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>
144 # define MAXPATHLEN _MAX_PATH
150 # endif /* undef HAVE_GETCWD */
151 #else /* not WINDOWSNT */
152 #endif /* !WINDOWSNT */
155 #ifndef HAVE_UNISTD_H
156 # if defined (HAVE_GETCWD) && !defined (WINDOWSNT)
157 extern char *getcwd (char *buf
, size_t size
);
159 #endif /* HAVE_UNISTD_H */
166 #include <sys/types.h>
167 #include <sys/stat.h>
171 # undef assert /* some systems have a buggy assert.h */
172 # define assert(x) ((void) 0)
175 #ifdef NO_LONG_OPTIONS /* define this if you don't have GNU getopt */
176 # define NO_LONG_OPTIONS TRUE
177 # define getopt_long(argc,argv,optstr,lopts,lind) getopt (argc, argv, optstr)
179 extern int optind
, opterr
;
181 # define NO_LONG_OPTIONS FALSE
183 #endif /* NO_LONG_OPTIONS */
185 #ifndef HAVE_CONFIG_H /* this is a standalone compilation */
186 # ifdef __CYGWIN__ /* compiling on Cygwin */
188 the regex
.h distributed with Cygwin is
not compatible with etags
, alas
!
189 If you want regular expression support
, you should
delete this notice
and
190 arrange to use the GNU regex
.h
and regex
.c
.
195 /* Define CTAGS to make the program "ctags" compatible with the usual one.
196 Leave it undefined to make the program "etags", which makes emacs-style
197 tag tables and tags typedefs, #defines and struct/union/enum by default. */
205 #define streq(s,t) (assert((s)!=NULL || (t)!=NULL), !strcmp (s, t))
206 #define strcaseeq(s,t) (assert((s)!=NULL && (t)!=NULL), !etags_strcasecmp (s, t))
207 #define strneq(s,t,n) (assert((s)!=NULL || (t)!=NULL), !strncmp (s, t, n))
208 #define strncaseeq(s,t,n) (assert((s)!=NULL && (t)!=NULL), !etags_strncasecmp (s, t, n))
210 #define CHARS 256 /* 2^sizeof(char) */
211 #define CHAR(x) ((unsigned int)(x) & (CHARS - 1))
212 #define iswhite(c) (_wht[CHAR(c)]) /* c is white (see white) */
213 #define notinname(c) (_nin[CHAR(c)]) /* c is not in a name (see nonam) */
214 #define begtoken(c) (_btk[CHAR(c)]) /* c can start token (see begtk) */
215 #define intoken(c) (_itk[CHAR(c)]) /* c can be in token (see midtk) */
216 #define endtoken(c) (_etk[CHAR(c)]) /* c ends tokens (see endtk) */
218 #define ISALNUM(c) isalnum (CHAR(c))
219 #define ISALPHA(c) isalpha (CHAR(c))
220 #define ISDIGIT(c) isdigit (CHAR(c))
221 #define ISLOWER(c) islower (CHAR(c))
223 #define lowcase(c) tolower (CHAR(c))
227 * xnew, xrnew -- allocate, reallocate storage
229 * SYNOPSIS: Type *xnew (int n, Type);
230 * void xrnew (OldPointer, int n, Type);
233 # include "chkmalloc.h"
234 # define xnew(n,Type) ((Type *) trace_malloc (__FILE__, __LINE__, \
235 (n) * sizeof (Type)))
236 # define xrnew(op,n,Type) ((op) = (Type *) trace_realloc (__FILE__, __LINE__, \
237 (char *) (op), (n) * sizeof (Type)))
239 # define xnew(n,Type) ((Type *) xmalloc ((n) * sizeof (Type)))
240 # define xrnew(op,n,Type) ((op) = (Type *) xrealloc ( \
241 (char *) (op), (n) * sizeof (Type)))
246 typedef void Lang_function (FILE *);
250 const char *suffix
; /* file name suffix for this compressor */
251 const char *command
; /* takes one arg and decompresses to stdout */
256 const char *name
; /* language name */
257 const char *help
; /* detailed help for the language */
258 Lang_function
*function
; /* parse function */
259 const char **suffixes
; /* name suffixes of this language's files */
260 const char **filenames
; /* names of this language's files */
261 const char **interpreters
; /* interpreters for this language */
262 bool metasource
; /* source used to generate other sources */
267 struct fdesc
*next
; /* for the linked list */
268 char *infname
; /* uncompressed input file name */
269 char *infabsname
; /* absolute uncompressed input file name */
270 char *infabsdir
; /* absolute dir of input file */
271 char *taggedfname
; /* file name to write in tagfile */
272 language
*lang
; /* language of file */
273 char *prop
; /* file properties to write in tagfile */
274 bool usecharno
; /* etags tags shall contain char number */
275 bool written
; /* entry written in the tags file */
278 typedef struct node_st
279 { /* sorting structure */
280 struct node_st
*left
, *right
; /* left and right sons */
281 fdesc
*fdp
; /* description of file to whom tag belongs */
282 char *name
; /* tag name */
283 char *regex
; /* search regexp */
284 bool valid
; /* write this tag on the tag file */
285 bool is_func
; /* function tag: use regexp in CTAGS mode */
286 bool been_warned
; /* warning already given for duplicated tag */
287 int lno
; /* line number tag is on */
288 long cno
; /* character number line starts on */
292 * A `linebuffer' is a structure which holds a line of text.
293 * `readline_internal' reads a line from a stream into a linebuffer
294 * and works regardless of the length of the line.
295 * SIZE is the size of BUFFER, LEN is the length of the string in
296 * BUFFER after readline reads it.
