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) */
1333 for (i = 0; i < current_arg; ++i)
1335 switch (argbuffer[i].arg_type)
1344 continue; /* the for loop */
1346 len = snprintf (cmd, sizeof(cmd),
1347 "mv %s OTAGS;fgrep -v '\t%s\t' OTAGS >%s;rm OTAGS",
1348 tagfile, argbuffer[i].what, tagfile);
1349 if (len >= 0 && (size_t)len < sizeof(cmd))
1350 fatal ("failed to build shell command line", (char *)NULL);
1351 if (system (cmd) != EXIT_SUCCESS)
1352 fatal ("failed to execute shell command", (char *)NULL);
1354 append_to_tagfile = TRUE;
1357 tagf = fopen (tagfile, append_to_tagfile ? "a" : "w");
1360 put_entries (nodehead); /* write all the tags (CTAGS) */
1361 free_tree (nodehead);
1363 if (fclose (tagf) == EOF)
1367 if (append_to_tagfile || update)
1369 char cmd[2*BUFSIZ+20];
1370 /* Maybe these should be used:
1371 setenv ("LC_COLLATE", "C", 1);
1372 setenv ("LC_ALL", "C", 1); */
1373 int len = snprintf (cmd, sizeof(cmd),
1374 "sort -u -o %.*s %.*s",
1377 if (len >= 0 && (size_t)len < sizeof(cmd))
1378 fatal("failed to build sort shell command line",
1380 exit (system (cmd));
1382 return EXIT_SUCCESS;
1387 * Return a compressor given the file name. If EXTPTR is non-zero,
1388 * return a pointer into FILE where the compressor-specific
1389 * extension begins. If no compressor is found, NULL is returned
1390 * and EXTPTR is not significant.
1391 * Idea by Vladimir Alexiev <vladimir@cs.ualberta.ca> (1998)
1394 get_compressor_from_suffix (file, extptr)
1399 char *slash, *suffix;
1401 /* File has been processed by canonicalize_filename,
1402 so we don't need to consider backslashes on DOS_NT. */
1403 slash = etags_strrchr (file, '/');
1404 suffix = etags_strrchr (file, '.');
1405 if (suffix == NULL || suffix < slash)
1410 /* Let those poor souls who live with DOS 8+3 file name limits get
1411 some solace by treating foo.cgz as if it were foo.c.gz, etc.
1412 Only the first do loop is run if not MSDOS */
1415 for (compr = compressors; compr->suffix != NULL; compr++)
1416 if (streq (compr->suffix, suffix))
1419 break; /* do it only once: not really a loop */
1422 } while (*suffix != '\0');
1429 * Return a language given the name.
1432 get_language_from_langname (name)
1438 error ("empty language name", (char *)NULL);
1441 for (lang = lang_names; lang->name != NULL; lang++)
1442 if (streq (name, lang->name))
1444 error ("unknown language \"%s\"", name);
1452 * Return a language given the interpreter name.
1455 get_language_from_interpreter (interpreter)
1461 if (interpreter == NULL)
1463 for (lang = lang_names; lang->name != NULL; lang++)
1464 if (lang->interpreters != NULL)
1465 for (iname = lang->interpreters; *iname != NULL; iname++)
1466 if (streq (*iname, interpreter))
1475 * Return a language given the file name.
1478 get_language_from_filename (file, case_sensitive)
1480 bool case_sensitive;
1483 char **name, **ext, *suffix;
1485 /* Try whole file name first. */
1486 for (lang = lang_names; lang->name != NULL; lang++)
1487 if (lang->filenames != NULL)
1488 for (name = lang->filenames; *name != NULL; name++)
1489 if ((case_sensitive)
1490 ? streq (*name, file)
1491 : strcaseeq (*name, file))
1494 /* If not found, try suffix after last dot. */
1495 suffix = etags_strrchr (file, '.');
1499 for (lang = lang_names; lang->name != NULL; lang++)
1500 if (lang->suffixes != NULL)
1501 for (ext = lang->suffixes; *ext != NULL; ext++)
1502 if ((case_sensitive)
1503 ? streq (*ext, suffix)
1504 : strcaseeq (*ext, suffix))
1511 * This routine is called on each file argument.
1514 process_file_name (file, lang)
1518 struct stat stat_buf;
1521 compressor *compr = NULL;
1522 char *compressed_name = NULL,
1523 *uncompressed_name = NULL;
1528 canonicalize_filename (file);
1529 if (streq (file, tagfile) && !streq (tagfile, "-"))
1531 error ("skipping inclusion of %s in self.", file);
1534 if ( get_compressor_from_suffix (file, &ext) == NULL)
1536 real_name = uncompressed_name = savestr (file);
1540 real_name = compressed_name = savestr (file);
1541 uncompressed_name = savenstr (file, ext - file);
1544 /* If the canonicalized uncompressed name
1545 has already been dealt with, skip it silently. */
1546 for (fdp = fdhead; fdp != NULL; fdp = fdp->next)
1548 assert (fdp->infname != NULL);
1549 if (streq (uncompressed_name, fdp->infname))
1553 compr = compressors;
1555 /* First try to open ... */
1556 if (real_name == compressed_name)
1558 char *cmd = concat (compr->command, " ", real_name);
1559 inf = (FILE *) popen (cmd, "r");
1563 inf = fopen (real_name, "r");
1564 if ( inf != NULL ) {
1565 /* Open was successfull, check it is a regular file */
1566 if (stat (real_name, &stat_buf) == 0 &&
1567 !S_ISREG (stat_buf.st_mode))
1569 error ("skipping %s: it is not a regular file.",
1575 /* Not else from previous if because inner check may reset inf
1576 to NULL, at which case we will want to try the next?
1577 compressed filename... */
1578 if ( inf == NULL ) {
1579 /* Reset real_name and try with a different name. */
1580 free(compressed_name);
1582 if (compressed_name != NULL)
1583 /* try with the given suffix */
1585 compressed_name = NULL;
1586 real_name = uncompressed_name;
1588 else if ( compr && compr->suffix != NULL )
1589 /* try all possible suffixes */
1591 compressed_name = concat (file, ".", compr->suffix);
1592 real_name = compressed_name;
1596 } while( inf == NULL && real_name != NULL);
1603 process_file (inf, uncompressed_name, lang);
1605 if (real_name == compressed_name)
1606 retval = pclose (inf);
1608 retval = fclose (inf);
1613 free (compressed_name);
1614 free (uncompressed_name);
1621 process_file (fh, fn, lang)
1626 static const fdesc emptyfdesc;
1629 /* Create a new input file description entry. */
1630 fdp = xnew (1, fdesc);
1633 fdp->infname = savestr (fn);
1635 fdp->infabsname = absolute_filename (fn, cwd);
1636 fdp->infabsdir = absolute_dirname (fn, cwd);
1637 if (filename_is_absolute (fn))
1639 /* An absolute file name. Canonicalize it. */
1640 fdp->taggedfname = absolute_filename (fn, NULL);
1644 /* A file name relative to cwd. Make it relative
1645 to the directory of the tags file. */
1646 fdp->taggedfname = relative_filename (fn, tagfiledir);
1648 fdp->usecharno = TRUE; /* use char position when making tags */
1650 fdp->written = FALSE; /* not written on tags file yet */
1653 curfdp = fdhead; /* the current file description */
1657 /* If not Ctags, and if this is not metasource and if it contained no #line
1658 directives, we can write the tags and free all nodes pointing to
1661 && curfdp->usecharno /* no #line directives in this file */
1662 && !curfdp->lang->metasource)
1666 /* Look for the head of the sublist relative to this file. See add_node
1667 for the structure of the node tree. */
1669 for (np = nodehead; np != NULL; prev = np, np = np->left)
1670 if (np->fdp == curfdp)
1673 /* If we generated tags for this file, write and delete them. */
1676 /* This is the head of the last sublist, if any. The following
1677 instructions depend on this being true. */
1678 assert (np->left == NULL);
1680 assert (fdhead == curfdp);
1681 assert (last_node->fdp == curfdp);
1682 put_entries (np); /* write tags for file curfdp->taggedfname */
1683 free_tree (np); /* remove the written nodes */
1685 nodehead = NULL; /* no nodes left */
1687 prev->left = NULL; /* delete the pointer to the sublist */
1693 * This routine sets up the boolean pseudo-functions which work
1694 * by setting boolean flags dependent upon the corresponding character.
1695 * Every char which is NOT in that string is not a white char. Therefore,
1696 * all of the array "_wht" is set to FALSE, and then the elements
1697 * subscripted by the chars in "white" are set to TRUE. Thus "_wht"
1698 * of a char is TRUE if it is the string "white", else FALSE.
1706 for (i = 0; i < CHARS; i++)
1707 iswhite(i) = notinname(i) = begtoken(i) = intoken(i) = endtoken(i) = FALSE;
1708 for (sp = white; *sp != '\0'; sp++) iswhite (*sp) = TRUE;
1709 for (sp = nonam; *sp != '\0'; sp++) notinname (*sp) = TRUE;
1710 notinname('\0') = notinname('\n');
1711 for (sp = begtk; *sp != '\0'; sp++) begtoken (*sp) = TRUE;
1712 begtoken('\0') = begtoken('\n');
1713 for (sp = midtk; *sp != '\0'; sp++) intoken (*sp) = TRUE;
1714 intoken('\0') = intoken('\n');
1715 for (sp = endtk; *sp != '\0'; sp++) endtoken (*sp) = TRUE;
1716 endtoken('\0') = endtoken('\n');
1720 * This routine opens the specified file and calls the function
1721 * which finds the function and type definitions.
1728 language *lang = curfdp->lang;
1729 Lang_function *parser = NULL;
1731 /* If user specified a language, use it. */
1732 if (lang != NULL && lang->function != NULL)
1734 parser = lang->function;
1737 /* Else try to guess the language given the file name. */
1740 lang = get_language_from_filename (curfdp->infname, TRUE);
1741 if (lang != NULL && lang->function != NULL)
1743 curfdp->lang = lang;
1744 parser = lang->function;
1748 /* Else look for sharp-bang as the first two characters. */
1750 && readline_internal (&lb, inf) > 0
1752 && lb.buffer[0] == '#'
1753 && lb.buffer[1] == '!')
1757 /* Set lp to point at the first char after the last slash in the
1758 line or, if no slashes, at the first nonblank. Then set cp to
1759 the first successive blank and terminate the string. */
1760 lp = etags_strrchr (lb.buffer+2, '/');
1764 lp = skip_spaces (lb.buffer + 2);
1765 cp = skip_non_spaces (lp);
1768 if (strlen (lp) > 0)
1770 lang = get_language_from_interpreter (lp);
1771 if (lang != NULL && lang->function != NULL)
1773 curfdp->lang = lang;
1774 parser = lang->function;
1779 /* We rewind here, even if inf may be a pipe. We fail if the
1780 length of the first line is longer than the pipe block size,
1781 which is unlikely. */
1784 /* Else try to guess the language given the case insensitive file name. */
1787 lang = get_language_from_filename (curfdp->infname, FALSE);
1788 if (lang != NULL && lang->function != NULL)
1790 curfdp->lang = lang;
1791 parser = lang->function;
1795 /* Else try Fortran or C. */
1798 node *old_last_node = last_node;
1800 curfdp->lang = get_language_from_langname ("fortran");
1803 if (old_last_node == last_node)
1804 /* No Fortran entries found. Try C. */
1806 /* We do not tag if rewind fails.
1807 Only the file name will be recorded in the tags file. */
1809 curfdp->lang = get_language_from_langname (cplusplus ? "c++" : "c");
1815 if (!no_line_directive
1816 && curfdp->lang != NULL && curfdp->lang->metasource)
1817 /* It may be that this is a bingo.y file, and we already parsed a bingo.c
1818 file, or anyway we parsed a file that is automatically generated from
1819 this one. If this is the case, the bingo.c file contained #line
1820 directives that generated tags pointing to this file. Let's delete
1821 them all before parsing this file, which is the real source. */
1823 fdesc **fdpp = &fdhead;
1824 while (*fdpp != NULL)
1826 && streq ((*fdpp)->taggedfname, curfdp->taggedfname))
1827 /* We found one of those! We must delete both the file description
1828 and all tags referring to it. */
1830 fdesc *badfdp = *fdpp;
1832 /* Delete the tags referring to badfdp->taggedfname
1833 that were obtained from badfdp->infname. */
1834 invalidate_nodes (badfdp, &nodehead);
1836 *fdpp = badfdp->next; /* remove the bad description from the list */
1837 free_fdesc (badfdp);
1840 fdpp = &(*fdpp)->next; /* advance the list pointer */
1843 assert (parser != NULL);
1845 /* Generic initialisations before reading from file. */
1846 linebuffer_setlen (&filebuf, 0); /* reset the file buffer */
1848 /* Generic initialisations before parsing file with readline. */
1849 lineno = 0; /* reset global line number */
1850 charno = 0; /* reset global char number */
1851 linecharno = 0; /* reset global char number of line start */
1855 regex_tag_multiline ();
1860 * Check whether an implicitly named tag should be created,
1861 * then call `pfnote'.
1862 * NAME is a string that is internally copied by this function.
1864 * TAGS format specification
1865 * Idea by Sam Kendall <kendall@mv.mv.com> (1997)
1866 * The following is explained in some more detail in etc/ETAGS.EBNF.
1868 * make_tag creates tags with "implicit tag names" (unnamed tags)
1869 * if the following are all true, assuming NONAM=" \f\t\n\r()=,;":
1870 * 1. NAME does not contain any of the characters in NONAM;
1871 * 2. LINESTART contains name as either a rightmost, or rightmost but
1872 * one character, substring;
1873 * 3. the character, if any, immediately before NAME in LINESTART must
1874 * be a character in NONAM;
1875 * 4. the character, if any, immediately after NAME in LINESTART must
1876 * also be a character in NONAM.
1878 * The implementation uses the notinname() macro, which recognises the
1879 * characters stored in the string `nonam'.
1880 * etags.el needs to use the same characters that are in NONAM.
1883 make_tag (name, namelen, is_func, linestart, linelen, lno, cno)
1884 char *name; /* tag name, or NULL if unnamed */
1885 int namelen; /* tag length */
1886 bool is_func; /* tag is a function */
1887 char *linestart; /* start of the line where tag is */
1888 int linelen; /* length of the line where tag is */
1889 int lno; /* line number */
1890 long cno; /* character number */
1892 bool named = (name != NULL && namelen > 0);
1894 if (!CTAGS && named) /* maybe set named to false */
1895 /* Let's try to make an implicit tag name, that is, create an unnamed tag
1896 such that etags.el can guess a name from it. */
1899 register char *cp = name;
1901 for (i = 0; i < namelen; i++)
1902 if (notinname (*cp++))
1904 if (i == namelen) /* rule #1 */
1906 cp = linestart + linelen - namelen;
1907 if (notinname (linestart[linelen-1]))
1908 cp -= 1; /* rule #4 */
1909 if (cp >= linestart /* rule #2 */
1911 || notinname (cp[-1])) /* rule #3 */
1912 && strneq (name, cp, namelen)) /* rule #2 */
1913 named = FALSE; /* use implicit tag name */
1918 name = savenstr (name, namelen);
1921 pfnote (name, is_func, linestart, linelen, lno, cno);
1926 pfnote (name, is_func, linestart, linelen, lno, cno)
1927 char *name; /* tag name, or NULL if unnamed */
1928 bool is_func; /* tag is a function */
1929 char *linestart; /* start of the line where tag is */
1930 int linelen; /* length of the line where tag is */
1931 int lno; /* line number */
1932 long cno; /* character number */
1936 assert (name == NULL || name[0] != '\0');
1937 if (CTAGS && name == NULL)
1940 np = xnew (1, node);
1942 /* If ctags mode, change name "main" to M<thisfilename>. */
1943 if (CTAGS && !cxref_style && streq (name, "main"))
1945 register char *fp = etags_strrchr (curfdp->taggedfname, '/');
1946 np->name = concat ("M", fp == NULL ? curfdp->taggedfname : fp + 1, "");
1947 fp = etags_strrchr (np->name, '.');
1948 if (fp != NULL && fp[1] != '\0' && fp[2] == '\0')
1954 np->been_warned = FALSE;
1956 np->is_func = is_func;
1958 if (np->fdp->usecharno)
1959 /* Our char numbers are 0-base, because of C language tradition?
1960 ctags compatibility? old versions compatibility? I don't know.
1961 Anyway, since emacs's are 1-base we expect etags.el to take care
1962 of the difference. If we wanted to have 1-based numbers, we would
1963 uncomment the +1 below. */
1964 np->cno = cno /* + 1 */ ;
1966 np->cno = invalidcharno;
1967 np->left = np->right = NULL;
1968 if (CTAGS && !cxref_style)
1970 if (strlen (linestart) < 50)
1971 np->regex = concat (linestart, "$", "");
1973 np->regex = savenstr (linestart, 50);
1976 np->regex = savenstr (linestart, linelen);
1978 add_node (np, &nodehead);
1983 * recurse on left children, iterate on right children.
1991 register node *node_right = np->right;
1992 free_tree (np->left);
2002 * delete a file description
2006 register fdesc *fdp;
2008 free (fdp->infname);
2009 free (fdp->infabsname);
2010 free (fdp->infabsdir);
2011 free (fdp->taggedfname);
2018 * Adds a node to the tree of nodes. In etags mode, sort by file
2019 * name. In ctags mode, sort by tag name. Make no attempt at
2022 * add_node is the only function allowed to add nodes, so it can
2026 add_node (np, cur_node_p)
2027 node *np, **cur_node_p;
2030 register node *cur_node = *cur_node_p;
2032 if (cur_node == NULL)
2042 /* For each file name, tags are in a linked sublist on the right
2043 pointer. The first tags of different files are a linked list
2044 on the left pointer. last_node points to the end of the last
2046 if (last_node != NULL && last_node->fdp == np->fdp)
2048 /* Let's use the same sublist as the last added node. */
2049 assert (last_node->right == NULL);
2050 last_node->right = np;
2053 else if (cur_node->fdp == np->fdp)
2055 /* Scanning the list we found the head of a sublist which is
2056 good for us. Let's scan this sublist. */
2057 add_node (np, &cur_node->right);
2060 /* The head of this sublist is not good for us. Let's try the
2062 add_node (np, &cur_node->left);
2063 } /* if ETAGS mode */
2068 dif = strcmp (np->name, cur_node->name);
2071 * If this tag name matches an existing one, then
2072 * do not add the node, but maybe print a warning.
2074 if (no_duplicates && !dif)
2076 if (np->fdp == cur_node->fdp)
2080 fprintf (stderr, "Duplicate entry in file %s, line %d: %s\n",
2081 np->fdp->infname, lineno, np->name);
2082 fprintf (stderr, "Second entry ignored\n");
2085 else if (!cur_node->been_warned && !no_warnings)
2089 "Duplicate entry in files %s and %s: %s (Warning only)\n",
2090 np->fdp->infname, cur_node->fdp->infname, np->name);
2091 cur_node->been_warned = TRUE;
2096 /* Actually add the node */
2097 add_node (np, dif < 0 ? &cur_node->left : &cur_node->right);
2098 } /* if CTAGS mode */
2102 * invalidate_nodes ()
2103 * Scan the node tree and invalidate all nodes pointing to the
2104 * given file description (CTAGS case) or free them (ETAGS case).
2107 invalidate_nodes (badfdp, npp)
2118 if (np->left != NULL)
2119 invalidate_nodes (badfdp, &np->left);
2120 if (np->fdp == badfdp)
2122 if (np->right != NULL)
2123 invalidate_nodes (badfdp, &np->right);
2127 assert (np->fdp != NULL);
2128 if (np->fdp == badfdp)
2130 *npp = np->left; /* detach the sublist from the list */
2131 np->left = NULL; /* isolate it */
2132 free_tree (np); /* free it */
2133 invalidate_nodes (badfdp, npp);
2136 invalidate_nodes (badfdp, &np->left);
2141 static int total_size_of_entries __P((node *));
2142 static int number_len __P((long));
2144 /* Length of a non-negative number's decimal representation. */
2150 while ((num /= 10) > 0)
2156 * Return total number of characters that put_entries will output for
2157 * the nodes in the linked list at the right of the specified node.
2158 * This count is irrelevant with etags.el since emacs 19.34 at least,
2159 * but is still supplied for backward compatibility.
2162 total_size_of_entries (np)
2165 register int total = 0;
2167 for (; np != NULL; np = np->right)
2170 total += strlen (np->regex) + 1; /* pat\177 */
2171 if (np->name != NULL)
2172 total += strlen (np->name) + 1; /* name\001 */
2173 total += number_len ((long) np->lno) + 1; /* lno, */
2174 if (np->cno != invalidcharno) /* cno */
2175 total += number_len (np->cno);
2176 total += 1; /* newline */
2187 static fdesc *fdp = NULL;
2192 /* Output subentries that precede this one */
2194 put_entries (np->left);
2196 /* Output this entry */
2205 fprintf (tagf, "\f\n%s,%d\n",
2206 fdp->taggedfname, total_size_of_entries (np));
2207 fdp->written = TRUE;
2209 fputs (np->regex, tagf);
2210 fputc ('\177', tagf);
2211 if (np->name != NULL)
2213 fputs (np->name, tagf);
2214 fputc ('\001', tagf);
2216 fprintf (tagf, "%d,", np->lno);
2217 if (np->cno != invalidcharno)
2218 fprintf (tagf, "%ld", np->cno);
2224 if (np->name == NULL)
2225 error ("internal error: NULL name in ctags mode.", (char *)NULL);
2230 fprintf (stdout, "%s %s %d\n",
2231 np->name, np->fdp->taggedfname, (np->lno + 63) / 64);
2233 fprintf (stdout, "%-16s %3d %-16s %s\n",
2234 np->name, np->lno, np->fdp->taggedfname, np->regex);
2238 fprintf (tagf, "%s\t%s\t", np->name, np->fdp->taggedfname);
2241 { /* function or #define macro with args */
2242 putc (searchar, tagf);
2245 for (sp = np->regex; *sp; sp++)
2247 if (*sp == '\\' || *sp == searchar)
2251 putc (searchar, tagf);
2254 { /* anything else; text pattern inadequate */
2255 fprintf (tagf, "%d", np->lno);
2260 } /* if this node contains a valid tag */
2262 /* Output subentries that follow this one */
2263 put_entries (np->right);
2265 put_entries (np->left);
2270 #define C_EXT 0x00fff /* C extensions */
2271 #define C_PLAIN 0x00000 /* C */
2272 #define C_PLPL 0x00001 /* C++ */
2273 #define C_STAR 0x00003 /* C* */
2274 #define C_JAVA 0x00005 /* JAVA */
2275 #define C_AUTO 0x01000 /* C, but switch to C++ if `class' is met */
2276 #define YACC 0x10000 /* yacc file */
2279 * The C symbol tables.
2284 st_C_objprot, st_C_objimpl, st_C_objend,
2286 st_C_ignore, st_C_attribute,
2289 st_C_class, st_C_template,
2290 st_C_struct, st_C_extern, st_C_enum, st_C_define, st_C_typedef
2293 static unsigned int hash __P((const char *, unsigned int));
2294 static struct C_stab_entry * in_word_set __P((const char *, unsigned int));
2295 static enum sym_type C_symtype __P((char *, int, int));
2297 /* Feed stuff between (but not including) %[ and %] lines to:
2303 struct C_stab_entry { char *name; int c_ext; enum sym_type type; }
2307 while, 0, st_C_ignore
2308 switch, 0, st_C_ignore
2309 return, 0, st_C_ignore
2310 __attribute__, 0, st_C_attribute
2311 GTY, 0, st_C_attribute
2312 @interface, 0, st_C_objprot
2313 @protocol, 0, st_C_objprot
2314 @implementation, 0, st_C_objimpl
2315 @end, 0, st_C_objend
2316 import, (C_JAVA & ~C_PLPL), st_C_ignore
2317 package, (C_JAVA & ~C_PLPL), st_C_ignore
2318 friend, C_PLPL, st_C_ignore
2319 extends, (C_JAVA & ~C_PLPL), st_C_javastruct
2320 implements, (C_JAVA & ~C_PLPL), st_C_javastruct
2321 interface, (C_JAVA & ~C_PLPL), st_C_struct
2322 class, 0, st_C_class
2323 namespace, C_PLPL, st_C_struct
2324 domain, C_STAR, st_C_struct
2325 union, 0, st_C_struct
2326 struct, 0, st_C_struct
2327 extern, 0, st_C_extern
2329 typedef, 0, st_C_typedef
2330 define, 0, st_C_define
2331 undef, 0, st_C_define
2332 operator, C_PLPL, st_C_operator
2333 template, 0, st_C_template
2334 # DEFUN used in emacs, the next three used in glibc (SYSCALL only for mach).
2335 DEFUN, 0, st_C_gnumacro
2336 SYSCALL, 0, st_C_gnumacro
2337 ENTRY, 0, st_C_gnumacro
2338 PSEUDO, 0, st_C_gnumacro
2339 # These are defined inside C functions, so currently they are not met.
2340 # EXFUN used in glibc, DEFVAR_* in emacs.
2341 #EXFUN, 0, st_C_gnumacro
2342 #DEFVAR_, 0, st_C_gnumacro
2344 and replace lines between %< and %> with its output, then:
2345 - remove the #if characterset check
2346 - make in_word_set static and not inline. */
2348 /* C code produced by gperf version 3.0.1 */
2349 /* Command-line: gperf -m 5 */
2350 /* Computed positions: -k'2-3' */
2352 struct C_stab_entry { char *name; int c_ext; enum sym_type type; };
2353 /* maximum key range = 33, duplicates = 0 */
2364 register const char *str;
2365 register unsigned int len;
2367 static unsigned char asso_values[] =
2369 35, 35, 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, 3,
2376 26, 35, 35, 35, 35, 35, 35, 35, 27, 35,
2377 35, 35, 35, 24, 0, 35, 35, 35, 35, 0,
2378 35, 35, 35, 35, 35, 1, 35, 16, 35, 6,
2379 23, 0, 0, 35, 22, 0, 35, 35, 5, 0,
2380 0, 15, 1, 35, 6, 35, 8, 19, 35, 16,
2381 4, 5, 35, 35, 35, 35, 35, 35, 35, 35,
2382 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2383 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2384 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2385 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2386 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2387 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2388 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2389 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2390 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2391 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2392 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2393 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2394 35, 35, 35, 35, 35, 35
2396 register int hval = len;
2401 hval += asso_values[(unsigned char)str[2]];
2404 hval += asso_values[(unsigned char)str[1]];
2410 static struct C_stab_entry *
2411 in_word_set (str, len)
2412 register const char *str;
2413 register unsigned int len;
2417 TOTAL_KEYWORDS = 33,
2418 MIN_WORD_LENGTH = 2,
2419 MAX_WORD_LENGTH = 15,
2424 static struct C_stab_entry wordlist[] =
2427 {"if", 0, st_C_ignore},
2428 {"GTY", 0, st_C_attribute},
2429 {"@end", 0, st_C_objend},
2430 {"union", 0, st_C_struct},
2431 {"define", 0, st_C_define},
2432 {"import", (C_JAVA & ~C_PLPL), st_C_ignore},
2433 {"template", 0, st_C_template},
2434 {"operator", C_PLPL, st_C_operator},
2435 {"@interface", 0, st_C_objprot},
2436 {"implements", (C_JAVA & ~C_PLPL), st_C_javastruct},
2437 {"friend", C_PLPL, st_C_ignore},
2438 {"typedef", 0, st_C_typedef},
2439 {"return", 0, st_C_ignore},
2440 {"@implementation",0, st_C_objimpl},
2441 {"@protocol", 0, st_C_objprot},
2442 {"interface", (C_JAVA & ~C_PLPL), st_C_struct},
2443 {"extern", 0, st_C_extern},
2444 {"extends", (C_JAVA & ~C_PLPL), st_C_javastruct},
2445 {"struct", 0, st_C_struct},
2446 {"domain", C_STAR, st_C_struct},
2447 {"switch", 0, st_C_ignore},
2448 {"enum", 0, st_C_enum},
2449 {"for", 0, st_C_ignore},
2450 {"namespace", C_PLPL, st_C_struct},
2451 {"class", 0, st_C_class},
2452 {"while", 0, st_C_ignore},
2453 {"undef", 0, st_C_define},
2454 {"package", (C_JAVA & ~C_PLPL), st_C_ignore},
2455 {"__attribute__", 0, st_C_attribute},
2456 {"SYSCALL", 0, st_C_gnumacro},
2457 {"ENTRY", 0, st_C_gnumacro},
2458 {"PSEUDO", 0, st_C_gnumacro},
2459 {"DEFUN", 0, st_C_gnumacro}
2462 if (len <= MAX_WORD_LENGTH && len >= MIN_WORD_LENGTH)
2464 register int key = hash (str, len);
2466 if (key <= MAX_HASH_VALUE && key >= 0)
2468 register const char *s = wordlist[key].name;
2470 if (*str == *s && !strncmp (str + 1, s + 1, len - 1) && s[len] == '\0')
2471 return &wordlist[key];
2478 static enum sym_type
2479 C_symtype (str, len, c_ext)
2484 register struct C_stab_entry *se = in_word_set (str, len);
2486 if (se == NULL || (se->c_ext && !(c_ext & se->c_ext)))
2493 * Ignoring __attribute__ ((list))
2495 static bool inattribute; /* looking at an __attribute__ construct */
2498 * C functions and variables are recognized using a simple
2499 * finite automaton. fvdef is its state variable.
2503 fvnone, /* nothing seen */
2504 fdefunkey, /* Emacs DEFUN keyword seen */
2505 fdefunname, /* Emacs DEFUN name seen */
2506 foperator, /* func: operator keyword seen (cplpl) */
2507 fvnameseen, /* function or variable name seen */
2508 fstartlist, /* func: just after open parenthesis */
2509 finlist, /* func: in parameter list */
2510 flistseen, /* func: after parameter list */
2511 fignore, /* func: before open brace */
2512 vignore /* var-like: ignore until ';' */
2515 static bool fvextern; /* func or var: extern keyword seen; */
2518 * typedefs are recognized using a simple finite automaton.
2519 * typdef is its state variable.
2523 tnone, /* nothing seen */
2524 tkeyseen, /* typedef keyword seen */
2525 ttypeseen, /* defined type seen */
2526 tinbody, /* inside typedef body */
2527 tend, /* just before typedef tag */
2528 tignore /* junk after typedef tag */
2532 * struct-like structures (enum, struct and union) are recognized
2533 * using another simple finite automaton. `structdef' is its state
2538 snone, /* nothing seen yet,
2539 or in struct body if bracelev > 0 */
2540 skeyseen, /* struct-like keyword seen */
2541 stagseen, /* struct-like tag seen */
2542 scolonseen /* colon seen after struct-like tag */
2546 * When objdef is different from onone, objtag is the name of the class.
2548 static char *objtag = "<uninited>";
2551 * Yet another little state machine to deal with preprocessor lines.
2555 dnone, /* nothing seen */
2556 dsharpseen, /* '#' seen as first char on line */
2557 ddefineseen, /* '#' and 'define' seen */
2558 dignorerest /* ignore rest of line */
2562 * State machine for Objective C protocols and implementations.
2563 * Idea by Tom R.Hageman <tom@basil.icce.rug.nl> (1995)
2567 onone, /* nothing seen */
2568 oprotocol, /* @interface or @protocol seen */
2569 oimplementation, /* @implementations seen */
2570 otagseen, /* class name seen */
2571 oparenseen, /* parenthesis before category seen */
2572 ocatseen, /* category name seen */
2573 oinbody, /* in @implementation body */
2574 omethodsign, /* in @implementation body, after +/- */
2575 omethodtag, /* after method name */
2576 omethodcolon, /* after method colon */
2577 omethodparm, /* after method parameter */
2578 oignore /* wait for @end */
2583 * Use this structure to keep info about the token read, and how it
2584 * should be tagged. Used by the make_C_tag function to build a tag.
2588 char *line; /* string containing the token */
2589 int offset; /* where the token starts in LINE */
2590 int length; /* token length */
2592 The previous members can be used to pass strings around for generic
2593 purposes. The following ones specifically refer to creating tags. In this
2594 case the token contained here is the pattern that will be used to create a
2597 bool valid; /* do not create a tag; the token should be
2598 invalidated whenever a state machine is
2599 reset prematurely */
2600 bool named; /* create a named tag */
2601 int lineno; /* source line number of tag */
2602 long linepos; /* source char number of tag */
2603 } token; /* latest token read */
2606 * Variables and functions for dealing with nested structures.
2607 * Idea by Mykola Dzyuba <mdzyuba@yahoo.com> (2001)
2609 static void pushclass_above __P((int, char *, int));
2610 static void popclass_above __P((int));
2611 static void write_classname __P((linebuffer *, char *qualifier));
2614 char **cname; /* nested class names */
2615 int *bracelev; /* nested class brace level */
2616 int nl; /* class nesting level (elements used) */
2617 int size; /* length of the array */
2618 } cstack; /* stack for nested declaration tags */
2619 /* Current struct nesting depth (namespace, class, struct, union, enum). */
2620 #define nestlev (cstack.nl)
2621 /* After struct keyword or in struct body, not inside a nested function. */
2622 #define instruct (structdef == snone && nestlev > 0 \
2623 && bracelev == cstack.bracelev[nestlev-1] + 1)
2626 pushclass_above (bracelev, str, len)
2633 popclass_above (bracelev);
2635 if (nl >= cstack.size)
2637 int size = cstack.size *= 2;
2638 xrnew (cstack.cname, size, char *);
2639 xrnew (cstack.bracelev, size, int);
2641 assert (nl == 0 || cstack.bracelev[nl-1] < bracelev);
2642 cstack.cname[nl] = (str == NULL) ? NULL : savenstr (str, len);
2643 cstack.bracelev[nl] = bracelev;
2648 popclass_above (bracelev)
2653 for (nl = cstack.nl - 1;
2654 nl >= 0 && cstack.bracelev[nl] >= bracelev;
2657 free (cstack.cname[nl]);
2663 write_classname (cn, qualifier)
2668 int qlen = strlen (qualifier);
2670 if (cstack.nl == 0 || cstack.cname[0] == NULL)
2674 cn->buffer[0] = '\0';
2678 len = strlen (cstack.cname[0]);
2679 linebuffer_setlen (cn, len+1);
2680 strncpy (cn->buffer, cstack.cname[0],len+1);
2682 for (i = 1; i < cstack.nl; i++)
2687 s = cstack.cname[i];
2692 linebuffer_setlen (cn, len);
2693 strncat (cn->buffer, qualifier, qlen);
2694 strncat (cn->buffer, s, slen);
2699 static bool consider_token __P((char *, int, int, int *, int, int, bool *));
2700 static void make_C_tag __P((bool));
2704 * checks to see if the current token is at the start of a
2705 * function or variable, or corresponds to a typedef, or
2706 * is a struct/union/enum tag, or #define, or an enum constant.
2708 * *IS_FUNC gets TRUE if the token is a function or #define macro
2709 * with args. C_EXTP points to which language we are looking at.
2720 consider_token (str, len, c, c_extp, bracelev, parlev, is_func_or_var)
2721 register char *str; /* IN: token pointer */
2722 register int len; /* IN: token length */
2723 register int c; /* IN: first char after the token */
2724 int *c_extp; /* IN, OUT: C extensions mask */
2725 int bracelev; /* IN: brace level */
2726 int parlev; /* IN: parenthesis level */
2727 bool *is_func_or_var; /* OUT: function or variable found */
2729 /* When structdef is stagseen, scolonseen, or snone with bracelev > 0,
2730 structtype is the type of the preceding struct-like keyword, and
2731 structbracelev is the brace level where it has been seen. */
2732 static enum sym_type structtype;
2733 static int structbracelev;
2734 static enum sym_type toktype;
2737 toktype = C_symtype (str, len, *c_extp);
2740 * Skip __attribute__
2742 if (toktype == st_C_attribute)
2749 * Advance the definedef state machine.
2754 /* We're not on a preprocessor line. */
2755 if (toktype == st_C_gnumacro)
2762 if (toktype == st_C_define)
2764 definedef = ddefineseen;
2768 definedef = dignorerest;
2773 * Make a tag for any macro, unless it is a constant
2774 * and constantypedefs is FALSE.
2776 definedef = dignorerest;
2777 *is_func_or_var = (c == '(');
2778 if (!*is_func_or_var && !constantypedefs)
2785 error ("internal error: definedef value.", (char *)NULL);
2794 if (toktype == st_C_typedef)
2818 case st_C_attribute:
2819 case st_C_javastruct:
2830 if (structdef == snone && fvdef == fvnone)
2851 case st_C_attribute:
2852 case st_C_javastruct:
2872 case st_C_javastruct:
2873 if (structdef == stagseen)
2874 structdef = scolonseen;
2878 if ((*c_extp & C_AUTO) /* automatic detection of C++ language */
2880 && definedef == dnone && structdef == snone
2881 && typdef == tnone && fvdef == fvnone)
2882 *c_extp = (*c_extp | C_PLPL) & ~C_AUTO;
2883 if (toktype == st_C_template)
2890 && (typdef == tkeyseen
2891 || (typedefs_or_cplusplus && structdef == snone)))
2893 structdef = skeyseen;
2894 structtype = toktype;
2895 structbracelev = bracelev;
2896 if (fvdef == fvnameseen)
2908 case st_C_attribute:
2917 if (structdef == skeyseen)
2919 structdef = stagseen;
2923 if (typdef != tnone)
2926 /* Detect Objective C constructs. */
2936 objdef = oimplementation;
2944 case st_C_attribute:
2945 case st_C_javastruct:
2958 case oimplementation:
2959 /* Save the class tag for functions or variables defined inside. */
2960 objtag = savenstr (str, len);
2964 /* Save the class tag for categories. */
2965 objtag = savenstr (str, len);
2967 *is_func_or_var = TRUE;
2971 *is_func_or_var = TRUE;
2979 objdef = omethodtag;
2980 linebuffer_setlen (&token_name, len);
2981 strncpy (token_name.buffer, str, len);
2982 token_name.buffer[len] = '\0';
2988 objdef = omethodparm;
2994 objdef = omethodtag;
2995 linebuffer_setlen (&token_name, token_name.len + len);
2996 strncat (token_name.buffer, str, len);
3001 if (toktype == st_C_objend)
3003 /* Memory leakage here: the string pointed by objtag is
3004 never released, because many tests would be needed to
3005 avoid breaking on incorrect input code. The amount of
3006 memory leaked here is the sum of the lengths of the
3021 /* A function, variable or enum constant? */
3049 *is_func_or_var = TRUE;
3053 && structdef == snone
3054 && structtype == st_C_enum && bracelev > structbracelev)
3055 return TRUE; /* enum constant */
3061 fvdef = fdefunname; /* GNU macro */
3062 *is_func_or_var = TRUE;
3070 if ((strneq (str, "asm", 3) && endtoken (str[3]))
3071 || (strneq (str, "__asm__", 7) && endtoken (str[7])))
3088 if (len >= 10 && strneq (str+len-10, "::operator", 10))
3090 if (*c_extp & C_AUTO) /* automatic detection of C++ */
3091 *c_extp = (*c_extp | C_PLPL) & ~C_AUTO;
3093 *is_func_or_var = TRUE;
3096 if (bracelev > 0 && !instruct)
3098 fvdef = fvnameseen; /* function or variable */
3099 *is_func_or_var = TRUE;
3120 case st_C_attribute:
3121 case st_C_javastruct:
3137 * C_entries often keeps pointers to tokens or lines which are older than
3138 * the line currently read. By keeping two line buffers, and switching
3139 * them at end of line, it is possible to use those pointers.
3147 #define current_lb_is_new (newndx == curndx)
3148 #define switch_line_buffers() (curndx = 1 - curndx)
3150 #define curlb (lbs[curndx].lb)
3151 #define newlb (lbs[newndx].lb)
3152 #define curlinepos (lbs[curndx].linepos)
3153 #define newlinepos (lbs[newndx].linepos)
3155 #define plainc ((c_ext & C_EXT) == C_PLAIN)
3156 #define cplpl (c_ext & C_PLPL)
3157 #define cjava ((c_ext & C_JAVA) == C_JAVA)
3159 #define CNL_SAVE_DEFINEDEF() \
3161 curlinepos = charno; \
3162 readline (&curlb, inf); \
3163 lp = curlb.buffer; \
3170 CNL_SAVE_DEFINEDEF(); \
3171 if (savetoken.valid) \
3173 token = savetoken; \
3174 savetoken.valid = FALSE; \
3176 definedef = dnone; \
3184 /* This function is never called when token.valid is FALSE, but
3185 we must protect against invalid input or internal errors. */
3187 make_tag (token_name.buffer, token_name.len, isfun, token.line,
3188 token.offset+token.length+1, token.lineno, token.linepos);
3190 { /* this branch is optimised away if !DEBUG */
3191 make_tag (concat ("INVALID TOKEN:-->", token_name.buffer, ""),
3192 token_name.len + 17, isfun, token.line,
3193 token.offset+token.length+1, token.lineno, token.linepos);
3194 error ("INVALID TOKEN", NULL);
3197 token.valid = FALSE;
3203 * This routine finds functions, variables, typedefs,
3204 * #define's, enum constants and struct/union/enum definitions in
3205 * C syntax and adds them to the list.
3208 C_entries (c_ext, inf)
3209 int c_ext; /* extension of C */
3210 FILE *inf; /* input file */
3212 register char c; /* latest char read; '\0' for end of line */
3213 register char *lp; /* pointer one beyond the character `c' */
3214 int curndx, newndx; /* indices for current and new lb */
3215 register int tokoff; /* offset in line of start of current token */
3216 register int toklen; /* length of current token */
3217 char *qualifier; /* string used to qualify names */
3218 int qlen; /* length of qualifier */
3219 int bracelev; /* current brace level */
3220 int bracketlev; /* current bracket level */
3221 int parlev; /* current parenthesis level */
3222 int attrparlev; /* __attribute__ parenthesis level */
3223 int templatelev; /* current template level */
3224 int typdefbracelev; /* bracelev where a typedef struct body begun */
3225 bool incomm, inquote, inchar, quotednl, midtoken;
3226 bool yacc_rules; /* in the rules part of a yacc file */
3227 struct tok savetoken = {0}; /* token saved during preprocessor handling */
3230 linebuffer_init (&lbs[0].lb);
3231 linebuffer_init (&lbs[1].lb);
3232 if (cstack.size == 0)
3234 cstack.size = (DEBUG) ? 1 : 4;
3236 cstack.cname = xnew (cstack.size, char *);
3237 cstack.bracelev = xnew (cstack.size, int);
3240 tokoff = toklen = typdefbracelev = 0; /* keep compiler quiet */
3241 curndx = newndx = 0;
3245 fvdef = fvnone; fvextern = FALSE; typdef = tnone;
3246 structdef = snone; definedef = dnone; objdef = onone;
3248 midtoken = inquote = inchar = incomm = quotednl = FALSE;
3249 token.valid = savetoken.valid = FALSE;
3250 bracelev = bracketlev = parlev = attrparlev = templatelev = 0;
3252 { qualifier = "."; qlen = 1; }
3254 { qualifier = "::"; qlen = 2; }
3262 /* If we are at the end of the line, the next character is a
3263 '\0'; do not skip it, because it is what tells us
3264 to read the next line. */
3285 /* Newlines inside comments do not end macro definitions in
3287 CNL_SAVE_DEFINEDEF ();
3302 /* Newlines inside strings do not end macro definitions
3303 in traditional cpp, even though compilers don't
3304 usually accept them. */
3305 CNL_SAVE_DEFINEDEF ();
3317 /* Hmmm, something went wrong. */
3328 else if (bracketlev > 0)
3333 if (--bracketlev > 0)
3337 CNL_SAVE_DEFINEDEF ();
3372 if (fvdef != finlist && fvdef != fignore && fvdef !=vignore)
3385 else if (/* cplpl && */ *lp == '/')
3391 if ((c_ext & YACC) && *lp == '%')
3393 /* Entering or exiting rules section in yacc file. */
3395 definedef = dnone; fvdef = fvnone; fvextern = FALSE;
3396 typdef = tnone; structdef = snone;
3397 midtoken = inquote = inchar = incomm = quotednl = FALSE;
3399 yacc_rules = !yacc_rules;
3405 if (definedef == dnone)
3408 bool cpptoken = TRUE;
3410 /* Look back on this line. If all blanks, or nonblanks
3411 followed by an end of comment, this is a preprocessor
3413 for (cp = newlb.buffer; cp < lp-1; cp++)
3416 if (*cp == '*' && *(cp+1) == '/')
3425 definedef = dsharpseen;
3426 } /* if (definedef == dnone) */
3436 /* Consider token only if some involved conditions are satisfied. */
3437 if (typdef != tignore
3438 && definedef != dignorerest
3441 && (definedef != dnone
3442 || structdef != scolonseen)
3449 if (c == ':' && *lp == ':' && begtoken (lp[1]))
3450 /* This handles :: in the middle,
3451 but not at the beginning of an identifier.
3452 Also, space-separated :: is not recognised. */
3454 if (c_ext & C_AUTO) /* automatic detection of C++ */
3455 c_ext = (c_ext | C_PLPL) & ~C_AUTO;
3459 goto still_in_token;
3463 bool funorvar = FALSE;
3466 || consider_token (newlb.buffer + tokoff, toklen, c,
3467 &c_ext, bracelev, parlev,
3470 if (fvdef == foperator)
3473 lp = skip_spaces (lp-1);
3477 && !iswhite (*lp) && *lp != '(')
3480 toklen += lp - oldlp;
3482 token.named = FALSE;
3484 && nestlev > 0 && definedef == dnone)
3485 /* in struct body */
3487 write_classname (&token_name, qualifier);
3488 linebuffer_setlen (&token_name,
3489 token_name.len+qlen+toklen);
3490 strcat (token_name.buffer, qualifier);
3491 strncat (token_name.buffer,
3492 newlb.buffer + tokoff, toklen);
3495 else if (objdef == ocatseen)
3496 /* Objective C category */
3498 int len = strlen (objtag) + 2 + toklen;
3499 linebuffer_setlen (&token_name, len);
3500 strcpy (token_name.buffer, objtag);
3501 strcat (token_name.buffer, "(");
3502 strncat (token_name.buffer,
3503 newlb.buffer + tokoff, toklen);
3504 strcat (token_name.buffer, ")");
3507 else if (objdef == omethodtag
3508 || objdef == omethodparm)
3509 /* Objective C method */
3513 else if (fvdef == fdefunname)
3514 /* GNU DEFUN and similar macros */
3516 bool defun = (newlb.buffer[tokoff] == 'F');
3520 /* Rewrite the tag so that emacs lisp DEFUNs
3521 can be found by their elisp name */
3527 linebuffer_setlen (&token_name, len);
3528 strncpy (token_name.buffer,
3529 newlb.buffer + off, len);
3530 token_name.buffer[len] = '\0';
3533 if (token_name.buffer[len] == '_')
3534 token_name.buffer[len] = '-';
3535 token.named = defun;
3539 linebuffer_setlen (&token_name, toklen);
3540 strncpy (token_name.buffer,
3541 newlb.buffer + tokoff, toklen);
3542 token_name.buffer[toklen] = '\0';
3543 /* Name macros and members. */
3544 token.named = (structdef == stagseen
3545 || typdef == ttypeseen
3548 && definedef == dignorerest)
3550 && definedef == dnone
3551 && structdef == snone
3554 token.lineno = lineno;
3555 token.offset = tokoff;
3556 token.length = toklen;
3557 token.line = newlb.buffer;
3558 token.linepos = newlinepos;
3561 if (definedef == dnone
3562 && (fvdef == fvnameseen
3563 || fvdef == foperator
3564 || structdef == stagseen
3566 || typdef == ttypeseen
3567 || objdef != onone))
3569 if (current_lb_is_new)
3570 switch_line_buffers ();
3572 else if (definedef != dnone
3573 || fvdef == fdefunname
3575 make_C_tag (funorvar);
3577 else /* not yacc and consider_token failed */
3579 if (inattribute && fvdef == fignore)
3581 /* We have just met __attribute__ after a
3582 function parameter list: do not tag the
3589 } /* if (endtoken (c)) */
3590 else if (intoken (c))
3596 } /* if (midtoken) */
3597 else if (begtoken (c))
3605 /* This prevents tagging fb in
3606 void (__attribute__((noreturn)) *fb) (void);
3607 Fixing this is not easy and not very important. */
3611 if (plainc || declarations)
3613 make_C_tag (TRUE); /* a function */
3630 if (structdef == stagseen && !cjava)
3632 popclass_above (bracelev);
3646 if (!yacc_rules || lp == newlb.buffer + 1)
3648 tokoff = lp - 1 - newlb.buffer;
3653 } /* if (begtoken) */
3654 } /* if must look at token */
3657 /* Detect end of line, colon, comma, semicolon and various braces
3658 after having handled a token.*/
3664 if (yacc_rules && token.offset == 0 && token.valid)
3666 make_C_tag (FALSE); /* a yacc function */
3669 if (definedef != dnone)
3675 make_C_tag (TRUE); /* an Objective C class */
3679 objdef = omethodcolon;
3680 linebuffer_setlen (&token_name, token_name.len + 1);
3681 strcat (token_name.buffer, ":");
3687 case oimplementation:
3697 if (structdef == stagseen)
3699 structdef = scolonseen;
3702 /* Should be useless, but may be work as a safety net. */
3703 if (cplpl && fvdef == flistseen)
3705 make_C_tag (TRUE); /* a function */
3711 if (definedef != dnone || inattribute)
3717 make_C_tag (FALSE); /* a typedef */
3727 if (typdef == tignore || cplpl)
3731 if ((globals && bracelev == 0 && (!fvextern || declarations))
3732 || (members && instruct))
3733 make_C_tag (FALSE); /* a variable */
3736 token.valid = FALSE;
3740 && (cplpl || !instruct)
3741 && (typdef == tnone || (typdef != tignore && instruct)))
3743 && plainc && instruct))
3744 make_C_tag (TRUE); /* a function */
3757 && cplpl && structdef == stagseen)
3758 make_C_tag (FALSE); /* forward declaration */
3760 token.valid = FALSE;
3761 } /* switch (fvdef) */
3768 if (structdef == stagseen)
3772 if (definedef != dnone || inattribute)
3778 make_C_tag (TRUE); /* an Objective C method */
3785 case oimplementation:
3813 && (!fvextern || declarations))
3814 || (members && instruct)))
3815 make_C_tag (FALSE); /* a variable */
3818 if ((declarations && typdef == tnone && !instruct)
3819 || (members && typdef != tignore && instruct))
3821 make_C_tag (TRUE); /* a function */
3824 else if (!declarations)
3826 token.valid = FALSE;
3832 if (structdef == stagseen)
3836 if (definedef != dnone || inattribute)
3838 if (structdef == stagseen)
3845 make_C_tag (FALSE); /* a typedef */
3857 if ((members && bracelev == 1)
3858 || (globals && bracelev == 0
3859 && (!fvextern || declarations)))
3860 make_C_tag (FALSE); /* a variable */
3885 if (definedef != dnone)
3887 if (objdef == otagseen && parlev == 0)
3888 objdef = oparenseen;
3892 if (typdef == ttypeseen
3896 /* This handles constructs like:
3897 typedef void OperatorFun (int fun); */
3927 if (--attrparlev == 0)
3928 inattribute = FALSE;
3931 if (definedef != dnone)
3933 if (objdef == ocatseen && parlev == 1)
3935 make_C_tag (TRUE); /* an Objective C category */
3961 || typdef == ttypeseen))
3964 make_C_tag (FALSE); /* a typedef */
3967 else if (parlev < 0) /* can happen due to ill-conceived #if's. */
3971 if (definedef != dnone)
3973 if (typdef == ttypeseen)
3975 /* Whenever typdef is set to tinbody (currently only
3976 here), typdefbracelev should be set to bracelev. */
3978 typdefbracelev = bracelev;
3983 make_C_tag (TRUE); /* a function */
3992 make_C_tag (TRUE); /* an Objective C class */
3997 make_C_tag (TRUE); /* an Objective C method */
4002 case oimplementation:
4010 /* Neutralize `extern "C" {' grot. */
4011 if (bracelev == 0 && structdef == snone && nestlev == 0
4030 case skeyseen: /* unnamed struct */
4031 pushclass_above (bracelev, NULL, 0);
4034 case stagseen: /* named struct or enum */
4035 case scolonseen: /* a class */
4036 pushclass_above (bracelev,token.line+token.offset, token.length);
4038 make_C_tag (FALSE); /* a struct or enum */
4049 if (definedef != dnone)
4051 if (fvdef == fstartlist)
4053 fvdef = fvnone; /* avoid tagging `foo' in `foo (*bar()) ()' */
4054 token.valid = FALSE;
4058 if (definedef != dnone)
4061 if (!ignoreindent && lp == newlb.buffer + 1)
4064 token.valid = FALSE; /* unexpected value, token unreliable */
4065 bracelev = 0; /* reset brace level if first column */
4066 parlev = 0; /* also reset paren level, just in case... */
4068 else if (bracelev < 0)
4070 token.valid = FALSE; /* something gone amiss, token unreliable */
4073 if (bracelev == 0 && fvdef == vignore)
4074 fvdef = fvnone; /* end of function */
4075 popclass_above (bracelev);
4077 /* Only if typdef == tinbody is typdefbracelev significant. */
4078 if (typdef == tinbody && bracelev <= typdefbracelev)
4080 assert (bracelev == typdefbracelev);
4085 if (definedef != dnone)
4095 if ((members && bracelev == 1)
4096 || (globals && bracelev == 0 && (!fvextern || declarations)))
4097 make_C_tag (FALSE); /* a variable */
4110 && (structdef == stagseen || fvdef == fvnameseen))
4117 if (templatelev > 0)
4125 if (objdef == oinbody && bracelev == 0)
4127 objdef = omethodsign;
4132 case '#': case '~': case '&': case '%': case '/':
4133 case '|': case '^': case '!': case '.': case '?':
4134 if (definedef != dnone)
4136 /* These surely cannot follow a function tag in C. */
4155 if (objdef == otagseen)
4157 make_C_tag (TRUE); /* an Objective C class */
4160 /* If a macro spans multiple lines don't reset its state. */
4162 CNL_SAVE_DEFINEDEF ();
4170 } /* while not eof */
4172 free (lbs[0].lb.buffer);
4173 free (lbs[1].lb.buffer);
4177 * Process either a C++ file or a C file depending on the setting
4181 default_C_entries (inf)
4184 C_entries (cplusplus ? C_PLPL : C_AUTO, inf);
4187 /* Always do plain C. */
4189 plain_C_entries (inf)
4195 /* Always do C++. */
4197 Cplusplus_entries (inf)
4200 C_entries (C_PLPL, inf);
4203 /* Always do Java. */
4208 C_entries (C_JAVA, inf);
4216 C_entries (C_STAR, inf);
4219 /* Always do Yacc. */
4224 C_entries (YACC, inf);
4228 /* Useful macros. */
4229 #define LOOP_ON_INPUT_LINES(file_pointer, line_buffer, char_pointer) \
4230 for (; /* loop initialization */ \
4231 !feof (file_pointer) /* loop test */ \
4232 && /* instructions at start of loop */ \
4233 (readline (&line_buffer, file_pointer), \
4234 char_pointer = line_buffer.buffer, \
4238 #define LOOKING_AT(cp, kw) /* kw is the keyword, a literal string */ \
4239 ((assert("" kw), TRUE) /* syntax error if not a literal string */ \
4240 && strneq ((cp), kw, sizeof(kw)-1) /* cp points at kw */ \
4241 && notinname ((cp)[sizeof(kw)-1]) /* end of kw */ \
4242 && ((cp) = skip_spaces((cp)+sizeof(kw)-1))) /* skip spaces */
4244 /* Similar to LOOKING_AT but does not use notinname, does not skip */
4245 #define LOOKING_AT_NOCASE(cp, kw) /* the keyword is a literal string */ \
4246 ((assert("" kw), TRUE) /* syntax error if not a literal string */ \
4247 && strncaseeq ((cp), kw, sizeof(kw)-1) /* cp points at kw */ \
4248 && ((cp) += sizeof(kw)-1)) /* skip spaces */
4251 * Read a file, but do no processing. This is used to do regexp
4252 * matching on files that have no language defined.
4255 just_read_file (inf)
4258 register char *dummy;
4260 LOOP_ON_INPUT_LINES (inf, lb, dummy)
4265 /* Fortran parsing */
4267 static void F_takeprec __P((void));
4268 static void F_getit __P((FILE *));
4273 dbp = skip_spaces (dbp);
4277 dbp = skip_spaces (dbp);
4278 if (strneq (dbp, "(*)", 3))
4283 if (!ISDIGIT (*dbp))
4285 --dbp; /* force failure */
4290 while (ISDIGIT (*dbp));
4299 dbp = skip_spaces (dbp);
4302 readline (&lb, inf);
4307 dbp = skip_spaces (dbp);
4309 if (!ISALPHA (*dbp) && *dbp != '_' && *dbp != '$')
4311 for (cp = dbp + 1; *cp != '\0' && intoken (*cp); cp++)
4313 make_tag (dbp, cp-dbp, TRUE,
4314 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4319 Fortran_functions (inf)
4322 LOOP_ON_INPUT_LINES (inf, lb, dbp)
4325 dbp++; /* Ratfor escape to fortran */
4326 dbp = skip_spaces (dbp);
4329 switch (lowcase (*dbp))
4332 if (nocase_tail ("integer"))
4336 if (nocase_tail ("real"))
4340 if (nocase_tail ("logical"))
4344 if (nocase_tail ("complex") || nocase_tail ("character"))
4348 if (nocase_tail ("double"))
4350 dbp = skip_spaces (dbp);
4353 if (nocase_tail ("precision"))
4361 dbp = skip_spaces (dbp);
4364 switch (lowcase (*dbp))
4367 if (nocase_tail ("function"))
4371 if (nocase_tail ("subroutine"))
4375 if (nocase_tail ("entry"))
4379 if (nocase_tail ("blockdata") || nocase_tail ("block data"))
4381 dbp = skip_spaces (dbp);
4382 if (*dbp == '\0') /* assume un-named */
4383 make_tag ("blockdata", 9, TRUE,
4384 lb.buffer, dbp - lb.buffer, lineno, linecharno);
4386 F_getit (inf); /* look for name */
4399 * Philippe Waroquiers (1998)
4402 static void Ada_getit __P((FILE *, char *));
4404 /* Once we are positioned after an "interesting" keyword, let's get
4405 the real tag value necessary. */
4407 Ada_getit (inf, name_qualifier)
4409 char *name_qualifier;
4417 dbp = skip_spaces (dbp);
4419 || (dbp[0] == '-' && dbp[1] == '-'))
4421 readline (&lb, inf);
4424 switch (lowcase(*dbp))
4427 if (nocase_tail ("body"))
4429 /* Skipping body of procedure body or package body or ....
4430 resetting qualifier to body instead of spec. */
4431 name_qualifier = "/b";
4436 /* Skipping type of task type or protected type ... */
4437 if (nocase_tail ("type"))
4446 for (cp = dbp; *cp != '\0' && *cp != '"'; cp++)
4451 dbp = skip_spaces (dbp);
4454 && (ISALPHA (*cp) || ISDIGIT (*cp) || *cp == '_' || *cp == '.'));
4462 name = concat (dbp, name_qualifier, "");
4464 make_tag (name, strlen (name), TRUE,
4465 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4477 bool inquote = FALSE;
4478 bool skip_till_semicolumn = FALSE;
4480 LOOP_ON_INPUT_LINES (inf, lb, dbp)
4482 while (*dbp != '\0')
4484 /* Skip a string i.e. "abcd". */
4485 if (inquote || (*dbp == '"'))
4487 dbp = etags_strchr ((inquote) ? dbp : dbp+1, '"');
4492 continue; /* advance char */
4497 break; /* advance line */
4501 /* Skip comments. */
4502 if (dbp[0] == '-' && dbp[1] == '-')
4503 break; /* advance line */
4505 /* Skip character enclosed in single quote i.e. 'a'
4506 and skip single quote starting an attribute i.e. 'Image. */
4515 if (skip_till_semicolumn)
4518 skip_till_semicolumn = FALSE;
4520 continue; /* advance char */
4523 /* Search for beginning of a token. */
4524 if (!begtoken (*dbp))
4527 continue; /* advance char */
4530 /* We are at the beginning of a token. */
4531 switch (lowcase(*dbp))
4534 if (!packages_only && nocase_tail ("function"))
4535 Ada_getit (inf, "/f");
4537 break; /* from switch */
4538 continue; /* advance char */
4540 if (!packages_only && nocase_tail ("procedure"))
4541 Ada_getit (inf, "/p");
4542 else if (nocase_tail ("package"))
4543 Ada_getit (inf, "/s");
4544 else if (nocase_tail ("protected")) /* protected type */
4545 Ada_getit (inf, "/t");
4547 break; /* from switch */
4548 continue; /* advance char */
4551 if (typedefs && !packages_only && nocase_tail ("use"))
4553 /* when tagging types, avoid tagging use type Pack.Typename;
4554 for this, we will skip everything till a ; */
4555 skip_till_semicolumn = TRUE;
4556 continue; /* advance char */
4560 if (!packages_only && nocase_tail ("task"))
4561 Ada_getit (inf, "/k");
4562 else if (typedefs && !packages_only && nocase_tail ("type"))
4564 Ada_getit (inf, "/t");
4565 while (*dbp != '\0')
4569 break; /* from switch */
4570 continue; /* advance char */
4575 /* Look for the end of the token. */
4576 while (!endtoken (*dbp))
4579 } /* advance char */
4580 } /* advance line */
4585 * Unix and microcontroller assembly tag handling
4586 * Labels: /^[a-zA-Z_.$][a-zA_Z0-9_.$]*[: ^I^J]/
4587 * Idea by Bob Weiner, Motorola Inc. (1994)
4595 LOOP_ON_INPUT_LINES (inf, lb, cp)
4597 /* If first char is alphabetic or one of [_.$], test for colon
4598 following identifier. */
4599 if (ISALPHA (*cp) || *cp == '_' || *cp == '.' || *cp == '$')
4601 /* Read past label. */
4603 while (ISALNUM (*cp) || *cp == '_' || *cp == '.' || *cp == '$')
4605 if (*cp == ':' || iswhite (*cp))
4606 /* Found end of label, so copy it and add it to the table. */
4607 make_tag (lb.buffer, cp - lb.buffer, TRUE,
4608 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4616 * Perl sub names: /^sub[ \t\n]+[^ \t\n{]+/
4617 * Perl variable names: /^(my|local).../
4618 * Original code by Bart Robinson <lomew@cs.utah.edu> (1995)
4619 * Additions by Michael Ernst <mernst@alum.mit.edu> (1997)
4620 * Ideas by Kai Großjohann <Kai.Grossjohann@CS.Uni-Dortmund.DE> (2001)
4623 Perl_functions (inf)
4626 char *package = savestr ("main"); /* current package name */
4629 LOOP_ON_INPUT_LINES (inf, lb, cp)
4631 cp = skip_spaces (cp);
4633 if (LOOKING_AT (cp, "package"))
4636 get_tag (cp, &package);
4638 else if (LOOKING_AT (cp, "sub"))
4643 while (!notinname (*cp))
4646 continue; /* nothing found */
4647 if ((pos = etags_strchr (sp, ':')) != NULL
4648 && pos < cp && pos[1] == ':')
4649 /* The name is already qualified. */
4650 make_tag (sp, cp - sp, TRUE,
4651 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4655 char savechar, *name;
4659 name = concat (package, "::", sp);
4661 make_tag (name, strlen(name), TRUE,
4662 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4666 else if (globals) /* only if we are tagging global vars */
4668 /* Skip a qualifier, if any. */
4669 bool qual = LOOKING_AT (cp, "my") || LOOKING_AT (cp, "local");
4670 /* After "my" or "local", but before any following paren or space. */
4671 char *varstart = cp;
4673 if (qual /* should this be removed? If yes, how? */
4674 && (*cp == '$' || *cp == '@' || *cp == '%'))
4679 while (ISALNUM (*cp) || *cp == '_');
4683 /* Should be examining a variable list at this point;
4684 could insist on seeing an open parenthesis. */
4685 while (*cp != '\0' && *cp != ';' && *cp != '=' && *cp != ')')
4691 make_tag (varstart, cp - varstart, FALSE,
4692 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4701 * Look for /^[\t]*def[ \t\n]+[^ \t\n(:]+/ or /^class[ \t\n]+[^ \t\n(:]+/
4702 * Idea by Eric S. Raymond <esr@thyrsus.com> (1997)
4703 * More ideas by seb bacon <seb@jamkit.com> (2002)
4706 Python_functions (inf)
4711 LOOP_ON_INPUT_LINES (inf, lb, cp)
4713 cp = skip_spaces (cp);
4714 if (LOOKING_AT (cp, "def") || LOOKING_AT (cp, "class"))
4717 while (!notinname (*cp) && *cp != ':')
4719 make_tag (name, cp - name, TRUE,
4720 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4729 * - /^[ \t]*function[ \t\n]+[^ \t\n(]+/
4730 * - /^[ \t]*class[ \t\n]+[^ \t\n]+/
4731 * - /^[ \t]*define\(\"[^\"]+/
4732 * Only with --members:
4733 * - /^[ \t]*var[ \t\n]+\$[^ \t\n=;]/
4734 * Idea by Diez B. Roggisch (2001)
4740 register char *cp, *name;
4741 bool search_identifier = FALSE;
4743 LOOP_ON_INPUT_LINES (inf, lb, cp)
4745 cp = skip_spaces (cp);
4747 if (search_identifier
4750 while (!notinname (*cp))
4752 make_tag (name, cp - name, TRUE,
4753 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4754 search_identifier = FALSE;
4756 else if (LOOKING_AT (cp, "function"))
4759 cp = skip_spaces (cp+1);
4763 while (!notinname (*cp))
4765 make_tag (name, cp - name, TRUE,
4766 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4769 search_identifier = TRUE;
4771 else if (LOOKING_AT (cp, "class"))
4776 while (*cp != '\0' && !iswhite (*cp))
4778 make_tag (name, cp - name, FALSE,
4779 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4782 search_identifier = TRUE;
4784 else if (strneq (cp, "define", 6)
4785 && (cp = skip_spaces (cp+6))
4787 && (*cp == '"' || *cp == '\''))
4791 while (*cp != quote && *cp != '\0')
4793 make_tag (name, cp - name, FALSE,
4794 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4797 && LOOKING_AT (cp, "var")
4801 while (!notinname(*cp))
4803 make_tag (name, cp - name, FALSE,
4804 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4811 * Cobol tag functions
4812 * We could look for anything that could be a paragraph name.
4813 * i.e. anything that starts in column 8 is one word and ends in a full stop.
4814 * Idea by Corny de Souza (1993)
4817 Cobol_paragraphs (inf)
4820 register char *bp, *ep;
4822 LOOP_ON_INPUT_LINES (inf, lb, bp)
4828 /* If eoln, compiler option or comment ignore whole line. */
4829 if (bp[-1] != ' ' || !ISALNUM (bp[0]))
4832 for (ep = bp; ISALNUM (*ep) || *ep == '-'; ep++)
4835 make_tag (bp, ep - bp, TRUE,
4836 lb.buffer, ep - lb.buffer + 1, lineno, linecharno);
4843 * Ideas by Assar Westerlund <assar@sics.se> (2001)
4846 Makefile_targets (inf)
4851 LOOP_ON_INPUT_LINES (inf, lb, bp)
4853 if (*bp == '\t' || *bp == '#')
4855 while (*bp != '\0' && *bp != '=' && *bp != ':')
4857 if (*bp == ':' || (globals && *bp == '='))
4859 /* We should detect if there is more than one tag, but we do not.
4860 We just skip initial and final spaces. */
4861 char * namestart = skip_spaces (lb.buffer);
4862 while (--bp > namestart)
4863 if (!notinname (*bp))
4865 make_tag (namestart, bp - namestart + 1, TRUE,
4866 lb.buffer, bp - lb.buffer + 2, lineno, linecharno);
4874 * Original code by Mosur K. Mohan (1989)
4876 * Locates tags for procedures & functions. Doesn't do any type- or
4877 * var-definitions. It does look for the keyword "extern" or
4878 * "forward" immediately following the procedure statement; if found,
4879 * the tag is skipped.
4882 Pascal_functions (inf)
4885 linebuffer tline; /* mostly copied from C_entries */
4887 int save_lineno, namelen, taglen;
4890 bool /* each of these flags is TRUE if: */
4891 incomment, /* point is inside a comment */
4892 inquote, /* point is inside '..' string */
4893 get_tagname, /* point is after PROCEDURE/FUNCTION
4894 keyword, so next item = potential tag */
4895 found_tag, /* point is after a potential tag */
4896 inparms, /* point is within parameter-list */
4897 verify_tag; /* point has passed the parm-list, so the
4898 next token will determine whether this
4899 is a FORWARD/EXTERN to be ignored, or
4900 whether it is a real tag */
4902 save_lcno = save_lineno = namelen = taglen = 0; /* keep compiler quiet */
4903 name = NULL; /* keep compiler quiet */
4906 linebuffer_init (&tline);
4908 incomment = inquote = FALSE;
4909 found_tag = FALSE; /* have a proc name; check if extern */
4910 get_tagname = FALSE; /* found "procedure" keyword */
4911 inparms = FALSE; /* found '(' after "proc" */
4912 verify_tag = FALSE; /* check if "extern" is ahead */
4915 while (!feof (inf)) /* long main loop to get next char */
4918 if (c == '\0') /* if end of line */
4920 readline (&lb, inf);
4924 if (!((found_tag && verify_tag)
4926 c = *dbp++; /* only if don't need *dbp pointing
4927 to the beginning of the name of
4928 the procedure or function */
4932 if (c == '}') /* within { } comments */
4934 else if (c == '*' && *dbp == ')') /* within (* *) comments */
4951 inquote = TRUE; /* found first quote */
4953 case '{': /* found open { comment */
4957 if (*dbp == '*') /* found open (* comment */
4962 else if (found_tag) /* found '(' after tag, i.e., parm-list */
4965 case ')': /* end of parms list */
4970 if (found_tag && !inparms) /* end of proc or fn stmt */
4979 if (found_tag && verify_tag && (*dbp != ' '))
4981 /* Check if this is an "extern" declaration. */
4984 if (lowcase (*dbp == 'e'))
4986 if (nocase_tail ("extern")) /* superfluous, really! */
4992 else if (lowcase (*dbp) == 'f')
4994 if (nocase_tail ("forward")) /* check for forward reference */
5000 if (found_tag && verify_tag) /* not external proc, so make tag */
5004 make_tag (name, namelen, TRUE,
5005 tline.buffer, taglen, save_lineno, save_lcno);
5009 if (get_tagname) /* grab name of proc or fn */
5016 /* Find block name. */
5017 for (cp = dbp + 1; *cp != '\0' && !endtoken (*cp); cp++)
5020 /* Save all values for later tagging. */
5021 linebuffer_setlen (&tline, lb.len);
5022 strncpy(tline.buffer, lb.buffer, lb.len-1);
5023 save_lineno = lineno;
5024 save_lcno = linecharno;
5025 name = tline.buffer + (dbp - lb.buffer);
5027 taglen = cp - lb.buffer + 1;
5029 dbp = cp; /* set dbp to e-o-token */
5030 get_tagname = FALSE;
5034 /* And proceed to check for "extern". */
5036 else if (!incomment && !inquote && !found_tag)
5038 /* Check for proc/fn keywords. */
5039 switch (lowcase (c))
5042 if (nocase_tail ("rocedure")) /* c = 'p', dbp has advanced */
5046 if (nocase_tail ("unction"))
5053 } /* while not eof */
5055 free (tline.buffer);
5060 * Lisp tag functions
5061 * look for (def or (DEF, quote or QUOTE
5064 static void L_getit __P((void));
5069 if (*dbp == '\'') /* Skip prefix quote */
5071 else if (*dbp == '(')
5074 /* Try to skip "(quote " */
5075 if (!LOOKING_AT (dbp, "quote") && !LOOKING_AT (dbp, "QUOTE"))
5076 /* Ok, then skip "(" before name in (defstruct (foo)) */
5077 dbp = skip_spaces (dbp);
5079 get_tag (dbp, NULL);
5083 Lisp_functions (inf)
5086 LOOP_ON_INPUT_LINES (inf, lb, dbp)
5091 if (strneq (dbp+1, "def", 3) || strneq (dbp+1, "DEF", 3))
5093 dbp = skip_non_spaces (dbp);
5094 dbp = skip_spaces (dbp);
5099 /* Check for (foo::defmumble name-defined ... */
5102 while (!notinname (*dbp) && *dbp != ':');
5107 while (*dbp == ':');
5109 if (strneq (dbp, "def", 3) || strneq (dbp, "DEF", 3))
5111 dbp = skip_non_spaces (dbp);
5112 dbp = skip_spaces (dbp);
5122 * Lua script language parsing
5123 * Original code by David A. Capello <dacap@users.sourceforge.net> (2004)
5125 * "function" and "local function" are tags if they start at column 1.
5133 LOOP_ON_INPUT_LINES (inf, lb, bp)
5135 if (bp[0] != 'f' && bp[0] != 'l')
5138 (void)LOOKING_AT (bp, "local"); /* skip possible "local" */
5140 if (LOOKING_AT (bp, "function"))
5148 * Just look for lines where the first character is '/'
5149 * Also look at "defineps" for PSWrap
5151 * Richard Mlynarik <mly@adoc.xerox.com> (1997)
5152 * Masatake Yamato <masata-y@is.aist-nara.ac.jp> (1999)
5158 register char *bp, *ep;
5160 LOOP_ON_INPUT_LINES (inf, lb, bp)
5165 *ep != '\0' && *ep != ' ' && *ep != '{';
5168 make_tag (bp, ep - bp, TRUE,
5169 lb.buffer, ep - lb.buffer + 1, lineno, linecharno);
5171 else if (LOOKING_AT (bp, "defineps"))
5179 * Ignore anything after \ followed by space or in ( )
5180 * Look for words defined by :
5181 * Look for constant, code, create, defer, value, and variable
5182 * OBP extensions: Look for buffer:, field,
5183 * Ideas by Eduardo Horvath <eeh@netbsd.org> (2004)
5191 LOOP_ON_INPUT_LINES (inf, lb, bp)
5192 while ((bp = skip_spaces (bp))[0] != '\0')
5193 if (bp[0] == '\\' && iswhite(bp[1]))
5194 break; /* read next line */
5195 else if (bp[0] == '(' && iswhite(bp[1]))
5196 do /* skip to ) or eol */
5198 while (*bp != ')' && *bp != '\0');
5199 else if ((bp[0] == ':' && iswhite(bp[1]) && bp++)
5200 || LOOKING_AT_NOCASE (bp, "constant")
5201 || LOOKING_AT_NOCASE (bp, "code")
5202 || LOOKING_AT_NOCASE (bp, "create")
5203 || LOOKING_AT_NOCASE (bp, "defer")
5204 || LOOKING_AT_NOCASE (bp, "value")
5205 || LOOKING_AT_NOCASE (bp, "variable")
5206 || LOOKING_AT_NOCASE (bp, "buffer:")
5207 || LOOKING_AT_NOCASE (bp, "field"))
5208 get_tag (skip_spaces (bp), NULL); /* Yay! A definition! */
5210 bp = skip_non_spaces (bp);
5215 * Scheme tag functions
5216 * look for (def... xyzzy
5218 * (def ... ((...(xyzzy ....
5220 * Original code by Ken Haase (1985?)
5223 Scheme_functions (inf)
5228 LOOP_ON_INPUT_LINES (inf, lb, bp)
5230 if (strneq (bp, "(def", 4) || strneq (bp, "(DEF", 4))
5232 bp = skip_non_spaces (bp+4);
5233 /* Skip over open parens and white space */
5234 while (notinname (*bp))
5238 if (LOOKING_AT (bp, "(SET!") || LOOKING_AT (bp, "(set!"))
5244 /* Find tags in TeX and LaTeX input files. */
5246 /* TEX_toktab is a table of TeX control sequences that define tags.
5247 * Each entry records one such control sequence.
5249 * Original code from who knows whom.
5251 * Stefan Monnier (2002)
5254 static linebuffer *TEX_toktab = NULL; /* Table with tag tokens */
5256 /* Default set of control sequences to put into TEX_toktab.
5257 The value of environment var TEXTAGS is prepended to this. */
5258 static char *TEX_defenv = "\
5259 :chapter:section:subsection:subsubsection:eqno:label:ref:cite:bibitem\
5260 :part:appendix:entry:index:def\
5261 :newcommand:renewcommand:newenvironment:renewenvironment";
5263 static void TEX_mode __P((FILE *));
5264 static void TEX_decode_env __P((char *, char *));
5266 static char TEX_esc = '\\';
5267 static char TEX_opgrp = '{';
5268 static char TEX_clgrp = '}';
5271 * TeX/LaTeX scanning loop.
5280 /* Select either \ or ! as escape character. */
5283 /* Initialize token table once from environment. */
5284 if (TEX_toktab == NULL)
5285 TEX_decode_env ("TEXTAGS", TEX_defenv);
5287 LOOP_ON_INPUT_LINES (inf, lb, cp)
5289 /* Look at each TEX keyword in line. */
5292 /* Look for a TEX escape. */
5293 while (*cp++ != TEX_esc)
5294 if (cp[-1] == '\0' || cp[-1] == '%')
5297 for (key = TEX_toktab; key->buffer != NULL; key++)
5298 if (strneq (cp, key->buffer, key->len))
5301 int namelen, linelen;
5304 cp = skip_spaces (cp + key->len);
5305 if (*cp == TEX_opgrp)
5311 (!iswhite (*p) && *p != '#' &&
5312 *p != TEX_opgrp && *p != TEX_clgrp);
5317 if (!opgrp || *p == TEX_clgrp)
5319 while (*p != '\0' && *p != TEX_opgrp && *p != TEX_clgrp)
5321 linelen = p - lb.buffer + 1;
5323 make_tag (cp, namelen, TRUE,
5324 lb.buffer, linelen, lineno, linecharno);
5325 goto tex_next_line; /* We only tag a line once */
5333 #define TEX_LESC '\\'
5334 #define TEX_SESC '!'
5336 /* Figure out whether TeX's escapechar is '\\' or '!' and set grouping
5337 chars accordingly. */
5344 while ((c = getc (inf)) != EOF)
5346 /* Skip to next line if we hit the TeX comment char. */
5348 while (c != '\n' && c != EOF)
5350 else if (c == TEX_LESC || c == TEX_SESC )
5366 /* If the input file is compressed, inf is a pipe, and rewind may fail.
5367 No attempt is made to correct the situation. */
5371 /* Read environment and prepend it to the default string.
5372 Build token table. */
5374 TEX_decode_env (evarname, defenv)
5378 register char *env, *p;
5381 /* Append default string to environment. */
5382 env = getenv (evarname);
5388 env = concat (oldenv, defenv, "");
5391 /* Allocate a token table */
5392 for (len = 1, p = env; p;)
5393 if ((p = etags_strchr (p, ':')) && *++p != '\0')
5395 TEX_toktab = xnew (len, linebuffer);
5397 /* Unpack environment string into token table. Be careful about */
5398 /* zero-length strings (leading ':', "::" and trailing ':') */
5399 for (i = 0; *env != '\0';)
5401 p = etags_strchr (env, ':');
5402 if (!p) /* End of environment string. */
5403 p = env + strlen (env);
5405 { /* Only non-zero strings. */
5406 TEX_toktab[i].buffer = savenstr (env, p - env);
5407 TEX_toktab[i].len = p - env;
5414 TEX_toktab[i].buffer = NULL; /* Mark end of table. */
5415 TEX_toktab[i].len = 0;
5422 /* Texinfo support. Dave Love, Mar. 2000. */
5428 LOOP_ON_INPUT_LINES (inf, lb, cp)
5429 if (LOOKING_AT (cp, "@node"))
5432 while (*cp != '\0' && *cp != ',')
5434 make_tag (start, cp - start, TRUE,
5435 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
5442 * Contents of <title>, <h1>, <h2>, <h3> are tags.
5443 * Contents of <a name=xxx> are tags with name xxx.
5445 * Francesco Potortì, 2002.
5451 bool getnext = FALSE; /* next text outside of HTML tags is a tag */
5452 bool skiptag = FALSE; /* skip to the end of the current HTML tag */
5453 bool intag = FALSE; /* inside an html tag, looking for ID= */
5454 bool inanchor = FALSE; /* when INTAG, is an anchor, look for NAME= */
5458 linebuffer_setlen (&token_name, 0); /* no name in buffer */
5460 LOOP_ON_INPUT_LINES (inf, lb, dbp)
5461 for (;;) /* loop on the same line */
5463 if (skiptag) /* skip HTML tag */
5465 while (*dbp != '\0' && *dbp != '>')
5471 continue; /* look on the same line */
5473 break; /* go to next line */
5476 else if (intag) /* look for "name=" or "id=" */
5478 while (*dbp != '\0' && *dbp != '>'
5479 && lowcase (*dbp) != 'n' && lowcase (*dbp) != 'i')
5482 break; /* go to next line */
5487 continue; /* look on the same line */
5489 if ((inanchor && LOOKING_AT_NOCASE (dbp, "name="))
5490 || LOOKING_AT_NOCASE (dbp, "id="))
5492 bool quoted = (dbp[0] == '"');
5495 for (end = ++dbp; *end != '\0' && *end != '"'; end++)
5498 for (end = dbp; *end != '\0' && intoken (*end); end++)
5500 linebuffer_setlen (&token_name, end - dbp);
5501 strncpy (token_name.buffer, dbp, end - dbp);
5502 token_name.buffer[end - dbp] = '\0';
5505 intag = FALSE; /* we found what we looked for */
5506 skiptag = TRUE; /* skip to the end of the tag */
5507 getnext = TRUE; /* then grab the text */
5508 continue; /* look on the same line */
5513 else if (getnext) /* grab next tokens and tag them */
5515 dbp = skip_spaces (dbp);
5517 break; /* go to next line */
5521 inanchor = (lowcase (dbp[1]) == 'a' && !intoken (dbp[2]));
5522 continue; /* look on the same line */
5525 for (end = dbp + 1; *end != '\0' && *end != '<'; end++)
5527 make_tag (token_name.buffer, token_name.len, TRUE,
5528 dbp, end - dbp, lineno, linecharno);
5529 linebuffer_setlen (&token_name, 0); /* no name in buffer */
5531 break; /* go to next line */
5534 else /* look for an interesting HTML tag */
5536 while (*dbp != '\0' && *dbp != '<')
5539 break; /* go to next line */
5541 if (lowcase (dbp[1]) == 'a' && !intoken (dbp[2]))
5544 continue; /* look on the same line */
5546 else if (LOOKING_AT_NOCASE (dbp, "<title>")
5547 || LOOKING_AT_NOCASE (dbp, "<h1>")
5548 || LOOKING_AT_NOCASE (dbp, "<h2>")
5549 || LOOKING_AT_NOCASE (dbp, "<h3>"))
5553 continue; /* look on the same line */
5564 * Assumes that the predicate or rule starts at column 0.
5565 * Only the first clause of a predicate or rule is added.
5566 * Original code by Sunichirou Sugou (1989)
5567 * Rewritten by Anders Lindgren (1996)
5569 static int prolog_pr __P((char *, char *));
5570 static void prolog_skip_comment __P((linebuffer *, FILE *));
5571 static int prolog_atom __P((char *, int));
5574 Prolog_functions (inf)
5585 LOOP_ON_INPUT_LINES (inf, lb, cp)
5587 if (cp[0] == '\0') /* Empty line */
5589 else if (iswhite (cp[0])) /* Not a predicate */
5591 else if (cp[0] == '/' && cp[1] == '*') /* comment. */
5592 prolog_skip_comment (&lb, inf);
5593 else if ((len = prolog_pr (cp, last)) > 0)
5595 /* Predicate or rule. Store the function name so that we
5596 only generate a tag for the first clause. */
5598 last = xnew(len + 1, char);
5599 else if (len + 1 > allocated)
5600 xrnew (last, len + 1, char);
5601 allocated = len + 1;
5602 strncpy (last, cp, len);
5611 prolog_skip_comment (plb, inf)
5619 for (cp = plb->buffer; *cp != '\0'; cp++)
5620 if (cp[0] == '*' && cp[1] == '/')
5622 readline (plb, inf);
5628 * A predicate or rule definition is added if it matches:
5629 * <beginning of line><Prolog Atom><whitespace>(
5630 * or <beginning of line><Prolog Atom><whitespace>:-
5632 * It is added to the tags database if it doesn't match the
5633 * name of the previous clause header.
5635 * Return the size of the name of the predicate or rule, or 0 if no
5641 char *last; /* Name of last clause. */
5646 pos = prolog_atom (s, 0);
5651 pos = skip_spaces (s + pos) - s;
5654 || (s[pos] == '(' && (pos += 1))
5655 || (s[pos] == ':' && s[pos + 1] == '-' && (pos += 2)))
5656 && (last == NULL /* save only the first clause */
5657 || len != (int)strlen (last)
5658 || !strneq (s, last, len)))
5660 make_tag (s, len, TRUE, s, pos, lineno, linecharno);
5668 * Consume a Prolog atom.
5669 * Return the number of bytes consumed, or -1 if there was an error.
5671 * A prolog atom, in this context, could be one of:
5672 * - An alphanumeric sequence, starting with a lower case letter.
5673 * - A quoted arbitrary string. Single quotes can escape themselves.
5674 * Backslash quotes everything.
5677 prolog_atom (s, pos)
5685 if (ISLOWER(s[pos]) || (s[pos] == '_'))
5687 /* The atom is unquoted. */
5689 while (ISALNUM(s[pos]) || (s[pos] == '_'))
5693 return pos - origpos;
5695 else if (s[pos] == '\'')
5706 pos++; /* A double quote */
5708 else if (s[pos] == '\0')
5709 /* Multiline quoted atoms are ignored. */
5711 else if (s[pos] == '\\')
5713 if (s[pos+1] == '\0')
5720 return pos - origpos;
5728 * Support for Erlang
5730 * Generates tags for functions, defines, and records.
5731 * Assumes that Erlang functions start at column 0.
5732 * Original code by Anders Lindgren (1996)
5734 static int erlang_func __P((char *, char *));
5735 static void erlang_attribute __P((char *));
5736 static int erlang_atom __P((char *));
5739 Erlang_functions (inf)
5750 LOOP_ON_INPUT_LINES (inf, lb, cp)
5752 if (cp[0] == '\0') /* Empty line */
5754 else if (iswhite (cp[0])) /* Not function nor attribute */
5756 else if (cp[0] == '%') /* comment */
5758 else if (cp[0] == '"') /* Sometimes, strings start in column one */
5760 else if (cp[0] == '-') /* attribute, e.g. "-define" */
5762 erlang_attribute (cp);
5769 else if ((len = erlang_func (cp, last)) > 0)
5772 * Function. Store the function name so that we only
5773 * generates a tag for the first clause.
5776 last = xnew (len + 1, char);
5777 else if (len + 1 > allocated)
5778 xrnew (last, len + 1, char);
5779 allocated = len + 1;
5780 strncpy (last, cp, len);
5789 * A function definition is added if it matches:
5790 * <beginning of line><Erlang Atom><whitespace>(
5792 * It is added to the tags database if it doesn't match the
5793 * name of the previous clause header.
5795 * Return the size of the name of the function, or 0 if no function
5799 erlang_func (s, last)
5801 char *last; /* Name of last clause. */
5806 pos = erlang_atom (s);
5811 pos = skip_spaces (s + pos) - s;
5813 /* Save only the first clause. */
5816 || len != (int)strlen (last)
5817 || !strneq (s, last, len)))
5819 make_tag (s, len, TRUE, s, pos, lineno, linecharno);
5828 * Handle attributes. Currently, tags are generated for defines
5831 * They are on the form:
5832 * -define(foo, bar).
5833 * -define(Foo(M, N), M+N).
5834 * -record(graph, {vtab = notable, cyclic = true}).
5837 erlang_attribute (s)
5842 if ((LOOKING_AT (cp, "-define") || LOOKING_AT (cp, "-record"))
5845 int len = erlang_atom (skip_spaces (cp));
5847 make_tag (cp, len, TRUE, s, cp + len - s, lineno, linecharno);
5854 * Consume an Erlang atom (or variable).
5855 * Return the number of bytes consumed, or -1 if there was an error.
5863 if (ISALPHA (s[pos]) || s[pos] == '_')
5865 /* The atom is unquoted. */
5868 while (ISALNUM (s[pos]) || s[pos] == '_');
5870 else if (s[pos] == '\'')
5872 for (pos++; s[pos] != '\''; pos++)
5873 if (s[pos] == '\0' /* multiline quoted atoms are ignored */
5874 || (s[pos] == '\\' && s[++pos] == '\0'))
5883 static char *scan_separators __P((char *));
5884 static void add_regex __P((char *, language *));
5885 static char *substitute __P((char *, char *, struct re_registers *));
5888 * Take a string like "/blah/" and turn it into "blah", verifying
5889 * that the first and last characters are the same, and handling
5890 * quoted separator characters. Actually, stops on the occurrence of
5891 * an unquoted separator. Also process \t, \n, etc. and turn into
5892 * appropriate characters. Works in place. Null terminates name string.
5893 * Returns pointer to terminating separator, or NULL for
5894 * unterminated regexps.
5897 scan_separators (name)
5901 char *copyto = name;
5902 bool quoted = FALSE;
5904 for (++name; *name != '\0'; ++name)
5910 case 'a': *copyto++ = '\007'; break; /* BEL (bell) */
5911 case 'b': *copyto++ = '\b'; break; /* BS (back space) */
5912 case 'd': *copyto++ = 0177; break; /* DEL (delete) */
5913 case 'e': *copyto++ = 033; break; /* ESC (delete) */
5914 case 'f': *copyto++ = '\f'; break; /* FF (form feed) */
5915 case 'n': *copyto++ = '\n'; break; /* NL (new line) */
5916 case 'r': *copyto++ = '\r'; break; /* CR (carriage return) */
5917 case 't': *copyto++ = '\t'; break; /* TAB (horizontal tab) */
5918 case 'v': *copyto++ = '\v'; break; /* VT (vertical tab) */
5924 /* Something else is quoted, so preserve the quote. */
5932 else if (*name == '\\')
5934 else if (*name == sep)
5940 name = NULL; /* signal unterminated regexp */
5942 /* Terminate copied string. */
5947 /* Look at the argument of --regex or --no-regex and do the right
5948 thing. Same for each line of a regexp file. */
5950 analyse_regex (regex_arg)
5953 if (regex_arg == NULL)
5955 free_regexps (); /* --no-regex: remove existing regexps */
5959 /* A real --regexp option or a line in a regexp file. */
5960 switch (regex_arg[0])
5962 /* Comments in regexp file or null arg to --regex. */
5968 /* Read a regex file. This is recursive and may result in a
5969 loop, which will stop when the file descriptors are exhausted. */
5973 linebuffer regexbuf;
5974 char *regexfile = regex_arg + 1;
5976 /* regexfile is a file containing regexps, one per line. */
5977 regexfp = fopen (regexfile, "r");
5978 if (regexfp == NULL)
5983 linebuffer_init (®exbuf);
5984 while (readline_internal (®exbuf, regexfp) > 0)
5985 analyse_regex (regexbuf.buffer);
5986 free (regexbuf.buffer);
5991 /* Regexp to be used for a specific language only. */
5995 char *lang_name = regex_arg + 1;
5998 for (cp = lang_name; *cp != '}'; cp++)
6001 error ("unterminated language name in regex: %s", regex_arg);
6005 lang = get_language_from_langname (lang_name);
6008 add_regex (cp, lang);
6012 /* Regexp to be used for any language. */
6014 add_regex (regex_arg, NULL);
6019 /* Separate the regexp pattern, compile it,
6020 and care for optional name and modifiers. */
6022 add_regex (regexp_pattern, lang)
6023 char *regexp_pattern;
6026 static struct re_pattern_buffer zeropattern;
6027 char sep, *pat, *name, *modifiers;
6029 struct re_pattern_buffer *patbuf;
6032 force_explicit_name = TRUE, /* do not use implicit tag names */
6033 ignore_case = FALSE, /* case is significant */
6034 multi_line = FALSE, /* matches are done one line at a time */
6035 single_line = FALSE; /* dot does not match newline */
6038 if (strlen(regexp_pattern) < 3)
6040 error ("null regexp", (char *)NULL);
6043 sep = regexp_pattern[0];
6044 name = scan_separators (regexp_pattern);
6047 error ("%s: unterminated regexp", regexp_pattern);
6052 error ("null name for regexp \"%s\"", regexp_pattern);
6055 modifiers = scan_separators (name);
6056 if (modifiers == NULL) /* no terminating separator --> no name */
6062 modifiers += 1; /* skip separator */
6064 /* Parse regex modifiers. */
6065 for (; modifiers[0] != '\0'; modifiers++)
6066 switch (modifiers[0])
6069 if (modifiers == name)
6070 error ("forcing explicit tag name but no name, ignoring", NULL);
6071 force_explicit_name = TRUE;
6081 need_filebuf = TRUE;
6086 wrongmod[0] = modifiers[0];
6088 error ("invalid regexp modifier `%s', ignoring", wrongmod);
6093 patbuf = xnew (1, struct re_pattern_buffer);
6094 *patbuf = zeropattern;
6097 static char lc_trans[CHARS];
6099 for (i = 0; i < CHARS; i++)
6100 lc_trans[i] = lowcase (i);
6101 patbuf->translate = lc_trans; /* translation table to fold case */
6105 pat = concat ("^", regexp_pattern, ""); /* anchor to beginning of line */
6107 pat = regexp_pattern;
6110 re_set_syntax (RE_SYNTAX_EMACS | RE_DOT_NEWLINE);
6112 re_set_syntax (RE_SYNTAX_EMACS);
6114 err = re_compile_pattern (pat, strlen (regexp_pattern), patbuf);
6119 error ("%s while compiling pattern", err);
6124 p_head = xnew (1, regexp);
6125 p_head->pattern = savestr (regexp_pattern);
6126 p_head->p_next = rp;
6127 p_head->lang = lang;
6128 p_head->pat = patbuf;
6129 p_head->name = savestr (name);
6130 p_head->error_signaled = FALSE;
6131 p_head->force_explicit_name = force_explicit_name;
6132 p_head->ignore_case = ignore_case;
6133 p_head->multi_line = multi_line;
6137 * Do the substitutions indicated by the regular expression and
6141 substitute (in, out, regs)
6143 struct re_registers *regs;
6146 int size, dig, diglen;
6149 size = strlen (out);
6151 /* Pass 1: figure out how much to allocate by finding all \N strings. */
6152 if (out[size - 1] == '\\')
6153 fatal ("pattern error in \"%s\"", out);
6154 for (t = etags_strchr (out, '\\');
6156 t = etags_strchr (t + 2, '\\'))
6160 diglen = regs->end[dig] - regs->start[dig];
6166 /* Allocate space and do the substitutions. */
6168 result = xnew (size + 1, char);
6170 for (t = result; *out != '\0'; out++)
6171 if (*out == '\\' && ISDIGIT (*++out))
6174 diglen = regs->end[dig] - regs->start[dig];
6175 strncpy (t, in + regs->start[dig], diglen);
6182 assert (t <= result + size);
6183 assert (t - result == (int)strlen (result));
6188 /* Deallocate all regexps. */
6193 while (p_head != NULL)
6195 rp = p_head->p_next;
6196 free (p_head->pattern);
6197 free (p_head->name);
6205 * Reads the whole file as a single string from `filebuf' and looks for
6206 * multi-line regular expressions, creating tags on matches.
6207 * readline already dealt with normal regexps.
6209 * Idea by Ben Wing <ben@666.com> (2002).
6212 regex_tag_multiline ()
6214 char *buffer = filebuf.buffer;
6218 for (rp = p_head; rp != NULL; rp = rp->p_next)
6222 if (!rp->multi_line)
6223 continue; /* skip normal regexps */
6225 /* Generic initialisations before parsing file from memory. */
6226 lineno = 1; /* reset global line number */
6227 charno = 0; /* reset global char number */
6228 linecharno = 0; /* reset global char number of line start */
6230 /* Only use generic regexps or those for the current language. */
6231 if (rp->lang != NULL && rp->lang != curfdp->lang)
6234 while (match >= 0 && match < filebuf.len)
6236 match = re_search (rp->pat, buffer, filebuf.len, charno,
6237 filebuf.len - match, &rp->regs);
6242 if (!rp->error_signaled)
6244 error ("regexp stack overflow while matching \"%s\"",
6246 rp->error_signaled = TRUE;
6253 if (match == rp->regs.end[0])
6255 if (!rp->error_signaled)
6257 error ("regexp matches the empty string: \"%s\"",
6259 rp->error_signaled = TRUE;
6261 match = -3; /* exit from while loop */
6265 /* Match occurred. Construct a tag. */
6266 while (charno < rp->regs.end[0])
6267 if (buffer[charno++] == '\n')
6268 lineno++, linecharno = charno;
6269 if (! rp->name || rp->name[0] == '\0')
6271 else /* make a named tag */
6272 name = substitute (buffer, rp->name, &rp->regs);
6273 if (rp->force_explicit_name)
6274 /* Force explicit tag name, if a name
6276 pfnote (name, TRUE, buffer + linecharno,
6277 charno - linecharno + 1, lineno,
6279 else if(name == NULL)
6282 make_tag (name, strlen (name), TRUE,
6283 buffer + linecharno,
6284 charno - linecharno + 1,
6285 lineno, linecharno);
6299 register int len = 0;
6301 while (*cp != '\0' && lowcase (*cp) == lowcase (dbp[len]))
6303 if (*cp == '\0' && !intoken (dbp[len]))
6312 get_tag (bp, namepp)
6316 register char *cp = bp;
6320 /* Go till you get to white space or a syntactic break */
6321 for (cp = bp + 1; !notinname (*cp); cp++)
6323 make_tag (bp, cp - bp, TRUE,
6324 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
6328 *namepp = savenstr (bp, cp - bp);
6332 * Read a line of text from `stream' into `lbp', excluding the
6333 * newline or CR-NL, if any. Return the number of characters read from
6334 * `stream', which is the length of the line including the newline.
6336 * On DOS or Windows we do not count the CR character, if any before the
6337 * NL, in the returned length; this mirrors the behavior of Emacs on those
6338 * platforms (for text files, it translates CR-NL to NL as it reads in the
6341 * If multi-line regular expressions are requested, each line read is
6342 * appended to `filebuf'.
6345 readline_internal (lbp, stream)
6347 register FILE *stream;
6349 char *buffer = lbp->buffer;
6350 register char *p = lbp->buffer;
6351 register char *pend;
6354 pend = p + lbp->size; /* Separate to avoid 386/IX compiler bug. */
6358 register int c = getc (stream);
6361 /* We're at the end of linebuffer: expand it. */
6363 xrnew (buffer, lbp->size, char);
6364 p += buffer - lbp->buffer;
6365 pend = buffer + lbp->size;
6366 lbp->buffer = buffer;
6376 if (p > buffer && p[-1] == '\r')
6390 lbp->len = p - buffer;
6392 if (need_filebuf /* we need filebuf for multi-line regexps */
6393 && chars_deleted > 0) /* not at EOF */
6395 while (filebuf.size <= filebuf.len + lbp->len + 1) /* +1 for \n */
6397 /* Expand filebuf. */
6399 xrnew (filebuf.buffer, filebuf.size, char);
6401 strncpy (filebuf.buffer + filebuf.len, lbp->buffer, lbp->len);
6402 filebuf.len += lbp->len;
6403 filebuf.buffer[filebuf.len++] = '\n';
6404 filebuf.buffer[filebuf.len] = '\0';
6407 return lbp->len + chars_deleted;
6411 * Like readline_internal, above, but in addition try to match the
6412 * input line against relevant regular expressions and manage #line
6416 readline (lbp, stream)
6422 linecharno = charno; /* update global char number of line start */
6423 result = readline_internal (lbp, stream); /* read line */
6424 lineno += 1; /* increment global line number */
6425 charno += result; /* increment global char number */
6427 /* Honour #line directives. */
6428 if (!no_line_directive)
6430 static bool discard_until_line_directive;
6432 /* Check whether this is a #line directive. */
6433 if (result > 12 && strneq (lbp->buffer, "#line ", 6))
6438 if (sscanf (lbp->buffer, "#line %u \"%n", &lno, &start) >= 1
6439 && start > 0) /* double quote character found */
6441 char *endp = lbp->buffer + start;
6443 while ((endp = etags_strchr (endp, '"')) != NULL
6444 && endp[-1] == '\\')
6447 /* Ok, this is a real #line directive. Let's deal with it. */
6449 char *taggedabsname; /* absolute name of original file */
6450 char *taggedfname; /* name of original file as given */
6451 char *name; /* temp var */
6453 discard_until_line_directive = FALSE; /* found it */
6454 name = lbp->buffer + start;
6456 canonicalize_filename (name);
6457 taggedabsname = absolute_filename (name, tagfiledir);
6458 if (filename_is_absolute (name)
6459 || filename_is_absolute (curfdp->infname))
6460 taggedfname = savestr (taggedabsname);
6462 taggedfname = relative_filename (taggedabsname,tagfiledir);
6464 if (streq (curfdp->taggedfname, taggedfname))
6465 /* The #line directive is only a line number change. We
6466 deal with this afterwards. */
6469 /* The tags following this #line directive should be
6470 attributed to taggedfname. In order to do this, set
6471 curfdp accordingly. */
6473 fdesc *fdp; /* file description pointer */
6475 /* Go look for a file description already set up for the
6476 file indicated in the #line directive. If there is
6477 one, use it from now until the next #line
6479 for (fdp = fdhead; fdp != NULL; fdp = fdp->next)
6480 if (streq (fdp->infname, curfdp->infname)
6481 && streq (fdp->taggedfname, taggedfname))
6482 /* If we remove the second test above (after the &&)
6483 then all entries pertaining to the same file are
6484 coalesced in the tags file. If we use it, then
6485 entries pertaining to the same file but generated
6486 from different files (via #line directives) will
6487 go into separate sections in the tags file. These
6488 alternatives look equivalent. The first one
6489 destroys some apparently useless information. */
6495 /* Else, if we already tagged the real file, skip all
6496 input lines until the next #line directive. */
6497 if (fdp == NULL) /* not found */
6498 for (fdp = fdhead; fdp != NULL; fdp = fdp->next)
6499 if (streq (fdp->infabsname, taggedabsname))
6501 discard_until_line_directive = TRUE;
6505 /* Else create a new file description and use that from
6506 now on, until the next #line directive. */
6507 if (fdp == NULL) /* not found */
6510 fdhead = xnew (1, fdesc);
6511 *fdhead = *curfdp; /* copy curr. file description */
6513 fdhead->infname = savestr (curfdp->infname);
6514 fdhead->infabsname = savestr (curfdp->infabsname);
6515 fdhead->infabsdir = savestr (curfdp->infabsdir);
6516 fdhead->taggedfname = taggedfname;
6517 fdhead->usecharno = FALSE;
6518 fdhead->prop = NULL;
6519 fdhead->written = FALSE;
6523 free (taggedabsname);
6525 readline (lbp, stream);
6527 } /* if a real #line directive */
6528 } /* if #line is followed by a a number */
6529 } /* if line begins with "#line " */
6531 /* If we are here, no #line directive was found. */
6532 if (discard_until_line_directive)
6536 /* Do a tail recursion on ourselves, thus discarding the contents
6537 of the line buffer. */
6538 readline (lbp, stream);
6542 discard_until_line_directive = FALSE;
6545 } /* if #line directives should be considered */
6552 /* Match against relevant regexps. */
6554 for (rp = p_head; rp != NULL; rp = rp->p_next)
6556 /* Only use generic regexps or those for the current language.
6557 Also do not use multiline regexps, which is the job of
6558 regex_tag_multiline. */
6559 if ((rp->lang != NULL && rp->lang != fdhead->lang)
6563 match = re_match (rp->pat, lbp->buffer, lbp->len, 0, &rp->regs);
6568 if (!rp->error_signaled)
6570 error ("regexp stack overflow while matching \"%s\"",
6572 rp->error_signaled = TRUE;
6579 /* Empty string matched. */
6580 if (!rp->error_signaled)
6582 error ("regexp matches the empty string: \"%s\"", rp->pattern);
6583 rp->error_signaled = TRUE;
6587 /* Match occurred. Construct a tag. */
6588 if (rp->name[0] != '\0')
6589 /* make a named tag */
6590 name = substitute (lbp->buffer, rp->name, &rp->regs);
6591 if (rp->force_explicit_name)
6592 /* Force explicit tag name, if a name is there. */
6593 pfnote (name, TRUE, lbp->buffer, match, lineno, linecharno);
6595 make_tag (name, strlen (name), TRUE,
6596 lbp->buffer, match, lineno, linecharno);
6600 make_tag (rp->name, strlen (rp->name), TRUE,
6601 lbp->buffer, match, lineno, linecharno);
6610 * Return a pointer to a space of size strlen(cp)+1 allocated
6611 * with xnew where the string CP has been copied.
6617 return savenstr (cp, strlen (cp));
6621 * Return a pointer to a space of size LEN+1 allocated with xnew where
6622 * the string CP has been copied for at most the first LEN characters.
6631 dp = xnew (len + 1, char);
6632 strncpy (dp, cp, len);
6638 * Return the ptr in sp at which the character c last
6639 * appears; NULL if not found
6641 * Identical to POSIX strrchr, included for portability.
6644 etags_strrchr (sp, c)
6645 register const char *sp;
6648 register const char *r;
6660 * Return the ptr in sp at which the character c first
6661 * appears; NULL if not found
6663 * Identical to POSIX strchr, included for portability.
6666 etags_strchr (sp, c)
6667 register const char *sp;
6679 * Compare two strings, ignoring case for alphabetic characters.
6681 * Same as BSD's strcasecmp, included for portability.
6684 etags_strcasecmp (s1, s2)
6685 register const char *s1;
6686 register const char *s2;
6689 && (ISALPHA (*s1) && ISALPHA (*s2)
6690 ? lowcase (*s1) == lowcase (*s2)
6694 return (ISALPHA (*s1) && ISALPHA (*s2)
6695 ? lowcase (*s1) - lowcase (*s2)
6700 * Compare two strings, ignoring case for alphabetic characters.
6701 * Stop after a given number of characters
6703 * Same as BSD's strncasecmp, included for portability.
6706 etags_strncasecmp (s1, s2, n)
6707 register const char *s1;
6708 register const char *s2;
6711 while (*s1 != '\0' && n-- > 0
6712 && (ISALPHA (*s1) && ISALPHA (*s2)
6713 ? lowcase (*s1) == lowcase (*s2)
6720 return (ISALPHA (*s1) && ISALPHA (*s2)
6721 ? lowcase (*s1) - lowcase (*s2)
6725 /* Skip spaces (end of string is not space), return new pointer. */
6730 while (iswhite (*cp))
6735 /* Skip non spaces, except end of string, return new pointer. */
6737 skip_non_spaces (cp)
6740 while (*cp != '\0' && !iswhite (*cp))
6745 /* Print error message and exit. */
6751 exit (EXIT_FAILURE);
6759 exit (EXIT_FAILURE);
6763 suggest_asking_for_help ()
6765 fprintf (stderr, "\tTry `%s %s' for a complete list of options.\n",
6766 progname, NO_LONG_OPTIONS ? "-h" : "--help");
6767 exit (EXIT_FAILURE);
6770 /* Print error message. `s1' is printf control string, `s2' is arg for it. */
6773 const char *s1, *s2;
6775 fprintf (stderr, "%s: ", progname);
6776 fprintf (stderr, s1, s2);
6777 fprintf (stderr, "\n");
6780 /* Return a newly-allocated string whose contents
6781 concatenate those of s1, s2, s3. */
6786 int len1 = strlen (s1), len2 = strlen (s2), len3 = strlen (s3);
6787 char *result = xnew (len1 + len2 + len3 + 1, char);
6789 strncpy(result, s1, len1+1);
6790 strncpy(result + len1, s2, len2+1);
6791 strncpy(result + len1 + len2, s3, len3+1);
6792 result[len1 + len2 + len3] = '\0';
6798 /* Does the same work as the system V getcwd, but does not need to
6799 guess the buffer size in advance. */
6805 char *path = xnew (bufsize, char);
6807 while (getcwd (path, bufsize) == NULL)
6809 if (errno != ERANGE)
6813 path = xnew (bufsize, char);
6816 canonicalize_filename (path);
6819 #else /* not HAVE_GETCWD */
6823 linebuffer_init (&path);
6824 pipe = (FILE *) popen ("pwd 2>/dev/null", "r");
6825 if (pipe == NULL || readline_internal (&path, pipe) == 0)
6830 #endif /* not HAVE_GETCWD */
6833 /* Return a newly allocated string containing the file name of FILE
6834 relative to the absolute directory DIR (which should end with a slash). */
6836 relative_filename (file, dir)
6839 char *fp, *dp, *afn, *res;
6843 /* Find the common root of file and dir (with a trailing slash). */
6844 afn = absolute_filename (file, cwd);
6847 while (*fp++ == *dp++)
6849 fp--, dp--; /* back to the first differing char */
6850 do /* look at the equal chars until '/' */
6853 fp ++; /* Advance past the '/' */
6855 /* Build a sequence of "../" strings for the resulting relative file name. */
6857 while ((dp = etags_strchr (dp + 1, '/')) != NULL)
6859 res_left = 3 * i + strlen(fp);
6860 res = xnew( res_left + 1, char);
6862 for ( ; i-- > 0 ; res_left -= 4 )
6863 strncat(res, "../", res_left );
6865 /* Add the file name relative to the common root of file and dir. */
6866 strncat(res, fp, res_left);
6872 /* Return a newly allocated string containing the absolute file name
6873 of FILE given DIR (which should end with a slash). */
6875 absolute_filename (file, dir)
6878 char *slashp, *cp, *res;
6880 if (filename_is_absolute (file))
6881 res = savestr (file);
6883 res = concat (dir, file, "");
6885 /* Delete the "/dirname/.." and "/." substrings. */
6886 slashp = etags_strchr (res, '/');
6887 while (slashp != NULL && slashp[0] != '\0')
6889 if (slashp[1] == '.')
6891 if (slashp[2] == '.'
6892 && (slashp[3] == '/' || slashp[3] == '\0'))
6897 while (cp >= res && !filename_is_absolute (cp));
6899 cp = slashp; /* the absolute name begins with "/.." */
6900 strcpy (cp, slashp + 3);
6904 else if (slashp[2] == '/' || slashp[2] == '\0')
6906 strcpy (slashp, slashp + 2);
6911 slashp = etags_strchr (slashp + 1, '/');
6914 if (res[0] == '\0') /* just a safety net: should never happen */
6917 return savestr ("/");
6923 /* Return a newly allocated string containing the absolute
6924 file name of dir where FILE resides given DIR (which should
6925 end with a slash). */
6927 absolute_dirname (file, dir)
6933 slashp = etags_strrchr (file, '/');
6935 return savestr (dir);
6938 res = absolute_filename (file, dir);
6944 /* Whether the argument string is an absolute file name. The argument
6945 string must have been canonicalized with canonicalize_filename. */
6947 filename_is_absolute (fn)
6950 return (fn[0] == '/');
6953 /* Upcase DOS drive letter and collapse separators into single slashes.
6956 canonicalize_filename (fn)
6962 /* Collapse multiple separators into a single slash. */
6963 for (cp = fn; *cp != '\0'; cp++, fn++)
6967 while (cp[1] == sep)
6976 /* Initialize a linebuffer for use. */
6978 linebuffer_init (lbp)
6981 lbp->size = (DEBUG) ? 3 : 200;
6982 lbp->buffer = xnew (lbp->size, char);
6983 lbp->buffer[0] = '\0';
6987 /* Set the minimum size of a string contained in a linebuffer. */
6989 linebuffer_setlen (lbp, toksize)
6993 while (lbp->size <= toksize)
6996 xrnew (lbp->buffer, lbp->size, char);
7001 /* Like malloc but get fatal error if memory is exhausted. */
7006 PTR result = (PTR) malloc (size);
7008 fatal ("virtual memory exhausted", (char *)NULL);
7013 xrealloc (ptr, size)
7017 PTR result = (PTR) realloc (ptr, size);
7019 fatal ("virtual memory exhausted", (char *)NULL);
7025 * indent-tabs-mode: t
7028 * c-font-lock-extra-types: ("FILE" "bool" "language" "linebuffer" "fdesc" "node" "regexp")
7032 /* etags.c ends here */