1 /* Tags file maker to go with GNU Emacs -*- coding: latin-1 -*-
3 Copyright (C) 1984 The Regents of the University of California
5 Redistribution and use in source and binary forms, with or without
6 modification, are permitted provided that the following conditions are
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
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.
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.
31 Copyright (C) 1984, 1987, 1988, 1989, 1993, 1994, 1995, 1998, 1999,
32 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008
33 Free Software Foundation, Inc.
35 This file is not considered part of GNU Emacs.
37 This program is free software: you can redistribute it and/or modify
38 it under the terms of the GNU General Public License as published by
39 the Free Software Foundation, either version 3 of the License, or
40 (at your option) any later version.
42 This program is distributed in the hope that it will be useful,
43 but WITHOUT ANY WARRANTY; without even the implied warranty of
44 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
45 GNU General Public License for more details.
47 You should have received a copy of the GNU General Public License
48 along with this program. If not, see <http://www.gnu.org/licenses/>. */
51 /* NB To comply with the above BSD license, copyright information is
52 reproduced in etc/ETAGS.README. That file should be updated when the
55 To the best of our knowledge, this code was originally based on the
56 ctags.c distributed with BSD4.2, which was copyrighted by the
57 University of California, as described above. */
62 * 1983 Ctags originally by Ken Arnold.
63 * 1984 Fortran added by Jim Kleckner.
64 * 1984 Ed Pelegri-Llopart added C typedefs.
65 * 1985 Emacs TAGS format by Richard Stallman.
66 * 1989 Sam Kendall added C++.
67 * 1992 Joseph B. Wells improved C and C++ parsing.
68 * 1993 Francesco Potortì reorganized C and C++.
69 * 1994 Line-by-line regexp tags by Tom Tromey.
70 * 2001 Nested classes by Francesco Potortì (concept by Mykola Dzyuba).
71 * 2002 #line directives by Francesco Potortì.
73 * Francesco Potortì <pot@gnu.org> has maintained and improved it since 1993.
77 * If you want to add support for a new language, start by looking at the LUA
78 * language, which is the simplest. Alternatively, consider distributing etags
79 * together with a configuration file containing regexp definitions for etags.
82 char pot_etags_version[] = "@(#) pot revision number is 17.38.1.4";
92 # define NDEBUG /* disable assert */
97 /* On some systems, Emacs defines static as nothing for the sake
98 of unexec. We don't want that here since we don't use unexec. */
100 # ifndef PTR /* for XEmacs */
103 # ifndef __P /* for XEmacs */
104 # define __P(args) args
106 #else /* no config.h */
107 # if defined(__STDC__) && (__STDC__ || defined(__SUNPRO_C))
108 # define __P(args) args /* use prototypes */
109 # define PTR void * /* for generic pointers */
110 # else /* not standard C */
111 # define __P(args) () /* no prototypes */
112 # define const /* remove const for old compilers' sake */
113 # define PTR long * /* don't use void* */
115 #endif /* !HAVE_CONFIG_H */
118 # define _GNU_SOURCE 1 /* enables some compiler checks on GNU */
126 #else /* no standard C headers */
127 extern char *getenv __P((const char *));
128 extern char *strcpy __P((char *, const char *));
129 extern char *strncpy __P((char *, const char *, unsigned long));
130 extern char *strcat __P((char *, const char *));
131 extern char *strncat __P((char *, const char *, unsigned long));
132 extern int strcmp __P((const char *, const char *));
133 extern int strncmp __P((const char *, const char *, unsigned long));
134 extern int system __P((const char *));
135 extern unsigned long strlen __P((const char *));
136 extern void *malloc __P((unsigned long));
137 extern void *realloc __P((void *, unsigned long));
138 extern void exit __P((int));
139 extern void free __P((void *));
140 extern void *memmove __P((void *, const void *, unsigned long));
141 # define EXIT_SUCCESS 0
142 # define EXIT_FAILURE 1
143 #endif /* STDC_HEADERS */
149 extern char *getcwd (char *buf, size_t size);
151 #endif /* HAVE_UNISTD_H */
159 #include <sys/types.h>
160 #include <sys/stat.h>
164 # undef assert /* some systems have a buggy assert.h */
165 # define assert(x) ((void) 0)
168 #if !defined (S_ISREG) && defined (S_IFREG)
169 # define S_ISREG(m) (((m) & S_IFMT) == S_IFREG)
172 #ifndef HAVE_GETOPT_LONG
173 # define NO_LONG_OPTIONS TRUE
174 # define getopt_long(argc,argv,optstr,lopts,lind) getopt (argc, argv, optstr)
176 extern int optind, opterr;
178 # define NO_LONG_OPTIONS FALSE
180 #endif /* HAVE_GETOPT_LONG */
184 /* Define CTAGS to make the program "ctags" compatible with the usual one.
185 Leave it undefined to make the program "etags", which makes emacs-style
186 tag tables and tags typedefs, #defines and struct/union/enum by default. */
194 #define streq(s,t) (assert((s)!=NULL || (t)!=NULL), !strcmp (s, t))
195 #define strcaseeq(s,t) (assert((s)!=NULL && (t)!=NULL), !etags_strcasecmp (s, t))
196 #define strneq(s,t,n) (assert((s)!=NULL || (t)!=NULL), !strncmp (s, t, n))
197 #define strncaseeq(s,t,n) (assert((s)!=NULL && (t)!=NULL), !etags_strncasecmp (s, t, n))
199 #define CHARS 256 /* 2^sizeof(char) */
200 #define CHAR(x) ((unsigned int)(x) & (CHARS - 1))
201 #define iswhite(c) (_wht[CHAR(c)]) /* c is white (see white) */
202 #define notinname(c) (_nin[CHAR(c)]) /* c is not in a name (see nonam) */
203 #define begtoken(c) (_btk[CHAR(c)]) /* c can start token (see begtk) */
204 #define intoken(c) (_itk[CHAR(c)]) /* c can be in token (see midtk) */
205 #define endtoken(c) (_etk[CHAR(c)]) /* c ends tokens (see endtk) */
207 #define ISALNUM(c) isalnum (CHAR(c))
208 #define ISALPHA(c) isalpha (CHAR(c))
209 #define ISDIGIT(c) isdigit (CHAR(c))
210 #define ISLOWER(c) islower (CHAR(c))
212 #define lowcase(c) tolower (CHAR(c))
213 #define upcase(c) toupper (CHAR(c))
217 * xnew, xrnew -- allocate, reallocate storage
219 * SYNOPSIS: Type *xnew (int n, Type);
220 * void xrnew (OldPointer, int n, Type);
223 # include "chkmalloc.h"
224 # define xnew(n,Type) ((Type *) trace_malloc (__FILE__, __LINE__, \
225 (n) * sizeof (Type)))
226 # define xrnew(op,n,Type) ((op) = (Type *) trace_realloc (__FILE__, __LINE__, \
227 (char *) (op), (n) * sizeof (Type)))
229 # define xnew(n,Type) ((Type *) xmalloc ((n) * sizeof (Type)))
230 # define xrnew(op,n,Type) ((op) = (Type *) xrealloc ( \
231 (char *) (op), (n) * sizeof (Type)))
236 typedef void Lang_function __P((FILE *));
240 char *suffix; /* file name suffix for this compressor */
241 char *command; /* takes one arg and decompresses to stdout */
246 char *name; /* language name */
247 char *help; /* detailed help for the language */
248 Lang_function *function; /* parse function */
249 char **suffixes; /* name suffixes of this language's files */
250 char **filenames; /* names of this language's files */
251 char **interpreters; /* interpreters for this language */
252 bool metasource; /* source used to generate other sources */
257 struct fdesc *next; /* for the linked list */
258 char *infname; /* uncompressed input file name */
259 char *infabsname; /* absolute uncompressed input file name */
260 char *infabsdir; /* absolute dir of input file */
261 char *taggedfname; /* file name to write in tagfile */
262 language *lang; /* language of file */
263 char *prop; /* file properties to write in tagfile */
264 bool usecharno; /* etags tags shall contain char number */
265 bool written; /* entry written in the tags file */
268 typedef struct node_st
269 { /* sorting structure */
270 struct node_st *left, *right; /* left and right sons */
271 fdesc *fdp; /* description of file to whom tag belongs */
272 char *name; /* tag name */
273 char *regex; /* search regexp */
274 bool valid; /* write this tag on the tag file */
275 bool is_func; /* function tag: use regexp in CTAGS mode */
276 bool been_warned; /* warning already given for duplicated tag */
277 int lno; /* line number tag is on */
278 long cno; /* character number line starts on */
282 * A `linebuffer' is a structure which holds a line of text.
283 * `readline_internal' reads a line from a stream into a linebuffer
284 * and works regardless of the length of the line.
285 * SIZE is the size of BUFFER, LEN is the length of the string in
286 * BUFFER after readline reads it.
295 /* Used to support mixing of --lang and file names. */
299 at_language, /* a language specification */
300 at_regexp, /* a regular expression */
301 at_filename, /* a file name */
302 at_stdin, /* read from stdin here */
303 at_end /* stop parsing the list */
304 } arg_type; /* argument type */
305 language *lang; /* language associated with the argument */
306 char *what; /* the argument itself */
309 /* Structure defining a regular expression. */
310 typedef struct regexp
312 struct regexp *p_next; /* pointer to next in list */
313 language *lang; /* if set, use only for this language */
314 char *pattern; /* the regexp pattern */
315 char *name; /* tag name */
316 struct re_pattern_buffer *pat; /* the compiled pattern */
317 struct re_registers regs; /* re registers */
318 bool error_signaled; /* already signaled for this regexp */
319 bool force_explicit_name; /* do not allow implict tag name */
320 bool ignore_case; /* ignore case when matching */
321 bool multi_line; /* do a multi-line match on the whole file */
325 /* Many compilers barf on this:
326 Lang_function Ada_funcs;
327 so let's write it this way */
328 static void Ada_funcs __P((FILE *));
329 static void Asm_labels __P((FILE *));
330 static void C_entries __P((int c_ext, FILE *));
331 static void default_C_entries __P((FILE *));
332 static void plain_C_entries __P((FILE *));
333 static void Cjava_entries __P((FILE *));
334 static void Cobol_paragraphs __P((FILE *));
335 static void Cplusplus_entries __P((FILE *));
336 static void Cstar_entries __P((FILE *));
337 static void Erlang_functions __P((FILE *));
338 static void Forth_words __P((FILE *));
339 static void Fortran_functions __P((FILE *));
340 static void HTML_labels __P((FILE *));
341 static void Lisp_functions __P((FILE *));
342 static void Lua_functions __P((FILE *));
343 static void Makefile_targets __P((FILE *));
344 static void Pascal_functions __P((FILE *));
345 static void Perl_functions __P((FILE *));
346 static void PHP_functions __P((FILE *));
347 static void PS_functions __P((FILE *));
348 static void Prolog_functions __P((FILE *));
349 static void Python_functions __P((FILE *));
350 static void Scheme_functions __P((FILE *));
351 static void TeX_commands __P((FILE *));
352 static void Texinfo_nodes __P((FILE *));
353 static void Yacc_entries __P((FILE *));
354 static void just_read_file __P((FILE *));
356 static void print_language_names __P((void));
357 static void print_version __P((void));
358 static void print_help __P((argument *));
359 int main __P((int, char **));
361 static compressor *get_compressor_from_suffix __P((char *, char **));
362 static language *get_language_from_langname __P((const char *));
363 static language *get_language_from_interpreter __P((char *));
364 static language *get_language_from_filename __P((char *, bool));
365 static void readline __P((linebuffer *, FILE *));
366 static long readline_internal __P((linebuffer *, FILE *));
367 static bool nocase_tail __P((char *));
368 static void get_tag __P((char *, char **));
370 static void analyse_regex __P((char *));
371 static void free_regexps __P((void));
372 static void regex_tag_multiline __P((void));
373 static void error __P((const char *, const char *));
374 static void suggest_asking_for_help __P((void));
375 void fatal __P((char *, char *));
376 static void pfatal __P((char *));
377 static void add_node __P((node *, node **));
379 static void init __P((void));
380 static void process_file_name __P((char *, language *));
381 static void process_file __P((FILE *, char *, language *));
382 static void find_entries __P((FILE *));
383 static void free_tree __P((node *));
384 static void free_fdesc __P((fdesc *));
385 static void pfnote __P((char *, bool, char *, int, int, long));
386 static void make_tag __P((char *, int, bool, char *, int, int, long));
387 static void invalidate_nodes __P((fdesc *, node **));
388 static void put_entries __P((node *));
390 static char *concat __P((char *, char *, char *));
391 static char *skip_spaces __P((char *));
392 static char *skip_non_spaces __P((char *));
393 static char *savenstr __P((char *, int));
394 static char *savestr __P((char *));
395 static char *etags_strchr __P((const char *, int));
396 static char *etags_strrchr __P((const char *, int));
397 static int etags_strcasecmp __P((const char *, const char *));
398 static int etags_strncasecmp __P((const char *, const char *, int));
399 static char *etags_getcwd __P((void));
400 static char *relative_filename __P((char *, char *));
401 static char *absolute_filename __P((char *, char *));
402 static char *absolute_dirname __P((char *, char *));
403 static bool filename_is_absolute __P((char *f));
404 static void canonicalize_filename __P((char *));
405 static void linebuffer_init __P((linebuffer *));
406 static void linebuffer_setlen __P((linebuffer *, int));
407 static PTR xmalloc __P((unsigned int));
408 static PTR xrealloc __P((char *, unsigned int));
411 static char searchar = '/'; /* use /.../ searches */
413 static char *tagfile; /* output file */
414 static char *progname; /* name this program was invoked with */
415 static char *cwd; /* current working directory */
416 static char *tagfiledir; /* directory of tagfile */
417 static FILE *tagf; /* ioptr for tags file */
419 static fdesc *fdhead; /* head of file description list */
420 static fdesc *curfdp; /* current file description */
421 static int lineno; /* line number of current line */
422 static long charno; /* current character number */
423 static long linecharno; /* charno of start of current line */
424 static char *dbp; /* pointer to start of current tag */
426 static const int invalidcharno = -1;
428 static node *nodehead; /* the head of the binary tree of tags */
429 static node *last_node; /* the last node created */
431 static linebuffer lb; /* the current line */
432 static linebuffer filebuf; /* a buffer containing the whole file */
433 static linebuffer token_name; /* a buffer containing a tag name */
435 /* boolean "functions" (see init) */
436 static bool _wht[CHARS], _nin[CHARS], _itk[CHARS], _btk[CHARS], _etk[CHARS];
439 *white = " \f\t\n\r\v",
441 *nonam = " \f\t\n\r()=,;", /* look at make_tag before modifying! */
442 /* token ending chars */
443 *endtk = " \t\n\r\"'#()[]{}=-+%*/&|^~!<>;,.:?",
444 /* token starting chars */
445 *begtk = "ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz$~@",
446 /* valid in-token chars */
447 *midtk = "ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz$0123456789";
449 static bool append_to_tagfile; /* -a: append to tags */
450 /* The next five default to TRUE in C and derived languages. */
451 static bool typedefs; /* -t: create tags for C and Ada typedefs */
452 static bool typedefs_or_cplusplus; /* -T: create tags for C typedefs, level */
453 /* 0 struct/enum/union decls, and C++ */
454 /* member functions. */
455 static bool constantypedefs; /* -d: create tags for C #define, enum */
456 /* constants and variables. */
457 /* -D: opposite of -d. Default under ctags. */
458 static bool globals; /* create tags for global variables */
459 static bool members; /* create tags for C member variables */
460 static bool declarations; /* --declarations: tag them and extern in C&Co*/
461 static bool no_line_directive; /* ignore #line directives (undocumented) */
462 static bool no_duplicates; /* no duplicate tags for ctags (undocumented) */
463 static bool update; /* -u: update tags */
464 static bool vgrind_style; /* -v: create vgrind style index output */
465 static bool no_warnings; /* -w: suppress warnings (undocumented) */
466 static bool cxref_style; /* -x: create cxref style output */
467 static bool cplusplus; /* .[hc] means C++, not C (undocumented) */
468 static bool ignoreindent; /* -I: ignore indentation in C */
469 static bool packages_only; /* --packages-only: in Ada, only tag packages*/
471 /* STDIN is defined in LynxOS system headers */
476 #define STDIN 0x1001 /* returned by getopt_long on --parse-stdin */
477 static bool parsing_stdin; /* --parse-stdin used */
479 static regexp *p_head; /* list of all regexps */
480 static bool need_filebuf; /* some regexes are multi-line */
482 #if NO_LONG_OPTIONS == FALSE
483 static struct option longopts[] =
485 { "append", no_argument, NULL, 'a' },
486 { "packages-only", no_argument, &packages_only, TRUE },
487 { "c++", no_argument, NULL, 'C' },
488 { "declarations", no_argument, &declarations, TRUE },
489 { "no-line-directive", no_argument, &no_line_directive, TRUE },
490 { "no-duplicates", no_argument, &no_duplicates, TRUE },
491 { "help", no_argument, NULL, 'h' },
492 { "help", no_argument, NULL, 'H' },
493 { "ignore-indentation", no_argument, NULL, 'I' },
494 { "language", required_argument, NULL, 'l' },
495 { "members", no_argument, &members, TRUE },
496 { "no-members", no_argument, &members, FALSE },
497 { "output", required_argument, NULL, 'o' },
498 { "regex", required_argument, NULL, 'r' },
499 { "no-regex", no_argument, NULL, 'R' },
500 { "ignore-case-regex", required_argument, NULL, 'c' },
501 { "parse-stdin", required_argument, NULL, STDIN },
502 { "version", no_argument, NULL, 'V' },
504 #if CTAGS /* Ctags options */
505 { "backward-search", no_argument, NULL, 'B' },
506 { "cxref", no_argument, NULL, 'x' },
507 { "defines", no_argument, NULL, 'd' },
508 { "globals", no_argument, &globals, TRUE },
509 { "typedefs", no_argument, NULL, 't' },
510 { "typedefs-and-c++", no_argument, NULL, 'T' },
511 { "update", no_argument, NULL, 'u' },
512 { "vgrind", no_argument, NULL, 'v' },
513 { "no-warn", no_argument, NULL, 'w' },
515 #else /* Etags options */
516 { "no-defines", no_argument, NULL, 'D' },
517 { "no-globals", no_argument, &globals, FALSE },
518 { "include", required_argument, NULL, 'i' },
525 static compressor compressors[] =
527 { "z", "gzip -d -c"},
528 { "Z", "gzip -d -c"},
529 { "gz", "gzip -d -c"},
530 { "GZ", "gzip -d -c"},
531 { "bz2", "bzip2 -d -c" },
540 static char *Ada_suffixes [] =
541 { "ads", "adb", "ada", NULL };
542 static char Ada_help [] =
543 "In Ada code, functions, procedures, packages, tasks and types are\n\
544 tags. Use the `--packages-only' option to create tags for\n\
546 Ada tag names have suffixes indicating the type of entity:\n\
547 Entity type: Qualifier:\n\
548 ------------ ----------\n\
555 Thus, `M-x find-tag <RET> bidule/b <RET>' will go directly to the\n\
556 body of the package `bidule', while `M-x find-tag <RET> bidule <RET>'\n\
557 will just search for any tag `bidule'.";
560 static char *Asm_suffixes [] =
561 { "a", /* Unix assembler */
562 "asm", /* Microcontroller assembly */
563 "def", /* BSO/Tasking definition includes */
564 "inc", /* Microcontroller include files */
565 "ins", /* Microcontroller include files */
566 "s", "sa", /* Unix assembler */
567 "S", /* cpp-processed Unix assembler */
568 "src", /* BSO/Tasking C compiler output */
571 static char Asm_help [] =
572 "In assembler code, labels appearing at the beginning of a line,\n\
573 followed by a colon, are tags.";
576 /* Note that .c and .h can be considered C++, if the --c++ flag was
577 given, or if the `class' or `template' keywords are met inside the file.
578 That is why default_C_entries is called for these. */
579 static char *default_C_suffixes [] =
581 #if CTAGS /* C help for Ctags */
582 static char default_C_help [] =
583 "In C code, any C function is a tag. Use -t to tag typedefs.\n\
584 Use -T to tag definitions of `struct', `union' and `enum'.\n\
585 Use -d to tag `#define' macro definitions and `enum' constants.\n\
586 Use --globals to tag global variables.\n\
587 You can tag function declarations and external variables by\n\
588 using `--declarations', and struct members by using `--members'.";
589 #else /* C help for Etags */
590 static char default_C_help [] =
591 "In C code, any C function or typedef is a tag, and so are\n\
592 definitions of `struct', `union' and `enum'. `#define' macro\n\
593 definitions and `enum' constants are tags unless you specify\n\
594 `--no-defines'. Global variables are tags unless you specify\n\
595 `--no-globals' and so are struct members unless you specify\n\
596 `--no-members'. Use of `--no-globals', `--no-defines' and\n\
597 `--no-members' can make the tags table file much smaller.\n\
598 You can tag function declarations and external variables by\n\
599 using `--declarations'.";
600 #endif /* C help for Ctags and Etags */
602 static char *Cplusplus_suffixes [] =
603 { "C", "c++", "cc", "cpp", "cxx", "H", "h++", "hh", "hpp", "hxx",
604 "M", /* Objective C++ */
605 "pdb", /* Postscript with C syntax */
607 static char Cplusplus_help [] =
608 "In C++ code, all the tag constructs of C code are tagged. (Use\n\
609 --help --lang=c --lang=c++ for full help.)\n\
610 In addition to C tags, member functions are also recognized. Member\n\
611 variables are recognized unless you use the `--no-members' option.\n\
612 Tags for variables and functions in classes are named `CLASS::VARIABLE'\n\
613 and `CLASS::FUNCTION'. `operator' definitions have tag names like\n\
616 static char *Cjava_suffixes [] =
618 static char Cjava_help [] =
619 "In Java code, all the tags constructs of C and C++ code are\n\
620 tagged. (Use --help --lang=c --lang=c++ --lang=java for full help.)";
623 static char *Cobol_suffixes [] =
624 { "COB", "cob", NULL };
625 static char Cobol_help [] =
626 "In Cobol code, tags are paragraph names; that is, any word\n\
627 starting in column 8 and followed by a period.";
629 static char *Cstar_suffixes [] =
630 { "cs", "hs", NULL };
632 static char *Erlang_suffixes [] =
633 { "erl", "hrl", NULL };
634 static char Erlang_help [] =
635 "In Erlang code, the tags are the functions, records and macros\n\
636 defined in the file.";
638 char *Forth_suffixes [] =
639 { "fth", "tok", NULL };
640 static char Forth_help [] =
641 "In Forth code, tags are words defined by `:',\n\
642 constant, code, create, defer, value, variable, buffer:, field.";
644 static char *Fortran_suffixes [] =
645 { "F", "f", "f90", "for", NULL };
646 static char Fortran_help [] =
647 "In Fortran code, functions, subroutines and block data are tags.";
649 static char *HTML_suffixes [] =
650 { "htm", "html", "shtml", NULL };
651 static char HTML_help [] =
652 "In HTML input files, the tags are the `title' and the `h1', `h2',\n\
653 `h3' headers. Also, tags are `name=' in anchors and all\n\
654 occurrences of `id='.";
656 static char *Lisp_suffixes [] =
657 { "cl", "clisp", "el", "l", "lisp", "LSP", "lsp", "ml", NULL };
658 static char Lisp_help [] =
659 "In Lisp code, any function defined with `defun', any variable\n\
660 defined with `defvar' or `defconst', and in general the first\n\
661 argument of any expression that starts with `(def' in column zero\n\
664 static char *Lua_suffixes [] =
665 { "lua", "LUA", NULL };
666 static char Lua_help [] =
667 "In Lua scripts, all functions are tags.";
669 static char *Makefile_filenames [] =
670 { "Makefile", "makefile", "GNUMakefile", "Makefile.in", "Makefile.am", NULL};
671 static char Makefile_help [] =
672 "In makefiles, targets are tags; additionally, variables are tags\n\
673 unless you specify `--no-globals'.";
675 static char *Objc_suffixes [] =
676 { "lm", /* Objective lex file */
677 "m", /* Objective C file */
679 static char Objc_help [] =
680 "In Objective C code, tags include Objective C definitions for classes,\n\
681 class categories, methods and protocols. Tags for variables and\n\
682 functions in classes are named `CLASS::VARIABLE' and `CLASS::FUNCTION'.\n\
683 (Use --help --lang=c --lang=objc --lang=java for full help.)";
685 static char *Pascal_suffixes [] =
686 { "p", "pas", NULL };
687 static char Pascal_help [] =
688 "In Pascal code, the tags are the functions and procedures defined\n\
690 /* " // this is for working around an Emacs highlighting bug... */
692 static char *Perl_suffixes [] =
693 { "pl", "pm", NULL };
694 static char *Perl_interpreters [] =
695 { "perl", "@PERL@", NULL };
696 static char Perl_help [] =
697 "In Perl code, the tags are the packages, subroutines and variables\n\
698 defined by the `package', `sub', `my' and `local' keywords. Use\n\
699 `--globals' if you want to tag global variables. Tags for\n\
700 subroutines are named `PACKAGE::SUB'. The name for subroutines\n\
701 defined in the default package is `main::SUB'.";
703 static char *PHP_suffixes [] =
704 { "php", "php3", "php4", NULL };
705 static char PHP_help [] =
706 "In PHP code, tags are functions, classes and defines. Unless you use\n\
707 the `--no-members' option, vars are tags too.";
709 static char *plain_C_suffixes [] =
710 { "pc", /* Pro*C file */
713 static char *PS_suffixes [] =
714 { "ps", "psw", NULL }; /* .psw is for PSWrap */
715 static char PS_help [] =
716 "In PostScript code, the tags are the functions.";
718 static char *Prolog_suffixes [] =
720 static char Prolog_help [] =
721 "In Prolog code, tags are predicates and rules at the beginning of\n\
724 static char *Python_suffixes [] =
726 static char Python_help [] =
727 "In Python code, `def' or `class' at the beginning of a line\n\
730 /* Can't do the `SCM' or `scm' prefix with a version number. */
731 static char *Scheme_suffixes [] =
732 { "oak", "sch", "scheme", "SCM", "scm", "SM", "sm", "ss", "t", NULL };
733 static char Scheme_help [] =
734 "In Scheme code, tags include anything defined with `def' or with a\n\
735 construct whose name starts with `def'. They also include\n\
736 variables set with `set!' at top level in the file.";
738 static char *TeX_suffixes [] =
739 { "bib", "clo", "cls", "ltx", "sty", "TeX", "tex", NULL };
740 static char TeX_help [] =
741 "In LaTeX text, the argument of any of the commands `\\chapter',\n\
742 `\\section', `\\subsection', `\\subsubsection', `\\eqno', `\\label',\n\
743 `\\ref', `\\cite', `\\bibitem', `\\part', `\\appendix', `\\entry',\n\
744 `\\index', `\\def', `\\newcommand', `\\renewcommand',\n\
745 `\\newenvironment' or `\\renewenvironment' is a tag.\n\
747 Other commands can be specified by setting the environment variable\n\
748 `TEXTAGS' to a colon-separated list like, for example,\n\
749 TEXTAGS=\"mycommand:myothercommand\".";
752 static char *Texinfo_suffixes [] =
753 { "texi", "texinfo", "txi", NULL };
754 static char Texinfo_help [] =
755 "for texinfo files, lines starting with @node are tagged.";
757 static char *Yacc_suffixes [] =
758 { "y", "y++", "ym", "yxx", "yy", NULL }; /* .ym is Objective yacc file */
759 static char Yacc_help [] =
760 "In Bison or Yacc input files, each rule defines as a tag the\n\
761 nonterminal it constructs. The portions of the file that contain\n\
762 C code are parsed as C code (use --help --lang=c --lang=yacc\n\
765 static char auto_help [] =
766 "`auto' is not a real language, it indicates to use\n\
767 a default language for files base on file name suffix and file contents.";
769 static char none_help [] =
770 "`none' is not a real language, it indicates to only do\n\
771 regexp processing on files.";
773 static char no_lang_help [] =
774 "No detailed help available for this language.";
778 * Table of languages.
780 * It is ok for a given function to be listed under more than one
781 * name. I just didn't.
784 static language lang_names [] =
786 { "ada", Ada_help, Ada_funcs, Ada_suffixes },
787 { "asm", Asm_help, Asm_labels, Asm_suffixes },
788 { "c", default_C_help, default_C_entries, default_C_suffixes },
789 { "c++", Cplusplus_help, Cplusplus_entries, Cplusplus_suffixes },
790 { "c*", no_lang_help, Cstar_entries, Cstar_suffixes },
791 { "cobol", Cobol_help, Cobol_paragraphs, Cobol_suffixes },
792 { "erlang", Erlang_help, Erlang_functions, Erlang_suffixes },
793 { "forth", Forth_help, Forth_words, Forth_suffixes },
794 { "fortran", Fortran_help, Fortran_functions, Fortran_suffixes },
795 { "html", HTML_help, HTML_labels, HTML_suffixes },
796 { "java", Cjava_help, Cjava_entries, Cjava_suffixes },
797 { "lisp", Lisp_help, Lisp_functions, Lisp_suffixes },
798 { "lua", Lua_help, Lua_functions, Lua_suffixes },
799 { "makefile", Makefile_help,Makefile_targets,NULL,Makefile_filenames},
800 { "objc", Objc_help, plain_C_entries, Objc_suffixes },
801 { "pascal", Pascal_help, Pascal_functions, Pascal_suffixes },
802 { "perl",Perl_help,Perl_functions,Perl_suffixes,NULL,Perl_interpreters},
803 { "php", PHP_help, PHP_functions, PHP_suffixes },
804 { "postscript",PS_help, PS_functions, PS_suffixes },
805 { "proc", no_lang_help, plain_C_entries, plain_C_suffixes },
806 { "prolog", Prolog_help, Prolog_functions, Prolog_suffixes },
807 { "python", Python_help, Python_functions, Python_suffixes },
808 { "scheme", Scheme_help, Scheme_functions, Scheme_suffixes },
809 { "tex", TeX_help, TeX_commands, TeX_suffixes },
810 { "texinfo", Texinfo_help, Texinfo_nodes, Texinfo_suffixes },
811 { "yacc", Yacc_help,Yacc_entries,Yacc_suffixes,NULL,NULL,TRUE},
812 { "auto", auto_help }, /* default guessing scheme */
813 { "none", none_help, just_read_file }, /* regexp matching only */
814 { NULL } /* end of list */
819 print_language_names ()
824 puts ("\nThese are the currently supported languages, along with the\n\
825 default file names and dot suffixes:");
826 for (lang = lang_names; lang->name != NULL; lang++)
828 printf (" %-*s", 10, lang->name);
829 if (lang->filenames != NULL)
830 for (name = lang->filenames; *name != NULL; name++)
831 printf (" %s", *name);
832 if (lang->suffixes != NULL)
833 for (ext = lang->suffixes; *ext != NULL; ext++)
834 printf (" .%s", *ext);
837 puts ("where `auto' means use default language for files based on file\n\
838 name suffix, and `none' means only do regexp processing on files.\n\
839 If no language is specified and no matching suffix is found,\n\
840 the first line of the file is read for a sharp-bang (#!) sequence\n\
841 followed by the name of an interpreter. If no such sequence is found,\n\
842 Fortran is tried first; if no tags are found, C is tried next.\n\
843 When parsing any C file, a \"class\" or \"template\" keyword\n\
845 puts ("Compressed files are supported using gzip and bzip2.\n\
847 For detailed help on a given language use, for example,\n\
848 etags --help --lang=ada.");
852 # define EMACS_NAME "standalone"
855 # define VERSION "17.38.1.4"
858 # define E_VERSION EMACS_VERSION
860 # define E_VERSION VERSION
865 /* Makes it easier to update automatically. */
866 char emacs_copyright[] = "Copyright (C) 2008 Free Software Foundation, Inc.";
868 printf ("%s (%s %s)\n", (CTAGS) ? "ctags" : "etags", EMACS_NAME, E_VERSION);
869 puts (pot_etags_version);
870 puts (emacs_copyright);
871 puts ("This program is distributed under the terms in ETAGS.README");
876 #ifndef PRINT_UNDOCUMENTED_OPTIONS_HELP
877 # define PRINT_UNDOCUMENTED_OPTIONS_HELP FALSE
881 print_help (argbuffer)
884 bool help_for_lang = FALSE;
886 for (; argbuffer->arg_type != at_end; argbuffer++)
887 if (argbuffer->arg_type == at_language)
891 puts (argbuffer->lang->help);
892 help_for_lang = TRUE;
898 printf ("Usage: %s [options] [[regex-option ...] file-name] ...\n\
900 These are the options accepted by %s.\n", progname, progname);
902 puts ("WARNING: long option names do not work with this executable,\n\
903 as it is not linked with GNU getopt.");
905 puts ("You may use unambiguous abbreviations for the long option names.");
906 puts (" A - as file name means read names from stdin (one per line).\n\
907 Absolute names are stored in the output file as they are.\n\
908 Relative ones are stored relative to the output file's directory.\n");
910 puts ("-a, --append\n\
911 Append tag entries to existing tags file.");
913 puts ("--packages-only\n\
914 For Ada files, only generate tags for packages.");
917 puts ("-B, --backward-search\n\
918 Write the search commands for the tag entries using '?', the\n\
919 backward-search command instead of '/', the forward-search command.");
921 /* This option is mostly obsolete, because etags can now automatically
922 detect C++. Retained for backward compatibility and for debugging and
923 experimentation. In principle, we could want to tag as C++ even
924 before any "class" or "template" keyword.
926 Treat files whose name suffix defaults to C language as C++ files.");
929 puts ("--declarations\n\
930 In C and derived languages, create tags for function declarations,");
932 puts ("\tand create tags for extern variables if --globals is used.");
935 ("\tand create tags for extern variables unless --no-globals is used.");
938 puts ("-d, --defines\n\
939 Create tag entries for C #define constants and enum constants, too.");
941 puts ("-D, --no-defines\n\
942 Don't create tag entries for C #define constants and enum constants.\n\
943 This makes the tags file smaller.");
946 puts ("-i FILE, --include=FILE\n\
947 Include a note in tag file indicating that, when searching for\n\
948 a tag, one should also consult the tags file FILE after\n\
949 checking the current file.");
951 puts ("-l LANG, --language=LANG\n\
952 Force the following files to be considered as written in the\n\
953 named language up to the next --language=LANG option.");
957 Create tag entries for global variables in some languages.");
959 puts ("--no-globals\n\
960 Do not create tag entries for global variables in some\n\
961 languages. This makes the tags file smaller.");
963 if (PRINT_UNDOCUMENTED_OPTIONS_HELP)
964 puts ("--no-line-directive\n\
965 Ignore #line preprocessor directives in C and derived languages.");
969 Create tag entries for members of structures in some languages.");
971 puts ("--no-members\n\
972 Do not create tag entries for members of structures\n\
973 in some languages.");
975 puts ("-r REGEXP, --regex=REGEXP or --regex=@regexfile\n\
976 Make a tag for each line matching a regular expression pattern\n\
977 in the following files. {LANGUAGE}REGEXP uses REGEXP for LANGUAGE\n\
978 files only. REGEXFILE is a file containing one REGEXP per line.\n\
979 REGEXP takes the form /TAGREGEXP/TAGNAME/MODS, where TAGNAME/ is\n\
980 optional. The TAGREGEXP pattern is anchored (as if preceded by ^).");
981 puts (" If TAGNAME/ is present, the tags created are named.\n\
982 For example Tcl named tags can be created with:\n\
983 --regex=\"/proc[ \\t]+\\([^ \\t]+\\)/\\1/.\".\n\
984 MODS are optional one-letter modifiers: `i' means to ignore case,\n\
985 `m' means to allow multi-line matches, `s' implies `m' and\n\
986 causes dot to match any character, including newline.");
988 puts ("-R, --no-regex\n\
989 Don't create tags from regexps for the following files.");
991 puts ("-I, --ignore-indentation\n\
992 In C and C++ do not assume that a closing brace in the first\n\
993 column is the final brace of a function or structure definition.");
995 puts ("-o FILE, --output=FILE\n\
996 Write the tags to FILE.");
998 puts ("--parse-stdin=NAME\n\
999 Read from standard input and record tags as belonging to file NAME.");
1003 puts ("-t, --typedefs\n\
1004 Generate tag entries for C and Ada typedefs.");
1005 puts ("-T, --typedefs-and-c++\n\
1006 Generate tag entries for C typedefs, C struct/enum/union tags,\n\
1007 and C++ member functions.");
1011 puts ("-u, --update\n\
1012 Update the tag entries for the given files, leaving tag\n\
1013 entries for other files in place. Currently, this is\n\
1014 implemented by deleting the existing entries for the given\n\
1015 files and then rewriting the new entries at the end of the\n\
1016 tags file. It is often faster to simply rebuild the entire\n\
1017 tag file than to use this.");
1021 puts ("-v, --vgrind\n\
1022 Print on the standard output an index of items intended for\n\
1023 human consumption, similar to the output of vgrind. The index\n\
1024 is sorted, and gives the page number of each item.");
1026 if (PRINT_UNDOCUMENTED_OPTIONS_HELP)
1027 puts ("-w, --no-duplicates\n\
1028 Do not create duplicate tag entries, for compatibility with\n\
1029 traditional ctags.");
1031 if (PRINT_UNDOCUMENTED_OPTIONS_HELP)
1032 puts ("-w, --no-warn\n\
1033 Suppress warning messages about duplicate tag entries.");
1035 puts ("-x, --cxref\n\
1036 Like --vgrind, but in the style of cxref, rather than vgrind.\n\
1037 The output uses line numbers instead of page numbers, but\n\
1038 beyond that the differences are cosmetic; try both to see\n\
1042 puts ("-V, --version\n\
1043 Print the version of the program.\n\
1045 Print this help message.\n\
1046 Followed by one or more `--language' options prints detailed\n\
1047 help about tag generation for the specified languages.");
1049 print_language_names ();
1052 puts ("Report bugs to bug-gnu-emacs@gnu.org");
1054 exit (EXIT_SUCCESS);
1064 unsigned int nincluded_files;
1065 char **included_files;
1066 argument *argbuffer;
1067 int current_arg, file_count;
1068 linebuffer filename_lb;
1069 bool help_asked = FALSE;
1075 nincluded_files = 0;
1076 included_files = xnew (argc, char *);
1080 /* Allocate enough no matter what happens. Overkill, but each one
1082 argbuffer = xnew (argc, argument);
1085 * Always find typedefs and structure tags.
1086 * Also default to find macro constants, enum constants, struct
1087 * members and global variables. Do it for both etags and ctags.
1089 typedefs = typedefs_or_cplusplus = constantypedefs = TRUE;
1090 globals = members = TRUE;
1092 /* When the optstring begins with a '-' getopt_long does not rearrange the
1093 non-options arguments to be at the end, but leaves them alone. */
1094 optstring = concat (NO_LONG_OPTIONS ? "" : "-",
1095 "ac:Cf:Il:o:r:RSVhH",
1096 (CTAGS) ? "BxdtTuvw" : "Di:");
1098 while ((opt = getopt_long (argc, argv, optstring, longopts, NULL)) != EOF)
1102 /* If getopt returns 0, then it has already processed a
1103 long-named option. We should do nothing. */
1107 /* This means that a file name has been seen. Record it. */
1108 argbuffer[current_arg].arg_type = at_filename;
1109 argbuffer[current_arg].what = optarg;
1115 /* Parse standard input. Idea by Vivek <vivek@etla.org>. */
1116 argbuffer[current_arg].arg_type = at_stdin;
1117 argbuffer[current_arg].what = optarg;
1121 fatal ("cannot parse standard input more than once", (char *)NULL);
1122 parsing_stdin = TRUE;
1125 /* Common options. */
1126 case 'a': append_to_tagfile = TRUE; break;
1127 case 'C': cplusplus = TRUE; break;
1128 case 'f': /* for compatibility with old makefiles */
1132 error ("-o option may only be given once.", (char *)NULL);
1133 suggest_asking_for_help ();
1139 case 'S': /* for backward compatibility */
1140 ignoreindent = TRUE;
1144 language *lang = get_language_from_langname (optarg);
1147 argbuffer[current_arg].lang = lang;
1148 argbuffer[current_arg].arg_type = at_language;
1154 /* Backward compatibility: support obsolete --ignore-case-regexp. */
1155 optarg = concat (optarg, "i", ""); /* memory leak here */
1158 argbuffer[current_arg].arg_type = at_regexp;
1159 argbuffer[current_arg].what = optarg;
1163 argbuffer[current_arg].arg_type = at_regexp;
1164 argbuffer[current_arg].what = NULL;
1176 case 'D': constantypedefs = FALSE; break;
1177 case 'i': included_files[nincluded_files++] = optarg; break;
1179 /* Ctags options. */
1180 case 'B': searchar = '?'; break;
1181 case 'd': constantypedefs = TRUE; break;
1182 case 't': typedefs = TRUE; break;
1183 case 'T': typedefs = typedefs_or_cplusplus = TRUE; break;
1184 case 'u': update = TRUE; break;
1185 case 'v': vgrind_style = TRUE; /*FALLTHRU*/
1186 case 'x': cxref_style = TRUE; break;
1187 case 'w': no_warnings = TRUE; break;
1189 suggest_asking_for_help ();
1193 /* No more options. Store the rest of arguments. */
1194 for (; optind < argc; optind++)
1196 argbuffer[current_arg].arg_type = at_filename;
1197 argbuffer[current_arg].what = argv[optind];
1202 argbuffer[current_arg].arg_type = at_end;
1205 print_help (argbuffer);
1208 if (nincluded_files == 0 && file_count == 0)
1210 error ("no input files specified.", (char *)NULL);
1211 suggest_asking_for_help ();
1215 if (tagfile == NULL)
1216 tagfile = savestr (CTAGS ? "tags" : "TAGS");
1217 cwd = etags_getcwd (); /* the current working directory */
1218 if (cwd[strlen (cwd) - 1] != '/')
1221 cwd = concat (oldcwd, "/", "");
1225 /* Compute base directory for relative file names. */
1226 if (streq (tagfile, "-")
1227 || strneq (tagfile, "/dev/", 5))
1228 tagfiledir = cwd; /* relative file names are relative to cwd */
1231 canonicalize_filename (tagfile);
1232 tagfiledir = absolute_dirname (tagfile, cwd);
1235 init (); /* set up boolean "functions" */
1237 linebuffer_init (&lb);
1238 linebuffer_init (&filename_lb);
1239 linebuffer_init (&filebuf);
1240 linebuffer_init (&token_name);
1244 if (streq (tagfile, "-"))
1249 tagf = fopen (tagfile, append_to_tagfile ? "a" : "w");
1255 * Loop through files finding functions.
1257 for (i = 0; i < current_arg; i++)
1259 static language *lang; /* non-NULL if language is forced */
1262 switch (argbuffer[i].arg_type)
1265 lang = argbuffer[i].lang;
1268 analyse_regex (argbuffer[i].what);
1271 this_file = argbuffer[i].what;
1272 /* Input file named "-" means read file names from stdin
1273 (one per line) and use them. */
1274 if (streq (this_file, "-"))
1277 fatal ("cannot parse standard input AND read file names from it",
1279 while (readline_internal (&filename_lb, stdin) > 0)
1280 process_file_name (filename_lb.buffer, lang);
1283 process_file_name (this_file, lang);
1286 this_file = argbuffer[i].what;
1287 process_file (stdin, this_file, lang);
1299 free (filebuf.buffer);
1300 free (token_name.buffer);
1302 if (!CTAGS || cxref_style)
1304 /* Write the remaining tags to tagf (ETAGS) or stdout (CXREF). */
1305 put_entries (nodehead);
1306 free_tree (nodehead);
1312 /* Output file entries that have no tags. */
1313 for (fdp = fdhead; fdp != NULL; fdp = fdp->next)
1315 fprintf (tagf, "\f\n%s,0\n", fdp->taggedfname);
1317 while (nincluded_files-- > 0)
1318 fprintf (tagf, "\f\n%s,include\n", *included_files++);
1320 if (fclose (tagf) == EOF)
1324 exit (EXIT_SUCCESS);
1327 /* From here on, we are in (CTAGS && !cxref_style) */
1331 for (i = 0; i < current_arg; ++i)
1333 switch (argbuffer[i].arg_type)
1339 continue; /* the for loop */
1342 "mv %s OTAGS;fgrep -v '\t%s\t' OTAGS >%s;rm OTAGS",
1343 tagfile, argbuffer[i].what, tagfile);
1344 if (system (cmd) != EXIT_SUCCESS)
1345 fatal ("failed to execute shell command", (char *)NULL);
1347 append_to_tagfile = TRUE;
1350 tagf = fopen (tagfile, append_to_tagfile ? "a" : "w");
1353 put_entries (nodehead); /* write all the tags (CTAGS) */
1354 free_tree (nodehead);
1356 if (fclose (tagf) == EOF)
1360 if (append_to_tagfile || update)
1362 char cmd[2*BUFSIZ+20];
1363 /* Maybe these should be used:
1364 setenv ("LC_COLLATE", "C", 1);
1365 setenv ("LC_ALL", "C", 1); */
1366 sprintf (cmd, "sort -u -o %.*s %.*s", BUFSIZ, tagfile, BUFSIZ, tagfile);
1367 exit (system (cmd));
1369 return EXIT_SUCCESS;
1374 * Return a compressor given the file name. If EXTPTR is non-zero,
1375 * return a pointer into FILE where the compressor-specific
1376 * extension begins. If no compressor is found, NULL is returned
1377 * and EXTPTR is not significant.
1378 * Idea by Vladimir Alexiev <vladimir@cs.ualberta.ca> (1998)
1381 get_compressor_from_suffix (file, extptr)
1386 char *slash, *suffix;
1388 /* File has been processed by canonicalize_filename,
1389 so we don't need to consider backslashes on DOS_NT. */
1390 slash = etags_strrchr (file, '/');
1391 suffix = etags_strrchr (file, '.');
1392 if (suffix == NULL || suffix < slash)
1397 /* Let those poor souls who live with DOS 8+3 file name limits get
1398 some solace by treating foo.cgz as if it were foo.c.gz, etc.
1399 Only the first do loop is run if not MSDOS */
1402 for (compr = compressors; compr->suffix != NULL; compr++)
1403 if (streq (compr->suffix, suffix))
1406 break; /* do it only once: not really a loop */
1409 } while (*suffix != '\0');
1416 * Return a language given the name.
1419 get_language_from_langname (name)
1425 error ("empty language name", (char *)NULL);
1428 for (lang = lang_names; lang->name != NULL; lang++)
1429 if (streq (name, lang->name))
1431 error ("unknown language \"%s\"", name);
1439 * Return a language given the interpreter name.
1442 get_language_from_interpreter (interpreter)
1448 if (interpreter == NULL)
1450 for (lang = lang_names; lang->name != NULL; lang++)
1451 if (lang->interpreters != NULL)
1452 for (iname = lang->interpreters; *iname != NULL; iname++)
1453 if (streq (*iname, interpreter))
1462 * Return a language given the file name.
1465 get_language_from_filename (file, case_sensitive)
1467 bool case_sensitive;
1470 char **name, **ext, *suffix;
1472 /* Try whole file name first. */
1473 for (lang = lang_names; lang->name != NULL; lang++)
1474 if (lang->filenames != NULL)
1475 for (name = lang->filenames; *name != NULL; name++)
1476 if ((case_sensitive)
1477 ? streq (*name, file)
1478 : strcaseeq (*name, file))
1481 /* If not found, try suffix after last dot. */
1482 suffix = etags_strrchr (file, '.');
1486 for (lang = lang_names; lang->name != NULL; lang++)
1487 if (lang->suffixes != NULL)
1488 for (ext = lang->suffixes; *ext != NULL; ext++)
1489 if ((case_sensitive)
1490 ? streq (*ext, suffix)
1491 : strcaseeq (*ext, suffix))
1498 * This routine is called on each file argument.
1501 process_file_name (file, lang)
1505 struct stat stat_buf;
1509 char *compressed_name, *uncompressed_name;
1510 char *ext, *real_name;
1513 canonicalize_filename (file);
1514 if (streq (file, tagfile) && !streq (tagfile, "-"))
1516 error ("skipping inclusion of %s in self.", file);
1519 if ((compr = get_compressor_from_suffix (file, &ext)) == NULL)
1521 compressed_name = NULL;
1522 real_name = uncompressed_name = savestr (file);
1526 real_name = compressed_name = savestr (file);
1527 uncompressed_name = savenstr (file, ext - file);
1530 /* If the canonicalized uncompressed name
1531 has already been dealt with, skip it silently. */
1532 for (fdp = fdhead; fdp != NULL; fdp = fdp->next)
1534 assert (fdp->infname != NULL);
1535 if (streq (uncompressed_name, fdp->infname))
1539 if (stat (real_name, &stat_buf) != 0)
1541 /* Reset real_name and try with a different name. */
1543 if (compressed_name != NULL) /* try with the given suffix */
1545 if (stat (uncompressed_name, &stat_buf) == 0)
1546 real_name = uncompressed_name;
1548 else /* try all possible suffixes */
1550 for (compr = compressors; compr->suffix != NULL; compr++)
1552 compressed_name = concat (file, ".", compr->suffix);
1553 if (stat (compressed_name, &stat_buf) != 0)
1555 free (compressed_name);
1556 compressed_name = NULL;
1560 real_name = compressed_name;
1565 if (real_name == NULL)
1570 } /* try with a different name */
1572 if (!S_ISREG (stat_buf.st_mode))
1574 error ("skipping %s: it is not a regular file.", real_name);
1577 if (real_name == compressed_name)
1579 char *cmd = concat (compr->command, " ", real_name);
1580 inf = (FILE *) popen (cmd, "r");
1584 inf = fopen (real_name, "r");
1591 process_file (inf, uncompressed_name, lang);
1593 if (real_name == compressed_name)
1594 retval = pclose (inf);
1596 retval = fclose (inf);
1601 free (compressed_name);
1602 free (uncompressed_name);
1609 process_file (fh, fn, lang)
1614 static const fdesc emptyfdesc;
1617 /* Create a new input file description entry. */
1618 fdp = xnew (1, fdesc);
1621 fdp->infname = savestr (fn);
1623 fdp->infabsname = absolute_filename (fn, cwd);
1624 fdp->infabsdir = absolute_dirname (fn, cwd);
1625 if (filename_is_absolute (fn))
1627 /* An absolute file name. Canonicalize it. */
1628 fdp->taggedfname = absolute_filename (fn, NULL);
1632 /* A file name relative to cwd. Make it relative
1633 to the directory of the tags file. */
1634 fdp->taggedfname = relative_filename (fn, tagfiledir);
1636 fdp->usecharno = TRUE; /* use char position when making tags */
1638 fdp->written = FALSE; /* not written on tags file yet */
1641 curfdp = fdhead; /* the current file description */
1645 /* If not Ctags, and if this is not metasource and if it contained no #line
1646 directives, we can write the tags and free all nodes pointing to
1649 && curfdp->usecharno /* no #line directives in this file */
1650 && !curfdp->lang->metasource)
1654 /* Look for the head of the sublist relative to this file. See add_node
1655 for the structure of the node tree. */
1657 for (np = nodehead; np != NULL; prev = np, np = np->left)
1658 if (np->fdp == curfdp)
1661 /* If we generated tags for this file, write and delete them. */
1664 /* This is the head of the last sublist, if any. The following
1665 instructions depend on this being true. */
1666 assert (np->left == NULL);
1668 assert (fdhead == curfdp);
1669 assert (last_node->fdp == curfdp);
1670 put_entries (np); /* write tags for file curfdp->taggedfname */
1671 free_tree (np); /* remove the written nodes */
1673 nodehead = NULL; /* no nodes left */
1675 prev->left = NULL; /* delete the pointer to the sublist */
1681 * This routine sets up the boolean pseudo-functions which work
1682 * by setting boolean flags dependent upon the corresponding character.
1683 * Every char which is NOT in that string is not a white char. Therefore,
1684 * all of the array "_wht" is set to FALSE, and then the elements
1685 * subscripted by the chars in "white" are set to TRUE. Thus "_wht"
1686 * of a char is TRUE if it is the string "white", else FALSE.
1694 for (i = 0; i < CHARS; i++)
1695 iswhite(i) = notinname(i) = begtoken(i) = intoken(i) = endtoken(i) = FALSE;
1696 for (sp = white; *sp != '\0'; sp++) iswhite (*sp) = TRUE;
1697 for (sp = nonam; *sp != '\0'; sp++) notinname (*sp) = TRUE;
1698 notinname('\0') = notinname('\n');
1699 for (sp = begtk; *sp != '\0'; sp++) begtoken (*sp) = TRUE;
1700 begtoken('\0') = begtoken('\n');
1701 for (sp = midtk; *sp != '\0'; sp++) intoken (*sp) = TRUE;
1702 intoken('\0') = intoken('\n');
1703 for (sp = endtk; *sp != '\0'; sp++) endtoken (*sp) = TRUE;
1704 endtoken('\0') = endtoken('\n');
1708 * This routine opens the specified file and calls the function
1709 * which finds the function and type definitions.
1716 language *lang = curfdp->lang;
1717 Lang_function *parser = NULL;
1719 /* If user specified a language, use it. */
1720 if (lang != NULL && lang->function != NULL)
1722 parser = lang->function;
1725 /* Else try to guess the language given the file name. */
1728 lang = get_language_from_filename (curfdp->infname, TRUE);
1729 if (lang != NULL && lang->function != NULL)
1731 curfdp->lang = lang;
1732 parser = lang->function;
1736 /* Else look for sharp-bang as the first two characters. */
1738 && readline_internal (&lb, inf) > 0
1740 && lb.buffer[0] == '#'
1741 && lb.buffer[1] == '!')
1745 /* Set lp to point at the first char after the last slash in the
1746 line or, if no slashes, at the first nonblank. Then set cp to
1747 the first successive blank and terminate the string. */
1748 lp = etags_strrchr (lb.buffer+2, '/');
1752 lp = skip_spaces (lb.buffer + 2);
1753 cp = skip_non_spaces (lp);
1756 if (strlen (lp) > 0)
1758 lang = get_language_from_interpreter (lp);
1759 if (lang != NULL && lang->function != NULL)
1761 curfdp->lang = lang;
1762 parser = lang->function;
1767 /* We rewind here, even if inf may be a pipe. We fail if the
1768 length of the first line is longer than the pipe block size,
1769 which is unlikely. */
1772 /* Else try to guess the language given the case insensitive file name. */
1775 lang = get_language_from_filename (curfdp->infname, FALSE);
1776 if (lang != NULL && lang->function != NULL)
1778 curfdp->lang = lang;
1779 parser = lang->function;
1783 /* Else try Fortran or C. */
1786 node *old_last_node = last_node;
1788 curfdp->lang = get_language_from_langname ("fortran");
1791 if (old_last_node == last_node)
1792 /* No Fortran entries found. Try C. */
1794 /* We do not tag if rewind fails.
1795 Only the file name will be recorded in the tags file. */
1797 curfdp->lang = get_language_from_langname (cplusplus ? "c++" : "c");
1803 if (!no_line_directive
1804 && curfdp->lang != NULL && curfdp->lang->metasource)
1805 /* It may be that this is a bingo.y file, and we already parsed a bingo.c
1806 file, or anyway we parsed a file that is automatically generated from
1807 this one. If this is the case, the bingo.c file contained #line
1808 directives that generated tags pointing to this file. Let's delete
1809 them all before parsing this file, which is the real source. */
1811 fdesc **fdpp = &fdhead;
1812 while (*fdpp != NULL)
1814 && streq ((*fdpp)->taggedfname, curfdp->taggedfname))
1815 /* We found one of those! We must delete both the file description
1816 and all tags referring to it. */
1818 fdesc *badfdp = *fdpp;
1820 /* Delete the tags referring to badfdp->taggedfname
1821 that were obtained from badfdp->infname. */
1822 invalidate_nodes (badfdp, &nodehead);
1824 *fdpp = badfdp->next; /* remove the bad description from the list */
1825 free_fdesc (badfdp);
1828 fdpp = &(*fdpp)->next; /* advance the list pointer */
1831 assert (parser != NULL);
1833 /* Generic initialisations before reading from file. */
1834 linebuffer_setlen (&filebuf, 0); /* reset the file buffer */
1836 /* Generic initialisations before parsing file with readline. */
1837 lineno = 0; /* reset global line number */
1838 charno = 0; /* reset global char number */
1839 linecharno = 0; /* reset global char number of line start */
1843 regex_tag_multiline ();
1848 * Check whether an implicitly named tag should be created,
1849 * then call `pfnote'.
1850 * NAME is a string that is internally copied by this function.
1852 * TAGS format specification
1853 * Idea by Sam Kendall <kendall@mv.mv.com> (1997)
1854 * The following is explained in some more detail in etc/ETAGS.EBNF.
1856 * make_tag creates tags with "implicit tag names" (unnamed tags)
1857 * if the following are all true, assuming NONAM=" \f\t\n\r()=,;":
1858 * 1. NAME does not contain any of the characters in NONAM;
1859 * 2. LINESTART contains name as either a rightmost, or rightmost but
1860 * one character, substring;
1861 * 3. the character, if any, immediately before NAME in LINESTART must
1862 * be a character in NONAM;
1863 * 4. the character, if any, immediately after NAME in LINESTART must
1864 * also be a character in NONAM.
1866 * The implementation uses the notinname() macro, which recognises the
1867 * characters stored in the string `nonam'.
1868 * etags.el needs to use the same characters that are in NONAM.
1871 make_tag (name, namelen, is_func, linestart, linelen, lno, cno)
1872 char *name; /* tag name, or NULL if unnamed */
1873 int namelen; /* tag length */
1874 bool is_func; /* tag is a function */
1875 char *linestart; /* start of the line where tag is */
1876 int linelen; /* length of the line where tag is */
1877 int lno; /* line number */
1878 long cno; /* character number */
1880 bool named = (name != NULL && namelen > 0);
1882 if (!CTAGS && named) /* maybe set named to false */
1883 /* Let's try to make an implicit tag name, that is, create an unnamed tag
1884 such that etags.el can guess a name from it. */
1887 register char *cp = name;
1889 for (i = 0; i < namelen; i++)
1890 if (notinname (*cp++))
1892 if (i == namelen) /* rule #1 */
1894 cp = linestart + linelen - namelen;
1895 if (notinname (linestart[linelen-1]))
1896 cp -= 1; /* rule #4 */
1897 if (cp >= linestart /* rule #2 */
1899 || notinname (cp[-1])) /* rule #3 */
1900 && strneq (name, cp, namelen)) /* rule #2 */
1901 named = FALSE; /* use implicit tag name */
1906 name = savenstr (name, namelen);
1909 pfnote (name, is_func, linestart, linelen, lno, cno);
1914 pfnote (name, is_func, linestart, linelen, lno, cno)
1915 char *name; /* tag name, or NULL if unnamed */
1916 bool is_func; /* tag is a function */
1917 char *linestart; /* start of the line where tag is */
1918 int linelen; /* length of the line where tag is */
1919 int lno; /* line number */
1920 long cno; /* character number */
1924 assert (name == NULL || name[0] != '\0');
1925 if (CTAGS && name == NULL)
1928 np = xnew (1, node);
1930 /* If ctags mode, change name "main" to M<thisfilename>. */
1931 if (CTAGS && !cxref_style && streq (name, "main"))
1933 register char *fp = etags_strrchr (curfdp->taggedfname, '/');
1934 np->name = concat ("M", fp == NULL ? curfdp->taggedfname : fp + 1, "");
1935 fp = etags_strrchr (np->name, '.');
1936 if (fp != NULL && fp[1] != '\0' && fp[2] == '\0')
1942 np->been_warned = FALSE;
1944 np->is_func = is_func;
1946 if (np->fdp->usecharno)
1947 /* Our char numbers are 0-base, because of C language tradition?
1948 ctags compatibility? old versions compatibility? I don't know.
1949 Anyway, since emacs's are 1-base we expect etags.el to take care
1950 of the difference. If we wanted to have 1-based numbers, we would
1951 uncomment the +1 below. */
1952 np->cno = cno /* + 1 */ ;
1954 np->cno = invalidcharno;
1955 np->left = np->right = NULL;
1956 if (CTAGS && !cxref_style)
1958 if (strlen (linestart) < 50)
1959 np->regex = concat (linestart, "$", "");
1961 np->regex = savenstr (linestart, 50);
1964 np->regex = savenstr (linestart, linelen);
1966 add_node (np, &nodehead);
1971 * recurse on left children, iterate on right children.
1979 register node *node_right = np->right;
1980 free_tree (np->left);
1990 * delete a file description
1994 register fdesc *fdp;
1996 free (fdp->infname);
1997 free (fdp->infabsname);
1998 free (fdp->infabsdir);
1999 free (fdp->taggedfname);
2006 * Adds a node to the tree of nodes. In etags mode, sort by file
2007 * name. In ctags mode, sort by tag name. Make no attempt at
2010 * add_node is the only function allowed to add nodes, so it can
2014 add_node (np, cur_node_p)
2015 node *np, **cur_node_p;
2018 register node *cur_node = *cur_node_p;
2020 if (cur_node == NULL)
2030 /* For each file name, tags are in a linked sublist on the right
2031 pointer. The first tags of different files are a linked list
2032 on the left pointer. last_node points to the end of the last
2034 if (last_node != NULL && last_node->fdp == np->fdp)
2036 /* Let's use the same sublist as the last added node. */
2037 assert (last_node->right == NULL);
2038 last_node->right = np;
2041 else if (cur_node->fdp == np->fdp)
2043 /* Scanning the list we found the head of a sublist which is
2044 good for us. Let's scan this sublist. */
2045 add_node (np, &cur_node->right);
2048 /* The head of this sublist is not good for us. Let's try the
2050 add_node (np, &cur_node->left);
2051 } /* if ETAGS mode */
2056 dif = strcmp (np->name, cur_node->name);
2059 * If this tag name matches an existing one, then
2060 * do not add the node, but maybe print a warning.
2062 if (no_duplicates && !dif)
2064 if (np->fdp == cur_node->fdp)
2068 fprintf (stderr, "Duplicate entry in file %s, line %d: %s\n",
2069 np->fdp->infname, lineno, np->name);
2070 fprintf (stderr, "Second entry ignored\n");
2073 else if (!cur_node->been_warned && !no_warnings)
2077 "Duplicate entry in files %s and %s: %s (Warning only)\n",
2078 np->fdp->infname, cur_node->fdp->infname, np->name);
2079 cur_node->been_warned = TRUE;
2084 /* Actually add the node */
2085 add_node (np, dif < 0 ? &cur_node->left : &cur_node->right);
2086 } /* if CTAGS mode */
2090 * invalidate_nodes ()
2091 * Scan the node tree and invalidate all nodes pointing to the
2092 * given file description (CTAGS case) or free them (ETAGS case).
2095 invalidate_nodes (badfdp, npp)
2106 if (np->left != NULL)
2107 invalidate_nodes (badfdp, &np->left);
2108 if (np->fdp == badfdp)
2110 if (np->right != NULL)
2111 invalidate_nodes (badfdp, &np->right);
2115 assert (np->fdp != NULL);
2116 if (np->fdp == badfdp)
2118 *npp = np->left; /* detach the sublist from the list */
2119 np->left = NULL; /* isolate it */
2120 free_tree (np); /* free it */
2121 invalidate_nodes (badfdp, npp);
2124 invalidate_nodes (badfdp, &np->left);
2129 static int total_size_of_entries __P((node *));
2130 static int number_len __P((long));
2132 /* Length of a non-negative number's decimal representation. */
2138 while ((num /= 10) > 0)
2144 * Return total number of characters that put_entries will output for
2145 * the nodes in the linked list at the right of the specified node.
2146 * This count is irrelevant with etags.el since emacs 19.34 at least,
2147 * but is still supplied for backward compatibility.
2150 total_size_of_entries (np)
2153 register int total = 0;
2155 for (; np != NULL; np = np->right)
2158 total += strlen (np->regex) + 1; /* pat\177 */
2159 if (np->name != NULL)
2160 total += strlen (np->name) + 1; /* name\001 */
2161 total += number_len ((long) np->lno) + 1; /* lno, */
2162 if (np->cno != invalidcharno) /* cno */
2163 total += number_len (np->cno);
2164 total += 1; /* newline */
2175 static fdesc *fdp = NULL;
2180 /* Output subentries that precede this one */
2182 put_entries (np->left);
2184 /* Output this entry */
2193 fprintf (tagf, "\f\n%s,%d\n",
2194 fdp->taggedfname, total_size_of_entries (np));
2195 fdp->written = TRUE;
2197 fputs (np->regex, tagf);
2198 fputc ('\177', tagf);
2199 if (np->name != NULL)
2201 fputs (np->name, tagf);
2202 fputc ('\001', tagf);
2204 fprintf (tagf, "%d,", np->lno);
2205 if (np->cno != invalidcharno)
2206 fprintf (tagf, "%ld", np->cno);
2212 if (np->name == NULL)
2213 error ("internal error: NULL name in ctags mode.", (char *)NULL);
2218 fprintf (stdout, "%s %s %d\n",
2219 np->name, np->fdp->taggedfname, (np->lno + 63) / 64);
2221 fprintf (stdout, "%-16s %3d %-16s %s\n",
2222 np->name, np->lno, np->fdp->taggedfname, np->regex);
2226 fprintf (tagf, "%s\t%s\t", np->name, np->fdp->taggedfname);
2229 { /* function or #define macro with args */
2230 putc (searchar, tagf);
2233 for (sp = np->regex; *sp; sp++)
2235 if (*sp == '\\' || *sp == searchar)
2239 putc (searchar, tagf);
2242 { /* anything else; text pattern inadequate */
2243 fprintf (tagf, "%d", np->lno);
2248 } /* if this node contains a valid tag */
2250 /* Output subentries that follow this one */
2251 put_entries (np->right);
2253 put_entries (np->left);
2258 #define C_EXT 0x00fff /* C extensions */
2259 #define C_PLAIN 0x00000 /* C */
2260 #define C_PLPL 0x00001 /* C++ */
2261 #define C_STAR 0x00003 /* C* */
2262 #define C_JAVA 0x00005 /* JAVA */
2263 #define C_AUTO 0x01000 /* C, but switch to C++ if `class' is met */
2264 #define YACC 0x10000 /* yacc file */
2267 * The C symbol tables.
2272 st_C_objprot, st_C_objimpl, st_C_objend,
2274 st_C_ignore, st_C_attribute,
2277 st_C_class, st_C_template,
2278 st_C_struct, st_C_extern, st_C_enum, st_C_define, st_C_typedef
2281 static unsigned int hash __P((const char *, unsigned int));
2282 static struct C_stab_entry * in_word_set __P((const char *, unsigned int));
2283 static enum sym_type C_symtype __P((char *, int, int));
2285 /* Feed stuff between (but not including) %[ and %] lines to:
2291 struct C_stab_entry { char *name; int c_ext; enum sym_type type; }
2295 while, 0, st_C_ignore
2296 switch, 0, st_C_ignore
2297 return, 0, st_C_ignore
2298 __attribute__, 0, st_C_attribute
2299 GTY, 0, st_C_attribute
2300 @interface, 0, st_C_objprot
2301 @protocol, 0, st_C_objprot
2302 @implementation, 0, st_C_objimpl
2303 @end, 0, st_C_objend
2304 import, (C_JAVA & ~C_PLPL), st_C_ignore
2305 package, (C_JAVA & ~C_PLPL), st_C_ignore
2306 friend, C_PLPL, st_C_ignore
2307 extends, (C_JAVA & ~C_PLPL), st_C_javastruct
2308 implements, (C_JAVA & ~C_PLPL), st_C_javastruct
2309 interface, (C_JAVA & ~C_PLPL), st_C_struct
2310 class, 0, st_C_class
2311 namespace, C_PLPL, st_C_struct
2312 domain, C_STAR, st_C_struct
2313 union, 0, st_C_struct
2314 struct, 0, st_C_struct
2315 extern, 0, st_C_extern
2317 typedef, 0, st_C_typedef
2318 define, 0, st_C_define
2319 undef, 0, st_C_define
2320 operator, C_PLPL, st_C_operator
2321 template, 0, st_C_template
2322 # DEFUN used in emacs, the next three used in glibc (SYSCALL only for mach).
2323 DEFUN, 0, st_C_gnumacro
2324 SYSCALL, 0, st_C_gnumacro
2325 ENTRY, 0, st_C_gnumacro
2326 PSEUDO, 0, st_C_gnumacro
2327 # These are defined inside C functions, so currently they are not met.
2328 # EXFUN used in glibc, DEFVAR_* in emacs.
2329 #EXFUN, 0, st_C_gnumacro
2330 #DEFVAR_, 0, st_C_gnumacro
2332 and replace lines between %< and %> with its output, then:
2333 - remove the #if characterset check
2334 - make in_word_set static and not inline. */
2336 /* C code produced by gperf version 3.0.1 */
2337 /* Command-line: gperf -m 5 */
2338 /* Computed positions: -k'2-3' */
2340 struct C_stab_entry { char *name; int c_ext; enum sym_type type; };
2341 /* maximum key range = 33, duplicates = 0 */
2352 register const char *str;
2353 register unsigned int len;
2355 static unsigned char asso_values[] =
2357 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2358 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2359 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2360 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2361 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2362 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2363 35, 35, 35, 35, 35, 35, 35, 35, 35, 3,
2364 26, 35, 35, 35, 35, 35, 35, 35, 27, 35,
2365 35, 35, 35, 24, 0, 35, 35, 35, 35, 0,
2366 35, 35, 35, 35, 35, 1, 35, 16, 35, 6,
2367 23, 0, 0, 35, 22, 0, 35, 35, 5, 0,
2368 0, 15, 1, 35, 6, 35, 8, 19, 35, 16,
2369 4, 5, 35, 35, 35, 35, 35, 35, 35, 35,
2370 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2371 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2372 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2373 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2374 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2375 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2376 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2377 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2378 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2379 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2380 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2381 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2382 35, 35, 35, 35, 35, 35
2384 register int hval = len;
2389 hval += asso_values[(unsigned char)str[2]];
2392 hval += asso_values[(unsigned char)str[1]];
2398 static struct C_stab_entry *
2399 in_word_set (str, len)
2400 register const char *str;
2401 register unsigned int len;
2405 TOTAL_KEYWORDS = 33,
2406 MIN_WORD_LENGTH = 2,
2407 MAX_WORD_LENGTH = 15,
2412 static struct C_stab_entry wordlist[] =
2415 {"if", 0, st_C_ignore},
2416 {"GTY", 0, st_C_attribute},
2417 {"@end", 0, st_C_objend},
2418 {"union", 0, st_C_struct},
2419 {"define", 0, st_C_define},
2420 {"import", (C_JAVA & ~C_PLPL), st_C_ignore},
2421 {"template", 0, st_C_template},
2422 {"operator", C_PLPL, st_C_operator},
2423 {"@interface", 0, st_C_objprot},
2424 {"implements", (C_JAVA & ~C_PLPL), st_C_javastruct},
2425 {"friend", C_PLPL, st_C_ignore},
2426 {"typedef", 0, st_C_typedef},
2427 {"return", 0, st_C_ignore},
2428 {"@implementation",0, st_C_objimpl},
2429 {"@protocol", 0, st_C_objprot},
2430 {"interface", (C_JAVA & ~C_PLPL), st_C_struct},
2431 {"extern", 0, st_C_extern},
2432 {"extends", (C_JAVA & ~C_PLPL), st_C_javastruct},
2433 {"struct", 0, st_C_struct},
2434 {"domain", C_STAR, st_C_struct},
2435 {"switch", 0, st_C_ignore},
2436 {"enum", 0, st_C_enum},
2437 {"for", 0, st_C_ignore},
2438 {"namespace", C_PLPL, st_C_struct},
2439 {"class", 0, st_C_class},
2440 {"while", 0, st_C_ignore},
2441 {"undef", 0, st_C_define},
2442 {"package", (C_JAVA & ~C_PLPL), st_C_ignore},
2443 {"__attribute__", 0, st_C_attribute},
2444 {"SYSCALL", 0, st_C_gnumacro},
2445 {"ENTRY", 0, st_C_gnumacro},
2446 {"PSEUDO", 0, st_C_gnumacro},
2447 {"DEFUN", 0, st_C_gnumacro}
2450 if (len <= MAX_WORD_LENGTH && len >= MIN_WORD_LENGTH)
2452 register int key = hash (str, len);
2454 if (key <= MAX_HASH_VALUE && key >= 0)
2456 register const char *s = wordlist[key].name;
2458 if (*str == *s && !strncmp (str + 1, s + 1, len - 1) && s[len] == '\0')
2459 return &wordlist[key];
2466 static enum sym_type
2467 C_symtype (str, len, c_ext)
2472 register struct C_stab_entry *se = in_word_set (str, len);
2474 if (se == NULL || (se->c_ext && !(c_ext & se->c_ext)))
2481 * Ignoring __attribute__ ((list))
2483 static bool inattribute; /* looking at an __attribute__ construct */
2486 * C functions and variables are recognized using a simple
2487 * finite automaton. fvdef is its state variable.
2491 fvnone, /* nothing seen */
2492 fdefunkey, /* Emacs DEFUN keyword seen */
2493 fdefunname, /* Emacs DEFUN name seen */
2494 foperator, /* func: operator keyword seen (cplpl) */
2495 fvnameseen, /* function or variable name seen */
2496 fstartlist, /* func: just after open parenthesis */
2497 finlist, /* func: in parameter list */
2498 flistseen, /* func: after parameter list */
2499 fignore, /* func: before open brace */
2500 vignore /* var-like: ignore until ';' */
2503 static bool fvextern; /* func or var: extern keyword seen; */
2506 * typedefs are recognized using a simple finite automaton.
2507 * typdef is its state variable.
2511 tnone, /* nothing seen */
2512 tkeyseen, /* typedef keyword seen */
2513 ttypeseen, /* defined type seen */
2514 tinbody, /* inside typedef body */
2515 tend, /* just before typedef tag */
2516 tignore /* junk after typedef tag */
2520 * struct-like structures (enum, struct and union) are recognized
2521 * using another simple finite automaton. `structdef' is its state
2526 snone, /* nothing seen yet,
2527 or in struct body if bracelev > 0 */
2528 skeyseen, /* struct-like keyword seen */
2529 stagseen, /* struct-like tag seen */
2530 scolonseen /* colon seen after struct-like tag */
2534 * When objdef is different from onone, objtag is the name of the class.
2536 static char *objtag = "<uninited>";
2539 * Yet another little state machine to deal with preprocessor lines.
2543 dnone, /* nothing seen */
2544 dsharpseen, /* '#' seen as first char on line */
2545 ddefineseen, /* '#' and 'define' seen */
2546 dignorerest /* ignore rest of line */
2550 * State machine for Objective C protocols and implementations.
2551 * Idea by Tom R.Hageman <tom@basil.icce.rug.nl> (1995)
2555 onone, /* nothing seen */
2556 oprotocol, /* @interface or @protocol seen */
2557 oimplementation, /* @implementations seen */
2558 otagseen, /* class name seen */
2559 oparenseen, /* parenthesis before category seen */
2560 ocatseen, /* category name seen */
2561 oinbody, /* in @implementation body */
2562 omethodsign, /* in @implementation body, after +/- */
2563 omethodtag, /* after method name */
2564 omethodcolon, /* after method colon */
2565 omethodparm, /* after method parameter */
2566 oignore /* wait for @end */
2571 * Use this structure to keep info about the token read, and how it
2572 * should be tagged. Used by the make_C_tag function to build a tag.
2576 char *line; /* string containing the token */
2577 int offset; /* where the token starts in LINE */
2578 int length; /* token length */
2580 The previous members can be used to pass strings around for generic
2581 purposes. The following ones specifically refer to creating tags. In this
2582 case the token contained here is the pattern that will be used to create a
2585 bool valid; /* do not create a tag; the token should be
2586 invalidated whenever a state machine is
2587 reset prematurely */
2588 bool named; /* create a named tag */
2589 int lineno; /* source line number of tag */
2590 long linepos; /* source char number of tag */
2591 } token; /* latest token read */
2594 * Variables and functions for dealing with nested structures.
2595 * Idea by Mykola Dzyuba <mdzyuba@yahoo.com> (2001)
2597 static void pushclass_above __P((int, char *, int));
2598 static void popclass_above __P((int));
2599 static void write_classname __P((linebuffer *, char *qualifier));
2602 char **cname; /* nested class names */
2603 int *bracelev; /* nested class brace level */
2604 int nl; /* class nesting level (elements used) */
2605 int size; /* length of the array */
2606 } cstack; /* stack for nested declaration tags */
2607 /* Current struct nesting depth (namespace, class, struct, union, enum). */
2608 #define nestlev (cstack.nl)
2609 /* After struct keyword or in struct body, not inside a nested function. */
2610 #define instruct (structdef == snone && nestlev > 0 \
2611 && bracelev == cstack.bracelev[nestlev-1] + 1)
2614 pushclass_above (bracelev, str, len)
2621 popclass_above (bracelev);
2623 if (nl >= cstack.size)
2625 int size = cstack.size *= 2;
2626 xrnew (cstack.cname, size, char *);
2627 xrnew (cstack.bracelev, size, int);
2629 assert (nl == 0 || cstack.bracelev[nl-1] < bracelev);
2630 cstack.cname[nl] = (str == NULL) ? NULL : savenstr (str, len);
2631 cstack.bracelev[nl] = bracelev;
2636 popclass_above (bracelev)
2641 for (nl = cstack.nl - 1;
2642 nl >= 0 && cstack.bracelev[nl] >= bracelev;
2645 free (cstack.cname[nl]);
2651 write_classname (cn, qualifier)
2656 int qlen = strlen (qualifier);
2658 if (cstack.nl == 0 || cstack.cname[0] == NULL)
2662 cn->buffer[0] = '\0';
2666 len = strlen (cstack.cname[0]);
2667 linebuffer_setlen (cn, len);
2668 strcpy (cn->buffer, cstack.cname[0]);
2670 for (i = 1; i < cstack.nl; i++)
2675 s = cstack.cname[i];
2680 linebuffer_setlen (cn, len);
2681 strncat (cn->buffer, qualifier, qlen);
2682 strncat (cn->buffer, s, slen);
2687 static bool consider_token __P((char *, int, int, int *, int, int, bool *));
2688 static void make_C_tag __P((bool));
2692 * checks to see if the current token is at the start of a
2693 * function or variable, or corresponds to a typedef, or
2694 * is a struct/union/enum tag, or #define, or an enum constant.
2696 * *IS_FUNC gets TRUE if the token is a function or #define macro
2697 * with args. C_EXTP points to which language we are looking at.
2708 consider_token (str, len, c, c_extp, bracelev, parlev, is_func_or_var)
2709 register char *str; /* IN: token pointer */
2710 register int len; /* IN: token length */
2711 register int c; /* IN: first char after the token */
2712 int *c_extp; /* IN, OUT: C extensions mask */
2713 int bracelev; /* IN: brace level */
2714 int parlev; /* IN: parenthesis level */
2715 bool *is_func_or_var; /* OUT: function or variable found */
2717 /* When structdef is stagseen, scolonseen, or snone with bracelev > 0,
2718 structtype is the type of the preceding struct-like keyword, and
2719 structbracelev is the brace level where it has been seen. */
2720 static enum sym_type structtype;
2721 static int structbracelev;
2722 static enum sym_type toktype;
2725 toktype = C_symtype (str, len, *c_extp);
2728 * Skip __attribute__
2730 if (toktype == st_C_attribute)
2737 * Advance the definedef state machine.
2742 /* We're not on a preprocessor line. */
2743 if (toktype == st_C_gnumacro)
2750 if (toktype == st_C_define)
2752 definedef = ddefineseen;
2756 definedef = dignorerest;
2761 * Make a tag for any macro, unless it is a constant
2762 * and constantypedefs is FALSE.
2764 definedef = dignorerest;
2765 *is_func_or_var = (c == '(');
2766 if (!*is_func_or_var && !constantypedefs)
2773 error ("internal error: definedef value.", (char *)NULL);
2782 if (toktype == st_C_typedef)
2806 if (structdef == snone && fvdef == fvnone)
2833 case st_C_javastruct:
2834 if (structdef == stagseen)
2835 structdef = scolonseen;
2839 if ((*c_extp & C_AUTO) /* automatic detection of C++ language */
2841 && definedef == dnone && structdef == snone
2842 && typdef == tnone && fvdef == fvnone)
2843 *c_extp = (*c_extp | C_PLPL) & ~C_AUTO;
2844 if (toktype == st_C_template)
2851 && (typdef == tkeyseen
2852 || (typedefs_or_cplusplus && structdef == snone)))
2854 structdef = skeyseen;
2855 structtype = toktype;
2856 structbracelev = bracelev;
2857 if (fvdef == fvnameseen)
2867 if (structdef == skeyseen)
2869 structdef = stagseen;
2873 if (typdef != tnone)
2876 /* Detect Objective C constructs. */
2886 objdef = oimplementation;
2894 case oimplementation:
2895 /* Save the class tag for functions or variables defined inside. */
2896 objtag = savenstr (str, len);
2900 /* Save the class tag for categories. */
2901 objtag = savenstr (str, len);
2903 *is_func_or_var = TRUE;
2907 *is_func_or_var = TRUE;
2915 objdef = omethodtag;
2916 linebuffer_setlen (&token_name, len);
2917 strncpy (token_name.buffer, str, len);
2918 token_name.buffer[len] = '\0';
2924 objdef = omethodparm;
2930 objdef = omethodtag;
2931 linebuffer_setlen (&token_name, token_name.len + len);
2932 strncat (token_name.buffer, str, len);
2937 if (toktype == st_C_objend)
2939 /* Memory leakage here: the string pointed by objtag is
2940 never released, because many tests would be needed to
2941 avoid breaking on incorrect input code. The amount of
2942 memory leaked here is the sum of the lengths of the
2954 /* A function, variable or enum constant? */
2976 *is_func_or_var = TRUE;
2980 && structdef == snone
2981 && structtype == st_C_enum && bracelev > structbracelev)
2982 return TRUE; /* enum constant */
2988 fvdef = fdefunname; /* GNU macro */
2989 *is_func_or_var = TRUE;
2997 if ((strneq (str, "asm", 3) && endtoken (str[3]))
2998 || (strneq (str, "__asm__", 7) && endtoken (str[7])))
3011 if (len >= 10 && strneq (str+len-10, "::operator", 10))
3013 if (*c_extp & C_AUTO) /* automatic detection of C++ */
3014 *c_extp = (*c_extp | C_PLPL) & ~C_AUTO;
3016 *is_func_or_var = TRUE;
3019 if (bracelev > 0 && !instruct)
3021 fvdef = fvnameseen; /* function or variable */
3022 *is_func_or_var = TRUE;
3041 * C_entries often keeps pointers to tokens or lines which are older than
3042 * the line currently read. By keeping two line buffers, and switching
3043 * them at end of line, it is possible to use those pointers.
3051 #define current_lb_is_new (newndx == curndx)
3052 #define switch_line_buffers() (curndx = 1 - curndx)
3054 #define curlb (lbs[curndx].lb)
3055 #define newlb (lbs[newndx].lb)
3056 #define curlinepos (lbs[curndx].linepos)
3057 #define newlinepos (lbs[newndx].linepos)
3059 #define plainc ((c_ext & C_EXT) == C_PLAIN)
3060 #define cplpl (c_ext & C_PLPL)
3061 #define cjava ((c_ext & C_JAVA) == C_JAVA)
3063 #define CNL_SAVE_DEFINEDEF() \
3065 curlinepos = charno; \
3066 readline (&curlb, inf); \
3067 lp = curlb.buffer; \
3074 CNL_SAVE_DEFINEDEF(); \
3075 if (savetoken.valid) \
3077 token = savetoken; \
3078 savetoken.valid = FALSE; \
3080 definedef = dnone; \
3088 /* This function is never called when token.valid is FALSE, but
3089 we must protect against invalid input or internal errors. */
3091 make_tag (token_name.buffer, token_name.len, isfun, token.line,
3092 token.offset+token.length+1, token.lineno, token.linepos);
3094 { /* this branch is optimised away if !DEBUG */
3095 make_tag (concat ("INVALID TOKEN:-->", token_name.buffer, ""),
3096 token_name.len + 17, isfun, token.line,
3097 token.offset+token.length+1, token.lineno, token.linepos);
3098 error ("INVALID TOKEN", NULL);
3101 token.valid = FALSE;
3107 * This routine finds functions, variables, typedefs,
3108 * #define's, enum constants and struct/union/enum definitions in
3109 * C syntax and adds them to the list.
3112 C_entries (c_ext, inf)
3113 int c_ext; /* extension of C */
3114 FILE *inf; /* input file */
3116 register char c; /* latest char read; '\0' for end of line */
3117 register char *lp; /* pointer one beyond the character `c' */
3118 int curndx, newndx; /* indices for current and new lb */
3119 register int tokoff; /* offset in line of start of current token */
3120 register int toklen; /* length of current token */
3121 char *qualifier; /* string used to qualify names */
3122 int qlen; /* length of qualifier */
3123 int bracelev; /* current brace level */
3124 int bracketlev; /* current bracket level */
3125 int parlev; /* current parenthesis level */
3126 int attrparlev; /* __attribute__ parenthesis level */
3127 int templatelev; /* current template level */
3128 int typdefbracelev; /* bracelev where a typedef struct body begun */
3129 bool incomm, inquote, inchar, quotednl, midtoken;
3130 bool yacc_rules; /* in the rules part of a yacc file */
3131 struct tok savetoken = {0}; /* token saved during preprocessor handling */
3134 linebuffer_init (&lbs[0].lb);
3135 linebuffer_init (&lbs[1].lb);
3136 if (cstack.size == 0)
3138 cstack.size = (DEBUG) ? 1 : 4;
3140 cstack.cname = xnew (cstack.size, char *);
3141 cstack.bracelev = xnew (cstack.size, int);
3144 tokoff = toklen = typdefbracelev = 0; /* keep compiler quiet */
3145 curndx = newndx = 0;
3149 fvdef = fvnone; fvextern = FALSE; typdef = tnone;
3150 structdef = snone; definedef = dnone; objdef = onone;
3152 midtoken = inquote = inchar = incomm = quotednl = FALSE;
3153 token.valid = savetoken.valid = FALSE;
3154 bracelev = bracketlev = parlev = attrparlev = templatelev = 0;
3156 { qualifier = "."; qlen = 1; }
3158 { qualifier = "::"; qlen = 2; }
3166 /* If we are at the end of the line, the next character is a
3167 '\0'; do not skip it, because it is what tells us
3168 to read the next line. */
3189 /* Newlines inside comments do not end macro definitions in
3191 CNL_SAVE_DEFINEDEF ();
3206 /* Newlines inside strings do not end macro definitions
3207 in traditional cpp, even though compilers don't
3208 usually accept them. */
3209 CNL_SAVE_DEFINEDEF ();
3221 /* Hmmm, something went wrong. */
3232 else if (bracketlev > 0)
3237 if (--bracketlev > 0)
3241 CNL_SAVE_DEFINEDEF ();
3271 if (fvdef != finlist && fvdef != fignore && fvdef !=vignore)
3284 else if (/* cplpl && */ *lp == '/')
3290 if ((c_ext & YACC) && *lp == '%')
3292 /* Entering or exiting rules section in yacc file. */
3294 definedef = dnone; fvdef = fvnone; fvextern = FALSE;
3295 typdef = tnone; structdef = snone;
3296 midtoken = inquote = inchar = incomm = quotednl = FALSE;
3298 yacc_rules = !yacc_rules;
3304 if (definedef == dnone)
3307 bool cpptoken = TRUE;
3309 /* Look back on this line. If all blanks, or nonblanks
3310 followed by an end of comment, this is a preprocessor
3312 for (cp = newlb.buffer; cp < lp-1; cp++)
3315 if (*cp == '*' && *(cp+1) == '/')
3324 definedef = dsharpseen;
3325 } /* if (definedef == dnone) */
3335 /* Consider token only if some involved conditions are satisfied. */
3336 if (typdef != tignore
3337 && definedef != dignorerest
3340 && (definedef != dnone
3341 || structdef != scolonseen)
3348 if (c == ':' && *lp == ':' && begtoken (lp[1]))
3349 /* This handles :: in the middle,
3350 but not at the beginning of an identifier.
3351 Also, space-separated :: is not recognised. */
3353 if (c_ext & C_AUTO) /* automatic detection of C++ */
3354 c_ext = (c_ext | C_PLPL) & ~C_AUTO;
3358 goto still_in_token;
3362 bool funorvar = FALSE;
3365 || consider_token (newlb.buffer + tokoff, toklen, c,
3366 &c_ext, bracelev, parlev,
3369 if (fvdef == foperator)
3372 lp = skip_spaces (lp-1);
3376 && !iswhite (*lp) && *lp != '(')
3379 toklen += lp - oldlp;
3381 token.named = FALSE;
3383 && nestlev > 0 && definedef == dnone)
3384 /* in struct body */
3386 write_classname (&token_name, qualifier);
3387 linebuffer_setlen (&token_name,
3388 token_name.len+qlen+toklen);
3389 strcat (token_name.buffer, qualifier);
3390 strncat (token_name.buffer,
3391 newlb.buffer + tokoff, toklen);
3394 else if (objdef == ocatseen)
3395 /* Objective C category */
3397 int len = strlen (objtag) + 2 + toklen;
3398 linebuffer_setlen (&token_name, len);
3399 strcpy (token_name.buffer, objtag);
3400 strcat (token_name.buffer, "(");
3401 strncat (token_name.buffer,
3402 newlb.buffer + tokoff, toklen);
3403 strcat (token_name.buffer, ")");
3406 else if (objdef == omethodtag
3407 || objdef == omethodparm)
3408 /* Objective C method */
3412 else if (fvdef == fdefunname)
3413 /* GNU DEFUN and similar macros */
3415 bool defun = (newlb.buffer[tokoff] == 'F');
3419 /* Rewrite the tag so that emacs lisp DEFUNs
3420 can be found by their elisp name */
3426 linebuffer_setlen (&token_name, len);
3427 strncpy (token_name.buffer,
3428 newlb.buffer + off, len);
3429 token_name.buffer[len] = '\0';
3432 if (token_name.buffer[len] == '_')
3433 token_name.buffer[len] = '-';
3434 token.named = defun;
3438 linebuffer_setlen (&token_name, toklen);
3439 strncpy (token_name.buffer,
3440 newlb.buffer + tokoff, toklen);
3441 token_name.buffer[toklen] = '\0';
3442 /* Name macros and members. */
3443 token.named = (structdef == stagseen
3444 || typdef == ttypeseen
3447 && definedef == dignorerest)
3449 && definedef == dnone
3450 && structdef == snone
3453 token.lineno = lineno;
3454 token.offset = tokoff;
3455 token.length = toklen;
3456 token.line = newlb.buffer;
3457 token.linepos = newlinepos;
3460 if (definedef == dnone
3461 && (fvdef == fvnameseen
3462 || fvdef == foperator
3463 || structdef == stagseen
3465 || typdef == ttypeseen
3466 || objdef != onone))
3468 if (current_lb_is_new)
3469 switch_line_buffers ();
3471 else if (definedef != dnone
3472 || fvdef == fdefunname
3474 make_C_tag (funorvar);
3476 else /* not yacc and consider_token failed */
3478 if (inattribute && fvdef == fignore)
3480 /* We have just met __attribute__ after a
3481 function parameter list: do not tag the
3488 } /* if (endtoken (c)) */
3489 else if (intoken (c))
3495 } /* if (midtoken) */
3496 else if (begtoken (c))
3504 /* This prevents tagging fb in
3505 void (__attribute__((noreturn)) *fb) (void);
3506 Fixing this is not easy and not very important. */
3510 if (plainc || declarations)
3512 make_C_tag (TRUE); /* a function */
3521 if (structdef == stagseen && !cjava)
3523 popclass_above (bracelev);
3535 if (!yacc_rules || lp == newlb.buffer + 1)
3537 tokoff = lp - 1 - newlb.buffer;
3542 } /* if (begtoken) */
3543 } /* if must look at token */
3546 /* Detect end of line, colon, comma, semicolon and various braces
3547 after having handled a token.*/
3553 if (yacc_rules && token.offset == 0 && token.valid)
3555 make_C_tag (FALSE); /* a yacc function */
3558 if (definedef != dnone)
3564 make_C_tag (TRUE); /* an Objective C class */
3568 objdef = omethodcolon;
3569 linebuffer_setlen (&token_name, token_name.len + 1);
3570 strcat (token_name.buffer, ":");
3577 if (structdef == stagseen)
3579 structdef = scolonseen;
3582 /* Should be useless, but may be work as a safety net. */
3583 if (cplpl && fvdef == flistseen)
3585 make_C_tag (TRUE); /* a function */
3591 if (definedef != dnone || inattribute)
3597 make_C_tag (FALSE); /* a typedef */
3607 if (typdef == tignore || cplpl)
3611 if ((globals && bracelev == 0 && (!fvextern || declarations))
3612 || (members && instruct))
3613 make_C_tag (FALSE); /* a variable */
3616 token.valid = FALSE;
3620 && (cplpl || !instruct)
3621 && (typdef == tnone || (typdef != tignore && instruct)))
3623 && plainc && instruct))
3624 make_C_tag (TRUE); /* a function */
3630 && cplpl && structdef == stagseen)
3631 make_C_tag (FALSE); /* forward declaration */
3633 token.valid = FALSE;
3634 } /* switch (fvdef) */
3640 if (structdef == stagseen)
3644 if (definedef != dnone || inattribute)
3650 make_C_tag (TRUE); /* an Objective C method */
3675 && (!fvextern || declarations))
3676 || (members && instruct)))
3677 make_C_tag (FALSE); /* a variable */
3680 if ((declarations && typdef == tnone && !instruct)
3681 || (members && typdef != tignore && instruct))
3683 make_C_tag (TRUE); /* a function */
3686 else if (!declarations)
3688 token.valid = FALSE;
3693 if (structdef == stagseen)
3697 if (definedef != dnone || inattribute)
3699 if (structdef == stagseen)
3706 make_C_tag (FALSE); /* a typedef */
3718 if ((members && bracelev == 1)
3719 || (globals && bracelev == 0
3720 && (!fvextern || declarations)))
3721 make_C_tag (FALSE); /* a variable */
3739 if (definedef != dnone)
3741 if (objdef == otagseen && parlev == 0)
3742 objdef = oparenseen;
3746 if (typdef == ttypeseen
3750 /* This handles constructs like:
3751 typedef void OperatorFun (int fun); */
3774 if (--attrparlev == 0)
3775 inattribute = FALSE;
3778 if (definedef != dnone)
3780 if (objdef == ocatseen && parlev == 1)
3782 make_C_tag (TRUE); /* an Objective C category */
3800 || typdef == ttypeseen))
3803 make_C_tag (FALSE); /* a typedef */
3806 else if (parlev < 0) /* can happen due to ill-conceived #if's. */
3810 if (definedef != dnone)
3812 if (typdef == ttypeseen)
3814 /* Whenever typdef is set to tinbody (currently only
3815 here), typdefbracelev should be set to bracelev. */
3817 typdefbracelev = bracelev;
3822 make_C_tag (TRUE); /* a function */
3831 make_C_tag (TRUE); /* an Objective C class */
3836 make_C_tag (TRUE); /* an Objective C method */
3840 /* Neutralize `extern "C" {' grot. */
3841 if (bracelev == 0 && structdef == snone && nestlev == 0
3853 case skeyseen: /* unnamed struct */
3854 pushclass_above (bracelev, NULL, 0);
3857 case stagseen: /* named struct or enum */
3858 case scolonseen: /* a class */
3859 pushclass_above (bracelev,token.line+token.offset, token.length);
3861 make_C_tag (FALSE); /* a struct or enum */
3871 if (definedef != dnone)
3873 if (fvdef == fstartlist)
3875 fvdef = fvnone; /* avoid tagging `foo' in `foo (*bar()) ()' */
3876 token.valid = FALSE;
3880 if (definedef != dnone)
3883 if (!ignoreindent && lp == newlb.buffer + 1)
3886 token.valid = FALSE; /* unexpected value, token unreliable */
3887 bracelev = 0; /* reset brace level if first column */
3888 parlev = 0; /* also reset paren level, just in case... */
3890 else if (bracelev < 0)
3892 token.valid = FALSE; /* something gone amiss, token unreliable */
3895 if (bracelev == 0 && fvdef == vignore)
3896 fvdef = fvnone; /* end of function */
3897 popclass_above (bracelev);
3899 /* Only if typdef == tinbody is typdefbracelev significant. */
3900 if (typdef == tinbody && bracelev <= typdefbracelev)
3902 assert (bracelev == typdefbracelev);
3907 if (definedef != dnone)
3917 if ((members && bracelev == 1)
3918 || (globals && bracelev == 0 && (!fvextern || declarations)))
3919 make_C_tag (FALSE); /* a variable */
3927 && (structdef == stagseen || fvdef == fvnameseen))
3934 if (templatelev > 0)
3942 if (objdef == oinbody && bracelev == 0)
3944 objdef = omethodsign;
3949 case '#': case '~': case '&': case '%': case '/':
3950 case '|': case '^': case '!': case '.': case '?':
3951 if (definedef != dnone)
3953 /* These surely cannot follow a function tag in C. */
3966 if (objdef == otagseen)
3968 make_C_tag (TRUE); /* an Objective C class */
3971 /* If a macro spans multiple lines don't reset its state. */
3973 CNL_SAVE_DEFINEDEF ();
3981 } /* while not eof */
3983 free (lbs[0].lb.buffer);
3984 free (lbs[1].lb.buffer);
3988 * Process either a C++ file or a C file depending on the setting
3992 default_C_entries (inf)
3995 C_entries (cplusplus ? C_PLPL : C_AUTO, inf);
3998 /* Always do plain C. */
4000 plain_C_entries (inf)
4006 /* Always do C++. */
4008 Cplusplus_entries (inf)
4011 C_entries (C_PLPL, inf);
4014 /* Always do Java. */
4019 C_entries (C_JAVA, inf);
4027 C_entries (C_STAR, inf);
4030 /* Always do Yacc. */
4035 C_entries (YACC, inf);
4039 /* Useful macros. */
4040 #define LOOP_ON_INPUT_LINES(file_pointer, line_buffer, char_pointer) \
4041 for (; /* loop initialization */ \
4042 !feof (file_pointer) /* loop test */ \
4043 && /* instructions at start of loop */ \
4044 (readline (&line_buffer, file_pointer), \
4045 char_pointer = line_buffer.buffer, \
4049 #define LOOKING_AT(cp, kw) /* kw is the keyword, a literal string */ \
4050 ((assert("" kw), TRUE) /* syntax error if not a literal string */ \
4051 && strneq ((cp), kw, sizeof(kw)-1) /* cp points at kw */ \
4052 && notinname ((cp)[sizeof(kw)-1]) /* end of kw */ \
4053 && ((cp) = skip_spaces((cp)+sizeof(kw)-1))) /* skip spaces */
4055 /* Similar to LOOKING_AT but does not use notinname, does not skip */
4056 #define LOOKING_AT_NOCASE(cp, kw) /* the keyword is a literal string */ \
4057 ((assert("" kw), TRUE) /* syntax error if not a literal string */ \
4058 && strncaseeq ((cp), kw, sizeof(kw)-1) /* cp points at kw */ \
4059 && ((cp) += sizeof(kw)-1)) /* skip spaces */
4062 * Read a file, but do no processing. This is used to do regexp
4063 * matching on files that have no language defined.
4066 just_read_file (inf)
4069 register char *dummy;
4071 LOOP_ON_INPUT_LINES (inf, lb, dummy)
4076 /* Fortran parsing */
4078 static void F_takeprec __P((void));
4079 static void F_getit __P((FILE *));
4084 dbp = skip_spaces (dbp);
4088 dbp = skip_spaces (dbp);
4089 if (strneq (dbp, "(*)", 3))
4094 if (!ISDIGIT (*dbp))
4096 --dbp; /* force failure */
4101 while (ISDIGIT (*dbp));
4110 dbp = skip_spaces (dbp);
4113 readline (&lb, inf);
4118 dbp = skip_spaces (dbp);
4120 if (!ISALPHA (*dbp) && *dbp != '_' && *dbp != '$')
4122 for (cp = dbp + 1; *cp != '\0' && intoken (*cp); cp++)
4124 make_tag (dbp, cp-dbp, TRUE,
4125 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4130 Fortran_functions (inf)
4133 LOOP_ON_INPUT_LINES (inf, lb, dbp)
4136 dbp++; /* Ratfor escape to fortran */
4137 dbp = skip_spaces (dbp);
4140 switch (lowcase (*dbp))
4143 if (nocase_tail ("integer"))
4147 if (nocase_tail ("real"))
4151 if (nocase_tail ("logical"))
4155 if (nocase_tail ("complex") || nocase_tail ("character"))
4159 if (nocase_tail ("double"))
4161 dbp = skip_spaces (dbp);
4164 if (nocase_tail ("precision"))
4172 dbp = skip_spaces (dbp);
4175 switch (lowcase (*dbp))
4178 if (nocase_tail ("function"))
4182 if (nocase_tail ("subroutine"))
4186 if (nocase_tail ("entry"))
4190 if (nocase_tail ("blockdata") || nocase_tail ("block data"))
4192 dbp = skip_spaces (dbp);
4193 if (*dbp == '\0') /* assume un-named */
4194 make_tag ("blockdata", 9, TRUE,
4195 lb.buffer, dbp - lb.buffer, lineno, linecharno);
4197 F_getit (inf); /* look for name */
4210 * Philippe Waroquiers (1998)
4213 static void Ada_getit __P((FILE *, char *));
4215 /* Once we are positioned after an "interesting" keyword, let's get
4216 the real tag value necessary. */
4218 Ada_getit (inf, name_qualifier)
4220 char *name_qualifier;
4228 dbp = skip_spaces (dbp);
4230 || (dbp[0] == '-' && dbp[1] == '-'))
4232 readline (&lb, inf);
4235 switch (lowcase(*dbp))
4238 if (nocase_tail ("body"))
4240 /* Skipping body of procedure body or package body or ....
4241 resetting qualifier to body instead of spec. */
4242 name_qualifier = "/b";
4247 /* Skipping type of task type or protected type ... */
4248 if (nocase_tail ("type"))
4257 for (cp = dbp; *cp != '\0' && *cp != '"'; cp++)
4262 dbp = skip_spaces (dbp);
4265 && (ISALPHA (*cp) || ISDIGIT (*cp) || *cp == '_' || *cp == '.'));
4273 name = concat (dbp, name_qualifier, "");
4275 make_tag (name, strlen (name), TRUE,
4276 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4288 bool inquote = FALSE;
4289 bool skip_till_semicolumn = FALSE;
4291 LOOP_ON_INPUT_LINES (inf, lb, dbp)
4293 while (*dbp != '\0')
4295 /* Skip a string i.e. "abcd". */
4296 if (inquote || (*dbp == '"'))
4298 dbp = etags_strchr ((inquote) ? dbp : dbp+1, '"');
4303 continue; /* advance char */
4308 break; /* advance line */
4312 /* Skip comments. */
4313 if (dbp[0] == '-' && dbp[1] == '-')
4314 break; /* advance line */
4316 /* Skip character enclosed in single quote i.e. 'a'
4317 and skip single quote starting an attribute i.e. 'Image. */
4326 if (skip_till_semicolumn)
4329 skip_till_semicolumn = FALSE;
4331 continue; /* advance char */
4334 /* Search for beginning of a token. */
4335 if (!begtoken (*dbp))
4338 continue; /* advance char */
4341 /* We are at the beginning of a token. */
4342 switch (lowcase(*dbp))
4345 if (!packages_only && nocase_tail ("function"))
4346 Ada_getit (inf, "/f");
4348 break; /* from switch */
4349 continue; /* advance char */
4351 if (!packages_only && nocase_tail ("procedure"))
4352 Ada_getit (inf, "/p");
4353 else if (nocase_tail ("package"))
4354 Ada_getit (inf, "/s");
4355 else if (nocase_tail ("protected")) /* protected type */
4356 Ada_getit (inf, "/t");
4358 break; /* from switch */
4359 continue; /* advance char */
4362 if (typedefs && !packages_only && nocase_tail ("use"))
4364 /* when tagging types, avoid tagging use type Pack.Typename;
4365 for this, we will skip everything till a ; */
4366 skip_till_semicolumn = TRUE;
4367 continue; /* advance char */
4371 if (!packages_only && nocase_tail ("task"))
4372 Ada_getit (inf, "/k");
4373 else if (typedefs && !packages_only && nocase_tail ("type"))
4375 Ada_getit (inf, "/t");
4376 while (*dbp != '\0')
4380 break; /* from switch */
4381 continue; /* advance char */
4386 /* Look for the end of the token. */
4387 while (!endtoken (*dbp))
4390 } /* advance char */
4391 } /* advance line */
4396 * Unix and microcontroller assembly tag handling
4397 * Labels: /^[a-zA-Z_.$][a-zA_Z0-9_.$]*[: ^I^J]/
4398 * Idea by Bob Weiner, Motorola Inc. (1994)
4406 LOOP_ON_INPUT_LINES (inf, lb, cp)
4408 /* If first char is alphabetic or one of [_.$], test for colon
4409 following identifier. */
4410 if (ISALPHA (*cp) || *cp == '_' || *cp == '.' || *cp == '$')
4412 /* Read past label. */
4414 while (ISALNUM (*cp) || *cp == '_' || *cp == '.' || *cp == '$')
4416 if (*cp == ':' || iswhite (*cp))
4417 /* Found end of label, so copy it and add it to the table. */
4418 make_tag (lb.buffer, cp - lb.buffer, TRUE,
4419 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4427 * Perl sub names: /^sub[ \t\n]+[^ \t\n{]+/
4428 * Perl variable names: /^(my|local).../
4429 * Original code by Bart Robinson <lomew@cs.utah.edu> (1995)
4430 * Additions by Michael Ernst <mernst@alum.mit.edu> (1997)
4431 * Ideas by Kai Großjohann <Kai.Grossjohann@CS.Uni-Dortmund.DE> (2001)
4434 Perl_functions (inf)
4437 char *package = savestr ("main"); /* current package name */
4440 LOOP_ON_INPUT_LINES (inf, lb, cp)
4442 cp = skip_spaces (cp);
4444 if (LOOKING_AT (cp, "package"))
4447 get_tag (cp, &package);
4449 else if (LOOKING_AT (cp, "sub"))
4454 while (!notinname (*cp))
4457 continue; /* nothing found */
4458 if ((pos = etags_strchr (sp, ':')) != NULL
4459 && pos < cp && pos[1] == ':')
4460 /* The name is already qualified. */
4461 make_tag (sp, cp - sp, TRUE,
4462 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4466 char savechar, *name;
4470 name = concat (package, "::", sp);
4472 make_tag (name, strlen(name), TRUE,
4473 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4477 else if (globals) /* only if we are tagging global vars */
4479 /* Skip a qualifier, if any. */
4480 bool qual = LOOKING_AT (cp, "my") || LOOKING_AT (cp, "local");
4481 /* After "my" or "local", but before any following paren or space. */
4482 char *varstart = cp;
4484 if (qual /* should this be removed? If yes, how? */
4485 && (*cp == '$' || *cp == '@' || *cp == '%'))
4490 while (ISALNUM (*cp) || *cp == '_');
4494 /* Should be examining a variable list at this point;
4495 could insist on seeing an open parenthesis. */
4496 while (*cp != '\0' && *cp != ';' && *cp != '=' && *cp != ')')
4502 make_tag (varstart, cp - varstart, FALSE,
4503 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4512 * Look for /^[\t]*def[ \t\n]+[^ \t\n(:]+/ or /^class[ \t\n]+[^ \t\n(:]+/
4513 * Idea by Eric S. Raymond <esr@thyrsus.com> (1997)
4514 * More ideas by seb bacon <seb@jamkit.com> (2002)
4517 Python_functions (inf)
4522 LOOP_ON_INPUT_LINES (inf, lb, cp)
4524 cp = skip_spaces (cp);
4525 if (LOOKING_AT (cp, "def") || LOOKING_AT (cp, "class"))
4528 while (!notinname (*cp) && *cp != ':')
4530 make_tag (name, cp - name, TRUE,
4531 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4540 * - /^[ \t]*function[ \t\n]+[^ \t\n(]+/
4541 * - /^[ \t]*class[ \t\n]+[^ \t\n]+/
4542 * - /^[ \t]*define\(\"[^\"]+/
4543 * Only with --members:
4544 * - /^[ \t]*var[ \t\n]+\$[^ \t\n=;]/
4545 * Idea by Diez B. Roggisch (2001)
4551 register char *cp, *name;
4552 bool search_identifier = FALSE;
4554 LOOP_ON_INPUT_LINES (inf, lb, cp)
4556 cp = skip_spaces (cp);
4558 if (search_identifier
4561 while (!notinname (*cp))
4563 make_tag (name, cp - name, TRUE,
4564 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4565 search_identifier = FALSE;
4567 else if (LOOKING_AT (cp, "function"))
4570 cp = skip_spaces (cp+1);
4574 while (!notinname (*cp))
4576 make_tag (name, cp - name, TRUE,
4577 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4580 search_identifier = TRUE;
4582 else if (LOOKING_AT (cp, "class"))
4587 while (*cp != '\0' && !iswhite (*cp))
4589 make_tag (name, cp - name, FALSE,
4590 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4593 search_identifier = TRUE;
4595 else if (strneq (cp, "define", 6)
4596 && (cp = skip_spaces (cp+6))
4598 && (*cp == '"' || *cp == '\''))
4602 while (*cp != quote && *cp != '\0')
4604 make_tag (name, cp - name, FALSE,
4605 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4608 && LOOKING_AT (cp, "var")
4612 while (!notinname(*cp))
4614 make_tag (name, cp - name, FALSE,
4615 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4622 * Cobol tag functions
4623 * We could look for anything that could be a paragraph name.
4624 * i.e. anything that starts in column 8 is one word and ends in a full stop.
4625 * Idea by Corny de Souza (1993)
4628 Cobol_paragraphs (inf)
4631 register char *bp, *ep;
4633 LOOP_ON_INPUT_LINES (inf, lb, bp)
4639 /* If eoln, compiler option or comment ignore whole line. */
4640 if (bp[-1] != ' ' || !ISALNUM (bp[0]))
4643 for (ep = bp; ISALNUM (*ep) || *ep == '-'; ep++)
4646 make_tag (bp, ep - bp, TRUE,
4647 lb.buffer, ep - lb.buffer + 1, lineno, linecharno);
4654 * Ideas by Assar Westerlund <assar@sics.se> (2001)
4657 Makefile_targets (inf)
4662 LOOP_ON_INPUT_LINES (inf, lb, bp)
4664 if (*bp == '\t' || *bp == '#')
4666 while (*bp != '\0' && *bp != '=' && *bp != ':')
4668 if (*bp == ':' || (globals && *bp == '='))
4670 /* We should detect if there is more than one tag, but we do not.
4671 We just skip initial and final spaces. */
4672 char * namestart = skip_spaces (lb.buffer);
4673 while (--bp > namestart)
4674 if (!notinname (*bp))
4676 make_tag (namestart, bp - namestart + 1, TRUE,
4677 lb.buffer, bp - lb.buffer + 2, lineno, linecharno);
4685 * Original code by Mosur K. Mohan (1989)
4687 * Locates tags for procedures & functions. Doesn't do any type- or
4688 * var-definitions. It does look for the keyword "extern" or
4689 * "forward" immediately following the procedure statement; if found,
4690 * the tag is skipped.
4693 Pascal_functions (inf)
4696 linebuffer tline; /* mostly copied from C_entries */
4698 int save_lineno, namelen, taglen;
4701 bool /* each of these flags is TRUE if: */
4702 incomment, /* point is inside a comment */
4703 inquote, /* point is inside '..' string */
4704 get_tagname, /* point is after PROCEDURE/FUNCTION
4705 keyword, so next item = potential tag */
4706 found_tag, /* point is after a potential tag */
4707 inparms, /* point is within parameter-list */
4708 verify_tag; /* point has passed the parm-list, so the
4709 next token will determine whether this
4710 is a FORWARD/EXTERN to be ignored, or
4711 whether it is a real tag */
4713 save_lcno = save_lineno = namelen = taglen = 0; /* keep compiler quiet */
4714 name = NULL; /* keep compiler quiet */
4717 linebuffer_init (&tline);
4719 incomment = inquote = FALSE;
4720 found_tag = FALSE; /* have a proc name; check if extern */
4721 get_tagname = FALSE; /* found "procedure" keyword */
4722 inparms = FALSE; /* found '(' after "proc" */
4723 verify_tag = FALSE; /* check if "extern" is ahead */
4726 while (!feof (inf)) /* long main loop to get next char */
4729 if (c == '\0') /* if end of line */
4731 readline (&lb, inf);
4735 if (!((found_tag && verify_tag)
4737 c = *dbp++; /* only if don't need *dbp pointing
4738 to the beginning of the name of
4739 the procedure or function */
4743 if (c == '}') /* within { } comments */
4745 else if (c == '*' && *dbp == ')') /* within (* *) comments */
4762 inquote = TRUE; /* found first quote */
4764 case '{': /* found open { comment */
4768 if (*dbp == '*') /* found open (* comment */
4773 else if (found_tag) /* found '(' after tag, i.e., parm-list */
4776 case ')': /* end of parms list */
4781 if (found_tag && !inparms) /* end of proc or fn stmt */
4790 if (found_tag && verify_tag && (*dbp != ' '))
4792 /* Check if this is an "extern" declaration. */
4795 if (lowcase (*dbp == 'e'))
4797 if (nocase_tail ("extern")) /* superfluous, really! */
4803 else if (lowcase (*dbp) == 'f')
4805 if (nocase_tail ("forward")) /* check for forward reference */
4811 if (found_tag && verify_tag) /* not external proc, so make tag */
4815 make_tag (name, namelen, TRUE,
4816 tline.buffer, taglen, save_lineno, save_lcno);
4820 if (get_tagname) /* grab name of proc or fn */
4827 /* Find block name. */
4828 for (cp = dbp + 1; *cp != '\0' && !endtoken (*cp); cp++)
4831 /* Save all values for later tagging. */
4832 linebuffer_setlen (&tline, lb.len);
4833 strcpy (tline.buffer, lb.buffer);
4834 save_lineno = lineno;
4835 save_lcno = linecharno;
4836 name = tline.buffer + (dbp - lb.buffer);
4838 taglen = cp - lb.buffer + 1;
4840 dbp = cp; /* set dbp to e-o-token */
4841 get_tagname = FALSE;
4845 /* And proceed to check for "extern". */
4847 else if (!incomment && !inquote && !found_tag)
4849 /* Check for proc/fn keywords. */
4850 switch (lowcase (c))
4853 if (nocase_tail ("rocedure")) /* c = 'p', dbp has advanced */
4857 if (nocase_tail ("unction"))
4864 } /* while not eof */
4866 free (tline.buffer);
4871 * Lisp tag functions
4872 * look for (def or (DEF, quote or QUOTE
4875 static void L_getit __P((void));
4880 if (*dbp == '\'') /* Skip prefix quote */
4882 else if (*dbp == '(')
4885 /* Try to skip "(quote " */
4886 if (!LOOKING_AT (dbp, "quote") && !LOOKING_AT (dbp, "QUOTE"))
4887 /* Ok, then skip "(" before name in (defstruct (foo)) */
4888 dbp = skip_spaces (dbp);
4890 get_tag (dbp, NULL);
4894 Lisp_functions (inf)
4897 LOOP_ON_INPUT_LINES (inf, lb, dbp)
4902 if (strneq (dbp+1, "def", 3) || strneq (dbp+1, "DEF", 3))
4904 dbp = skip_non_spaces (dbp);
4905 dbp = skip_spaces (dbp);
4910 /* Check for (foo::defmumble name-defined ... */
4913 while (!notinname (*dbp) && *dbp != ':');
4918 while (*dbp == ':');
4920 if (strneq (dbp, "def", 3) || strneq (dbp, "DEF", 3))
4922 dbp = skip_non_spaces (dbp);
4923 dbp = skip_spaces (dbp);
4933 * Lua script language parsing
4934 * Original code by David A. Capello <dacap@users.sourceforge.net> (2004)
4936 * "function" and "local function" are tags if they start at column 1.
4944 LOOP_ON_INPUT_LINES (inf, lb, bp)
4946 if (bp[0] != 'f' && bp[0] != 'l')
4949 (void)LOOKING_AT (bp, "local"); /* skip possible "local" */
4951 if (LOOKING_AT (bp, "function"))
4959 * Just look for lines where the first character is '/'
4960 * Also look at "defineps" for PSWrap
4962 * Richard Mlynarik <mly@adoc.xerox.com> (1997)
4963 * Masatake Yamato <masata-y@is.aist-nara.ac.jp> (1999)
4969 register char *bp, *ep;
4971 LOOP_ON_INPUT_LINES (inf, lb, bp)
4976 *ep != '\0' && *ep != ' ' && *ep != '{';
4979 make_tag (bp, ep - bp, TRUE,
4980 lb.buffer, ep - lb.buffer + 1, lineno, linecharno);
4982 else if (LOOKING_AT (bp, "defineps"))
4990 * Ignore anything after \ followed by space or in ( )
4991 * Look for words defined by :
4992 * Look for constant, code, create, defer, value, and variable
4993 * OBP extensions: Look for buffer:, field,
4994 * Ideas by Eduardo Horvath <eeh@netbsd.org> (2004)
5002 LOOP_ON_INPUT_LINES (inf, lb, bp)
5003 while ((bp = skip_spaces (bp))[0] != '\0')
5004 if (bp[0] == '\\' && iswhite(bp[1]))
5005 break; /* read next line */
5006 else if (bp[0] == '(' && iswhite(bp[1]))
5007 do /* skip to ) or eol */
5009 while (*bp != ')' && *bp != '\0');
5010 else if ((bp[0] == ':' && iswhite(bp[1]) && bp++)
5011 || LOOKING_AT_NOCASE (bp, "constant")
5012 || LOOKING_AT_NOCASE (bp, "code")
5013 || LOOKING_AT_NOCASE (bp, "create")
5014 || LOOKING_AT_NOCASE (bp, "defer")
5015 || LOOKING_AT_NOCASE (bp, "value")
5016 || LOOKING_AT_NOCASE (bp, "variable")
5017 || LOOKING_AT_NOCASE (bp, "buffer:")
5018 || LOOKING_AT_NOCASE (bp, "field"))
5019 get_tag (skip_spaces (bp), NULL); /* Yay! A definition! */
5021 bp = skip_non_spaces (bp);
5026 * Scheme tag functions
5027 * look for (def... xyzzy
5029 * (def ... ((...(xyzzy ....
5031 * Original code by Ken Haase (1985?)
5034 Scheme_functions (inf)
5039 LOOP_ON_INPUT_LINES (inf, lb, bp)
5041 if (strneq (bp, "(def", 4) || strneq (bp, "(DEF", 4))
5043 bp = skip_non_spaces (bp+4);
5044 /* Skip over open parens and white space */
5045 while (notinname (*bp))
5049 if (LOOKING_AT (bp, "(SET!") || LOOKING_AT (bp, "(set!"))
5055 /* Find tags in TeX and LaTeX input files. */
5057 /* TEX_toktab is a table of TeX control sequences that define tags.
5058 * Each entry records one such control sequence.
5060 * Original code from who knows whom.
5062 * Stefan Monnier (2002)
5065 static linebuffer *TEX_toktab = NULL; /* Table with tag tokens */
5067 /* Default set of control sequences to put into TEX_toktab.
5068 The value of environment var TEXTAGS is prepended to this. */
5069 static char *TEX_defenv = "\
5070 :chapter:section:subsection:subsubsection:eqno:label:ref:cite:bibitem\
5071 :part:appendix:entry:index:def\
5072 :newcommand:renewcommand:newenvironment:renewenvironment";
5074 static void TEX_mode __P((FILE *));
5075 static void TEX_decode_env __P((char *, char *));
5077 static char TEX_esc = '\\';
5078 static char TEX_opgrp = '{';
5079 static char TEX_clgrp = '}';
5082 * TeX/LaTeX scanning loop.
5091 /* Select either \ or ! as escape character. */
5094 /* Initialize token table once from environment. */
5095 if (TEX_toktab == NULL)
5096 TEX_decode_env ("TEXTAGS", TEX_defenv);
5098 LOOP_ON_INPUT_LINES (inf, lb, cp)
5100 /* Look at each TEX keyword in line. */
5103 /* Look for a TEX escape. */
5104 while (*cp++ != TEX_esc)
5105 if (cp[-1] == '\0' || cp[-1] == '%')
5108 for (key = TEX_toktab; key->buffer != NULL; key++)
5109 if (strneq (cp, key->buffer, key->len))
5112 int namelen, linelen;
5115 cp = skip_spaces (cp + key->len);
5116 if (*cp == TEX_opgrp)
5122 (!iswhite (*p) && *p != '#' &&
5123 *p != TEX_opgrp && *p != TEX_clgrp);
5128 if (!opgrp || *p == TEX_clgrp)
5130 while (*p != '\0' && *p != TEX_opgrp && *p != TEX_clgrp)
5132 linelen = p - lb.buffer + 1;
5134 make_tag (cp, namelen, TRUE,
5135 lb.buffer, linelen, lineno, linecharno);
5136 goto tex_next_line; /* We only tag a line once */
5144 #define TEX_LESC '\\'
5145 #define TEX_SESC '!'
5147 /* Figure out whether TeX's escapechar is '\\' or '!' and set grouping
5148 chars accordingly. */
5155 while ((c = getc (inf)) != EOF)
5157 /* Skip to next line if we hit the TeX comment char. */
5159 while (c != '\n' && c != EOF)
5161 else if (c == TEX_LESC || c == TEX_SESC )
5177 /* If the input file is compressed, inf is a pipe, and rewind may fail.
5178 No attempt is made to correct the situation. */
5182 /* Read environment and prepend it to the default string.
5183 Build token table. */
5185 TEX_decode_env (evarname, defenv)
5189 register char *env, *p;
5192 /* Append default string to environment. */
5193 env = getenv (evarname);
5199 env = concat (oldenv, defenv, "");
5202 /* Allocate a token table */
5203 for (len = 1, p = env; p;)
5204 if ((p = etags_strchr (p, ':')) && *++p != '\0')
5206 TEX_toktab = xnew (len, linebuffer);
5208 /* Unpack environment string into token table. Be careful about */
5209 /* zero-length strings (leading ':', "::" and trailing ':') */
5210 for (i = 0; *env != '\0';)
5212 p = etags_strchr (env, ':');
5213 if (!p) /* End of environment string. */
5214 p = env + strlen (env);
5216 { /* Only non-zero strings. */
5217 TEX_toktab[i].buffer = savenstr (env, p - env);
5218 TEX_toktab[i].len = p - env;
5225 TEX_toktab[i].buffer = NULL; /* Mark end of table. */
5226 TEX_toktab[i].len = 0;
5233 /* Texinfo support. Dave Love, Mar. 2000. */
5239 LOOP_ON_INPUT_LINES (inf, lb, cp)
5240 if (LOOKING_AT (cp, "@node"))
5243 while (*cp != '\0' && *cp != ',')
5245 make_tag (start, cp - start, TRUE,
5246 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
5253 * Contents of <title>, <h1>, <h2>, <h3> are tags.
5254 * Contents of <a name=xxx> are tags with name xxx.
5256 * Francesco Potortì, 2002.
5262 bool getnext = FALSE; /* next text outside of HTML tags is a tag */
5263 bool skiptag = FALSE; /* skip to the end of the current HTML tag */
5264 bool intag = FALSE; /* inside an html tag, looking for ID= */
5265 bool inanchor = FALSE; /* when INTAG, is an anchor, look for NAME= */
5269 linebuffer_setlen (&token_name, 0); /* no name in buffer */
5271 LOOP_ON_INPUT_LINES (inf, lb, dbp)
5272 for (;;) /* loop on the same line */
5274 if (skiptag) /* skip HTML tag */
5276 while (*dbp != '\0' && *dbp != '>')
5282 continue; /* look on the same line */
5284 break; /* go to next line */
5287 else if (intag) /* look for "name=" or "id=" */
5289 while (*dbp != '\0' && *dbp != '>'
5290 && lowcase (*dbp) != 'n' && lowcase (*dbp) != 'i')
5293 break; /* go to next line */
5298 continue; /* look on the same line */
5300 if ((inanchor && LOOKING_AT_NOCASE (dbp, "name="))
5301 || LOOKING_AT_NOCASE (dbp, "id="))
5303 bool quoted = (dbp[0] == '"');
5306 for (end = ++dbp; *end != '\0' && *end != '"'; end++)
5309 for (end = dbp; *end != '\0' && intoken (*end); end++)
5311 linebuffer_setlen (&token_name, end - dbp);
5312 strncpy (token_name.buffer, dbp, end - dbp);
5313 token_name.buffer[end - dbp] = '\0';
5316 intag = FALSE; /* we found what we looked for */
5317 skiptag = TRUE; /* skip to the end of the tag */
5318 getnext = TRUE; /* then grab the text */
5319 continue; /* look on the same line */
5324 else if (getnext) /* grab next tokens and tag them */
5326 dbp = skip_spaces (dbp);
5328 break; /* go to next line */
5332 inanchor = (lowcase (dbp[1]) == 'a' && !intoken (dbp[2]));
5333 continue; /* look on the same line */
5336 for (end = dbp + 1; *end != '\0' && *end != '<'; end++)
5338 make_tag (token_name.buffer, token_name.len, TRUE,
5339 dbp, end - dbp, lineno, linecharno);
5340 linebuffer_setlen (&token_name, 0); /* no name in buffer */
5342 break; /* go to next line */
5345 else /* look for an interesting HTML tag */
5347 while (*dbp != '\0' && *dbp != '<')
5350 break; /* go to next line */
5352 if (lowcase (dbp[1]) == 'a' && !intoken (dbp[2]))
5355 continue; /* look on the same line */
5357 else if (LOOKING_AT_NOCASE (dbp, "<title>")
5358 || LOOKING_AT_NOCASE (dbp, "<h1>")
5359 || LOOKING_AT_NOCASE (dbp, "<h2>")
5360 || LOOKING_AT_NOCASE (dbp, "<h3>"))
5364 continue; /* look on the same line */
5375 * Assumes that the predicate or rule starts at column 0.
5376 * Only the first clause of a predicate or rule is added.
5377 * Original code by Sunichirou Sugou (1989)
5378 * Rewritten by Anders Lindgren (1996)
5380 static int prolog_pr __P((char *, char *));
5381 static void prolog_skip_comment __P((linebuffer *, FILE *));
5382 static int prolog_atom __P((char *, int));
5385 Prolog_functions (inf)
5396 LOOP_ON_INPUT_LINES (inf, lb, cp)
5398 if (cp[0] == '\0') /* Empty line */
5400 else if (iswhite (cp[0])) /* Not a predicate */
5402 else if (cp[0] == '/' && cp[1] == '*') /* comment. */
5403 prolog_skip_comment (&lb, inf);
5404 else if ((len = prolog_pr (cp, last)) > 0)
5406 /* Predicate or rule. Store the function name so that we
5407 only generate a tag for the first clause. */
5409 last = xnew(len + 1, char);
5410 else if (len + 1 > allocated)
5411 xrnew (last, len + 1, char);
5412 allocated = len + 1;
5413 strncpy (last, cp, len);
5422 prolog_skip_comment (plb, inf)
5430 for (cp = plb->buffer; *cp != '\0'; cp++)
5431 if (cp[0] == '*' && cp[1] == '/')
5433 readline (plb, inf);
5439 * A predicate or rule definition is added if it matches:
5440 * <beginning of line><Prolog Atom><whitespace>(
5441 * or <beginning of line><Prolog Atom><whitespace>:-
5443 * It is added to the tags database if it doesn't match the
5444 * name of the previous clause header.
5446 * Return the size of the name of the predicate or rule, or 0 if no
5452 char *last; /* Name of last clause. */
5457 pos = prolog_atom (s, 0);
5462 pos = skip_spaces (s + pos) - s;
5465 || (s[pos] == '(' && (pos += 1))
5466 || (s[pos] == ':' && s[pos + 1] == '-' && (pos += 2)))
5467 && (last == NULL /* save only the first clause */
5468 || len != (int)strlen (last)
5469 || !strneq (s, last, len)))
5471 make_tag (s, len, TRUE, s, pos, lineno, linecharno);
5479 * Consume a Prolog atom.
5480 * Return the number of bytes consumed, or -1 if there was an error.
5482 * A prolog atom, in this context, could be one of:
5483 * - An alphanumeric sequence, starting with a lower case letter.
5484 * - A quoted arbitrary string. Single quotes can escape themselves.
5485 * Backslash quotes everything.
5488 prolog_atom (s, pos)
5496 if (ISLOWER(s[pos]) || (s[pos] == '_'))
5498 /* The atom is unquoted. */
5500 while (ISALNUM(s[pos]) || (s[pos] == '_'))
5504 return pos - origpos;
5506 else if (s[pos] == '\'')
5517 pos++; /* A double quote */
5519 else if (s[pos] == '\0')
5520 /* Multiline quoted atoms are ignored. */
5522 else if (s[pos] == '\\')
5524 if (s[pos+1] == '\0')
5531 return pos - origpos;
5539 * Support for Erlang
5541 * Generates tags for functions, defines, and records.
5542 * Assumes that Erlang functions start at column 0.
5543 * Original code by Anders Lindgren (1996)
5545 static int erlang_func __P((char *, char *));
5546 static void erlang_attribute __P((char *));
5547 static int erlang_atom __P((char *));
5550 Erlang_functions (inf)
5561 LOOP_ON_INPUT_LINES (inf, lb, cp)
5563 if (cp[0] == '\0') /* Empty line */
5565 else if (iswhite (cp[0])) /* Not function nor attribute */
5567 else if (cp[0] == '%') /* comment */
5569 else if (cp[0] == '"') /* Sometimes, strings start in column one */
5571 else if (cp[0] == '-') /* attribute, e.g. "-define" */
5573 erlang_attribute (cp);
5580 else if ((len = erlang_func (cp, last)) > 0)
5583 * Function. Store the function name so that we only
5584 * generates a tag for the first clause.
5587 last = xnew (len + 1, char);
5588 else if (len + 1 > allocated)
5589 xrnew (last, len + 1, char);
5590 allocated = len + 1;
5591 strncpy (last, cp, len);
5600 * A function definition is added if it matches:
5601 * <beginning of line><Erlang Atom><whitespace>(
5603 * It is added to the tags database if it doesn't match the
5604 * name of the previous clause header.
5606 * Return the size of the name of the function, or 0 if no function
5610 erlang_func (s, last)
5612 char *last; /* Name of last clause. */
5617 pos = erlang_atom (s);
5622 pos = skip_spaces (s + pos) - s;
5624 /* Save only the first clause. */
5627 || len != (int)strlen (last)
5628 || !strneq (s, last, len)))
5630 make_tag (s, len, TRUE, s, pos, lineno, linecharno);
5639 * Handle attributes. Currently, tags are generated for defines
5642 * They are on the form:
5643 * -define(foo, bar).
5644 * -define(Foo(M, N), M+N).
5645 * -record(graph, {vtab = notable, cyclic = true}).
5648 erlang_attribute (s)
5653 if ((LOOKING_AT (cp, "-define") || LOOKING_AT (cp, "-record"))
5656 int len = erlang_atom (skip_spaces (cp));
5658 make_tag (cp, len, TRUE, s, cp + len - s, lineno, linecharno);
5665 * Consume an Erlang atom (or variable).
5666 * Return the number of bytes consumed, or -1 if there was an error.
5674 if (ISALPHA (s[pos]) || s[pos] == '_')
5676 /* The atom is unquoted. */
5679 while (ISALNUM (s[pos]) || s[pos] == '_');
5681 else if (s[pos] == '\'')
5683 for (pos++; s[pos] != '\''; pos++)
5684 if (s[pos] == '\0' /* multiline quoted atoms are ignored */
5685 || (s[pos] == '\\' && s[++pos] == '\0'))
5694 static char *scan_separators __P((char *));
5695 static void add_regex __P((char *, language *));
5696 static char *substitute __P((char *, char *, struct re_registers *));
5699 * Take a string like "/blah/" and turn it into "blah", verifying
5700 * that the first and last characters are the same, and handling
5701 * quoted separator characters. Actually, stops on the occurrence of
5702 * an unquoted separator. Also process \t, \n, etc. and turn into
5703 * appropriate characters. Works in place. Null terminates name string.
5704 * Returns pointer to terminating separator, or NULL for
5705 * unterminated regexps.
5708 scan_separators (name)
5712 char *copyto = name;
5713 bool quoted = FALSE;
5715 for (++name; *name != '\0'; ++name)
5721 case 'a': *copyto++ = '\007'; break; /* BEL (bell) */
5722 case 'b': *copyto++ = '\b'; break; /* BS (back space) */
5723 case 'd': *copyto++ = 0177; break; /* DEL (delete) */
5724 case 'e': *copyto++ = 033; break; /* ESC (delete) */
5725 case 'f': *copyto++ = '\f'; break; /* FF (form feed) */
5726 case 'n': *copyto++ = '\n'; break; /* NL (new line) */
5727 case 'r': *copyto++ = '\r'; break; /* CR (carriage return) */
5728 case 't': *copyto++ = '\t'; break; /* TAB (horizontal tab) */
5729 case 'v': *copyto++ = '\v'; break; /* VT (vertical tab) */
5735 /* Something else is quoted, so preserve the quote. */
5743 else if (*name == '\\')
5745 else if (*name == sep)
5751 name = NULL; /* signal unterminated regexp */
5753 /* Terminate copied string. */
5758 /* Look at the argument of --regex or --no-regex and do the right
5759 thing. Same for each line of a regexp file. */
5761 analyse_regex (regex_arg)
5764 if (regex_arg == NULL)
5766 free_regexps (); /* --no-regex: remove existing regexps */
5770 /* A real --regexp option or a line in a regexp file. */
5771 switch (regex_arg[0])
5773 /* Comments in regexp file or null arg to --regex. */
5779 /* Read a regex file. This is recursive and may result in a
5780 loop, which will stop when the file descriptors are exhausted. */
5784 linebuffer regexbuf;
5785 char *regexfile = regex_arg + 1;
5787 /* regexfile is a file containing regexps, one per line. */
5788 regexfp = fopen (regexfile, "r");
5789 if (regexfp == NULL)
5794 linebuffer_init (®exbuf);
5795 while (readline_internal (®exbuf, regexfp) > 0)
5796 analyse_regex (regexbuf.buffer);
5797 free (regexbuf.buffer);
5802 /* Regexp to be used for a specific language only. */
5806 char *lang_name = regex_arg + 1;
5809 for (cp = lang_name; *cp != '}'; cp++)
5812 error ("unterminated language name in regex: %s", regex_arg);
5816 lang = get_language_from_langname (lang_name);
5819 add_regex (cp, lang);
5823 /* Regexp to be used for any language. */
5825 add_regex (regex_arg, NULL);
5830 /* Separate the regexp pattern, compile it,
5831 and care for optional name and modifiers. */
5833 add_regex (regexp_pattern, lang)
5834 char *regexp_pattern;
5837 static struct re_pattern_buffer zeropattern;
5838 char sep, *pat, *name, *modifiers;
5840 struct re_pattern_buffer *patbuf;
5843 force_explicit_name = TRUE, /* do not use implicit tag names */
5844 ignore_case = FALSE, /* case is significant */
5845 multi_line = FALSE, /* matches are done one line at a time */
5846 single_line = FALSE; /* dot does not match newline */
5849 if (strlen(regexp_pattern) < 3)
5851 error ("null regexp", (char *)NULL);
5854 sep = regexp_pattern[0];
5855 name = scan_separators (regexp_pattern);
5858 error ("%s: unterminated regexp", regexp_pattern);
5863 error ("null name for regexp \"%s\"", regexp_pattern);
5866 modifiers = scan_separators (name);
5867 if (modifiers == NULL) /* no terminating separator --> no name */
5873 modifiers += 1; /* skip separator */
5875 /* Parse regex modifiers. */
5876 for (; modifiers[0] != '\0'; modifiers++)
5877 switch (modifiers[0])
5880 if (modifiers == name)
5881 error ("forcing explicit tag name but no name, ignoring", NULL);
5882 force_explicit_name = TRUE;
5892 need_filebuf = TRUE;
5897 wrongmod[0] = modifiers[0];
5899 error ("invalid regexp modifier `%s', ignoring", wrongmod);
5904 patbuf = xnew (1, struct re_pattern_buffer);
5905 *patbuf = zeropattern;
5908 static char lc_trans[CHARS];
5910 for (i = 0; i < CHARS; i++)
5911 lc_trans[i] = lowcase (i);
5912 patbuf->translate = lc_trans; /* translation table to fold case */
5916 pat = concat ("^", regexp_pattern, ""); /* anchor to beginning of line */
5918 pat = regexp_pattern;
5921 re_set_syntax (RE_SYNTAX_EMACS | RE_DOT_NEWLINE);
5923 re_set_syntax (RE_SYNTAX_EMACS);
5925 err = re_compile_pattern (pat, strlen (regexp_pattern), patbuf);
5930 error ("%s while compiling pattern", err);
5935 p_head = xnew (1, regexp);
5936 p_head->pattern = savestr (regexp_pattern);
5937 p_head->p_next = rp;
5938 p_head->lang = lang;
5939 p_head->pat = patbuf;
5940 p_head->name = savestr (name);
5941 p_head->error_signaled = FALSE;
5942 p_head->force_explicit_name = force_explicit_name;
5943 p_head->ignore_case = ignore_case;
5944 p_head->multi_line = multi_line;
5948 * Do the substitutions indicated by the regular expression and
5952 substitute (in, out, regs)
5954 struct re_registers *regs;
5957 int size, dig, diglen;
5960 size = strlen (out);
5962 /* Pass 1: figure out how much to allocate by finding all \N strings. */
5963 if (out[size - 1] == '\\')
5964 fatal ("pattern error in \"%s\"", out);
5965 for (t = etags_strchr (out, '\\');
5967 t = etags_strchr (t + 2, '\\'))
5971 diglen = regs->end[dig] - regs->start[dig];
5977 /* Allocate space and do the substitutions. */
5979 result = xnew (size + 1, char);
5981 for (t = result; *out != '\0'; out++)
5982 if (*out == '\\' && ISDIGIT (*++out))
5985 diglen = regs->end[dig] - regs->start[dig];
5986 strncpy (t, in + regs->start[dig], diglen);
5993 assert (t <= result + size);
5994 assert (t - result == (int)strlen (result));
5999 /* Deallocate all regexps. */
6004 while (p_head != NULL)
6006 rp = p_head->p_next;
6007 free (p_head->pattern);
6008 free (p_head->name);
6016 * Reads the whole file as a single string from `filebuf' and looks for
6017 * multi-line regular expressions, creating tags on matches.
6018 * readline already dealt with normal regexps.
6020 * Idea by Ben Wing <ben@666.com> (2002).
6023 regex_tag_multiline ()
6025 char *buffer = filebuf.buffer;
6029 for (rp = p_head; rp != NULL; rp = rp->p_next)
6033 if (!rp->multi_line)
6034 continue; /* skip normal regexps */
6036 /* Generic initialisations before parsing file from memory. */
6037 lineno = 1; /* reset global line number */
6038 charno = 0; /* reset global char number */
6039 linecharno = 0; /* reset global char number of line start */
6041 /* Only use generic regexps or those for the current language. */
6042 if (rp->lang != NULL && rp->lang != curfdp->lang)
6045 while (match >= 0 && match < filebuf.len)
6047 match = re_search (rp->pat, buffer, filebuf.len, charno,
6048 filebuf.len - match, &rp->regs);
6053 if (!rp->error_signaled)
6055 error ("regexp stack overflow while matching \"%s\"",
6057 rp->error_signaled = TRUE;
6064 if (match == rp->regs.end[0])
6066 if (!rp->error_signaled)
6068 error ("regexp matches the empty string: \"%s\"",
6070 rp->error_signaled = TRUE;
6072 match = -3; /* exit from while loop */
6076 /* Match occurred. Construct a tag. */
6077 while (charno < rp->regs.end[0])
6078 if (buffer[charno++] == '\n')
6079 lineno++, linecharno = charno;
6081 if (name[0] == '\0')
6083 else /* make a named tag */
6084 name = substitute (buffer, rp->name, &rp->regs);
6085 if (rp->force_explicit_name)
6086 /* Force explicit tag name, if a name is there. */
6087 pfnote (name, TRUE, buffer + linecharno,
6088 charno - linecharno + 1, lineno, linecharno);
6090 make_tag (name, strlen (name), TRUE, buffer + linecharno,
6091 charno - linecharno + 1, lineno, linecharno);
6103 register int len = 0;
6105 while (*cp != '\0' && lowcase (*cp) == lowcase (dbp[len]))
6107 if (*cp == '\0' && !intoken (dbp[len]))
6116 get_tag (bp, namepp)
6120 register char *cp = bp;
6124 /* Go till you get to white space or a syntactic break */
6125 for (cp = bp + 1; !notinname (*cp); cp++)
6127 make_tag (bp, cp - bp, TRUE,
6128 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
6132 *namepp = savenstr (bp, cp - bp);
6136 * Read a line of text from `stream' into `lbp', excluding the
6137 * newline or CR-NL, if any. Return the number of characters read from
6138 * `stream', which is the length of the line including the newline.
6140 * On DOS or Windows we do not count the CR character, if any before the
6141 * NL, in the returned length; this mirrors the behavior of Emacs on those
6142 * platforms (for text files, it translates CR-NL to NL as it reads in the
6145 * If multi-line regular expressions are requested, each line read is
6146 * appended to `filebuf'.
6149 readline_internal (lbp, stream)
6151 register FILE *stream;
6153 char *buffer = lbp->buffer;
6154 register char *p = lbp->buffer;
6155 register char *pend;
6158 pend = p + lbp->size; /* Separate to avoid 386/IX compiler bug. */
6162 register int c = getc (stream);
6165 /* We're at the end of linebuffer: expand it. */
6167 xrnew (buffer, lbp->size, char);
6168 p += buffer - lbp->buffer;
6169 pend = buffer + lbp->size;
6170 lbp->buffer = buffer;
6180 if (p > buffer && p[-1] == '\r')
6194 lbp->len = p - buffer;
6196 if (need_filebuf /* we need filebuf for multi-line regexps */
6197 && chars_deleted > 0) /* not at EOF */
6199 while (filebuf.size <= filebuf.len + lbp->len + 1) /* +1 for \n */
6201 /* Expand filebuf. */
6203 xrnew (filebuf.buffer, filebuf.size, char);
6205 strncpy (filebuf.buffer + filebuf.len, lbp->buffer, lbp->len);
6206 filebuf.len += lbp->len;
6207 filebuf.buffer[filebuf.len++] = '\n';
6208 filebuf.buffer[filebuf.len] = '\0';
6211 return lbp->len + chars_deleted;
6215 * Like readline_internal, above, but in addition try to match the
6216 * input line against relevant regular expressions and manage #line
6220 readline (lbp, stream)
6226 linecharno = charno; /* update global char number of line start */
6227 result = readline_internal (lbp, stream); /* read line */
6228 lineno += 1; /* increment global line number */
6229 charno += result; /* increment global char number */
6231 /* Honour #line directives. */
6232 if (!no_line_directive)
6234 static bool discard_until_line_directive;
6236 /* Check whether this is a #line directive. */
6237 if (result > 12 && strneq (lbp->buffer, "#line ", 6))
6242 if (sscanf (lbp->buffer, "#line %u \"%n", &lno, &start) >= 1
6243 && start > 0) /* double quote character found */
6245 char *endp = lbp->buffer + start;
6247 while ((endp = etags_strchr (endp, '"')) != NULL
6248 && endp[-1] == '\\')
6251 /* Ok, this is a real #line directive. Let's deal with it. */
6253 char *taggedabsname; /* absolute name of original file */
6254 char *taggedfname; /* name of original file as given */
6255 char *name; /* temp var */
6257 discard_until_line_directive = FALSE; /* found it */
6258 name = lbp->buffer + start;
6260 canonicalize_filename (name);
6261 taggedabsname = absolute_filename (name, tagfiledir);
6262 if (filename_is_absolute (name)
6263 || filename_is_absolute (curfdp->infname))
6264 taggedfname = savestr (taggedabsname);
6266 taggedfname = relative_filename (taggedabsname,tagfiledir);
6268 if (streq (curfdp->taggedfname, taggedfname))
6269 /* The #line directive is only a line number change. We
6270 deal with this afterwards. */
6273 /* The tags following this #line directive should be
6274 attributed to taggedfname. In order to do this, set
6275 curfdp accordingly. */
6277 fdesc *fdp; /* file description pointer */
6279 /* Go look for a file description already set up for the
6280 file indicated in the #line directive. If there is
6281 one, use it from now until the next #line
6283 for (fdp = fdhead; fdp != NULL; fdp = fdp->next)
6284 if (streq (fdp->infname, curfdp->infname)
6285 && streq (fdp->taggedfname, taggedfname))
6286 /* If we remove the second test above (after the &&)
6287 then all entries pertaining to the same file are
6288 coalesced in the tags file. If we use it, then
6289 entries pertaining to the same file but generated
6290 from different files (via #line directives) will
6291 go into separate sections in the tags file. These
6292 alternatives look equivalent. The first one
6293 destroys some apparently useless information. */
6299 /* Else, if we already tagged the real file, skip all
6300 input lines until the next #line directive. */
6301 if (fdp == NULL) /* not found */
6302 for (fdp = fdhead; fdp != NULL; fdp = fdp->next)
6303 if (streq (fdp->infabsname, taggedabsname))
6305 discard_until_line_directive = TRUE;
6309 /* Else create a new file description and use that from
6310 now on, until the next #line directive. */
6311 if (fdp == NULL) /* not found */
6314 fdhead = xnew (1, fdesc);
6315 *fdhead = *curfdp; /* copy curr. file description */
6317 fdhead->infname = savestr (curfdp->infname);
6318 fdhead->infabsname = savestr (curfdp->infabsname);
6319 fdhead->infabsdir = savestr (curfdp->infabsdir);
6320 fdhead->taggedfname = taggedfname;
6321 fdhead->usecharno = FALSE;
6322 fdhead->prop = NULL;
6323 fdhead->written = FALSE;
6327 free (taggedabsname);
6329 readline (lbp, stream);
6331 } /* if a real #line directive */
6332 } /* if #line is followed by a a number */
6333 } /* if line begins with "#line " */
6335 /* If we are here, no #line directive was found. */
6336 if (discard_until_line_directive)
6340 /* Do a tail recursion on ourselves, thus discarding the contents
6341 of the line buffer. */
6342 readline (lbp, stream);
6346 discard_until_line_directive = FALSE;
6349 } /* if #line directives should be considered */
6356 /* Match against relevant regexps. */
6358 for (rp = p_head; rp != NULL; rp = rp->p_next)
6360 /* Only use generic regexps or those for the current language.
6361 Also do not use multiline regexps, which is the job of
6362 regex_tag_multiline. */
6363 if ((rp->lang != NULL && rp->lang != fdhead->lang)
6367 match = re_match (rp->pat, lbp->buffer, lbp->len, 0, &rp->regs);
6372 if (!rp->error_signaled)
6374 error ("regexp stack overflow while matching \"%s\"",
6376 rp->error_signaled = TRUE;
6383 /* Empty string matched. */
6384 if (!rp->error_signaled)
6386 error ("regexp matches the empty string: \"%s\"", rp->pattern);
6387 rp->error_signaled = TRUE;
6391 /* Match occurred. Construct a tag. */
6393 if (name[0] == '\0')
6395 else /* make a named tag */
6396 name = substitute (lbp->buffer, rp->name, &rp->regs);
6397 if (rp->force_explicit_name)
6398 /* Force explicit tag name, if a name is there. */
6399 pfnote (name, TRUE, lbp->buffer, match, lineno, linecharno);
6401 make_tag (name, strlen (name), TRUE,
6402 lbp->buffer, match, lineno, linecharno);
6411 * Return a pointer to a space of size strlen(cp)+1 allocated
6412 * with xnew where the string CP has been copied.
6418 return savenstr (cp, strlen (cp));
6422 * Return a pointer to a space of size LEN+1 allocated with xnew where
6423 * the string CP has been copied for at most the first LEN characters.
6432 dp = xnew (len + 1, char);
6433 strncpy (dp, cp, len);
6439 * Return the ptr in sp at which the character c last
6440 * appears; NULL if not found
6442 * Identical to POSIX strrchr, included for portability.
6445 etags_strrchr (sp, c)
6446 register const char *sp;
6449 register const char *r;
6461 * Return the ptr in sp at which the character c first
6462 * appears; NULL if not found
6464 * Identical to POSIX strchr, included for portability.
6467 etags_strchr (sp, c)
6468 register const char *sp;
6480 * Compare two strings, ignoring case for alphabetic characters.
6482 * Same as BSD's strcasecmp, included for portability.
6485 etags_strcasecmp (s1, s2)
6486 register const char *s1;
6487 register const char *s2;
6490 && (ISALPHA (*s1) && ISALPHA (*s2)
6491 ? lowcase (*s1) == lowcase (*s2)
6495 return (ISALPHA (*s1) && ISALPHA (*s2)
6496 ? lowcase (*s1) - lowcase (*s2)
6501 * Compare two strings, ignoring case for alphabetic characters.
6502 * Stop after a given number of characters
6504 * Same as BSD's strncasecmp, included for portability.
6507 etags_strncasecmp (s1, s2, n)
6508 register const char *s1;
6509 register const char *s2;
6512 while (*s1 != '\0' && n-- > 0
6513 && (ISALPHA (*s1) && ISALPHA (*s2)
6514 ? lowcase (*s1) == lowcase (*s2)
6521 return (ISALPHA (*s1) && ISALPHA (*s2)
6522 ? lowcase (*s1) - lowcase (*s2)
6526 /* Skip spaces (end of string is not space), return new pointer. */
6531 while (iswhite (*cp))
6536 /* Skip non spaces, except end of string, return new pointer. */
6538 skip_non_spaces (cp)
6541 while (*cp != '\0' && !iswhite (*cp))
6546 /* Print error message and exit. */
6552 exit (EXIT_FAILURE);
6560 exit (EXIT_FAILURE);
6564 suggest_asking_for_help ()
6566 fprintf (stderr, "\tTry `%s %s' for a complete list of options.\n",
6567 progname, NO_LONG_OPTIONS ? "-h" : "--help");
6568 exit (EXIT_FAILURE);
6571 /* Print error message. `s1' is printf control string, `s2' is arg for it. */
6574 const char *s1, *s2;
6576 fprintf (stderr, "%s: ", progname);
6577 fprintf (stderr, s1, s2);
6578 fprintf (stderr, "\n");
6581 /* Return a newly-allocated string whose contents
6582 concatenate those of s1, s2, s3. */
6587 int len1 = strlen (s1), len2 = strlen (s2), len3 = strlen (s3);
6588 char *result = xnew (len1 + len2 + len3 + 1, char);
6590 strcpy (result, s1);
6591 strcpy (result + len1, s2);
6592 strcpy (result + len1 + len2, s3);
6593 result[len1 + len2 + len3] = '\0';
6599 /* Does the same work as the system V getcwd, but does not need to
6600 guess the buffer size in advance. */
6606 char *path = xnew (bufsize, char);
6608 while (getcwd (path, bufsize) == NULL)
6610 if (errno != ERANGE)
6614 path = xnew (bufsize, char);
6617 canonicalize_filename (path);
6620 #else /* not HAVE_GETCWD */
6624 linebuffer_init (&path);
6625 pipe = (FILE *) popen ("pwd 2>/dev/null", "r");
6626 if (pipe == NULL || readline_internal (&path, pipe) == 0)
6631 #endif /* not HAVE_GETCWD */
6634 /* Return a newly allocated string containing the file name of FILE
6635 relative to the absolute directory DIR (which should end with a slash). */
6637 relative_filename (file, dir)
6640 char *fp, *dp, *afn, *res;
6643 /* Find the common root of file and dir (with a trailing slash). */
6644 afn = absolute_filename (file, cwd);
6647 while (*fp++ == *dp++)
6649 fp--, dp--; /* back to the first differing char */
6650 do /* look at the equal chars until '/' */
6654 /* Build a sequence of "../" strings for the resulting relative file name. */
6656 while ((dp = etags_strchr (dp + 1, '/')) != NULL)
6658 res = xnew (3*i + strlen (fp + 1) + 1, char);
6661 strcat (res, "../");
6663 /* Add the file name relative to the common root of file and dir. */
6664 strcat (res, fp + 1);
6670 /* Return a newly allocated string containing the absolute file name
6671 of FILE given DIR (which should end with a slash). */
6673 absolute_filename (file, dir)
6676 char *slashp, *cp, *res;
6678 if (filename_is_absolute (file))
6679 res = savestr (file);
6681 res = concat (dir, file, "");
6683 /* Delete the "/dirname/.." and "/." substrings. */
6684 slashp = etags_strchr (res, '/');
6685 while (slashp != NULL && slashp[0] != '\0')
6687 if (slashp[1] == '.')
6689 if (slashp[2] == '.'
6690 && (slashp[3] == '/' || slashp[3] == '\0'))
6695 while (cp >= res && !filename_is_absolute (cp));
6697 cp = slashp; /* the absolute name begins with "/.." */
6698 strcpy (cp, slashp + 3);
6702 else if (slashp[2] == '/' || slashp[2] == '\0')
6704 strcpy (slashp, slashp + 2);
6709 slashp = etags_strchr (slashp + 1, '/');
6712 if (res[0] == '\0') /* just a safety net: should never happen */
6715 return savestr ("/");
6721 /* Return a newly allocated string containing the absolute
6722 file name of dir where FILE resides given DIR (which should
6723 end with a slash). */
6725 absolute_dirname (file, dir)
6731 slashp = etags_strrchr (file, '/');
6733 return savestr (dir);
6736 res = absolute_filename (file, dir);
6742 /* Whether the argument string is an absolute file name. The argument
6743 string must have been canonicalized with canonicalize_filename. */
6745 filename_is_absolute (fn)
6748 return (fn[0] == '/');
6751 /* Upcase DOS drive letter and collapse separators into single slashes.
6754 canonicalize_filename (fn)
6760 /* Collapse multiple separators into a single slash. */
6761 for (cp = fn; *cp != '\0'; cp++, fn++)
6765 while (cp[1] == sep)
6774 /* Initialize a linebuffer for use. */
6776 linebuffer_init (lbp)
6779 lbp->size = (DEBUG) ? 3 : 200;
6780 lbp->buffer = xnew (lbp->size, char);
6781 lbp->buffer[0] = '\0';
6785 /* Set the minimum size of a string contained in a linebuffer. */
6787 linebuffer_setlen (lbp, toksize)
6791 while (lbp->size <= toksize)
6794 xrnew (lbp->buffer, lbp->size, char);
6799 /* Like malloc but get fatal error if memory is exhausted. */
6804 PTR result = (PTR) malloc (size);
6806 fatal ("virtual memory exhausted", (char *)NULL);
6811 xrealloc (ptr, size)
6815 PTR result = (PTR) realloc (ptr, size);
6817 fatal ("virtual memory exhausted", (char *)NULL);
6823 * indent-tabs-mode: t
6826 * c-font-lock-extra-types: ("FILE" "bool" "language" "linebuffer" "fdesc" "node" "regexp")
6830 /* etags.c ends here */