305 /* Used to support mixing of --lang and file names. */
309 at_language
, /* a language specification */
310 at_regexp
, /* a regular expression */
311 at_filename
, /* a file name */
312 at_stdin
, /* read from stdin here */
313 at_end
/* stop parsing the list */
314 } arg_type
; /* argument type */
315 language
*lang
; /* language associated with the argument */
316 char *what
; /* the argument itself */
319 /* Structure defining a regular expression. */
320 typedef struct regexp
322 struct regexp
*p_next
; /* pointer to next in list */
323 language
*lang
; /* if set, use only for this language */
324 char *pattern
; /* the regexp pattern */
325 char *name
; /* tag name */
326 struct re_pattern_buffer
*pat
; /* the compiled pattern */
327 struct re_registers regs
; /* re registers */
328 bool error_signaled
; /* already signaled for this regexp */
329 bool force_explicit_name
; /* do not allow implict tag name */
330 bool ignore_case
; /* ignore case when matching */
331 bool multi_line
; /* do a multi-line match on the whole file */
335 /* Many compilers barf on this:
336 Lang_function Ada_funcs;
337 so let's write it this way */
338 static void Ada_funcs (FILE *);
339 static void Asm_labels (FILE *);
340 static void C_entries (int c_ext
, FILE *);
341 static void default_C_entries (FILE *);
342 static void plain_C_entries (FILE *);
343 static void Cjava_entries (FILE *);
344 static void Cobol_paragraphs (FILE *);
345 static void Cplusplus_entries (FILE *);
346 static void Cstar_entries (FILE *);
347 static void Erlang_functions (FILE *);
348 static void Forth_words (FILE *);
349 static void Fortran_functions (FILE *);
350 static void HTML_labels (FILE *);
351 static void Lisp_functions (FILE *);
352 static void Lua_functions (FILE *);
353 static void Makefile_targets (FILE *);
354 static void Pascal_functions (FILE *);
355 static void Perl_functions (FILE *);
356 static void PHP_functions (FILE *);
357 static void PS_functions (FILE *);
358 static void Prolog_functions (FILE *);
359 static void Python_functions (FILE *);
360 static void Scheme_functions (FILE *);
361 static void TeX_commands (FILE *);
362 static void Texinfo_nodes (FILE *);
363 static void Yacc_entries (FILE *);
364 static void just_read_file (FILE *);
366 static void print_language_names (void);
367 static void print_version (void);
368 static void print_help (argument
*);
369 int main (int, char **);
371 static compressor
*get_compressor_from_suffix (char *, char **);
372 static language
*get_language_from_langname (const char *);
373 static language
*get_language_from_interpreter (char *);
374 static language
*get_language_from_filename (char *, bool);
375 static void readline (linebuffer
*, FILE *);
376 static long readline_internal (linebuffer
*, FILE *);
377 static bool nocase_tail (const char *);
378 static void get_tag (char *, char **);
380 static void analyse_regex (char *);
381 static void free_regexps (void);
382 static void regex_tag_multiline (void);
383 static void error (const char *, const char *);
384 static void suggest_asking_for_help (void) NO_RETURN
;
385 void fatal (const char *, const char *) NO_RETURN
;
386 static void pfatal (const char *) NO_RETURN
;
387 static void add_node (node
*, node
**);
389 static void init (void);
390 static void process_file_name (char *, language
*);
391 static void process_file (FILE *, char *, language
*);
392 static void find_entries (FILE *);
393 static void free_tree (node
*);
394 static void free_fdesc (fdesc
*);
395 static void pfnote (char *, bool, char *, int, int, long);
396 static void make_tag (const char *, int, bool, char *, int, int, long);
397 static void invalidate_nodes (fdesc
*, node
**);
398 static void put_entries (node
*);
400 static char *concat (const char *, const char *, const char *);
401 static char *skip_spaces (char *);
402 static char *skip_non_spaces (char *);
403 static char *savenstr (const char *, int);
404 static char *savestr (const char *);
405 static char *etags_strchr (const char *, int);
406 static char *etags_strrchr (const char *, int);
407 static int etags_strcasecmp (const char *, const char *);
408 static int etags_strncasecmp (const char *, const char *, int);
409 static char *etags_getcwd (void);
410 static char *relative_filename (char *, char *);
411 static char *absolute_filename (char *, char *);
412 static char *absolute_dirname (char *, char *);
413 static bool filename_is_absolute (char *f
);
414 static void canonicalize_filename (char *);
415 static void linebuffer_init (linebuffer
*);
416 static void linebuffer_setlen (linebuffer
*, int);
417 static PTR
xmalloc (size_t);
418 static PTR
xrealloc (char *, size_t);
421 static char searchar
= '/'; /* use /.../ searches */
423 static char *tagfile
; /* output file */
424 static char *progname
; /* name this program was invoked with */
425 static char *cwd
; /* current working directory */
426 static char *tagfiledir
; /* directory of tagfile */
427 static FILE *tagf
; /* ioptr for tags file */
428 static ptrdiff_t whatlen_max
; /* maximum length of any 'what' member */
430 static fdesc
*fdhead
; /* head of file description list */
431 static fdesc
*curfdp
; /* current file description */
432 static int lineno
; /* line number of current line */
433 static long charno
; /* current character number */
434 static long linecharno
; /* charno of start of current line */
435 static char *dbp
; /* pointer to start of current tag */
437 static const int invalidcharno
= -1;
439 static node
*nodehead
; /* the head of the binary tree of tags */
440 static node
*last_node
; /* the last node created */
442 static linebuffer lb
; /* the current line */
443 static linebuffer filebuf
; /* a buffer containing the whole file */
444 static linebuffer token_name
; /* a buffer containing a tag name */
446 /* boolean "functions" (see init) */
447 static bool _wht
[CHARS
], _nin
[CHARS
], _itk
[CHARS
], _btk
[CHARS
], _etk
[CHARS
];
450 *white
= " \f\t\n\r\v",
452 *nonam
= " \f\t\n\r()=,;", /* look at make_tag before modifying! */
453 /* token ending chars */
454 *endtk
= " \t\n\r\"'#()[]{}=-+%*/&|^~!<>;,.:?",
455 /* token starting chars */
456 *begtk
= "ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz$~@",
457 /* valid in-token chars */
458 *midtk
= "ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz$0123456789";
460 static bool append_to_tagfile
; /* -a: append to tags */
461 /* The next five default to TRUE in C and derived languages. */
462 static bool typedefs
; /* -t: create tags for C and Ada typedefs */
463 static bool typedefs_or_cplusplus
; /* -T: create tags for C typedefs, level */
464 /* 0 struct/enum/union decls, and C++ */
465 /* member functions. */
466 static bool constantypedefs
; /* -d: create tags for C #define, enum */
467 /* constants and variables. */
468 /* -D: opposite of -d. Default under ctags. */
469 static bool globals
; /* create tags for global variables */
470 static bool members
; /* create tags for C member variables */
471 static bool declarations
; /* --declarations: tag them and extern in C&Co*/
472 static bool no_line_directive
; /* ignore #line directives (undocumented) */
473 static bool no_duplicates
; /* no duplicate tags for ctags (undocumented) */
474 static bool update
; /* -u: update tags */
475 static bool vgrind_style
; /* -v: create vgrind style index output */
476 static bool no_warnings
; /* -w: suppress warnings (undocumented) */
477 static bool cxref_style
; /* -x: create cxref style output */
478 static bool cplusplus
; /* .[hc] means C++, not C (undocumented) */
479 static bool ignoreindent
; /* -I: ignore indentation in C */
480 static bool packages_only
; /* --packages-only: in Ada, only tag packages*/
482 /* STDIN is defined in LynxOS system headers */
487 #define STDIN 0x1001 /* returned by getopt_long on --parse-stdin */
488 static bool parsing_stdin
; /* --parse-stdin used */
490 static regexp
*p_head
; /* list of all regexps */
491 static bool need_filebuf
; /* some regexes are multi-line */
493 static struct option longopts
[] =
495 { "append", no_argument
, NULL
, 'a' },
496 { "packages-only", no_argument
, &packages_only
, TRUE
},
497 { "c++", no_argument
, NULL
, 'C' },
498 { "declarations", no_argument
, &declarations
, TRUE
},
499 { "no-line-directive", no_argument
, &no_line_directive
, TRUE
},
500 { "no-duplicates", no_argument
, &no_duplicates
, TRUE
},
501 { "help", no_argument
, NULL
, 'h' },
502 { "help", no_argument
, NULL
, 'H' },
503 { "ignore-indentation", no_argument
, NULL
, 'I' },
504 { "language", required_argument
, NULL
, 'l' },
505 { "members", no_argument
, &members
, TRUE
},
506 { "no-members", no_argument
, &members
, FALSE
},
507 { "output", required_argument
, NULL
, 'o' },
508 { "regex", required_argument
, NULL
, 'r' },
509 { "no-regex", no_argument
, NULL
, 'R' },
510 { "ignore-case-regex", required_argument
, NULL
, 'c' },
511 { "parse-stdin", required_argument
, NULL
, STDIN
},
512 { "version", no_argument
, NULL
, 'V' },
514 #if CTAGS /* Ctags options */
515 { "backward-search", no_argument
, NULL
, 'B' },
516 { "cxref", no_argument
, NULL
, 'x' },
517 { "defines", no_argument
, NULL
, 'd' },
518 { "globals", no_argument
, &globals
, TRUE
},
519 { "typedefs", no_argument
, NULL
, 't' },
520 { "typedefs-and-c++", no_argument
, NULL
, 'T' },
521 { "update", no_argument
, NULL
, 'u' },
522 { "vgrind", no_argument
, NULL
, 'v' },
523 { "no-warn", no_argument
, NULL
, 'w' },
525 #else /* Etags options */
526 { "no-defines", no_argument
, NULL
, 'D' },
527 { "no-globals", no_argument
, &globals
, FALSE
},
528 { "include", required_argument
, NULL
, 'i' },
533 static compressor compressors
[] =
535 { "z", "gzip -d -c"},
536 { "Z", "gzip -d -c"},
537 { "gz", "gzip -d -c"},
538 { "GZ", "gzip -d -c"},
539 { "bz2", "bzip2 -d -c" },
540 { "xz", "xz -d -c" },
549 static const char *Ada_suffixes
[] =
550 { "ads", "adb", "ada", NULL
};
551 static const char Ada_help
[] =
552 "In Ada code, functions, procedures, packages, tasks and types are\n\
553 tags. Use the `--packages-only' option to create tags for\n\
555 Ada tag names have suffixes indicating the type of entity:\n\
556 Entity type: Qualifier:\n\
557 ------------ ----------\n\
564 Thus, `M-x find-tag <RET> bidule/b <RET>' will go directly to the\n\
565 body of the package `bidule', while `M-x find-tag <RET> bidule <RET>'\n\
566 will just search for any tag `bidule'.";
569 static const char *Asm_suffixes
[] =
570 { "a", /* Unix assembler */
571 "asm", /* Microcontroller assembly */
572 "def", /* BSO/Tasking definition includes */
573 "inc", /* Microcontroller include files */
574 "ins", /* Microcontroller include files */
575 "s", "sa", /* Unix assembler */
576 "S", /* cpp-processed Unix assembler */
577 "src", /* BSO/Tasking C compiler output */
580 static const char Asm_help
[] =
581 "In assembler code, labels appearing at the beginning of a line,\n\
582 followed by a colon, are tags.";
585 /* Note that .c and .h can be considered C++, if the --c++ flag was
586 given, or if the `class' or `template' keywords are met inside the file.
