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;
1522 char *compressed_name, *uncompressed_name;
1523 char *ext, *real_name;
1526 canonicalize_filename (file);
1527 if (streq (file, tagfile) && !streq (tagfile, "-"))
1529 error ("skipping inclusion of %s in self.", file);
1532 if ((compr = get_compressor_from_suffix (file, &ext)) == NULL)
1534 compressed_name = NULL;
1535 real_name = uncompressed_name = savestr (file);
1539 real_name = compressed_name = savestr (file);
1540 uncompressed_name = savenstr (file, ext - file);
1543 /* If the canonicalized uncompressed name
1544 has already been dealt with, skip it silently. */
1545 for (fdp = fdhead; fdp != NULL; fdp = fdp->next)
1547 assert (fdp->infname != NULL);
1548 if (streq (uncompressed_name, fdp->infname))
1552 if (stat (real_name, &stat_buf) != 0)
1554 /* Reset real_name and try with a different name. */
1556 if (compressed_name != NULL) /* try with the given suffix */
1558 if (stat (uncompressed_name, &stat_buf) == 0)
1559 real_name = uncompressed_name;
1561 else /* try all possible suffixes */
1563 for (compr = compressors; compr->suffix != NULL; compr++)
1565 compressed_name = concat (file, ".", compr->suffix);
1566 if (stat (compressed_name, &stat_buf) != 0)
1568 free (compressed_name);
1569 compressed_name = NULL;
1573 real_name = compressed_name;
1578 if (real_name == NULL)
1583 } /* try with a different name */
1585 if (!S_ISREG (stat_buf.st_mode))
1587 error ("skipping %s: it is not a regular file.", real_name);
1590 if (real_name == compressed_name)
1592 char *cmd = concat (compr->command, " ", real_name);
1593 inf = (FILE *) popen (cmd, "r");
1597 inf = fopen (real_name, "r");
1604 process_file (inf, uncompressed_name, lang);
1606 if (real_name == compressed_name)
1607 retval = pclose (inf);
1609 retval = fclose (inf);
1614 free (compressed_name);
1615 free (uncompressed_name);
1622 process_file (fh, fn, lang)
1627 static const fdesc emptyfdesc;
1630 /* Create a new input file description entry. */
1631 fdp = xnew (1, fdesc);
1634 fdp->infname = savestr (fn);
1636 fdp->infabsname = absolute_filename (fn, cwd);
1637 fdp->infabsdir = absolute_dirname (fn, cwd);
1638 if (filename_is_absolute (fn))
1640 /* An absolute file name. Canonicalize it. */
1641 fdp->taggedfname = absolute_filename (fn, NULL);
1645 /* A file name relative to cwd. Make it relative
1646 to the directory of the tags file. */
1647 fdp->taggedfname = relative_filename (fn, tagfiledir);
1649 fdp->usecharno = TRUE; /* use char position when making tags */
1651 fdp->written = FALSE; /* not written on tags file yet */
1654 curfdp = fdhead; /* the current file description */
1658 /* If not Ctags, and if this is not metasource and if it contained no #line
1659 directives, we can write the tags and free all nodes pointing to
1662 && curfdp->usecharno /* no #line directives in this file */
1663 && !curfdp->lang->metasource)
1667 /* Look for the head of the sublist relative to this file. See add_node
1668 for the structure of the node tree. */
1670 for (np = nodehead; np != NULL; prev = np, np = np->left)
1671 if (np->fdp == curfdp)
1674 /* If we generated tags for this file, write and delete them. */
1677 /* This is the head of the last sublist, if any. The following
1678 instructions depend on this being true. */
1679 assert (np->left == NULL);
1681 assert (fdhead == curfdp);
1682 assert (last_node->fdp == curfdp);
1683 put_entries (np); /* write tags for file curfdp->taggedfname */
1684 free_tree (np); /* remove the written nodes */
1686 nodehead = NULL; /* no nodes left */
1688 prev->left = NULL; /* delete the pointer to the sublist */
1694 * This routine sets up the boolean pseudo-functions which work
1695 * by setting boolean flags dependent upon the corresponding character.
1696 * Every char which is NOT in that string is not a white char. Therefore,
1697 * all of the array "_wht" is set to FALSE, and then the elements
1698 * subscripted by the chars in "white" are set to TRUE. Thus "_wht"
1699 * of a char is TRUE if it is the string "white", else FALSE.
1707 for (i = 0; i < CHARS; i++)
1708 iswhite(i) = notinname(i) = begtoken(i) = intoken(i) = endtoken(i) = FALSE;
1709 for (sp = white; *sp != '\0'; sp++) iswhite (*sp) = TRUE;
1710 for (sp = nonam; *sp != '\0'; sp++) notinname (*sp) = TRUE;
1711 notinname('\0') = notinname('\n');
1712 for (sp = begtk; *sp != '\0'; sp++) begtoken (*sp) = TRUE;
1713 begtoken('\0') = begtoken('\n');
1714 for (sp = midtk; *sp != '\0'; sp++) intoken (*sp) = TRUE;
1715 intoken('\0') = intoken('\n');
1716 for (sp = endtk; *sp != '\0'; sp++) endtoken (*sp) = TRUE;
1717 endtoken('\0') = endtoken('\n');
1721 * This routine opens the specified file and calls the function
1722 * which finds the function and type definitions.
1729 language *lang = curfdp->lang;
1730 Lang_function *parser = NULL;
1732 /* If user specified a language, use it. */
1733 if (lang != NULL && lang->function != NULL)
1735 parser = lang->function;
1738 /* Else try to guess the language given the file name. */
1741 lang = get_language_from_filename (curfdp->infname, TRUE);
1742 if (lang != NULL && lang->function != NULL)
1744 curfdp->lang = lang;
1745 parser = lang->function;
1749 /* Else look for sharp-bang as the first two characters. */
1751 && readline_internal (&lb, inf) > 0
1753 && lb.buffer[0] == '#'
1754 && lb.buffer[1] == '!')
1758 /* Set lp to point at the first char after the last slash in the
1759 line or, if no slashes, at the first nonblank. Then set cp to
1760 the first successive blank and terminate the string. */
1761 lp = etags_strrchr (lb.buffer+2, '/');
1765 lp = skip_spaces (lb.buffer + 2);
1766 cp = skip_non_spaces (lp);
1769 if (strlen (lp) > 0)
1771 lang = get_language_from_interpreter (lp);
1772 if (lang != NULL && lang->function != NULL)
1774 curfdp->lang = lang;
1775 parser = lang->function;
1780 /* We rewind here, even if inf may be a pipe. We fail if the
1781 length of the first line is longer than the pipe block size,
1782 which is unlikely. */
1785 /* Else try to guess the language given the case insensitive file name. */
1788 lang = get_language_from_filename (curfdp->infname, FALSE);
1789 if (lang != NULL && lang->function != NULL)
1791 curfdp->lang = lang;
1792 parser = lang->function;
1796 /* Else try Fortran or C. */
1799 node *old_last_node = last_node;
1801 curfdp->lang = get_language_from_langname ("fortran");
1804 if (old_last_node == last_node)
1805 /* No Fortran entries found. Try C. */
1807 /* We do not tag if rewind fails.
1808 Only the file name will be recorded in the tags file. */
1810 curfdp->lang = get_language_from_langname (cplusplus ? "c++" : "c");
1816 if (!no_line_directive
1817 && curfdp->lang != NULL && curfdp->lang->metasource)
1818 /* It may be that this is a bingo.y file, and we already parsed a bingo.c
1819 file, or anyway we parsed a file that is automatically generated from
1820 this one. If this is the case, the bingo.c file contained #line
1821 directives that generated tags pointing to this file. Let's delete
1822 them all before parsing this file, which is the real source. */
1824 fdesc **fdpp = &fdhead;
1825 while (*fdpp != NULL)
1827 && streq ((*fdpp)->taggedfname, curfdp->taggedfname))
1828 /* We found one of those! We must delete both the file description
1829 and all tags referring to it. */
1831 fdesc *badfdp = *fdpp;
1833 /* Delete the tags referring to badfdp->taggedfname
1834 that were obtained from badfdp->infname. */
1835 invalidate_nodes (badfdp, &nodehead);
1837 *fdpp = badfdp->next; /* remove the bad description from the list */
1838 free_fdesc (badfdp);
1841 fdpp = &(*fdpp)->next; /* advance the list pointer */
1844 assert (parser != NULL);
1846 /* Generic initialisations before reading from file. */
1847 linebuffer_setlen (&filebuf, 0); /* reset the file buffer */
1849 /* Generic initialisations before parsing file with readline. */
1850 lineno = 0; /* reset global line number */
1851 charno = 0; /* reset global char number */
1852 linecharno = 0; /* reset global char number of line start */
1856 regex_tag_multiline ();
1861 * Check whether an implicitly named tag should be created,
1862 * then call `pfnote'.
1863 * NAME is a string that is internally copied by this function.
1865 * TAGS format specification
1866 * Idea by Sam Kendall <kendall@mv.mv.com> (1997)
1867 * The following is explained in some more detail in etc/ETAGS.EBNF.
1869 * make_tag creates tags with "implicit tag names" (unnamed tags)
1870 * if the following are all true, assuming NONAM=" \f\t\n\r()=,;":
1871 * 1. NAME does not contain any of the characters in NONAM;
1872 * 2. LINESTART contains name as either a rightmost, or rightmost but
1873 * one character, substring;
1874 * 3. the character, if any, immediately before NAME in LINESTART must
1875 * be a character in NONAM;
1876 * 4. the character, if any, immediately after NAME in LINESTART must
1877 * also be a character in NONAM.
1879 * The implementation uses the notinname() macro, which recognises the
1880 * characters stored in the string `nonam'.
1881 * etags.el needs to use the same characters that are in NONAM.
1884 make_tag (name, namelen, is_func, linestart, linelen, lno, cno)
1885 char *name; /* tag name, or NULL if unnamed */
1886 int namelen; /* tag length */
1887 bool is_func; /* tag is a function */
1888 char *linestart; /* start of the line where tag is */
1889 int linelen; /* length of the line where tag is */
1890 int lno; /* line number */
1891 long cno; /* character number */
1893 bool named = (name != NULL && namelen > 0);
1895 if (!CTAGS && named) /* maybe set named to false */
1896 /* Let's try to make an implicit tag name, that is, create an unnamed tag
1897 such that etags.el can guess a name from it. */
1900 register char *cp = name;
1902 for (i = 0; i < namelen; i++)
1903 if (notinname (*cp++))
1905 if (i == namelen) /* rule #1 */
1907 cp = linestart + linelen - namelen;
1908 if (notinname (linestart[linelen-1]))
1909 cp -= 1; /* rule #4 */
1910 if (cp >= linestart /* rule #2 */
1912 || notinname (cp[-1])) /* rule #3 */
1913 && strneq (name, cp, namelen)) /* rule #2 */
1914 named = FALSE; /* use implicit tag name */
1919 name = savenstr (name, namelen);
1922 pfnote (name, is_func, linestart, linelen, lno, cno);
1927 pfnote (name, is_func, linestart, linelen, lno, cno)
1928 char *name; /* tag name, or NULL if unnamed */
1929 bool is_func; /* tag is a function */
1930 char *linestart; /* start of the line where tag is */
1931 int linelen; /* length of the line where tag is */
1932 int lno; /* line number */
1933 long cno; /* character number */
1937 assert (name == NULL || name[0] != '\0');
1938 if (CTAGS && name == NULL)
1941 np = xnew (1, node);
1943 /* If ctags mode, change name "main" to M<thisfilename>. */
1944 if (CTAGS && !cxref_style && streq (name, "main"))
1946 register char *fp = etags_strrchr (curfdp->taggedfname, '/');
1947 np->name = concat ("M", fp == NULL ? curfdp->taggedfname : fp + 1, "");
1948 fp = etags_strrchr (np->name, '.');
1949 if (fp != NULL && fp[1] != '\0' && fp[2] == '\0')
1955 np->been_warned = FALSE;
1957 np->is_func = is_func;
1959 if (np->fdp->usecharno)
1960 /* Our char numbers are 0-base, because of C language tradition?
1961 ctags compatibility? old versions compatibility? I don't know.
1962 Anyway, since emacs's are 1-base we expect etags.el to take care
1963 of the difference. If we wanted to have 1-based numbers, we would
1964 uncomment the +1 below. */
1965 np->cno = cno /* + 1 */ ;
1967 np->cno = invalidcharno;
1968 np->left = np->right = NULL;
1969 if (CTAGS && !cxref_style)
1971 if (strlen (linestart) < 50)
1972 np->regex = concat (linestart, "$", "");
1974 np->regex = savenstr (linestart, 50);
1977 np->regex = savenstr (linestart, linelen);
1979 add_node (np, &nodehead);
1984 * recurse on left children, iterate on right children.
1992 register node *node_right = np->right;
1993 free_tree (np->left);
2003 * delete a file description
2007 register fdesc *fdp;
2009 free (fdp->infname);
2010 free (fdp->infabsname);
2011 free (fdp->infabsdir);
2012 free (fdp->taggedfname);
2019 * Adds a node to the tree of nodes. In etags mode, sort by file
2020 * name. In ctags mode, sort by tag name. Make no attempt at
2023 * add_node is the only function allowed to add nodes, so it can
2027 add_node (np, cur_node_p)
2028 node *np, **cur_node_p;
2031 register node *cur_node = *cur_node_p;
2033 if (cur_node == NULL)
2043 /* For each file name, tags are in a linked sublist on the right
2044 pointer. The first tags of different files are a linked list
2045 on the left pointer. last_node points to the end of the last
2047 if (last_node != NULL && last_node->fdp == np->fdp)
2049 /* Let's use the same sublist as the last added node. */
2050 assert (last_node->right == NULL);
2051 last_node->right = np;
2054 else if (cur_node->fdp == np->fdp)
2056 /* Scanning the list we found the head of a sublist which is
2057 good for us. Let's scan this sublist. */
2058 add_node (np, &cur_node->right);
2061 /* The head of this sublist is not good for us. Let's try the
2063 add_node (np, &cur_node->left);
2064 } /* if ETAGS mode */
2069 dif = strcmp (np->name, cur_node->name);
2072 * If this tag name matches an existing one, then
2073 * do not add the node, but maybe print a warning.
2075 if (no_duplicates && !dif)
2077 if (np->fdp == cur_node->fdp)
2081 fprintf (stderr, "Duplicate entry in file %s, line %d: %s\n",
2082 np->fdp->infname, lineno, np->name);
2083 fprintf (stderr, "Second entry ignored\n");
2086 else if (!cur_node->been_warned && !no_warnings)
2090 "Duplicate entry in files %s and %s: %s (Warning only)\n",
2091 np->fdp->infname, cur_node->fdp->infname, np->name);
2092 cur_node->been_warned = TRUE;
2097 /* Actually add the node */
2098 add_node (np, dif < 0 ? &cur_node->left : &cur_node->right);
2099 } /* if CTAGS mode */
2103 * invalidate_nodes ()
2104 * Scan the node tree and invalidate all nodes pointing to the
2105 * given file description (CTAGS case) or free them (ETAGS case).
2108 invalidate_nodes (badfdp, npp)
2119 if (np->left != NULL)
2120 invalidate_nodes (badfdp, &np->left);
2121 if (np->fdp == badfdp)
2123 if (np->right != NULL)
2124 invalidate_nodes (badfdp, &np->right);
2128 assert (np->fdp != NULL);
2129 if (np->fdp == badfdp)
2131 *npp = np->left; /* detach the sublist from the list */
2132 np->left = NULL; /* isolate it */
2133 free_tree (np); /* free it */
2134 invalidate_nodes (badfdp, npp);
2137 invalidate_nodes (badfdp, &np->left);
2142 static int total_size_of_entries __P((node *));
2143 static int number_len __P((long));
2145 /* Length of a non-negative number's decimal representation. */
2151 while ((num /= 10) > 0)
2157 * Return total number of characters that put_entries will output for
2158 * the nodes in the linked list at the right of the specified node.
2159 * This count is irrelevant with etags.el since emacs 19.34 at least,
2160 * but is still supplied for backward compatibility.
2163 total_size_of_entries (np)
2166 register int total = 0;
2168 for (; np != NULL; np = np->right)
2171 total += strlen (np->regex) + 1; /* pat\177 */
2172 if (np->name != NULL)
2173 total += strlen (np->name) + 1; /* name\001 */
2174 total += number_len ((long) np->lno) + 1; /* lno, */
2175 if (np->cno != invalidcharno) /* cno */
2176 total += number_len (np->cno);
2177 total += 1; /* newline */
2188 static fdesc *fdp = NULL;
2193 /* Output subentries that precede this one */
2195 put_entries (np->left);
2197 /* Output this entry */
2206 fprintf (tagf, "\f\n%s,%d\n",
2207 fdp->taggedfname, total_size_of_entries (np));
2208 fdp->written = TRUE;
2210 fputs (np->regex, tagf);
2211 fputc ('\177', tagf);
2212 if (np->name != NULL)
2214 fputs (np->name, tagf);
2215 fputc ('\001', tagf);
2217 fprintf (tagf, "%d,", np->lno);
2218 if (np->cno != invalidcharno)
2219 fprintf (tagf, "%ld", np->cno);
2225 if (np->name == NULL)
2226 error ("internal error: NULL name in ctags mode.", (char *)NULL);
2231 fprintf (stdout, "%s %s %d\n",
2232 np->name, np->fdp->taggedfname, (np->lno + 63) / 64);
2234 fprintf (stdout, "%-16s %3d %-16s %s\n",
2235 np->name, np->lno, np->fdp->taggedfname, np->regex);
2239 fprintf (tagf, "%s\t%s\t", np->name, np->fdp->taggedfname);
2242 { /* function or #define macro with args */
2243 putc (searchar, tagf);
2246 for (sp = np->regex; *sp; sp++)
2248 if (*sp == '\\' || *sp == searchar)
2252 putc (searchar, tagf);
2255 { /* anything else; text pattern inadequate */
2256 fprintf (tagf, "%d", np->lno);
2261 } /* if this node contains a valid tag */
2263 /* Output subentries that follow this one */
2264 put_entries (np->right);
2266 put_entries (np->left);
2271 #define C_EXT 0x00fff /* C extensions */
2272 #define C_PLAIN 0x00000 /* C */
2273 #define C_PLPL 0x00001 /* C++ */
2274 #define C_STAR 0x00003 /* C* */
2275 #define C_JAVA 0x00005 /* JAVA */
2276 #define C_AUTO 0x01000 /* C, but switch to C++ if `class' is met */
2277 #define YACC 0x10000 /* yacc file */
2280 * The C symbol tables.
2285 st_C_objprot, st_C_objimpl, st_C_objend,
2287 st_C_ignore, st_C_attribute,
2290 st_C_class, st_C_template,
2291 st_C_struct, st_C_extern, st_C_enum, st_C_define, st_C_typedef
2294 static unsigned int hash __P((const char *, unsigned int));
2295 static struct C_stab_entry * in_word_set __P((const char *, unsigned int));
2296 static enum sym_type C_symtype __P((char *, int, int));
2298 /* Feed stuff between (but not including) %[ and %] lines to:
2304 struct C_stab_entry { char *name; int c_ext; enum sym_type type; }
2308 while, 0, st_C_ignore
2309 switch, 0, st_C_ignore
2310 return, 0, st_C_ignore
2311 __attribute__, 0, st_C_attribute
2312 GTY, 0, st_C_attribute
2313 @interface, 0, st_C_objprot
2314 @protocol, 0, st_C_objprot
2315 @implementation, 0, st_C_objimpl
2316 @end, 0, st_C_objend
2317 import, (C_JAVA & ~C_PLPL), st_C_ignore
2318 package, (C_JAVA & ~C_PLPL), st_C_ignore
2319 friend, C_PLPL, st_C_ignore
2320 extends, (C_JAVA & ~C_PLPL), st_C_javastruct
2321 implements, (C_JAVA & ~C_PLPL), st_C_javastruct
2322 interface, (C_JAVA & ~C_PLPL), st_C_struct
2323 class, 0, st_C_class
2324 namespace, C_PLPL, st_C_struct
2325 domain, C_STAR, st_C_struct
2326 union, 0, st_C_struct
2327 struct, 0, st_C_struct
2328 extern, 0, st_C_extern
2330 typedef, 0, st_C_typedef
2331 define, 0, st_C_define
2332 undef, 0, st_C_define
2333 operator, C_PLPL, st_C_operator
2334 template, 0, st_C_template
2335 # DEFUN used in emacs, the next three used in glibc (SYSCALL only for mach).
2336 DEFUN, 0, st_C_gnumacro
2337 SYSCALL, 0, st_C_gnumacro
2338 ENTRY, 0, st_C_gnumacro
2339 PSEUDO, 0, st_C_gnumacro
2340 # These are defined inside C functions, so currently they are not met.
2341 # EXFUN used in glibc, DEFVAR_* in emacs.
2342 #EXFUN, 0, st_C_gnumacro
2343 #DEFVAR_, 0, st_C_gnumacro
2345 and replace lines between %< and %> with its output, then:
2346 - remove the #if characterset check
2347 - make in_word_set static and not inline. */
2349 /* C code produced by gperf version 3.0.1 */
2350 /* Command-line: gperf -m 5 */
2351 /* Computed positions: -k'2-3' */
2353 struct C_stab_entry { char *name; int c_ext; enum sym_type type; };
2354 /* maximum key range = 33, duplicates = 0 */
2365 register const char *str;
2366 register unsigned int len;
2368 static unsigned char asso_values[] =
2370 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2371 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2372 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2373 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2374 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2375 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2376 35, 35, 35, 35, 35, 35, 35, 35, 35, 3,
2377 26, 35, 35, 35, 35, 35, 35, 35, 27, 35,
2378 35, 35, 35, 24, 0, 35, 35, 35, 35, 0,
2379 35, 35, 35, 35, 35, 1, 35, 16, 35, 6,
2380 23, 0, 0, 35, 22, 0, 35, 35, 5, 0,
2381 0, 15, 1, 35, 6, 35, 8, 19, 35, 16,
2382 4, 5, 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, 35, 35, 35, 35,
2395 35, 35, 35, 35, 35, 35
2397 register int hval = len;
2402 hval += asso_values[(unsigned char)str[2]];
2405 hval += asso_values[(unsigned char)str[1]];
2411 static struct C_stab_entry *
2412 in_word_set (str, len)
2413 register const char *str;
2414 register unsigned int len;
2418 TOTAL_KEYWORDS = 33,
2419 MIN_WORD_LENGTH = 2,
2420 MAX_WORD_LENGTH = 15,
2425 static struct C_stab_entry wordlist[] =
2428 {"if", 0, st_C_ignore},
2429 {"GTY", 0, st_C_attribute},
2430 {"@end", 0, st_C_objend},
2431 {"union", 0, st_C_struct},
2432 {"define", 0, st_C_define},
2433 {"import", (C_JAVA & ~C_PLPL), st_C_ignore},
2434 {"template", 0, st_C_template},
2435 {"operator", C_PLPL, st_C_operator},
2436 {"@interface", 0, st_C_objprot},
2437 {"implements", (C_JAVA & ~C_PLPL), st_C_javastruct},
2438 {"friend", C_PLPL, st_C_ignore},
2439 {"typedef", 0, st_C_typedef},
2440 {"return", 0, st_C_ignore},
2441 {"@implementation",0, st_C_objimpl},
2442 {"@protocol", 0, st_C_objprot},
2443 {"interface", (C_JAVA & ~C_PLPL), st_C_struct},
2444 {"extern", 0, st_C_extern},
2445 {"extends", (C_JAVA & ~C_PLPL), st_C_javastruct},
2446 {"struct", 0, st_C_struct},
2447 {"domain", C_STAR, st_C_struct},
2448 {"switch", 0, st_C_ignore},
2449 {"enum", 0, st_C_enum},
2450 {"for", 0, st_C_ignore},
2451 {"namespace", C_PLPL, st_C_struct},
2452 {"class", 0, st_C_class},
2453 {"while", 0, st_C_ignore},
2454 {"undef", 0, st_C_define},
2455 {"package", (C_JAVA & ~C_PLPL), st_C_ignore},
2456 {"__attribute__", 0, st_C_attribute},
2457 {"SYSCALL", 0, st_C_gnumacro},
2458 {"ENTRY", 0, st_C_gnumacro},
2459 {"PSEUDO", 0, st_C_gnumacro},
2460 {"DEFUN", 0, st_C_gnumacro}
2463 if (len <= MAX_WORD_LENGTH && len >= MIN_WORD_LENGTH)
2465 register int key = hash (str, len);
2467 if (key <= MAX_HASH_VALUE && key >= 0)
2469 register const char *s = wordlist[key].name;
2471 if (*str == *s && !strncmp (str + 1, s + 1, len - 1) && s[len] == '\0')
2472 return &wordlist[key];
2479 static enum sym_type
2480 C_symtype (str, len, c_ext)
2485 register struct C_stab_entry *se = in_word_set (str, len);
2487 if (se == NULL || (se->c_ext && !(c_ext & se->c_ext)))
2494 * Ignoring __attribute__ ((list))
2496 static bool inattribute; /* looking at an __attribute__ construct */
2499 * C functions and variables are recognized using a simple
2500 * finite automaton. fvdef is its state variable.
2504 fvnone, /* nothing seen */
2505 fdefunkey, /* Emacs DEFUN keyword seen */
2506 fdefunname, /* Emacs DEFUN name seen */
2507 foperator, /* func: operator keyword seen (cplpl) */
2508 fvnameseen, /* function or variable name seen */
2509 fstartlist, /* func: just after open parenthesis */
2510 finlist, /* func: in parameter list */
2511 flistseen, /* func: after parameter list */
2512 fignore, /* func: before open brace */
2513 vignore /* var-like: ignore until ';' */
2516 static bool fvextern; /* func or var: extern keyword seen; */
2519 * typedefs are recognized using a simple finite automaton.
2520 * typdef is its state variable.
2524 tnone, /* nothing seen */
2525 tkeyseen, /* typedef keyword seen */
2526 ttypeseen, /* defined type seen */
2527 tinbody, /* inside typedef body */
2528 tend, /* just before typedef tag */
2529 tignore /* junk after typedef tag */
2533 * struct-like structures (enum, struct and union) are recognized
2534 * using another simple finite automaton. `structdef' is its state
2539 snone, /* nothing seen yet,
2540 or in struct body if bracelev > 0 */
2541 skeyseen, /* struct-like keyword seen */
2542 stagseen, /* struct-like tag seen */
2543 scolonseen /* colon seen after struct-like tag */
2547 * When objdef is different from onone, objtag is the name of the class.
2549 static char *objtag = "<uninited>";
2552 * Yet another little state machine to deal with preprocessor lines.
2556 dnone, /* nothing seen */
2557 dsharpseen, /* '#' seen as first char on line */
2558 ddefineseen, /* '#' and 'define' seen */
2559 dignorerest /* ignore rest of line */
2563 * State machine for Objective C protocols and implementations.
2564 * Idea by Tom R.Hageman <tom@basil.icce.rug.nl> (1995)
2568 onone, /* nothing seen */
2569 oprotocol, /* @interface or @protocol seen */
2570 oimplementation, /* @implementations seen */
2571 otagseen, /* class name seen */
2572 oparenseen, /* parenthesis before category seen */
2573 ocatseen, /* category name seen */
2574 oinbody, /* in @implementation body */
2575 omethodsign, /* in @implementation body, after +/- */
2576 omethodtag, /* after method name */
2577 omethodcolon, /* after method colon */
2578 omethodparm, /* after method parameter */
2579 oignore /* wait for @end */
2584 * Use this structure to keep info about the token read, and how it
2585 * should be tagged. Used by the make_C_tag function to build a tag.
2589 char *line; /* string containing the token */
2590 int offset; /* where the token starts in LINE */
2591 int length; /* token length */
2593 The previous members can be used to pass strings around for generic
2594 purposes. The following ones specifically refer to creating tags. In this
2595 case the token contained here is the pattern that will be used to create a
2598 bool valid; /* do not create a tag; the token should be
2599 invalidated whenever a state machine is
2600 reset prematurely */
2601 bool named; /* create a named tag */
2602 int lineno; /* source line number of tag */
2603 long linepos; /* source char number of tag */
2604 } token; /* latest token read */
2607 * Variables and functions for dealing with nested structures.
2608 * Idea by Mykola Dzyuba <mdzyuba@yahoo.com> (2001)
2610 static void pushclass_above __P((int, char *, int));
2611 static void popclass_above __P((int));
2612 static void write_classname __P((linebuffer *, char *qualifier));
2615 char **cname; /* nested class names */
2616 int *bracelev; /* nested class brace level */
2617 int nl; /* class nesting level (elements used) */
2618 int size; /* length of the array */
2619 } cstack; /* stack for nested declaration tags */
2620 /* Current struct nesting depth (namespace, class, struct, union, enum). */
2621 #define nestlev (cstack.nl)
2622 /* After struct keyword or in struct body, not inside a nested function. */
2623 #define instruct (structdef == snone && nestlev > 0 \
2624 && bracelev == cstack.bracelev[nestlev-1] + 1)
2627 pushclass_above (bracelev, str, len)
2634 popclass_above (bracelev);
2636 if (nl >= cstack.size)
2638 int size = cstack.size *= 2;
2639 xrnew (cstack.cname, size, char *);
2640 xrnew (cstack.bracelev, size, int);
2642 assert (nl == 0 || cstack.bracelev[nl-1] < bracelev);
2643 cstack.cname[nl] = (str == NULL) ? NULL : savenstr (str, len);
2644 cstack.bracelev[nl] = bracelev;
2649 popclass_above (bracelev)
2654 for (nl = cstack.nl - 1;
2655 nl >= 0 && cstack.bracelev[nl] >= bracelev;
2658 free (cstack.cname[nl]);
2664 write_classname (cn, qualifier)
2669 int qlen = strlen (qualifier);
2671 if (cstack.nl == 0 || cstack.cname[0] == NULL)
2675 cn->buffer[0] = '\0';
2679 len = strlen (cstack.cname[0]);
2680 linebuffer_setlen (cn, len+1);
2681 strncpy (cn->buffer, cstack.cname[0],len+1);
2683 for (i = 1; i < cstack.nl; i++)
2688 s = cstack.cname[i];
2693 linebuffer_setlen (cn, len);
2694 strncat (cn->buffer, qualifier, qlen);
2695 strncat (cn->buffer, s, slen);
2700 static bool consider_token __P((char *, int, int, int *, int, int, bool *));
2701 static void make_C_tag __P((bool));
2705 * checks to see if the current token is at the start of a
2706 * function or variable, or corresponds to a typedef, or
2707 * is a struct/union/enum tag, or #define, or an enum constant.
2709 * *IS_FUNC gets TRUE if the token is a function or #define macro
2710 * with args. C_EXTP points to which language we are looking at.
2721 consider_token (str, len, c, c_extp, bracelev, parlev, is_func_or_var)
2722 register char *str; /* IN: token pointer */
2723 register int len; /* IN: token length */
2724 register int c; /* IN: first char after the token */
2725 int *c_extp; /* IN, OUT: C extensions mask */
2726 int bracelev; /* IN: brace level */
2727 int parlev; /* IN: parenthesis level */
2728 bool *is_func_or_var; /* OUT: function or variable found */
2730 /* When structdef is stagseen, scolonseen, or snone with bracelev > 0,
2731 structtype is the type of the preceding struct-like keyword, and
2732 structbracelev is the brace level where it has been seen. */
2733 static enum sym_type structtype;
2734 static int structbracelev;
2735 static enum sym_type toktype;
2738 toktype = C_symtype (str, len, *c_extp);
2741 * Skip __attribute__
2743 if (toktype == st_C_attribute)
2750 * Advance the definedef state machine.
2755 /* We're not on a preprocessor line. */
2756 if (toktype == st_C_gnumacro)
2763 if (toktype == st_C_define)
2765 definedef = ddefineseen;
2769 definedef = dignorerest;
2774 * Make a tag for any macro, unless it is a constant
2775 * and constantypedefs is FALSE.
2777 definedef = dignorerest;
2778 *is_func_or_var = (c == '(');
2779 if (!*is_func_or_var && !constantypedefs)
2786 error ("internal error: definedef value.", (char *)NULL);
2795 if (toktype == st_C_typedef)
2819 case st_C_attribute:
2820 case st_C_javastruct:
2831 if (structdef == snone && fvdef == fvnone)
2852 case st_C_attribute:
2853 case st_C_javastruct:
2873 case st_C_javastruct:
2874 if (structdef == stagseen)
2875 structdef = scolonseen;
2879 if ((*c_extp & C_AUTO) /* automatic detection of C++ language */
2881 && definedef == dnone && structdef == snone
2882 && typdef == tnone && fvdef == fvnone)
2883 *c_extp = (*c_extp | C_PLPL) & ~C_AUTO;
2884 if (toktype == st_C_template)
2891 && (typdef == tkeyseen
2892 || (typedefs_or_cplusplus && structdef == snone)))
2894 structdef = skeyseen;
2895 structtype = toktype;
2896 structbracelev = bracelev;
2897 if (fvdef == fvnameseen)
2909 case st_C_attribute:
2918 if (structdef == skeyseen)
2920 structdef = stagseen;
2924 if (typdef != tnone)
2927 /* Detect Objective C constructs. */
2937 objdef = oimplementation;
2945 case st_C_attribute:
2946 case st_C_javastruct:
2959 case oimplementation:
2960 /* Save the class tag for functions or variables defined inside. */
2961 objtag = savenstr (str, len);
2965 /* Save the class tag for categories. */
2966 objtag = savenstr (str, len);
2968 *is_func_or_var = TRUE;
2972 *is_func_or_var = TRUE;
2980 objdef = omethodtag;
2981 linebuffer_setlen (&token_name, len);
2982 strncpy (token_name.buffer, str, len);
2983 token_name.buffer[len] = '\0';
2989 objdef = omethodparm;
2995 objdef = omethodtag;
2996 linebuffer_setlen (&token_name, token_name.len + len);
2997 strncat (token_name.buffer, str, len);
3002 if (toktype == st_C_objend)
3004 /* Memory leakage here: the string pointed by objtag is
3005 never released, because many tests would be needed to
3006 avoid breaking on incorrect input code. The amount of
3007 memory leaked here is the sum of the lengths of the
3022 /* A function, variable or enum constant? */
3050 *is_func_or_var = TRUE;
3054 && structdef == snone
3055 && structtype == st_C_enum && bracelev > structbracelev)
3056 return TRUE; /* enum constant */
3062 fvdef = fdefunname; /* GNU macro */
3063 *is_func_or_var = TRUE;
3071 if ((strneq (str, "asm", 3) && endtoken (str[3]))
3072 || (strneq (str, "__asm__", 7) && endtoken (str[7])))
3089 if (len >= 10 && strneq (str+len-10, "::operator", 10))
3091 if (*c_extp & C_AUTO) /* automatic detection of C++ */
3092 *c_extp = (*c_extp | C_PLPL) & ~C_AUTO;
3094 *is_func_or_var = TRUE;
3097 if (bracelev > 0 && !instruct)
3099 fvdef = fvnameseen; /* function or variable */
3100 *is_func_or_var = TRUE;
3121 case st_C_attribute:
3122 case st_C_javastruct:
3138 * C_entries often keeps pointers to tokens or lines which are older than
3139 * the line currently read. By keeping two line buffers, and switching
3140 * them at end of line, it is possible to use those pointers.
3148 #define current_lb_is_new (newndx == curndx)
3149 #define switch_line_buffers() (curndx = 1 - curndx)
3151 #define curlb (lbs[curndx].lb)
3152 #define newlb (lbs[newndx].lb)
3153 #define curlinepos (lbs[curndx].linepos)
3154 #define newlinepos (lbs[newndx].linepos)
3156 #define plainc ((c_ext & C_EXT) == C_PLAIN)
3157 #define cplpl (c_ext & C_PLPL)
3158 #define cjava ((c_ext & C_JAVA) == C_JAVA)
3160 #define CNL_SAVE_DEFINEDEF() \
3162 curlinepos = charno; \
3163 readline (&curlb, inf); \
3164 lp = curlb.buffer; \
3171 CNL_SAVE_DEFINEDEF(); \
3172 if (savetoken.valid) \
3174 token = savetoken; \
3175 savetoken.valid = FALSE; \
3177 definedef = dnone; \
3185 /* This function is never called when token.valid is FALSE, but
3186 we must protect against invalid input or internal errors. */
3188 make_tag (token_name.buffer, token_name.len, isfun, token.line,
3189 token.offset+token.length+1, token.lineno, token.linepos);
3191 { /* this branch is optimised away if !DEBUG */
3192 make_tag (concat ("INVALID TOKEN:-->", token_name.buffer, ""),
3193 token_name.len + 17, isfun, token.line,
3194 token.offset+token.length+1, token.lineno, token.linepos);
3195 error ("INVALID TOKEN", NULL);
3198 token.valid = FALSE;
3204 * This routine finds functions, variables, typedefs,
3205 * #define's, enum constants and struct/union/enum definitions in
3206 * C syntax and adds them to the list.
3209 C_entries (c_ext, inf)
3210 int c_ext; /* extension of C */
3211 FILE *inf; /* input file */
3213 register char c; /* latest char read; '\0' for end of line */
3214 register char *lp; /* pointer one beyond the character `c' */
3215 int curndx, newndx; /* indices for current and new lb */
3216 register int tokoff; /* offset in line of start of current token */
3217 register int toklen; /* length of current token */
3218 char *qualifier; /* string used to qualify names */
3219 int qlen; /* length of qualifier */
3220 int bracelev; /* current brace level */
3221 int bracketlev; /* current bracket level */
3222 int parlev; /* current parenthesis level */
3223 int attrparlev; /* __attribute__ parenthesis level */
3224 int templatelev; /* current template level */
3225 int typdefbracelev; /* bracelev where a typedef struct body begun */
3226 bool incomm, inquote, inchar, quotednl, midtoken;
3227 bool yacc_rules; /* in the rules part of a yacc file */
3228 struct tok savetoken = {0}; /* token saved during preprocessor handling */
3231 linebuffer_init (&lbs[0].lb);
3232 linebuffer_init (&lbs[1].lb);
3233 if (cstack.size == 0)
3235 cstack.size = (DEBUG) ? 1 : 4;
3237 cstack.cname = xnew (cstack.size, char *);
3238 cstack.bracelev = xnew (cstack.size, int);
3241 tokoff = toklen = typdefbracelev = 0; /* keep compiler quiet */
3242 curndx = newndx = 0;
3246 fvdef = fvnone; fvextern = FALSE; typdef = tnone;
3247 structdef = snone; definedef = dnone; objdef = onone;
3249 midtoken = inquote = inchar = incomm = quotednl = FALSE;
3250 token.valid = savetoken.valid = FALSE;
3251 bracelev = bracketlev = parlev = attrparlev = templatelev = 0;
3253 { qualifier = "."; qlen = 1; }
3255 { qualifier = "::"; qlen = 2; }
3263 /* If we are at the end of the line, the next character is a
3264 '\0'; do not skip it, because it is what tells us
3265 to read the next line. */
3286 /* Newlines inside comments do not end macro definitions in
3288 CNL_SAVE_DEFINEDEF ();
3303 /* Newlines inside strings do not end macro definitions
3304 in traditional cpp, even though compilers don't
3305 usually accept them. */
3306 CNL_SAVE_DEFINEDEF ();
3318 /* Hmmm, something went wrong. */
3329 else if (bracketlev > 0)
3334 if (--bracketlev > 0)
3338 CNL_SAVE_DEFINEDEF ();
3373 if (fvdef != finlist && fvdef != fignore && fvdef !=vignore)
3386 else if (/* cplpl && */ *lp == '/')
3392 if ((c_ext & YACC) && *lp == '%')
3394 /* Entering or exiting rules section in yacc file. */
3396 definedef = dnone; fvdef = fvnone; fvextern = FALSE;
3397 typdef = tnone; structdef = snone;
3398 midtoken = inquote = inchar = incomm = quotednl = FALSE;
3400 yacc_rules = !yacc_rules;
3406 if (definedef == dnone)
3409 bool cpptoken = TRUE;
3411 /* Look back on this line. If all blanks, or nonblanks
3412 followed by an end of comment, this is a preprocessor
3414 for (cp = newlb.buffer; cp < lp-1; cp++)
3417 if (*cp == '*' && *(cp+1) == '/')
3426 definedef = dsharpseen;
3427 } /* if (definedef == dnone) */
3437 /* Consider token only if some involved conditions are satisfied. */
3438 if (typdef != tignore
3439 && definedef != dignorerest
3442 && (definedef != dnone
3443 || structdef != scolonseen)
3450 if (c == ':' && *lp == ':' && begtoken (lp[1]))
3451 /* This handles :: in the middle,
3452 but not at the beginning of an identifier.
3453 Also, space-separated :: is not recognised. */
3455 if (c_ext & C_AUTO) /* automatic detection of C++ */
3456 c_ext = (c_ext | C_PLPL) & ~C_AUTO;
3460 goto still_in_token;
3464 bool funorvar = FALSE;
3467 || consider_token (newlb.buffer + tokoff, toklen, c,
3468 &c_ext, bracelev, parlev,
3471 if (fvdef == foperator)
3474 lp = skip_spaces (lp-1);
3478 && !iswhite (*lp) && *lp != '(')
3481 toklen += lp - oldlp;
3483 token.named = FALSE;
3485 && nestlev > 0 && definedef == dnone)
3486 /* in struct body */
3488 write_classname (&token_name, qualifier);
3489 linebuffer_setlen (&token_name,
3490 token_name.len+qlen+toklen);
3491 strcat (token_name.buffer, qualifier);
3492 strncat (token_name.buffer,
3493 newlb.buffer + tokoff, toklen);
3496 else if (objdef == ocatseen)
3497 /* Objective C category */
3499 int len = strlen (objtag) + 2 + toklen;
3500 linebuffer_setlen (&token_name, len);
3501 strcpy (token_name.buffer, objtag);
3502 strcat (token_name.buffer, "(");
3503 strncat (token_name.buffer,
3504 newlb.buffer + tokoff, toklen);
3505 strcat (token_name.buffer, ")");
3508 else if (objdef == omethodtag
3509 || objdef == omethodparm)
3510 /* Objective C method */
3514 else if (fvdef == fdefunname)
3515 /* GNU DEFUN and similar macros */
3517 bool defun = (newlb.buffer[tokoff] == 'F');
3521 /* Rewrite the tag so that emacs lisp DEFUNs
3522 can be found by their elisp name */
3528 linebuffer_setlen (&token_name, len);
3529 strncpy (token_name.buffer,
3530 newlb.buffer + off, len);
3531 token_name.buffer[len] = '\0';
3534 if (token_name.buffer[len] == '_')
3535 token_name.buffer[len] = '-';
3536 token.named = defun;
3540 linebuffer_setlen (&token_name, toklen);
3541 strncpy (token_name.buffer,
3542 newlb.buffer + tokoff, toklen);
3543 token_name.buffer[toklen] = '\0';
3544 /* Name macros and members. */
3545 token.named = (structdef == stagseen
3546 || typdef == ttypeseen
3549 && definedef == dignorerest)
3551 && definedef == dnone
3552 && structdef == snone
3555 token.lineno = lineno;
3556 token.offset = tokoff;
3557 token.length = toklen;
3558 token.line = newlb.buffer;
3559 token.linepos = newlinepos;
3562 if (definedef == dnone
3563 && (fvdef == fvnameseen
3564 || fvdef == foperator
3565 || structdef == stagseen
3567 || typdef == ttypeseen
3568 || objdef != onone))
3570 if (current_lb_is_new)
3571 switch_line_buffers ();
3573 else if (definedef != dnone
3574 || fvdef == fdefunname
3576 make_C_tag (funorvar);
3578 else /* not yacc and consider_token failed */
3580 if (inattribute && fvdef == fignore)
3582 /* We have just met __attribute__ after a
3583 function parameter list: do not tag the
3590 } /* if (endtoken (c)) */
3591 else if (intoken (c))
3597 } /* if (midtoken) */
3598 else if (begtoken (c))
3606 /* This prevents tagging fb in
3607 void (__attribute__((noreturn)) *fb) (void);
3608 Fixing this is not easy and not very important. */
3612 if (plainc || declarations)
3614 make_C_tag (TRUE); /* a function */
3631 if (structdef == stagseen && !cjava)
3633 popclass_above (bracelev);
3647 if (!yacc_rules || lp == newlb.buffer + 1)
3649 tokoff = lp - 1 - newlb.buffer;
3654 } /* if (begtoken) */
3655 } /* if must look at token */
3658 /* Detect end of line, colon, comma, semicolon and various braces
3659 after having handled a token.*/
3665 if (yacc_rules && token.offset == 0 && token.valid)
3667 make_C_tag (FALSE); /* a yacc function */
3670 if (definedef != dnone)
3676 make_C_tag (TRUE); /* an Objective C class */
3680 objdef = omethodcolon;
3681 linebuffer_setlen (&token_name, token_name.len + 1);
3682 strcat (token_name.buffer, ":");
3688 case oimplementation:
3698 if (structdef == stagseen)
3700 structdef = scolonseen;
3703 /* Should be useless, but may be work as a safety net. */
3704 if (cplpl && fvdef == flistseen)
3706 make_C_tag (TRUE); /* a function */
3712 if (definedef != dnone || inattribute)
3718 make_C_tag (FALSE); /* a typedef */
3728 if (typdef == tignore || cplpl)
3732 if ((globals && bracelev == 0 && (!fvextern || declarations))
3733 || (members && instruct))
3734 make_C_tag (FALSE); /* a variable */
3737 token.valid = FALSE;
3741 && (cplpl || !instruct)
3742 && (typdef == tnone || (typdef != tignore && instruct)))
3744 && plainc && instruct))
3745 make_C_tag (TRUE); /* a function */
3758 && cplpl && structdef == stagseen)
3759 make_C_tag (FALSE); /* forward declaration */
3761 token.valid = FALSE;
3762 } /* switch (fvdef) */
3769 if (structdef == stagseen)
3773 if (definedef != dnone || inattribute)
3779 make_C_tag (TRUE); /* an Objective C method */
3786 case oimplementation:
3814 && (!fvextern || declarations))
3815 || (members && instruct)))
3816 make_C_tag (FALSE); /* a variable */
3819 if ((declarations && typdef == tnone && !instruct)
3820 || (members && typdef != tignore && instruct))
3822 make_C_tag (TRUE); /* a function */
3825 else if (!declarations)
3827 token.valid = FALSE;
3833 if (structdef == stagseen)
3837 if (definedef != dnone || inattribute)
3839 if (structdef == stagseen)
3846 make_C_tag (FALSE); /* a typedef */
3858 if ((members && bracelev == 1)
3859 || (globals && bracelev == 0
3860 && (!fvextern || declarations)))
3861 make_C_tag (FALSE); /* a variable */
3886 if (definedef != dnone)
3888 if (objdef == otagseen && parlev == 0)
3889 objdef = oparenseen;
3893 if (typdef == ttypeseen
3897 /* This handles constructs like:
3898 typedef void OperatorFun (int fun); */
3928 if (--attrparlev == 0)
3929 inattribute = FALSE;
3932 if (definedef != dnone)
3934 if (objdef == ocatseen && parlev == 1)
3936 make_C_tag (TRUE); /* an Objective C category */
3962 || typdef == ttypeseen))
3965 make_C_tag (FALSE); /* a typedef */
3968 else if (parlev < 0) /* can happen due to ill-conceived #if's. */
3972 if (definedef != dnone)
3974 if (typdef == ttypeseen)
3976 /* Whenever typdef is set to tinbody (currently only
3977 here), typdefbracelev should be set to bracelev. */
3979 typdefbracelev = bracelev;
3984 make_C_tag (TRUE); /* a function */
3993 make_C_tag (TRUE); /* an Objective C class */
3998 make_C_tag (TRUE); /* an Objective C method */
4003 case oimplementation:
4011 /* Neutralize `extern "C" {' grot. */
4012 if (bracelev == 0 && structdef == snone && nestlev == 0
4031 case skeyseen: /* unnamed struct */
4032 pushclass_above (bracelev, NULL, 0);
4035 case stagseen: /* named struct or enum */
4036 case scolonseen: /* a class */
4037 pushclass_above (bracelev,token.line+token.offset, token.length);
4039 make_C_tag (FALSE); /* a struct or enum */
4050 if (definedef != dnone)
4052 if (fvdef == fstartlist)
4054 fvdef = fvnone; /* avoid tagging `foo' in `foo (*bar()) ()' */
4055 token.valid = FALSE;
4059 if (definedef != dnone)
4062 if (!ignoreindent && lp == newlb.buffer + 1)
4065 token.valid = FALSE; /* unexpected value, token unreliable */
4066 bracelev = 0; /* reset brace level if first column */
4067 parlev = 0; /* also reset paren level, just in case... */
4069 else if (bracelev < 0)
4071 token.valid = FALSE; /* something gone amiss, token unreliable */
4074 if (bracelev == 0 && fvdef == vignore)
4075 fvdef = fvnone; /* end of function */
4076 popclass_above (bracelev);
4078 /* Only if typdef == tinbody is typdefbracelev significant. */
4079 if (typdef == tinbody && bracelev <= typdefbracelev)
4081 assert (bracelev == typdefbracelev);
4086 if (definedef != dnone)
4096 if ((members && bracelev == 1)
4097 || (globals && bracelev == 0 && (!fvextern || declarations)))
4098 make_C_tag (FALSE); /* a variable */
4111 && (structdef == stagseen || fvdef == fvnameseen))
4118 if (templatelev > 0)
4126 if (objdef == oinbody && bracelev == 0)
4128 objdef = omethodsign;
4133 case '#': case '~': case '&': case '%': case '/':
4134 case '|': case '^': case '!': case '.': case '?':
4135 if (definedef != dnone)
4137 /* These surely cannot follow a function tag in C. */
4156 if (objdef == otagseen)
4158 make_C_tag (TRUE); /* an Objective C class */
4161 /* If a macro spans multiple lines don't reset its state. */
4163 CNL_SAVE_DEFINEDEF ();
4171 } /* while not eof */
4173 free (lbs[0].lb.buffer);
4174 free (lbs[1].lb.buffer);
4178 * Process either a C++ file or a C file depending on the setting
4182 default_C_entries (inf)
4185 C_entries (cplusplus ? C_PLPL : C_AUTO, inf);
4188 /* Always do plain C. */
4190 plain_C_entries (inf)
4196 /* Always do C++. */
4198 Cplusplus_entries (inf)
4201 C_entries (C_PLPL, inf);
4204 /* Always do Java. */
4209 C_entries (C_JAVA, inf);
4217 C_entries (C_STAR, inf);
4220 /* Always do Yacc. */
4225 C_entries (YACC, inf);
4229 /* Useful macros. */
4230 #define LOOP_ON_INPUT_LINES(file_pointer, line_buffer, char_pointer) \
4231 for (; /* loop initialization */ \
4232 !feof (file_pointer) /* loop test */ \
4233 && /* instructions at start of loop */ \
4234 (readline (&line_buffer, file_pointer), \
4235 char_pointer = line_buffer.buffer, \
4239 #define LOOKING_AT(cp, kw) /* kw is the keyword, a literal string */ \
4240 ((assert("" kw), TRUE) /* syntax error if not a literal string */ \
4241 && strneq ((cp), kw, sizeof(kw)-1) /* cp points at kw */ \
4242 && notinname ((cp)[sizeof(kw)-1]) /* end of kw */ \
4243 && ((cp) = skip_spaces((cp)+sizeof(kw)-1))) /* skip spaces */
4245 /* Similar to LOOKING_AT but does not use notinname, does not skip */
4246 #define LOOKING_AT_NOCASE(cp, kw) /* the keyword is a literal string */ \
4247 ((assert("" kw), TRUE) /* syntax error if not a literal string */ \
4248 && strncaseeq ((cp), kw, sizeof(kw)-1) /* cp points at kw */ \
4249 && ((cp) += sizeof(kw)-1)) /* skip spaces */
4252 * Read a file, but do no processing. This is used to do regexp
4253 * matching on files that have no language defined.
4256 just_read_file (inf)
4259 register char *dummy;
4261 LOOP_ON_INPUT_LINES (inf, lb, dummy)
4266 /* Fortran parsing */
4268 static void F_takeprec __P((void));
4269 static void F_getit __P((FILE *));
4274 dbp = skip_spaces (dbp);
4278 dbp = skip_spaces (dbp);
4279 if (strneq (dbp, "(*)", 3))
4284 if (!ISDIGIT (*dbp))
4286 --dbp; /* force failure */
4291 while (ISDIGIT (*dbp));
4300 dbp = skip_spaces (dbp);
4303 readline (&lb, inf);
4308 dbp = skip_spaces (dbp);
4310 if (!ISALPHA (*dbp) && *dbp != '_' && *dbp != '$')
4312 for (cp = dbp + 1; *cp != '\0' && intoken (*cp); cp++)
4314 make_tag (dbp, cp-dbp, TRUE,
4315 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4320 Fortran_functions (inf)
4323 LOOP_ON_INPUT_LINES (inf, lb, dbp)
4326 dbp++; /* Ratfor escape to fortran */
4327 dbp = skip_spaces (dbp);
4330 switch (lowcase (*dbp))
4333 if (nocase_tail ("integer"))
4337 if (nocase_tail ("real"))
4341 if (nocase_tail ("logical"))
4345 if (nocase_tail ("complex") || nocase_tail ("character"))
4349 if (nocase_tail ("double"))
4351 dbp = skip_spaces (dbp);
4354 if (nocase_tail ("precision"))
4362 dbp = skip_spaces (dbp);
4365 switch (lowcase (*dbp))
4368 if (nocase_tail ("function"))
4372 if (nocase_tail ("subroutine"))
4376 if (nocase_tail ("entry"))
4380 if (nocase_tail ("blockdata") || nocase_tail ("block data"))
4382 dbp = skip_spaces (dbp);
4383 if (*dbp == '\0') /* assume un-named */
4384 make_tag ("blockdata", 9, TRUE,
4385 lb.buffer, dbp - lb.buffer, lineno, linecharno);
4387 F_getit (inf); /* look for name */
4400 * Philippe Waroquiers (1998)
4403 static void Ada_getit __P((FILE *, char *));
4405 /* Once we are positioned after an "interesting" keyword, let's get
4406 the real tag value necessary. */
4408 Ada_getit (inf, name_qualifier)
4410 char *name_qualifier;
4418 dbp = skip_spaces (dbp);
4420 || (dbp[0] == '-' && dbp[1] == '-'))
4422 readline (&lb, inf);
4425 switch (lowcase(*dbp))
4428 if (nocase_tail ("body"))
4430 /* Skipping body of procedure body or package body or ....
4431 resetting qualifier to body instead of spec. */
4432 name_qualifier = "/b";
4437 /* Skipping type of task type or protected type ... */
4438 if (nocase_tail ("type"))
4447 for (cp = dbp; *cp != '\0' && *cp != '"'; cp++)
4452 dbp = skip_spaces (dbp);
4455 && (ISALPHA (*cp) || ISDIGIT (*cp) || *cp == '_' || *cp == '.'));
4463 name = concat (dbp, name_qualifier, "");
4465 make_tag (name, strlen (name), TRUE,
4466 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4478 bool inquote = FALSE;
4479 bool skip_till_semicolumn = FALSE;
4481 LOOP_ON_INPUT_LINES (inf, lb, dbp)
4483 while (*dbp != '\0')
4485 /* Skip a string i.e. "abcd". */
4486 if (inquote || (*dbp == '"'))
4488 dbp = etags_strchr ((inquote) ? dbp : dbp+1, '"');
4493 continue; /* advance char */
4498 break; /* advance line */
4502 /* Skip comments. */
4503 if (dbp[0] == '-' && dbp[1] == '-')
4504 break; /* advance line */
4506 /* Skip character enclosed in single quote i.e. 'a'
4507 and skip single quote starting an attribute i.e. 'Image. */
4516 if (skip_till_semicolumn)
4519 skip_till_semicolumn = FALSE;
4521 continue; /* advance char */
4524 /* Search for beginning of a token. */
4525 if (!begtoken (*dbp))
4528 continue; /* advance char */
4531 /* We are at the beginning of a token. */
4532 switch (lowcase(*dbp))
4535 if (!packages_only && nocase_tail ("function"))
4536 Ada_getit (inf, "/f");
4538 break; /* from switch */
4539 continue; /* advance char */
4541 if (!packages_only && nocase_tail ("procedure"))
4542 Ada_getit (inf, "/p");
4543 else if (nocase_tail ("package"))
4544 Ada_getit (inf, "/s");
4545 else if (nocase_tail ("protected")) /* protected type */
4546 Ada_getit (inf, "/t");
4548 break; /* from switch */
4549 continue; /* advance char */
4552 if (typedefs && !packages_only && nocase_tail ("use"))
4554 /* when tagging types, avoid tagging use type Pack.Typename;
4555 for this, we will skip everything till a ; */
4556 skip_till_semicolumn = TRUE;
4557 continue; /* advance char */
4561 if (!packages_only && nocase_tail ("task"))
4562 Ada_getit (inf, "/k");
4563 else if (typedefs && !packages_only && nocase_tail ("type"))
4565 Ada_getit (inf, "/t");
4566 while (*dbp != '\0')
4570 break; /* from switch */
4571 continue; /* advance char */
4576 /* Look for the end of the token. */
4577 while (!endtoken (*dbp))
4580 } /* advance char */
4581 } /* advance line */
4586 * Unix and microcontroller assembly tag handling
4587 * Labels: /^[a-zA-Z_.$][a-zA_Z0-9_.$]*[: ^I^J]/
4588 * Idea by Bob Weiner, Motorola Inc. (1994)
4596 LOOP_ON_INPUT_LINES (inf, lb, cp)
4598 /* If first char is alphabetic or one of [_.$], test for colon
4599 following identifier. */
4600 if (ISALPHA (*cp) || *cp == '_' || *cp == '.' || *cp == '$')
4602 /* Read past label. */
4604 while (ISALNUM (*cp) || *cp == '_' || *cp == '.' || *cp == '$')
4606 if (*cp == ':' || iswhite (*cp))
4607 /* Found end of label, so copy it and add it to the table. */
4608 make_tag (lb.buffer, cp - lb.buffer, TRUE,
4609 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4617 * Perl sub names: /^sub[ \t\n]+[^ \t\n{]+/
4618 * Perl variable names: /^(my|local).../
4619 * Original code by Bart Robinson <lomew@cs.utah.edu> (1995)
4620 * Additions by Michael Ernst <mernst@alum.mit.edu> (1997)
4621 * Ideas by Kai Großjohann <Kai.Grossjohann@CS.Uni-Dortmund.DE> (2001)
4624 Perl_functions (inf)
4627 char *package = savestr ("main"); /* current package name */
4630 LOOP_ON_INPUT_LINES (inf, lb, cp)
4632 cp = skip_spaces (cp);
4634 if (LOOKING_AT (cp, "package"))
4637 get_tag (cp, &package);
4639 else if (LOOKING_AT (cp, "sub"))
4644 while (!notinname (*cp))
4647 continue; /* nothing found */
4648 if ((pos = etags_strchr (sp, ':')) != NULL
4649 && pos < cp && pos[1] == ':')
4650 /* The name is already qualified. */
4651 make_tag (sp, cp - sp, TRUE,
4652 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4656 char savechar, *name;
4660 name = concat (package, "::", sp);
4662 make_tag (name, strlen(name), TRUE,
4663 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4667 else if (globals) /* only if we are tagging global vars */
4669 /* Skip a qualifier, if any. */
4670 bool qual = LOOKING_AT (cp, "my") || LOOKING_AT (cp, "local");
4671 /* After "my" or "local", but before any following paren or space. */
4672 char *varstart = cp;
4674 if (qual /* should this be removed? If yes, how? */
4675 && (*cp == '$' || *cp == '@' || *cp == '%'))
4680 while (ISALNUM (*cp) || *cp == '_');
4684 /* Should be examining a variable list at this point;
4685 could insist on seeing an open parenthesis. */
4686 while (*cp != '\0' && *cp != ';' && *cp != '=' && *cp != ')')
4692 make_tag (varstart, cp - varstart, FALSE,
4693 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4702 * Look for /^[\t]*def[ \t\n]+[^ \t\n(:]+/ or /^class[ \t\n]+[^ \t\n(:]+/
4703 * Idea by Eric S. Raymond <esr@thyrsus.com> (1997)
4704 * More ideas by seb bacon <seb@jamkit.com> (2002)
4707 Python_functions (inf)
4712 LOOP_ON_INPUT_LINES (inf, lb, cp)
4714 cp = skip_spaces (cp);
4715 if (LOOKING_AT (cp, "def") || LOOKING_AT (cp, "class"))
4718 while (!notinname (*cp) && *cp != ':')
4720 make_tag (name, cp - name, TRUE,
4721 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4730 * - /^[ \t]*function[ \t\n]+[^ \t\n(]+/
4731 * - /^[ \t]*class[ \t\n]+[^ \t\n]+/
4732 * - /^[ \t]*define\(\"[^\"]+/
4733 * Only with --members:
4734 * - /^[ \t]*var[ \t\n]+\$[^ \t\n=;]/
4735 * Idea by Diez B. Roggisch (2001)
4741 register char *cp, *name;
4742 bool search_identifier = FALSE;
4744 LOOP_ON_INPUT_LINES (inf, lb, cp)
4746 cp = skip_spaces (cp);
4748 if (search_identifier
4751 while (!notinname (*cp))
4753 make_tag (name, cp - name, TRUE,
4754 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4755 search_identifier = FALSE;
4757 else if (LOOKING_AT (cp, "function"))
4760 cp = skip_spaces (cp+1);
4764 while (!notinname (*cp))
4766 make_tag (name, cp - name, TRUE,
4767 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4770 search_identifier = TRUE;
4772 else if (LOOKING_AT (cp, "class"))
4777 while (*cp != '\0' && !iswhite (*cp))
4779 make_tag (name, cp - name, FALSE,
4780 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4783 search_identifier = TRUE;
4785 else if (strneq (cp, "define", 6)
4786 && (cp = skip_spaces (cp+6))
4788 && (*cp == '"' || *cp == '\''))
4792 while (*cp != quote && *cp != '\0')
4794 make_tag (name, cp - name, FALSE,
4795 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4798 && LOOKING_AT (cp, "var")
4802 while (!notinname(*cp))
4804 make_tag (name, cp - name, FALSE,
4805 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
4812 * Cobol tag functions
4813 * We could look for anything that could be a paragraph name.
4814 * i.e. anything that starts in column 8 is one word and ends in a full stop.
4815 * Idea by Corny de Souza (1993)
4818 Cobol_paragraphs (inf)
4821 register char *bp, *ep;
4823 LOOP_ON_INPUT_LINES (inf, lb, bp)
4829 /* If eoln, compiler option or comment ignore whole line. */
4830 if (bp[-1] != ' ' || !ISALNUM (bp[0]))
4833 for (ep = bp; ISALNUM (*ep) || *ep == '-'; ep++)
4836 make_tag (bp, ep - bp, TRUE,
4837 lb.buffer, ep - lb.buffer + 1, lineno, linecharno);
4844 * Ideas by Assar Westerlund <assar@sics.se> (2001)
4847 Makefile_targets (inf)
4852 LOOP_ON_INPUT_LINES (inf, lb, bp)
4854 if (*bp == '\t' || *bp == '#')
4856 while (*bp != '\0' && *bp != '=' && *bp != ':')
4858 if (*bp == ':' || (globals && *bp == '='))
4860 /* We should detect if there is more than one tag, but we do not.
4861 We just skip initial and final spaces. */
4862 char * namestart = skip_spaces (lb.buffer);
4863 while (--bp > namestart)
4864 if (!notinname (*bp))
4866 make_tag (namestart, bp - namestart + 1, TRUE,
4867 lb.buffer, bp - lb.buffer + 2, lineno, linecharno);
4875 * Original code by Mosur K. Mohan (1989)
4877 * Locates tags for procedures & functions. Doesn't do any type- or
4878 * var-definitions. It does look for the keyword "extern" or
4879 * "forward" immediately following the procedure statement; if found,
4880 * the tag is skipped.
4883 Pascal_functions (inf)
4886 linebuffer tline; /* mostly copied from C_entries */
4888 int save_lineno, namelen, taglen;
4891 bool /* each of these flags is TRUE if: */
4892 incomment, /* point is inside a comment */
4893 inquote, /* point is inside '..' string */
4894 get_tagname, /* point is after PROCEDURE/FUNCTION
4895 keyword, so next item = potential tag */
4896 found_tag, /* point is after a potential tag */
4897 inparms, /* point is within parameter-list */
4898 verify_tag; /* point has passed the parm-list, so the
4899 next token will determine whether this
4900 is a FORWARD/EXTERN to be ignored, or
4901 whether it is a real tag */
4903 save_lcno = save_lineno = namelen = taglen = 0; /* keep compiler quiet */
4904 name = NULL; /* keep compiler quiet */
4907 linebuffer_init (&tline);
4909 incomment = inquote = FALSE;
4910 found_tag = FALSE; /* have a proc name; check if extern */
4911 get_tagname = FALSE; /* found "procedure" keyword */
4912 inparms = FALSE; /* found '(' after "proc" */
4913 verify_tag = FALSE; /* check if "extern" is ahead */
4916 while (!feof (inf)) /* long main loop to get next char */
4919 if (c == '\0') /* if end of line */
4921 readline (&lb, inf);
4925 if (!((found_tag && verify_tag)
4927 c = *dbp++; /* only if don't need *dbp pointing
4928 to the beginning of the name of
4929 the procedure or function */
4933 if (c == '}') /* within { } comments */
4935 else if (c == '*' && *dbp == ')') /* within (* *) comments */
4952 inquote = TRUE; /* found first quote */
4954 case '{': /* found open { comment */
4958 if (*dbp == '*') /* found open (* comment */
4963 else if (found_tag) /* found '(' after tag, i.e., parm-list */
4966 case ')': /* end of parms list */
4971 if (found_tag && !inparms) /* end of proc or fn stmt */
4980 if (found_tag && verify_tag && (*dbp != ' '))
4982 /* Check if this is an "extern" declaration. */
4985 if (lowcase (*dbp == 'e'))
4987 if (nocase_tail ("extern")) /* superfluous, really! */
4993 else if (lowcase (*dbp) == 'f')
4995 if (nocase_tail ("forward")) /* check for forward reference */
5001 if (found_tag && verify_tag) /* not external proc, so make tag */
5005 make_tag (name, namelen, TRUE,
5006 tline.buffer, taglen, save_lineno, save_lcno);
5010 if (get_tagname) /* grab name of proc or fn */
5017 /* Find block name. */
5018 for (cp = dbp + 1; *cp != '\0' && !endtoken (*cp); cp++)
5021 /* Save all values for later tagging. */
5022 linebuffer_setlen (&tline, lb.len);
5023 strncpy(tline.buffer, lb.buffer, lb.len-1);
5024 save_lineno = lineno;
5025 save_lcno = linecharno;
5026 name = tline.buffer + (dbp - lb.buffer);
5028 taglen = cp - lb.buffer + 1;
5030 dbp = cp; /* set dbp to e-o-token */
5031 get_tagname = FALSE;
5035 /* And proceed to check for "extern". */
5037 else if (!incomment && !inquote && !found_tag)
5039 /* Check for proc/fn keywords. */
5040 switch (lowcase (c))
5043 if (nocase_tail ("rocedure")) /* c = 'p', dbp has advanced */
5047 if (nocase_tail ("unction"))
5054 } /* while not eof */
5056 free (tline.buffer);
5061 * Lisp tag functions
5062 * look for (def or (DEF, quote or QUOTE
5065 static void L_getit __P((void));
5070 if (*dbp == '\'') /* Skip prefix quote */
5072 else if (*dbp == '(')
5075 /* Try to skip "(quote " */
5076 if (!LOOKING_AT (dbp, "quote") && !LOOKING_AT (dbp, "QUOTE"))
5077 /* Ok, then skip "(" before name in (defstruct (foo)) */
5078 dbp = skip_spaces (dbp);
5080 get_tag (dbp, NULL);
5084 Lisp_functions (inf)
5087 LOOP_ON_INPUT_LINES (inf, lb, dbp)
5092 if (strneq (dbp+1, "def", 3) || strneq (dbp+1, "DEF", 3))
5094 dbp = skip_non_spaces (dbp);
5095 dbp = skip_spaces (dbp);
5100 /* Check for (foo::defmumble name-defined ... */
5103 while (!notinname (*dbp) && *dbp != ':');
5108 while (*dbp == ':');
5110 if (strneq (dbp, "def", 3) || strneq (dbp, "DEF", 3))
5112 dbp = skip_non_spaces (dbp);
5113 dbp = skip_spaces (dbp);
5123 * Lua script language parsing
5124 * Original code by David A. Capello <dacap@users.sourceforge.net> (2004)
5126 * "function" and "local function" are tags if they start at column 1.
5134 LOOP_ON_INPUT_LINES (inf, lb, bp)
5136 if (bp[0] != 'f' && bp[0] != 'l')
5139 (void)LOOKING_AT (bp, "local"); /* skip possible "local" */
5141 if (LOOKING_AT (bp, "function"))
5149 * Just look for lines where the first character is '/'
5150 * Also look at "defineps" for PSWrap
5152 * Richard Mlynarik <mly@adoc.xerox.com> (1997)
5153 * Masatake Yamato <masata-y@is.aist-nara.ac.jp> (1999)
5159 register char *bp, *ep;
5161 LOOP_ON_INPUT_LINES (inf, lb, bp)
5166 *ep != '\0' && *ep != ' ' && *ep != '{';
5169 make_tag (bp, ep - bp, TRUE,
5170 lb.buffer, ep - lb.buffer + 1, lineno, linecharno);
5172 else if (LOOKING_AT (bp, "defineps"))
5180 * Ignore anything after \ followed by space or in ( )
5181 * Look for words defined by :
5182 * Look for constant, code, create, defer, value, and variable
5183 * OBP extensions: Look for buffer:, field,
5184 * Ideas by Eduardo Horvath <eeh@netbsd.org> (2004)
5192 LOOP_ON_INPUT_LINES (inf, lb, bp)
5193 while ((bp = skip_spaces (bp))[0] != '\0')
5194 if (bp[0] == '\\' && iswhite(bp[1]))
5195 break; /* read next line */
5196 else if (bp[0] == '(' && iswhite(bp[1]))
5197 do /* skip to ) or eol */
5199 while (*bp != ')' && *bp != '\0');
5200 else if ((bp[0] == ':' && iswhite(bp[1]) && bp++)
5201 || LOOKING_AT_NOCASE (bp, "constant")
5202 || LOOKING_AT_NOCASE (bp, "code")
5203 || LOOKING_AT_NOCASE (bp, "create")
5204 || LOOKING_AT_NOCASE (bp, "defer")
5205 || LOOKING_AT_NOCASE (bp, "value")
5206 || LOOKING_AT_NOCASE (bp, "variable")
5207 || LOOKING_AT_NOCASE (bp, "buffer:")
5208 || LOOKING_AT_NOCASE (bp, "field"))
5209 get_tag (skip_spaces (bp), NULL); /* Yay! A definition! */
5211 bp = skip_non_spaces (bp);
5216 * Scheme tag functions
5217 * look for (def... xyzzy
5219 * (def ... ((...(xyzzy ....
5221 * Original code by Ken Haase (1985?)
5224 Scheme_functions (inf)
5229 LOOP_ON_INPUT_LINES (inf, lb, bp)
5231 if (strneq (bp, "(def", 4) || strneq (bp, "(DEF", 4))
5233 bp = skip_non_spaces (bp+4);
5234 /* Skip over open parens and white space */
5235 while (notinname (*bp))
5239 if (LOOKING_AT (bp, "(SET!") || LOOKING_AT (bp, "(set!"))
5245 /* Find tags in TeX and LaTeX input files. */
5247 /* TEX_toktab is a table of TeX control sequences that define tags.
5248 * Each entry records one such control sequence.
5250 * Original code from who knows whom.
5252 * Stefan Monnier (2002)
5255 static linebuffer *TEX_toktab = NULL; /* Table with tag tokens */
5257 /* Default set of control sequences to put into TEX_toktab.
5258 The value of environment var TEXTAGS is prepended to this. */
5259 static char *TEX_defenv = "\
5260 :chapter:section:subsection:subsubsection:eqno:label:ref:cite:bibitem\
5261 :part:appendix:entry:index:def\
5262 :newcommand:renewcommand:newenvironment:renewenvironment";
5264 static void TEX_mode __P((FILE *));
5265 static void TEX_decode_env __P((char *, char *));
5267 static char TEX_esc = '\\';
5268 static char TEX_opgrp = '{';
5269 static char TEX_clgrp = '}';
5272 * TeX/LaTeX scanning loop.
5281 /* Select either \ or ! as escape character. */
5284 /* Initialize token table once from environment. */
5285 if (TEX_toktab == NULL)
5286 TEX_decode_env ("TEXTAGS", TEX_defenv);
5288 LOOP_ON_INPUT_LINES (inf, lb, cp)
5290 /* Look at each TEX keyword in line. */
5293 /* Look for a TEX escape. */
5294 while (*cp++ != TEX_esc)
5295 if (cp[-1] == '\0' || cp[-1] == '%')
5298 for (key = TEX_toktab; key->buffer != NULL; key++)
5299 if (strneq (cp, key->buffer, key->len))
5302 int namelen, linelen;
5305 cp = skip_spaces (cp + key->len);
5306 if (*cp == TEX_opgrp)
5312 (!iswhite (*p) && *p != '#' &&
5313 *p != TEX_opgrp && *p != TEX_clgrp);
5318 if (!opgrp || *p == TEX_clgrp)
5320 while (*p != '\0' && *p != TEX_opgrp && *p != TEX_clgrp)
5322 linelen = p - lb.buffer + 1;
5324 make_tag (cp, namelen, TRUE,
5325 lb.buffer, linelen, lineno, linecharno);
5326 goto tex_next_line; /* We only tag a line once */
5334 #define TEX_LESC '\\'
5335 #define TEX_SESC '!'
5337 /* Figure out whether TeX's escapechar is '\\' or '!' and set grouping
5338 chars accordingly. */
5345 while ((c = getc (inf)) != EOF)
5347 /* Skip to next line if we hit the TeX comment char. */
5349 while (c != '\n' && c != EOF)
5351 else if (c == TEX_LESC || c == TEX_SESC )
5367 /* If the input file is compressed, inf is a pipe, and rewind may fail.
5368 No attempt is made to correct the situation. */
5372 /* Read environment and prepend it to the default string.
5373 Build token table. */
5375 TEX_decode_env (evarname, defenv)
5379 register char *env, *p;
5382 /* Append default string to environment. */
5383 env = getenv (evarname);
5389 env = concat (oldenv, defenv, "");
5392 /* Allocate a token table */
5393 for (len = 1, p = env; p;)
5394 if ((p = etags_strchr (p, ':')) && *++p != '\0')
5396 TEX_toktab = xnew (len, linebuffer);
5398 /* Unpack environment string into token table. Be careful about */
5399 /* zero-length strings (leading ':', "::" and trailing ':') */
5400 for (i = 0; *env != '\0';)
5402 p = etags_strchr (env, ':');
5403 if (!p) /* End of environment string. */
5404 p = env + strlen (env);
5406 { /* Only non-zero strings. */
5407 TEX_toktab[i].buffer = savenstr (env, p - env);
5408 TEX_toktab[i].len = p - env;
5415 TEX_toktab[i].buffer = NULL; /* Mark end of table. */
5416 TEX_toktab[i].len = 0;
5423 /* Texinfo support. Dave Love, Mar. 2000. */
5429 LOOP_ON_INPUT_LINES (inf, lb, cp)
5430 if (LOOKING_AT (cp, "@node"))
5433 while (*cp != '\0' && *cp != ',')
5435 make_tag (start, cp - start, TRUE,
5436 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
5443 * Contents of <title>, <h1>, <h2>, <h3> are tags.
5444 * Contents of <a name=xxx> are tags with name xxx.
5446 * Francesco Potortì, 2002.
5452 bool getnext = FALSE; /* next text outside of HTML tags is a tag */
5453 bool skiptag = FALSE; /* skip to the end of the current HTML tag */
5454 bool intag = FALSE; /* inside an html tag, looking for ID= */
5455 bool inanchor = FALSE; /* when INTAG, is an anchor, look for NAME= */
5459 linebuffer_setlen (&token_name, 0); /* no name in buffer */
5461 LOOP_ON_INPUT_LINES (inf, lb, dbp)
5462 for (;;) /* loop on the same line */
5464 if (skiptag) /* skip HTML tag */
5466 while (*dbp != '\0' && *dbp != '>')
5472 continue; /* look on the same line */
5474 break; /* go to next line */
5477 else if (intag) /* look for "name=" or "id=" */
5479 while (*dbp != '\0' && *dbp != '>'
5480 && lowcase (*dbp) != 'n' && lowcase (*dbp) != 'i')
5483 break; /* go to next line */
5488 continue; /* look on the same line */
5490 if ((inanchor && LOOKING_AT_NOCASE (dbp, "name="))
5491 || LOOKING_AT_NOCASE (dbp, "id="))
5493 bool quoted = (dbp[0] == '"');
5496 for (end = ++dbp; *end != '\0' && *end != '"'; end++)
5499 for (end = dbp; *end != '\0' && intoken (*end); end++)
5501 linebuffer_setlen (&token_name, end - dbp);
5502 strncpy (token_name.buffer, dbp, end - dbp);
5503 token_name.buffer[end - dbp] = '\0';
5506 intag = FALSE; /* we found what we looked for */
5507 skiptag = TRUE; /* skip to the end of the tag */
5508 getnext = TRUE; /* then grab the text */
5509 continue; /* look on the same line */
5514 else if (getnext) /* grab next tokens and tag them */
5516 dbp = skip_spaces (dbp);
5518 break; /* go to next line */
5522 inanchor = (lowcase (dbp[1]) == 'a' && !intoken (dbp[2]));
5523 continue; /* look on the same line */
5526 for (end = dbp + 1; *end != '\0' && *end != '<'; end++)
5528 make_tag (token_name.buffer, token_name.len, TRUE,
5529 dbp, end - dbp, lineno, linecharno);
5530 linebuffer_setlen (&token_name, 0); /* no name in buffer */
5532 break; /* go to next line */
5535 else /* look for an interesting HTML tag */
5537 while (*dbp != '\0' && *dbp != '<')
5540 break; /* go to next line */
5542 if (lowcase (dbp[1]) == 'a' && !intoken (dbp[2]))
5545 continue; /* look on the same line */
5547 else if (LOOKING_AT_NOCASE (dbp, "<title>")
5548 || LOOKING_AT_NOCASE (dbp, "<h1>")
5549 || LOOKING_AT_NOCASE (dbp, "<h2>")
5550 || LOOKING_AT_NOCASE (dbp, "<h3>"))
5554 continue; /* look on the same line */
5565 * Assumes that the predicate or rule starts at column 0.
5566 * Only the first clause of a predicate or rule is added.
5567 * Original code by Sunichirou Sugou (1989)
5568 * Rewritten by Anders Lindgren (1996)
5570 static int prolog_pr __P((char *, char *));
5571 static void prolog_skip_comment __P((linebuffer *, FILE *));
5572 static int prolog_atom __P((char *, int));
5575 Prolog_functions (inf)
5586 LOOP_ON_INPUT_LINES (inf, lb, cp)
5588 if (cp[0] == '\0') /* Empty line */
5590 else if (iswhite (cp[0])) /* Not a predicate */
5592 else if (cp[0] == '/' && cp[1] == '*') /* comment. */
5593 prolog_skip_comment (&lb, inf);
5594 else if ((len = prolog_pr (cp, last)) > 0)
5596 /* Predicate or rule. Store the function name so that we
5597 only generate a tag for the first clause. */
5599 last = xnew(len + 1, char);
5600 else if (len + 1 > allocated)
5601 xrnew (last, len + 1, char);
5602 allocated = len + 1;
5603 strncpy (last, cp, len);
5612 prolog_skip_comment (plb, inf)
5620 for (cp = plb->buffer; *cp != '\0'; cp++)
5621 if (cp[0] == '*' && cp[1] == '/')
5623 readline (plb, inf);
5629 * A predicate or rule definition is added if it matches:
5630 * <beginning of line><Prolog Atom><whitespace>(
5631 * or <beginning of line><Prolog Atom><whitespace>:-
5633 * It is added to the tags database if it doesn't match the
5634 * name of the previous clause header.
5636 * Return the size of the name of the predicate or rule, or 0 if no
5642 char *last; /* Name of last clause. */
5647 pos = prolog_atom (s, 0);
5652 pos = skip_spaces (s + pos) - s;
5655 || (s[pos] == '(' && (pos += 1))
5656 || (s[pos] == ':' && s[pos + 1] == '-' && (pos += 2)))
5657 && (last == NULL /* save only the first clause */
5658 || len != (int)strlen (last)
5659 || !strneq (s, last, len)))
5661 make_tag (s, len, TRUE, s, pos, lineno, linecharno);
5669 * Consume a Prolog atom.
5670 * Return the number of bytes consumed, or -1 if there was an error.
5672 * A prolog atom, in this context, could be one of:
5673 * - An alphanumeric sequence, starting with a lower case letter.
5674 * - A quoted arbitrary string. Single quotes can escape themselves.
5675 * Backslash quotes everything.
5678 prolog_atom (s, pos)
5686 if (ISLOWER(s[pos]) || (s[pos] == '_'))
5688 /* The atom is unquoted. */
5690 while (ISALNUM(s[pos]) || (s[pos] == '_'))
5694 return pos - origpos;
5696 else if (s[pos] == '\'')
5707 pos++; /* A double quote */
5709 else if (s[pos] == '\0')
5710 /* Multiline quoted atoms are ignored. */
5712 else if (s[pos] == '\\')
5714 if (s[pos+1] == '\0')
5721 return pos - origpos;
5729 * Support for Erlang
5731 * Generates tags for functions, defines, and records.
5732 * Assumes that Erlang functions start at column 0.
5733 * Original code by Anders Lindgren (1996)
5735 static int erlang_func __P((char *, char *));
5736 static void erlang_attribute __P((char *));
5737 static int erlang_atom __P((char *));
5740 Erlang_functions (inf)
5751 LOOP_ON_INPUT_LINES (inf, lb, cp)
5753 if (cp[0] == '\0') /* Empty line */
5755 else if (iswhite (cp[0])) /* Not function nor attribute */
5757 else if (cp[0] == '%') /* comment */
5759 else if (cp[0] == '"') /* Sometimes, strings start in column one */
5761 else if (cp[0] == '-') /* attribute, e.g. "-define" */
5763 erlang_attribute (cp);
5770 else if ((len = erlang_func (cp, last)) > 0)
5773 * Function. Store the function name so that we only
5774 * generates a tag for the first clause.
5777 last = xnew (len + 1, char);
5778 else if (len + 1 > allocated)
5779 xrnew (last, len + 1, char);
5780 allocated = len + 1;
5781 strncpy (last, cp, len);
5790 * A function definition is added if it matches:
5791 * <beginning of line><Erlang Atom><whitespace>(
5793 * It is added to the tags database if it doesn't match the
5794 * name of the previous clause header.
5796 * Return the size of the name of the function, or 0 if no function
5800 erlang_func (s, last)
5802 char *last; /* Name of last clause. */
5807 pos = erlang_atom (s);
5812 pos = skip_spaces (s + pos) - s;
5814 /* Save only the first clause. */
5817 || len != (int)strlen (last)
5818 || !strneq (s, last, len)))
5820 make_tag (s, len, TRUE, s, pos, lineno, linecharno);
5829 * Handle attributes. Currently, tags are generated for defines
5832 * They are on the form:
5833 * -define(foo, bar).
5834 * -define(Foo(M, N), M+N).
5835 * -record(graph, {vtab = notable, cyclic = true}).
5838 erlang_attribute (s)
5843 if ((LOOKING_AT (cp, "-define") || LOOKING_AT (cp, "-record"))
5846 int len = erlang_atom (skip_spaces (cp));
5848 make_tag (cp, len, TRUE, s, cp + len - s, lineno, linecharno);
5855 * Consume an Erlang atom (or variable).
5856 * Return the number of bytes consumed, or -1 if there was an error.
5864 if (ISALPHA (s[pos]) || s[pos] == '_')
5866 /* The atom is unquoted. */
5869 while (ISALNUM (s[pos]) || s[pos] == '_');
5871 else if (s[pos] == '\'')
5873 for (pos++; s[pos] != '\''; pos++)
5874 if (s[pos] == '\0' /* multiline quoted atoms are ignored */
5875 || (s[pos] == '\\' && s[++pos] == '\0'))
5884 static char *scan_separators __P((char *));
5885 static void add_regex __P((char *, language *));
5886 static char *substitute __P((char *, char *, struct re_registers *));
5889 * Take a string like "/blah/" and turn it into "blah", verifying
5890 * that the first and last characters are the same, and handling
5891 * quoted separator characters. Actually, stops on the occurrence of
5892 * an unquoted separator. Also process \t, \n, etc. and turn into
5893 * appropriate characters. Works in place. Null terminates name string.
5894 * Returns pointer to terminating separator, or NULL for
5895 * unterminated regexps.
5898 scan_separators (name)
5902 char *copyto = name;
5903 bool quoted = FALSE;
5905 for (++name; *name != '\0'; ++name)
5911 case 'a': *copyto++ = '\007'; break; /* BEL (bell) */
5912 case 'b': *copyto++ = '\b'; break; /* BS (back space) */
5913 case 'd': *copyto++ = 0177; break; /* DEL (delete) */
5914 case 'e': *copyto++ = 033; break; /* ESC (delete) */
5915 case 'f': *copyto++ = '\f'; break; /* FF (form feed) */
5916 case 'n': *copyto++ = '\n'; break; /* NL (new line) */
5917 case 'r': *copyto++ = '\r'; break; /* CR (carriage return) */
5918 case 't': *copyto++ = '\t'; break; /* TAB (horizontal tab) */
5919 case 'v': *copyto++ = '\v'; break; /* VT (vertical tab) */
5925 /* Something else is quoted, so preserve the quote. */
5933 else if (*name == '\\')
5935 else if (*name == sep)
5941 name = NULL; /* signal unterminated regexp */
5943 /* Terminate copied string. */
5948 /* Look at the argument of --regex or --no-regex and do the right
5949 thing. Same for each line of a regexp file. */
5951 analyse_regex (regex_arg)
5954 if (regex_arg == NULL)
5956 free_regexps (); /* --no-regex: remove existing regexps */
5960 /* A real --regexp option or a line in a regexp file. */
5961 switch (regex_arg[0])
5963 /* Comments in regexp file or null arg to --regex. */
5969 /* Read a regex file. This is recursive and may result in a
5970 loop, which will stop when the file descriptors are exhausted. */
5974 linebuffer regexbuf;
5975 char *regexfile = regex_arg + 1;
5977 /* regexfile is a file containing regexps, one per line. */
5978 regexfp = fopen (regexfile, "r");
5979 if (regexfp == NULL)
5984 linebuffer_init (®exbuf);
5985 while (readline_internal (®exbuf, regexfp) > 0)
5986 analyse_regex (regexbuf.buffer);
5987 free (regexbuf.buffer);
5992 /* Regexp to be used for a specific language only. */
5996 char *lang_name = regex_arg + 1;
5999 for (cp = lang_name; *cp != '}'; cp++)
6002 error ("unterminated language name in regex: %s", regex_arg);
6006 lang = get_language_from_langname (lang_name);
6009 add_regex (cp, lang);
6013 /* Regexp to be used for any language. */
6015 add_regex (regex_arg, NULL);
6020 /* Separate the regexp pattern, compile it,
6021 and care for optional name and modifiers. */
6023 add_regex (regexp_pattern, lang)
6024 char *regexp_pattern;
6027 static struct re_pattern_buffer zeropattern;
6028 char sep, *pat, *name, *modifiers;
6030 struct re_pattern_buffer *patbuf;
6033 force_explicit_name = TRUE, /* do not use implicit tag names */
6034 ignore_case = FALSE, /* case is significant */
6035 multi_line = FALSE, /* matches are done one line at a time */
6036 single_line = FALSE; /* dot does not match newline */
6039 if (strlen(regexp_pattern) < 3)
6041 error ("null regexp", (char *)NULL);
6044 sep = regexp_pattern[0];
6045 name = scan_separators (regexp_pattern);
6048 error ("%s: unterminated regexp", regexp_pattern);
6053 error ("null name for regexp \"%s\"", regexp_pattern);
6056 modifiers = scan_separators (name);
6057 if (modifiers == NULL) /* no terminating separator --> no name */
6063 modifiers += 1; /* skip separator */
6065 /* Parse regex modifiers. */
6066 for (; modifiers[0] != '\0'; modifiers++)
6067 switch (modifiers[0])
6070 if (modifiers == name)
6071 error ("forcing explicit tag name but no name, ignoring", NULL);
6072 force_explicit_name = TRUE;
6082 need_filebuf = TRUE;
6087 wrongmod[0] = modifiers[0];
6089 error ("invalid regexp modifier `%s', ignoring", wrongmod);
6094 patbuf = xnew (1, struct re_pattern_buffer);
6095 *patbuf = zeropattern;
6098 static char lc_trans[CHARS];
6100 for (i = 0; i < CHARS; i++)
6101 lc_trans[i] = lowcase (i);
6102 patbuf->translate = lc_trans; /* translation table to fold case */
6106 pat = concat ("^", regexp_pattern, ""); /* anchor to beginning of line */
6108 pat = regexp_pattern;
6111 re_set_syntax (RE_SYNTAX_EMACS | RE_DOT_NEWLINE);
6113 re_set_syntax (RE_SYNTAX_EMACS);
6115 err = re_compile_pattern (pat, strlen (regexp_pattern), patbuf);
6120 error ("%s while compiling pattern", err);
6125 p_head = xnew (1, regexp);
6126 p_head->pattern = savestr (regexp_pattern);
6127 p_head->p_next = rp;
6128 p_head->lang = lang;
6129 p_head->pat = patbuf;
6130 p_head->name = savestr (name);
6131 p_head->error_signaled = FALSE;
6132 p_head->force_explicit_name = force_explicit_name;
6133 p_head->ignore_case = ignore_case;
6134 p_head->multi_line = multi_line;
6138 * Do the substitutions indicated by the regular expression and
6142 substitute (in, out, regs)
6144 struct re_registers *regs;
6147 int size, dig, diglen;
6150 size = strlen (out);
6152 /* Pass 1: figure out how much to allocate by finding all \N strings. */
6153 if (out[size - 1] == '\\')
6154 fatal ("pattern error in \"%s\"", out);
6155 for (t = etags_strchr (out, '\\');
6157 t = etags_strchr (t + 2, '\\'))
6161 diglen = regs->end[dig] - regs->start[dig];
6167 /* Allocate space and do the substitutions. */
6169 result = xnew (size + 1, char);
6171 for (t = result; *out != '\0'; out++)
6172 if (*out == '\\' && ISDIGIT (*++out))
6175 diglen = regs->end[dig] - regs->start[dig];
6176 strncpy (t, in + regs->start[dig], diglen);
6183 assert (t <= result + size);
6184 assert (t - result == (int)strlen (result));
6189 /* Deallocate all regexps. */
6194 while (p_head != NULL)
6196 rp = p_head->p_next;
6197 free (p_head->pattern);
6198 free (p_head->name);
6206 * Reads the whole file as a single string from `filebuf' and looks for
6207 * multi-line regular expressions, creating tags on matches.
6208 * readline already dealt with normal regexps.
6210 * Idea by Ben Wing <ben@666.com> (2002).
6213 regex_tag_multiline ()
6215 char *buffer = filebuf.buffer;
6219 for (rp = p_head; rp != NULL; rp = rp->p_next)
6223 if (!rp->multi_line)
6224 continue; /* skip normal regexps */
6226 /* Generic initialisations before parsing file from memory. */
6227 lineno = 1; /* reset global line number */
6228 charno = 0; /* reset global char number */
6229 linecharno = 0; /* reset global char number of line start */
6231 /* Only use generic regexps or those for the current language. */
6232 if (rp->lang != NULL && rp->lang != curfdp->lang)
6235 while (match >= 0 && match < filebuf.len)
6237 match = re_search (rp->pat, buffer, filebuf.len, charno,
6238 filebuf.len - match, &rp->regs);
6243 if (!rp->error_signaled)
6245 error ("regexp stack overflow while matching \"%s\"",
6247 rp->error_signaled = TRUE;
6254 if (match == rp->regs.end[0])
6256 if (!rp->error_signaled)
6258 error ("regexp matches the empty string: \"%s\"",
6260 rp->error_signaled = TRUE;
6262 match = -3; /* exit from while loop */
6266 /* Match occurred. Construct a tag. */
6267 while (charno < rp->regs.end[0])
6268 if (buffer[charno++] == '\n')
6269 lineno++, linecharno = charno;
6270 if (! rp->name || rp->name[0] == '\0')
6272 else /* make a named tag */
6273 name = substitute (buffer, rp->name, &rp->regs);
6274 if (rp->force_explicit_name)
6275 /* Force explicit tag name, if a name
6277 pfnote (name, TRUE, buffer + linecharno,
6278 charno - linecharno + 1, lineno,
6280 else if(name == NULL)
6283 make_tag (name, strlen (name), TRUE,
6284 buffer + linecharno,
6285 charno - linecharno + 1,
6286 lineno, linecharno);
6300 register int len = 0;
6302 while (*cp != '\0' && lowcase (*cp) == lowcase (dbp[len]))
6304 if (*cp == '\0' && !intoken (dbp[len]))
6313 get_tag (bp, namepp)
6317 register char *cp = bp;
6321 /* Go till you get to white space or a syntactic break */
6322 for (cp = bp + 1; !notinname (*cp); cp++)
6324 make_tag (bp, cp - bp, TRUE,
6325 lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
6329 *namepp = savenstr (bp, cp - bp);
6333 * Read a line of text from `stream' into `lbp', excluding the
6334 * newline or CR-NL, if any. Return the number of characters read from
6335 * `stream', which is the length of the line including the newline.
6337 * On DOS or Windows we do not count the CR character, if any before the
6338 * NL, in the returned length; this mirrors the behavior of Emacs on those
6339 * platforms (for text files, it translates CR-NL to NL as it reads in the
6342 * If multi-line regular expressions are requested, each line read is
6343 * appended to `filebuf'.
6346 readline_internal (lbp, stream)
6348 register FILE *stream;
6350 char *buffer = lbp->buffer;
6351 register char *p = lbp->buffer;
6352 register char *pend;
6355 pend = p + lbp->size; /* Separate to avoid 386/IX compiler bug. */
6359 register int c = getc (stream);
6362 /* We're at the end of linebuffer: expand it. */
6364 xrnew (buffer, lbp->size, char);
6365 p += buffer - lbp->buffer;
6366 pend = buffer + lbp->size;
6367 lbp->buffer = buffer;
6377 if (p > buffer && p[-1] == '\r')
6391 lbp->len = p - buffer;
6393 if (need_filebuf /* we need filebuf for multi-line regexps */
6394 && chars_deleted > 0) /* not at EOF */
6396 while (filebuf.size <= filebuf.len + lbp->len + 1) /* +1 for \n */
6398 /* Expand filebuf. */
6400 xrnew (filebuf.buffer, filebuf.size, char);
6402 strncpy (filebuf.buffer + filebuf.len, lbp->buffer, lbp->len);
6403 filebuf.len += lbp->len;
6404 filebuf.buffer[filebuf.len++] = '\n';
6405 filebuf.buffer[filebuf.len] = '\0';
6408 return lbp->len + chars_deleted;
6412 * Like readline_internal, above, but in addition try to match the
6413 * input line against relevant regular expressions and manage #line
6417 readline (lbp, stream)
6423 linecharno = charno; /* update global char number of line start */
6424 result = readline_internal (lbp, stream); /* read line */
6425 lineno += 1; /* increment global line number */
6426 charno += result; /* increment global char number */
6428 /* Honour #line directives. */
6429 if (!no_line_directive)
6431 static bool discard_until_line_directive;
6433 /* Check whether this is a #line directive. */
6434 if (result > 12 && strneq (lbp->buffer, "#line ", 6))
6439 if (sscanf (lbp->buffer, "#line %u \"%n", &lno, &start) >= 1
6440 && start > 0) /* double quote character found */
6442 char *endp = lbp->buffer + start;
6444 while ((endp = etags_strchr (endp, '"')) != NULL
6445 && endp[-1] == '\\')
6448 /* Ok, this is a real #line directive. Let's deal with it. */
6450 char *taggedabsname; /* absolute name of original file */
6451 char *taggedfname; /* name of original file as given */
6452 char *name; /* temp var */
6454 discard_until_line_directive = FALSE; /* found it */
6455 name = lbp->buffer + start;
6457 canonicalize_filename (name);
6458 taggedabsname = absolute_filename (name, tagfiledir);
6459 if (filename_is_absolute (name)
6460 || filename_is_absolute (curfdp->infname))
6461 taggedfname = savestr (taggedabsname);
6463 taggedfname = relative_filename (taggedabsname,tagfiledir);
6465 if (streq (curfdp->taggedfname, taggedfname))
6466 /* The #line directive is only a line number change. We
6467 deal with this afterwards. */
6470 /* The tags following this #line directive should be
6471 attributed to taggedfname. In order to do this, set
6472 curfdp accordingly. */
6474 fdesc *fdp; /* file description pointer */
6476 /* Go look for a file description already set up for the
6477 file indicated in the #line directive. If there is
6478 one, use it from now until the next #line
6480 for (fdp = fdhead; fdp != NULL; fdp = fdp->next)
6481 if (streq (fdp->infname, curfdp->infname)
6482 && streq (fdp->taggedfname, taggedfname))
6483 /* If we remove the second test above (after the &&)
6484 then all entries pertaining to the same file are
6485 coalesced in the tags file. If we use it, then
6486 entries pertaining to the same file but generated
6487 from different files (via #line directives) will
6488 go into separate sections in the tags file. These
6489 alternatives look equivalent. The first one
6490 destroys some apparently useless information. */
6496 /* Else, if we already tagged the real file, skip all
6497 input lines until the next #line directive. */
6498 if (fdp == NULL) /* not found */
6499 for (fdp = fdhead; fdp != NULL; fdp = fdp->next)
6500 if (streq (fdp->infabsname, taggedabsname))
6502 discard_until_line_directive = TRUE;
6506 /* Else create a new file description and use that from
6507 now on, until the next #line directive. */
6508 if (fdp == NULL) /* not found */
6511 fdhead = xnew (1, fdesc);
6512 *fdhead = *curfdp; /* copy curr. file description */
6514 fdhead->infname = savestr (curfdp->infname);
6515 fdhead->infabsname = savestr (curfdp->infabsname);
6516 fdhead->infabsdir = savestr (curfdp->infabsdir);
6517 fdhead->taggedfname = taggedfname;
6518 fdhead->usecharno = FALSE;
6519 fdhead->prop = NULL;
6520 fdhead->written = FALSE;
6524 free (taggedabsname);
6526 readline (lbp, stream);
6528 } /* if a real #line directive */
6529 } /* if #line is followed by a a number */
6530 } /* if line begins with "#line " */
6532 /* If we are here, no #line directive was found. */
6533 if (discard_until_line_directive)
6537 /* Do a tail recursion on ourselves, thus discarding the contents
6538 of the line buffer. */
6539 readline (lbp, stream);
6543 discard_until_line_directive = FALSE;
6546 } /* if #line directives should be considered */
6553 /* Match against relevant regexps. */
6555 for (rp = p_head; rp != NULL; rp = rp->p_next)
6557 /* Only use generic regexps or those for the current language.
6558 Also do not use multiline regexps, which is the job of
6559 regex_tag_multiline. */
6560 if ((rp->lang != NULL && rp->lang != fdhead->lang)
6564 match = re_match (rp->pat, lbp->buffer, lbp->len, 0, &rp->regs);
6569 if (!rp->error_signaled)
6571 error ("regexp stack overflow while matching \"%s\"",
6573 rp->error_signaled = TRUE;
6580 /* Empty string matched. */
6581 if (!rp->error_signaled)
6583 error ("regexp matches the empty string: \"%s\"", rp->pattern);
6584 rp->error_signaled = TRUE;
6588 /* Match occurred. Construct a tag. */
6589 if (rp->name[0] != '\0')
6590 /* make a named tag */
6591 name = substitute (lbp->buffer, rp->name, &rp->regs);
6592 if (rp->force_explicit_name)
6593 /* Force explicit tag name, if a name is there. */
6594 pfnote (name, TRUE, lbp->buffer, match, lineno, linecharno);
6596 make_tag (name, strlen (name), TRUE,
6597 lbp->buffer, match, lineno, linecharno);
6601 make_tag (rp->name, strlen (rp->name), TRUE,
6602 lbp->buffer, match, lineno, linecharno);
6611 * Return a pointer to a space of size strlen(cp)+1 allocated
6612 * with xnew where the string CP has been copied.
6618 return savenstr (cp, strlen (cp));
6622 * Return a pointer to a space of size LEN+1 allocated with xnew where
6623 * the string CP has been copied for at most the first LEN characters.
6632 dp = xnew (len + 1, char);
6633 strncpy (dp, cp, len);
6639 * Return the ptr in sp at which the character c last
6640 * appears; NULL if not found
6642 * Identical to POSIX strrchr, included for portability.
6645 etags_strrchr (sp, c)
6646 register const char *sp;
6649 register const char *r;
6661 * Return the ptr in sp at which the character c first
6662 * appears; NULL if not found
6664 * Identical to POSIX strchr, included for portability.
6667 etags_strchr (sp, c)
6668 register const char *sp;
6680 * Compare two strings, ignoring case for alphabetic characters.
6682 * Same as BSD's strcasecmp, included for portability.
6685 etags_strcasecmp (s1, s2)
6686 register const char *s1;
6687 register const char *s2;
6690 && (ISALPHA (*s1) && ISALPHA (*s2)
6691 ? lowcase (*s1) == lowcase (*s2)
6695 return (ISALPHA (*s1) && ISALPHA (*s2)
6696 ? lowcase (*s1) - lowcase (*s2)
6701 * Compare two strings, ignoring case for alphabetic characters.
6702 * Stop after a given number of characters
6704 * Same as BSD's strncasecmp, included for portability.
6707 etags_strncasecmp (s1, s2, n)
6708 register const char *s1;
6709 register const char *s2;
6712 while (*s1 != '\0' && n-- > 0
6713 && (ISALPHA (*s1) && ISALPHA (*s2)
6714 ? lowcase (*s1) == lowcase (*s2)
6721 return (ISALPHA (*s1) && ISALPHA (*s2)
6722 ? lowcase (*s1) - lowcase (*s2)
6726 /* Skip spaces (end of string is not space), return new pointer. */
6731 while (iswhite (*cp))
6736 /* Skip non spaces, except end of string, return new pointer. */
6738 skip_non_spaces (cp)
6741 while (*cp != '\0' && !iswhite (*cp))
6746 /* Print error message and exit. */
6752 exit (EXIT_FAILURE);
6760 exit (EXIT_FAILURE);
6764 suggest_asking_for_help ()
6766 fprintf (stderr, "\tTry `%s %s' for a complete list of options.\n",
6767 progname, NO_LONG_OPTIONS ? "-h" : "--help");
6768 exit (EXIT_FAILURE);
6771 /* Print error message. `s1' is printf control string, `s2' is arg for it. */
6774 const char *s1, *s2;
6776 fprintf (stderr, "%s: ", progname);
6777 fprintf (stderr, s1, s2);
6778 fprintf (stderr, "\n");
6781 /* Return a newly-allocated string whose contents
6782 concatenate those of s1, s2, s3. */
6787 int len1 = strlen (s1), len2 = strlen (s2), len3 = strlen (s3);
6788 char *result = xnew (len1 + len2 + len3 + 1, char);
6790 strncpy(result, s1, len1+1);
6791 strncpy(result + len1, s2, len2+1);
6792 strncpy(result + len1 + len2, s3, len3+1);
6793 result[len1 + len2 + len3] = '\0';
6799 /* Does the same work as the system V getcwd, but does not need to
6800 guess the buffer size in advance. */
6806 char *path = xnew (bufsize, char);
6808 while (getcwd (path, bufsize) == NULL)
6810 if (errno != ERANGE)
6814 path = xnew (bufsize, char);
6817 canonicalize_filename (path);
6820 #else /* not HAVE_GETCWD */
6824 linebuffer_init (&path);
6825 pipe = (FILE *) popen ("pwd 2>/dev/null", "r");
6826 if (pipe == NULL || readline_internal (&path, pipe) == 0)
6831 #endif /* not HAVE_GETCWD */
6834 /* Return a newly allocated string containing the file name of FILE
6835 relative to the absolute directory DIR (which should end with a slash). */
6837 relative_filename (file, dir)
6840 char *fp, *dp, *afn, *res;
6844 /* Find the common root of file and dir (with a trailing slash). */
6845 afn = absolute_filename (file, cwd);
6848 while (*fp++ == *dp++)
6850 fp--, dp--; /* back to the first differing char */
6851 do /* look at the equal chars until '/' */
6854 fp ++; /* Advance past the '/' */
6856 /* Build a sequence of "../" strings for the resulting relative file name. */
6858 while ((dp = etags_strchr (dp + 1, '/')) != NULL)
6860 res_left = 3 * i + strlen(fp);
6861 res = xnew( res_left + 1, char);
6863 for ( ; i-- > 0 ; res_left -= 4 )
6864 strncat(res, "../", res_left );
6866 /* Add the file name relative to the common root of file and dir. */
6867 strncat(res, fp, res_left);
6873 /* Return a newly allocated string containing the absolute file name
6874 of FILE given DIR (which should end with a slash). */
6876 absolute_filename (file, dir)
6879 char *slashp, *cp, *res;
6881 if (filename_is_absolute (file))
6882 res = savestr (file);
6884 res = concat (dir, file, "");
6886 /* Delete the "/dirname/.." and "/." substrings. */
6887 slashp = etags_strchr (res, '/');
6888 while (slashp != NULL && slashp[0] != '\0')
6890 if (slashp[1] == '.')
6892 if (slashp[2] == '.'
6893 && (slashp[3] == '/' || slashp[3] == '\0'))
6898 while (cp >= res && !filename_is_absolute (cp));
6900 cp = slashp; /* the absolute name begins with "/.." */
6901 strcpy (cp, slashp + 3);
6905 else if (slashp[2] == '/' || slashp[2] == '\0')
6907 strcpy (slashp, slashp + 2);
6912 slashp = etags_strchr (slashp + 1, '/');
6915 if (res[0] == '\0') /* just a safety net: should never happen */
6918 return savestr ("/");
6924 /* Return a newly allocated string containing the absolute
6925 file name of dir where FILE resides given DIR (which should
6926 end with a slash). */
6928 absolute_dirname (file, dir)
6934 slashp = etags_strrchr (file, '/');
6936 return savestr (dir);
6939 res = absolute_filename (file, dir);
6945 /* Whether the argument string is an absolute file name. The argument
6946 string must have been canonicalized with canonicalize_filename. */
6948 filename_is_absolute (fn)
6951 return (fn[0] == '/');
6954 /* Upcase DOS drive letter and collapse separators into single slashes.
6957 canonicalize_filename (fn)
6963 /* Collapse multiple separators into a single slash. */
6964 for (cp = fn; *cp != '\0'; cp++, fn++)
6968 while (cp[1] == sep)
6977 /* Initialize a linebuffer for use. */
6979 linebuffer_init (lbp)
6982 lbp->size = (DEBUG) ? 3 : 200;
6983 lbp->buffer = xnew (lbp->size, char);
6984 lbp->buffer[0] = '\0';
6988 /* Set the minimum size of a string contained in a linebuffer. */
6990 linebuffer_setlen (lbp, toksize)
6994 while (lbp->size <= toksize)
6997 xrnew (lbp->buffer, lbp->size, char);
7002 /* Like malloc but get fatal error if memory is exhausted. */
7007 PTR result = (PTR) malloc (size);
7009 fatal ("virtual memory exhausted", (char *)NULL);
7014 xrealloc (ptr, size)
7018 PTR result = (PTR) realloc (ptr, size);
7020 fatal ("virtual memory exhausted", (char *)NULL);
7026 * indent-tabs-mode: t
7029 * c-font-lock-extra-types: ("FILE" "bool" "language" "linebuffer" "fdesc" "node" "regexp")
7033 /* etags.c ends here */