| 3042 | curndx = newndx = 0; |
| 3043 | lp = curlb.buffer; |
| 3044 | *lp = 0; |
| 3045 | |
| 3046 | fvdef = fvnone; fvextern = FALSE; typdef = tnone; |
| 3047 | structdef = snone; definedef = dnone; objdef = onone; |
| 3048 | yacc_rules = FALSE; |
| 3049 | midtoken = inquote = inchar = incomm = quotednl = FALSE; |
| 3050 | token.valid = savetoken.valid = FALSE; |
| 3051 | bracelev = bracketlev = parlev = attrparlev = templatelev = 0; |
| 3052 | if (cjava) |
| 3053 | { qualifier = "."; qlen = 1; } |
| 3054 | else |
| 3055 | { qualifier = "::"; qlen = 2; } |
| 3056 | |
| 3057 | |
| 3058 | while (!feof (inf)) |
| 3059 | { |
| 3060 | c = *lp++; |
| 3061 | if (c == '\\') |
| 3062 | { |
| 3063 | /* If we are at the end of the line, the next character is a |
| 3064 | '\0'; do not skip it, because it is what tells us |
| 3065 | to read the next line. */ |
| 3066 | if (*lp == '\0') |
| 3067 | { |
| 3068 | quotednl = TRUE; |
| 3069 | continue; |
| 3070 | } |
| 3071 | lp++; |
| 3072 | c = ' '; |
| 3073 | } |
| 3074 | else if (incomm) |
| 3075 | { |
| 3076 | switch (c) |
| 3077 | { |
| 3078 | case '*': |
| 3079 | if (*lp == '/') |
| 3080 | { |
| 3081 | c = *lp++; |
| 3082 | incomm = FALSE; |
| 3083 | } |
| 3084 | break; |
| 3085 | case '\0': |
| 3086 | /* Newlines inside comments do not end macro definitions in |
| 3087 | traditional cpp. */ |
| 3088 | CNL_SAVE_DEFINEDEF (); |
| 3089 | break; |
| 3090 | } |
| 3091 | continue; |
| 3092 | } |
| 3093 | else if (inquote) |
| 3094 | { |
| 3095 | switch (c) |
| 3096 | { |
| 3097 | case '"': |
| 3098 | inquote = FALSE; |
| 3099 | break; |
| 3100 | case '\0': |
| 3101 | /* Newlines inside strings do not end macro definitions |
| 3102 | in traditional cpp, even though compilers don't |
| 3103 | usually accept them. */ |
| 3104 | CNL_SAVE_DEFINEDEF (); |
| 3105 | break; |
| 3106 | } |
| 3107 | continue; |
| 3108 | } |
| 3109 | else if (inchar) |
| 3110 | { |
| 3111 | switch (c) |
| 3112 | { |
| 3113 | case '\0': |
| 3114 | /* Hmmm, something went wrong. */ |
| 3115 | CNL (); |
| 3116 | /* FALLTHRU */ |
| 3117 | case '\'': |
| 3118 | inchar = FALSE; |
| 3119 | break; |
| 3120 | } |
| 3121 | continue; |
| 3122 | } |
| 3123 | else switch (c) |
| 3124 | { |
| 3125 | case '"': |
| 3126 | inquote = TRUE; |
| 3127 | if (bracketlev > 0) |
| 3128 | continue; |
| 3129 | if (inattribute) |
| 3130 | break; |
| 3131 | switch (fvdef) |
| 3132 | { |
| 3133 | case fdefunkey: |
| 3134 | case fstartlist: |
| 3135 | case finlist: |
| 3136 | case fignore: |
| 3137 | case vignore: |
| 3138 | break; |
| 3139 | default: |
| 3140 | fvextern = FALSE; |
| 3141 | fvdef = fvnone; |
| 3142 | } |
| 3143 | continue; |
| 3144 | case '\'': |
| 3145 | inchar = TRUE; |
| 3146 | if (bracketlev > 0) |
| 3147 | continue; |
| 3148 | if (inattribute) |
| 3149 | break; |
| 3150 | if (fvdef != finlist && fvdef != fignore && fvdef != vignore) |
| 3151 | { |
| 3152 | fvextern = FALSE; |
| 3153 | fvdef = fvnone; |
| 3154 | } |
| 3155 | continue; |
| 3156 | case '/': |
| 3157 | if (*lp == '*') |
| 3158 | { |
| 3159 | incomm = TRUE; |
| 3160 | lp++; |
| 3161 | c = ' '; |
| 3162 | if (bracketlev > 0) |
| 3163 | continue; |
| 3164 | } |
| 3165 | else if (/* cplpl && */ *lp == '/') |
| 3166 | { |
| 3167 | c = '\0'; |
| 3168 | } |
| 3169 | break; |
| 3170 | case '%': |
| 3171 | if ((c_ext & YACC) && *lp == '%') |
| 3172 | { |
| 3173 | /* Entering or exiting rules section in yacc file. */ |
| 3174 | lp++; |
| 3175 | definedef = dnone; fvdef = fvnone; fvextern = FALSE; |
| 3176 | typdef = tnone; structdef = snone; |
| 3177 | midtoken = inquote = inchar = incomm = quotednl = FALSE; |
| 3178 | bracelev = 0; |
| 3179 | yacc_rules = !yacc_rules; |
| 3180 | continue; |
| 3181 | } |
| 3182 | else |
| 3183 | break; |
| 3184 | case '#': |
| 3185 | if (definedef == dnone) |
| 3186 | { |
| 3187 | char *cp; |
| 3188 | bool cpptoken = TRUE; |
| 3189 | |
| 3190 | /* Look back on this line. If all blanks, or nonblanks |
| 3191 | followed by an end of comment, this is a preprocessor |
| 3192 | token. */ |
| 3193 | for (cp = newlb.buffer; cp < lp-1; cp++) |
| 3194 | if (!iswhite (*cp)) |
| 3195 | { |
| 3196 | if (*cp == '*' && cp[1] == '/') |
| 3197 | { |
| 3198 | cp++; |
| 3199 | cpptoken = TRUE; |
| 3200 | } |
| 3201 | else |
| 3202 | cpptoken = FALSE; |
| 3203 | } |
| 3204 | if (cpptoken) |
| 3205 | definedef = dsharpseen; |
| 3206 | } /* if (definedef == dnone) */ |
| 3207 | continue; |
| 3208 | case '[': |
| 3209 | bracketlev++; |
| 3210 | continue; |
| 3211 | default: |
| 3212 | if (bracketlev > 0) |
| 3213 | { |
| 3214 | if (c == ']') |
| 3215 | --bracketlev; |
| 3216 | else if (c == '\0') |
| 3217 | CNL_SAVE_DEFINEDEF (); |
| 3218 | continue; |
| 3219 | } |
| 3220 | break; |
| 3221 | } /* switch (c) */ |
| 3222 | |
| 3223 | |
| 3224 | /* Consider token only if some involved conditions are satisfied. */ |
| 3225 | if (typdef != tignore |
| 3226 | && definedef != dignorerest |
| 3227 | && fvdef != finlist |
| 3228 | && templatelev == 0 |
| 3229 | && (definedef != dnone |
| 3230 | || structdef != scolonseen) |
| 3231 | && !inattribute) |
| 3232 | { |
| 3233 | if (midtoken) |
| 3234 | { |
| 3235 | if (endtoken (c)) |
| 3236 | { |
| 3237 | if (c == ':' && *lp == ':' && begtoken (lp[1])) |
| 3238 | /* This handles :: in the middle, |
| 3239 | but not at the beginning of an identifier. |
| 3240 | Also, space-separated :: is not recognized. */ |
| 3241 | { |
| 3242 | if (c_ext & C_AUTO) /* automatic detection of C++ */ |
| 3243 | c_ext = (c_ext | C_PLPL) & ~C_AUTO; |
| 3244 | lp += 2; |
| 3245 | toklen += 2; |
| 3246 | c = lp[-1]; |
| 3247 | goto still_in_token; |
| 3248 | } |
| 3249 | else |
| 3250 | { |
| 3251 | bool funorvar = FALSE; |
| 3252 | |
| 3253 | if (yacc_rules |
| 3254 | || consider_token (newlb.buffer + tokoff, toklen, c, |
| 3255 | &c_ext, bracelev, parlev, |
| 3256 | &funorvar)) |
| 3257 | { |
| 3258 | if (fvdef == foperator) |
| 3259 | { |
| 3260 | char *oldlp = lp; |
| 3261 | lp = skip_spaces (lp-1); |
| 3262 | if (*lp != '\0') |
| 3263 | lp += 1; |
| 3264 | while (*lp != '\0' |
| 3265 | && !iswhite (*lp) && *lp != '(') |
| 3266 | lp += 1; |
| 3267 | c = *lp++; |
| 3268 | toklen += lp - oldlp; |
| 3269 | } |
| 3270 | token.named = FALSE; |
| 3271 | if (!plainc |
| 3272 | && nestlev > 0 && definedef == dnone) |
| 3273 | /* in struct body */ |
| 3274 | { |
| 3275 | int len; |
| 3276 | write_classname (&token_name, qualifier); |
| 3277 | len = token_name.len; |
| 3278 | linebuffer_setlen (&token_name, len+qlen+toklen); |
| 3279 | sprintf (token_name.buffer + len, "%s%.*s", |
| 3280 | qualifier, toklen, newlb.buffer + tokoff); |
| 3281 | token.named = TRUE; |
| 3282 | } |
| 3283 | else if (objdef == ocatseen) |
| 3284 | /* Objective C category */ |
| 3285 | { |
| 3286 | int len = strlen (objtag) + 2 + toklen; |
| 3287 | linebuffer_setlen (&token_name, len); |
| 3288 | sprintf (token_name.buffer, "%s(%.*s)", |
| 3289 | objtag, toklen, newlb.buffer + tokoff); |
| 3290 | token.named = TRUE; |
| 3291 | } |
| 3292 | else if (objdef == omethodtag |
| 3293 | || objdef == omethodparm) |
| 3294 | /* Objective C method */ |
| 3295 | { |
| 3296 | token.named = TRUE; |
| 3297 | } |
| 3298 | else if (fvdef == fdefunname) |
| 3299 | /* GNU DEFUN and similar macros */ |
| 3300 | { |
| 3301 | bool defun = (newlb.buffer[tokoff] == 'F'); |
| 3302 | int off = tokoff; |
| 3303 | int len = toklen; |
| 3304 | |
| 3305 | /* Rewrite the tag so that emacs lisp DEFUNs |
| 3306 | can be found by their elisp name */ |
| 3307 | if (defun) |
| 3308 | { |
| 3309 | off += 1; |
| 3310 | len -= 1; |
| 3311 | } |
| 3312 | linebuffer_setlen (&token_name, len); |
| 3313 | memcpy (token_name.buffer, |
| 3314 | newlb.buffer + off, len); |
| 3315 | token_name.buffer[len] = '\0'; |
| 3316 | if (defun) |
| 3317 | while (--len >= 0) |
| 3318 | if (token_name.buffer[len] == '_') |
| 3319 | token_name.buffer[len] = '-'; |
| 3320 | token.named = defun; |
| 3321 | } |
| 3322 | else |
| 3323 | { |
| 3324 | linebuffer_setlen (&token_name, toklen); |
| 3325 | memcpy (token_name.buffer, |
| 3326 | newlb.buffer + tokoff, toklen); |
| 3327 | token_name.buffer[toklen] = '\0'; |
| 3328 | /* Name macros and members. */ |
| 3329 | token.named = (structdef == stagseen |
| 3330 | || typdef == ttypeseen |
| 3331 | || typdef == tend |
| 3332 | || (funorvar |
| 3333 | && definedef == dignorerest) |
| 3334 | || (funorvar |
| 3335 | && definedef == dnone |
| 3336 | && structdef == snone |
| 3337 | && bracelev > 0)); |
| 3338 | } |
| 3339 | token.lineno = lineno; |
| 3340 | token.offset = tokoff; |
| 3341 | token.length = toklen; |
| 3342 | token.line = newlb.buffer; |
| 3343 | token.linepos = newlinepos; |
| 3344 | token.valid = TRUE; |
| 3345 | |
| 3346 | if (definedef == dnone |
| 3347 | && (fvdef == fvnameseen |
| 3348 | || fvdef == foperator |
| 3349 | || structdef == stagseen |
| 3350 | || typdef == tend |
| 3351 | || typdef == ttypeseen |
| 3352 | || objdef != onone)) |
| 3353 | { |
| 3354 | if (current_lb_is_new) |
| 3355 | switch_line_buffers (); |
| 3356 | } |
| 3357 | else if (definedef != dnone |
| 3358 | || fvdef == fdefunname |
| 3359 | || instruct) |
| 3360 | make_C_tag (funorvar); |
| 3361 | } |
| 3362 | else /* not yacc and consider_token failed */ |
| 3363 | { |
| 3364 | if (inattribute && fvdef == fignore) |
| 3365 | { |
| 3366 | /* We have just met __attribute__ after a |
| 3367 | function parameter list: do not tag the |
| 3368 | function again. */ |
| 3369 | fvdef = fvnone; |
| 3370 | } |
| 3371 | } |
| 3372 | midtoken = FALSE; |
| 3373 | } |
| 3374 | } /* if (endtoken (c)) */ |
| 3375 | else if (intoken (c)) |
| 3376 | still_in_token: |
| 3377 | { |
| 3378 | toklen++; |
| 3379 | continue; |
| 3380 | } |
| 3381 | } /* if (midtoken) */ |
| 3382 | else if (begtoken (c)) |
| 3383 | { |
| 3384 | switch (definedef) |
| 3385 | { |
| 3386 | case dnone: |
| 3387 | switch (fvdef) |
| 3388 | { |
| 3389 | case fstartlist: |
| 3390 | /* This prevents tagging fb in |
| 3391 | void (__attribute__((noreturn)) *fb) (void); |
| 3392 | Fixing this is not easy and not very important. */ |
| 3393 | fvdef = finlist; |
| 3394 | continue; |
| 3395 | case flistseen: |
| 3396 | if (plainc || declarations) |
| 3397 | { |
| 3398 | make_C_tag (TRUE); /* a function */ |
| 3399 | fvdef = fignore; |
| 3400 | } |
| 3401 | break; |
| 3402 | } |
| 3403 | if (structdef == stagseen && !cjava) |
| 3404 | { |
| 3405 | popclass_above (bracelev); |
| 3406 | structdef = snone; |
| 3407 | } |
| 3408 | break; |
| 3409 | case dsharpseen: |
| 3410 | savetoken = token; |
| 3411 | break; |
| 3412 | } |
| 3413 | if (!yacc_rules || lp == newlb.buffer + 1) |
| 3414 | { |
| 3415 | tokoff = lp - 1 - newlb.buffer; |
| 3416 | toklen = 1; |
| 3417 | midtoken = TRUE; |
| 3418 | } |
| 3419 | continue; |
| 3420 | } /* if (begtoken) */ |
| 3421 | } /* if must look at token */ |
| 3422 | |
| 3423 | |
| 3424 | /* Detect end of line, colon, comma, semicolon and various braces |
| 3425 | after having handled a token.*/ |
| 3426 | switch (c) |
| 3427 | { |
| 3428 | case ':': |
| 3429 | if (inattribute) |
| 3430 | break; |
| 3431 | if (yacc_rules && token.offset == 0 && token.valid) |
| 3432 | { |
| 3433 | make_C_tag (FALSE); /* a yacc function */ |
| 3434 | break; |
| 3435 | } |
| 3436 | if (definedef != dnone) |
| 3437 | break; |
| 3438 | switch (objdef) |
| 3439 | { |
| 3440 | case otagseen: |
| 3441 | objdef = oignore; |
| 3442 | make_C_tag (TRUE); /* an Objective C class */ |
| 3443 | break; |
| 3444 | case omethodtag: |
| 3445 | case omethodparm: |
| 3446 | objdef = omethodcolon; |
| 3447 | linebuffer_setlen (&token_name, token_name.len + 1); |
| 3448 | strcat (token_name.buffer, ":"); |
| 3449 | break; |
| 3450 | } |
| 3451 | if (structdef == stagseen) |
| 3452 | { |
| 3453 | structdef = scolonseen; |
| 3454 | break; |
| 3455 | } |
| 3456 | /* Should be useless, but may be work as a safety net. */ |
| 3457 | if (cplpl && fvdef == flistseen) |
| 3458 | { |
| 3459 | make_C_tag (TRUE); /* a function */ |
| 3460 | fvdef = fignore; |
| 3461 | break; |
| 3462 | } |
| 3463 | break; |
| 3464 | case ';': |
| 3465 | if (definedef != dnone || inattribute) |
| 3466 | break; |
| 3467 | switch (typdef) |
| 3468 | { |
| 3469 | case tend: |
| 3470 | case ttypeseen: |
| 3471 | make_C_tag (FALSE); /* a typedef */ |
| 3472 | typdef = tnone; |
| 3473 | fvdef = fvnone; |
| 3474 | break; |
| 3475 | case tnone: |
| 3476 | case tinbody: |
| 3477 | case tignore: |
| 3478 | switch (fvdef) |
| 3479 | { |
| 3480 | case fignore: |
| 3481 | if (typdef == tignore || cplpl) |
| 3482 | fvdef = fvnone; |
| 3483 | break; |
| 3484 | case fvnameseen: |
| 3485 | if ((globals && bracelev == 0 && (!fvextern || declarations)) |
| 3486 | || (members && instruct)) |
| 3487 | make_C_tag (FALSE); /* a variable */ |
| 3488 | fvextern = FALSE; |
| 3489 | fvdef = fvnone; |
| 3490 | token.valid = FALSE; |
| 3491 | break; |
| 3492 | case flistseen: |
| 3493 | if ((declarations |
| 3494 | && (cplpl || !instruct) |
| 3495 | && (typdef == tnone || (typdef != tignore && instruct))) |
| 3496 | || (members |
| 3497 | && plainc && instruct)) |
| 3498 | make_C_tag (TRUE); /* a function */ |
| 3499 | /* FALLTHRU */ |
| 3500 | default: |
| 3501 | fvextern = FALSE; |
| 3502 | fvdef = fvnone; |
| 3503 | if (declarations |
| 3504 | && cplpl && structdef == stagseen) |
| 3505 | make_C_tag (FALSE); /* forward declaration */ |
| 3506 | else |
| 3507 | token.valid = FALSE; |
| 3508 | } /* switch (fvdef) */ |
| 3509 | /* FALLTHRU */ |
| 3510 | default: |
| 3511 | if (!instruct) |
| 3512 | typdef = tnone; |
| 3513 | } |
| 3514 | if (structdef == stagseen) |
| 3515 | structdef = snone; |
| 3516 | break; |
| 3517 | case ',': |
| 3518 | if (definedef != dnone || inattribute) |
| 3519 | break; |
| 3520 | switch (objdef) |
| 3521 | { |
| 3522 | case omethodtag: |
| 3523 | case omethodparm: |
| 3524 | make_C_tag (TRUE); /* an Objective C method */ |
| 3525 | objdef = oinbody; |
| 3526 | break; |
| 3527 | } |
| 3528 | switch (fvdef) |
| 3529 | { |
| 3530 | case fdefunkey: |
| 3531 | case foperator: |
| 3532 | case fstartlist: |
| 3533 | case finlist: |
| 3534 | case fignore: |
| 3535 | case vignore: |
| 3536 | break; |
| 3537 | case fdefunname: |
| 3538 | fvdef = fignore; |
| 3539 | break; |
| 3540 | case fvnameseen: |
| 3541 | if (parlev == 0 |
| 3542 | && ((globals |
| 3543 | && bracelev == 0 |
| 3544 | && templatelev == 0 |
| 3545 | && (!fvextern || declarations)) |
| 3546 | || (members && instruct))) |
| 3547 | make_C_tag (FALSE); /* a variable */ |
| 3548 | break; |
| 3549 | case flistseen: |
| 3550 | if ((declarations && typdef == tnone && !instruct) |
| 3551 | || (members && typdef != tignore && instruct)) |
| 3552 | { |
| 3553 | make_C_tag (TRUE); /* a function */ |
| 3554 | fvdef = fvnameseen; |
| 3555 | } |
| 3556 | else if (!declarations) |
| 3557 | fvdef = fvnone; |
| 3558 | token.valid = FALSE; |
| 3559 | break; |
| 3560 | default: |
| 3561 | fvdef = fvnone; |
| 3562 | } |
| 3563 | if (structdef == stagseen) |
| 3564 | structdef = snone; |
| 3565 | break; |
| 3566 | case ']': |
| 3567 | if (definedef != dnone || inattribute) |
| 3568 | break; |
| 3569 | if (structdef == stagseen) |
| 3570 | structdef = snone; |
| 3571 | switch (typdef) |
| 3572 | { |
| 3573 | case ttypeseen: |
| 3574 | case tend: |
| 3575 | typdef = tignore; |
| 3576 | make_C_tag (FALSE); /* a typedef */ |
| 3577 | break; |
| 3578 | case tnone: |
| 3579 | case tinbody: |
| 3580 | switch (fvdef) |
| 3581 | { |
| 3582 | case foperator: |
| 3583 | case finlist: |
| 3584 | case fignore: |
| 3585 | case vignore: |
| 3586 | break; |
| 3587 | case fvnameseen: |
| 3588 | if ((members && bracelev == 1) |
| 3589 | || (globals && bracelev == 0 |
| 3590 | && (!fvextern || declarations))) |
| 3591 | make_C_tag (FALSE); /* a variable */ |
| 3592 | /* FALLTHRU */ |
| 3593 | default: |
| 3594 | fvdef = fvnone; |
| 3595 | } |
| 3596 | break; |
| 3597 | } |
| 3598 | break; |
| 3599 | case '(': |
| 3600 | if (inattribute) |
| 3601 | { |
| 3602 | attrparlev++; |
| 3603 | break; |
| 3604 | } |
| 3605 | if (definedef != dnone) |
| 3606 | break; |
| 3607 | if (objdef == otagseen && parlev == 0) |
| 3608 | objdef = oparenseen; |
| 3609 | switch (fvdef) |
| 3610 | { |
| 3611 | case fvnameseen: |
| 3612 | if (typdef == ttypeseen |
| 3613 | && *lp != '*' |
| 3614 | && !instruct) |
| 3615 | { |
| 3616 | /* This handles constructs like: |
| 3617 | typedef void OperatorFun (int fun); */ |
| 3618 | make_C_tag (FALSE); |
| 3619 | typdef = tignore; |
| 3620 | fvdef = fignore; |
| 3621 | break; |
| 3622 | } |
| 3623 | /* FALLTHRU */ |
| 3624 | case foperator: |
| 3625 | fvdef = fstartlist; |
| 3626 | break; |
| 3627 | case flistseen: |
| 3628 | fvdef = finlist; |
| 3629 | break; |
| 3630 | } |
| 3631 | parlev++; |
| 3632 | break; |
| 3633 | case ')': |
| 3634 | if (inattribute) |
| 3635 | { |
| 3636 | if (--attrparlev == 0) |
| 3637 | inattribute = FALSE; |
| 3638 | break; |
| 3639 | } |
| 3640 | if (definedef != dnone) |
| 3641 | break; |
| 3642 | if (objdef == ocatseen && parlev == 1) |
| 3643 | { |
| 3644 | make_C_tag (TRUE); /* an Objective C category */ |
| 3645 | objdef = oignore; |
| 3646 | } |
| 3647 | if (--parlev == 0) |
| 3648 | { |
| 3649 | switch (fvdef) |
| 3650 | { |
| 3651 | case fstartlist: |
| 3652 | case finlist: |
| 3653 | fvdef = flistseen; |
| 3654 | break; |
| 3655 | } |
| 3656 | if (!instruct |
| 3657 | && (typdef == tend |
| 3658 | || typdef == ttypeseen)) |
| 3659 | { |
| 3660 | typdef = tignore; |
| 3661 | make_C_tag (FALSE); /* a typedef */ |
| 3662 | } |
| 3663 | } |
| 3664 | else if (parlev < 0) /* can happen due to ill-conceived #if's. */ |
| 3665 | parlev = 0; |
| 3666 | break; |
| 3667 | case '{': |
| 3668 | if (definedef != dnone) |
| 3669 | break; |
| 3670 | if (typdef == ttypeseen) |
| 3671 | { |
| 3672 | /* Whenever typdef is set to tinbody (currently only |
| 3673 | here), typdefbracelev should be set to bracelev. */ |
| 3674 | typdef = tinbody; |
| 3675 | typdefbracelev = bracelev; |
| 3676 | } |
| 3677 | switch (fvdef) |
| 3678 | { |
| 3679 | case flistseen: |
| 3680 | make_C_tag (TRUE); /* a function */ |
| 3681 | /* FALLTHRU */ |
| 3682 | case fignore: |
| 3683 | fvdef = fvnone; |
| 3684 | break; |
| 3685 | case fvnone: |
| 3686 | switch (objdef) |
| 3687 | { |
| 3688 | case otagseen: |
| 3689 | make_C_tag (TRUE); /* an Objective C class */ |
| 3690 | objdef = oignore; |
| 3691 | break; |
| 3692 | case omethodtag: |
| 3693 | case omethodparm: |
| 3694 | make_C_tag (TRUE); /* an Objective C method */ |
| 3695 | objdef = oinbody; |
| 3696 | break; |
| 3697 | default: |
| 3698 | /* Neutralize `extern "C" {' grot. */ |
| 3699 | if (bracelev == 0 && structdef == snone && nestlev == 0 |
| 3700 | && typdef == tnone) |
| 3701 | bracelev = -1; |
| 3702 | } |
| 3703 | break; |
| 3704 | } |
| 3705 | switch (structdef) |
| 3706 | { |
| 3707 | case skeyseen: /* unnamed struct */ |
| 3708 | pushclass_above (bracelev, NULL, 0); |
| 3709 | structdef = snone; |
| 3710 | break; |
| 3711 | case stagseen: /* named struct or enum */ |
| 3712 | case scolonseen: /* a class */ |
| 3713 | pushclass_above (bracelev,token.line+token.offset, token.length); |
| 3714 | structdef = snone; |
| 3715 | make_C_tag (FALSE); /* a struct or enum */ |
| 3716 | break; |
| 3717 | } |
| 3718 | bracelev += 1; |
| 3719 | break; |
| 3720 | case '*': |
| 3721 | if (definedef != dnone) |
| 3722 | break; |
| 3723 | if (fvdef == fstartlist) |
| 3724 | { |
| 3725 | fvdef = fvnone; /* avoid tagging `foo' in `foo (*bar()) ()' */ |
| 3726 | token.valid = FALSE; |
| 3727 | } |
| 3728 | break; |
| 3729 | case '}': |
| 3730 | if (definedef != dnone) |
| 3731 | break; |
| 3732 | bracelev -= 1; |
| 3733 | if (!ignoreindent && lp == newlb.buffer + 1) |
| 3734 | { |
| 3735 | if (bracelev != 0) |
| 3736 | token.valid = FALSE; /* unexpected value, token unreliable */ |
| 3737 | bracelev = 0; /* reset brace level if first column */ |
| 3738 | parlev = 0; /* also reset paren level, just in case... */ |
| 3739 | } |
| 3740 | else if (bracelev < 0) |
| 3741 | { |
| 3742 | token.valid = FALSE; /* something gone amiss, token unreliable */ |
| 3743 | bracelev = 0; |
| 3744 | } |
| 3745 | if (bracelev == 0 && fvdef == vignore) |
| 3746 | fvdef = fvnone; /* end of function */ |
| 3747 | popclass_above (bracelev); |
| 3748 | structdef = snone; |
| 3749 | /* Only if typdef == tinbody is typdefbracelev significant. */ |
| 3750 | if (typdef == tinbody && bracelev <= typdefbracelev) |
| 3751 | { |
| 3752 | assert (bracelev == typdefbracelev); |
| 3753 | typdef = tend; |
| 3754 | } |
| 3755 | break; |
| 3756 | case '=': |
| 3757 | if (definedef != dnone) |
| 3758 | break; |
| 3759 | switch (fvdef) |
| 3760 | { |
| 3761 | case foperator: |
| 3762 | case finlist: |
| 3763 | case fignore: |
| 3764 | case vignore: |
| 3765 | break; |
| 3766 | case fvnameseen: |
| 3767 | if ((members && bracelev == 1) |
| 3768 | || (globals && bracelev == 0 && (!fvextern || declarations))) |
| 3769 | make_C_tag (FALSE); /* a variable */ |
| 3770 | /* FALLTHRU */ |
| 3771 | default: |
| 3772 | fvdef = vignore; |
| 3773 | } |
| 3774 | break; |
| 3775 | case '<': |
| 3776 | if (cplpl |
| 3777 | && (structdef == stagseen || fvdef == fvnameseen)) |
| 3778 | { |
| 3779 | templatelev++; |
| 3780 | break; |
| 3781 | } |
| 3782 | goto resetfvdef; |
| 3783 | case '>': |
| 3784 | if (templatelev > 0) |
| 3785 | { |
| 3786 | templatelev--; |
| 3787 | break; |
| 3788 | } |
| 3789 | goto resetfvdef; |
| 3790 | case '+': |
| 3791 | case '-': |
| 3792 | if (objdef == oinbody && bracelev == 0) |
| 3793 | { |
| 3794 | objdef = omethodsign; |
| 3795 | break; |
| 3796 | } |
| 3797 | /* FALLTHRU */ |
| 3798 | resetfvdef: |
| 3799 | case '#': case '~': case '&': case '%': case '/': |
| 3800 | case '|': case '^': case '!': case '.': case '?': |
| 3801 | if (definedef != dnone) |
| 3802 | break; |
| 3803 | /* These surely cannot follow a function tag in C. */ |
| 3804 | switch (fvdef) |
| 3805 | { |
| 3806 | case foperator: |
| 3807 | case finlist: |
| 3808 | case fignore: |
| 3809 | case vignore: |
| 3810 | break; |
| 3811 | default: |
| 3812 | fvdef = fvnone; |
| 3813 | } |
| 3814 | break; |
| 3815 | case '\0': |
| 3816 | if (objdef == otagseen) |
| 3817 | { |
| 3818 | make_C_tag (TRUE); /* an Objective C class */ |
| 3819 | objdef = oignore; |
| 3820 | } |
| 3821 | /* If a macro spans multiple lines don't reset its state. */ |
| 3822 | if (quotednl) |
| 3823 | CNL_SAVE_DEFINEDEF (); |
| 3824 | else |
| 3825 | CNL (); |
| 3826 | break; |
| 3827 | } /* switch (c) */ |
| 3828 | |
| 3829 | } /* while not eof */ |
| 3830 | |
| 3831 | free (lbs[0].lb.buffer); |
| 3832 | free (lbs[1].lb.buffer); |
| 3833 | } |
| 3834 | |
| 3835 | /* |
| 3836 | * Process either a C++ file or a C file depending on the setting |
| 3837 | * of a global flag. |
| 3838 | */ |
| 3839 | static void |
| 3840 | default_C_entries (FILE *inf) |
| 3841 | { |
| 3842 | C_entries (cplusplus ? C_PLPL : C_AUTO, inf); |
| 3843 | } |
| 3844 | |
| 3845 | /* Always do plain C. */ |
| 3846 | static void |
| 3847 | plain_C_entries (FILE *inf) |
| 3848 | { |
| 3849 | C_entries (0, inf); |
| 3850 | } |
| 3851 | |
| 3852 | /* Always do C++. */ |
| 3853 | static void |
| 3854 | Cplusplus_entries (FILE *inf) |
| 3855 | { |
| 3856 | C_entries (C_PLPL, inf); |
| 3857 | } |
| 3858 | |
| 3859 | /* Always do Java. */ |
| 3860 | static void |
| 3861 | Cjava_entries (FILE *inf) |
| 3862 | { |
| 3863 | C_entries (C_JAVA, inf); |
| 3864 | } |
| 3865 | |
| 3866 | /* Always do C*. */ |
| 3867 | static void |
| 3868 | Cstar_entries (FILE *inf) |
| 3869 | { |
| 3870 | C_entries (C_STAR, inf); |
| 3871 | } |
| 3872 | |
| 3873 | /* Always do Yacc. */ |
| 3874 | static void |
| 3875 | Yacc_entries (FILE *inf) |
| 3876 | { |
| 3877 | C_entries (YACC, inf); |
| 3878 | } |
| 3879 | |
| 3880 | \f |
| 3881 | /* Useful macros. */ |
| 3882 | #define LOOP_ON_INPUT_LINES(file_pointer, line_buffer, char_pointer) \ |
| 3883 | for (; /* loop initialization */ \ |
| 3884 | !feof (file_pointer) /* loop test */ \ |
| 3885 | && /* instructions at start of loop */ \ |
| 3886 | (readline (&line_buffer, file_pointer), \ |
| 3887 | char_pointer = line_buffer.buffer, \ |
| 3888 | TRUE); \ |
| 3889 | ) |
| 3890 | |
| 3891 | #define LOOKING_AT(cp, kw) /* kw is the keyword, a literal string */ \ |
| 3892 | ((assert ("" kw), TRUE) /* syntax error if not a literal string */ \ |
| 3893 | && strneq ((cp), kw, sizeof (kw)-1) /* cp points at kw */ \ |
| 3894 | && notinname ((cp)[sizeof (kw)-1]) /* end of kw */ \ |
| 3895 | && ((cp) = skip_spaces ((cp)+sizeof (kw)-1))) /* skip spaces */ |
| 3896 | |
| 3897 | /* Similar to LOOKING_AT but does not use notinname, does not skip */ |
| 3898 | #define LOOKING_AT_NOCASE(cp, kw) /* the keyword is a literal string */ \ |
| 3899 | ((assert ("" kw), TRUE) /* syntax error if not a literal string */ \ |
| 3900 | && strncaseeq ((cp), kw, sizeof (kw)-1) /* cp points at kw */ \ |
| 3901 | && ((cp) += sizeof (kw)-1)) /* skip spaces */ |
| 3902 | |
| 3903 | /* |
| 3904 | * Read a file, but do no processing. This is used to do regexp |
| 3905 | * matching on files that have no language defined. |
| 3906 | */ |
| 3907 | static void |
| 3908 | just_read_file (FILE *inf) |
| 3909 | { |
| 3910 | while (!feof (inf)) |
| 3911 | readline (&lb, inf); |
| 3912 | } |
| 3913 | |
| 3914 | \f |
| 3915 | /* Fortran parsing */ |
| 3916 | |
| 3917 | static void F_takeprec (void); |
| 3918 | static void F_getit (FILE *); |
| 3919 | |
| 3920 | static void |
| 3921 | F_takeprec (void) |
| 3922 | { |
| 3923 | dbp = skip_spaces (dbp); |
| 3924 | if (*dbp != '*') |
| 3925 | return; |
| 3926 | dbp++; |
| 3927 | dbp = skip_spaces (dbp); |
| 3928 | if (strneq (dbp, "(*)", 3)) |
| 3929 | { |
| 3930 | dbp += 3; |
| 3931 | return; |
| 3932 | } |
| 3933 | if (!ISDIGIT (*dbp)) |
| 3934 | { |
| 3935 | --dbp; /* force failure */ |
| 3936 | return; |
| 3937 | } |
| 3938 | do |
| 3939 | dbp++; |
| 3940 | while (ISDIGIT (*dbp)); |
| 3941 | } |
| 3942 | |
| 3943 | static void |
| 3944 | F_getit (FILE *inf) |
| 3945 | { |
| 3946 | register char *cp; |
| 3947 | |
| 3948 | dbp = skip_spaces (dbp); |
| 3949 | if (*dbp == '\0') |
| 3950 | { |
| 3951 | readline (&lb, inf); |
| 3952 | dbp = lb.buffer; |
| 3953 | if (dbp[5] != '&') |
| 3954 | return; |
| 3955 | dbp += 6; |
| 3956 | dbp = skip_spaces (dbp); |
| 3957 | } |
| 3958 | if (!ISALPHA (*dbp) && *dbp != '_' && *dbp != '$') |
| 3959 | return; |
| 3960 | for (cp = dbp + 1; *cp != '\0' && intoken (*cp); cp++) |
| 3961 | continue; |
| 3962 | make_tag (dbp, cp-dbp, TRUE, |
| 3963 | lb.buffer, cp - lb.buffer + 1, lineno, linecharno); |
| 3964 | } |
| 3965 | |
| 3966 | |
| 3967 | static void |
| 3968 | Fortran_functions (FILE *inf) |
| 3969 | { |
| 3970 | LOOP_ON_INPUT_LINES (inf, lb, dbp) |
| 3971 | { |
| 3972 | if (*dbp == '%') |
| 3973 | dbp++; /* Ratfor escape to fortran */ |
| 3974 | dbp = skip_spaces (dbp); |
| 3975 | if (*dbp == '\0') |
| 3976 | continue; |
| 3977 | |
| 3978 | if (LOOKING_AT_NOCASE (dbp, "recursive")) |
| 3979 | dbp = skip_spaces (dbp); |
| 3980 | |
| 3981 | if (LOOKING_AT_NOCASE (dbp, "pure")) |
| 3982 | dbp = skip_spaces (dbp); |
| 3983 | |
| 3984 | if (LOOKING_AT_NOCASE (dbp, "elemental")) |
| 3985 | dbp = skip_spaces (dbp); |
| 3986 | |
| 3987 | switch (lowcase (*dbp)) |
| 3988 | { |
| 3989 | case 'i': |
| 3990 | if (nocase_tail ("integer")) |
| 3991 | F_takeprec (); |
| 3992 | break; |
| 3993 | case 'r': |
| 3994 | if (nocase_tail ("real")) |
| 3995 | F_takeprec (); |
| 3996 | break; |
| 3997 | case 'l': |
| 3998 | if (nocase_tail ("logical")) |
| 3999 | F_takeprec (); |
| 4000 | break; |
| 4001 | case 'c': |
| 4002 | if (nocase_tail ("complex") || nocase_tail ("character")) |
| 4003 | F_takeprec (); |
| 4004 | break; |
| 4005 | case 'd': |
| 4006 | if (nocase_tail ("double")) |
| 4007 | { |
| 4008 | dbp = skip_spaces (dbp); |
| 4009 | if (*dbp == '\0') |
| 4010 | continue; |
| 4011 | if (nocase_tail ("precision")) |
| 4012 | break; |
| 4013 | continue; |
| 4014 | } |
| 4015 | break; |
| 4016 | } |
| 4017 | dbp = skip_spaces (dbp); |
| 4018 | if (*dbp == '\0') |
| 4019 | continue; |
| 4020 | switch (lowcase (*dbp)) |
| 4021 | { |
| 4022 | case 'f': |
| 4023 | if (nocase_tail ("function")) |
| 4024 | F_getit (inf); |
| 4025 | continue; |
| 4026 | case 's': |
| 4027 | if (nocase_tail ("subroutine")) |
| 4028 | F_getit (inf); |
| 4029 | continue; |
| 4030 | case 'e': |
| 4031 | if (nocase_tail ("entry")) |
| 4032 | F_getit (inf); |
| 4033 | continue; |
| 4034 | case 'b': |
| 4035 | if (nocase_tail ("blockdata") || nocase_tail ("block data")) |
| 4036 | { |
| 4037 | dbp = skip_spaces (dbp); |
| 4038 | if (*dbp == '\0') /* assume un-named */ |
| 4039 | make_tag ("blockdata", 9, TRUE, |
| 4040 | lb.buffer, dbp - lb.buffer, lineno, linecharno); |
| 4041 | else |
| 4042 | F_getit (inf); /* look for name */ |
| 4043 | } |
| 4044 | continue; |
| 4045 | } |
| 4046 | } |
| 4047 | } |
| 4048 | |
| 4049 | \f |
| 4050 | /* |
| 4051 | * Ada parsing |
| 4052 | * Original code by |
| 4053 | * Philippe Waroquiers (1998) |
| 4054 | */ |
| 4055 | |
| 4056 | /* Once we are positioned after an "interesting" keyword, let's get |
| 4057 | the real tag value necessary. */ |
| 4058 | static void |
| 4059 | Ada_getit (FILE *inf, const char *name_qualifier) |
| 4060 | { |
| 4061 | register char *cp; |
| 4062 | char *name; |
| 4063 | char c; |
| 4064 | |
| 4065 | while (!feof (inf)) |
| 4066 | { |
| 4067 | dbp = skip_spaces (dbp); |
| 4068 | if (*dbp == '\0' |
| 4069 | || (dbp[0] == '-' && dbp[1] == '-')) |
| 4070 | { |
| 4071 | readline (&lb, inf); |
| 4072 | dbp = lb.buffer; |
| 4073 | } |
| 4074 | switch (lowcase (*dbp)) |
| 4075 | { |
| 4076 | case 'b': |
| 4077 | if (nocase_tail ("body")) |
| 4078 | { |
| 4079 | /* Skipping body of procedure body or package body or .... |
| 4080 | resetting qualifier to body instead of spec. */ |
| 4081 | name_qualifier = "/b"; |
| 4082 | continue; |
| 4083 | } |
| 4084 | break; |
| 4085 | case 't': |
| 4086 | /* Skipping type of task type or protected type ... */ |
| 4087 | if (nocase_tail ("type")) |
| 4088 | continue; |
| 4089 | break; |
| 4090 | } |
| 4091 | if (*dbp == '"') |
| 4092 | { |
| 4093 | dbp += 1; |
| 4094 | for (cp = dbp; *cp != '\0' && *cp != '"'; cp++) |
| 4095 | continue; |
| 4096 | } |
| 4097 | else |
| 4098 | { |
| 4099 | dbp = skip_spaces (dbp); |
| 4100 | for (cp = dbp; |
| 4101 | (*cp != '\0' |
| 4102 | && (ISALPHA (*cp) || ISDIGIT (*cp) || *cp == '_' || *cp == '.')); |
| 4103 | cp++) |
| 4104 | continue; |
| 4105 | if (cp == dbp) |
| 4106 | return; |
| 4107 | } |
| 4108 | c = *cp; |
| 4109 | *cp = '\0'; |
| 4110 | name = concat (dbp, name_qualifier, ""); |
| 4111 | *cp = c; |
| 4112 | make_tag (name, strlen (name), TRUE, |
| 4113 | lb.buffer, cp - lb.buffer + 1, lineno, linecharno); |
| 4114 | free (name); |
| 4115 | if (c == '"') |
| 4116 | dbp = cp + 1; |
| 4117 | return; |
| 4118 | } |
| 4119 | } |
| 4120 | |
| 4121 | static void |
| 4122 | Ada_funcs (FILE *inf) |
| 4123 | { |
| 4124 | bool inquote = FALSE; |
| 4125 | bool skip_till_semicolumn = FALSE; |
| 4126 | |
| 4127 | LOOP_ON_INPUT_LINES (inf, lb, dbp) |
| 4128 | { |
| 4129 | while (*dbp != '\0') |
| 4130 | { |
| 4131 | /* Skip a string i.e. "abcd". */ |
| 4132 | if (inquote || (*dbp == '"')) |
| 4133 | { |
| 4134 | dbp = etags_strchr (dbp + !inquote, '"'); |
| 4135 | if (dbp != NULL) |
| 4136 | { |
| 4137 | inquote = FALSE; |
| 4138 | dbp += 1; |
| 4139 | continue; /* advance char */ |
| 4140 | } |
| 4141 | else |
| 4142 | { |
| 4143 | inquote = TRUE; |
| 4144 | break; /* advance line */ |
| 4145 | } |
| 4146 | } |
| 4147 | |
| 4148 | /* Skip comments. */ |
| 4149 | if (dbp[0] == '-' && dbp[1] == '-') |
| 4150 | break; /* advance line */ |
| 4151 | |
| 4152 | /* Skip character enclosed in single quote i.e. 'a' |
| 4153 | and skip single quote starting an attribute i.e. 'Image. */ |
| 4154 | if (*dbp == '\'') |
| 4155 | { |
| 4156 | dbp++ ; |
| 4157 | if (*dbp != '\0') |
| 4158 | dbp++; |
| 4159 | continue; |
| 4160 | } |
| 4161 | |
| 4162 | if (skip_till_semicolumn) |
| 4163 | { |
| 4164 | if (*dbp == ';') |
| 4165 | skip_till_semicolumn = FALSE; |
| 4166 | dbp++; |
| 4167 | continue; /* advance char */ |
| 4168 | } |
| 4169 | |
| 4170 | /* Search for beginning of a token. */ |
| 4171 | if (!begtoken (*dbp)) |
| 4172 | { |
| 4173 | dbp++; |
| 4174 | continue; /* advance char */ |
| 4175 | } |
| 4176 | |
| 4177 | /* We are at the beginning of a token. */ |
| 4178 | switch (lowcase (*dbp)) |
| 4179 | { |
| 4180 | case 'f': |
| 4181 | if (!packages_only && nocase_tail ("function")) |
| 4182 | Ada_getit (inf, "/f"); |
| 4183 | else |
| 4184 | break; /* from switch */ |
| 4185 | continue; /* advance char */ |
| 4186 | case 'p': |
| 4187 | if (!packages_only && nocase_tail ("procedure")) |
| 4188 | Ada_getit (inf, "/p"); |
| 4189 | else if (nocase_tail ("package")) |
| 4190 | Ada_getit (inf, "/s"); |
| 4191 | else if (nocase_tail ("protected")) /* protected type */ |
| 4192 | Ada_getit (inf, "/t"); |
| 4193 | else |
| 4194 | break; /* from switch */ |
| 4195 | continue; /* advance char */ |
| 4196 | |
| 4197 | case 'u': |
| 4198 | if (typedefs && !packages_only && nocase_tail ("use")) |
| 4199 | { |
| 4200 | /* when tagging types, avoid tagging use type Pack.Typename; |
| 4201 | for this, we will skip everything till a ; */ |
| 4202 | skip_till_semicolumn = TRUE; |
| 4203 | continue; /* advance char */ |
| 4204 | } |
| 4205 | |
| 4206 | case 't': |
| 4207 | if (!packages_only && nocase_tail ("task")) |
| 4208 | Ada_getit (inf, "/k"); |
| 4209 | else if (typedefs && !packages_only && nocase_tail ("type")) |
| 4210 | { |
| 4211 | Ada_getit (inf, "/t"); |
| 4212 | while (*dbp != '\0') |
| 4213 | dbp += 1; |
| 4214 | } |
| 4215 | else |
| 4216 | break; /* from switch */ |
| 4217 | continue; /* advance char */ |
| 4218 | } |
| 4219 | |
| 4220 | /* Look for the end of the token. */ |
| 4221 | while (!endtoken (*dbp)) |
| 4222 | dbp++; |
| 4223 | |
| 4224 | } /* advance char */ |
| 4225 | } /* advance line */ |
| 4226 | } |
| 4227 | |
| 4228 | \f |
| 4229 | /* |
| 4230 | * Unix and microcontroller assembly tag handling |
| 4231 | * Labels: /^[a-zA-Z_.$][a-zA_Z0-9_.$]*[: ^I^J]/ |
| 4232 | * Idea by Bob Weiner, Motorola Inc. (1994) |
| 4233 | */ |
| 4234 | static void |
| 4235 | Asm_labels (FILE *inf) |
| 4236 | { |
| 4237 | register char *cp; |
| 4238 | |
| 4239 | LOOP_ON_INPUT_LINES (inf, lb, cp) |
| 4240 | { |
| 4241 | /* If first char is alphabetic or one of [_.$], test for colon |
| 4242 | following identifier. */ |
| 4243 | if (ISALPHA (*cp) || *cp == '_' || *cp == '.' || *cp == '$') |
| 4244 | { |
| 4245 | /* Read past label. */ |
| 4246 | cp++; |
| 4247 | while (ISALNUM (*cp) || *cp == '_' || *cp == '.' || *cp == '$') |
| 4248 | cp++; |
| 4249 | if (*cp == ':' || iswhite (*cp)) |
| 4250 | /* Found end of label, so copy it and add it to the table. */ |
| 4251 | make_tag (lb.buffer, cp - lb.buffer, TRUE, |
| 4252 | lb.buffer, cp - lb.buffer + 1, lineno, linecharno); |
| 4253 | } |
| 4254 | } |
| 4255 | } |
| 4256 | |
| 4257 | \f |
| 4258 | /* |
| 4259 | * Perl support |
| 4260 | * Perl sub names: /^sub[ \t\n]+[^ \t\n{]+/ |
| 4261 | * /^use constant[ \t\n]+[^ \t\n{=,;]+/ |
| 4262 | * Perl variable names: /^(my|local).../ |
| 4263 | * Original code by Bart Robinson <lomew@cs.utah.edu> (1995) |
| 4264 | * Additions by Michael Ernst <mernst@alum.mit.edu> (1997) |
| 4265 | * Ideas by Kai Großjohann <Kai.Grossjohann@CS.Uni-Dortmund.DE> (2001) |
| 4266 | */ |
| 4267 | static void |
| 4268 | Perl_functions (FILE *inf) |
| 4269 | { |
| 4270 | char *package = savestr ("main"); /* current package name */ |
| 4271 | register char *cp; |
| 4272 | |
| 4273 | LOOP_ON_INPUT_LINES (inf, lb, cp) |
| 4274 | { |
| 4275 | cp = skip_spaces (cp); |
| 4276 | |
| 4277 | if (LOOKING_AT (cp, "package")) |
| 4278 | { |
| 4279 | free (package); |
| 4280 | get_tag (cp, &package); |
| 4281 | } |
| 4282 | else if (LOOKING_AT (cp, "sub")) |
| 4283 | { |
| 4284 | char *pos, *sp; |
| 4285 | |
| 4286 | subr: |
| 4287 | sp = cp; |
| 4288 | while (!notinname (*cp)) |
| 4289 | cp++; |
| 4290 | if (cp == sp) |
| 4291 | continue; /* nothing found */ |
| 4292 | if ((pos = etags_strchr (sp, ':')) != NULL |
| 4293 | && pos < cp && pos[1] == ':') |
| 4294 | /* The name is already qualified. */ |
| 4295 | make_tag (sp, cp - sp, TRUE, |
| 4296 | lb.buffer, cp - lb.buffer + 1, lineno, linecharno); |
| 4297 | else |
| 4298 | /* Qualify it. */ |
| 4299 | { |
| 4300 | char savechar, *name; |
| 4301 | |
| 4302 | savechar = *cp; |
| 4303 | *cp = '\0'; |
| 4304 | name = concat (package, "::", sp); |
| 4305 | *cp = savechar; |
| 4306 | make_tag (name, strlen (name), TRUE, |
| 4307 | lb.buffer, cp - lb.buffer + 1, lineno, linecharno); |
| 4308 | free (name); |
| 4309 | } |
| 4310 | } |
| 4311 | else if (LOOKING_AT (cp, "use constant") |
| 4312 | || LOOKING_AT (cp, "use constant::defer")) |
| 4313 | { |
| 4314 | /* For hash style multi-constant like |
| 4315 | use constant { FOO => 123, |
| 4316 | BAR => 456 }; |
| 4317 | only the first FOO is picked up. Parsing across the value |
| 4318 | expressions would be difficult in general, due to possible nested |
| 4319 | hashes, here-documents, etc. */ |
| 4320 | if (*cp == '{') |
| 4321 | cp = skip_spaces (cp+1); |
| 4322 | goto subr; |
| 4323 | } |
| 4324 | else if (globals) /* only if we are tagging global vars */ |
| 4325 | { |
| 4326 | /* Skip a qualifier, if any. */ |
| 4327 | bool qual = LOOKING_AT (cp, "my") || LOOKING_AT (cp, "local"); |
| 4328 | /* After "my" or "local", but before any following paren or space. */ |
| 4329 | char *varstart = cp; |
| 4330 | |
| 4331 | if (qual /* should this be removed? If yes, how? */ |
| 4332 | && (*cp == '$' || *cp == '@' || *cp == '%')) |
| 4333 | { |
| 4334 | varstart += 1; |
| 4335 | do |
| 4336 | cp++; |
| 4337 | while (ISALNUM (*cp) || *cp == '_'); |
| 4338 | } |
| 4339 | else if (qual) |
| 4340 | { |
| 4341 | /* Should be examining a variable list at this point; |
| 4342 | could insist on seeing an open parenthesis. */ |
| 4343 | while (*cp != '\0' && *cp != ';' && *cp != '=' && *cp != ')') |
| 4344 | cp++; |
| 4345 | } |
| 4346 | else |
| 4347 | continue; |
| 4348 | |
| 4349 | make_tag (varstart, cp - varstart, FALSE, |
| 4350 | lb.buffer, cp - lb.buffer + 1, lineno, linecharno); |
| 4351 | } |
| 4352 | } |
| 4353 | free (package); |
| 4354 | } |
| 4355 | |
| 4356 | |
| 4357 | /* |
| 4358 | * Python support |
| 4359 | * Look for /^[\t]*def[ \t\n]+[^ \t\n(:]+/ or /^class[ \t\n]+[^ \t\n(:]+/ |
| 4360 | * Idea by Eric S. Raymond <esr@thyrsus.com> (1997) |
| 4361 | * More ideas by seb bacon <seb@jamkit.com> (2002) |
| 4362 | */ |
| 4363 | static void |
| 4364 | Python_functions (FILE *inf) |
| 4365 | { |
| 4366 | register char *cp; |
| 4367 | |
| 4368 | LOOP_ON_INPUT_LINES (inf, lb, cp) |
| 4369 | { |
| 4370 | cp = skip_spaces (cp); |
| 4371 | if (LOOKING_AT (cp, "def") || LOOKING_AT (cp, "class")) |
| 4372 | { |
| 4373 | char *name = cp; |
| 4374 | while (!notinname (*cp) && *cp != ':') |
| 4375 | cp++; |
| 4376 | make_tag (name, cp - name, TRUE, |
| 4377 | lb.buffer, cp - lb.buffer + 1, lineno, linecharno); |
| 4378 | } |
| 4379 | } |
| 4380 | } |
| 4381 | |
| 4382 | \f |
| 4383 | /* |
| 4384 | * PHP support |
| 4385 | * Look for: |
| 4386 | * - /^[ \t]*function[ \t\n]+[^ \t\n(]+/ |
| 4387 | * - /^[ \t]*class[ \t\n]+[^ \t\n]+/ |
| 4388 | * - /^[ \t]*define\(\"[^\"]+/ |
| 4389 | * Only with --members: |
| 4390 | * - /^[ \t]*var[ \t\n]+\$[^ \t\n=;]/ |
| 4391 | * Idea by Diez B. Roggisch (2001) |
| 4392 | */ |
| 4393 | static void |
| 4394 | PHP_functions (FILE *inf) |
| 4395 | { |
| 4396 | register char *cp, *name; |
| 4397 | bool search_identifier = FALSE; |
| 4398 | |
| 4399 | LOOP_ON_INPUT_LINES (inf, lb, cp) |
| 4400 | { |
| 4401 | cp = skip_spaces (cp); |
| 4402 | name = cp; |
| 4403 | if (search_identifier |
| 4404 | && *cp != '\0') |
| 4405 | { |
| 4406 | while (!notinname (*cp)) |
| 4407 | cp++; |
| 4408 | make_tag (name, cp - name, TRUE, |
| 4409 | lb.buffer, cp - lb.buffer + 1, lineno, linecharno); |
| 4410 | search_identifier = FALSE; |
| 4411 | } |
| 4412 | else if (LOOKING_AT (cp, "function")) |
| 4413 | { |
| 4414 | if (*cp == '&') |
| 4415 | cp = skip_spaces (cp+1); |
| 4416 | if (*cp != '\0') |
| 4417 | { |
| 4418 | name = cp; |
| 4419 | while (!notinname (*cp)) |
| 4420 | cp++; |
| 4421 | make_tag (name, cp - name, TRUE, |
| 4422 | lb.buffer, cp - lb.buffer + 1, lineno, linecharno); |
| 4423 | } |
| 4424 | else |
| 4425 | search_identifier = TRUE; |
| 4426 | } |
| 4427 | else if (LOOKING_AT (cp, "class")) |
| 4428 | { |
| 4429 | if (*cp != '\0') |
| 4430 | { |
| 4431 | name = cp; |
| 4432 | while (*cp != '\0' && !iswhite (*cp)) |
| 4433 | cp++; |
| 4434 | make_tag (name, cp - name, FALSE, |
| 4435 | lb.buffer, cp - lb.buffer + 1, lineno, linecharno); |
| 4436 | } |
| 4437 | else |
| 4438 | search_identifier = TRUE; |
| 4439 | } |
| 4440 | else if (strneq (cp, "define", 6) |
| 4441 | && (cp = skip_spaces (cp+6)) |
| 4442 | && *cp++ == '(' |
| 4443 | && (*cp == '"' || *cp == '\'')) |
| 4444 | { |
| 4445 | char quote = *cp++; |
| 4446 | name = cp; |
| 4447 | while (*cp != quote && *cp != '\0') |
| 4448 | cp++; |
| 4449 | make_tag (name, cp - name, FALSE, |
| 4450 | lb.buffer, cp - lb.buffer + 1, lineno, linecharno); |
| 4451 | } |
| 4452 | else if (members |
| 4453 | && LOOKING_AT (cp, "var") |
| 4454 | && *cp == '$') |
| 4455 | { |
| 4456 | name = cp; |
| 4457 | while (!notinname (*cp)) |
| 4458 | cp++; |
| 4459 | make_tag (name, cp - name, FALSE, |
| 4460 | lb.buffer, cp - lb.buffer + 1, lineno, linecharno); |
| 4461 | } |
| 4462 | } |
| 4463 | } |
| 4464 | |
| 4465 | \f |
| 4466 | /* |
| 4467 | * Cobol tag functions |
| 4468 | * We could look for anything that could be a paragraph name. |
| 4469 | * i.e. anything that starts in column 8 is one word and ends in a full stop. |
| 4470 | * Idea by Corny de Souza (1993) |
| 4471 | */ |
| 4472 | static void |
| 4473 | Cobol_paragraphs (FILE *inf) |
| 4474 | { |
| 4475 | register char *bp, *ep; |
| 4476 | |
| 4477 | LOOP_ON_INPUT_LINES (inf, lb, bp) |
| 4478 | { |
| 4479 | if (lb.len < 9) |
| 4480 | continue; |
| 4481 | bp += 8; |
| 4482 | |
| 4483 | /* If eoln, compiler option or comment ignore whole line. */ |
| 4484 | if (bp[-1] != ' ' || !ISALNUM (bp[0])) |
| 4485 | continue; |
| 4486 | |
| 4487 | for (ep = bp; ISALNUM (*ep) || *ep == '-'; ep++) |
| 4488 | continue; |
| 4489 | if (*ep++ == '.') |
| 4490 | make_tag (bp, ep - bp, TRUE, |
| 4491 | lb.buffer, ep - lb.buffer + 1, lineno, linecharno); |
| 4492 | } |
| 4493 | } |
| 4494 | |
| 4495 | \f |
| 4496 | /* |
| 4497 | * Makefile support |
| 4498 | * Ideas by Assar Westerlund <assar@sics.se> (2001) |
| 4499 | */ |
| 4500 | static void |
| 4501 | Makefile_targets (FILE *inf) |
| 4502 | { |
| 4503 | register char *bp; |
| 4504 | |
| 4505 | LOOP_ON_INPUT_LINES (inf, lb, bp) |
| 4506 | { |
| 4507 | if (*bp == '\t' || *bp == '#') |
| 4508 | continue; |
| 4509 | while (*bp != '\0' && *bp != '=' && *bp != ':') |
| 4510 | bp++; |
| 4511 | if (*bp == ':' || (globals && *bp == '=')) |
| 4512 | { |
| 4513 | /* We should detect if there is more than one tag, but we do not. |
| 4514 | We just skip initial and final spaces. */ |
| 4515 | char * namestart = skip_spaces (lb.buffer); |
| 4516 | while (--bp > namestart) |
| 4517 | if (!notinname (*bp)) |
| 4518 | break; |
| 4519 | make_tag (namestart, bp - namestart + 1, TRUE, |
| 4520 | lb.buffer, bp - lb.buffer + 2, lineno, linecharno); |
| 4521 | } |
| 4522 | } |
| 4523 | } |
| 4524 | |
| 4525 | \f |
| 4526 | /* |
| 4527 | * Pascal parsing |
| 4528 | * Original code by Mosur K. Mohan (1989) |
| 4529 | * |
| 4530 | * Locates tags for procedures & functions. Doesn't do any type- or |
| 4531 | * var-definitions. It does look for the keyword "extern" or |
| 4532 | * "forward" immediately following the procedure statement; if found, |
| 4533 | * the tag is skipped. |
| 4534 | */ |
| 4535 | static void |
| 4536 | Pascal_functions (FILE *inf) |
| 4537 | { |
| 4538 | linebuffer tline; /* mostly copied from C_entries */ |
| 4539 | long save_lcno; |
| 4540 | int save_lineno, namelen, taglen; |
| 4541 | char c, *name; |
| 4542 | |
| 4543 | bool /* each of these flags is TRUE if: */ |
| 4544 | incomment, /* point is inside a comment */ |
| 4545 | inquote, /* point is inside '..' string */ |
| 4546 | get_tagname, /* point is after PROCEDURE/FUNCTION |
| 4547 | keyword, so next item = potential tag */ |
| 4548 | found_tag, /* point is after a potential tag */ |
| 4549 | inparms, /* point is within parameter-list */ |
| 4550 | verify_tag; /* point has passed the parm-list, so the |
| 4551 | next token will determine whether this |
| 4552 | is a FORWARD/EXTERN to be ignored, or |
| 4553 | whether it is a real tag */ |
| 4554 | |
| 4555 | save_lcno = save_lineno = namelen = taglen = 0; /* keep compiler quiet */ |
| 4556 | name = NULL; /* keep compiler quiet */ |
| 4557 | dbp = lb.buffer; |
| 4558 | *dbp = '\0'; |
| 4559 | linebuffer_init (&tline); |
| 4560 | |
| 4561 | incomment = inquote = FALSE; |
| 4562 | found_tag = FALSE; /* have a proc name; check if extern */ |
| 4563 | get_tagname = FALSE; /* found "procedure" keyword */ |
| 4564 | inparms = FALSE; /* found '(' after "proc" */ |
| 4565 | verify_tag = FALSE; /* check if "extern" is ahead */ |
| 4566 | |
| 4567 | |
| 4568 | while (!feof (inf)) /* long main loop to get next char */ |
| 4569 | { |
| 4570 | c = *dbp++; |
| 4571 | if (c == '\0') /* if end of line */ |
| 4572 | { |
| 4573 | readline (&lb, inf); |
| 4574 | dbp = lb.buffer; |
| 4575 | if (*dbp == '\0') |
| 4576 | continue; |
| 4577 | if (!((found_tag && verify_tag) |
| 4578 | || get_tagname)) |
| 4579 | c = *dbp++; /* only if don't need *dbp pointing |
| 4580 | to the beginning of the name of |
| 4581 | the procedure or function */ |
| 4582 | } |
| 4583 | if (incomment) |
| 4584 | { |
| 4585 | if (c == '}') /* within { } comments */ |
| 4586 | incomment = FALSE; |
| 4587 | else if (c == '*' && *dbp == ')') /* within (* *) comments */ |
| 4588 | { |
| 4589 | dbp++; |
| 4590 | incomment = FALSE; |
| 4591 | } |
| 4592 | continue; |
| 4593 | } |
| 4594 | else if (inquote) |
| 4595 | { |
| 4596 | if (c == '\'') |
| 4597 | inquote = FALSE; |
| 4598 | continue; |
| 4599 | } |
| 4600 | else |
| 4601 | switch (c) |
| 4602 | { |
| 4603 | case '\'': |
| 4604 | inquote = TRUE; /* found first quote */ |
| 4605 | continue; |
| 4606 | case '{': /* found open { comment */ |
| 4607 | incomment = TRUE; |
| 4608 | continue; |
| 4609 | case '(': |
| 4610 | if (*dbp == '*') /* found open (* comment */ |
| 4611 | { |
| 4612 | incomment = TRUE; |
| 4613 | dbp++; |
| 4614 | } |
| 4615 | else if (found_tag) /* found '(' after tag, i.e., parm-list */ |
| 4616 | inparms = TRUE; |
| 4617 | continue; |
| 4618 | case ')': /* end of parms list */ |
| 4619 | if (inparms) |
| 4620 | inparms = FALSE; |
| 4621 | continue; |
| 4622 | case ';': |
| 4623 | if (found_tag && !inparms) /* end of proc or fn stmt */ |
| 4624 | { |
| 4625 | verify_tag = TRUE; |
| 4626 | break; |
| 4627 | } |
| 4628 | continue; |
| 4629 | } |
| 4630 | if (found_tag && verify_tag && (*dbp != ' ')) |
| 4631 | { |
| 4632 | /* Check if this is an "extern" declaration. */ |
| 4633 | if (*dbp == '\0') |
| 4634 | continue; |
| 4635 | if (lowcase (*dbp) == 'e') |
| 4636 | { |
| 4637 | if (nocase_tail ("extern")) /* superfluous, really! */ |
| 4638 | { |
| 4639 | found_tag = FALSE; |
| 4640 | verify_tag = FALSE; |
| 4641 | } |
| 4642 | } |
| 4643 | else if (lowcase (*dbp) == 'f') |
| 4644 | { |
| 4645 | if (nocase_tail ("forward")) /* check for forward reference */ |
| 4646 | { |
| 4647 | found_tag = FALSE; |
| 4648 | verify_tag = FALSE; |
| 4649 | } |
| 4650 | } |
| 4651 | if (found_tag && verify_tag) /* not external proc, so make tag */ |
| 4652 | { |
| 4653 | found_tag = FALSE; |
| 4654 | verify_tag = FALSE; |
| 4655 | make_tag (name, namelen, TRUE, |
| 4656 | tline.buffer, taglen, save_lineno, save_lcno); |
| 4657 | continue; |
| 4658 | } |
| 4659 | } |
| 4660 | if (get_tagname) /* grab name of proc or fn */ |
| 4661 | { |
| 4662 | char *cp; |
| 4663 | |
| 4664 | if (*dbp == '\0') |
| 4665 | continue; |
| 4666 | |
| 4667 | /* Find block name. */ |
| 4668 | for (cp = dbp + 1; *cp != '\0' && !endtoken (*cp); cp++) |
| 4669 | continue; |
| 4670 | |
| 4671 | /* Save all values for later tagging. */ |
| 4672 | linebuffer_setlen (&tline, lb.len); |
| 4673 | strcpy (tline.buffer, lb.buffer); |
| 4674 | save_lineno = lineno; |
| 4675 | save_lcno = linecharno; |
| 4676 | name = tline.buffer + (dbp - lb.buffer); |
| 4677 | namelen = cp - dbp; |
| 4678 | taglen = cp - lb.buffer + 1; |
| 4679 | |
| 4680 | dbp = cp; /* set dbp to e-o-token */ |
| 4681 | get_tagname = FALSE; |
| 4682 | found_tag = TRUE; |
| 4683 | continue; |
| 4684 | |
| 4685 | /* And proceed to check for "extern". */ |
| 4686 | } |
| 4687 | else if (!incomment && !inquote && !found_tag) |
| 4688 | { |
| 4689 | /* Check for proc/fn keywords. */ |
| 4690 | switch (lowcase (c)) |
| 4691 | { |
| 4692 | case 'p': |
| 4693 | if (nocase_tail ("rocedure")) /* c = 'p', dbp has advanced */ |
| 4694 | get_tagname = TRUE; |
| 4695 | continue; |
| 4696 | case 'f': |
| 4697 | if (nocase_tail ("unction")) |
| 4698 | get_tagname = TRUE; |
| 4699 | continue; |
| 4700 | } |
| 4701 | } |
| 4702 | } /* while not eof */ |
| 4703 | |
| 4704 | free (tline.buffer); |
| 4705 | } |
| 4706 | |
| 4707 | \f |
| 4708 | /* |
| 4709 | * Lisp tag functions |
| 4710 | * look for (def or (DEF, quote or QUOTE |
| 4711 | */ |
| 4712 | |
| 4713 | static void L_getit (void); |
| 4714 | |
| 4715 | static void |
| 4716 | L_getit (void) |
| 4717 | { |
| 4718 | if (*dbp == '\'') /* Skip prefix quote */ |
| 4719 | dbp++; |
| 4720 | else if (*dbp == '(') |
| 4721 | { |
| 4722 | dbp++; |
| 4723 | /* Try to skip "(quote " */ |
| 4724 | if (!LOOKING_AT (dbp, "quote") && !LOOKING_AT (dbp, "QUOTE")) |
| 4725 | /* Ok, then skip "(" before name in (defstruct (foo)) */ |
| 4726 | dbp = skip_spaces (dbp); |
| 4727 | } |
| 4728 | get_tag (dbp, NULL); |
| 4729 | } |
| 4730 | |
| 4731 | static void |
| 4732 | Lisp_functions (FILE *inf) |
| 4733 | { |
| 4734 | LOOP_ON_INPUT_LINES (inf, lb, dbp) |
| 4735 | { |
| 4736 | if (dbp[0] != '(') |
| 4737 | continue; |
| 4738 | |
| 4739 | /* "(defvar foo)" is a declaration rather than a definition. */ |
| 4740 | if (! declarations) |
| 4741 | { |
| 4742 | char *p = dbp + 1; |
| 4743 | if (LOOKING_AT (p, "defvar")) |
| 4744 | { |
| 4745 | p = skip_name (p); /* past var name */ |
| 4746 | p = skip_spaces (p); |
| 4747 | if (*p == ')') |
| 4748 | continue; |
| 4749 | } |
| 4750 | } |
| 4751 | |
| 4752 | if (strneq (dbp+1, "def", 3) || strneq (dbp+1, "DEF", 3)) |
| 4753 | { |
| 4754 | dbp = skip_non_spaces (dbp); |
| 4755 | dbp = skip_spaces (dbp); |
| 4756 | L_getit (); |
| 4757 | } |
| 4758 | else |
| 4759 | { |
| 4760 | /* Check for (foo::defmumble name-defined ... */ |
| 4761 | do |
| 4762 | dbp++; |
| 4763 | while (!notinname (*dbp) && *dbp != ':'); |
| 4764 | if (*dbp == ':') |
| 4765 | { |
| 4766 | do |
| 4767 | dbp++; |
| 4768 | while (*dbp == ':'); |
| 4769 | |
| 4770 | if (strneq (dbp, "def", 3) || strneq (dbp, "DEF", 3)) |
| 4771 | { |
| 4772 | dbp = skip_non_spaces (dbp); |
| 4773 | dbp = skip_spaces (dbp); |
| 4774 | L_getit (); |
| 4775 | } |
| 4776 | } |
| 4777 | } |
| 4778 | } |
| 4779 | } |
| 4780 | |
| 4781 | \f |
| 4782 | /* |
| 4783 | * Lua script language parsing |
| 4784 | * Original code by David A. Capello <dacap@users.sourceforge.net> (2004) |
| 4785 | * |
| 4786 | * "function" and "local function" are tags if they start at column 1. |
| 4787 | */ |
| 4788 | static void |
| 4789 | Lua_functions (FILE *inf) |
| 4790 | { |
| 4791 | register char *bp; |
| 4792 | |
| 4793 | LOOP_ON_INPUT_LINES (inf, lb, bp) |
| 4794 | { |
| 4795 | if (bp[0] != 'f' && bp[0] != 'l') |
| 4796 | continue; |
| 4797 | |
| 4798 | (void)LOOKING_AT (bp, "local"); /* skip possible "local" */ |
| 4799 | |
| 4800 | if (LOOKING_AT (bp, "function")) |
| 4801 | get_tag (bp, NULL); |
| 4802 | } |
| 4803 | } |
| 4804 | |
| 4805 | \f |
| 4806 | /* |
| 4807 | * PostScript tags |
| 4808 | * Just look for lines where the first character is '/' |
| 4809 | * Also look at "defineps" for PSWrap |
| 4810 | * Ideas by: |
| 4811 | * Richard Mlynarik <mly@adoc.xerox.com> (1997) |
| 4812 | * Masatake Yamato <masata-y@is.aist-nara.ac.jp> (1999) |
| 4813 | */ |
| 4814 | static void |
| 4815 | PS_functions (FILE *inf) |
| 4816 | { |
| 4817 | register char *bp, *ep; |
| 4818 | |
| 4819 | LOOP_ON_INPUT_LINES (inf, lb, bp) |
| 4820 | { |
| 4821 | if (bp[0] == '/') |
| 4822 | { |
| 4823 | for (ep = bp+1; |
| 4824 | *ep != '\0' && *ep != ' ' && *ep != '{'; |
| 4825 | ep++) |
| 4826 | continue; |
| 4827 | make_tag (bp, ep - bp, TRUE, |
| 4828 | lb.buffer, ep - lb.buffer + 1, lineno, linecharno); |
| 4829 | } |
| 4830 | else if (LOOKING_AT (bp, "defineps")) |
| 4831 | get_tag (bp, NULL); |
| 4832 | } |
| 4833 | } |
| 4834 | |
| 4835 | \f |
| 4836 | /* |
| 4837 | * Forth tags |
| 4838 | * Ignore anything after \ followed by space or in ( ) |
| 4839 | * Look for words defined by : |
| 4840 | * Look for constant, code, create, defer, value, and variable |
| 4841 | * OBP extensions: Look for buffer:, field, |
| 4842 | * Ideas by Eduardo Horvath <eeh@netbsd.org> (2004) |
| 4843 | */ |
| 4844 | static void |
| 4845 | Forth_words (FILE *inf) |
| 4846 | { |
| 4847 | register char *bp; |
| 4848 | |
| 4849 | LOOP_ON_INPUT_LINES (inf, lb, bp) |
| 4850 | while ((bp = skip_spaces (bp))[0] != '\0') |
| 4851 | if (bp[0] == '\\' && iswhite (bp[1])) |
| 4852 | break; /* read next line */ |
| 4853 | else if (bp[0] == '(' && iswhite (bp[1])) |
| 4854 | do /* skip to ) or eol */ |
| 4855 | bp++; |
| 4856 | while (*bp != ')' && *bp != '\0'); |
| 4857 | else if ((bp[0] == ':' && iswhite (bp[1]) && bp++) |
| 4858 | || LOOKING_AT_NOCASE (bp, "constant") |
| 4859 | || LOOKING_AT_NOCASE (bp, "code") |
| 4860 | || LOOKING_AT_NOCASE (bp, "create") |
| 4861 | || LOOKING_AT_NOCASE (bp, "defer") |
| 4862 | || LOOKING_AT_NOCASE (bp, "value") |
| 4863 | || LOOKING_AT_NOCASE (bp, "variable") |
| 4864 | || LOOKING_AT_NOCASE (bp, "buffer:") |
| 4865 | || LOOKING_AT_NOCASE (bp, "field")) |
| 4866 | get_tag (skip_spaces (bp), NULL); /* Yay! A definition! */ |
| 4867 | else |
| 4868 | bp = skip_non_spaces (bp); |
| 4869 | } |
| 4870 | |
| 4871 | \f |
| 4872 | /* |
| 4873 | * Scheme tag functions |
| 4874 | * look for (def... xyzzy |
| 4875 | * (def... (xyzzy |
| 4876 | * (def ... ((...(xyzzy .... |
| 4877 | * (set! xyzzy |
| 4878 | * Original code by Ken Haase (1985?) |
| 4879 | */ |
| 4880 | static void |
| 4881 | Scheme_functions (FILE *inf) |
| 4882 | { |
| 4883 | register char *bp; |
| 4884 | |
| 4885 | LOOP_ON_INPUT_LINES (inf, lb, bp) |
| 4886 | { |
| 4887 | if (strneq (bp, "(def", 4) || strneq (bp, "(DEF", 4)) |
| 4888 | { |
| 4889 | bp = skip_non_spaces (bp+4); |
| 4890 | /* Skip over open parens and white space. Don't continue past |
| 4891 | '\0'. */ |
| 4892 | while (*bp && notinname (*bp)) |
| 4893 | bp++; |
| 4894 | get_tag (bp, NULL); |
| 4895 | } |
| 4896 | if (LOOKING_AT (bp, "(SET!") || LOOKING_AT (bp, "(set!")) |
| 4897 | get_tag (bp, NULL); |
| 4898 | } |
| 4899 | } |
| 4900 | |
| 4901 | \f |
| 4902 | /* Find tags in TeX and LaTeX input files. */ |
| 4903 | |
| 4904 | /* TEX_toktab is a table of TeX control sequences that define tags. |
| 4905 | * Each entry records one such control sequence. |
| 4906 | * |
| 4907 | * Original code from who knows whom. |
| 4908 | * Ideas by: |
| 4909 | * Stefan Monnier (2002) |
| 4910 | */ |
| 4911 | |
| 4912 | static linebuffer *TEX_toktab = NULL; /* Table with tag tokens */ |
| 4913 | |
| 4914 | /* Default set of control sequences to put into TEX_toktab. |
| 4915 | The value of environment var TEXTAGS is prepended to this. */ |
| 4916 | static const char *TEX_defenv = "\ |
| 4917 | :chapter:section:subsection:subsubsection:eqno:label:ref:cite:bibitem\ |
| 4918 | :part:appendix:entry:index:def\ |
| 4919 | :newcommand:renewcommand:newenvironment:renewenvironment"; |
| 4920 | |
| 4921 | static void TEX_mode (FILE *); |
| 4922 | static void TEX_decode_env (const char *, const char *); |
| 4923 | |
| 4924 | static char TEX_esc = '\\'; |
| 4925 | static char TEX_opgrp = '{'; |
| 4926 | static char TEX_clgrp = '}'; |
| 4927 | |
| 4928 | /* |
| 4929 | * TeX/LaTeX scanning loop. |
| 4930 | */ |
| 4931 | static void |
| 4932 | TeX_commands (FILE *inf) |
| 4933 | { |
| 4934 | char *cp; |
| 4935 | linebuffer *key; |
| 4936 | |
| 4937 | /* Select either \ or ! as escape character. */ |
| 4938 | TEX_mode (inf); |
| 4939 | |
| 4940 | /* Initialize token table once from environment. */ |
| 4941 | if (TEX_toktab == NULL) |
| 4942 | TEX_decode_env ("TEXTAGS", TEX_defenv); |
| 4943 | |
| 4944 | LOOP_ON_INPUT_LINES (inf, lb, cp) |
| 4945 | { |
| 4946 | /* Look at each TEX keyword in line. */ |
| 4947 | for (;;) |
| 4948 | { |
| 4949 | /* Look for a TEX escape. */ |
| 4950 | while (*cp++ != TEX_esc) |
| 4951 | if (cp[-1] == '\0' || cp[-1] == '%') |
| 4952 | goto tex_next_line; |
| 4953 | |
| 4954 | for (key = TEX_toktab; key->buffer != NULL; key++) |
| 4955 | if (strneq (cp, key->buffer, key->len)) |
| 4956 | { |
| 4957 | register char *p; |
| 4958 | int namelen, linelen; |
| 4959 | bool opgrp = FALSE; |
| 4960 | |
| 4961 | cp = skip_spaces (cp + key->len); |
| 4962 | if (*cp == TEX_opgrp) |
| 4963 | { |
| 4964 | opgrp = TRUE; |
| 4965 | cp++; |
| 4966 | } |
| 4967 | for (p = cp; |
| 4968 | (!iswhite (*p) && *p != '#' && |
| 4969 | *p != TEX_opgrp && *p != TEX_clgrp); |
| 4970 | p++) |
| 4971 | continue; |
| 4972 | namelen = p - cp; |
| 4973 | linelen = lb.len; |
| 4974 | if (!opgrp || *p == TEX_clgrp) |
| 4975 | { |
| 4976 | while (*p != '\0' && *p != TEX_opgrp && *p != TEX_clgrp) |
| 4977 | p++; |
| 4978 | linelen = p - lb.buffer + 1; |
| 4979 | } |
| 4980 | make_tag (cp, namelen, TRUE, |
| 4981 | lb.buffer, linelen, lineno, linecharno); |
| 4982 | goto tex_next_line; /* We only tag a line once */ |
| 4983 | } |
| 4984 | } |
| 4985 | tex_next_line: |
| 4986 | ; |
| 4987 | } |
| 4988 | } |
| 4989 | |
| 4990 | #define TEX_LESC '\\' |
| 4991 | #define TEX_SESC '!' |
| 4992 | |
| 4993 | /* Figure out whether TeX's escapechar is '\\' or '!' and set grouping |
| 4994 | chars accordingly. */ |
| 4995 | static void |
| 4996 | TEX_mode (FILE *inf) |
| 4997 | { |
| 4998 | int c; |
| 4999 | |
| 5000 | while ((c = getc (inf)) != EOF) |
| 5001 | { |
| 5002 | /* Skip to next line if we hit the TeX comment char. */ |
| 5003 | if (c == '%') |
| 5004 | while (c != '\n' && c != EOF) |
| 5005 | c = getc (inf); |
| 5006 | else if (c == TEX_LESC || c == TEX_SESC ) |
| 5007 | break; |
| 5008 | } |
| 5009 | |
| 5010 | if (c == TEX_LESC) |
| 5011 | { |
| 5012 | TEX_esc = TEX_LESC; |
| 5013 | TEX_opgrp = '{'; |
| 5014 | TEX_clgrp = '}'; |
| 5015 | } |
| 5016 | else |
| 5017 | { |
| 5018 | TEX_esc = TEX_SESC; |
| 5019 | TEX_opgrp = '<'; |
| 5020 | TEX_clgrp = '>'; |
| 5021 | } |
| 5022 | /* If the input file is compressed, inf is a pipe, and rewind may fail. |
| 5023 | No attempt is made to correct the situation. */ |
| 5024 | rewind (inf); |
| 5025 | } |
| 5026 | |
| 5027 | /* Read environment and prepend it to the default string. |
| 5028 | Build token table. */ |
| 5029 | static void |
| 5030 | TEX_decode_env (const char *evarname, const char *defenv) |
| 5031 | { |
| 5032 | register const char *env, *p; |
| 5033 | int i, len; |
| 5034 | |
| 5035 | /* Append default string to environment. */ |
| 5036 | env = getenv (evarname); |
| 5037 | if (!env) |
| 5038 | env = defenv; |
| 5039 | else |
| 5040 | env = concat (env, defenv, ""); |
| 5041 | |
| 5042 | /* Allocate a token table */ |
| 5043 | for (len = 1, p = env; p;) |
| 5044 | if ((p = etags_strchr (p, ':')) && *++p != '\0') |
| 5045 | len++; |
| 5046 | TEX_toktab = xnew (len, linebuffer); |
| 5047 | |
| 5048 | /* Unpack environment string into token table. Be careful about */ |
| 5049 | /* zero-length strings (leading ':', "::" and trailing ':') */ |
| 5050 | for (i = 0; *env != '\0';) |
| 5051 | { |
| 5052 | p = etags_strchr (env, ':'); |
| 5053 | if (!p) /* End of environment string. */ |
| 5054 | p = env + strlen (env); |
| 5055 | if (p - env > 0) |
| 5056 | { /* Only non-zero strings. */ |
| 5057 | TEX_toktab[i].buffer = savenstr (env, p - env); |
| 5058 | TEX_toktab[i].len = p - env; |
| 5059 | i++; |
| 5060 | } |
| 5061 | if (*p) |
| 5062 | env = p + 1; |
| 5063 | else |
| 5064 | { |
| 5065 | TEX_toktab[i].buffer = NULL; /* Mark end of table. */ |
| 5066 | TEX_toktab[i].len = 0; |
| 5067 | break; |
| 5068 | } |
| 5069 | } |
| 5070 | } |
| 5071 | |
| 5072 | \f |
| 5073 | /* Texinfo support. Dave Love, Mar. 2000. */ |
| 5074 | static void |
| 5075 | Texinfo_nodes (FILE *inf) |
| 5076 | { |
| 5077 | char *cp, *start; |
| 5078 | LOOP_ON_INPUT_LINES (inf, lb, cp) |
| 5079 | if (LOOKING_AT (cp, "@node")) |
| 5080 | { |
| 5081 | start = cp; |
| 5082 | while (*cp != '\0' && *cp != ',') |
| 5083 | cp++; |
| 5084 | make_tag (start, cp - start, TRUE, |
| 5085 | lb.buffer, cp - lb.buffer + 1, lineno, linecharno); |
| 5086 | } |
| 5087 | } |
| 5088 | |
| 5089 | \f |
| 5090 | /* |
| 5091 | * HTML support. |
| 5092 | * Contents of <title>, <h1>, <h2>, <h3> are tags. |
| 5093 | * Contents of <a name=xxx> are tags with name xxx. |
| 5094 | * |
| 5095 | * Francesco Potortì, 2002. |
| 5096 | */ |
| 5097 | static void |
| 5098 | HTML_labels (FILE *inf) |
| 5099 | { |
| 5100 | bool getnext = FALSE; /* next text outside of HTML tags is a tag */ |
| 5101 | bool skiptag = FALSE; /* skip to the end of the current HTML tag */ |
| 5102 | bool intag = FALSE; /* inside an html tag, looking for ID= */ |
| 5103 | bool inanchor = FALSE; /* when INTAG, is an anchor, look for NAME= */ |
| 5104 | char *end; |
| 5105 | |
| 5106 | |
| 5107 | linebuffer_setlen (&token_name, 0); /* no name in buffer */ |
| 5108 | |
| 5109 | LOOP_ON_INPUT_LINES (inf, lb, dbp) |
| 5110 | for (;;) /* loop on the same line */ |
| 5111 | { |
| 5112 | if (skiptag) /* skip HTML tag */ |
| 5113 | { |
| 5114 | while (*dbp != '\0' && *dbp != '>') |
| 5115 | dbp++; |
| 5116 | if (*dbp == '>') |
| 5117 | { |
| 5118 | dbp += 1; |
| 5119 | skiptag = FALSE; |
| 5120 | continue; /* look on the same line */ |
| 5121 | } |
| 5122 | break; /* go to next line */ |
| 5123 | } |
| 5124 | |
| 5125 | else if (intag) /* look for "name=" or "id=" */ |
| 5126 | { |
| 5127 | while (*dbp != '\0' && *dbp != '>' |
| 5128 | && lowcase (*dbp) != 'n' && lowcase (*dbp) != 'i') |
| 5129 | dbp++; |
| 5130 | if (*dbp == '\0') |
| 5131 | break; /* go to next line */ |
| 5132 | if (*dbp == '>') |
| 5133 | { |
| 5134 | dbp += 1; |
| 5135 | intag = FALSE; |
| 5136 | continue; /* look on the same line */ |
| 5137 | } |
| 5138 | if ((inanchor && LOOKING_AT_NOCASE (dbp, "name=")) |
| 5139 | || LOOKING_AT_NOCASE (dbp, "id=")) |
| 5140 | { |
| 5141 | bool quoted = (dbp[0] == '"'); |
| 5142 | |
| 5143 | if (quoted) |
| 5144 | for (end = ++dbp; *end != '\0' && *end != '"'; end++) |
| 5145 | continue; |
| 5146 | else |
| 5147 | for (end = dbp; *end != '\0' && intoken (*end); end++) |
| 5148 | continue; |
| 5149 | linebuffer_setlen (&token_name, end - dbp); |
| 5150 | memcpy (token_name.buffer, dbp, end - dbp); |
| 5151 | token_name.buffer[end - dbp] = '\0'; |
| 5152 | |
| 5153 | dbp = end; |
| 5154 | intag = FALSE; /* we found what we looked for */ |
| 5155 | skiptag = TRUE; /* skip to the end of the tag */ |
| 5156 | getnext = TRUE; /* then grab the text */ |
| 5157 | continue; /* look on the same line */ |
| 5158 | } |
| 5159 | dbp += 1; |
| 5160 | } |
| 5161 | |
| 5162 | else if (getnext) /* grab next tokens and tag them */ |
| 5163 | { |
| 5164 | dbp = skip_spaces (dbp); |
| 5165 | if (*dbp == '\0') |
| 5166 | break; /* go to next line */ |
| 5167 | if (*dbp == '<') |
| 5168 | { |
| 5169 | intag = TRUE; |
| 5170 | inanchor = (lowcase (dbp[1]) == 'a' && !intoken (dbp[2])); |
| 5171 | continue; /* look on the same line */ |
| 5172 | } |
| 5173 | |
| 5174 | for (end = dbp + 1; *end != '\0' && *end != '<'; end++) |
| 5175 | continue; |
| 5176 | make_tag (token_name.buffer, token_name.len, TRUE, |
| 5177 | dbp, end - dbp, lineno, linecharno); |
| 5178 | linebuffer_setlen (&token_name, 0); /* no name in buffer */ |
| 5179 | getnext = FALSE; |
| 5180 | break; /* go to next line */ |
| 5181 | } |
| 5182 | |
| 5183 | else /* look for an interesting HTML tag */ |
| 5184 | { |
| 5185 | while (*dbp != '\0' && *dbp != '<') |
| 5186 | dbp++; |
| 5187 | if (*dbp == '\0') |
| 5188 | break; /* go to next line */ |
| 5189 | intag = TRUE; |
| 5190 | if (lowcase (dbp[1]) == 'a' && !intoken (dbp[2])) |
| 5191 | { |
| 5192 | inanchor = TRUE; |
| 5193 | continue; /* look on the same line */ |
| 5194 | } |
| 5195 | else if (LOOKING_AT_NOCASE (dbp, "<title>") |
| 5196 | || LOOKING_AT_NOCASE (dbp, "<h1>") |
| 5197 | || LOOKING_AT_NOCASE (dbp, "<h2>") |
| 5198 | || LOOKING_AT_NOCASE (dbp, "<h3>")) |
| 5199 | { |
| 5200 | intag = FALSE; |
| 5201 | getnext = TRUE; |
| 5202 | continue; /* look on the same line */ |
| 5203 | } |
| 5204 | dbp += 1; |
| 5205 | } |
| 5206 | } |
| 5207 | } |
| 5208 | |
| 5209 | \f |
| 5210 | /* |
| 5211 | * Prolog support |
| 5212 | * |
| 5213 | * Assumes that the predicate or rule starts at column 0. |
| 5214 | * Only the first clause of a predicate or rule is added. |
| 5215 | * Original code by Sunichirou Sugou (1989) |
| 5216 | * Rewritten by Anders Lindgren (1996) |
| 5217 | */ |
| 5218 | static size_t prolog_pr (char *, char *); |
| 5219 | static void prolog_skip_comment (linebuffer *, FILE *); |
| 5220 | static size_t prolog_atom (char *, size_t); |
| 5221 | |
| 5222 | static void |
| 5223 | Prolog_functions (FILE *inf) |
| 5224 | { |
| 5225 | char *cp, *last; |
| 5226 | size_t len; |
| 5227 | size_t allocated; |
| 5228 | |
| 5229 | allocated = 0; |
| 5230 | len = 0; |
| 5231 | last = NULL; |
| 5232 | |
| 5233 | LOOP_ON_INPUT_LINES (inf, lb, cp) |
| 5234 | { |
| 5235 | if (cp[0] == '\0') /* Empty line */ |
| 5236 | continue; |
| 5237 | else if (iswhite (cp[0])) /* Not a predicate */ |
| 5238 | continue; |
| 5239 | else if (cp[0] == '/' && cp[1] == '*') /* comment. */ |
| 5240 | prolog_skip_comment (&lb, inf); |
| 5241 | else if ((len = prolog_pr (cp, last)) > 0) |
| 5242 | { |
| 5243 | /* Predicate or rule. Store the function name so that we |
| 5244 | only generate a tag for the first clause. */ |
| 5245 | if (last == NULL) |
| 5246 | last = xnew (len + 1, char); |
| 5247 | else if (len + 1 > allocated) |
| 5248 | xrnew (last, len + 1, char); |
| 5249 | allocated = len + 1; |
| 5250 | memcpy (last, cp, len); |
| 5251 | last[len] = '\0'; |
| 5252 | } |
| 5253 | } |
| 5254 | free (last); |
| 5255 | } |
| 5256 | |
| 5257 | |
| 5258 | static void |
| 5259 | prolog_skip_comment (linebuffer *plb, FILE *inf) |
| 5260 | { |
| 5261 | char *cp; |
| 5262 | |
| 5263 | do |
| 5264 | { |
| 5265 | for (cp = plb->buffer; *cp != '\0'; cp++) |
| 5266 | if (cp[0] == '*' && cp[1] == '/') |
| 5267 | return; |
| 5268 | readline (plb, inf); |
| 5269 | } |
| 5270 | while (!feof (inf)); |
| 5271 | } |
| 5272 | |
| 5273 | /* |
| 5274 | * A predicate or rule definition is added if it matches: |
| 5275 | * <beginning of line><Prolog Atom><whitespace>( |
| 5276 | * or <beginning of line><Prolog Atom><whitespace>:- |
| 5277 | * |
| 5278 | * It is added to the tags database if it doesn't match the |
| 5279 | * name of the previous clause header. |
| 5280 | * |
| 5281 | * Return the size of the name of the predicate or rule, or 0 if no |
| 5282 | * header was found. |
| 5283 | */ |
| 5284 | static size_t |
| 5285 | prolog_pr (char *s, char *last) |
| 5286 | |
| 5287 | /* Name of last clause. */ |
| 5288 | { |
| 5289 | size_t pos; |
| 5290 | size_t len; |
| 5291 | |
| 5292 | pos = prolog_atom (s, 0); |
| 5293 | if (! pos) |
| 5294 | return 0; |
| 5295 | |
| 5296 | len = pos; |
| 5297 | pos = skip_spaces (s + pos) - s; |
| 5298 | |
| 5299 | if ((s[pos] == '.' |
| 5300 | || (s[pos] == '(' && (pos += 1)) |
| 5301 | || (s[pos] == ':' && s[pos + 1] == '-' && (pos += 2))) |
| 5302 | && (last == NULL /* save only the first clause */ |
| 5303 | || len != strlen (last) |
| 5304 | || !strneq (s, last, len))) |
| 5305 | { |
| 5306 | make_tag (s, len, TRUE, s, pos, lineno, linecharno); |
| 5307 | return len; |
| 5308 | } |
| 5309 | else |
| 5310 | return 0; |
| 5311 | } |
| 5312 | |
| 5313 | /* |
| 5314 | * Consume a Prolog atom. |
| 5315 | * Return the number of bytes consumed, or 0 if there was an error. |
| 5316 | * |
| 5317 | * A prolog atom, in this context, could be one of: |
| 5318 | * - An alphanumeric sequence, starting with a lower case letter. |
| 5319 | * - A quoted arbitrary string. Single quotes can escape themselves. |
| 5320 | * Backslash quotes everything. |
| 5321 | */ |
| 5322 | static size_t |
| 5323 | prolog_atom (char *s, size_t pos) |
| 5324 | { |
| 5325 | size_t origpos; |
| 5326 | |
| 5327 | origpos = pos; |
| 5328 | |
| 5329 | if (ISLOWER (s[pos]) || (s[pos] == '_')) |
| 5330 | { |
| 5331 | /* The atom is unquoted. */ |
| 5332 | pos++; |
| 5333 | while (ISALNUM (s[pos]) || (s[pos] == '_')) |
| 5334 | { |
| 5335 | pos++; |
| 5336 | } |
| 5337 | return pos - origpos; |
| 5338 | } |
| 5339 | else if (s[pos] == '\'') |
| 5340 | { |
| 5341 | pos++; |
| 5342 | |
| 5343 | for (;;) |
| 5344 | { |
| 5345 | if (s[pos] == '\'') |
| 5346 | { |
| 5347 | pos++; |
| 5348 | if (s[pos] != '\'') |
| 5349 | break; |
| 5350 | pos++; /* A double quote */ |
| 5351 | } |
| 5352 | else if (s[pos] == '\0') |
| 5353 | /* Multiline quoted atoms are ignored. */ |
| 5354 | return 0; |
| 5355 | else if (s[pos] == '\\') |
| 5356 | { |
| 5357 | if (s[pos+1] == '\0') |
| 5358 | return 0; |
| 5359 | pos += 2; |
| 5360 | } |
| 5361 | else |
| 5362 | pos++; |
| 5363 | } |
| 5364 | return pos - origpos; |
| 5365 | } |
| 5366 | else |
| 5367 | return 0; |
| 5368 | } |
| 5369 | |
| 5370 | \f |
| 5371 | /* |
| 5372 | * Support for Erlang |
| 5373 | * |
| 5374 | * Generates tags for functions, defines, and records. |
| 5375 | * Assumes that Erlang functions start at column 0. |
| 5376 | * Original code by Anders Lindgren (1996) |
| 5377 | */ |
| 5378 | static int erlang_func (char *, char *); |
| 5379 | static void erlang_attribute (char *); |
| 5380 | static int erlang_atom (char *); |
| 5381 | |
| 5382 | static void |
| 5383 | Erlang_functions (FILE *inf) |
| 5384 | { |
| 5385 | char *cp, *last; |
| 5386 | int len; |
| 5387 | int allocated; |
| 5388 | |
| 5389 | allocated = 0; |
| 5390 | len = 0; |
| 5391 | last = NULL; |
| 5392 | |
| 5393 | LOOP_ON_INPUT_LINES (inf, lb, cp) |
| 5394 | { |
| 5395 | if (cp[0] == '\0') /* Empty line */ |
| 5396 | continue; |
| 5397 | else if (iswhite (cp[0])) /* Not function nor attribute */ |
| 5398 | continue; |
| 5399 | else if (cp[0] == '%') /* comment */ |
| 5400 | continue; |
| 5401 | else if (cp[0] == '"') /* Sometimes, strings start in column one */ |
| 5402 | continue; |
| 5403 | else if (cp[0] == '-') /* attribute, e.g. "-define" */ |
| 5404 | { |
| 5405 | erlang_attribute (cp); |
| 5406 | if (last != NULL) |
| 5407 | { |
| 5408 | free (last); |
| 5409 | last = NULL; |
| 5410 | } |
| 5411 | } |
| 5412 | else if ((len = erlang_func (cp, last)) > 0) |
| 5413 | { |
| 5414 | /* |
| 5415 | * Function. Store the function name so that we only |
| 5416 | * generates a tag for the first clause. |
| 5417 | */ |
| 5418 | if (last == NULL) |
| 5419 | last = xnew (len + 1, char); |
| 5420 | else if (len + 1 > allocated) |
| 5421 | xrnew (last, len + 1, char); |
| 5422 | allocated = len + 1; |
| 5423 | memcpy (last, cp, len); |
| 5424 | last[len] = '\0'; |
| 5425 | } |
| 5426 | } |
| 5427 | free (last); |
| 5428 | } |
| 5429 | |
| 5430 | |
| 5431 | /* |
| 5432 | * A function definition is added if it matches: |
| 5433 | * <beginning of line><Erlang Atom><whitespace>( |
| 5434 | * |
| 5435 | * It is added to the tags database if it doesn't match the |
| 5436 | * name of the previous clause header. |
| 5437 | * |
| 5438 | * Return the size of the name of the function, or 0 if no function |
| 5439 | * was found. |
| 5440 | */ |
| 5441 | static int |
| 5442 | erlang_func (char *s, char *last) |
| 5443 | |
| 5444 | /* Name of last clause. */ |
| 5445 | { |
| 5446 | int pos; |
| 5447 | int len; |
| 5448 | |
| 5449 | pos = erlang_atom (s); |
| 5450 | if (pos < 1) |
| 5451 | return 0; |
| 5452 | |
| 5453 | len = pos; |
| 5454 | pos = skip_spaces (s + pos) - s; |
| 5455 | |
| 5456 | /* Save only the first clause. */ |
| 5457 | if (s[pos++] == '(' |
| 5458 | && (last == NULL |
| 5459 | || len != (int)strlen (last) |
| 5460 | || !strneq (s, last, len))) |
| 5461 | { |
| 5462 | make_tag (s, len, TRUE, s, pos, lineno, linecharno); |
| 5463 | return len; |
| 5464 | } |
| 5465 | |
| 5466 | return 0; |
| 5467 | } |
| 5468 | |
| 5469 | |
| 5470 | /* |
| 5471 | * Handle attributes. Currently, tags are generated for defines |
| 5472 | * and records. |
| 5473 | * |
| 5474 | * They are on the form: |
| 5475 | * -define(foo, bar). |
| 5476 | * -define(Foo(M, N), M+N). |
| 5477 | * -record(graph, {vtab = notable, cyclic = true}). |
| 5478 | */ |
| 5479 | static void |
| 5480 | erlang_attribute (char *s) |
| 5481 | { |
| 5482 | char *cp = s; |
| 5483 | |
| 5484 | if ((LOOKING_AT (cp, "-define") || LOOKING_AT (cp, "-record")) |
| 5485 | && *cp++ == '(') |
| 5486 | { |
| 5487 | int len = erlang_atom (skip_spaces (cp)); |
| 5488 | if (len > 0) |
| 5489 | make_tag (cp, len, TRUE, s, cp + len - s, lineno, linecharno); |
| 5490 | } |
| 5491 | return; |
| 5492 | } |
| 5493 | |
| 5494 | |
| 5495 | /* |
| 5496 | * Consume an Erlang atom (or variable). |
| 5497 | * Return the number of bytes consumed, or -1 if there was an error. |
| 5498 | */ |
| 5499 | static int |
| 5500 | erlang_atom (char *s) |
| 5501 | { |
| 5502 | int pos = 0; |
| 5503 | |
| 5504 | if (ISALPHA (s[pos]) || s[pos] == '_') |
| 5505 | { |
| 5506 | /* The atom is unquoted. */ |
| 5507 | do |
| 5508 | pos++; |
| 5509 | while (ISALNUM (s[pos]) || s[pos] == '_'); |
| 5510 | } |
| 5511 | else if (s[pos] == '\'') |
| 5512 | { |
| 5513 | for (pos++; s[pos] != '\''; pos++) |
| 5514 | if (s[pos] == '\0' /* multiline quoted atoms are ignored */ |
| 5515 | || (s[pos] == '\\' && s[++pos] == '\0')) |
| 5516 | return 0; |
| 5517 | pos++; |
| 5518 | } |
| 5519 | |
| 5520 | return pos; |
| 5521 | } |
| 5522 | |
| 5523 | \f |
| 5524 | static char *scan_separators (char *); |
| 5525 | static void add_regex (char *, language *); |
| 5526 | static char *substitute (char *, char *, struct re_registers *); |
| 5527 | |
| 5528 | /* |
| 5529 | * Take a string like "/blah/" and turn it into "blah", verifying |
| 5530 | * that the first and last characters are the same, and handling |
| 5531 | * quoted separator characters. Actually, stops on the occurrence of |
| 5532 | * an unquoted separator. Also process \t, \n, etc. and turn into |
| 5533 | * appropriate characters. Works in place. Null terminates name string. |
| 5534 | * Returns pointer to terminating separator, or NULL for |
| 5535 | * unterminated regexps. |
| 5536 | */ |
| 5537 | static char * |
| 5538 | scan_separators (char *name) |
| 5539 | { |
| 5540 | char sep = name[0]; |
| 5541 | char *copyto = name; |
| 5542 | bool quoted = FALSE; |
| 5543 | |
| 5544 | for (++name; *name != '\0'; ++name) |
| 5545 | { |
| 5546 | if (quoted) |
| 5547 | { |
| 5548 | switch (*name) |
| 5549 | { |
| 5550 | case 'a': *copyto++ = '\007'; break; /* BEL (bell) */ |
| 5551 | case 'b': *copyto++ = '\b'; break; /* BS (back space) */ |
| 5552 | case 'd': *copyto++ = 0177; break; /* DEL (delete) */ |
| 5553 | case 'e': *copyto++ = 033; break; /* ESC (delete) */ |
| 5554 | case 'f': *copyto++ = '\f'; break; /* FF (form feed) */ |
| 5555 | case 'n': *copyto++ = '\n'; break; /* NL (new line) */ |
| 5556 | case 'r': *copyto++ = '\r'; break; /* CR (carriage return) */ |
| 5557 | case 't': *copyto++ = '\t'; break; /* TAB (horizontal tab) */ |
| 5558 | case 'v': *copyto++ = '\v'; break; /* VT (vertical tab) */ |
| 5559 | default: |
| 5560 | if (*name == sep) |
| 5561 | *copyto++ = sep; |
| 5562 | else |
| 5563 | { |
| 5564 | /* Something else is quoted, so preserve the quote. */ |
| 5565 | *copyto++ = '\\'; |
| 5566 | *copyto++ = *name; |
| 5567 | } |
| 5568 | break; |
| 5569 | } |
| 5570 | quoted = FALSE; |
| 5571 | } |
| 5572 | else if (*name == '\\') |
| 5573 | quoted = TRUE; |
| 5574 | else if (*name == sep) |
| 5575 | break; |
| 5576 | else |
| 5577 | *copyto++ = *name; |
| 5578 | } |
| 5579 | if (*name != sep) |
| 5580 | name = NULL; /* signal unterminated regexp */ |
| 5581 | |
| 5582 | /* Terminate copied string. */ |
| 5583 | *copyto = '\0'; |
| 5584 | return name; |
| 5585 | } |
| 5586 | |
| 5587 | /* Look at the argument of --regex or --no-regex and do the right |
| 5588 | thing. Same for each line of a regexp file. */ |
| 5589 | static void |
| 5590 | analyse_regex (char *regex_arg) |
| 5591 | { |
| 5592 | if (regex_arg == NULL) |
| 5593 | { |
| 5594 | free_regexps (); /* --no-regex: remove existing regexps */ |
| 5595 | return; |
| 5596 | } |
| 5597 | |
| 5598 | /* A real --regexp option or a line in a regexp file. */ |
| 5599 | switch (regex_arg[0]) |
| 5600 | { |
| 5601 | /* Comments in regexp file or null arg to --regex. */ |
| 5602 | case '\0': |
| 5603 | case ' ': |
| 5604 | case '\t': |
| 5605 | break; |
| 5606 | |
| 5607 | /* Read a regex file. This is recursive and may result in a |
| 5608 | loop, which will stop when the file descriptors are exhausted. */ |
| 5609 | case '@': |
| 5610 | { |
| 5611 | FILE *regexfp; |
| 5612 | linebuffer regexbuf; |
| 5613 | char *regexfile = regex_arg + 1; |
| 5614 | |
| 5615 | /* regexfile is a file containing regexps, one per line. */ |
| 5616 | regexfp = fopen (regexfile, "r"); |
| 5617 | if (regexfp == NULL) |
| 5618 | pfatal (regexfile); |
| 5619 | linebuffer_init (®exbuf); |
| 5620 | while (readline_internal (®exbuf, regexfp) > 0) |
| 5621 | analyse_regex (regexbuf.buffer); |
| 5622 | free (regexbuf.buffer); |
| 5623 | fclose (regexfp); |
| 5624 | } |
| 5625 | break; |
| 5626 | |
| 5627 | /* Regexp to be used for a specific language only. */ |
| 5628 | case '{': |
| 5629 | { |
| 5630 | language *lang; |
| 5631 | char *lang_name = regex_arg + 1; |
| 5632 | char *cp; |
| 5633 | |
| 5634 | for (cp = lang_name; *cp != '}'; cp++) |
| 5635 | if (*cp == '\0') |
| 5636 | { |
| 5637 | error ("unterminated language name in regex: %s", regex_arg); |
| 5638 | return; |
| 5639 | } |
| 5640 | *cp++ = '\0'; |
| 5641 | lang = get_language_from_langname (lang_name); |
| 5642 | if (lang == NULL) |
| 5643 | return; |
| 5644 | add_regex (cp, lang); |
| 5645 | } |
| 5646 | break; |
| 5647 | |
| 5648 | /* Regexp to be used for any language. */ |
| 5649 | default: |
| 5650 | add_regex (regex_arg, NULL); |
| 5651 | break; |
| 5652 | } |
| 5653 | } |
| 5654 | |
| 5655 | /* Separate the regexp pattern, compile it, |
| 5656 | and care for optional name and modifiers. */ |
| 5657 | static void |
| 5658 | add_regex (char *regexp_pattern, language *lang) |
| 5659 | { |
| 5660 | static struct re_pattern_buffer zeropattern; |
| 5661 | char sep, *pat, *name, *modifiers; |
| 5662 | char empty = '\0'; |
| 5663 | const char *err; |
| 5664 | struct re_pattern_buffer *patbuf; |
| 5665 | regexp *rp; |
| 5666 | bool |
| 5667 | force_explicit_name = TRUE, /* do not use implicit tag names */ |
| 5668 | ignore_case = FALSE, /* case is significant */ |
| 5669 | multi_line = FALSE, /* matches are done one line at a time */ |
| 5670 | single_line = FALSE; /* dot does not match newline */ |
| 5671 | |
| 5672 | |
| 5673 | if (strlen (regexp_pattern) < 3) |
| 5674 | { |
| 5675 | error ("null regexp"); |
| 5676 | return; |
| 5677 | } |
| 5678 | sep = regexp_pattern[0]; |
| 5679 | name = scan_separators (regexp_pattern); |
| 5680 | if (name == NULL) |
| 5681 | { |
| 5682 | error ("%s: unterminated regexp", regexp_pattern); |
| 5683 | return; |
| 5684 | } |
| 5685 | if (name[1] == sep) |
| 5686 | { |
| 5687 | error ("null name for regexp \"%s\"", regexp_pattern); |
| 5688 | return; |
| 5689 | } |
| 5690 | modifiers = scan_separators (name); |
| 5691 | if (modifiers == NULL) /* no terminating separator --> no name */ |
| 5692 | { |
| 5693 | modifiers = name; |
| 5694 | name = ∅ |
| 5695 | } |
| 5696 | else |
| 5697 | modifiers += 1; /* skip separator */ |
| 5698 | |
| 5699 | /* Parse regex modifiers. */ |
| 5700 | for (; modifiers[0] != '\0'; modifiers++) |
| 5701 | switch (modifiers[0]) |
| 5702 | { |
| 5703 | case 'N': |
| 5704 | if (modifiers == name) |
| 5705 | error ("forcing explicit tag name but no name, ignoring"); |
| 5706 | force_explicit_name = TRUE; |
| 5707 | break; |
| 5708 | case 'i': |
| 5709 | ignore_case = TRUE; |
| 5710 | break; |
| 5711 | case 's': |
| 5712 | single_line = TRUE; |
| 5713 | /* FALLTHRU */ |
| 5714 | case 'm': |
| 5715 | multi_line = TRUE; |
| 5716 | need_filebuf = TRUE; |
| 5717 | break; |
| 5718 | default: |
| 5719 | error ("invalid regexp modifier `%c', ignoring", modifiers[0]); |
| 5720 | break; |
| 5721 | } |
| 5722 | |
| 5723 | patbuf = xnew (1, struct re_pattern_buffer); |
| 5724 | *patbuf = zeropattern; |
| 5725 | if (ignore_case) |
| 5726 | { |
| 5727 | static char lc_trans[CHARS]; |
| 5728 | int i; |
| 5729 | for (i = 0; i < CHARS; i++) |
| 5730 | lc_trans[i] = lowcase (i); |
| 5731 | patbuf->translate = lc_trans; /* translation table to fold case */ |
| 5732 | } |
| 5733 | |
| 5734 | if (multi_line) |
| 5735 | pat = concat ("^", regexp_pattern, ""); /* anchor to beginning of line */ |
| 5736 | else |
| 5737 | pat = regexp_pattern; |
| 5738 | |
| 5739 | if (single_line) |
| 5740 | re_set_syntax (RE_SYNTAX_EMACS | RE_DOT_NEWLINE); |
| 5741 | else |
| 5742 | re_set_syntax (RE_SYNTAX_EMACS); |
| 5743 | |
| 5744 | err = re_compile_pattern (pat, strlen (pat), patbuf); |
| 5745 | if (multi_line) |
| 5746 | free (pat); |
| 5747 | if (err != NULL) |
| 5748 | { |
| 5749 | error ("%s while compiling pattern", err); |
| 5750 | return; |
| 5751 | } |
| 5752 | |
| 5753 | rp = p_head; |
| 5754 | p_head = xnew (1, regexp); |
| 5755 | p_head->pattern = savestr (regexp_pattern); |
| 5756 | p_head->p_next = rp; |
| 5757 | p_head->lang = lang; |
| 5758 | p_head->pat = patbuf; |
| 5759 | p_head->name = savestr (name); |
| 5760 | p_head->error_signaled = FALSE; |
| 5761 | p_head->force_explicit_name = force_explicit_name; |
| 5762 | p_head->ignore_case = ignore_case; |
| 5763 | p_head->multi_line = multi_line; |
| 5764 | } |
| 5765 | |
| 5766 | /* |
| 5767 | * Do the substitutions indicated by the regular expression and |
| 5768 | * arguments. |
| 5769 | */ |
| 5770 | static char * |
| 5771 | substitute (char *in, char *out, struct re_registers *regs) |
| 5772 | { |
| 5773 | char *result, *t; |
| 5774 | int size, dig, diglen; |
| 5775 | |
| 5776 | result = NULL; |
| 5777 | size = strlen (out); |
| 5778 | |
| 5779 | /* Pass 1: figure out how much to allocate by finding all \N strings. */ |
| 5780 | if (out[size - 1] == '\\') |
| 5781 | fatal ("pattern error in \"%s\"", out); |
| 5782 | for (t = etags_strchr (out, '\\'); |
| 5783 | t != NULL; |
| 5784 | t = etags_strchr (t + 2, '\\')) |
| 5785 | if (ISDIGIT (t[1])) |
| 5786 | { |
| 5787 | dig = t[1] - '0'; |
| 5788 | diglen = regs->end[dig] - regs->start[dig]; |
| 5789 | size += diglen - 2; |
| 5790 | } |
| 5791 | else |
| 5792 | size -= 1; |
| 5793 | |
| 5794 | /* Allocate space and do the substitutions. */ |
| 5795 | assert (size >= 0); |
| 5796 | result = xnew (size + 1, char); |
| 5797 | |
| 5798 | for (t = result; *out != '\0'; out++) |
| 5799 | if (*out == '\\' && ISDIGIT (*++out)) |
| 5800 | { |
| 5801 | dig = *out - '0'; |
| 5802 | diglen = regs->end[dig] - regs->start[dig]; |
| 5803 | memcpy (t, in + regs->start[dig], diglen); |
| 5804 | t += diglen; |
| 5805 | } |
| 5806 | else |
| 5807 | *t++ = *out; |
| 5808 | *t = '\0'; |
| 5809 | |
| 5810 | assert (t <= result + size); |
| 5811 | assert (t - result == (int)strlen (result)); |
| 5812 | |
| 5813 | return result; |
| 5814 | } |
| 5815 | |
| 5816 | /* Deallocate all regexps. */ |
| 5817 | static void |
| 5818 | free_regexps (void) |
| 5819 | { |
| 5820 | regexp *rp; |
| 5821 | while (p_head != NULL) |
| 5822 | { |
| 5823 | rp = p_head->p_next; |
| 5824 | free (p_head->pattern); |
| 5825 | free (p_head->name); |
| 5826 | free (p_head); |
| 5827 | p_head = rp; |
| 5828 | } |
| 5829 | return; |
| 5830 | } |
| 5831 | |
| 5832 | /* |
| 5833 | * Reads the whole file as a single string from `filebuf' and looks for |
| 5834 | * multi-line regular expressions, creating tags on matches. |
| 5835 | * readline already dealt with normal regexps. |
| 5836 | * |
| 5837 | * Idea by Ben Wing <ben@666.com> (2002). |
| 5838 | */ |
| 5839 | static void |
| 5840 | regex_tag_multiline (void) |
| 5841 | { |
| 5842 | char *buffer = filebuf.buffer; |
| 5843 | regexp *rp; |
| 5844 | char *name; |
| 5845 | |
| 5846 | for (rp = p_head; rp != NULL; rp = rp->p_next) |
| 5847 | { |
| 5848 | int match = 0; |
| 5849 | |
| 5850 | if (!rp->multi_line) |
| 5851 | continue; /* skip normal regexps */ |
| 5852 | |
| 5853 | /* Generic initializations before parsing file from memory. */ |
| 5854 | lineno = 1; /* reset global line number */ |
| 5855 | charno = 0; /* reset global char number */ |
| 5856 | linecharno = 0; /* reset global char number of line start */ |
| 5857 | |
| 5858 | /* Only use generic regexps or those for the current language. */ |
| 5859 | if (rp->lang != NULL && rp->lang != curfdp->lang) |
| 5860 | continue; |
| 5861 | |
| 5862 | while (match >= 0 && match < filebuf.len) |
| 5863 | { |
| 5864 | match = re_search (rp->pat, buffer, filebuf.len, charno, |
| 5865 | filebuf.len - match, &rp->regs); |
| 5866 | switch (match) |
| 5867 | { |
| 5868 | case -2: |
| 5869 | /* Some error. */ |
| 5870 | if (!rp->error_signaled) |
| 5871 | { |
| 5872 | error ("regexp stack overflow while matching \"%s\"", |
| 5873 | rp->pattern); |
| 5874 | rp->error_signaled = TRUE; |
| 5875 | } |
| 5876 | break; |
| 5877 | case -1: |
| 5878 | /* No match. */ |
| 5879 | break; |
| 5880 | default: |
| 5881 | if (match == rp->regs.end[0]) |
| 5882 | { |
| 5883 | if (!rp->error_signaled) |
| 5884 | { |
| 5885 | error ("regexp matches the empty string: \"%s\"", |
| 5886 | rp->pattern); |
| 5887 | rp->error_signaled = TRUE; |
| 5888 | } |
| 5889 | match = -3; /* exit from while loop */ |
| 5890 | break; |
| 5891 | } |
| 5892 | |
| 5893 | /* Match occurred. Construct a tag. */ |
| 5894 | while (charno < rp->regs.end[0]) |
| 5895 | if (buffer[charno++] == '\n') |
| 5896 | lineno++, linecharno = charno; |
| 5897 | name = rp->name; |
| 5898 | if (name[0] == '\0') |
| 5899 | name = NULL; |
| 5900 | else /* make a named tag */ |
| 5901 | name = substitute (buffer, rp->name, &rp->regs); |
| 5902 | if (rp->force_explicit_name) |
| 5903 | /* Force explicit tag name, if a name is there. */ |
| 5904 | pfnote (name, TRUE, buffer + linecharno, |
| 5905 | charno - linecharno + 1, lineno, linecharno); |
| 5906 | else |
| 5907 | make_tag (name, strlen (name), TRUE, buffer + linecharno, |
| 5908 | charno - linecharno + 1, lineno, linecharno); |
| 5909 | break; |
| 5910 | } |
| 5911 | } |
| 5912 | } |
| 5913 | } |
| 5914 | |
| 5915 | \f |
| 5916 | static bool |
| 5917 | nocase_tail (const char *cp) |
| 5918 | { |
| 5919 | register int len = 0; |
| 5920 | |
| 5921 | while (*cp != '\0' && lowcase (*cp) == lowcase (dbp[len])) |
| 5922 | cp++, len++; |
| 5923 | if (*cp == '\0' && !intoken (dbp[len])) |
| 5924 | { |
| 5925 | dbp += len; |
| 5926 | return TRUE; |
| 5927 | } |
| 5928 | return FALSE; |
| 5929 | } |
| 5930 | |
| 5931 | static void |
| 5932 | get_tag (register char *bp, char **namepp) |
| 5933 | { |
| 5934 | register char *cp = bp; |
| 5935 | |
| 5936 | if (*bp != '\0') |
| 5937 | { |
| 5938 | /* Go till you get to white space or a syntactic break */ |
| 5939 | for (cp = bp + 1; !notinname (*cp); cp++) |
| 5940 | continue; |
| 5941 | make_tag (bp, cp - bp, TRUE, |
| 5942 | lb.buffer, cp - lb.buffer + 1, lineno, linecharno); |
| 5943 | } |
| 5944 | |
| 5945 | if (namepp != NULL) |
| 5946 | *namepp = savenstr (bp, cp - bp); |
| 5947 | } |
| 5948 | |
| 5949 | /* |
| 5950 | * Read a line of text from `stream' into `lbp', excluding the |
| 5951 | * newline or CR-NL, if any. Return the number of characters read from |
| 5952 | * `stream', which is the length of the line including the newline. |
| 5953 | * |
| 5954 | * On DOS or Windows we do not count the CR character, if any before the |
| 5955 | * NL, in the returned length; this mirrors the behavior of Emacs on those |
| 5956 | * platforms (for text files, it translates CR-NL to NL as it reads in the |
| 5957 | * file). |
| 5958 | * |
| 5959 | * If multi-line regular expressions are requested, each line read is |
| 5960 | * appended to `filebuf'. |
| 5961 | */ |
| 5962 | static long |
| 5963 | readline_internal (linebuffer *lbp, register FILE *stream) |
| 5964 | { |
| 5965 | char *buffer = lbp->buffer; |
| 5966 | register char *p = lbp->buffer; |
| 5967 | register char *pend; |
| 5968 | int chars_deleted; |
| 5969 | |
| 5970 | pend = p + lbp->size; /* Separate to avoid 386/IX compiler bug. */ |
| 5971 | |
| 5972 | for (;;) |
| 5973 | { |
| 5974 | register int c = getc (stream); |
| 5975 | if (p == pend) |
| 5976 | { |
| 5977 | /* We're at the end of linebuffer: expand it. */ |
| 5978 | lbp->size *= 2; |
| 5979 | xrnew (buffer, lbp->size, char); |
| 5980 | p += buffer - lbp->buffer; |
| 5981 | pend = buffer + lbp->size; |
| 5982 | lbp->buffer = buffer; |
| 5983 | } |
| 5984 | if (c == EOF) |
| 5985 | { |
| 5986 | *p = '\0'; |
| 5987 | chars_deleted = 0; |
| 5988 | break; |
| 5989 | } |
| 5990 | if (c == '\n') |
| 5991 | { |
| 5992 | if (p > buffer && p[-1] == '\r') |
| 5993 | { |
| 5994 | p -= 1; |
| 5995 | #ifdef DOS_NT |
| 5996 | /* Assume CRLF->LF translation will be performed by Emacs |
| 5997 | when loading this file, so CRs won't appear in the buffer. |
| 5998 | It would be cleaner to compensate within Emacs; |
| 5999 | however, Emacs does not know how many CRs were deleted |
| 6000 | before any given point in the file. */ |
| 6001 | chars_deleted = 1; |
| 6002 | #else |
| 6003 | chars_deleted = 2; |
| 6004 | #endif |
| 6005 | } |
| 6006 | else |
| 6007 | { |
| 6008 | chars_deleted = 1; |
| 6009 | } |
| 6010 | *p = '\0'; |
| 6011 | break; |
| 6012 | } |
| 6013 | *p++ = c; |
| 6014 | } |
| 6015 | lbp->len = p - buffer; |
| 6016 | |
| 6017 | if (need_filebuf /* we need filebuf for multi-line regexps */ |
| 6018 | && chars_deleted > 0) /* not at EOF */ |
| 6019 | { |
| 6020 | while (filebuf.size <= filebuf.len + lbp->len + 1) /* +1 for \n */ |
| 6021 | { |
| 6022 | /* Expand filebuf. */ |
| 6023 | filebuf.size *= 2; |
| 6024 | xrnew (filebuf.buffer, filebuf.size, char); |
| 6025 | } |
| 6026 | memcpy (filebuf.buffer + filebuf.len, lbp->buffer, lbp->len); |
| 6027 | filebuf.len += lbp->len; |
| 6028 | filebuf.buffer[filebuf.len++] = '\n'; |
| 6029 | filebuf.buffer[filebuf.len] = '\0'; |
| 6030 | } |
| 6031 | |
| 6032 | return lbp->len + chars_deleted; |
| 6033 | } |
| 6034 | |
| 6035 | /* |
| 6036 | * Like readline_internal, above, but in addition try to match the |
| 6037 | * input line against relevant regular expressions and manage #line |
| 6038 | * directives. |
| 6039 | */ |
| 6040 | static void |
| 6041 | readline (linebuffer *lbp, FILE *stream) |
| 6042 | { |
| 6043 | long result; |
| 6044 | |
| 6045 | linecharno = charno; /* update global char number of line start */ |
| 6046 | result = readline_internal (lbp, stream); /* read line */ |
| 6047 | lineno += 1; /* increment global line number */ |
| 6048 | charno += result; /* increment global char number */ |
| 6049 | |
| 6050 | /* Honor #line directives. */ |
| 6051 | if (!no_line_directive) |
| 6052 | { |
| 6053 | static bool discard_until_line_directive; |
| 6054 | |
| 6055 | /* Check whether this is a #line directive. */ |
| 6056 | if (result > 12 && strneq (lbp->buffer, "#line ", 6)) |
| 6057 | { |
| 6058 | unsigned int lno; |
| 6059 | int start = 0; |
| 6060 | |
| 6061 | if (sscanf (lbp->buffer, "#line %u \"%n", &lno, &start) >= 1 |
| 6062 | && start > 0) /* double quote character found */ |
| 6063 | { |
| 6064 | char *endp = lbp->buffer + start; |
| 6065 | |
| 6066 | while ((endp = etags_strchr (endp, '"')) != NULL |
| 6067 | && endp[-1] == '\\') |
| 6068 | endp++; |
| 6069 | if (endp != NULL) |
| 6070 | /* Ok, this is a real #line directive. Let's deal with it. */ |
| 6071 | { |
| 6072 | char *taggedabsname; /* absolute name of original file */ |
| 6073 | char *taggedfname; /* name of original file as given */ |
| 6074 | char *name; /* temp var */ |
| 6075 | |
| 6076 | discard_until_line_directive = FALSE; /* found it */ |
| 6077 | name = lbp->buffer + start; |
| 6078 | *endp = '\0'; |
| 6079 | canonicalize_filename (name); |
| 6080 | taggedabsname = absolute_filename (name, tagfiledir); |
| 6081 | if (filename_is_absolute (name) |
| 6082 | || filename_is_absolute (curfdp->infname)) |
| 6083 | taggedfname = savestr (taggedabsname); |
| 6084 | else |
| 6085 | taggedfname = relative_filename (taggedabsname,tagfiledir); |
| 6086 | |
| 6087 | if (streq (curfdp->taggedfname, taggedfname)) |
| 6088 | /* The #line directive is only a line number change. We |
| 6089 | deal with this afterwards. */ |
| 6090 | free (taggedfname); |
| 6091 | else |
| 6092 | /* The tags following this #line directive should be |
| 6093 | attributed to taggedfname. In order to do this, set |
| 6094 | curfdp accordingly. */ |
| 6095 | { |
| 6096 | fdesc *fdp; /* file description pointer */ |
| 6097 | |
| 6098 | /* Go look for a file description already set up for the |
| 6099 | file indicated in the #line directive. If there is |
| 6100 | one, use it from now until the next #line |
| 6101 | directive. */ |
| 6102 | for (fdp = fdhead; fdp != NULL; fdp = fdp->next) |
| 6103 | if (streq (fdp->infname, curfdp->infname) |
| 6104 | && streq (fdp->taggedfname, taggedfname)) |
| 6105 | /* If we remove the second test above (after the &&) |
| 6106 | then all entries pertaining to the same file are |
| 6107 | coalesced in the tags file. If we use it, then |
| 6108 | entries pertaining to the same file but generated |
| 6109 | from different files (via #line directives) will |
| 6110 | go into separate sections in the tags file. These |
| 6111 | alternatives look equivalent. The first one |
| 6112 | destroys some apparently useless information. */ |
| 6113 | { |
| 6114 | curfdp = fdp; |
| 6115 | free (taggedfname); |
| 6116 | break; |
| 6117 | } |
| 6118 | /* Else, if we already tagged the real file, skip all |
| 6119 | input lines until the next #line directive. */ |
| 6120 | if (fdp == NULL) /* not found */ |
| 6121 | for (fdp = fdhead; fdp != NULL; fdp = fdp->next) |
| 6122 | if (streq (fdp->infabsname, taggedabsname)) |
| 6123 | { |
| 6124 | discard_until_line_directive = TRUE; |
| 6125 | free (taggedfname); |
| 6126 | break; |
| 6127 | } |
| 6128 | /* Else create a new file description and use that from |
| 6129 | now on, until the next #line directive. */ |
| 6130 | if (fdp == NULL) /* not found */ |
| 6131 | { |
| 6132 | fdp = fdhead; |
| 6133 | fdhead = xnew (1, fdesc); |
| 6134 | *fdhead = *curfdp; /* copy curr. file description */ |
| 6135 | fdhead->next = fdp; |
| 6136 | fdhead->infname = savestr (curfdp->infname); |
| 6137 | fdhead->infabsname = savestr (curfdp->infabsname); |
| 6138 | fdhead->infabsdir = savestr (curfdp->infabsdir); |
| 6139 | fdhead->taggedfname = taggedfname; |
| 6140 | fdhead->usecharno = FALSE; |
| 6141 | fdhead->prop = NULL; |
| 6142 | fdhead->written = FALSE; |
| 6143 | curfdp = fdhead; |
| 6144 | } |
| 6145 | } |
| 6146 | free (taggedabsname); |
| 6147 | lineno = lno - 1; |
| 6148 | readline (lbp, stream); |
| 6149 | return; |
| 6150 | } /* if a real #line directive */ |
| 6151 | } /* if #line is followed by a number */ |
| 6152 | } /* if line begins with "#line " */ |
| 6153 | |
| 6154 | /* If we are here, no #line directive was found. */ |
| 6155 | if (discard_until_line_directive) |
| 6156 | { |
| 6157 | if (result > 0) |
| 6158 | { |
| 6159 | /* Do a tail recursion on ourselves, thus discarding the contents |
| 6160 | of the line buffer. */ |
| 6161 | readline (lbp, stream); |
| 6162 | return; |
| 6163 | } |
| 6164 | /* End of file. */ |
| 6165 | discard_until_line_directive = FALSE; |
| 6166 | return; |
| 6167 | } |
| 6168 | } /* if #line directives should be considered */ |
| 6169 | |
| 6170 | { |
| 6171 | int match; |
| 6172 | regexp *rp; |
| 6173 | char *name; |
| 6174 | |
| 6175 | /* Match against relevant regexps. */ |
| 6176 | if (lbp->len > 0) |
| 6177 | for (rp = p_head; rp != NULL; rp = rp->p_next) |
| 6178 | { |
| 6179 | /* Only use generic regexps or those for the current language. |
| 6180 | Also do not use multiline regexps, which is the job of |
| 6181 | regex_tag_multiline. */ |
| 6182 | if ((rp->lang != NULL && rp->lang != fdhead->lang) |
| 6183 | || rp->multi_line) |
| 6184 | continue; |
| 6185 | |
| 6186 | match = re_match (rp->pat, lbp->buffer, lbp->len, 0, &rp->regs); |
| 6187 | switch (match) |
| 6188 | { |
| 6189 | case -2: |
| 6190 | /* Some error. */ |
| 6191 | if (!rp->error_signaled) |
| 6192 | { |
| 6193 | error ("regexp stack overflow while matching \"%s\"", |
| 6194 | rp->pattern); |
| 6195 | rp->error_signaled = TRUE; |
| 6196 | } |
| 6197 | break; |
| 6198 | case -1: |
| 6199 | /* No match. */ |
| 6200 | break; |
| 6201 | case 0: |
| 6202 | /* Empty string matched. */ |
| 6203 | if (!rp->error_signaled) |
| 6204 | { |
| 6205 | error ("regexp matches the empty string: \"%s\"", rp->pattern); |
| 6206 | rp->error_signaled = TRUE; |
| 6207 | } |
| 6208 | break; |
| 6209 | default: |
| 6210 | /* Match occurred. Construct a tag. */ |
| 6211 | name = rp->name; |
| 6212 | if (name[0] == '\0') |
| 6213 | name = NULL; |
| 6214 | else /* make a named tag */ |
| 6215 | name = substitute (lbp->buffer, rp->name, &rp->regs); |
| 6216 | if (rp->force_explicit_name) |
| 6217 | /* Force explicit tag name, if a name is there. */ |
| 6218 | pfnote (name, TRUE, lbp->buffer, match, lineno, linecharno); |
| 6219 | else |
| 6220 | make_tag (name, strlen (name), TRUE, |
| 6221 | lbp->buffer, match, lineno, linecharno); |
| 6222 | break; |
| 6223 | } |
| 6224 | } |
| 6225 | } |
| 6226 | } |
| 6227 | |
| 6228 | \f |
| 6229 | /* |
| 6230 | * Return a pointer to a space of size strlen(cp)+1 allocated |
| 6231 | * with xnew where the string CP has been copied. |
| 6232 | */ |
| 6233 | static char * |
| 6234 | savestr (const char *cp) |
| 6235 | { |
| 6236 | return savenstr (cp, strlen (cp)); |
| 6237 | } |
| 6238 | |
| 6239 | /* |
| 6240 | * Return a pointer to a space of size LEN+1 allocated with xnew where |
| 6241 | * the string CP has been copied for at most the first LEN characters. |
| 6242 | */ |
| 6243 | static char * |
| 6244 | savenstr (const char *cp, int len) |
| 6245 | { |
| 6246 | char *dp = xnew (len + 1, char); |
| 6247 | dp[len] = '\0'; |
| 6248 | return memcpy (dp, cp, len); |
| 6249 | } |
| 6250 | |
| 6251 | /* |
| 6252 | * Return the ptr in sp at which the character c last |
| 6253 | * appears; NULL if not found |
| 6254 | * |
| 6255 | * Identical to POSIX strrchr, included for portability. |
| 6256 | */ |
| 6257 | static char * |
| 6258 | etags_strrchr (register const char *sp, register int c) |
| 6259 | { |
| 6260 | register const char *r; |
| 6261 | |
| 6262 | r = NULL; |
| 6263 | do |
| 6264 | { |
| 6265 | if (*sp == c) |
| 6266 | r = sp; |
| 6267 | } while (*sp++); |
| 6268 | return (char *)r; |
| 6269 | } |
| 6270 | |
| 6271 | /* |
| 6272 | * Return the ptr in sp at which the character c first |
| 6273 | * appears; NULL if not found |
| 6274 | * |
| 6275 | * Identical to POSIX strchr, included for portability. |
| 6276 | */ |
| 6277 | static char * |
| 6278 | etags_strchr (register const char *sp, register int c) |
| 6279 | { |
| 6280 | do |
| 6281 | { |
| 6282 | if (*sp == c) |
| 6283 | return (char *)sp; |
| 6284 | } while (*sp++); |
| 6285 | return NULL; |
| 6286 | } |
| 6287 | |
| 6288 | /* Skip spaces (end of string is not space), return new pointer. */ |
| 6289 | static char * |
| 6290 | skip_spaces (char *cp) |
| 6291 | { |
| 6292 | while (iswhite (*cp)) |
| 6293 | cp++; |
| 6294 | return cp; |
| 6295 | } |
| 6296 | |
| 6297 | /* Skip non spaces, except end of string, return new pointer. */ |
| 6298 | static char * |
| 6299 | skip_non_spaces (char *cp) |
| 6300 | { |
| 6301 | while (*cp != '\0' && !iswhite (*cp)) |
| 6302 | cp++; |
| 6303 | return cp; |
| 6304 | } |
| 6305 | |
| 6306 | /* Skip any chars in the "name" class.*/ |
| 6307 | static char * |
| 6308 | skip_name (char *cp) |
| 6309 | { |
| 6310 | /* '\0' is a notinname() so loop stops there too */ |
| 6311 | while (! notinname (*cp)) |
| 6312 | cp++; |
| 6313 | return cp; |
| 6314 | } |
| 6315 | |
| 6316 | /* Print error message and exit. */ |
| 6317 | void |
| 6318 | fatal (const char *s1, const char *s2) |
| 6319 | { |
| 6320 | error (s1, s2); |
| 6321 | exit (EXIT_FAILURE); |
| 6322 | } |
| 6323 | |
| 6324 | static void |
| 6325 | pfatal (const char *s1) |
| 6326 | { |
| 6327 | perror (s1); |
| 6328 | exit (EXIT_FAILURE); |
| 6329 | } |
| 6330 | |
| 6331 | static void |
| 6332 | suggest_asking_for_help (void) |
| 6333 | { |
| 6334 | fprintf (stderr, "\tTry `%s --help' for a complete list of options.\n", |
| 6335 | progname); |
| 6336 | exit (EXIT_FAILURE); |
| 6337 | } |
| 6338 | |
| 6339 | /* Output a diagnostic with printf-style FORMAT and args. */ |
| 6340 | static void |
| 6341 | error (const char *format, ...) |
| 6342 | { |
| 6343 | va_list ap; |
| 6344 | va_start (ap, format); |
| 6345 | fprintf (stderr, "%s: ", progname); |
| 6346 | vfprintf (stderr, format, ap); |
| 6347 | fprintf (stderr, "\n"); |
| 6348 | va_end (ap); |
| 6349 | } |
| 6350 | |
| 6351 | /* Return a newly-allocated string whose contents |
| 6352 | concatenate those of s1, s2, s3. */ |
| 6353 | static char * |
| 6354 | concat (const char *s1, const char *s2, const char *s3) |
| 6355 | { |
| 6356 | int len1 = strlen (s1), len2 = strlen (s2), len3 = strlen (s3); |
| 6357 | char *result = xnew (len1 + len2 + len3 + 1, char); |
| 6358 | |
| 6359 | strcpy (result, s1); |
| 6360 | strcpy (result + len1, s2); |
| 6361 | strcpy (result + len1 + len2, s3); |
| 6362 | |
| 6363 | return result; |
| 6364 | } |
| 6365 | |
| 6366 | \f |
| 6367 | /* Does the same work as the system V getcwd, but does not need to |
| 6368 | guess the buffer size in advance. */ |
| 6369 | static char * |
| 6370 | etags_getcwd (void) |
| 6371 | { |
| 6372 | int bufsize = 200; |
| 6373 | char *path = xnew (bufsize, char); |
| 6374 | |
| 6375 | while (getcwd (path, bufsize) == NULL) |
| 6376 | { |
| 6377 | if (errno != ERANGE) |
| 6378 | pfatal ("getcwd"); |
| 6379 | bufsize *= 2; |
| 6380 | free (path); |
| 6381 | path = xnew (bufsize, char); |
| 6382 | } |
| 6383 | |
| 6384 | canonicalize_filename (path); |
| 6385 | return path; |
| 6386 | } |
| 6387 | |
| 6388 | /* Return a newly allocated string containing the file name of FILE |
| 6389 | relative to the absolute directory DIR (which should end with a slash). */ |
| 6390 | static char * |
| 6391 | relative_filename (char *file, char *dir) |
| 6392 | { |
| 6393 | char *fp, *dp, *afn, *res; |
| 6394 | int i; |
| 6395 | |
| 6396 | /* Find the common root of file and dir (with a trailing slash). */ |
| 6397 | afn = absolute_filename (file, cwd); |
| 6398 | fp = afn; |
| 6399 | dp = dir; |
| 6400 | while (*fp++ == *dp++) |
| 6401 | continue; |
| 6402 | fp--, dp--; /* back to the first differing char */ |
| 6403 | #ifdef DOS_NT |
| 6404 | if (fp == afn && afn[0] != '/') /* cannot build a relative name */ |
| 6405 | return afn; |
| 6406 | #endif |
| 6407 | do /* look at the equal chars until '/' */ |
| 6408 | fp--, dp--; |
| 6409 | while (*fp != '/'); |
| 6410 | |
| 6411 | /* Build a sequence of "../" strings for the resulting relative file name. */ |
| 6412 | i = 0; |
| 6413 | while ((dp = etags_strchr (dp + 1, '/')) != NULL) |
| 6414 | i += 1; |
| 6415 | res = xnew (3*i + strlen (fp + 1) + 1, char); |
| 6416 | res[0] = '\0'; |
| 6417 | while (i-- > 0) |
| 6418 | strcat (res, "../"); |
| 6419 | |
| 6420 | /* Add the file name relative to the common root of file and dir. */ |
| 6421 | strcat (res, fp + 1); |
| 6422 | free (afn); |
| 6423 | |
| 6424 | return res; |
| 6425 | } |
| 6426 | |
| 6427 | /* Return a newly allocated string containing the absolute file name |
| 6428 | of FILE given DIR (which should end with a slash). */ |
| 6429 | static char * |
| 6430 | absolute_filename (char *file, char *dir) |
| 6431 | { |
| 6432 | char *slashp, *cp, *res; |
| 6433 | |
| 6434 | if (filename_is_absolute (file)) |
| 6435 | res = savestr (file); |
| 6436 | #ifdef DOS_NT |
| 6437 | /* We don't support non-absolute file names with a drive |
| 6438 | letter, like `d:NAME' (it's too much hassle). */ |
| 6439 | else if (file[1] == ':') |
| 6440 | fatal ("%s: relative file names with drive letters not supported", file); |
| 6441 | #endif |
| 6442 | else |
| 6443 | res = concat (dir, file, ""); |
| 6444 | |
| 6445 | /* Delete the "/dirname/.." and "/." substrings. */ |
| 6446 | slashp = etags_strchr (res, '/'); |
| 6447 | while (slashp != NULL && slashp[0] != '\0') |
| 6448 | { |
| 6449 | if (slashp[1] == '.') |
| 6450 | { |
| 6451 | if (slashp[2] == '.' |
| 6452 | && (slashp[3] == '/' || slashp[3] == '\0')) |
| 6453 | { |
| 6454 | cp = slashp; |
| 6455 | do |
| 6456 | cp--; |
| 6457 | while (cp >= res && !filename_is_absolute (cp)); |
| 6458 | if (cp < res) |
| 6459 | cp = slashp; /* the absolute name begins with "/.." */ |
| 6460 | #ifdef DOS_NT |
| 6461 | /* Under MSDOS and NT we get `d:/NAME' as absolute |
| 6462 | file name, so the luser could say `d:/../NAME'. |
| 6463 | We silently treat this as `d:/NAME'. */ |
| 6464 | else if (cp[0] != '/') |
| 6465 | cp = slashp; |
| 6466 | #endif |
| 6467 | memmove (cp, slashp + 3, strlen (slashp + 2)); |
| 6468 | slashp = cp; |
| 6469 | continue; |
| 6470 | } |
| 6471 | else if (slashp[2] == '/' || slashp[2] == '\0') |
| 6472 | { |
| 6473 | memmove (slashp, slashp + 2, strlen (slashp + 1)); |
| 6474 | continue; |
| 6475 | } |
| 6476 | } |
| 6477 | |
| 6478 | slashp = etags_strchr (slashp + 1, '/'); |
| 6479 | } |
| 6480 | |
| 6481 | if (res[0] == '\0') /* just a safety net: should never happen */ |
| 6482 | { |
| 6483 | free (res); |
| 6484 | return savestr ("/"); |
| 6485 | } |
| 6486 | else |
| 6487 | return res; |
| 6488 | } |
| 6489 | |
| 6490 | /* Return a newly allocated string containing the absolute |
| 6491 | file name of dir where FILE resides given DIR (which should |
| 6492 | end with a slash). */ |
| 6493 | static char * |
| 6494 | absolute_dirname (char *file, char *dir) |
| 6495 | { |
| 6496 | char *slashp, *res; |
| 6497 | char save; |
| 6498 | |
| 6499 | slashp = etags_strrchr (file, '/'); |
| 6500 | if (slashp == NULL) |
| 6501 | return savestr (dir); |
| 6502 | save = slashp[1]; |
| 6503 | slashp[1] = '\0'; |
| 6504 | res = absolute_filename (file, dir); |
| 6505 | slashp[1] = save; |
| 6506 | |
| 6507 | return res; |
| 6508 | } |
| 6509 | |
| 6510 | /* Whether the argument string is an absolute file name. The argument |
| 6511 | string must have been canonicalized with canonicalize_filename. */ |
| 6512 | static bool |
| 6513 | filename_is_absolute (char *fn) |
| 6514 | { |
| 6515 | return (fn[0] == '/' |
| 6516 | #ifdef DOS_NT |
| 6517 | || (ISALPHA (fn[0]) && fn[1] == ':' && fn[2] == '/') |
| 6518 | #endif |
| 6519 | ); |
| 6520 | } |
| 6521 | |
| 6522 | /* Downcase DOS drive letter and collapse separators into single slashes. |
| 6523 | Works in place. */ |
| 6524 | static void |
| 6525 | canonicalize_filename (register char *fn) |
| 6526 | { |
| 6527 | register char* cp; |
| 6528 | char sep = '/'; |
| 6529 | |
| 6530 | #ifdef DOS_NT |
| 6531 | /* Canonicalize drive letter case. */ |
| 6532 | # define ISUPPER(c) isupper (CHAR (c)) |
| 6533 | if (fn[0] != '\0' && fn[1] == ':' && ISUPPER (fn[0])) |
| 6534 | fn[0] = lowcase (fn[0]); |
| 6535 | |
| 6536 | sep = '\\'; |
| 6537 | #endif |
| 6538 | |
| 6539 | /* Collapse multiple separators into a single slash. */ |
| 6540 | for (cp = fn; *cp != '\0'; cp++, fn++) |
| 6541 | if (*cp == sep) |
| 6542 | { |
| 6543 | *fn = '/'; |
| 6544 | while (cp[1] == sep) |
| 6545 | cp++; |
| 6546 | } |
| 6547 | else |
| 6548 | *fn = *cp; |
| 6549 | *fn = '\0'; |
| 6550 | } |
| 6551 | |
| 6552 | \f |
| 6553 | /* Initialize a linebuffer for use. */ |
| 6554 | static void |
| 6555 | linebuffer_init (linebuffer *lbp) |
| 6556 | { |
| 6557 | lbp->size = (DEBUG) ? 3 : 200; |
| 6558 | lbp->buffer = xnew (lbp->size, char); |
| 6559 | lbp->buffer[0] = '\0'; |
| 6560 | lbp->len = 0; |
| 6561 | } |
| 6562 | |
| 6563 | /* Set the minimum size of a string contained in a linebuffer. */ |
| 6564 | static void |
| 6565 | linebuffer_setlen (linebuffer *lbp, int toksize) |
| 6566 | { |
| 6567 | while (lbp->size <= toksize) |
| 6568 | { |
| 6569 | lbp->size *= 2; |
| 6570 | xrnew (lbp->buffer, lbp->size, char); |
| 6571 | } |
| 6572 | lbp->len = toksize; |
| 6573 | } |
| 6574 | |
| 6575 | /* Like malloc but get fatal error if memory is exhausted. */ |
| 6576 | static void * |
| 6577 | xmalloc (size_t size) |
| 6578 | { |
| 6579 | void *result = malloc (size); |
| 6580 | if (result == NULL) |
| 6581 | fatal ("virtual memory exhausted", (char *)NULL); |
| 6582 | return result; |
| 6583 | } |
| 6584 | |
| 6585 | static void * |
| 6586 | xrealloc (char *ptr, size_t size) |
| 6587 | { |
| 6588 | void *result = realloc (ptr, size); |
| 6589 | if (result == NULL) |
| 6590 | fatal ("virtual memory exhausted", (char *)NULL); |
| 6591 | return result; |
| 6592 | } |
| 6593 | |
| 6594 | /* |
| 6595 | * Local Variables: |
| 6596 | * indent-tabs-mode: t |
| 6597 | * tab-width: 8 |
| 6598 | * fill-column: 79 |
| 6599 | * c-font-lock-extra-types: ("FILE" "bool" "language" "linebuffer" "fdesc" "node" "regexp") |
| 6600 | * c-file-style: "gnu" |
| 6601 | * End: |
| 6602 | */ |
| 6603 | |
| 6604 | /* etags.c ends here */ |