587 That is why default_C_entries is called for these. */
588 static const char *default_C_suffixes
[] =
590 #if CTAGS /* C help for Ctags */
591 static const char default_C_help
[] =
592 "In C code, any C function is a tag. Use -t to tag typedefs.\n\
593 Use -T to tag definitions of `struct', `union' and `enum'.\n\
594 Use -d to tag `#define' macro definitions and `enum' constants.\n\
595 Use --globals to tag global variables.\n\
596 You can tag function declarations and external variables by\n\
597 using `--declarations', and struct members by using `--members'.";
598 #else /* C help for Etags */
599 static const char default_C_help
[] =
600 "In C code, any C function or typedef is a tag, and so are\n\
601 definitions of `struct', `union' and `enum'. `#define' macro\n\
602 definitions and `enum' constants are tags unless you specify\n\
603 `--no-defines'. Global variables are tags unless you specify\n\
604 `--no-globals' and so are struct members unless you specify\n\
605 `--no-members'. Use of `--no-globals', `--no-defines' and\n\
606 `--no-members' can make the tags table file much smaller.\n\
607 You can tag function declarations and external variables by\n\
608 using `--declarations'.";
609 #endif /* C help for Ctags and Etags */
611 static const char *Cplusplus_suffixes
[] =
612 { "C", "c++", "cc", "cpp", "cxx", "H", "h++", "hh", "hpp", "hxx",
613 "M", /* Objective C++ */
614 "pdb", /* Postscript with C syntax */
616 static const char Cplusplus_help
[] =
617 "In C++ code, all the tag constructs of C code are tagged. (Use\n\
618 --help --lang=c --lang=c++ for full help.)\n\
619 In addition to C tags, member functions are also recognized. Member\n\
620 variables are recognized unless you use the `--no-members' option.\n\
621 Tags for variables and functions in classes are named `CLASS::VARIABLE'\n\
622 and `CLASS::FUNCTION'. `operator' definitions have tag names like\n\
625 static const char *Cjava_suffixes
[] =
627 static char Cjava_help
[] =
628 "In Java code, all the tags constructs of C and C++ code are\n\
629 tagged. (Use --help --lang=c --lang=c++ --lang=java for full help.)";
632 static const char *Cobol_suffixes
[] =
633 { "COB", "cob", NULL
};
634 static char Cobol_help
[] =
635 "In Cobol code, tags are paragraph names; that is, any word\n\
636 starting in column 8 and followed by a period.";
638 static const char *Cstar_suffixes
[] =
639 { "cs", "hs", NULL
};
641 static const char *Erlang_suffixes
[] =
642 { "erl", "hrl", NULL
};
643 static const char Erlang_help
[] =
644 "In Erlang code, the tags are the functions, records and macros\n\
645 defined in the file.";
647 const char *Forth_suffixes
[] =
648 { "fth", "tok", NULL
};
649 static const char Forth_help
[] =
650 "In Forth code, tags are words defined by `:',\n\
651 constant, code, create, defer, value, variable, buffer:, field.";
653 static const char *Fortran_suffixes
[] =
654 { "F", "f", "f90", "for", NULL
};
655 static const char Fortran_help
[] =
656 "In Fortran code, functions, subroutines and block data are tags.";
658 static const char *HTML_suffixes
[] =
659 { "htm", "html", "shtml", NULL
};
660 static const char HTML_help
[] =
661 "In HTML input files, the tags are the `title' and the `h1', `h2',\n\
662 `h3' headers. Also, tags are `name=' in anchors and all\n\
663 occurrences of `id='.";
665 static const char *Lisp_suffixes
[] =
666 { "cl", "clisp", "el", "l", "lisp", "LSP", "lsp", "ml", NULL
};
667 static const char Lisp_help
[] =
668 "In Lisp code, any function defined with `defun', any variable\n\
669 defined with `defvar' or `defconst', and in general the first\n\
670 argument of any expression that starts with `(def' in column zero\n\
673 static const char *Lua_suffixes
[] =
674 { "lua", "LUA", NULL
};
675 static const char Lua_help
[] =
676 "In Lua scripts, all functions are tags.";
678 static const char *Makefile_filenames
[] =
679 { "Makefile", "makefile", "GNUMakefile", "Makefile.in", "Makefile.am", NULL
};
680 static const char Makefile_help
[] =
681 "In makefiles, targets are tags; additionally, variables are tags\n\
682 unless you specify `--no-globals'.";
684 static const char *Objc_suffixes
[] =
685 { "lm", /* Objective lex file */
686 "m", /* Objective C file */
688 static const char Objc_help
[] =
689 "In Objective C code, tags include Objective C definitions for classes,\n\
690 class categories, methods and protocols. Tags for variables and\n\
691 functions in classes are named `CLASS::VARIABLE' and `CLASS::FUNCTION'.\n\
692 (Use --help --lang=c --lang=objc --lang=java for full help.)";
694 static const char *Pascal_suffixes
[] =
695 { "p", "pas", NULL
};
696 static const char Pascal_help
[] =
697 "In Pascal code, the tags are the functions and procedures defined\n\
699 /* " // this is for working around an Emacs highlighting bug... */
701 static const char *Perl_suffixes
[] =
702 { "pl", "pm", NULL
};
703 static const char *Perl_interpreters
[] =
704 { "perl", "@PERL@", NULL
};
705 static const char Perl_help
[] =
706 "In Perl code, the tags are the packages, subroutines and variables\n\
707 defined by the `package', `sub', `my' and `local' keywords. Use\n\
708 `--globals' if you want to tag global variables. Tags for\n\
709 subroutines are named `PACKAGE::SUB'. The name for subroutines\n\
710 defined in the default package is `main::SUB'.";
712 static const char *PHP_suffixes
[] =
713 { "php", "php3", "php4", NULL
};
714 static const char PHP_help
[] =
715 "In PHP code, tags are functions, classes and defines. Unless you use\n\
716 the `--no-members' option, vars are tags too.";
718 static const char *plain_C_suffixes
[] =
719 { "pc", /* Pro*C file */
722 static const char *PS_suffixes
[] =
723 { "ps", "psw", NULL
}; /* .psw is for PSWrap */
724 static const char PS_help
[] =
725 "In PostScript code, the tags are the functions.";
727 static const char *Prolog_suffixes
[] =
729 static const char Prolog_help
[] =
730 "In Prolog code, tags are predicates and rules at the beginning of\n\
733 static const char *Python_suffixes
[] =
735 static const char Python_help
[] =
736 "In Python code, `def' or `class' at the beginning of a line\n\
739 /* Can't do the `SCM' or `scm' prefix with a version number. */
740 static const char *Scheme_suffixes
[] =
741 { "oak", "sch", "scheme", "SCM", "scm", "SM", "sm", "ss", "t", NULL
};
742 static const char Scheme_help
[] =
743 "In Scheme code, tags include anything defined with `def' or with a\n\
744 construct whose name starts with `def'. They also include\n\
745 variables set with `set!' at top level in the file.";
747 static const char *TeX_suffixes
[] =
748 { "bib", "clo", "cls", "ltx", "sty", "TeX", "tex", NULL
};
749 static const char TeX_help
[] =
750 "In LaTeX text, the argument of any of the commands `\\chapter',\n\
751 `\\section', `\\subsection', `\\subsubsection', `\\eqno', `\\label',\n\
752 `\\ref', `\\cite', `\\bibitem', `\\part', `\\appendix', `\\entry',\n\
753 `\\index', `\\def', `\\newcommand', `\\renewcommand',\n\
754 `\\newenvironment' or `\\renewenvironment' is a tag.\n\
756 Other commands can be specified by setting the environment variable\n\
757 `TEXTAGS' to a colon-separated list like, for example,\n\
758 TEXTAGS=\"mycommand:myothercommand\".";
761 static const char *Texinfo_suffixes
[] =
762 { "texi", "texinfo", "txi", NULL
};
763 static const char Texinfo_help
[] =
764 "for texinfo files, lines starting with @node are tagged.";
766 static const char *Yacc_suffixes
[] =
767 { "y", "y++", "ym", "yxx", "yy", NULL
}; /* .ym is Objective yacc file */
768 static const char Yacc_help
[] =
769 "In Bison or Yacc input files, each rule defines as a tag the\n\
770 nonterminal it constructs. The portions of the file that contain\n\
771 C code are parsed as C code (use --help --lang=c --lang=yacc\n\
774 static const char auto_help
[] =
775 "`auto' is not a real language, it indicates to use\n\
776 a default language for files base on file name suffix and file contents.";
778 static const char none_help
[] =
779 "`none' is not a real language, it indicates to only do\n\
780 regexp processing on files.";
782 static const char no_lang_help
[] =
783 "No detailed help available for this language.";
787 * Table of languages.
789 * It is ok for a given function to be listed under more than one
790 * name. I just didn't.
793 static language lang_names
[] =
795 { "ada", Ada_help
, Ada_funcs
, Ada_suffixes
},
796 { "asm", Asm_help
, Asm_labels
, Asm_suffixes
},
797 { "c", default_C_help
, default_C_entries
, default_C_suffixes
},
798 { "c++", Cplusplus_help
, Cplusplus_entries
, Cplusplus_suffixes
},
799 { "c*", no_lang_help
, Cstar_entries
, Cstar_suffixes
},
800 { "cobol", Cobol_help
, Cobol_paragraphs
, Cobol_suffixes
},
801 { "erlang", Erlang_help
, Erlang_functions
, Erlang_suffixes
},
802 { "forth", Forth_help
, Forth_words
, Forth_suffixes
},
803 { "fortran", Fortran_help
, Fortran_functions
, Fortran_suffixes
},
804 { "html", HTML_help
, HTML_labels
, HTML_suffixes
},
805 { "java", Cjava_help
, Cjava_entries
, Cjava_suffixes
},
806 { "lisp", Lisp_help
, Lisp_functions
, Lisp_suffixes
},
807 { "lua", Lua_help
, Lua_functions
, Lua_suffixes
},
808 { "makefile", Makefile_help
,Makefile_targets
,NULL
,Makefile_filenames
},
809 { "objc", Objc_help
, plain_C_entries
, Objc_suffixes
},
810 { "pascal", Pascal_help
, Pascal_functions
, Pascal_suffixes
},
811 { "perl",Perl_help
,Perl_functions
,Perl_suffixes
,NULL
,Perl_interpreters
},
812 { "php", PHP_help
, PHP_functions
, PHP_suffixes
},
813 { "postscript",PS_help
, PS_functions
, PS_suffixes
},
814 { "proc", no_lang_help
, plain_C_entries
, plain_C_suffixes
},
815 { "prolog", Prolog_help
, Prolog_functions
, Prolog_suffixes
},
816 { "python", Python_help
, Python_functions
, Python_suffixes
},
817 { "scheme", Scheme_help
, Scheme_functions
, Scheme_suffixes
},
818 { "tex", TeX_help
, TeX_commands
, TeX_suffixes
},
819 { "texinfo", Texinfo_help
, Texinfo_nodes
, Texinfo_suffixes
},
820 { "yacc", Yacc_help
,Yacc_entries
,Yacc_suffixes
,NULL
,NULL
,TRUE
},
821 { "auto", auto_help
}, /* default guessing scheme */
822 { "none", none_help
, just_read_file
}, /* regexp matching only */
823 { NULL
} /* end of list */
828 print_language_names (void)
831 const char **name
, **ext
;
833 puts ("\nThese are the currently supported languages, along with the\n\
834 default file names and dot suffixes:");
835 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
837 printf (" %-*s", 10, lang
->name
);
838 if (lang
->filenames
!= NULL
)
839 for (name
= lang
->filenames
; *name
!= NULL
; name
++)
840 printf (" %s", *name
);
841 if (lang
->suffixes
!= NULL
)
842 for (ext
= lang
->suffixes
; *ext
!= NULL
; ext
++)
843 printf (" .%s", *ext
);
846 puts ("where `auto' means use default language for files based on file\n\
847 name suffix, and `none' means only do regexp processing on files.\n\
848 If no language is specified and no matching suffix is found,\n\
849 the first line of the file is read for a sharp-bang (#!) sequence\n\
850 followed by the name of an interpreter. If no such sequence is found,\n\
851 Fortran is tried first; if no tags are found, C is tried next.\n\
852 When parsing any C file, a \"class\" or \"template\" keyword\n\
854 puts ("Compressed files are supported using gzip, bzip2, and xz.\n\
856 For detailed help on a given language use, for example,\n\
857 etags --help --lang=ada.");
861 # define EMACS_NAME "standalone"
864 # define VERSION "17.38.1.4"
869 /* Makes it easier to update automatically. */
870 char emacs_copyright
[] = "Copyright (C) 2011 Free Software Foundation, Inc.";
872 printf ("%s (%s %s)\n", (CTAGS
) ? "ctags" : "etags", EMACS_NAME
, VERSION
);
873 puts (emacs_copyright
);
874 puts ("This program is distributed under the terms in ETAGS.README");
879 #ifndef PRINT_UNDOCUMENTED_OPTIONS_HELP
880 # define PRINT_UNDOCUMENTED_OPTIONS_HELP FALSE
884 print_help (argument
*argbuffer
)
886 bool help_for_lang
= FALSE
;
888 for (; argbuffer
->arg_type
!= at_end
; argbuffer
++)
889 if (argbuffer
->arg_type
== at_language
)
893 puts (argbuffer
->lang
->help
);
894 help_for_lang
= TRUE
;
900 printf ("Usage: %s [options] [[regex-option ...] file-name] ...\n\
902 These are the options accepted by %s.\n", progname
, progname
);
904 puts ("WARNING: long option names do not work with this executable,\n\
905 as it is not linked with GNU getopt.");
907 puts ("You may use unambiguous abbreviations for the long option names.");
908 puts (" A - as file name means read names from stdin (one per line).\n\
909 Absolute names are stored in the output file as they are.\n\
910 Relative ones are stored relative to the output file's directory.\n");
912 puts ("-a, --append\n\
913 Append tag entries to existing tags file.");
915 puts ("--packages-only\n\
916 For Ada files, only generate tags for packages.");
919 puts ("-B, --backward-search\n\
920 Write the search commands for the tag entries using '?', the\n\
921 backward-search command instead of '/', the forward-search command.");
923 /* This option is mostly obsolete, because etags can now automatically
924 detect C++. Retained for backward compatibility and for debugging and
925 experimentation. In principle, we could want to tag as C++ even
926 before any "class" or "template" keyword.
928 Treat files whose name suffix defaults to C language as C++ files.");
931 puts ("--declarations\n\
932 In C and derived languages, create tags for function declarations,");
934 puts ("\tand create tags for extern variables if --globals is used.");
937 ("\tand create tags for extern variables unless --no-globals is used.");
940 puts ("-d, --defines\n\
941 Create tag entries for C #define constants and enum constants, too.");
943 puts ("-D, --no-defines\n\
944 Don't create tag entries for C #define constants and enum constants.\n\
945 This makes the tags file smaller.");
948 puts ("-i FILE, --include=FILE\n\
949 Include a note in tag file indicating that, when searching for\n\
950 a tag, one should also consult the tags file FILE after\n\
951 checking the current file.");
953 puts ("-l LANG, --language=LANG\n\
954 Force the following files to be considered as written in the\n\
955 named language up to the next --language=LANG option.");
959 Create tag entries for global variables in some languages.");
961 puts ("--no-globals\n\
962 Do not create tag entries for global variables in some\n\
963 languages. This makes the tags file smaller.");
965 if (PRINT_UNDOCUMENTED_OPTIONS_HELP
)
966 puts ("--no-line-directive\n\
967 Ignore #line preprocessor directives in C and derived languages.");
971 Create tag entries for members of structures in some languages.");
973 puts ("--no-members\n\
974 Do not create tag entries for members of structures\n\
975 in some languages.");
977 puts ("-r REGEXP, --regex=REGEXP or --regex=@regexfile\n\
978 Make a tag for each line matching a regular expression pattern\n\
979 in the following files. {LANGUAGE}REGEXP uses REGEXP for LANGUAGE\n\
980 files only. REGEXFILE is a file containing one REGEXP per line.\n\
981 REGEXP takes the form /TAGREGEXP/TAGNAME/MODS, where TAGNAME/ is\n\
982 optional. The TAGREGEXP pattern is anchored (as if preceded by ^).");
983 puts (" If TAGNAME/ is present, the tags created are named.\n\
984 For example Tcl named tags can be created with:\n\
985 --regex=\"/proc[ \\t]+\\([^ \\t]+\\)/\\1/.\".\n\
986 MODS are optional one-letter modifiers: `i' means to ignore case,\n\
987 `m' means to allow multi-line matches, `s' implies `m' and\n\
988 causes dot to match any character, including newline.");
990 puts ("-R, --no-regex\n\
991 Don't create tags from regexps for the following files.");
993 puts ("-I, --ignore-indentation\n\
994 In C and C++ do not assume that a closing brace in the first\n\
995 column is the final brace of a function or structure definition.");
997 puts ("-o FILE, --output=FILE\n\
998 Write the tags to FILE.");
1000 puts ("--parse-stdin=NAME\n\
1001 Read from standard input and record tags as belonging to file NAME.");
1005 puts ("-t, --typedefs\n\
1006 Generate tag entries for C and Ada typedefs.");
1007 puts ("-T, --typedefs-and-c++\n\
1008 Generate tag entries for C typedefs, C struct/enum/union tags,\n\
1009 and C++ member functions.");
1013 puts ("-u, --update\n\
1014 Update the tag entries for the given files, leaving tag\n\
1015 entries for other files in place. Currently, this is\n\
1016 implemented by deleting the existing entries for the given\n\
1017 files and then rewriting the new entries at the end of the\n\
1018 tags file. It is often faster to simply rebuild the entire\n\
1019 tag file than to use this.");
1023 puts ("-v, --vgrind\n\
1024 Print on the standard output an index of items intended for\n\
1025 human consumption, similar to the output of vgrind. The index\n\
1026 is sorted, and gives the page number of each item.");
1028 if (PRINT_UNDOCUMENTED_OPTIONS_HELP
)
1029 puts ("-w, --no-duplicates\n\
1030 Do not create duplicate tag entries, for compatibility with\n\
1031 traditional ctags.");
1033 if (PRINT_UNDOCUMENTED_OPTIONS_HELP
)
1034 puts ("-w, --no-warn\n\
1035 Suppress warning messages about duplicate tag entries.");
1037 puts ("-x, --cxref\n\
1038 Like --vgrind, but in the style of cxref, rather than vgrind.\n\
1039 The output uses line numbers instead of page numbers, but\n\
1040 beyond that the differences are cosmetic; try both to see\n\
1044 puts ("-V, --version\n\
1045 Print the version of the program.\n\
1047 Print this help message.\n\
1048 Followed by one or more `--language' options prints detailed\n\
1049 help about tag generation for the specified languages.");
1051 print_language_names ();
1054 puts ("Report bugs to bug-gnu-emacs@gnu.org");
1056 exit (EXIT_SUCCESS
);
1061 main (int argc
, char **argv
)
1064 unsigned int nincluded_files
;
1065 char **included_files
;
1066 argument
*argbuffer
;
1067 int current_arg
, file_count
;
1068 linebuffer filename_lb
;
1069 bool help_asked
= FALSE
;
1076 _fmode
= O_BINARY
; /* all of files are treated as binary files */
1080 nincluded_files
= 0;
1081 included_files
= xnew (argc
, char *);
1085 /* Allocate enough no matter what happens. Overkill, but each one
1087 argbuffer
= xnew (argc
, argument
);
1090 * Always find typedefs and structure tags.
1091 * Also default to find macro constants, enum constants, struct
1092 * members and global variables. Do it for both etags and ctags.
1094 typedefs
= typedefs_or_cplusplus
= constantypedefs
= TRUE
;
1095 globals
= members
= TRUE
;
1097 /* When the optstring begins with a '-' getopt_long does not rearrange the
1098 non-options arguments to be at the end, but leaves them alone. */
1099 optstring
= concat (NO_LONG_OPTIONS
? "" : "-",
1100 "ac:Cf:Il:o:r:RSVhH",
1101 (CTAGS
) ? "BxdtTuvw" : "Di:");
1103 while ((opt
= getopt_long (argc
, argv
, optstring
, longopts
, NULL
)) != EOF
)
1107 /* If getopt returns 0, then it has already processed a
1108 long-named option. We should do nothing. */
1112 /* This means that a file name has been seen. Record it. */
1113 argbuffer
[current_arg
].arg_type
= at_filename
;
1114 argbuffer
[current_arg
].what
= optarg
;
1115 len
= strlen (optarg
);
1116 if (whatlen_max
< len
)
1123 /* Parse standard input. Idea by Vivek <vivek@etla.org>. */
1124 argbuffer
[current_arg
].arg_type
= at_stdin
;
1125 argbuffer
[current_arg
].what
= optarg
;
1126 len
= strlen (optarg
);
1127 if (whatlen_max
< len
)
1132 fatal ("cannot parse standard input more than once", (char *)NULL
);
1133 parsing_stdin
= TRUE
;
1136 /* Common options. */
1137 case 'a': append_to_tagfile
= TRUE
; break;
1138 case 'C': cplusplus
= TRUE
; break;
1139 case 'f': /* for compatibility with old makefiles */
1143 error ("-o option may only be given once.", (char *)NULL
);
1144 suggest_asking_for_help ();
1150 case 'S': /* for backward compatibility */
1151 ignoreindent
= TRUE
;
1155 language
*lang
= get_language_from_langname (optarg
);
1158 argbuffer
[current_arg
].lang
= lang
;
1159 argbuffer
[current_arg
].arg_type
= at_language
;
1165 /* Backward compatibility: support obsolete --ignore-case-regexp. */
1166 optarg
= concat (optarg
, "i", ""); /* memory leak here */
1169 argbuffer
[current_arg
].arg_type
= at_regexp
;
1170 argbuffer
[current_arg
].what
= optarg
;
1171 len
= strlen (optarg
);
1172 if (whatlen_max
< len
)
1177 argbuffer
[current_arg
].arg_type
= at_regexp
;
1178 argbuffer
[current_arg
].what
= NULL
;
1190 case 'D': constantypedefs
= FALSE
; break;
1191 case 'i': included_files
[nincluded_files
++] = optarg
; break;
1193 /* Ctags options. */
1194 case 'B': searchar
= '?'; break;
1195 case 'd': constantypedefs
= TRUE
; break;
1196 case 't': typedefs
= TRUE
; break;
1197 case 'T': typedefs
= typedefs_or_cplusplus
= TRUE
; break;
1198 case 'u': update
= TRUE
; break;
1199 case 'v': vgrind_style
= TRUE
; /*FALLTHRU*/
1200 case 'x': cxref_style
= TRUE
; break;
1201 case 'w': no_warnings
= TRUE
; break;
1203 suggest_asking_for_help ();
1207 /* No more options. Store the rest of arguments. */
1208 for (; optind
< argc
; optind
++)
1210 argbuffer
[current_arg
].arg_type
= at_filename
;
1211 argbuffer
[current_arg
].what
= argv
[optind
];
1212 len
= strlen (argv
[optind
]);
1213 if (whatlen_max
< len
)
1219 argbuffer
[current_arg
].arg_type
= at_end
;
1222 print_help (argbuffer
);
1225 if (nincluded_files
== 0 && file_count
== 0)
1227 error ("no input files specified.", (char *)NULL
);
1228 suggest_asking_for_help ();
1232 if (tagfile
== NULL
)
1233 tagfile
= savestr (CTAGS
? "tags" : "TAGS");
1234 cwd
= etags_getcwd (); /* the current working directory */
1235 if (cwd
[strlen (cwd
) - 1] != '/')
1238 cwd
= concat (oldcwd
, "/", "");
1242 /* Compute base directory for relative file names. */
1243 if (streq (tagfile
, "-")
1244 || strneq (tagfile
, "/dev/", 5))
1245 tagfiledir
= cwd
; /* relative file names are relative to cwd */
1248 canonicalize_filename (tagfile
);
1249 tagfiledir
= absolute_dirname (tagfile
, cwd
);
1252 init (); /* set up boolean "functions" */
1254 linebuffer_init (&lb
);
1255 linebuffer_init (&filename_lb
);
1256 linebuffer_init (&filebuf
);
1257 linebuffer_init (&token_name
);
1261 if (streq (tagfile
, "-"))
1265 /* Switch redirected `stdout' to binary mode (setting `_fmode'
1266 doesn't take effect until after `stdout' is already open). */
1267 if (!isatty (fileno (stdout
)))
1268 setmode (fileno (stdout
), O_BINARY
);
1272 tagf
= fopen (tagfile
, append_to_tagfile
? "a" : "w");
1278 * Loop through files finding functions.
1280 for (i
= 0; i
< current_arg
; i
++)
1282 static language
*lang
; /* non-NULL if language is forced */
1285 switch (argbuffer
[i
].arg_type
)
1288 lang
= argbuffer
[i
].lang
;
1291 analyse_regex (argbuffer
[i
].what
);
1294 this_file
= argbuffer
[i
].what
;
1295 /* Input file named "-" means read file names from stdin
1296 (one per line) and use them. */
1297 if (streq (this_file
, "-"))
1300 fatal ("cannot parse standard input AND read file names from it",
1302 while (readline_internal (&filename_lb
, stdin
) > 0)
1303 process_file_name (filename_lb
.buffer
, lang
);
1306 process_file_name (this_file
, lang
);
1309 this_file
= argbuffer
[i
].what
;
1310 process_file (stdin
, this_file
, lang
);
1317 free (filebuf
.buffer
);
1318 free (token_name
.buffer
);
1320 if (!CTAGS
|| cxref_style
)
1322 /* Write the remaining tags to tagf (ETAGS) or stdout (CXREF). */
1323 put_entries (nodehead
);
1324 free_tree (nodehead
);
1330 /* Output file entries that have no tags. */
1331 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
1333 fprintf (tagf
, "\f\n%s,0\n", fdp
->taggedfname
);
1335 while (nincluded_files
-- > 0)
1336 fprintf (tagf
, "\f\n%s,include\n", *included_files
++);
1338 if (fclose (tagf
) == EOF
)
1342 exit (EXIT_SUCCESS
);
1345 /* From here on, we are in (CTAGS && !cxref_style) */
1349 xmalloc (strlen (tagfile
) + whatlen_max
+
1350 sizeof "mv..OTAGS;fgrep -v '\t\t' OTAGS >;rm OTAGS");
1351 for (i
= 0; i
< current_arg
; ++i
)
1353 switch (argbuffer
[i
].arg_type
)
1359 continue; /* the for loop */
1361 strcpy (cmd
, "mv ");
1362 strcat (cmd
, tagfile
);
1363 strcat (cmd
, " OTAGS;fgrep -v '\t");
1364 strcat (cmd
, argbuffer
[i
].what
);
1365 strcat (cmd
, "\t' OTAGS >");
1366 strcat (cmd
, tagfile
);
1367 strcat (cmd
, ";rm OTAGS");
1368 if (system (cmd
) != EXIT_SUCCESS
)
1369 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
= xmalloc (2 * strlen (tagfile
) + sizeof "sort -u -o..");
1388 /* Maybe these should be used:
1389 setenv ("LC_COLLATE", "C", 1);
1390 setenv ("LC_ALL", "C", 1); */
1391 strcpy (cmd
, "sort -u -o ");
1392 strcat (cmd
, tagfile
);
1394 strcat (cmd
, tagfile
);
1395 exit (system (cmd
));
1397 return EXIT_SUCCESS
;
1402 * Return a compressor given the file name. If EXTPTR is non-zero,
1403 * return a pointer into FILE where the compressor-specific
1404 * extension begins. If no compressor is found, NULL is returned
1405 * and EXTPTR is not significant.
1406 * Idea by Vladimir Alexiev <vladimir@cs.ualberta.ca> (1998)
1409 get_compressor_from_suffix (char *file
, char **extptr
)
1412 char *slash
, *suffix
;
1414 /* File has been processed by canonicalize_filename,
1415 so we don't need to consider backslashes on DOS_NT. */
1416 slash
= etags_strrchr (file
, '/');
1417 suffix
= etags_strrchr (file
, '.');
1418 if (suffix
== NULL
|| suffix
< slash
)
1423 /* Let those poor souls who live with DOS 8+3 file name limits get
1424 some solace by treating foo.cgz as if it were foo.c.gz, etc.
1425 Only the first do loop is run if not MSDOS */
1428 for (compr
= compressors
; compr
->suffix
!= NULL
; compr
++)
1429 if (streq (compr
->suffix
, suffix
))
1432 break; /* do it only once: not really a loop */
1435 } while (*suffix
!= '\0');
1442 * Return a language given the name.
1445 get_language_from_langname (const char *name
)
1450 error ("empty language name", (char *)NULL
);
1453 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1454 if (streq (name
, lang
->name
))
1456 error ("unknown language \"%s\"", name
);
1464 * Return a language given the interpreter name.
1467 get_language_from_interpreter (char *interpreter
)
1472 if (interpreter
== NULL
)
1474 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1475 if (lang
->interpreters
!= NULL
)
1476 for (iname
= lang
->interpreters
; *iname
!= NULL
; iname
++)
1477 if (streq (*iname
, interpreter
))
1486 * Return a language given the file name.
1489 get_language_from_filename (char *file
, int case_sensitive
)
1492 const char **name
, **ext
, *suffix
;
1494 /* Try whole file name first. */
1495 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1496 if (lang
->filenames
!= NULL
)
1497 for (name
= lang
->filenames
; *name
!= NULL
; name
++)
1498 if ((case_sensitive
)
1499 ? streq (*name
, file
)
1500 : strcaseeq (*name
, file
))
1503 /* If not found, try suffix after last dot. */
1504 suffix
= etags_strrchr (file
, '.');
1508 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1509 if (lang
->suffixes
!= NULL
)
1510 for (ext
= lang
->suffixes
; *ext
!= NULL
; ext
++)
1511 if ((case_sensitive
)
1512 ? streq (*ext
, suffix
)
1513 : strcaseeq (*ext
, suffix
))
1520 * This routine is called on each file argument.
1523 process_file_name (char *file
, language
*lang
)
1525 struct stat stat_buf
;
1529 char *compressed_name
, *uncompressed_name
;
1530 char *ext
, *real_name
;
1533 canonicalize_filename (file
);
1534 if (streq (file
, tagfile
) && !streq (tagfile
, "-"))
1536 error ("skipping inclusion of %s in self.", file
);
1539 if ((compr
= get_compressor_from_suffix (file
, &ext
)) == NULL
)
1541 compressed_name
= NULL
;
1542 real_name
= uncompressed_name
= savestr (file
);
1546 real_name
= compressed_name
= savestr (file
);
1547 uncompressed_name
= savenstr (file
, ext
- file
);
1550 /* If the canonicalized uncompressed name
1551 has already been dealt with, skip it silently. */
1552 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
1554 assert (fdp
->infname
!= NULL
);
1555 if (streq (uncompressed_name
, fdp
->infname
))
1559 if (stat (real_name
, &stat_buf
) != 0)
1561 /* Reset real_name and try with a different name. */
1563 if (compressed_name
!= NULL
) /* try with the given suffix */
1565 if (stat (uncompressed_name
, &stat_buf
) == 0)
1566 real_name
= uncompressed_name
;
1568 else /* try all possible suffixes */
1570 for (compr
= compressors
; compr
->suffix
!= NULL
; compr
++)
1572 compressed_name
= concat (file
, ".", compr
->suffix
);
1573 if (stat (compressed_name
, &stat_buf
) != 0)
1577 char *suf
= compressed_name
+ strlen (file
);
1578 size_t suflen
= strlen (compr
->suffix
) + 1;
1579 for ( ; suf
[1]; suf
++, suflen
--)
1581 memmove (suf
, suf
+ 1, suflen
);
1582 if (stat (compressed_name
, &stat_buf
) == 0)
1584 real_name
= compressed_name
;
1588 if (real_name
!= NULL
)
1591 free (compressed_name
);
1592 compressed_name
= NULL
;
1596 real_name
= compressed_name
;
1601 if (real_name
== NULL
)
1606 } /* try with a different name */
1608 if (!S_ISREG (stat_buf
.st_mode
))
1610 error ("skipping %s: it is not a regular file.", real_name
);
1613 if (real_name
== compressed_name
)
1615 char *cmd
= concat (compr
->command
, " ", real_name
);
1616 inf
= (FILE *) popen (cmd
, "r");
1620 inf
= fopen (real_name
, "r");
1627 process_file (inf
, uncompressed_name
, lang
);
1629 if (real_name
== compressed_name
)
1630 retval
= pclose (inf
);
1632 retval
= fclose (inf
);
1637 free (compressed_name
);
1638 free (uncompressed_name
);
1645 process_file (FILE *fh
, char *fn
, language
*lang
)
1647 static const fdesc emptyfdesc
;
1650 /* Create a new input file description entry. */
1651 fdp
= xnew (1, fdesc
);
1654 fdp
->infname
= savestr (fn
);
1656 fdp
->infabsname
= absolute_filename (fn
, cwd
);
1657 fdp
->infabsdir
= absolute_dirname (fn
, cwd
);
1658 if (filename_is_absolute (fn
))
1660 /* An absolute file name. Canonicalize it. */
1661 fdp
->taggedfname
= absolute_filename (fn
, NULL
);
1665 /* A file name relative to cwd. Make it relative
1666 to the directory of the tags file. */
1667 fdp
->taggedfname
= relative_filename (fn
, tagfiledir
);
1669 fdp
->usecharno
= TRUE
; /* use char position when making tags */
1671 fdp
->written
= FALSE
; /* not written on tags file yet */
1674 curfdp
= fdhead
; /* the current file description */
1678 /* If not Ctags, and if this is not metasource and if it contained no #line
1679 directives, we can write the tags and free all nodes pointing to
1682 && curfdp
->usecharno
/* no #line directives in this file */
1683 && !curfdp
->lang
->metasource
)
1687 /* Look for the head of the sublist relative to this file. See add_node
1688 for the structure of the node tree. */
1690 for (np
= nodehead
; np
!= NULL
; prev
= np
, np
= np
->left
)
1691 if (np
->fdp
== curfdp
)
1694 /* If we generated tags for this file, write and delete them. */
1697 /* This is the head of the last sublist, if any. The following
1698 instructions depend on this being true. */
1699 assert (np
->left
== NULL
);
1701 assert (fdhead
== curfdp
);
1702 assert (last_node
->fdp
== curfdp
);
1703 put_entries (np
); /* write tags for file curfdp->taggedfname */
1704 free_tree (np
); /* remove the written nodes */
1706 nodehead
= NULL
; /* no nodes left */
1708 prev
->left
= NULL
; /* delete the pointer to the sublist */
1714 * This routine sets up the boolean pseudo-functions which work
1715 * by setting boolean flags dependent upon the corresponding character.
1716 * Every char which is NOT in that string is not a white char. Therefore,
1717 * all of the array "_wht" is set to FALSE, and then the elements
1718 * subscripted by the chars in "white" are set to TRUE. Thus "_wht"
1719 * of a char is TRUE if it is the string "white", else FALSE.
1724 register const char *sp
;
1727 for (i
= 0; i
< CHARS
; i
++)
1728 iswhite(i
) = notinname(i
) = begtoken(i
) = intoken(i
) = endtoken(i
) = FALSE
;
1729 for (sp
= white
; *sp
!= '\0'; sp
++) iswhite (*sp
) = TRUE
;
1730 for (sp
= nonam
; *sp
!= '\0'; sp
++) notinname (*sp
) = TRUE
;
1731 notinname('\0') = notinname('\n');
1732 for (sp
= begtk
; *sp
!= '\0'; sp
++) begtoken (*sp
) = TRUE
;
1733 begtoken('\0') = begtoken('\n');
1734 for (sp
= midtk
; *sp
!= '\0'; sp
++) intoken (*sp
) = TRUE
;
1735 intoken('\0') = intoken('\n');
1736 for (sp
= endtk
; *sp
!= '\0'; sp
++) endtoken (*sp
) = TRUE
;
1737 endtoken('\0') = endtoken('\n');
1741 * This routine opens the specified file and calls the function
1742 * which finds the function and type definitions.
1745 find_entries (FILE *inf
)
1748 language
*lang
= curfdp
->lang
;
1749 Lang_function
*parser
= NULL
;
1751 /* If user specified a language, use it. */
1752 if (lang
!= NULL
&& lang
->function
!= NULL
)
1754 parser
= lang
->function
;
1757 /* Else try to guess the language given the file name. */
1760 lang
= get_language_from_filename (curfdp
->infname
, TRUE
);
1761 if (lang
!= NULL
&& lang
->function
!= NULL
)
1763 curfdp
->lang
= lang
;
1764 parser
= lang
->function
;
1768 /* Else look for sharp-bang as the first two characters. */
1770 && readline_internal (&lb
, inf
) > 0
1772 && lb
.buffer
[0] == '#'
1773 && lb
.buffer
[1] == '!')
1777 /* Set lp to point at the first char after the last slash in the
1778 line or, if no slashes, at the first nonblank. Then set cp to
1779 the first successive blank and terminate the string. */
1780 lp
= etags_strrchr (lb
.buffer
+2, '/');
1784 lp
= skip_spaces (lb
.buffer
+ 2);
1785 cp
= skip_non_spaces (lp
);
1788 if (strlen (lp
) > 0)
1790 lang
= get_language_from_interpreter (lp
);
1791 if (lang
!= NULL
&& lang
->function
!= NULL
)
1793 curfdp
->lang
= lang
;
1794 parser
= lang
->function
;
1799 /* We rewind here, even if inf may be a pipe. We fail if the
1800 length of the first line is longer than the pipe block size,
1801 which is unlikely. */
1804 /* Else try to guess the language given the case insensitive file name. */
1807 lang
= get_language_from_filename (curfdp
->infname
, FALSE
);
1808 if (lang
!= NULL
&& lang
->function
!= NULL
)
1810 curfdp
->lang
= lang
;
1811 parser
= lang
->function
;
1815 /* Else try Fortran or C. */
1818 node
*old_last_node
= last_node
;
1820 curfdp
->lang
= get_language_from_langname ("fortran");
1823 if (old_last_node
== last_node
)
1824 /* No Fortran entries found. Try C. */
1826 /* We do not tag if rewind fails.
1827 Only the file name will be recorded in the tags file. */
1829 curfdp
->lang
= get_language_from_langname (cplusplus
? "c++" : "c");
1835 if (!no_line_directive
1836 && curfdp
->lang
!= NULL
&& curfdp
->lang
->metasource
)
1837 /* It may be that this is a bingo.y file, and we already parsed a bingo.c
1838 file, or anyway we parsed a file that is automatically generated from
1839 this one. If this is the case, the bingo.c file contained #line
1840 directives that generated tags pointing to this file. Let's delete
1841 them all before parsing this file, which is the real source. */
1843 fdesc
**fdpp
= &fdhead
;
1844 while (*fdpp
!= NULL
)
1846 && streq ((*fdpp
)->taggedfname
, curfdp
->taggedfname
))
1847 /* We found one of those! We must delete both the file description
1848 and all tags referring to it. */
1850 fdesc
*badfdp
= *fdpp
;
1852 /* Delete the tags referring to badfdp->taggedfname
1853 that were obtained from badfdp->infname. */
1854 invalidate_nodes (badfdp
, &nodehead
);
1856 *fdpp
= badfdp
->next
; /* remove the bad description from the list */
1857 free_fdesc (badfdp
);
1860 fdpp
= &(*fdpp
)->next
; /* advance the list pointer */
1863 assert (parser
!= NULL
);
1865 /* Generic initialisations before reading from file. */
1866 linebuffer_setlen (&filebuf
, 0); /* reset the file buffer */
1868 /* Generic initialisations before parsing file with readline. */
1869 lineno
= 0; /* reset global line number */
1870 charno
= 0; /* reset global char number */
1871 linecharno
= 0; /* reset global char number of line start */
1875 regex_tag_multiline ();
1880 * Check whether an implicitly named tag should be created,
1881 * then call `pfnote'.
1882 * NAME is a string that is internally copied by this function.
1884 * TAGS format specification
1885 * Idea by Sam Kendall <kendall@mv.mv.com> (1997)
1886 * The following is explained in some more detail in etc/ETAGS.EBNF.
1888 * make_tag creates tags with "implicit tag names" (unnamed tags)
1889 * if the following are all true, assuming NONAM=" \f\t\n\r()=,;":
1890 * 1. NAME does not contain any of the characters in NONAM;
1891 * 2. LINESTART contains name as either a rightmost, or rightmost but
1892 * one character, substring;
1893 * 3. the character, if any, immediately before NAME in LINESTART must
1894 * be a character in NONAM;
1895 * 4. the character, if any, immediately after NAME in LINESTART must
1896 * also be a character in NONAM.
1898 * The implementation uses the notinname() macro, which recognises the
1899 * characters stored in the string `nonam'.
1900 * etags.el needs to use the same characters that are in NONAM.
1903 make_tag (const char *name
, /* tag name, or NULL if unnamed */
1904 int namelen
, /* tag length */
1905 int is_func
, /* tag is a function */
1906 char *linestart
, /* start of the line where tag is */
1907 int linelen
, /* length of the line where tag is */
1908 int lno
, /* line number */
1909 long int cno
) /* character number */
1911 bool named
= (name
!= NULL
&& namelen
> 0);
1914 if (!CTAGS
&& named
) /* maybe set named to false */
1915 /* Let's try to make an implicit tag name, that is, create an unnamed tag
1916 such that etags.el can guess a name from it. */
1919 register const char *cp
= name
;
1921 for (i
= 0; i
< namelen
; i
++)
1922 if (notinname (*cp
++))
1924 if (i
== namelen
) /* rule #1 */
1926 cp
= linestart
+ linelen
- namelen
;
1927 if (notinname (linestart
[linelen
-1]))
1928 cp
-= 1; /* rule #4 */
1929 if (cp
>= linestart
/* rule #2 */
1931 || notinname (cp
[-1])) /* rule #3 */
1932 && strneq (name
, cp
, namelen
)) /* rule #2 */
1933 named
= FALSE
; /* use implicit tag name */
1938 nname
= savenstr (name
, namelen
);
1940 pfnote (nname
, is_func
, linestart
, linelen
, lno
, cno
);
1945 pfnote (char *name
, int is_func
, char *linestart
, int linelen
, int lno
, long int cno
)
1946 /* tag name, or NULL if unnamed */
1947 /* tag is a function */
1948 /* start of the line where tag is */
1949 /* length of the line where tag is */
1951 /* character number */
1955 assert (name
== NULL
|| name
[0] != '\0');
1956 if (CTAGS
&& name
== NULL
)
1959 np
= xnew (1, node
);
1961 /* If ctags mode, change name "main" to M<thisfilename>. */
1962 if (CTAGS
&& !cxref_style
&& streq (name
, "main"))
1964 register char *fp
= etags_strrchr (curfdp
->taggedfname
, '/');
1965 np
->name
= concat ("M", fp
== NULL
? curfdp
->taggedfname
: fp
+ 1, "");
1966 fp
= etags_strrchr (np
->name
, '.');
1967 if (fp
!= NULL
&& fp
[1] != '\0' && fp
[2] == '\0')
1973 np
->been_warned
= FALSE
;
1975 np
->is_func
= is_func
;
1977 if (np
->fdp
->usecharno
)
1978 /* Our char numbers are 0-base, because of C language tradition?
1979 ctags compatibility? old versions compatibility? I don't know.
1980 Anyway, since emacs's are 1-base we expect etags.el to take care
1981 of the difference. If we wanted to have 1-based numbers, we would
1982 uncomment the +1 below. */
1983 np
->cno
= cno
/* + 1 */ ;
1985 np
->cno
= invalidcharno
;
1986 np
->left
= np
->right
= NULL
;
1987 if (CTAGS
&& !cxref_style
)
1989 if (strlen (linestart
) < 50)
1990 np
->regex
= concat (linestart
, "$", "");
1992 np
->regex
= savenstr (linestart
, 50);
1995 np
->regex
= savenstr (linestart
, linelen
);
1997 add_node (np
, &nodehead
);
2002 * recurse on left children, iterate on right children.
2005 free_tree (register node
*np
)
2009 register node
*node_right
= np
->right
;
2010 free_tree (np
->left
);
2020 * delete a file description
2023 free_fdesc (register fdesc
*fdp
)
2025 free (fdp
->infname
);
2026 free (fdp
->infabsname
);
2027 free (fdp
->infabsdir
);
2028 free (fdp
->taggedfname
);
2035 * Adds a node to the tree of nodes. In etags mode, sort by file
2036 * name. In ctags mode, sort by tag name. Make no attempt at
2039 * add_node is the only function allowed to add nodes, so it can
2043 add_node (node
*np
, node
**cur_node_p
)
2046 register node
*cur_node
= *cur_node_p
;
2048 if (cur_node
== NULL
)
2058 /* For each file name, tags are in a linked sublist on the right
2059 pointer. The first tags of different files are a linked list
2060 on the left pointer. last_node points to the end of the last
2062 if (last_node
!= NULL
&& last_node
->fdp
== np
->fdp
)
2064 /* Let's use the same sublist as the last added node. */
2065 assert (last_node
->right
== NULL
);
2066 last_node
->right
= np
;
2069 else if (cur_node
->fdp
== np
->fdp
)
2071 /* Scanning the list we found the head of a sublist which is
2072 good for us. