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, 1987-1989, 1993-1995, 1998-2011
32 Free Software Foundation, Inc.
34 This file is not considered part of GNU Emacs.
36 This program is free software: you can redistribute it and/or modify
37 it under the terms of the GNU General Public License as published by
38 the Free Software Foundation, either version 3 of the License, or
39 (at your option) any later version.
41 This program is distributed in the hope that it will be useful,
42 but WITHOUT ANY WARRANTY; without even the implied warranty of
43 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
44 GNU General Public License for more details.
46 You should have received a copy of the GNU General Public License
47 along with this program. If not, see <http://www.gnu.org/licenses/>. */
50 /* NB To comply with the above BSD license, copyright information is
51 reproduced in etc/ETAGS.README. That file should be updated when the
54 To the best of our knowledge, this code was originally based on the
55 ctags.c distributed with BSD4.2, which was copyrighted by the
56 University of California, as described above. */
61 * 1983 Ctags originally by Ken Arnold.
62 * 1984 Fortran added by Jim Kleckner.
63 * 1984 Ed Pelegri-Llopart added C typedefs.
64 * 1985 Emacs TAGS format by Richard Stallman.
65 * 1989 Sam Kendall added C++.
66 * 1992 Joseph B. Wells improved C and C++ parsing.
67 * 1993 Francesco Potortì reorganized C and C++.
68 * 1994 Line-by-line regexp tags by Tom Tromey.
69 * 2001 Nested classes by Francesco Potortì (concept by Mykola Dzyuba).
70 * 2002 #line directives by Francesco Potortì.
72 * Francesco Potortì <pot@gnu.org> has maintained and improved it since 1993.
76 * If you want to add support for a new language, start by looking at the LUA
77 * language, which is the simplest. Alternatively, consider distributing etags
78 * together with a configuration file containing regexp definitions for etags.
81 char pot_etags_version
[] = "@(#) pot revision number is 17.38.1.4";
91 # define NDEBUG /* disable assert */
96 /* This is probably not necessary any more. On some systems, config.h
97 used to define static as nothing for the sake of unexec. We don't
98 want that here since we don't use unexec. None of these systems
99 are supported any more, but the idea is still mentioned in
102 # ifndef PTR /* for XEmacs */
105 #else /* no config.h */
106 # if defined(__STDC__) && (__STDC__ || defined(__SUNPRO_C))
107 # define PTR void * /* for generic pointers */
108 # else /* not standard C */
109 # define const /* remove const for old compilers' sake */
110 # define PTR long * /* don't use void* */
112 #endif /* !HAVE_CONFIG_H */
115 # define _GNU_SOURCE 1 /* enables some compiler checks on GNU */
118 /* WIN32_NATIVE is for XEmacs.
119 MSDOS, WINDOWSNT, DOS_NT are for Emacs. */
124 #endif /* WIN32_NATIVE */
130 # include <sys/param.h>
132 # ifndef HAVE_CONFIG_H
134 # include <sys/config.h>
146 # define MAXPATHLEN _MAX_PATH
152 # endif /* undef HAVE_GETCWD */
153 #else /* not WINDOWSNT */
157 # else /* no standard C headers */
158 extern char *getenv (const char *);
159 extern char *strcpy (char *, const char *);
160 extern char *strncpy (char *, const char *, unsigned long);
161 extern char *strcat (char *, const char *);
162 extern char *strncat (char *, const char *, unsigned long);
163 extern int strcmp (const char *, const char *);
164 extern int strncmp (const char *, const char *, unsigned long);
165 extern int system (const char *);
166 extern unsigned long strlen (const char *);
167 extern void *malloc (unsigned long);
168 extern void *realloc (void *, unsigned long);
169 extern void exit (int);
170 extern void free (void *);
171 extern void *memmove (void *, const void *, unsigned long);
172 # define EXIT_SUCCESS 0
173 # define EXIT_FAILURE 1
175 #endif /* !WINDOWSNT */
178 #ifndef HAVE_UNISTD_H
179 # if defined (HAVE_GETCWD) && !defined (WINDOWSNT)
180 extern char *getcwd (char *buf
, size_t size
);
182 #endif /* HAVE_UNISTD_H */
187 #include <sys/types.h>
188 #include <sys/stat.h>
192 # undef assert /* some systems have a buggy assert.h */
193 # define assert(x) ((void) 0)
196 #ifdef NO_LONG_OPTIONS /* define this if you don't have GNU getopt */
197 # define NO_LONG_OPTIONS TRUE
198 # define getopt_long(argc,argv,optstr,lopts,lind) getopt (argc, argv, optstr)
200 extern int optind
, opterr
;
202 # define NO_LONG_OPTIONS FALSE
204 #endif /* NO_LONG_OPTIONS */
206 #ifndef HAVE_CONFIG_H /* this is a standalone compilation */
207 # ifdef __CYGWIN__ /* compiling on Cygwin */
209 the regex
.h distributed with Cygwin is
not compatible with etags
, alas
!
210 If you want regular expression support
, you should
delete this notice
and
211 arrange to use the GNU regex
.h
and regex
.c
.
216 /* Define CTAGS to make the program "ctags" compatible with the usual one.
217 Leave it undefined to make the program "etags", which makes emacs-style
218 tag tables and tags typedefs, #defines and struct/union/enum by default. */
226 #define streq(s,t) (assert((s)!=NULL || (t)!=NULL), !strcmp (s, t))
227 #define strcaseeq(s,t) (assert((s)!=NULL && (t)!=NULL), !etags_strcasecmp (s, t))
228 #define strneq(s,t,n) (assert((s)!=NULL || (t)!=NULL), !strncmp (s, t, n))
229 #define strncaseeq(s,t,n) (assert((s)!=NULL && (t)!=NULL), !etags_strncasecmp (s, t, n))
231 #define CHARS 256 /* 2^sizeof(char) */
232 #define CHAR(x) ((unsigned int)(x) & (CHARS - 1))
233 #define iswhite(c) (_wht[CHAR(c)]) /* c is white (see white) */
234 #define notinname(c) (_nin[CHAR(c)]) /* c is not in a name (see nonam) */
235 #define begtoken(c) (_btk[CHAR(c)]) /* c can start token (see begtk) */
236 #define intoken(c) (_itk[CHAR(c)]) /* c can be in token (see midtk) */
237 #define endtoken(c) (_etk[CHAR(c)]) /* c ends tokens (see endtk) */
239 #define ISALNUM(c) isalnum (CHAR(c))
240 #define ISALPHA(c) isalpha (CHAR(c))
241 #define ISDIGIT(c) isdigit (CHAR(c))
242 #define ISLOWER(c) islower (CHAR(c))
244 #define lowcase(c) tolower (CHAR(c))
248 * xnew, xrnew -- allocate, reallocate storage
250 * SYNOPSIS: Type *xnew (int n, Type);
251 * void xrnew (OldPointer, int n, Type);
254 # include "chkmalloc.h"
255 # define xnew(n,Type) ((Type *) trace_malloc (__FILE__, __LINE__, \
256 (n) * sizeof (Type)))
257 # define xrnew(op,n,Type) ((op) = (Type *) trace_realloc (__FILE__, __LINE__, \
258 (char *) (op), (n) * sizeof (Type)))
260 # define xnew(n,Type) ((Type *) xmalloc ((n) * sizeof (Type)))
261 # define xrnew(op,n,Type) ((op) = (Type *) xrealloc ( \
262 (char *) (op), (n) * sizeof (Type)))
267 typedef void Lang_function (FILE *);
271 const char *suffix
; /* file name suffix for this compressor */
272 const char *command
; /* takes one arg and decompresses to stdout */
277 const char *name
; /* language name */
278 const char *help
; /* detailed help for the language */
279 Lang_function
*function
; /* parse function */
280 const char **suffixes
; /* name suffixes of this language's files */
281 const char **filenames
; /* names of this language's files */
282 const char **interpreters
; /* interpreters for this language */
283 bool metasource
; /* source used to generate other sources */
288 struct fdesc
*next
; /* for the linked list */
289 char *infname
; /* uncompressed input file name */
290 char *infabsname
; /* absolute uncompressed input file name */
291 char *infabsdir
; /* absolute dir of input file */
292 char *taggedfname
; /* file name to write in tagfile */
293 language
*lang
; /* language of file */
294 char *prop
; /* file properties to write in tagfile */
295 bool usecharno
; /* etags tags shall contain char number */
296 bool written
; /* entry written in the tags file */
299 typedef struct node_st
300 { /* sorting structure */
301 struct node_st
*left
, *right
; /* left and right sons */
302 fdesc
*fdp
; /* description of file to whom tag belongs */
303 char *name
; /* tag name */
304 char *regex
; /* search regexp */
305 bool valid
; /* write this tag on the tag file */
306 bool is_func
; /* function tag: use regexp in CTAGS mode */
307 bool been_warned
; /* warning already given for duplicated tag */
308 int lno
; /* line number tag is on */
309 long cno
; /* character number line starts on */
313 * A `linebuffer' is a structure which holds a line of text.
314 * `readline_internal' reads a line from a stream into a linebuffer
315 * and works regardless of the length of the line.
316 * SIZE is the size of BUFFER, LEN is the length of the string in
317 * BUFFER after readline reads it.
326 /* Used to support mixing of --lang and file names. */
330 at_language
, /* a language specification */
331 at_regexp
, /* a regular expression */
332 at_filename
, /* a file name */
333 at_stdin
, /* read from stdin here */
334 at_end
/* stop parsing the list */
335 } arg_type
; /* argument type */
336 language
*lang
; /* language associated with the argument */
337 char *what
; /* the argument itself */
340 /* Structure defining a regular expression. */
341 typedef struct regexp
343 struct regexp
*p_next
; /* pointer to next in list */
344 language
*lang
; /* if set, use only for this language */
345 char *pattern
; /* the regexp pattern */
346 char *name
; /* tag name */
347 struct re_pattern_buffer
*pat
; /* the compiled pattern */
348 struct re_registers regs
; /* re registers */
349 bool error_signaled
; /* already signaled for this regexp */
350 bool force_explicit_name
; /* do not allow implict tag name */
351 bool ignore_case
; /* ignore case when matching */
352 bool multi_line
; /* do a multi-line match on the whole file */
356 /* Many compilers barf on this:
357 Lang_function Ada_funcs;
358 so let's write it this way */
359 static void Ada_funcs (FILE *);
360 static void Asm_labels (FILE *);
361 static void C_entries (int c_ext
, FILE *);
362 static void default_C_entries (FILE *);
363 static void plain_C_entries (FILE *);
364 static void Cjava_entries (FILE *);
365 static void Cobol_paragraphs (FILE *);
366 static void Cplusplus_entries (FILE *);
367 static void Cstar_entries (FILE *);
368 static void Erlang_functions (FILE *);
369 static void Forth_words (FILE *);
370 static void Fortran_functions (FILE *);
371 static void HTML_labels (FILE *);
372 static void Lisp_functions (FILE *);
373 static void Lua_functions (FILE *);
374 static void Makefile_targets (FILE *);
375 static void Pascal_functions (FILE *);
376 static void Perl_functions (FILE *);
377 static void PHP_functions (FILE *);
378 static void PS_functions (FILE *);
379 static void Prolog_functions (FILE *);
380 static void Python_functions (FILE *);
381 static void Scheme_functions (FILE *);
382 static void TeX_commands (FILE *);
383 static void Texinfo_nodes (FILE *);
384 static void Yacc_entries (FILE *);
385 static void just_read_file (FILE *);
387 static void print_language_names (void);
388 static void print_version (void);
389 static void print_help (argument
*);
390 int main (int, char **);
392 static compressor
*get_compressor_from_suffix (char *, char **);
393 static language
*get_language_from_langname (const char *);
394 static language
*get_language_from_interpreter (char *);
395 static language
*get_language_from_filename (char *, bool);
396 static void readline (linebuffer
*, FILE *);
397 static long readline_internal (linebuffer
*, FILE *);
398 static bool nocase_tail (const char *);
399 static void get_tag (char *, char **);
401 static void analyse_regex (char *);
402 static void free_regexps (void);
403 static void regex_tag_multiline (void);
404 static void error (const char *, const char *);
405 static void suggest_asking_for_help (void) NO_RETURN
;
406 void fatal (const char *, const char *) NO_RETURN
;
407 static void pfatal (const char *) NO_RETURN
;
408 static void add_node (node
*, node
**);
410 static void init (void);
411 static void process_file_name (char *, language
*);
412 static void process_file (FILE *, char *, language
*);
413 static void find_entries (FILE *);
414 static void free_tree (node
*);
415 static void free_fdesc (fdesc
*);
416 static void pfnote (char *, bool, char *, int, int, long);
417 static void make_tag (const char *, int, bool, char *, int, int, long);
418 static void invalidate_nodes (fdesc
*, node
**);
419 static void put_entries (node
*);
421 static char *concat (const char *, const char *, const char *);
422 static char *skip_spaces (char *);
423 static char *skip_non_spaces (char *);
424 static char *savenstr (const char *, int);
425 static char *savestr (const char *);
426 static char *etags_strchr (const char *, int);
427 static char *etags_strrchr (const char *, int);
428 static int etags_strcasecmp (const char *, const char *);
429 static int etags_strncasecmp (const char *, const char *, int);
430 static char *etags_getcwd (void);
431 static char *relative_filename (char *, char *);
432 static char *absolute_filename (char *, char *);
433 static char *absolute_dirname (char *, char *);
434 static bool filename_is_absolute (char *f
);
435 static void canonicalize_filename (char *);
436 static void linebuffer_init (linebuffer
*);
437 static void linebuffer_setlen (linebuffer
*, int);
438 static PTR
xmalloc (unsigned int);
439 static PTR
xrealloc (char *, unsigned int);
442 static char searchar
= '/'; /* use /.../ searches */
444 static char *tagfile
; /* output file */
445 static char *progname
; /* name this program was invoked with */
446 static char *cwd
; /* current working directory */
447 static char *tagfiledir
; /* directory of tagfile */
448 static FILE *tagf
; /* ioptr for tags file */
450 static fdesc
*fdhead
; /* head of file description list */
451 static fdesc
*curfdp
; /* current file description */
452 static int lineno
; /* line number of current line */
453 static long charno
; /* current character number */
454 static long linecharno
; /* charno of start of current line */
455 static char *dbp
; /* pointer to start of current tag */
457 static const int invalidcharno
= -1;
459 static node
*nodehead
; /* the head of the binary tree of tags */
460 static node
*last_node
; /* the last node created */
462 static linebuffer lb
; /* the current line */
463 static linebuffer filebuf
; /* a buffer containing the whole file */
464 static linebuffer token_name
; /* a buffer containing a tag name */
466 /* boolean "functions" (see init) */
467 static bool _wht
[CHARS
], _nin
[CHARS
], _itk
[CHARS
], _btk
[CHARS
], _etk
[CHARS
];
470 *white
= " \f\t\n\r\v",
472 *nonam
= " \f\t\n\r()=,;", /* look at make_tag before modifying! */
473 /* token ending chars */
474 *endtk
= " \t\n\r\"'#()[]{}=-+%*/&|^~!<>;,.:?",
475 /* token starting chars */
476 *begtk
= "ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz$~@",
477 /* valid in-token chars */
478 *midtk
= "ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz$0123456789";
480 static bool append_to_tagfile
; /* -a: append to tags */
481 /* The next five default to TRUE in C and derived languages. */
482 static bool typedefs
; /* -t: create tags for C and Ada typedefs */
483 static bool typedefs_or_cplusplus
; /* -T: create tags for C typedefs, level */
484 /* 0 struct/enum/union decls, and C++ */
485 /* member functions. */
486 static bool constantypedefs
; /* -d: create tags for C #define, enum */
487 /* constants and variables. */
488 /* -D: opposite of -d. Default under ctags. */
489 static bool globals
; /* create tags for global variables */
490 static bool members
; /* create tags for C member variables */
491 static bool declarations
; /* --declarations: tag them and extern in C&Co*/
492 static bool no_line_directive
; /* ignore #line directives (undocumented) */
493 static bool no_duplicates
; /* no duplicate tags for ctags (undocumented) */
494 static bool update
; /* -u: update tags */
495 static bool vgrind_style
; /* -v: create vgrind style index output */
496 static bool no_warnings
; /* -w: suppress warnings (undocumented) */
497 static bool cxref_style
; /* -x: create cxref style output */
498 static bool cplusplus
; /* .[hc] means C++, not C (undocumented) */
499 static bool ignoreindent
; /* -I: ignore indentation in C */
500 static bool packages_only
; /* --packages-only: in Ada, only tag packages*/
502 /* STDIN is defined in LynxOS system headers */
507 #define STDIN 0x1001 /* returned by getopt_long on --parse-stdin */
508 static bool parsing_stdin
; /* --parse-stdin used */
510 static regexp
*p_head
; /* list of all regexps */
511 static bool need_filebuf
; /* some regexes are multi-line */
513 static struct option longopts
[] =
515 { "append", no_argument
, NULL
, 'a' },
516 { "packages-only", no_argument
, &packages_only
, TRUE
},
517 { "c++", no_argument
, NULL
, 'C' },
518 { "declarations", no_argument
, &declarations
, TRUE
},
519 { "no-line-directive", no_argument
, &no_line_directive
, TRUE
},
520 { "no-duplicates", no_argument
, &no_duplicates
, TRUE
},
521 { "help", no_argument
, NULL
, 'h' },
522 { "help", no_argument
, NULL
, 'H' },
523 { "ignore-indentation", no_argument
, NULL
, 'I' },
524 { "language", required_argument
, NULL
, 'l' },
525 { "members", no_argument
, &members
, TRUE
},
526 { "no-members", no_argument
, &members
, FALSE
},
527 { "output", required_argument
, NULL
, 'o' },
528 { "regex", required_argument
, NULL
, 'r' },
529 { "no-regex", no_argument
, NULL
, 'R' },
530 { "ignore-case-regex", required_argument
, NULL
, 'c' },
531 { "parse-stdin", required_argument
, NULL
, STDIN
},
532 { "version", no_argument
, NULL
, 'V' },
534 #if CTAGS /* Ctags options */
535 { "backward-search", no_argument
, NULL
, 'B' },
536 { "cxref", no_argument
, NULL
, 'x' },
537 { "defines", no_argument
, NULL
, 'd' },
538 { "globals", no_argument
, &globals
, TRUE
},
539 { "typedefs", no_argument
, NULL
, 't' },
540 { "typedefs-and-c++", no_argument
, NULL
, 'T' },
541 { "update", no_argument
, NULL
, 'u' },
542 { "vgrind", no_argument
, NULL
, 'v' },
543 { "no-warn", no_argument
, NULL
, 'w' },
545 #else /* Etags options */
546 { "no-defines", no_argument
, NULL
, 'D' },
547 { "no-globals", no_argument
, &globals
, FALSE
},
548 { "include", required_argument
, NULL
, 'i' },
553 static compressor compressors
[] =
555 { "z", "gzip -d -c"},
556 { "Z", "gzip -d -c"},
557 { "gz", "gzip -d -c"},
558 { "GZ", "gzip -d -c"},
559 { "bz2", "bzip2 -d -c" },
560 { "xz", "xz -d -c" },
569 static const char *Ada_suffixes
[] =
570 { "ads", "adb", "ada", NULL
};
571 static const char Ada_help
[] =
572 "In Ada code, functions, procedures, packages, tasks and types are\n\
573 tags. Use the `--packages-only' option to create tags for\n\
575 Ada tag names have suffixes indicating the type of entity:\n\
576 Entity type: Qualifier:\n\
577 ------------ ----------\n\
584 Thus, `M-x find-tag <RET> bidule/b <RET>' will go directly to the\n\
585 body of the package `bidule', while `M-x find-tag <RET> bidule <RET>'\n\
586 will just search for any tag `bidule'.";
589 static const char *Asm_suffixes
[] =
590 { "a", /* Unix assembler */
591 "asm", /* Microcontroller assembly */
592 "def", /* BSO/Tasking definition includes */
593 "inc", /* Microcontroller include files */
594 "ins", /* Microcontroller include files */
595 "s", "sa", /* Unix assembler */
596 "S", /* cpp-processed Unix assembler */
597 "src", /* BSO/Tasking C compiler output */
600 static const char Asm_help
[] =
601 "In assembler code, labels appearing at the beginning of a line,\n\
602 followed by a colon, are tags.";
605 /* Note that .c and .h can be considered C++, if the --c++ flag was
606 given, or if the `class' or `template' keywords are met inside the file.
607 That is why default_C_entries is called for these. */
608 static const char *default_C_suffixes
[] =
610 #if CTAGS /* C help for Ctags */
611 static const char default_C_help
[] =
612 "In C code, any C function is a tag. Use -t to tag typedefs.\n\
613 Use -T to tag definitions of `struct', `union' and `enum'.\n\
614 Use -d to tag `#define' macro definitions and `enum' constants.\n\
615 Use --globals to tag global variables.\n\
616 You can tag function declarations and external variables by\n\
617 using `--declarations', and struct members by using `--members'.";
618 #else /* C help for Etags */
619 static const char default_C_help
[] =
620 "In C code, any C function or typedef is a tag, and so are\n\
621 definitions of `struct', `union' and `enum'. `#define' macro\n\
622 definitions and `enum' constants are tags unless you specify\n\
623 `--no-defines'. Global variables are tags unless you specify\n\
624 `--no-globals' and so are struct members unless you specify\n\
625 `--no-members'. Use of `--no-globals', `--no-defines' and\n\
626 `--no-members' can make the tags table file much smaller.\n\
627 You can tag function declarations and external variables by\n\
628 using `--declarations'.";
629 #endif /* C help for Ctags and Etags */
631 static const char *Cplusplus_suffixes
[] =
632 { "C", "c++", "cc", "cpp", "cxx", "H", "h++", "hh", "hpp", "hxx",
633 "M", /* Objective C++ */
634 "pdb", /* Postscript with C syntax */
636 static const char Cplusplus_help
[] =
637 "In C++ code, all the tag constructs of C code are tagged. (Use\n\
638 --help --lang=c --lang=c++ for full help.)\n\
639 In addition to C tags, member functions are also recognized. Member\n\
640 variables are recognized unless you use the `--no-members' option.\n\
641 Tags for variables and functions in classes are named `CLASS::VARIABLE'\n\
642 and `CLASS::FUNCTION'. `operator' definitions have tag names like\n\
645 static const char *Cjava_suffixes
[] =
647 static char Cjava_help
[] =
648 "In Java code, all the tags constructs of C and C++ code are\n\
649 tagged. (Use --help --lang=c --lang=c++ --lang=java for full help.)";
652 static const char *Cobol_suffixes
[] =
653 { "COB", "cob", NULL
};
654 static char Cobol_help
[] =
655 "In Cobol code, tags are paragraph names; that is, any word\n\
656 starting in column 8 and followed by a period.";
658 static const char *Cstar_suffixes
[] =
659 { "cs", "hs", NULL
};
661 static const char *Erlang_suffixes
[] =
662 { "erl", "hrl", NULL
};
663 static const char Erlang_help
[] =
664 "In Erlang code, the tags are the functions, records and macros\n\
665 defined in the file.";
667 const char *Forth_suffixes
[] =
668 { "fth", "tok", NULL
};
669 static const char Forth_help
[] =
670 "In Forth code, tags are words defined by `:',\n\
671 constant, code, create, defer, value, variable, buffer:, field.";
673 static const char *Fortran_suffixes
[] =
674 { "F", "f", "f90", "for", NULL
};
675 static const char Fortran_help
[] =
676 "In Fortran code, functions, subroutines and block data are tags.";
678 static const char *HTML_suffixes
[] =
679 { "htm", "html", "shtml", NULL
};
680 static const char HTML_help
[] =
681 "In HTML input files, the tags are the `title' and the `h1', `h2',\n\
682 `h3' headers. Also, tags are `name=' in anchors and all\n\
683 occurrences of `id='.";
685 static const char *Lisp_suffixes
[] =
686 { "cl", "clisp", "el", "l", "lisp", "LSP", "lsp", "ml", NULL
};
687 static const char Lisp_help
[] =
688 "In Lisp code, any function defined with `defun', any variable\n\
689 defined with `defvar' or `defconst', and in general the first\n\
690 argument of any expression that starts with `(def' in column zero\n\
693 static const char *Lua_suffixes
[] =
694 { "lua", "LUA", NULL
};
695 static const char Lua_help
[] =
696 "In Lua scripts, all functions are tags.";
698 static const char *Makefile_filenames
[] =
699 { "Makefile", "makefile", "GNUMakefile", "Makefile.in", "Makefile.am", NULL
};
700 static const char Makefile_help
[] =
701 "In makefiles, targets are tags; additionally, variables are tags\n\
702 unless you specify `--no-globals'.";
704 static const char *Objc_suffixes
[] =
705 { "lm", /* Objective lex file */
706 "m", /* Objective C file */
708 static const char Objc_help
[] =
709 "In Objective C code, tags include Objective C definitions for classes,\n\
710 class categories, methods and protocols. Tags for variables and\n\
711 functions in classes are named `CLASS::VARIABLE' and `CLASS::FUNCTION'.\n\
712 (Use --help --lang=c --lang=objc --lang=java for full help.)";
714 static const char *Pascal_suffixes
[] =
715 { "p", "pas", NULL
};
716 static const char Pascal_help
[] =
717 "In Pascal code, the tags are the functions and procedures defined\n\
719 /* " // this is for working around an Emacs highlighting bug... */
721 static const char *Perl_suffixes
[] =
722 { "pl", "pm", NULL
};
723 static const char *Perl_interpreters
[] =
724 { "perl", "@PERL@", NULL
};
725 static const char Perl_help
[] =
726 "In Perl code, the tags are the packages, subroutines and variables\n\
727 defined by the `package', `sub', `my' and `local' keywords. Use\n\
728 `--globals' if you want to tag global variables. Tags for\n\
729 subroutines are named `PACKAGE::SUB'. The name for subroutines\n\
730 defined in the default package is `main::SUB'.";
732 static const char *PHP_suffixes
[] =
733 { "php", "php3", "php4", NULL
};
734 static const char PHP_help
[] =
735 "In PHP code, tags are functions, classes and defines. Unless you use\n\
736 the `--no-members' option, vars are tags too.";
738 static const char *plain_C_suffixes
[] =
739 { "pc", /* Pro*C file */
742 static const char *PS_suffixes
[] =
743 { "ps", "psw", NULL
}; /* .psw is for PSWrap */
744 static const char PS_help
[] =
745 "In PostScript code, the tags are the functions.";
747 static const char *Prolog_suffixes
[] =
749 static const char Prolog_help
[] =
750 "In Prolog code, tags are predicates and rules at the beginning of\n\
753 static const char *Python_suffixes
[] =
755 static const char Python_help
[] =
756 "In Python code, `def' or `class' at the beginning of a line\n\
759 /* Can't do the `SCM' or `scm' prefix with a version number. */
760 static const char *Scheme_suffixes
[] =
761 { "oak", "sch", "scheme", "SCM", "scm", "SM", "sm", "ss", "t", NULL
};
762 static const char Scheme_help
[] =
763 "In Scheme code, tags include anything defined with `def' or with a\n\
764 construct whose name starts with `def'. They also include\n\
765 variables set with `set!' at top level in the file.";
767 static const char *TeX_suffixes
[] =
768 { "bib", "clo", "cls", "ltx", "sty", "TeX", "tex", NULL
};
769 static const char TeX_help
[] =
770 "In LaTeX text, the argument of any of the commands `\\chapter',\n\
771 `\\section', `\\subsection', `\\subsubsection', `\\eqno', `\\label',\n\
772 `\\ref', `\\cite', `\\bibitem', `\\part', `\\appendix', `\\entry',\n\
773 `\\index', `\\def', `\\newcommand', `\\renewcommand',\n\
774 `\\newenvironment' or `\\renewenvironment' is a tag.\n\
776 Other commands can be specified by setting the environment variable\n\
777 `TEXTAGS' to a colon-separated list like, for example,\n\
778 TEXTAGS=\"mycommand:myothercommand\".";
781 static const char *Texinfo_suffixes
[] =
782 { "texi", "texinfo", "txi", NULL
};
783 static const char Texinfo_help
[] =
784 "for texinfo files, lines starting with @node are tagged.";
786 static const char *Yacc_suffixes
[] =
787 { "y", "y++", "ym", "yxx", "yy", NULL
}; /* .ym is Objective yacc file */
788 static const char Yacc_help
[] =
789 "In Bison or Yacc input files, each rule defines as a tag the\n\
790 nonterminal it constructs. The portions of the file that contain\n\
791 C code are parsed as C code (use --help --lang=c --lang=yacc\n\
794 static const char auto_help
[] =
795 "`auto' is not a real language, it indicates to use\n\
796 a default language for files base on file name suffix and file contents.";
798 static const char none_help
[] =
799 "`none' is not a real language, it indicates to only do\n\
800 regexp processing on files.";
802 static const char no_lang_help
[] =
803 "No detailed help available for this language.";
807 * Table of languages.
809 * It is ok for a given function to be listed under more than one
810 * name. I just didn't.
813 static language lang_names
[] =
815 { "ada", Ada_help
, Ada_funcs
, Ada_suffixes
},
816 { "asm", Asm_help
, Asm_labels
, Asm_suffixes
},
817 { "c", default_C_help
, default_C_entries
, default_C_suffixes
},
818 { "c++", Cplusplus_help
, Cplusplus_entries
, Cplusplus_suffixes
},
819 { "c*", no_lang_help
, Cstar_entries
, Cstar_suffixes
},
820 { "cobol", Cobol_help
, Cobol_paragraphs
, Cobol_suffixes
},
821 { "erlang", Erlang_help
, Erlang_functions
, Erlang_suffixes
},
822 { "forth", Forth_help
, Forth_words
, Forth_suffixes
},
823 { "fortran", Fortran_help
, Fortran_functions
, Fortran_suffixes
},
824 { "html", HTML_help
, HTML_labels
, HTML_suffixes
},
825 { "java", Cjava_help
, Cjava_entries
, Cjava_suffixes
},
826 { "lisp", Lisp_help
, Lisp_functions
, Lisp_suffixes
},
827 { "lua", Lua_help
, Lua_functions
, Lua_suffixes
},
828 { "makefile", Makefile_help
,Makefile_targets
,NULL
,Makefile_filenames
},
829 { "objc", Objc_help
, plain_C_entries
, Objc_suffixes
},
830 { "pascal", Pascal_help
, Pascal_functions
, Pascal_suffixes
},
831 { "perl",Perl_help
,Perl_functions
,Perl_suffixes
,NULL
,Perl_interpreters
},
832 { "php", PHP_help
, PHP_functions
, PHP_suffixes
},
833 { "postscript",PS_help
, PS_functions
, PS_suffixes
},
834 { "proc", no_lang_help
, plain_C_entries
, plain_C_suffixes
},
835 { "prolog", Prolog_help
, Prolog_functions
, Prolog_suffixes
},
836 { "python", Python_help
, Python_functions
, Python_suffixes
},
837 { "scheme", Scheme_help
, Scheme_functions
, Scheme_suffixes
},
838 { "tex", TeX_help
, TeX_commands
, TeX_suffixes
},
839 { "texinfo", Texinfo_help
, Texinfo_nodes
, Texinfo_suffixes
},
840 { "yacc", Yacc_help
,Yacc_entries
,Yacc_suffixes
,NULL
,NULL
,TRUE
},
841 { "auto", auto_help
}, /* default guessing scheme */
842 { "none", none_help
, just_read_file
}, /* regexp matching only */
843 { NULL
} /* end of list */
848 print_language_names (void)
851 const char **name
, **ext
;
853 puts ("\nThese are the currently supported languages, along with the\n\
854 default file names and dot suffixes:");
855 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
857 printf (" %-*s", 10, lang
->name
);
858 if (lang
->filenames
!= NULL
)
859 for (name
= lang
->filenames
; *name
!= NULL
; name
++)
860 printf (" %s", *name
);
861 if (lang
->suffixes
!= NULL
)
862 for (ext
= lang
->suffixes
; *ext
!= NULL
; ext
++)
863 printf (" .%s", *ext
);
866 puts ("where `auto' means use default language for files based on file\n\
867 name suffix, and `none' means only do regexp processing on files.\n\
868 If no language is specified and no matching suffix is found,\n\
869 the first line of the file is read for a sharp-bang (#!) sequence\n\
870 followed by the name of an interpreter. If no such sequence is found,\n\
871 Fortran is tried first; if no tags are found, C is tried next.\n\
872 When parsing any C file, a \"class\" or \"template\" keyword\n\
874 puts ("Compressed files are supported using gzip, bzip2, and xz.\n\
876 For detailed help on a given language use, for example,\n\
877 etags --help --lang=ada.");
881 # define EMACS_NAME "standalone"
884 # define VERSION "17.38.1.4"
889 /* Makes it easier to update automatically. */
890 char emacs_copyright
[] = "Copyright (C) 2011 Free Software Foundation, Inc.";
892 printf ("%s (%s %s)\n", (CTAGS
) ? "ctags" : "etags", EMACS_NAME
, VERSION
);
893 puts (emacs_copyright
);
894 puts ("This program is distributed under the terms in ETAGS.README");
899 #ifndef PRINT_UNDOCUMENTED_OPTIONS_HELP
900 # define PRINT_UNDOCUMENTED_OPTIONS_HELP FALSE
904 print_help (argument
*argbuffer
)
906 bool help_for_lang
= FALSE
;
908 for (; argbuffer
->arg_type
!= at_end
; argbuffer
++)
909 if (argbuffer
->arg_type
== at_language
)
913 puts (argbuffer
->lang
->help
);
914 help_for_lang
= TRUE
;
920 printf ("Usage: %s [options] [[regex-option ...] file-name] ...\n\
922 These are the options accepted by %s.\n", progname
, progname
);
924 puts ("WARNING: long option names do not work with this executable,\n\
925 as it is not linked with GNU getopt.");
927 puts ("You may use unambiguous abbreviations for the long option names.");
928 puts (" A - as file name means read names from stdin (one per line).\n\
929 Absolute names are stored in the output file as they are.\n\
930 Relative ones are stored relative to the output file's directory.\n");
932 puts ("-a, --append\n\
933 Append tag entries to existing tags file.");
935 puts ("--packages-only\n\
936 For Ada files, only generate tags for packages.");
939 puts ("-B, --backward-search\n\
940 Write the search commands for the tag entries using '?', the\n\
941 backward-search command instead of '/', the forward-search command.");
943 /* This option is mostly obsolete, because etags can now automatically
944 detect C++. Retained for backward compatibility and for debugging and
945 experimentation. In principle, we could want to tag as C++ even
946 before any "class" or "template" keyword.
948 Treat files whose name suffix defaults to C language as C++ files.");
951 puts ("--declarations\n\
952 In C and derived languages, create tags for function declarations,");
954 puts ("\tand create tags for extern variables if --globals is used.");
957 ("\tand create tags for extern variables unless --no-globals is used.");
960 puts ("-d, --defines\n\
961 Create tag entries for C #define constants and enum constants, too.");
963 puts ("-D, --no-defines\n\
964 Don't create tag entries for C #define constants and enum constants.\n\
965 This makes the tags file smaller.");
968 puts ("-i FILE, --include=FILE\n\
969 Include a note in tag file indicating that, when searching for\n\
970 a tag, one should also consult the tags file FILE after\n\
971 checking the current file.");
973 puts ("-l LANG, --language=LANG\n\
974 Force the following files to be considered as written in the\n\
975 named language up to the next --language=LANG option.");
979 Create tag entries for global variables in some languages.");
981 puts ("--no-globals\n\
982 Do not create tag entries for global variables in some\n\
983 languages. This makes the tags file smaller.");
985 if (PRINT_UNDOCUMENTED_OPTIONS_HELP
)
986 puts ("--no-line-directive\n\
987 Ignore #line preprocessor directives in C and derived languages.");
991 Create tag entries for members of structures in some languages.");
993 puts ("--no-members\n\
994 Do not create tag entries for members of structures\n\
995 in some languages.");
997 puts ("-r REGEXP, --regex=REGEXP or --regex=@regexfile\n\
998 Make a tag for each line matching a regular expression pattern\n\
999 in the following files. {LANGUAGE}REGEXP uses REGEXP for LANGUAGE\n\
1000 files only. REGEXFILE is a file containing one REGEXP per line.\n\
1001 REGEXP takes the form /TAGREGEXP/TAGNAME/MODS, where TAGNAME/ is\n\
1002 optional. The TAGREGEXP pattern is anchored (as if preceded by ^).");
1003 puts (" If TAGNAME/ is present, the tags created are named.\n\
1004 For example Tcl named tags can be created with:\n\
1005 --regex=\"/proc[ \\t]+\\([^ \\t]+\\)/\\1/.\".\n\
1006 MODS are optional one-letter modifiers: `i' means to ignore case,\n\
1007 `m' means to allow multi-line matches, `s' implies `m' and\n\
1008 causes dot to match any character, including newline.");
1010 puts ("-R, --no-regex\n\
1011 Don't create tags from regexps for the following files.");
1013 puts ("-I, --ignore-indentation\n\
1014 In C and C++ do not assume that a closing brace in the first\n\
1015 column is the final brace of a function or structure definition.");
1017 puts ("-o FILE, --output=FILE\n\
1018 Write the tags to FILE.");
1020 puts ("--parse-stdin=NAME\n\
1021 Read from standard input and record tags as belonging to file NAME.");
1025 puts ("-t, --typedefs\n\
1026 Generate tag entries for C and Ada typedefs.");
1027 puts ("-T, --typedefs-and-c++\n\
1028 Generate tag entries for C typedefs, C struct/enum/union tags,\n\
1029 and C++ member functions.");
1033 puts ("-u, --update\n\
1034 Update the tag entries for the given files, leaving tag\n\
1035 entries for other files in place. Currently, this is\n\
1036 implemented by deleting the existing entries for the given\n\
1037 files and then rewriting the new entries at the end of the\n\
1038 tags file. It is often faster to simply rebuild the entire\n\
1039 tag file than to use this.");
1043 puts ("-v, --vgrind\n\
1044 Print on the standard output an index of items intended for\n\
1045 human consumption, similar to the output of vgrind. The index\n\
1046 is sorted, and gives the page number of each item.");
1048 if (PRINT_UNDOCUMENTED_OPTIONS_HELP
)
1049 puts ("-w, --no-duplicates\n\
1050 Do not create duplicate tag entries, for compatibility with\n\
1051 traditional ctags.");
1053 if (PRINT_UNDOCUMENTED_OPTIONS_HELP
)
1054 puts ("-w, --no-warn\n\
1055 Suppress warning messages about duplicate tag entries.");
1057 puts ("-x, --cxref\n\
1058 Like --vgrind, but in the style of cxref, rather than vgrind.\n\
1059 The output uses line numbers instead of page numbers, but\n\
1060 beyond that the differences are cosmetic; try both to see\n\
1064 puts ("-V, --version\n\
1065 Print the version of the program.\n\
1067 Print this help message.\n\
1068 Followed by one or more `--language' options prints detailed\n\
1069 help about tag generation for the specified languages.");
1071 print_language_names ();
1074 puts ("Report bugs to bug-gnu-emacs@gnu.org");
1076 exit (EXIT_SUCCESS
);
1081 main (int argc
, char **argv
)
1084 unsigned int nincluded_files
;
1085 char **included_files
;
1086 argument
*argbuffer
;
1087 int current_arg
, file_count
;
1088 linebuffer filename_lb
;
1089 bool help_asked
= FALSE
;
1095 _fmode
= O_BINARY
; /* all of files are treated as binary files */
1099 nincluded_files
= 0;
1100 included_files
= xnew (argc
, char *);
1104 /* Allocate enough no matter what happens. Overkill, but each one
1106 argbuffer
= xnew (argc
, argument
);
1109 * Always find typedefs and structure tags.
1110 * Also default to find macro constants, enum constants, struct
1111 * members and global variables. Do it for both etags and ctags.
1113 typedefs
= typedefs_or_cplusplus
= constantypedefs
= TRUE
;
1114 globals
= members
= TRUE
;
1116 /* When the optstring begins with a '-' getopt_long does not rearrange the
1117 non-options arguments to be at the end, but leaves them alone. */
1118 optstring
= concat (NO_LONG_OPTIONS
? "" : "-",
1119 "ac:Cf:Il:o:r:RSVhH",
1120 (CTAGS
) ? "BxdtTuvw" : "Di:");
1122 while ((opt
= getopt_long (argc
, argv
, optstring
, longopts
, NULL
)) != EOF
)
1126 /* If getopt returns 0, then it has already processed a
1127 long-named option. We should do nothing. */
1131 /* This means that a file name has been seen. Record it. */
1132 argbuffer
[current_arg
].arg_type
= at_filename
;
1133 argbuffer
[current_arg
].what
= optarg
;
1139 /* Parse standard input. Idea by Vivek <vivek@etla.org>. */
1140 argbuffer
[current_arg
].arg_type
= at_stdin
;
1141 argbuffer
[current_arg
].what
= optarg
;
1145 fatal ("cannot parse standard input more than once", (char *)NULL
);
1146 parsing_stdin
= TRUE
;
1149 /* Common options. */
1150 case 'a': append_to_tagfile
= TRUE
; break;
1151 case 'C': cplusplus
= TRUE
; break;
1152 case 'f': /* for compatibility with old makefiles */
1156 error ("-o option may only be given once.", (char *)NULL
);
1157 suggest_asking_for_help ();
1163 case 'S': /* for backward compatibility */
1164 ignoreindent
= TRUE
;
1168 language
*lang
= get_language_from_langname (optarg
);
1171 argbuffer
[current_arg
].lang
= lang
;
1172 argbuffer
[current_arg
].arg_type
= at_language
;
1178 /* Backward compatibility: support obsolete --ignore-case-regexp. */
1179 optarg
= concat (optarg
, "i", ""); /* memory leak here */
1182 argbuffer
[current_arg
].arg_type
= at_regexp
;
1183 argbuffer
[current_arg
].what
= optarg
;
1187 argbuffer
[current_arg
].arg_type
= at_regexp
;
1188 argbuffer
[current_arg
].what
= NULL
;
1200 case 'D': constantypedefs
= FALSE
; break;
1201 case 'i': included_files
[nincluded_files
++] = optarg
; break;
1203 /* Ctags options. */
1204 case 'B': searchar
= '?'; break;
1205 case 'd': constantypedefs
= TRUE
; break;
1206 case 't': typedefs
= TRUE
; break;
1207 case 'T': typedefs
= typedefs_or_cplusplus
= TRUE
; break;
1208 case 'u': update
= TRUE
; break;
1209 case 'v': vgrind_style
= TRUE
; /*FALLTHRU*/
1210 case 'x': cxref_style
= TRUE
; break;
1211 case 'w': no_warnings
= TRUE
; break;
1213 suggest_asking_for_help ();
1217 /* No more options. Store the rest of arguments. */
1218 for (; optind
< argc
; optind
++)
1220 argbuffer
[current_arg
].arg_type
= at_filename
;
1221 argbuffer
[current_arg
].what
= argv
[optind
];
1226 argbuffer
[current_arg
].arg_type
= at_end
;
1229 print_help (argbuffer
);
1232 if (nincluded_files
== 0 && file_count
== 0)
1234 error ("no input files specified.", (char *)NULL
);
1235 suggest_asking_for_help ();
1239 if (tagfile
== NULL
)
1240 tagfile
= savestr (CTAGS
? "tags" : "TAGS");
1241 cwd
= etags_getcwd (); /* the current working directory */
1242 if (cwd
[strlen (cwd
) - 1] != '/')
1245 cwd
= concat (oldcwd
, "/", "");
1249 /* Compute base directory for relative file names. */
1250 if (streq (tagfile
, "-")
1251 || strneq (tagfile
, "/dev/", 5))
1252 tagfiledir
= cwd
; /* relative file names are relative to cwd */
1255 canonicalize_filename (tagfile
);
1256 tagfiledir
= absolute_dirname (tagfile
, cwd
);
1259 init (); /* set up boolean "functions" */
1261 linebuffer_init (&lb
);
1262 linebuffer_init (&filename_lb
);
1263 linebuffer_init (&filebuf
);
1264 linebuffer_init (&token_name
);
1268 if (streq (tagfile
, "-"))
1272 /* Switch redirected `stdout' to binary mode (setting `_fmode'
1273 doesn't take effect until after `stdout' is already open). */
1274 if (!isatty (fileno (stdout
)))
1275 setmode (fileno (stdout
), O_BINARY
);
1279 tagf
= fopen (tagfile
, append_to_tagfile
? "a" : "w");
1285 * Loop through files finding functions.
1287 for (i
= 0; i
< current_arg
; i
++)
1289 static language
*lang
; /* non-NULL if language is forced */
1292 switch (argbuffer
[i
].arg_type
)
1295 lang
= argbuffer
[i
].lang
;
1298 analyse_regex (argbuffer
[i
].what
);
1301 this_file
= argbuffer
[i
].what
;
1302 /* Input file named "-" means read file names from stdin
1303 (one per line) and use them. */
1304 if (streq (this_file
, "-"))
1307 fatal ("cannot parse standard input AND read file names from it",
1309 while (readline_internal (&filename_lb
, stdin
) > 0)
1310 process_file_name (filename_lb
.buffer
, lang
);
1313 process_file_name (this_file
, lang
);
1316 this_file
= argbuffer
[i
].what
;
1317 process_file (stdin
, this_file
, lang
);
1324 free (filebuf
.buffer
);
1325 free (token_name
.buffer
);
1327 if (!CTAGS
|| cxref_style
)
1329 /* Write the remaining tags to tagf (ETAGS) or stdout (CXREF). */
1330 put_entries (nodehead
);
1331 free_tree (nodehead
);
1337 /* Output file entries that have no tags. */
1338 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
1340 fprintf (tagf
, "\f\n%s,0\n", fdp
->taggedfname
);
1342 while (nincluded_files
-- > 0)
1343 fprintf (tagf
, "\f\n%s,include\n", *included_files
++);
1345 if (fclose (tagf
) == EOF
)
1349 exit (EXIT_SUCCESS
);
1352 /* From here on, we are in (CTAGS && !cxref_style) */
1356 for (i
= 0; i
< current_arg
; ++i
)
1358 switch (argbuffer
[i
].arg_type
)
1364 continue; /* the for loop */
1367 "mv %s OTAGS;fgrep -v '\t%s\t' OTAGS >%s;rm OTAGS",
1368 tagfile
, argbuffer
[i
].what
, tagfile
);
1369 if (system (cmd
) != EXIT_SUCCESS
)
1370 fatal ("failed to execute shell command", (char *)NULL
);
1372 append_to_tagfile
= TRUE
;
1375 tagf
= fopen (tagfile
, append_to_tagfile
? "a" : "w");
1378 put_entries (nodehead
); /* write all the tags (CTAGS) */
1379 free_tree (nodehead
);
1381 if (fclose (tagf
) == EOF
)
1385 if (append_to_tagfile
|| update
)
1387 char cmd
[2*BUFSIZ
+20];
1388 /* Maybe these should be used:
1389 setenv ("LC_COLLATE", "C", 1);
1390 setenv ("LC_ALL", "C", 1); */
1391 sprintf (cmd
, "sort -u -o %.*s %.*s", BUFSIZ
, tagfile
, BUFSIZ
, tagfile
);
1392 exit (system (cmd
));
1394 return EXIT_SUCCESS
;
1399 * Return a compressor given the file name. If EXTPTR is non-zero,
1400 * return a pointer into FILE where the compressor-specific
1401 * extension begins. If no compressor is found, NULL is returned
1402 * and EXTPTR is not significant.
1403 * Idea by Vladimir Alexiev <vladimir@cs.ualberta.ca> (1998)
1406 get_compressor_from_suffix (char *file
, char **extptr
)
1409 char *slash
, *suffix
;
1411 /* File has been processed by canonicalize_filename,
1412 so we don't need to consider backslashes on DOS_NT. */
1413 slash
= etags_strrchr (file
, '/');
1414 suffix
= etags_strrchr (file
, '.');
1415 if (suffix
== NULL
|| suffix
< slash
)
1420 /* Let those poor souls who live with DOS 8+3 file name limits get
1421 some solace by treating foo.cgz as if it were foo.c.gz, etc.
1422 Only the first do loop is run if not MSDOS */
1425 for (compr
= compressors
; compr
->suffix
!= NULL
; compr
++)
1426 if (streq (compr
->suffix
, suffix
))
1429 break; /* do it only once: not really a loop */
1432 } while (*suffix
!= '\0');
1439 * Return a language given the name.
1442 get_language_from_langname (const char *name
)
1447 error ("empty language name", (char *)NULL
);
1450 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1451 if (streq (name
, lang
->name
))
1453 error ("unknown language \"%s\"", name
);
1461 * Return a language given the interpreter name.
1464 get_language_from_interpreter (char *interpreter
)
1469 if (interpreter
== NULL
)
1471 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1472 if (lang
->interpreters
!= NULL
)
1473 for (iname
= lang
->interpreters
; *iname
!= NULL
; iname
++)
1474 if (streq (*iname
, interpreter
))
1483 * Return a language given the file name.
1486 get_language_from_filename (char *file
, int case_sensitive
)
1489 const char **name
, **ext
, *suffix
;
1491 /* Try whole file name first. */
1492 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1493 if (lang
->filenames
!= NULL
)
1494 for (name
= lang
->filenames
; *name
!= NULL
; name
++)
1495 if ((case_sensitive
)
1496 ? streq (*name
, file
)
1497 : strcaseeq (*name
, file
))
1500 /* If not found, try suffix after last dot. */
1501 suffix
= etags_strrchr (file
, '.');
1505 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1506 if (lang
->suffixes
!= NULL
)
1507 for (ext
= lang
->suffixes
; *ext
!= NULL
; ext
++)
1508 if ((case_sensitive
)
1509 ? streq (*ext
, suffix
)
1510 : strcaseeq (*ext
, suffix
))
1517 * This routine is called on each file argument.
1520 process_file_name (char *file
, language
*lang
)
1522 struct stat stat_buf
;
1526 char *compressed_name
, *uncompressed_name
;
1527 char *ext
, *real_name
;
1530 canonicalize_filename (file
);
1531 if (streq (file
, tagfile
) && !streq (tagfile
, "-"))
1533 error ("skipping inclusion of %s in self.", file
);
1536 if ((compr
= get_compressor_from_suffix (file
, &ext
)) == NULL
)
1538 compressed_name
= NULL
;
1539 real_name
= uncompressed_name
= savestr (file
);
1543 real_name
= compressed_name
= savestr (file
);
1544 uncompressed_name
= savenstr (file
, ext
- file
);
1547 /* If the canonicalized uncompressed name
1548 has already been dealt with, skip it silently. */
1549 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
1551 assert (fdp
->infname
!= NULL
);
1552 if (streq (uncompressed_name
, fdp
->infname
))
1556 if (stat (real_name
, &stat_buf
) != 0)
1558 /* Reset real_name and try with a different name. */
1560 if (compressed_name
!= NULL
) /* try with the given suffix */
1562 if (stat (uncompressed_name
, &stat_buf
) == 0)
1563 real_name
= uncompressed_name
;
1565 else /* try all possible suffixes */
1567 for (compr
= compressors
; compr
->suffix
!= NULL
; compr
++)
1569 compressed_name
= concat (file
, ".", compr
->suffix
);
1570 if (stat (compressed_name
, &stat_buf
) != 0)
1574 char *suf
= compressed_name
+ strlen (file
);
1575 size_t suflen
= strlen (compr
->suffix
) + 1;
1576 for ( ; suf
[1]; suf
++, suflen
--)
1578 memmove (suf
, suf
+ 1, suflen
);
1579 if (stat (compressed_name
, &stat_buf
) == 0)
1581 real_name
= compressed_name
;
1585 if (real_name
!= NULL
)
1588 free (compressed_name
);
1589 compressed_name
= NULL
;
1593 real_name
= compressed_name
;
1598 if (real_name
== NULL
)
1603 } /* try with a different name */
1605 if (!S_ISREG (stat_buf
.st_mode
))
1607 error ("skipping %s: it is not a regular file.", real_name
);
1610 if (real_name
== compressed_name
)
1612 char *cmd
= concat (compr
->command
, " ", real_name
);
1613 inf
= (FILE *) popen (cmd
, "r");
1617 inf
= fopen (real_name
, "r");
1624 process_file (inf
, uncompressed_name
, lang
);
1626 if (real_name
== compressed_name
)
1627 retval
= pclose (inf
);
1629 retval
= fclose (inf
);
1634 free (compressed_name
);
1635 free (uncompressed_name
);
1642 process_file (FILE *fh
, char *fn
, language
*lang
)
1644 static const fdesc emptyfdesc
;
1647 /* Create a new input file description entry. */
1648 fdp
= xnew (1, fdesc
);
1651 fdp
->infname
= savestr (fn
);
1653 fdp
->infabsname
= absolute_filename (fn
, cwd
);
1654 fdp
->infabsdir
= absolute_dirname (fn
, cwd
);
1655 if (filename_is_absolute (fn
))
1657 /* An absolute file name. Canonicalize it. */
1658 fdp
->taggedfname
= absolute_filename (fn
, NULL
);
1662 /* A file name relative to cwd. Make it relative
1663 to the directory of the tags file. */
1664 fdp
->taggedfname
= relative_filename (fn
, tagfiledir
);
1666 fdp
->usecharno
= TRUE
; /* use char position when making tags */
1668 fdp
->written
= FALSE
; /* not written on tags file yet */
1671 curfdp
= fdhead
; /* the current file description */
1675 /* If not Ctags, and if this is not metasource and if it contained no #line
1676 directives, we can write the tags and free all nodes pointing to
1679 && curfdp
->usecharno
/* no #line directives in this file */
1680 && !curfdp
->lang
->metasource
)
1684 /* Look for the head of the sublist relative to this file. See add_node
1685 for the structure of the node tree. */
1687 for (np
= nodehead
; np
!= NULL
; prev
= np
, np
= np
->left
)
1688 if (np
->fdp
== curfdp
)
1691 /* If we generated tags for this file, write and delete them. */
1694 /* This is the head of the last sublist, if any. The following
1695 instructions depend on this being true. */
1696 assert (np
->left
== NULL
);
1698 assert (fdhead
== curfdp
);
1699 assert (last_node
->fdp
== curfdp
);
1700 put_entries (np
); /* write tags for file curfdp->taggedfname */
1701 free_tree (np
); /* remove the written nodes */
1703 nodehead
= NULL
; /* no nodes left */
1705 prev
->left
= NULL
; /* delete the pointer to the sublist */
1711 * This routine sets up the boolean pseudo-functions which work
1712 * by setting boolean flags dependent upon the corresponding character.
1713 * Every char which is NOT in that string is not a white char. Therefore,
1714 * all of the array "_wht" is set to FALSE, and then the elements
1715 * subscripted by the chars in "white" are set to TRUE. Thus "_wht"
1716 * of a char is TRUE if it is the string "white", else FALSE.
1721 register const char *sp
;
1724 for (i
= 0; i
< CHARS
; i
++)
1725 iswhite(i
) = notinname(i
) = begtoken(i
) = intoken(i
) = endtoken(i
) = FALSE
;
1726 for (sp
= white
; *sp
!= '\0'; sp
++) iswhite (*sp
) = TRUE
;
1727 for (sp
= nonam
; *sp
!= '\0'; sp
++) notinname (*sp
) = TRUE
;
1728 notinname('\0') = notinname('\n');
1729 for (sp
= begtk
; *sp
!= '\0'; sp
++) begtoken (*sp
) = TRUE
;
1730 begtoken('\0') = begtoken('\n');
1731 for (sp
= midtk
; *sp
!= '\0'; sp
++) intoken (*sp
) = TRUE
;
1732 intoken('\0') = intoken('\n');
1733 for (sp
= endtk
; *sp
!= '\0'; sp
++) endtoken (*sp
) = TRUE
;
1734 endtoken('\0') = endtoken('\n');
1738 * This routine opens the specified file and calls the function
1739 * which finds the function and type definitions.
1742 find_entries (FILE *inf
)
1745 language
*lang
= curfdp
->lang
;
1746 Lang_function
*parser
= NULL
;
1748 /* If user specified a language, use it. */
1749 if (lang
!= NULL
&& lang
->function
!= NULL
)
1751 parser
= lang
->function
;
1754 /* Else try to guess the language given the file name. */
1757 lang
= get_language_from_filename (curfdp
->infname
, TRUE
);
1758 if (lang
!= NULL
&& lang
->function
!= NULL
)
1760 curfdp
->lang
= lang
;
1761 parser
= lang
->function
;
1765 /* Else look for sharp-bang as the first two characters. */
1767 && readline_internal (&lb
, inf
) > 0
1769 && lb
.buffer
[0] == '#'
1770 && lb
.buffer
[1] == '!')
1774 /* Set lp to point at the first char after the last slash in the
1775 line or, if no slashes, at the first nonblank. Then set cp to
1776 the first successive blank and terminate the string. */
1777 lp
= etags_strrchr (lb
.buffer
+2, '/');
1781 lp
= skip_spaces (lb
.buffer
+ 2);
1782 cp
= skip_non_spaces (lp
);
1785 if (strlen (lp
) > 0)
1787 lang
= get_language_from_interpreter (lp
);
1788 if (lang
!= NULL
&& lang
->function
!= NULL
)
1790 curfdp
->lang
= lang
;
1791 parser
= lang
->function
;
1796 /* We rewind here, even if inf may be a pipe. We fail if the
1797 length of the first line is longer than the pipe block size,
1798 which is unlikely. */
1801 /* Else try to guess the language given the case insensitive file name. */
1804 lang
= get_language_from_filename (curfdp
->infname
, FALSE
);
1805 if (lang
!= NULL
&& lang
->function
!= NULL
)
1807 curfdp
->lang
= lang
;
1808 parser
= lang
->function
;
1812 /* Else try Fortran or C. */
1815 node
*old_last_node
= last_node
;
1817 curfdp
->lang
= get_language_from_langname ("fortran");
1820 if (old_last_node
== last_node
)
1821 /* No Fortran entries found. Try C. */
1823 /* We do not tag if rewind fails.
1824 Only the file name will be recorded in the tags file. */
1826 curfdp
->lang
= get_language_from_langname (cplusplus
? "c++" : "c");
1832 if (!no_line_directive
1833 && curfdp
->lang
!= NULL
&& curfdp
->lang
->metasource
)
1834 /* It may be that this is a bingo.y file, and we already parsed a bingo.c
1835 file, or anyway we parsed a file that is automatically generated from
1836 this one. If this is the case, the bingo.c file contained #line
1837 directives that generated tags pointing to this file. Let's delete
1838 them all before parsing this file, which is the real source. */
1840 fdesc
**fdpp
= &fdhead
;
1841 while (*fdpp
!= NULL
)
1843 && streq ((*fdpp
)->taggedfname
, curfdp
->taggedfname
))
1844 /* We found one of those! We must delete both the file description
1845 and all tags referring to it. */
1847 fdesc
*badfdp
= *fdpp
;
1849 /* Delete the tags referring to badfdp->taggedfname
1850 that were obtained from badfdp->infname. */
1851 invalidate_nodes (badfdp
, &nodehead
);
1853 *fdpp
= badfdp
->next
; /* remove the bad description from the list */
1854 free_fdesc (badfdp
);
1857 fdpp
= &(*fdpp
)->next
; /* advance the list pointer */
1860 assert (parser
!= NULL
);
1862 /* Generic initialisations before reading from file. */
1863 linebuffer_setlen (&filebuf
, 0); /* reset the file buffer */
1865 /* Generic initialisations before parsing file with readline. */
1866 lineno
= 0; /* reset global line number */
1867 charno
= 0; /* reset global char number */
1868 linecharno
= 0; /* reset global char number of line start */
1872 regex_tag_multiline ();
1877 * Check whether an implicitly named tag should be created,
1878 * then call `pfnote'.
1879 * NAME is a string that is internally copied by this function.
1881 * TAGS format specification
1882 * Idea by Sam Kendall <kendall@mv.mv.com> (1997)
1883 * The following is explained in some more detail in etc/ETAGS.EBNF.
1885 * make_tag creates tags with "implicit tag names" (unnamed tags)
1886 * if the following are all true, assuming NONAM=" \f\t\n\r()=,;":
1887 * 1. NAME does not contain any of the characters in NONAM;
1888 * 2. LINESTART contains name as either a rightmost, or rightmost but
1889 * one character, substring;
1890 * 3. the character, if any, immediately before NAME in LINESTART must
1891 * be a character in NONAM;
1892 * 4. the character, if any, immediately after NAME in LINESTART must
1893 * also be a character in NONAM.
1895 * The implementation uses the notinname() macro, which recognises the
1896 * characters stored in the string `nonam'.
1897 * etags.el needs to use the same characters that are in NONAM.
1900 make_tag (const char *name
, /* tag name, or NULL if unnamed */
1901 int namelen
, /* tag length */
1902 int is_func
, /* tag is a function */
1903 char *linestart
, /* start of the line where tag is */
1904 int linelen
, /* length of the line where tag is */
1905 int lno
, /* line number */
1906 long int cno
) /* character number */
1908 bool named
= (name
!= NULL
&& namelen
> 0);
1911 if (!CTAGS
&& named
) /* maybe set named to false */
1912 /* Let's try to make an implicit tag name, that is, create an unnamed tag
1913 such that etags.el can guess a name from it. */
1916 register const char *cp
= name
;
1918 for (i
= 0; i
< namelen
; i
++)
1919 if (notinname (*cp
++))
1921 if (i
== namelen
) /* rule #1 */
1923 cp
= linestart
+ linelen
- namelen
;
1924 if (notinname (linestart
[linelen
-1]))
1925 cp
-= 1; /* rule #4 */
1926 if (cp
>= linestart
/* rule #2 */
1928 || notinname (cp
[-1])) /* rule #3 */
1929 && strneq (name
, cp
, namelen
)) /* rule #2 */
1930 named
= FALSE
; /* use implicit tag name */
1935 nname
= savenstr (name
, namelen
);
1937 pfnote (nname
, is_func
, linestart
, linelen
, lno
, cno
);
1942 pfnote (char *name
, int is_func
, char *linestart
, int linelen
, int lno
, long int cno
)
1943 /* tag name, or NULL if unnamed */
1944 /* tag is a function */
1945 /* start of the line where tag is */
1946 /* length of the line where tag is */
1948 /* character number */
1952 assert (name
== NULL
|| name
[0] != '\0');
1953 if (CTAGS
&& name
== NULL
)
1956 np
= xnew (1, node
);
1958 /* If ctags mode, change name "main" to M<thisfilename>. */
1959 if (CTAGS
&& !cxref_style
&& streq (name
, "main"))
1961 register char *fp
= etags_strrchr (curfdp
->taggedfname
, '/');
1962 np
->name
= concat ("M", fp
== NULL
? curfdp
->taggedfname
: fp
+ 1, "");
1963 fp
= etags_strrchr (np
->name
, '.');
1964 if (fp
!= NULL
&& fp
[1] != '\0' && fp
[2] == '\0')
1970 np
->been_warned
= FALSE
;
1972 np
->is_func
= is_func
;
1974 if (np
->fdp
->usecharno
)
1975 /* Our char numbers are 0-base, because of C language tradition?
1976 ctags compatibility? old versions compatibility? I don't know.
1977 Anyway, since emacs's are 1-base we expect etags.el to take care
1978 of the difference. If we wanted to have 1-based numbers, we would
1979 uncomment the +1 below. */
1980 np
->cno
= cno
/* + 1 */ ;
1982 np
->cno
= invalidcharno
;
1983 np
->left
= np
->right
= NULL
;
1984 if (CTAGS
&& !cxref_style
)
1986 if (strlen (linestart
) < 50)
1987 np
->regex
= concat (linestart
, "$", "");
1989 np
->regex
= savenstr (linestart
, 50);
1992 np
->regex
= savenstr (linestart
, linelen
);
1994 add_node (np
, &nodehead
);
1999 * recurse on left children, iterate on right children.
2002 free_tree (register node
*np
)
2006 register node
*node_right
= np
->right
;
2007 free_tree (np
->left
);
2017 * delete a file description
2020 free_fdesc (register fdesc
*fdp
)
2022 free (fdp
->infname
);
2023 free (fdp
->infabsname
);
2024 free (fdp
->infabsdir
);
2025 free (fdp
->taggedfname
);
2032 * Adds a node to the tree of nodes. In etags mode, sort by file
2033 * name. In ctags mode, sort by tag name. Make no attempt at
2036 * add_node is the only function allowed to add nodes, so it can
2040 add_node (node
*np
, node
**cur_node_p
)
2043 register node
*cur_node
= *cur_node_p
;
2045 if (cur_node
== NULL
)
2055 /* For each file name, tags are in a linked sublist on the right
2056 pointer. The first tags of different files are a linked list
2057 on the left pointer. last_node points to the end of the last
2059 if (last_node
!= NULL
&& last_node
->fdp
== np
->fdp
)
2061 /* Let's use the same sublist as the last added node. */
2062 assert (last_node
->right
== NULL
);
2063 last_node
->right
= np
;
2066 else if (cur_node
->fdp
== np
->fdp
)
2068 /* Scanning the list we found the head of a sublist which is
2069 good for us. Let's scan this sublist. */
2070 add_node (np
, &cur_node
->right
);
2073 /* The head of this sublist is not good for us. Let's try the
2075 add_node (np
, &cur_node
->left
);
2076 } /* if ETAGS mode */
2081 dif
= strcmp (np
->name
, cur_node
->name
);
2084 * If this tag name matches an existing one, then
2085 * do not add the node, but maybe print a warning.
2087 if (no_duplicates
&& !dif
)
2089 if (np
->fdp
== cur_node
->fdp
)
2093 fprintf (stderr
, "Duplicate entry in file %s, line %d: %s\n",
2094 np
->fdp
->infname
, lineno
, np
->name
);
2095 fprintf (stderr
, "Second entry ignored\n");
2098 else if (!cur_node
->been_warned
&& !no_warnings
)
2102 "Duplicate entry in files %s and %s: %s (Warning only)\n",
2103 np
->fdp
->infname
, cur_node
->fdp
->infname
, np
->name
);
2104 cur_node
->been_warned
= TRUE
;
2109 /* Actually add the node */
2110 add_node (np
, dif
< 0 ? &cur_node
->left
: &cur_node
->right
);
2111 } /* if CTAGS mode */
2115 * invalidate_nodes ()
2116 * Scan the node tree and invalidate all nodes pointing to the
2117 * given file description (CTAGS case) or free them (ETAGS case).
2120 invalidate_nodes (fdesc
*badfdp
, node
**npp
)
2129 if (np
->left
!= NULL
)
2130 invalidate_nodes (badfdp
, &np
->left
);
2131 if (np
->fdp
== badfdp
)
2133 if (np
->right
!= NULL
)
2134 invalidate_nodes (badfdp
, &np
->right
);
2138 assert (np
->fdp
!= NULL
);
2139 if (np
->fdp
== badfdp
)
2141 *npp
= np
->left
; /* detach the sublist from the list */
2142 np
->left
= NULL
; /* isolate it */
2143 free_tree (np
); /* free it */
2144 invalidate_nodes (badfdp
, npp
);
2147 invalidate_nodes (badfdp
, &np
->left
);
2152 static int total_size_of_entries (node
*);
2153 static int number_len (long);
2155 /* Length of a non-negative number's decimal representation. */
2157 number_len (long int num
)
2160 while ((num
/= 10) > 0)
2166 * Return total number of characters that put_entries will output for
2167 * the nodes in the linked list at the right of the specified node.
2168 * This count is irrelevant with etags.el since emacs 19.34 at least,
2169 * but is still supplied for backward compatibility.
2172 total_size_of_entries (register node
*np
)
2174 register int total
= 0;
2176 for (; np
!= NULL
; np
= np
->right
)
2179 total
+= strlen (np
->regex
) + 1; /* pat\177 */
2180 if (np
->name
!= NULL
)
2181 total
+= strlen (np
->name
) + 1; /* name\001 */
2182 total
+= number_len ((long) np
->lno
) + 1; /* lno, */
2183 if (np
->cno
!= invalidcharno
) /* cno */
2184 total
+= number_len (np
->cno
);
2185 total
+= 1; /* newline */
2192 put_entries (register node
*np
)
2195 static fdesc
*fdp
= NULL
;
2200 /* Output subentries that precede this one */
2202 put_entries (np
->left
);
2204 /* Output this entry */
2213 fprintf (tagf
, "\f\n%s,%d\n",
2214 fdp
->taggedfname
, total_size_of_entries (np
));
2215 fdp
->written
= TRUE
;
2217 fputs (np
->regex
, tagf
);
2218 fputc ('\177', tagf
);
2219 if (np
->name
!= NULL
)
2221 fputs (np
->name
, tagf
);
2222 fputc ('\001', tagf
);
2224 fprintf (tagf
, "%d,", np
->lno
);
2225 if (np
->cno
!= invalidcharno
)
2226 fprintf (tagf
, "%ld", np
->cno
);
2232 if (np
->name
== NULL
)
2233 error ("internal error: NULL name in ctags mode.", (char *)NULL
);
2238 fprintf (stdout
, "%s %s %d\n",
2239 np
->name
, np
->fdp
->taggedfname
, (np
->lno
+ 63) / 64);
2241 fprintf (stdout
, "%-16s %3d %-16s %s\n",
2242 np
->name
, np
->lno
, np
->fdp
->taggedfname
, np
->regex
);
2246 fprintf (tagf
, "%s\t%s\t", np
->name
, np
->fdp
->taggedfname
);
2249 { /* function or #define macro with args */
2250 putc (searchar
, tagf
);
2253 for (sp
= np
->regex
; *sp
; sp
++)
2255 if (*sp
== '\\' || *sp
== searchar
)
2259 putc (searchar
, tagf
);
2262 { /* anything else; text pattern inadequate */
2263 fprintf (tagf
, "%d", np
->lno
);
2268 } /* if this node contains a valid tag */
2270 /* Output subentries that follow this one */
2271 put_entries (np
->right
);
2273 put_entries (np
->left
);
2278 #define C_EXT 0x00fff /* C extensions */
2279 #define C_PLAIN 0x00000 /* C */
2280 #define C_PLPL 0x00001 /* C++ */
2281 #define C_STAR 0x00003 /* C* */
2282 #define C_JAVA 0x00005 /* JAVA */
2283 #define C_AUTO 0x01000 /* C, but switch to C++ if `class' is met */
2284 #define YACC 0x10000 /* yacc file */
2287 * The C symbol tables.
2292 st_C_objprot
, st_C_objimpl
, st_C_objend
,
2294 st_C_ignore
, st_C_attribute
,
2297 st_C_class
, st_C_template
,
2298 st_C_struct
, st_C_extern
, st_C_enum
, st_C_define
, st_C_typedef
2301 static unsigned int hash (const char *, unsigned int);
2302 static struct C_stab_entry
* in_word_set (const char *, unsigned int);
2303 static enum sym_type
C_symtype (char *, int, int);
2305 /* Feed stuff between (but not including) %[ and %] lines to:
2311 struct C_stab_entry { char *name; int c_ext; enum sym_type type; }
2315 while, 0, st_C_ignore
2316 switch, 0, st_C_ignore
2317 return, 0, st_C_ignore
2318 __attribute__, 0, st_C_attribute
2319 GTY, 0, st_C_attribute
2320 @interface, 0, st_C_objprot
2321 @protocol, 0, st_C_objprot
2322 @implementation,0, st_C_objimpl
2323 @end, 0, st_C_objend
2324 import, (C_JAVA & ~C_PLPL), st_C_ignore
2325 package, (C_JAVA & ~C_PLPL), st_C_ignore
2326 friend, C_PLPL, st_C_ignore
2327 extends, (C_JAVA & ~C_PLPL), st_C_javastruct
2328 implements, (C_JAVA & ~C_PLPL), st_C_javastruct
2329 interface, (C_JAVA & ~C_PLPL), st_C_struct
2330 class, 0, st_C_class
2331 namespace, C_PLPL, st_C_struct
2332 domain, C_STAR, st_C_struct
2333 union, 0, st_C_struct
2334 struct, 0, st_C_struct
2335 extern, 0, st_C_extern
2337 typedef, 0, st_C_typedef
2338 define, 0, st_C_define
2339 undef, 0, st_C_define
2340 operator, C_PLPL, st_C_operator
2341 template, 0, st_C_template
2342 # DEFUN used in emacs, the next three used in glibc (SYSCALL only for mach).
2343 DEFUN, 0, st_C_gnumacro
2344 SYSCALL, 0, st_C_gnumacro
2345 ENTRY, 0, st_C_gnumacro
2346 PSEUDO, 0, st_C_gnumacro
2347 # These are defined inside C functions, so currently they are not met.
2348 # EXFUN used in glibc, DEFVAR_* in emacs.
2349 #EXFUN, 0, st_C_gnumacro
2350 #DEFVAR_, 0, st_C_gnumacro
2352 and replace lines between %< and %> with its output, then:
2353 - remove the #if characterset check
2354 - make in_word_set static and not inline. */
2356 /* C code produced by gperf version 3.0.1 */
2357 /* Command-line: gperf -m 5 */
2358 /* Computed positions: -k'2-3' */
2360 struct C_stab_entry
{ const char *name
; int c_ext
; enum sym_type type
; };
2361 /* maximum key range = 33, duplicates = 0 */
2371 hash (register const char *str
, register unsigned int len
)
2373 static unsigned char asso_values
[] =
2375 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2376 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2377 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2378 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2379 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2380 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2381 35, 35, 35, 35, 35, 35, 35, 35, 35, 3,
2382 26, 35, 35, 35, 35, 35, 35, 35, 27, 35,
2383 35, 35, 35, 24, 0, 35, 35, 35, 35, 0,
2384 35, 35, 35, 35, 35, 1, 35, 16, 35, 6,
2385 23, 0, 0, 35, 22, 0, 35, 35, 5, 0,
2386 0, 15, 1, 35, 6, 35, 8, 19, 35, 16,
2387 4, 5, 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, 35, 35, 35, 35,
2396 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2397 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2398 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2399 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2400 35, 35, 35, 35, 35, 35
2402 register int hval
= len
;
2407 hval
+= asso_values
[(unsigned char)str
[2]];
2410 hval
+= asso_values
[(unsigned char)str
[1]];
2416 static struct C_stab_entry
*
2417 in_word_set (register const char *str
, register unsigned int len
)
2421 TOTAL_KEYWORDS
= 33,
2422 MIN_WORD_LENGTH
= 2,
2423 MAX_WORD_LENGTH
= 15,
2428 static struct C_stab_entry wordlist
[] =
2431 {"if", 0, st_C_ignore
},
2432 {"GTY", 0, st_C_attribute
},
2433 {"@end", 0, st_C_objend
},
2434 {"union", 0, st_C_struct
},
2435 {"define", 0, st_C_define
},
2436 {"import", (C_JAVA
& ~C_PLPL
), st_C_ignore
},
2437 {"template", 0, st_C_template
},
2438 {"operator", C_PLPL
, st_C_operator
},
2439 {"@interface", 0, st_C_objprot
},
2440 {"implements", (C_JAVA
& ~C_PLPL
), st_C_javastruct
},
2441 {"friend", C_PLPL
, st_C_ignore
},
2442 {"typedef", 0, st_C_typedef
},
2443 {"return", 0, st_C_ignore
},
2444 {"@implementation",0, st_C_objimpl
},
2445 {"@protocol", 0, st_C_objprot
},
2446 {"interface", (C_JAVA
& ~C_PLPL
), st_C_struct
},
2447 {"extern", 0, st_C_extern
},
2448 {"extends", (C_JAVA
& ~C_PLPL
), st_C_javastruct
},
2449 {"struct", 0, st_C_struct
},
2450 {"domain", C_STAR
, st_C_struct
},
2451 {"switch", 0, st_C_ignore
},
2452 {"enum", 0, st_C_enum
},
2453 {"for", 0, st_C_ignore
},
2454 {"namespace", C_PLPL
, st_C_struct
},
2455 {"class", 0, st_C_class
},
2456 {"while", 0, st_C_ignore
},
2457 {"undef", 0, st_C_define
},
2458 {"package", (C_JAVA
& ~C_PLPL
), st_C_ignore
},
2459 {"__attribute__", 0, st_C_attribute
},
2460 {"SYSCALL", 0, st_C_gnumacro
},
2461 {"ENTRY", 0, st_C_gnumacro
},
2462 {"PSEUDO", 0, st_C_gnumacro
},
2463 {"DEFUN", 0, st_C_gnumacro
}
2466 if (len
<= MAX_WORD_LENGTH
&& len
>= MIN_WORD_LENGTH
)
2468 register int key
= hash (str
, len
);
2470 if (key
<= MAX_HASH_VALUE
&& key
>= 0)
2472 register const char *s
= wordlist
[key
].name
;
2474 if (*str
== *s
&& !strncmp (str
+ 1, s
+ 1, len
- 1) && s
[len
] == '\0')
2475 return &wordlist
[key
];
2482 static enum sym_type
2483 C_symtype (char *str
, int len
, int 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 const 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 (int, char *, int);
2611 static void popclass_above (int);
2612 static void write_classname (linebuffer
*, const 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 (int bracelev
, char *str
, int len
)
2631 popclass_above (bracelev
);
2633 if (nl
>= cstack
.size
)
2635 int size
= cstack
.size
*= 2;
2636 xrnew (cstack
.cname
, size
, char *);
2637 xrnew (cstack
.bracelev
, size
, int);
2639 assert (nl
== 0 || cstack
.bracelev
[nl
-1] < bracelev
);
2640 cstack
.cname
[nl
] = (str
== NULL
) ? NULL
: savenstr (str
, len
);
2641 cstack
.bracelev
[nl
] = bracelev
;
2646 popclass_above (int bracelev
)
2650 for (nl
= cstack
.nl
- 1;
2651 nl
>= 0 && cstack
.bracelev
[nl
] >= bracelev
;
2654 free (cstack
.cname
[nl
]);
2660 write_classname (linebuffer
*cn
, const char *qualifier
)
2663 int qlen
= strlen (qualifier
);
2665 if (cstack
.nl
== 0 || cstack
.cname
[0] == NULL
)
2669 cn
->buffer
[0] = '\0';
2673 len
= strlen (cstack
.cname
[0]);
2674 linebuffer_setlen (cn
, len
);
2675 strcpy (cn
->buffer
, cstack
.cname
[0]);
2677 for (i
= 1; i
< cstack
.nl
; i
++)
2682 s
= cstack
.cname
[i
];
2687 linebuffer_setlen (cn
, len
);
2688 strncat (cn
->buffer
, qualifier
, qlen
);
2689 strncat (cn
->buffer
, s
, slen
);
2694 static bool consider_token (char *, int, int, int *, int, int, bool *);
2695 static void make_C_tag (bool);
2699 * checks to see if the current token is at the start of a
2700 * function or variable, or corresponds to a typedef, or
2701 * is a struct/union/enum tag, or #define, or an enum constant.
2703 * *IS_FUNC gets TRUE if the token is a function or #define macro
2704 * with args. C_EXTP points to which language we are looking at.
2715 consider_token (register char *str
, register int len
, register int c
, int *c_extp
, int bracelev
, int parlev
, int *is_func_or_var
)
2716 /* IN: token pointer */
2717 /* IN: token length */
2718 /* IN: first char after the token */
2719 /* IN, OUT: C extensions mask */
2720 /* IN: brace level */
2721 /* IN: parenthesis level */
2722 /* OUT: function or variable found */
2724 /* When structdef is stagseen, scolonseen, or snone with bracelev > 0,
2725 structtype is the type of the preceding struct-like keyword, and
2726 structbracelev is the brace level where it has been seen. */
2727 static enum sym_type structtype
;
2728 static int structbracelev
;
2729 static enum sym_type toktype
;
2732 toktype
= C_symtype (str
, len
, *c_extp
);
2735 * Skip __attribute__
2737 if (toktype
== st_C_attribute
)
2744 * Advance the definedef state machine.
2749 /* We're not on a preprocessor line. */
2750 if (toktype
== st_C_gnumacro
)
2757 if (toktype
== st_C_define
)
2759 definedef
= ddefineseen
;
2763 definedef
= dignorerest
;
2768 * Make a tag for any macro, unless it is a constant
2769 * and constantypedefs is FALSE.
2771 definedef
= dignorerest
;
2772 *is_func_or_var
= (c
== '(');
2773 if (!*is_func_or_var
&& !constantypedefs
)
2780 error ("internal error: definedef value.", (char *)NULL
);
2789 if (toktype
== st_C_typedef
)
2809 if (structdef
== snone
&& fvdef
== fvnone
)
2828 case st_C_javastruct
:
2829 if (structdef
== stagseen
)
2830 structdef
= scolonseen
;
2834 if ((*c_extp
& C_AUTO
) /* automatic detection of C++ language */
2836 && definedef
== dnone
&& structdef
== snone
2837 && typdef
== tnone
&& fvdef
== fvnone
)
2838 *c_extp
= (*c_extp
| C_PLPL
) & ~C_AUTO
;
2839 if (toktype
== st_C_template
)
2846 && (typdef
== tkeyseen
2847 || (typedefs_or_cplusplus
&& structdef
== snone
)))
2849 structdef
= skeyseen
;
2850 structtype
= toktype
;
2851 structbracelev
= bracelev
;
2852 if (fvdef
== fvnameseen
)
2858 if (structdef
== skeyseen
)
2860 structdef
= stagseen
;
2864 if (typdef
!= tnone
)
2867 /* Detect Objective C constructs. */
2877 objdef
= oimplementation
;
2881 case oimplementation
:
2882 /* Save the class tag for functions or variables defined inside. */
2883 objtag
= savenstr (str
, len
);
2887 /* Save the class tag for categories. */
2888 objtag
= savenstr (str
, len
);
2890 *is_func_or_var
= TRUE
;
2894 *is_func_or_var
= TRUE
;
2902 objdef
= omethodtag
;
2903 linebuffer_setlen (&token_name
, len
);
2904 strncpy (token_name
.buffer
, str
, len
);
2905 token_name
.buffer
[len
] = '\0';
2911 objdef
= omethodparm
;
2917 objdef
= omethodtag
;
2918 linebuffer_setlen (&token_name
, token_name
.len
+ len
);
2919 strncat (token_name
.buffer
, str
, len
);
2924 if (toktype
== st_C_objend
)
2926 /* Memory leakage here: the string pointed by objtag is
2927 never released, because many tests would be needed to
2928 avoid breaking on incorrect input code. The amount of
2929 memory leaked here is the sum of the lengths of the
2937 /* A function, variable or enum constant? */
2959 *is_func_or_var
= TRUE
;
2963 && structdef
== snone
2964 && structtype
== st_C_enum
&& bracelev
> structbracelev
)
2965 return TRUE
; /* enum constant */
2971 fvdef
= fdefunname
; /* GNU macro */
2972 *is_func_or_var
= TRUE
;
2980 if ((strneq (str
, "asm", 3) && endtoken (str
[3]))
2981 || (strneq (str
, "__asm__", 7) && endtoken (str
[7])))
2990 if (len
>= 10 && strneq (str
+len
-10, "::operator", 10))
2992 if (*c_extp
& C_AUTO
) /* automatic detection of C++ */
2993 *c_extp
= (*c_extp
| C_PLPL
) & ~C_AUTO
;
2995 *is_func_or_var
= TRUE
;
2998 if (bracelev
> 0 && !instruct
)
3000 fvdef
= fvnameseen
; /* function or variable */
3001 *is_func_or_var
= TRUE
;
3012 * C_entries often keeps pointers to tokens or lines which are older than
3013 * the line currently read. By keeping two line buffers, and switching
3014 * them at end of line, it is possible to use those pointers.
3022 #define current_lb_is_new (newndx == curndx)
3023 #define switch_line_buffers() (curndx = 1 - curndx)
3025 #define curlb (lbs[curndx].lb)
3026 #define newlb (lbs[newndx].lb)
3027 #define curlinepos (lbs[curndx].linepos)
3028 #define newlinepos (lbs[newndx].linepos)
3030 #define plainc ((c_ext & C_EXT) == C_PLAIN)
3031 #define cplpl (c_ext & C_PLPL)
3032 #define cjava ((c_ext & C_JAVA) == C_JAVA)
3034 #define CNL_SAVE_DEFINEDEF() \
3036 curlinepos = charno; \
3037 readline (&curlb, inf); \
3038 lp = curlb.buffer; \
3045 CNL_SAVE_DEFINEDEF(); \
3046 if (savetoken.valid) \
3048 token = savetoken; \
3049 savetoken.valid = FALSE; \
3051 definedef = dnone; \
3056 make_C_tag (int isfun
)
3058 /* This function is never called when token.valid is FALSE, but
3059 we must protect against invalid input or internal errors. */
3061 make_tag (token_name
.buffer
, token_name
.len
, isfun
, token
.line
,
3062 token
.offset
+token
.length
+1, token
.lineno
, token
.linepos
);
3064 { /* this branch is optimised away if !DEBUG */
3065 make_tag (concat ("INVALID TOKEN:-->", token_name
.buffer
, ""),
3066 token_name
.len
+ 17, isfun
, token
.line
,
3067 token
.offset
+token
.length
+1, token
.lineno
, token
.linepos
);
3068 error ("INVALID TOKEN", NULL
);
3071 token
.valid
= FALSE
;
3077 * This routine finds functions, variables, typedefs,
3078 * #define's, enum constants and struct/union/enum definitions in
3079 * C syntax and adds them to the list.
3082 C_entries (int c_ext
, FILE *inf
)
3083 /* extension of C */
3086 register char c
; /* latest char read; '\0' for end of line */
3087 register char *lp
; /* pointer one beyond the character `c' */
3088 int curndx
, newndx
; /* indices for current and new lb */
3089 register int tokoff
; /* offset in line of start of current token */
3090 register int toklen
; /* length of current token */
3091 const char *qualifier
; /* string used to qualify names */
3092 int qlen
; /* length of qualifier */
3093 int bracelev
; /* current brace level */
3094 int bracketlev
; /* current bracket level */
3095 int parlev
; /* current parenthesis level */
3096 int attrparlev
; /* __attribute__ parenthesis level */
3097 int templatelev
; /* current template level */
3098 int typdefbracelev
; /* bracelev where a typedef struct body begun */
3099 bool incomm
, inquote
, inchar
, quotednl
, midtoken
;
3100 bool yacc_rules
; /* in the rules part of a yacc file */
3101 struct tok savetoken
= {0}; /* token saved during preprocessor handling */
3104 linebuffer_init (&lbs
[0].lb
);
3105 linebuffer_init (&lbs
[1].lb
);
3106 if (cstack
.size
== 0)
3108 cstack
.size
= (DEBUG
) ? 1 : 4;
3110 cstack
.cname
= xnew (cstack
.size
, char *);
3111 cstack
.bracelev
= xnew (cstack
.size
, int);
3114 tokoff
= toklen
= typdefbracelev
= 0; /* keep compiler quiet */
3115 curndx
= newndx
= 0;
3119 fvdef
= fvnone
; fvextern
= FALSE
; typdef
= tnone
;
3120 structdef
= snone
; definedef
= dnone
; objdef
= onone
;
3122 midtoken
= inquote
= inchar
= incomm
= quotednl
= FALSE
;
3123 token
.valid
= savetoken
.valid
= FALSE
;
3124 bracelev
= bracketlev
= parlev
= attrparlev
= templatelev
= 0;
3126 { qualifier
= "."; qlen
= 1; }
3128 { qualifier
= "::"; qlen
= 2; }
3136 /* If we are at the end of the line, the next character is a
3137 '\0'; do not skip it, because it is what tells us
3138 to read the next line. */
3159 /* Newlines inside comments do not end macro definitions in
3161 CNL_SAVE_DEFINEDEF ();
3174 /* Newlines inside strings do not end macro definitions
3175 in traditional cpp, even though compilers don't
3176 usually accept them. */
3177 CNL_SAVE_DEFINEDEF ();
3187 /* Hmmm, something went wrong. */
3196 else if (bracketlev
> 0)
3201 if (--bracketlev
> 0)
3205 CNL_SAVE_DEFINEDEF ();
3233 if (fvdef
!= finlist
&& fvdef
!= fignore
&& fvdef
!=vignore
)
3246 else if (/* cplpl && */ *lp
== '/')
3252 if ((c_ext
& YACC
) && *lp
== '%')
3254 /* Entering or exiting rules section in yacc file. */
3256 definedef
= dnone
; fvdef
= fvnone
; fvextern
= FALSE
;
3257 typdef
= tnone
; structdef
= snone
;
3258 midtoken
= inquote
= inchar
= incomm
= quotednl
= FALSE
;
3260 yacc_rules
= !yacc_rules
;
3266 if (definedef
== dnone
)
3269 bool cpptoken
= TRUE
;
3271 /* Look back on this line. If all blanks, or nonblanks
3272 followed by an end of comment, this is a preprocessor
3274 for (cp
= newlb
.buffer
; cp
< lp
-1; cp
++)
3277 if (*cp
== '*' && *(cp
+1) == '/')
3286 definedef
= dsharpseen
;
3287 } /* if (definedef == dnone) */
3295 /* Consider token only if some involved conditions are satisfied. */
3296 if (typdef
!= tignore
3297 && definedef
!= dignorerest
3300 && (definedef
!= dnone
3301 || structdef
!= scolonseen
)
3308 if (c
== ':' && *lp
== ':' && begtoken (lp
[1]))
3309 /* This handles :: in the middle,
3310 but not at the beginning of an identifier.
3311 Also, space-separated :: is not recognised. */
3313 if (c_ext
& C_AUTO
) /* automatic detection of C++ */
3314 c_ext
= (c_ext
| C_PLPL
) & ~C_AUTO
;
3318 goto still_in_token
;
3322 bool funorvar
= FALSE
;
3325 || consider_token (newlb
.buffer
+ tokoff
, toklen
, c
,
3326 &c_ext
, bracelev
, parlev
,
3329 if (fvdef
== foperator
)
3332 lp
= skip_spaces (lp
-1);
3336 && !iswhite (*lp
) && *lp
!= '(')
3339 toklen
+= lp
- oldlp
;
3341 token
.named
= FALSE
;
3343 && nestlev
> 0 && definedef
== dnone
)
3344 /* in struct body */
3346 write_classname (&token_name
, qualifier
);
3347 linebuffer_setlen (&token_name
,
3348 token_name
.len
+qlen
+toklen
);
3349 strcat (token_name
.buffer
, qualifier
);
3350 strncat (token_name
.buffer
,
3351 newlb
.buffer
+ tokoff
, toklen
);
3354 else if (objdef
== ocatseen
)
3355 /* Objective C category */
3357 int len
= strlen (objtag
) + 2 + toklen
;
3358 linebuffer_setlen (&token_name
, len
);
3359 strcpy (token_name
.buffer
, objtag
);
3360 strcat (token_name
.buffer
, "(");
3361 strncat (token_name
.buffer
,
3362 newlb
.buffer
+ tokoff
, toklen
);
3363 strcat (token_name
.buffer
, ")");
3366 else if (objdef
== omethodtag
3367 || objdef
== omethodparm
)
3368 /* Objective C method */
3372 else if (fvdef
== fdefunname
)
3373 /* GNU DEFUN and similar macros */
3375 bool defun
= (newlb
.buffer
[tokoff
] == 'F');
3379 /* Rewrite the tag so that emacs lisp DEFUNs
3380 can be found by their elisp name */
3386 linebuffer_setlen (&token_name
, len
);
3387 strncpy (token_name
.buffer
,
3388 newlb
.buffer
+ off
, len
);
3389 token_name
.buffer
[len
] = '\0';
3392 if (token_name
.buffer
[len
] == '_')
3393 token_name
.buffer
[len
] = '-';
3394 token
.named
= defun
;
3398 linebuffer_setlen (&token_name
, toklen
);
3399 strncpy (token_name
.buffer
,
3400 newlb
.buffer
+ tokoff
, toklen
);
3401 token_name
.buffer
[toklen
] = '\0';
3402 /* Name macros and members. */
3403 token
.named
= (structdef
== stagseen
3404 || typdef
== ttypeseen
3407 && definedef
== dignorerest
)
3409 && definedef
== dnone
3410 && structdef
== snone
3413 token
.lineno
= lineno
;
3414 token
.offset
= tokoff
;
3415 token
.length
= toklen
;
3416 token
.line
= newlb
.buffer
;
3417 token
.linepos
= newlinepos
;
3420 if (definedef
== dnone
3421 && (fvdef
== fvnameseen
3422 || fvdef
== foperator
3423 || structdef
== stagseen
3425 || typdef
== ttypeseen
3426 || objdef
!= onone
))
3428 if (current_lb_is_new
)
3429 switch_line_buffers ();
3431 else if (definedef
!= dnone
3432 || fvdef
== fdefunname
3434 make_C_tag (funorvar
);
3436 else /* not yacc and consider_token failed */
3438 if (inattribute
&& fvdef
== fignore
)
3440 /* We have just met __attribute__ after a
3441 function parameter list: do not tag the
3448 } /* if (endtoken (c)) */
3449 else if (intoken (c
))
3455 } /* if (midtoken) */
3456 else if (begtoken (c
))
3464 /* This prevents tagging fb in
3465 void (__attribute__((noreturn)) *fb) (void);
3466 Fixing this is not easy and not very important. */
3470 if (plainc
|| declarations
)
3472 make_C_tag (TRUE
); /* a function */
3477 if (structdef
== stagseen
&& !cjava
)
3479 popclass_above (bracelev
);
3487 if (!yacc_rules
|| lp
== newlb
.buffer
+ 1)
3489 tokoff
= lp
- 1 - newlb
.buffer
;
3494 } /* if (begtoken) */
3495 } /* if must look at token */
3498 /* Detect end of line, colon, comma, semicolon and various braces
3499 after having handled a token.*/
3505 if (yacc_rules
&& token
.offset
== 0 && token
.valid
)
3507 make_C_tag (FALSE
); /* a yacc function */
3510 if (definedef
!= dnone
)
3516 make_C_tag (TRUE
); /* an Objective C class */
3520 objdef
= omethodcolon
;
3521 linebuffer_setlen (&token_name
, token_name
.len
+ 1);
3522 strcat (token_name
.buffer
, ":");
3525 if (structdef
== stagseen
)
3527 structdef
= scolonseen
;
3530 /* Should be useless, but may be work as a safety net. */
3531 if (cplpl
&& fvdef
== flistseen
)
3533 make_C_tag (TRUE
); /* a function */
3539 if (definedef
!= dnone
|| inattribute
)
3545 make_C_tag (FALSE
); /* a typedef */
3555 if (typdef
== tignore
|| cplpl
)
3559 if ((globals
&& bracelev
== 0 && (!fvextern
|| declarations
))
3560 || (members
&& instruct
))
3561 make_C_tag (FALSE
); /* a variable */
3564 token
.valid
= FALSE
;
3568 && (cplpl
|| !instruct
)
3569 && (typdef
== tnone
|| (typdef
!= tignore
&& instruct
)))
3571 && plainc
&& instruct
))
3572 make_C_tag (TRUE
); /* a function */
3578 && cplpl
&& structdef
== stagseen
)
3579 make_C_tag (FALSE
); /* forward declaration */
3581 token
.valid
= FALSE
;
3582 } /* switch (fvdef) */
3588 if (structdef
== stagseen
)
3592 if (definedef
!= dnone
|| inattribute
)
3598 make_C_tag (TRUE
); /* an Objective C method */
3619 && (!fvextern
|| declarations
))
3620 || (members
&& instruct
)))
3621 make_C_tag (FALSE
); /* a variable */
3624 if ((declarations
&& typdef
== tnone
&& !instruct
)
3625 || (members
&& typdef
!= tignore
&& instruct
))
3627 make_C_tag (TRUE
); /* a function */
3630 else if (!declarations
)
3632 token
.valid
= FALSE
;
3637 if (structdef
== stagseen
)
3641 if (definedef
!= dnone
|| inattribute
)
3643 if (structdef
== stagseen
)
3650 make_C_tag (FALSE
); /* a typedef */
3662 if ((members
&& bracelev
== 1)
3663 || (globals
&& bracelev
== 0
3664 && (!fvextern
|| declarations
)))
3665 make_C_tag (FALSE
); /* a variable */
3679 if (definedef
!= dnone
)
3681 if (objdef
== otagseen
&& parlev
== 0)
3682 objdef
= oparenseen
;
3686 if (typdef
== ttypeseen
3690 /* This handles constructs like:
3691 typedef void OperatorFun (int fun); */
3710 if (--attrparlev
== 0)
3711 inattribute
= FALSE
;
3714 if (definedef
!= dnone
)
3716 if (objdef
== ocatseen
&& parlev
== 1)
3718 make_C_tag (TRUE
); /* an Objective C category */
3732 || typdef
== ttypeseen
))
3735 make_C_tag (FALSE
); /* a typedef */
3738 else if (parlev
< 0) /* can happen due to ill-conceived #if's. */
3742 if (definedef
!= dnone
)
3744 if (typdef
== ttypeseen
)
3746 /* Whenever typdef is set to tinbody (currently only
3747 here), typdefbracelev should be set to bracelev. */
3749 typdefbracelev
= bracelev
;
3754 make_C_tag (TRUE
); /* a function */
3763 make_C_tag (TRUE
); /* an Objective C class */
3768 make_C_tag (TRUE
); /* an Objective C method */
3772 /* Neutralize `extern "C" {' grot. */
3773 if (bracelev
== 0 && structdef
== snone
&& nestlev
== 0
3781 case skeyseen
: /* unnamed struct */
3782 pushclass_above (bracelev
, NULL
, 0);
3785 case stagseen
: /* named struct or enum */
3786 case scolonseen
: /* a class */
3787 pushclass_above (bracelev
,token
.line
+token
.offset
, token
.length
);
3789 make_C_tag (FALSE
); /* a struct or enum */
3795 if (definedef
!= dnone
)
3797 if (fvdef
== fstartlist
)
3799 fvdef
= fvnone
; /* avoid tagging `foo' in `foo (*bar()) ()' */
3800 token
.valid
= FALSE
;
3804 if (definedef
!= dnone
)
3807 if (!ignoreindent
&& lp
== newlb
.buffer
+ 1)
3810 token
.valid
= FALSE
; /* unexpected value, token unreliable */
3811 bracelev
= 0; /* reset brace level if first column */
3812 parlev
= 0; /* also reset paren level, just in case... */
3814 else if (bracelev
< 0)
3816 token
.valid
= FALSE
; /* something gone amiss, token unreliable */
3819 if (bracelev
== 0 && fvdef
== vignore
)
3820 fvdef
= fvnone
; /* end of function */
3821 popclass_above (bracelev
);
3823 /* Only if typdef == tinbody is typdefbracelev significant. */
3824 if (typdef
== tinbody
&& bracelev
<= typdefbracelev
)
3826 assert (bracelev
== typdefbracelev
);
3831 if (definedef
!= dnone
)
3841 if ((members
&& bracelev
== 1)
3842 || (globals
&& bracelev
== 0 && (!fvextern
|| declarations
)))
3843 make_C_tag (FALSE
); /* a variable */
3851 && (structdef
== stagseen
|| fvdef
== fvnameseen
))
3858 if (templatelev
> 0)
3866 if (objdef
== oinbody
&& bracelev
== 0)
3868 objdef
= omethodsign
;
3873 case '#': case '~': case '&': case '%': case '/':
3874 case '|': case '^': case '!': case '.': case '?':
3875 if (definedef
!= dnone
)
3877 /* These surely cannot follow a function tag in C. */
3890 if (objdef
== otagseen
)
3892 make_C_tag (TRUE
); /* an Objective C class */
3895 /* If a macro spans multiple lines don't reset its state. */
3897 CNL_SAVE_DEFINEDEF ();
3903 } /* while not eof */
3905 free (lbs
[0].lb
.buffer
);
3906 free (lbs
[1].lb
.buffer
);
3910 * Process either a C++ file or a C file depending on the setting
3914 default_C_entries (FILE *inf
)
3916 C_entries (cplusplus
? C_PLPL
: C_AUTO
, inf
);
3919 /* Always do plain C. */
3921 plain_C_entries (FILE *inf
)
3926 /* Always do C++. */
3928 Cplusplus_entries (FILE *inf
)
3930 C_entries (C_PLPL
, inf
);
3933 /* Always do Java. */
3935 Cjava_entries (FILE *inf
)
3937 C_entries (C_JAVA
, inf
);
3942 Cstar_entries (FILE *inf
)
3944 C_entries (C_STAR
, inf
);
3947 /* Always do Yacc. */
3949 Yacc_entries (FILE *inf
)
3951 C_entries (YACC
, inf
);
3955 /* Useful macros. */
3956 #define LOOP_ON_INPUT_LINES(file_pointer, line_buffer, char_pointer) \
3957 for (; /* loop initialization */ \
3958 !feof (file_pointer) /* loop test */ \
3959 && /* instructions at start of loop */ \
3960 (readline (&line_buffer, file_pointer), \
3961 char_pointer = line_buffer.buffer, \
3965 #define LOOKING_AT(cp, kw) /* kw is the keyword, a literal string */ \
3966 ((assert("" kw), TRUE) /* syntax error if not a literal string */ \
3967 && strneq ((cp), kw, sizeof(kw)-1) /* cp points at kw */ \
3968 && notinname ((cp)[sizeof(kw)-1]) /* end of kw */ \
3969 && ((cp) = skip_spaces((cp)+sizeof(kw)-1))) /* skip spaces */
3971 /* Similar to LOOKING_AT but does not use notinname, does not skip */
3972 #define LOOKING_AT_NOCASE(cp, kw) /* the keyword is a literal string */ \
3973 ((assert("" kw), TRUE) /* syntax error if not a literal string */ \
3974 && strncaseeq ((cp), kw, sizeof(kw)-1) /* cp points at kw */ \
3975 && ((cp) += sizeof(kw)-1)) /* skip spaces */
3978 * Read a file, but do no processing. This is used to do regexp
3979 * matching on files that have no language defined.
3982 just_read_file (FILE *inf
)
3985 readline (&lb
, inf
);
3989 /* Fortran parsing */
3991 static void F_takeprec (void);
3992 static void F_getit (FILE *);
3997 dbp
= skip_spaces (dbp
);
4001 dbp
= skip_spaces (dbp
);
4002 if (strneq (dbp
, "(*)", 3))
4007 if (!ISDIGIT (*dbp
))
4009 --dbp
; /* force failure */
4014 while (ISDIGIT (*dbp
));
4022 dbp
= skip_spaces (dbp
);
4025 readline (&lb
, inf
);
4030 dbp
= skip_spaces (dbp
);
4032 if (!ISALPHA (*dbp
) && *dbp
!= '_' && *dbp
!= '$')
4034 for (cp
= dbp
+ 1; *cp
!= '\0' && intoken (*cp
); cp
++)
4036 make_tag (dbp
, cp
-dbp
, TRUE
,
4037 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4042 Fortran_functions (FILE *inf
)
4044 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
4047 dbp
++; /* Ratfor escape to fortran */
4048 dbp
= skip_spaces (dbp
);
4052 if (LOOKING_AT_NOCASE (dbp
, "recursive"))
4053 dbp
= skip_spaces (dbp
);
4055 switch (lowcase (*dbp
))
4058 if (nocase_tail ("integer"))
4062 if (nocase_tail ("real"))
4066 if (nocase_tail ("logical"))
4070 if (nocase_tail ("complex") || nocase_tail ("character"))
4074 if (nocase_tail ("double"))
4076 dbp
= skip_spaces (dbp
);
4079 if (nocase_tail ("precision"))
4085 dbp
= skip_spaces (dbp
);
4088 switch (lowcase (*dbp
))
4091 if (nocase_tail ("function"))
4095 if (nocase_tail ("subroutine"))
4099 if (nocase_tail ("entry"))
4103 if (nocase_tail ("blockdata") || nocase_tail ("block data"))
4105 dbp
= skip_spaces (dbp
);
4106 if (*dbp
== '\0') /* assume un-named */
4107 make_tag ("blockdata", 9, TRUE
,
4108 lb
.buffer
, dbp
- lb
.buffer
, lineno
, linecharno
);
4110 F_getit (inf
); /* look for name */
4121 * Philippe Waroquiers (1998)
4124 /* Once we are positioned after an "interesting" keyword, let's get
4125 the real tag value necessary. */
4127 Ada_getit (FILE *inf
, const char *name_qualifier
)
4135 dbp
= skip_spaces (dbp
);
4137 || (dbp
[0] == '-' && dbp
[1] == '-'))
4139 readline (&lb
, inf
);
4142 switch (lowcase(*dbp
))
4145 if (nocase_tail ("body"))
4147 /* Skipping body of procedure body or package body or ....
4148 resetting qualifier to body instead of spec. */
4149 name_qualifier
= "/b";
4154 /* Skipping type of task type or protected type ... */
4155 if (nocase_tail ("type"))
4162 for (cp
= dbp
; *cp
!= '\0' && *cp
!= '"'; cp
++)
4167 dbp
= skip_spaces (dbp
);
4170 && (ISALPHA (*cp
) || ISDIGIT (*cp
) || *cp
== '_' || *cp
== '.'));
4178 name
= concat (dbp
, name_qualifier
, "");
4180 make_tag (name
, strlen (name
), TRUE
,
4181 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4190 Ada_funcs (FILE *inf
)
4192 bool inquote
= FALSE
;
4193 bool skip_till_semicolumn
= FALSE
;
4195 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
4197 while (*dbp
!= '\0')
4199 /* Skip a string i.e. "abcd". */
4200 if (inquote
|| (*dbp
== '"'))
4202 dbp
= etags_strchr (dbp
+ !inquote
, '"');
4207 continue; /* advance char */
4212 break; /* advance line */
4216 /* Skip comments. */
4217 if (dbp
[0] == '-' && dbp
[1] == '-')
4218 break; /* advance line */
4220 /* Skip character enclosed in single quote i.e. 'a'
4221 and skip single quote starting an attribute i.e. 'Image. */
4230 if (skip_till_semicolumn
)
4233 skip_till_semicolumn
= FALSE
;
4235 continue; /* advance char */
4238 /* Search for beginning of a token. */
4239 if (!begtoken (*dbp
))
4242 continue; /* advance char */
4245 /* We are at the beginning of a token. */
4246 switch (lowcase(*dbp
))
4249 if (!packages_only
&& nocase_tail ("function"))
4250 Ada_getit (inf
, "/f");
4252 break; /* from switch */
4253 continue; /* advance char */
4255 if (!packages_only
&& nocase_tail ("procedure"))
4256 Ada_getit (inf
, "/p");
4257 else if (nocase_tail ("package"))
4258 Ada_getit (inf
, "/s");
4259 else if (nocase_tail ("protected")) /* protected type */
4260 Ada_getit (inf
, "/t");
4262 break; /* from switch */
4263 continue; /* advance char */
4266 if (typedefs
&& !packages_only
&& nocase_tail ("use"))
4268 /* when tagging types, avoid tagging use type Pack.Typename;
4269 for this, we will skip everything till a ; */
4270 skip_till_semicolumn
= TRUE
;
4271 continue; /* advance char */
4275 if (!packages_only
&& nocase_tail ("task"))
4276 Ada_getit (inf
, "/k");
4277 else if (typedefs
&& !packages_only
&& nocase_tail ("type"))
4279 Ada_getit (inf
, "/t");
4280 while (*dbp
!= '\0')
4284 break; /* from switch */
4285 continue; /* advance char */
4288 /* Look for the end of the token. */
4289 while (!endtoken (*dbp
))
4292 } /* advance char */
4293 } /* advance line */
4298 * Unix and microcontroller assembly tag handling
4299 * Labels: /^[a-zA-Z_.$][a-zA_Z0-9_.$]*[: ^I^J]/
4300 * Idea by Bob Weiner, Motorola Inc. (1994)
4303 Asm_labels (FILE *inf
)
4307 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4309 /* If first char is alphabetic or one of [_.$], test for colon
4310 following identifier. */
4311 if (ISALPHA (*cp
) || *cp
== '_' || *cp
== '.' || *cp
== '$')
4313 /* Read past label. */
4315 while (ISALNUM (*cp
) || *cp
== '_' || *cp
== '.' || *cp
== '$')
4317 if (*cp
== ':' || iswhite (*cp
))
4318 /* Found end of label, so copy it and add it to the table. */
4319 make_tag (lb
.buffer
, cp
- lb
.buffer
, TRUE
,
4320 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4328 * Perl sub names: /^sub[ \t\n]+[^ \t\n{]+/
4329 * Perl variable names: /^(my|local).../
4330 * Original code by Bart Robinson <lomew@cs.utah.edu> (1995)
4331 * Additions by Michael Ernst <mernst@alum.mit.edu> (1997)
4332 * Ideas by Kai Großjohann <Kai.Grossjohann@CS.Uni-Dortmund.DE> (2001)
4335 Perl_functions (FILE *inf
)
4337 char *package
= savestr ("main"); /* current package name */
4340 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4342 cp
= skip_spaces (cp
);
4344 if (LOOKING_AT (cp
, "package"))
4347 get_tag (cp
, &package
);
4349 else if (LOOKING_AT (cp
, "sub"))
4354 while (!notinname (*cp
))
4357 continue; /* nothing found */
4358 if ((pos
= etags_strchr (sp
, ':')) != NULL
4359 && pos
< cp
&& pos
[1] == ':')
4360 /* The name is already qualified. */
4361 make_tag (sp
, cp
- sp
, TRUE
,
4362 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4366 char savechar
, *name
;
4370 name
= concat (package
, "::", sp
);
4372 make_tag (name
, strlen(name
), TRUE
,
4373 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4377 else if (globals
) /* only if we are tagging global vars */
4379 /* Skip a qualifier, if any. */
4380 bool qual
= LOOKING_AT (cp
, "my") || LOOKING_AT (cp
, "local");
4381 /* After "my" or "local", but before any following paren or space. */
4382 char *varstart
= cp
;
4384 if (qual
/* should this be removed? If yes, how? */
4385 && (*cp
== '$' || *cp
== '@' || *cp
== '%'))
4390 while (ISALNUM (*cp
) || *cp
== '_');
4394 /* Should be examining a variable list at this point;
4395 could insist on seeing an open parenthesis. */
4396 while (*cp
!= '\0' && *cp
!= ';' && *cp
!= '=' && *cp
!= ')')
4402 make_tag (varstart
, cp
- varstart
, FALSE
,
4403 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4412 * Look for /^[\t]*def[ \t\n]+[^ \t\n(:]+/ or /^class[ \t\n]+[^ \t\n(:]+/
4413 * Idea by Eric S. Raymond <esr@thyrsus.com> (1997)
4414 * More ideas by seb bacon <seb@jamkit.com> (2002)
4417 Python_functions (FILE *inf
)
4421 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4423 cp
= skip_spaces (cp
);
4424 if (LOOKING_AT (cp
, "def") || LOOKING_AT (cp
, "class"))
4427 while (!notinname (*cp
) && *cp
!= ':')
4429 make_tag (name
, cp
- name
, TRUE
,
4430 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4439 * - /^[ \t]*function[ \t\n]+[^ \t\n(]+/
4440 * - /^[ \t]*class[ \t\n]+[^ \t\n]+/
4441 * - /^[ \t]*define\(\"[^\"]+/
4442 * Only with --members:
4443 * - /^[ \t]*var[ \t\n]+\$[^ \t\n=;]/
4444 * Idea by Diez B. Roggisch (2001)
4447 PHP_functions (FILE *inf
)
4449 register char *cp
, *name
;
4450 bool search_identifier
= FALSE
;
4452 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4454 cp
= skip_spaces (cp
);
4456 if (search_identifier
4459 while (!notinname (*cp
))
4461 make_tag (name
, cp
- name
, TRUE
,
4462 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4463 search_identifier
= FALSE
;
4465 else if (LOOKING_AT (cp
, "function"))
4468 cp
= skip_spaces (cp
+1);
4472 while (!notinname (*cp
))
4474 make_tag (name
, cp
- name
, TRUE
,
4475 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4478 search_identifier
= TRUE
;
4480 else if (LOOKING_AT (cp
, "class"))
4485 while (*cp
!= '\0' && !iswhite (*cp
))
4487 make_tag (name
, cp
- name
, FALSE
,
4488 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4491 search_identifier
= TRUE
;
4493 else if (strneq (cp
, "define", 6)
4494 && (cp
= skip_spaces (cp
+6))
4496 && (*cp
== '"' || *cp
== '\''))
4500 while (*cp
!= quote
&& *cp
!= '\0')
4502 make_tag (name
, cp
- name
, FALSE
,
4503 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4506 && LOOKING_AT (cp
, "var")
4510 while (!notinname(*cp
))
4512 make_tag (name
, cp
- name
, FALSE
,
4513 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4520 * Cobol tag functions
4521 * We could look for anything that could be a paragraph name.
4522 * i.e. anything that starts in column 8 is one word and ends in a full stop.
4523 * Idea by Corny de Souza (1993)
4526 Cobol_paragraphs (FILE *inf
)
4528 register char *bp
, *ep
;
4530 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4536 /* If eoln, compiler option or comment ignore whole line. */
4537 if (bp
[-1] != ' ' || !ISALNUM (bp
[0]))
4540 for (ep
= bp
; ISALNUM (*ep
) || *ep
== '-'; ep
++)
4543 make_tag (bp
, ep
- bp
, TRUE
,
4544 lb
.buffer
, ep
- lb
.buffer
+ 1, lineno
, linecharno
);
4551 * Ideas by Assar Westerlund <assar@sics.se> (2001)
4554 Makefile_targets (FILE *inf
)
4558 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4560 if (*bp
== '\t' || *bp
== '#')
4562 while (*bp
!= '\0' && *bp
!= '=' && *bp
!= ':')
4564 if (*bp
== ':' || (globals
&& *bp
== '='))
4566 /* We should detect if there is more than one tag, but we do not.
4567 We just skip initial and final spaces. */
4568 char * namestart
= skip_spaces (lb
.buffer
);
4569 while (--bp
> namestart
)
4570 if (!notinname (*bp
))
4572 make_tag (namestart
, bp
- namestart
+ 1, TRUE
,
4573 lb
.buffer
, bp
- lb
.buffer
+ 2, lineno
, linecharno
);
4581 * Original code by Mosur K. Mohan (1989)
4583 * Locates tags for procedures & functions. Doesn't do any type- or
4584 * var-definitions. It does look for the keyword "extern" or
4585 * "forward" immediately following the procedure statement; if found,
4586 * the tag is skipped.
4589 Pascal_functions (FILE *inf
)
4591 linebuffer tline
; /* mostly copied from C_entries */
4593 int save_lineno
, namelen
, taglen
;
4596 bool /* each of these flags is TRUE if: */
4597 incomment
, /* point is inside a comment */
4598 inquote
, /* point is inside '..' string */
4599 get_tagname
, /* point is after PROCEDURE/FUNCTION
4600 keyword, so next item = potential tag */
4601 found_tag
, /* point is after a potential tag */
4602 inparms
, /* point is within parameter-list */
4603 verify_tag
; /* point has passed the parm-list, so the
4604 next token will determine whether this
4605 is a FORWARD/EXTERN to be ignored, or
4606 whether it is a real tag */
4608 save_lcno
= save_lineno
= namelen
= taglen
= 0; /* keep compiler quiet */
4609 name
= NULL
; /* keep compiler quiet */
4612 linebuffer_init (&tline
);
4614 incomment
= inquote
= FALSE
;
4615 found_tag
= FALSE
; /* have a proc name; check if extern */
4616 get_tagname
= FALSE
; /* found "procedure" keyword */
4617 inparms
= FALSE
; /* found '(' after "proc" */
4618 verify_tag
= FALSE
; /* check if "extern" is ahead */
4621 while (!feof (inf
)) /* long main loop to get next char */
4624 if (c
== '\0') /* if end of line */
4626 readline (&lb
, inf
);
4630 if (!((found_tag
&& verify_tag
)
4632 c
= *dbp
++; /* only if don't need *dbp pointing
4633 to the beginning of the name of
4634 the procedure or function */
4638 if (c
== '}') /* within { } comments */
4640 else if (c
== '*' && *dbp
== ')') /* within (* *) comments */
4657 inquote
= TRUE
; /* found first quote */
4659 case '{': /* found open { comment */
4663 if (*dbp
== '*') /* found open (* comment */
4668 else if (found_tag
) /* found '(' after tag, i.e., parm-list */
4671 case ')': /* end of parms list */
4676 if (found_tag
&& !inparms
) /* end of proc or fn stmt */
4683 if (found_tag
&& verify_tag
&& (*dbp
!= ' '))
4685 /* Check if this is an "extern" declaration. */
4688 if (lowcase (*dbp
== 'e'))
4690 if (nocase_tail ("extern")) /* superfluous, really! */
4696 else if (lowcase (*dbp
) == 'f')
4698 if (nocase_tail ("forward")) /* check for forward reference */
4704 if (found_tag
&& verify_tag
) /* not external proc, so make tag */
4708 make_tag (name
, namelen
, TRUE
,
4709 tline
.buffer
, taglen
, save_lineno
, save_lcno
);
4713 if (get_tagname
) /* grab name of proc or fn */
4720 /* Find block name. */
4721 for (cp
= dbp
+ 1; *cp
!= '\0' && !endtoken (*cp
); cp
++)
4724 /* Save all values for later tagging. */
4725 linebuffer_setlen (&tline
, lb
.len
);
4726 strcpy (tline
.buffer
, lb
.buffer
);
4727 save_lineno
= lineno
;
4728 save_lcno
= linecharno
;
4729 name
= tline
.buffer
+ (dbp
- lb
.buffer
);
4731 taglen
= cp
- lb
.buffer
+ 1;
4733 dbp
= cp
; /* set dbp to e-o-token */
4734 get_tagname
= FALSE
;
4738 /* And proceed to check for "extern". */
4740 else if (!incomment
&& !inquote
&& !found_tag
)
4742 /* Check for proc/fn keywords. */
4743 switch (lowcase (c
))
4746 if (nocase_tail ("rocedure")) /* c = 'p', dbp has advanced */
4750 if (nocase_tail ("unction"))
4755 } /* while not eof */
4757 free (tline
.buffer
);
4762 * Lisp tag functions
4763 * look for (def or (DEF, quote or QUOTE
4766 static void L_getit (void);
4771 if (*dbp
== '\'') /* Skip prefix quote */
4773 else if (*dbp
== '(')
4776 /* Try to skip "(quote " */
4777 if (!LOOKING_AT (dbp
, "quote") && !LOOKING_AT (dbp
, "QUOTE"))
4778 /* Ok, then skip "(" before name in (defstruct (foo)) */
4779 dbp
= skip_spaces (dbp
);
4781 get_tag (dbp
, NULL
);
4785 Lisp_functions (FILE *inf
)
4787 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
4792 if (strneq (dbp
+1, "def", 3) || strneq (dbp
+1, "DEF", 3))
4794 dbp
= skip_non_spaces (dbp
);
4795 dbp
= skip_spaces (dbp
);
4800 /* Check for (foo::defmumble name-defined ... */
4803 while (!notinname (*dbp
) && *dbp
!= ':');
4808 while (*dbp
== ':');
4810 if (strneq (dbp
, "def", 3) || strneq (dbp
, "DEF", 3))
4812 dbp
= skip_non_spaces (dbp
);
4813 dbp
= skip_spaces (dbp
);
4823 * Lua script language parsing
4824 * Original code by David A. Capello <dacap@users.sourceforge.net> (2004)
4826 * "function" and "local function" are tags if they start at column 1.
4829 Lua_functions (FILE *inf
)
4833 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4835 if (bp
[0] != 'f' && bp
[0] != 'l')
4838 (void)LOOKING_AT (bp
, "local"); /* skip possible "local" */
4840 if (LOOKING_AT (bp
, "function"))
4848 * Just look for lines where the first character is '/'
4849 * Also look at "defineps" for PSWrap
4851 * Richard Mlynarik <mly@adoc.xerox.com> (1997)
4852 * Masatake Yamato <masata-y@is.aist-nara.ac.jp> (1999)
4855 PS_functions (FILE *inf
)
4857 register char *bp
, *ep
;
4859 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4864 *ep
!= '\0' && *ep
!= ' ' && *ep
!= '{';
4867 make_tag (bp
, ep
- bp
, TRUE
,
4868 lb
.buffer
, ep
- lb
.buffer
+ 1, lineno
, linecharno
);
4870 else if (LOOKING_AT (bp
, "defineps"))
4878 * Ignore anything after \ followed by space or in ( )
4879 * Look for words defined by :
4880 * Look for constant, code, create, defer, value, and variable
4881 * OBP extensions: Look for buffer:, field,
4882 * Ideas by Eduardo Horvath <eeh@netbsd.org> (2004)
4885 Forth_words (FILE *inf
)
4889 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4890 while ((bp
= skip_spaces (bp
))[0] != '\0')
4891 if (bp
[0] == '\\' && iswhite(bp
[1]))
4892 break; /* read next line */
4893 else if (bp
[0] == '(' && iswhite(bp
[1]))
4894 do /* skip to ) or eol */
4896 while (*bp
!= ')' && *bp
!= '\0');
4897 else if ((bp
[0] == ':' && iswhite(bp
[1]) && bp
++)
4898 || LOOKING_AT_NOCASE (bp
, "constant")
4899 || LOOKING_AT_NOCASE (bp
, "code")
4900 || LOOKING_AT_NOCASE (bp
, "create")
4901 || LOOKING_AT_NOCASE (bp
, "defer")
4902 || LOOKING_AT_NOCASE (bp
, "value")
4903 || LOOKING_AT_NOCASE (bp
, "variable")
4904 || LOOKING_AT_NOCASE (bp
, "buffer:")
4905 || LOOKING_AT_NOCASE (bp
, "field"))
4906 get_tag (skip_spaces (bp
), NULL
); /* Yay! A definition! */
4908 bp
= skip_non_spaces (bp
);
4913 * Scheme tag functions
4914 * look for (def... xyzzy
4916 * (def ... ((...(xyzzy ....
4918 * Original code by Ken Haase (1985?)
4921 Scheme_functions (FILE *inf
)
4925 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4927 if (strneq (bp
, "(def", 4) || strneq (bp
, "(DEF", 4))
4929 bp
= skip_non_spaces (bp
+4);
4930 /* Skip over open parens and white space. Don't continue past
4932 while (*bp
&& notinname (*bp
))
4936 if (LOOKING_AT (bp
, "(SET!") || LOOKING_AT (bp
, "(set!"))
4942 /* Find tags in TeX and LaTeX input files. */
4944 /* TEX_toktab is a table of TeX control sequences that define tags.
4945 * Each entry records one such control sequence.
4947 * Original code from who knows whom.
4949 * Stefan Monnier (2002)
4952 static linebuffer
*TEX_toktab
= NULL
; /* Table with tag tokens */
4954 /* Default set of control sequences to put into TEX_toktab.
4955 The value of environment var TEXTAGS is prepended to this. */
4956 static const char *TEX_defenv
= "\
4957 :chapter:section:subsection:subsubsection:eqno:label:ref:cite:bibitem\
4958 :part:appendix:entry:index:def\
4959 :newcommand:renewcommand:newenvironment:renewenvironment";
4961 static void TEX_mode (FILE *);
4962 static void TEX_decode_env (const char *, const char *);
4964 static char TEX_esc
= '\\';
4965 static char TEX_opgrp
= '{';
4966 static char TEX_clgrp
= '}';
4969 * TeX/LaTeX scanning loop.
4972 TeX_commands (FILE *inf
)
4977 /* Select either \ or ! as escape character. */
4980 /* Initialize token table once from environment. */
4981 if (TEX_toktab
== NULL
)
4982 TEX_decode_env ("TEXTAGS", TEX_defenv
);
4984 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4986 /* Look at each TEX keyword in line. */
4989 /* Look for a TEX escape. */
4990 while (*cp
++ != TEX_esc
)
4991 if (cp
[-1] == '\0' || cp
[-1] == '%')
4994 for (key
= TEX_toktab
; key
->buffer
!= NULL
; key
++)
4995 if (strneq (cp
, key
->buffer
, key
->len
))
4998 int namelen
, linelen
;
5001 cp
= skip_spaces (cp
+ key
->len
);
5002 if (*cp
== TEX_opgrp
)
5008 (!iswhite (*p
) && *p
!= '#' &&
5009 *p
!= TEX_opgrp
&& *p
!= TEX_clgrp
);
5014 if (!opgrp
|| *p
== TEX_clgrp
)
5016 while (*p
!= '\0' && *p
!= TEX_opgrp
&& *p
!= TEX_clgrp
)
5018 linelen
= p
- lb
.buffer
+ 1;
5020 make_tag (cp
, namelen
, TRUE
,
5021 lb
.buffer
, linelen
, lineno
, linecharno
);
5022 goto tex_next_line
; /* We only tag a line once */
5030 #define TEX_LESC '\\'
5031 #define TEX_SESC '!'
5033 /* Figure out whether TeX's escapechar is '\\' or '!' and set grouping
5034 chars accordingly. */
5036 TEX_mode (FILE *inf
)
5040 while ((c
= getc (inf
)) != EOF
)
5042 /* Skip to next line if we hit the TeX comment char. */
5044 while (c
!= '\n' && c
!= EOF
)
5046 else if (c
== TEX_LESC
|| c
== TEX_SESC
)
5062 /* If the input file is compressed, inf is a pipe, and rewind may fail.
5063 No attempt is made to correct the situation. */
5067 /* Read environment and prepend it to the default string.
5068 Build token table. */
5070 TEX_decode_env (const char *evarname
, const char *defenv
)
5072 register const char *env
, *p
;
5075 /* Append default string to environment. */
5076 env
= getenv (evarname
);
5080 env
= concat (env
, defenv
, "");
5082 /* Allocate a token table */
5083 for (len
= 1, p
= env
; p
;)
5084 if ((p
= etags_strchr (p
, ':')) && *++p
!= '\0')
5086 TEX_toktab
= xnew (len
, linebuffer
);
5088 /* Unpack environment string into token table. Be careful about */
5089 /* zero-length strings (leading ':', "::" and trailing ':') */
5090 for (i
= 0; *env
!= '\0';)
5092 p
= etags_strchr (env
, ':');
5093 if (!p
) /* End of environment string. */
5094 p
= env
+ strlen (env
);
5096 { /* Only non-zero strings. */
5097 TEX_toktab
[i
].buffer
= savenstr (env
, p
- env
);
5098 TEX_toktab
[i
].len
= p
- env
;
5105 TEX_toktab
[i
].buffer
= NULL
; /* Mark end of table. */
5106 TEX_toktab
[i
].len
= 0;
5113 /* Texinfo support. Dave Love, Mar. 2000. */
5115 Texinfo_nodes (FILE *inf
)
5118 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5119 if (LOOKING_AT (cp
, "@node"))
5122 while (*cp
!= '\0' && *cp
!= ',')
5124 make_tag (start
, cp
- start
, TRUE
,
5125 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
5132 * Contents of <title>, <h1>, <h2>, <h3> are tags.
5133 * Contents of <a name=xxx> are tags with name xxx.
5135 * Francesco Potortì, 2002.
5138 HTML_labels (FILE *inf
)
5140 bool getnext
= FALSE
; /* next text outside of HTML tags is a tag */
5141 bool skiptag
= FALSE
; /* skip to the end of the current HTML tag */
5142 bool intag
= FALSE
; /* inside an html tag, looking for ID= */
5143 bool inanchor
= FALSE
; /* when INTAG, is an anchor, look for NAME= */
5147 linebuffer_setlen (&token_name
, 0); /* no name in buffer */
5149 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
5150 for (;;) /* loop on the same line */
5152 if (skiptag
) /* skip HTML tag */
5154 while (*dbp
!= '\0' && *dbp
!= '>')
5160 continue; /* look on the same line */
5162 break; /* go to next line */
5165 else if (intag
) /* look for "name=" or "id=" */
5167 while (*dbp
!= '\0' && *dbp
!= '>'
5168 && lowcase (*dbp
) != 'n' && lowcase (*dbp
) != 'i')
5171 break; /* go to next line */
5176 continue; /* look on the same line */
5178 if ((inanchor
&& LOOKING_AT_NOCASE (dbp
, "name="))
5179 || LOOKING_AT_NOCASE (dbp
, "id="))
5181 bool quoted
= (dbp
[0] == '"');
5184 for (end
= ++dbp
; *end
!= '\0' && *end
!= '"'; end
++)
5187 for (end
= dbp
; *end
!= '\0' && intoken (*end
); end
++)
5189 linebuffer_setlen (&token_name
, end
- dbp
);
5190 strncpy (token_name
.buffer
, dbp
, end
- dbp
);
5191 token_name
.buffer
[end
- dbp
] = '\0';
5194 intag
= FALSE
; /* we found what we looked for */
5195 skiptag
= TRUE
; /* skip to the end of the tag */
5196 getnext
= TRUE
; /* then grab the text */
5197 continue; /* look on the same line */
5202 else if (getnext
) /* grab next tokens and tag them */
5204 dbp
= skip_spaces (dbp
);
5206 break; /* go to next line */
5210 inanchor
= (lowcase (dbp
[1]) == 'a' && !intoken (dbp
[2]));
5211 continue; /* look on the same line */
5214 for (end
= dbp
+ 1; *end
!= '\0' && *end
!= '<'; end
++)
5216 make_tag (token_name
.buffer
, token_name
.len
, TRUE
,
5217 dbp
, end
- dbp
, lineno
, linecharno
);
5218 linebuffer_setlen (&token_name
, 0); /* no name in buffer */
5220 break; /* go to next line */
5223 else /* look for an interesting HTML tag */
5225 while (*dbp
!= '\0' && *dbp
!= '<')
5228 break; /* go to next line */
5230 if (lowcase (dbp
[1]) == 'a' && !intoken (dbp
[2]))
5233 continue; /* look on the same line */
5235 else if (LOOKING_AT_NOCASE (dbp
, "<title>")
5236 || LOOKING_AT_NOCASE (dbp
, "<h1>")
5237 || LOOKING_AT_NOCASE (dbp
, "<h2>")
5238 || LOOKING_AT_NOCASE (dbp
, "<h3>"))
5242 continue; /* look on the same line */
5253 * Assumes that the predicate or rule starts at column 0.
5254 * Only the first clause of a predicate or rule is added.
5255 * Original code by Sunichirou Sugou (1989)
5256 * Rewritten by Anders Lindgren (1996)
5258 static size_t prolog_pr (char *, char *);
5259 static void prolog_skip_comment (linebuffer
*, FILE *);
5260 static size_t prolog_atom (char *, size_t);
5263 Prolog_functions (FILE *inf
)
5273 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5275 if (cp
[0] == '\0') /* Empty line */
5277 else if (iswhite (cp
[0])) /* Not a predicate */
5279 else if (cp
[0] == '/' && cp
[1] == '*') /* comment. */
5280 prolog_skip_comment (&lb
, inf
);
5281 else if ((len
= prolog_pr (cp
, last
)) > 0)
5283 /* Predicate or rule. Store the function name so that we
5284 only generate a tag for the first clause. */
5286 last
= xnew(len
+ 1, char);
5287 else if (len
+ 1 > allocated
)
5288 xrnew (last
, len
+ 1, char);
5289 allocated
= len
+ 1;
5290 strncpy (last
, cp
, len
);
5299 prolog_skip_comment (linebuffer
*plb
, FILE *inf
)
5305 for (cp
= plb
->buffer
; *cp
!= '\0'; cp
++)
5306 if (cp
[0] == '*' && cp
[1] == '/')
5308 readline (plb
, inf
);
5314 * A predicate or rule definition is added if it matches:
5315 * <beginning of line><Prolog Atom><whitespace>(
5316 * or <beginning of line><Prolog Atom><whitespace>:-
5318 * It is added to the tags database if it doesn't match the
5319 * name of the previous clause header.
5321 * Return the size of the name of the predicate or rule, or 0 if no
5325 prolog_pr (char *s
, char *last
)
5327 /* Name of last clause. */
5332 pos
= prolog_atom (s
, 0);
5337 pos
= skip_spaces (s
+ pos
) - s
;
5340 || (s
[pos
] == '(' && (pos
+= 1))
5341 || (s
[pos
] == ':' && s
[pos
+ 1] == '-' && (pos
+= 2)))
5342 && (last
== NULL
/* save only the first clause */
5343 || len
!= strlen (last
)
5344 || !strneq (s
, last
, len
)))
5346 make_tag (s
, len
, TRUE
, s
, pos
, lineno
, linecharno
);
5354 * Consume a Prolog atom.
5355 * Return the number of bytes consumed, or 0 if there was an error.
5357 * A prolog atom, in this context, could be one of:
5358 * - An alphanumeric sequence, starting with a lower case letter.
5359 * - A quoted arbitrary string. Single quotes can escape themselves.
5360 * Backslash quotes everything.
5363 prolog_atom (char *s
, size_t pos
)
5369 if (ISLOWER(s
[pos
]) || (s
[pos
] == '_'))
5371 /* The atom is unquoted. */
5373 while (ISALNUM(s
[pos
]) || (s
[pos
] == '_'))
5377 return pos
- origpos
;
5379 else if (s
[pos
] == '\'')
5390 pos
++; /* A double quote */
5392 else if (s
[pos
] == '\0')
5393 /* Multiline quoted atoms are ignored. */
5395 else if (s
[pos
] == '\\')
5397 if (s
[pos
+1] == '\0')
5404 return pos
- origpos
;
5412 * Support for Erlang
5414 * Generates tags for functions, defines, and records.
5415 * Assumes that Erlang functions start at column 0.
5416 * Original code by Anders Lindgren (1996)
5418 static int erlang_func (char *, char *);
5419 static void erlang_attribute (char *);
5420 static int erlang_atom (char *);
5423 Erlang_functions (FILE *inf
)
5433 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5435 if (cp
[0] == '\0') /* Empty line */
5437 else if (iswhite (cp
[0])) /* Not function nor attribute */
5439 else if (cp
[0] == '%') /* comment */
5441 else if (cp
[0] == '"') /* Sometimes, strings start in column one */
5443 else if (cp
[0] == '-') /* attribute, e.g. "-define" */
5445 erlang_attribute (cp
);
5452 else if ((len
= erlang_func (cp
, last
)) > 0)
5455 * Function. Store the function name so that we only
5456 * generates a tag for the first clause.
5459 last
= xnew (len
+ 1, char);
5460 else if (len
+ 1 > allocated
)
5461 xrnew (last
, len
+ 1, char);
5462 allocated
= len
+ 1;
5463 strncpy (last
, cp
, len
);
5472 * A function definition is added if it matches:
5473 * <beginning of line><Erlang Atom><whitespace>(
5475 * It is added to the tags database if it doesn't match the
5476 * name of the previous clause header.
5478 * Return the size of the name of the function, or 0 if no function
5482 erlang_func (char *s
, char *last
)
5484 /* Name of last clause. */
5489 pos
= erlang_atom (s
);
5494 pos
= skip_spaces (s
+ pos
) - s
;
5496 /* Save only the first clause. */
5499 || len
!= (int)strlen (last
)
5500 || !strneq (s
, last
, len
)))
5502 make_tag (s
, len
, TRUE
, s
, pos
, lineno
, linecharno
);
5511 * Handle attributes. Currently, tags are generated for defines
5514 * They are on the form:
5515 * -define(foo, bar).
5516 * -define(Foo(M, N), M+N).
5517 * -record(graph, {vtab = notable, cyclic = true}).
5520 erlang_attribute (char *s
)
5524 if ((LOOKING_AT (cp
, "-define") || LOOKING_AT (cp
, "-record"))
5527 int len
= erlang_atom (skip_spaces (cp
));
5529 make_tag (cp
, len
, TRUE
, s
, cp
+ len
- s
, lineno
, linecharno
);
5536 * Consume an Erlang atom (or variable).
5537 * Return the number of bytes consumed, or -1 if there was an error.
5540 erlang_atom (char *s
)
5544 if (ISALPHA (s
[pos
]) || s
[pos
] == '_')
5546 /* The atom is unquoted. */
5549 while (ISALNUM (s
[pos
]) || s
[pos
] == '_');
5551 else if (s
[pos
] == '\'')
5553 for (pos
++; s
[pos
] != '\''; pos
++)
5554 if (s
[pos
] == '\0' /* multiline quoted atoms are ignored */
5555 || (s
[pos
] == '\\' && s
[++pos
] == '\0'))
5564 static char *scan_separators (char *);
5565 static void add_regex (char *, language
*);
5566 static char *substitute (char *, char *, struct re_registers
*);
5569 * Take a string like "/blah/" and turn it into "blah", verifying
5570 * that the first and last characters are the same, and handling
5571 * quoted separator characters. Actually, stops on the occurrence of
5572 * an unquoted separator. Also process \t, \n, etc. and turn into
5573 * appropriate characters. Works in place. Null terminates name string.
5574 * Returns pointer to terminating separator, or NULL for
5575 * unterminated regexps.
5578 scan_separators (char *name
)
5581 char *copyto
= name
;
5582 bool quoted
= FALSE
;
5584 for (++name
; *name
!= '\0'; ++name
)
5590 case 'a': *copyto
++ = '\007'; break; /* BEL (bell) */
5591 case 'b': *copyto
++ = '\b'; break; /* BS (back space) */
5592 case 'd': *copyto
++ = 0177; break; /* DEL (delete) */
5593 case 'e': *copyto
++ = 033; break; /* ESC (delete) */
5594 case 'f': *copyto
++ = '\f'; break; /* FF (form feed) */
5595 case 'n': *copyto
++ = '\n'; break; /* NL (new line) */
5596 case 'r': *copyto
++ = '\r'; break; /* CR (carriage return) */
5597 case 't': *copyto
++ = '\t'; break; /* TAB (horizontal tab) */
5598 case 'v': *copyto
++ = '\v'; break; /* VT (vertical tab) */
5604 /* Something else is quoted, so preserve the quote. */
5612 else if (*name
== '\\')
5614 else if (*name
== sep
)
5620 name
= NULL
; /* signal unterminated regexp */
5622 /* Terminate copied string. */
5627 /* Look at the argument of --regex or --no-regex and do the right
5628 thing. Same for each line of a regexp file. */
5630 analyse_regex (char *regex_arg
)
5632 if (regex_arg
== NULL
)
5634 free_regexps (); /* --no-regex: remove existing regexps */
5638 /* A real --regexp option or a line in a regexp file. */
5639 switch (regex_arg
[0])
5641 /* Comments in regexp file or null arg to --regex. */
5647 /* Read a regex file. This is recursive and may result in a
5648 loop, which will stop when the file descriptors are exhausted. */
5652 linebuffer regexbuf
;
5653 char *regexfile
= regex_arg
+ 1;
5655 /* regexfile is a file containing regexps, one per line. */
5656 regexfp
= fopen (regexfile
, "r");
5657 if (regexfp
== NULL
)
5662 linebuffer_init (®exbuf
);
5663 while (readline_internal (®exbuf
, regexfp
) > 0)
5664 analyse_regex (regexbuf
.buffer
);
5665 free (regexbuf
.buffer
);
5670 /* Regexp to be used for a specific language only. */
5674 char *lang_name
= regex_arg
+ 1;
5677 for (cp
= lang_name
; *cp
!= '}'; cp
++)
5680 error ("unterminated language name in regex: %s", regex_arg
);
5684 lang
= get_language_from_langname (lang_name
);
5687 add_regex (cp
, lang
);
5691 /* Regexp to be used for any language. */
5693 add_regex (regex_arg
, NULL
);
5698 /* Separate the regexp pattern, compile it,
5699 and care for optional name and modifiers. */
5701 add_regex (char *regexp_pattern
, language
*lang
)
5703 static struct re_pattern_buffer zeropattern
;
5704 char sep
, *pat
, *name
, *modifiers
;
5707 struct re_pattern_buffer
*patbuf
;
5710 force_explicit_name
= TRUE
, /* do not use implicit tag names */
5711 ignore_case
= FALSE
, /* case is significant */
5712 multi_line
= FALSE
, /* matches are done one line at a time */
5713 single_line
= FALSE
; /* dot does not match newline */
5716 if (strlen(regexp_pattern
) < 3)
5718 error ("null regexp", (char *)NULL
);
5721 sep
= regexp_pattern
[0];
5722 name
= scan_separators (regexp_pattern
);
5725 error ("%s: unterminated regexp", regexp_pattern
);
5730 error ("null name for regexp \"%s\"", regexp_pattern
);
5733 modifiers
= scan_separators (name
);
5734 if (modifiers
== NULL
) /* no terminating separator --> no name */
5740 modifiers
+= 1; /* skip separator */
5742 /* Parse regex modifiers. */
5743 for (; modifiers
[0] != '\0'; modifiers
++)
5744 switch (modifiers
[0])
5747 if (modifiers
== name
)
5748 error ("forcing explicit tag name but no name, ignoring", NULL
);
5749 force_explicit_name
= TRUE
;
5759 need_filebuf
= TRUE
;
5764 wrongmod
[0] = modifiers
[0];
5766 error ("invalid regexp modifier `%s', ignoring", wrongmod
);
5771 patbuf
= xnew (1, struct re_pattern_buffer
);
5772 *patbuf
= zeropattern
;
5775 static char lc_trans
[CHARS
];
5777 for (i
= 0; i
< CHARS
; i
++)
5778 lc_trans
[i
] = lowcase (i
);
5779 patbuf
->translate
= lc_trans
; /* translation table to fold case */
5783 pat
= concat ("^", regexp_pattern
, ""); /* anchor to beginning of line */
5785 pat
= regexp_pattern
;
5788 re_set_syntax (RE_SYNTAX_EMACS
| RE_DOT_NEWLINE
);
5790 re_set_syntax (RE_SYNTAX_EMACS
);
5792 err
= re_compile_pattern (pat
, strlen (pat
), patbuf
);
5797 error ("%s while compiling pattern", err
);
5802 p_head
= xnew (1, regexp
);
5803 p_head
->pattern
= savestr (regexp_pattern
);
5804 p_head
->p_next
= rp
;
5805 p_head
->lang
= lang
;
5806 p_head
->pat
= patbuf
;
5807 p_head
->name
= savestr (name
);
5808 p_head
->error_signaled
= FALSE
;
5809 p_head
->force_explicit_name
= force_explicit_name
;
5810 p_head
->ignore_case
= ignore_case
;
5811 p_head
->multi_line
= multi_line
;
5815 * Do the substitutions indicated by the regular expression and
5819 substitute (char *in
, char *out
, struct re_registers
*regs
)
5822 int size
, dig
, diglen
;
5825 size
= strlen (out
);
5827 /* Pass 1: figure out how much to allocate by finding all \N strings. */
5828 if (out
[size
- 1] == '\\')
5829 fatal ("pattern error in \"%s\"", out
);
5830 for (t
= etags_strchr (out
, '\\');
5832 t
= etags_strchr (t
+ 2, '\\'))
5836 diglen
= regs
->end
[dig
] - regs
->start
[dig
];
5842 /* Allocate space and do the substitutions. */
5844 result
= xnew (size
+ 1, char);
5846 for (t
= result
; *out
!= '\0'; out
++)
5847 if (*out
== '\\' && ISDIGIT (*++out
))
5850 diglen
= regs
->end
[dig
] - regs
->start
[dig
];
5851 strncpy (t
, in
+ regs
->start
[dig
], diglen
);
5858 assert (t
<= result
+ size
);
5859 assert (t
- result
== (int)strlen (result
));
5864 /* Deallocate all regexps. */
5869 while (p_head
!= NULL
)
5871 rp
= p_head
->p_next
;
5872 free (p_head
->pattern
);
5873 free (p_head
->name
);
5881 * Reads the whole file as a single string from `filebuf' and looks for
5882 * multi-line regular expressions, creating tags on matches.
5883 * readline already dealt with normal regexps.
5885 * Idea by Ben Wing <ben@666.com> (2002).
5888 regex_tag_multiline (void)
5890 char *buffer
= filebuf
.buffer
;
5894 for (rp
= p_head
; rp
!= NULL
; rp
= rp
->p_next
)
5898 if (!rp
->multi_line
)
5899 continue; /* skip normal regexps */
5901 /* Generic initialisations before parsing file from memory. */
5902 lineno
= 1; /* reset global line number */
5903 charno
= 0; /* reset global char number */
5904 linecharno
= 0; /* reset global char number of line start */
5906 /* Only use generic regexps or those for the current language. */
5907 if (rp
->lang
!= NULL
&& rp
->lang
!= curfdp
->lang
)
5910 while (match
>= 0 && match
< filebuf
.len
)
5912 match
= re_search (rp
->pat
, buffer
, filebuf
.len
, charno
,
5913 filebuf
.len
- match
, &rp
->regs
);
5918 if (!rp
->error_signaled
)
5920 error ("regexp stack overflow while matching \"%s\"",
5922 rp
->error_signaled
= TRUE
;
5929 if (match
== rp
->regs
.end
[0])
5931 if (!rp
->error_signaled
)
5933 error ("regexp matches the empty string: \"%s\"",
5935 rp
->error_signaled
= TRUE
;
5937 match
= -3; /* exit from while loop */
5941 /* Match occurred. Construct a tag. */
5942 while (charno
< rp
->regs
.end
[0])
5943 if (buffer
[charno
++] == '\n')
5944 lineno
++, linecharno
= charno
;
5946 if (name
[0] == '\0')
5948 else /* make a named tag */
5949 name
= substitute (buffer
, rp
->name
, &rp
->regs
);
5950 if (rp
->force_explicit_name
)
5951 /* Force explicit tag name, if a name is there. */
5952 pfnote (name
, TRUE
, buffer
+ linecharno
,
5953 charno
- linecharno
+ 1, lineno
, linecharno
);
5955 make_tag (name
, strlen (name
), TRUE
, buffer
+ linecharno
,
5956 charno
- linecharno
+ 1, lineno
, linecharno
);
5965 nocase_tail (const char *cp
)
5967 register int len
= 0;
5969 while (*cp
!= '\0' && lowcase (*cp
) == lowcase (dbp
[len
]))
5971 if (*cp
== '\0' && !intoken (dbp
[len
]))
5980 get_tag (register char *bp
, char **namepp
)
5982 register char *cp
= bp
;
5986 /* Go till you get to white space or a syntactic break */
5987 for (cp
= bp
+ 1; !notinname (*cp
); cp
++)
5989 make_tag (bp
, cp
- bp
, TRUE
,
5990 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
5994 *namepp
= savenstr (bp
, cp
- bp
);
5998 * Read a line of text from `stream' into `lbp', excluding the
5999 * newline or CR-NL, if any. Return the number of characters read from
6000 * `stream', which is the length of the line including the newline.
6002 * On DOS or Windows we do not count the CR character, if any before the
6003 * NL, in the returned length; this mirrors the behavior of Emacs on those
6004 * platforms (for text files, it translates CR-NL to NL as it reads in the
6007 * If multi-line regular expressions are requested, each line read is
6008 * appended to `filebuf'.
6011 readline_internal (linebuffer
*lbp
, register FILE *stream
)
6013 char *buffer
= lbp
->buffer
;
6014 register char *p
= lbp
->buffer
;
6015 register char *pend
;
6018 pend
= p
+ lbp
->size
; /* Separate to avoid 386/IX compiler bug. */
6022 register int c
= getc (stream
);
6025 /* We're at the end of linebuffer: expand it. */
6027 xrnew (buffer
, lbp
->size
, char);
6028 p
+= buffer
- lbp
->buffer
;
6029 pend
= buffer
+ lbp
->size
;
6030 lbp
->buffer
= buffer
;
6040 if (p
> buffer
&& p
[-1] == '\r')
6044 /* Assume CRLF->LF translation will be performed by Emacs
6045 when loading this file, so CRs won't appear in the buffer.
6046 It would be cleaner to compensate within Emacs;
6047 however, Emacs does not know how many CRs were deleted
6048 before any given point in the file. */
6063 lbp
->len
= p
- buffer
;
6065 if (need_filebuf
/* we need filebuf for multi-line regexps */
6066 && chars_deleted
> 0) /* not at EOF */
6068 while (filebuf
.size
<= filebuf
.len
+ lbp
->len
+ 1) /* +1 for \n */
6070 /* Expand filebuf. */
6072 xrnew (filebuf
.buffer
, filebuf
.size
, char);
6074 strncpy (filebuf
.buffer
+ filebuf
.len
, lbp
->buffer
, lbp
->len
);
6075 filebuf
.len
+= lbp
->len
;
6076 filebuf
.buffer
[filebuf
.len
++] = '\n';
6077 filebuf
.buffer
[filebuf
.len
] = '\0';
6080 return lbp
->len
+ chars_deleted
;
6084 * Like readline_internal, above, but in addition try to match the
6085 * input line against relevant regular expressions and manage #line
6089 readline (linebuffer
*lbp
, FILE *stream
)
6093 linecharno
= charno
; /* update global char number of line start */
6094 result
= readline_internal (lbp
, stream
); /* read line */
6095 lineno
+= 1; /* increment global line number */
6096 charno
+= result
; /* increment global char number */
6098 /* Honour #line directives. */
6099 if (!no_line_directive
)
6101 static bool discard_until_line_directive
;
6103 /* Check whether this is a #line directive. */
6104 if (result
> 12 && strneq (lbp
->buffer
, "#line ", 6))
6109 if (sscanf (lbp
->buffer
, "#line %u \"%n", &lno
, &start
) >= 1
6110 && start
> 0) /* double quote character found */
6112 char *endp
= lbp
->buffer
+ start
;
6114 while ((endp
= etags_strchr (endp
, '"')) != NULL
6115 && endp
[-1] == '\\')
6118 /* Ok, this is a real #line directive. Let's deal with it. */
6120 char *taggedabsname
; /* absolute name of original file */
6121 char *taggedfname
; /* name of original file as given */
6122 char *name
; /* temp var */
6124 discard_until_line_directive
= FALSE
; /* found it */
6125 name
= lbp
->buffer
+ start
;
6127 canonicalize_filename (name
);
6128 taggedabsname
= absolute_filename (name
, tagfiledir
);
6129 if (filename_is_absolute (name
)
6130 || filename_is_absolute (curfdp
->infname
))
6131 taggedfname
= savestr (taggedabsname
);
6133 taggedfname
= relative_filename (taggedabsname
,tagfiledir
);
6135 if (streq (curfdp
->taggedfname
, taggedfname
))
6136 /* The #line directive is only a line number change. We
6137 deal with this afterwards. */
6140 /* The tags following this #line directive should be
6141 attributed to taggedfname. In order to do this, set
6142 curfdp accordingly. */
6144 fdesc
*fdp
; /* file description pointer */
6146 /* Go look for a file description already set up for the
6147 file indicated in the #line directive. If there is
6148 one, use it from now until the next #line
6150 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
6151 if (streq (fdp
->infname
, curfdp
->infname
)
6152 && streq (fdp
->taggedfname
, taggedfname
))
6153 /* If we remove the second test above (after the &&)
6154 then all entries pertaining to the same file are
6155 coalesced in the tags file. If we use it, then
6156 entries pertaining to the same file but generated
6157 from different files (via #line directives) will
6158 go into separate sections in the tags file. These
6159 alternatives look equivalent. The first one
6160 destroys some apparently useless information. */
6166 /* Else, if we already tagged the real file, skip all
6167 input lines until the next #line directive. */
6168 if (fdp
== NULL
) /* not found */
6169 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
6170 if (streq (fdp
->infabsname
, taggedabsname
))
6172 discard_until_line_directive
= TRUE
;
6176 /* Else create a new file description and use that from
6177 now on, until the next #line directive. */
6178 if (fdp
== NULL
) /* not found */
6181 fdhead
= xnew (1, fdesc
);
6182 *fdhead
= *curfdp
; /* copy curr. file description */
6184 fdhead
->infname
= savestr (curfdp
->infname
);
6185 fdhead
->infabsname
= savestr (curfdp
->infabsname
);
6186 fdhead
->infabsdir
= savestr (curfdp
->infabsdir
);
6187 fdhead
->taggedfname
= taggedfname
;
6188 fdhead
->usecharno
= FALSE
;
6189 fdhead
->prop
= NULL
;
6190 fdhead
->written
= FALSE
;
6194 free (taggedabsname
);
6196 readline (lbp
, stream
);
6198 } /* if a real #line directive */
6199 } /* if #line is followed by a number */
6200 } /* if line begins with "#line " */
6202 /* If we are here, no #line directive was found. */
6203 if (discard_until_line_directive
)
6207 /* Do a tail recursion on ourselves, thus discarding the contents
6208 of the line buffer. */
6209 readline (lbp
, stream
);
6213 discard_until_line_directive
= FALSE
;
6216 } /* if #line directives should be considered */
6223 /* Match against relevant regexps. */
6225 for (rp
= p_head
; rp
!= NULL
; rp
= rp
->p_next
)
6227 /* Only use generic regexps or those for the current language.
6228 Also do not use multiline regexps, which is the job of
6229 regex_tag_multiline. */
6230 if ((rp
->lang
!= NULL
&& rp
->lang
!= fdhead
->lang
)
6234 match
= re_match (rp
->pat
, lbp
->buffer
, lbp
->len
, 0, &rp
->regs
);
6239 if (!rp
->error_signaled
)
6241 error ("regexp stack overflow while matching \"%s\"",
6243 rp
->error_signaled
= TRUE
;
6250 /* Empty string matched. */
6251 if (!rp
->error_signaled
)
6253 error ("regexp matches the empty string: \"%s\"", rp
->pattern
);
6254 rp
->error_signaled
= TRUE
;
6258 /* Match occurred. Construct a tag. */
6260 if (name
[0] == '\0')
6262 else /* make a named tag */
6263 name
= substitute (lbp
->buffer
, rp
->name
, &rp
->regs
);
6264 if (rp
->force_explicit_name
)
6265 /* Force explicit tag name, if a name is there. */
6266 pfnote (name
, TRUE
, lbp
->buffer
, match
, lineno
, linecharno
);
6268 make_tag (name
, strlen (name
), TRUE
,
6269 lbp
->buffer
, match
, lineno
, linecharno
);
6278 * Return a pointer to a space of size strlen(cp)+1 allocated
6279 * with xnew where the string CP has been copied.
6282 savestr (const char *cp
)
6284 return savenstr (cp
, strlen (cp
));
6288 * Return a pointer to a space of size LEN+1 allocated with xnew where
6289 * the string CP has been copied for at most the first LEN characters.
6292 savenstr (const char *cp
, int len
)
6296 dp
= xnew (len
+ 1, char);
6297 strncpy (dp
, cp
, len
);
6303 * Return the ptr in sp at which the character c last
6304 * appears; NULL if not found
6306 * Identical to POSIX strrchr, included for portability.
6309 etags_strrchr (register const char *sp
, register int c
)
6311 register const char *r
;
6323 * Return the ptr in sp at which the character c first
6324 * appears; NULL if not found
6326 * Identical to POSIX strchr, included for portability.
6329 etags_strchr (register const char *sp
, register int c
)
6340 * Compare two strings, ignoring case for alphabetic characters.
6342 * Same as BSD's strcasecmp, included for portability.
6345 etags_strcasecmp (register const char *s1
, register const char *s2
)
6348 && (ISALPHA (*s1
) && ISALPHA (*s2
)
6349 ? lowcase (*s1
) == lowcase (*s2
)
6353 return (ISALPHA (*s1
) && ISALPHA (*s2
)
6354 ? lowcase (*s1
) - lowcase (*s2
)
6359 * Compare two strings, ignoring case for alphabetic characters.
6360 * Stop after a given number of characters
6362 * Same as BSD's strncasecmp, included for portability.
6365 etags_strncasecmp (register const char *s1
, register const char *s2
, register int n
)
6367 while (*s1
!= '\0' && n
-- > 0
6368 && (ISALPHA (*s1
) && ISALPHA (*s2
)
6369 ? lowcase (*s1
) == lowcase (*s2
)
6376 return (ISALPHA (*s1
) && ISALPHA (*s2
)
6377 ? lowcase (*s1
) - lowcase (*s2
)
6381 /* Skip spaces (end of string is not space), return new pointer. */
6383 skip_spaces (char *cp
)
6385 while (iswhite (*cp
))
6390 /* Skip non spaces, except end of string, return new pointer. */
6392 skip_non_spaces (char *cp
)
6394 while (*cp
!= '\0' && !iswhite (*cp
))
6399 /* Print error message and exit. */
6401 fatal (const char *s1
, const char *s2
)
6404 exit (EXIT_FAILURE
);
6408 pfatal (const char *s1
)
6411 exit (EXIT_FAILURE
);
6415 suggest_asking_for_help (void)
6417 fprintf (stderr
, "\tTry `%s %s' for a complete list of options.\n",
6418 progname
, NO_LONG_OPTIONS
? "-h" : "--help");
6419 exit (EXIT_FAILURE
);
6422 /* Print error message. `s1' is printf control string, `s2' is arg for it. */
6424 error (const char *s1
, const char *s2
)
6426 fprintf (stderr
, "%s: ", progname
);
6427 fprintf (stderr
, s1
, s2
);
6428 fprintf (stderr
, "\n");
6431 /* Return a newly-allocated string whose contents
6432 concatenate those of s1, s2, s3. */
6434 concat (const char *s1
, const char *s2
, const char *s3
)
6436 int len1
= strlen (s1
), len2
= strlen (s2
), len3
= strlen (s3
);
6437 char *result
= xnew (len1
+ len2
+ len3
+ 1, char);
6439 strcpy (result
, s1
);
6440 strcpy (result
+ len1
, s2
);
6441 strcpy (result
+ len1
+ len2
, s3
);
6442 result
[len1
+ len2
+ len3
] = '\0';
6448 /* Does the same work as the system V getcwd, but does not need to
6449 guess the buffer size in advance. */
6455 char *path
= xnew (bufsize
, char);
6457 while (getcwd (path
, bufsize
) == NULL
)
6459 if (errno
!= ERANGE
)
6463 path
= xnew (bufsize
, char);
6466 canonicalize_filename (path
);
6469 #else /* not HAVE_GETCWD */
6472 char *p
, path
[MAXPATHLEN
+ 1]; /* Fixed size is safe on MSDOS. */
6476 for (p
= path
; *p
!= '\0'; p
++)
6482 return strdup (path
);
6483 #else /* not MSDOS */
6487 linebuffer_init (&path
);
6488 pipe
= (FILE *) popen ("pwd 2>/dev/null", "r");
6489 if (pipe
== NULL
|| readline_internal (&path
, pipe
) == 0)
6494 #endif /* not MSDOS */
6495 #endif /* not HAVE_GETCWD */
6498 /* Return a newly allocated string containing the file name of FILE
6499 relative to the absolute directory DIR (which should end with a slash). */
6501 relative_filename (char *file
, char *dir
)
6503 char *fp
, *dp
, *afn
, *res
;
6506 /* Find the common root of file and dir (with a trailing slash). */
6507 afn
= absolute_filename (file
, cwd
);
6510 while (*fp
++ == *dp
++)
6512 fp
--, dp
--; /* back to the first differing char */
6514 if (fp
== afn
&& afn
[0] != '/') /* cannot build a relative name */
6517 do /* look at the equal chars until '/' */
6521 /* Build a sequence of "../" strings for the resulting relative file name. */
6523 while ((dp
= etags_strchr (dp
+ 1, '/')) != NULL
)
6525 res
= xnew (3*i
+ strlen (fp
+ 1) + 1, char);
6528 strcat (res
, "../");
6530 /* Add the file name relative to the common root of file and dir. */
6531 strcat (res
, fp
+ 1);
6537 /* Return a newly allocated string containing the absolute file name
6538 of FILE given DIR (which should end with a slash). */
6540 absolute_filename (char *file
, char *dir
)
6542 char *slashp
, *cp
, *res
;
6544 if (filename_is_absolute (file
))
6545 res
= savestr (file
);
6547 /* We don't support non-absolute file names with a drive
6548 letter, like `d:NAME' (it's too much hassle). */
6549 else if (file
[1] == ':')
6550 fatal ("%s: relative file names with drive letters not supported", file
);
6553 res
= concat (dir
, file
, "");
6555 /* Delete the "/dirname/.." and "/." substrings. */
6556 slashp
= etags_strchr (res
, '/');
6557 while (slashp
!= NULL
&& slashp
[0] != '\0')
6559 if (slashp
[1] == '.')
6561 if (slashp
[2] == '.'
6562 && (slashp
[3] == '/' || slashp
[3] == '\0'))
6567 while (cp
>= res
&& !filename_is_absolute (cp
));
6569 cp
= slashp
; /* the absolute name begins with "/.." */
6571 /* Under MSDOS and NT we get `d:/NAME' as absolute
6572 file name, so the luser could say `d:/../NAME'.
6573 We silently treat this as `d:/NAME'. */
6574 else if (cp
[0] != '/')
6578 memmove (cp
, slashp
+ 3, strlen (slashp
+ 2));
6580 /* Overlapping copy isn't really okay */
6581 strcpy (cp
, slashp
+ 3);
6586 else if (slashp
[2] == '/' || slashp
[2] == '\0')
6589 memmove (slashp
, slashp
+ 2, strlen (slashp
+ 1));
6591 strcpy (slashp
, slashp
+ 2);
6597 slashp
= etags_strchr (slashp
+ 1, '/');
6600 if (res
[0] == '\0') /* just a safety net: should never happen */
6603 return savestr ("/");
6609 /* Return a newly allocated string containing the absolute
6610 file name of dir where FILE resides given DIR (which should
6611 end with a slash). */
6613 absolute_dirname (char *file
, char *dir
)
6618 slashp
= etags_strrchr (file
, '/');
6620 return savestr (dir
);
6623 res
= absolute_filename (file
, dir
);
6629 /* Whether the argument string is an absolute file name. The argument
6630 string must have been canonicalized with canonicalize_filename. */
6632 filename_is_absolute (char *fn
)
6634 return (fn
[0] == '/'
6636 || (ISALPHA(fn
[0]) && fn
[1] == ':' && fn
[2] == '/')
6641 /* Downcase DOS drive letter and collapse separators into single slashes.
6644 canonicalize_filename (register char *fn
)
6650 /* Canonicalize drive letter case. */
6651 # define ISUPPER(c) isupper (CHAR(c))
6652 if (fn
[0] != '\0' && fn
[1] == ':' && ISUPPER (fn
[0]))
6653 fn
[0] = lowcase (fn
[0]);
6658 /* Collapse multiple separators into a single slash. */
6659 for (cp
= fn
; *cp
!= '\0'; cp
++, fn
++)
6663 while (cp
[1] == sep
)
6672 /* Initialize a linebuffer for use. */
6674 linebuffer_init (linebuffer
*lbp
)
6676 lbp
->size
= (DEBUG
) ? 3 : 200;
6677 lbp
->buffer
= xnew (lbp
->size
, char);
6678 lbp
->buffer
[0] = '\0';
6682 /* Set the minimum size of a string contained in a linebuffer. */
6684 linebuffer_setlen (linebuffer
*lbp
, int toksize
)
6686 while (lbp
->size
<= toksize
)
6689 xrnew (lbp
->buffer
, lbp
->size
, char);
6694 /* Like malloc but get fatal error if memory is exhausted. */
6696 xmalloc (unsigned int size
)
6698 PTR result
= (PTR
) malloc (size
);
6700 fatal ("virtual memory exhausted", (char *)NULL
);
6705 xrealloc (char *ptr
, unsigned int size
)
6707 PTR result
= (PTR
) realloc (ptr
, size
);
6709 fatal ("virtual memory exhausted", (char *)NULL
);
6715 * indent-tabs-mode: t
6718 * c-font-lock-extra-types: ("FILE" "bool" "language" "linebuffer" "fdesc" "node" "regexp")
6719 * c-file-style: "gnu"
6723 /* etags.c ends here */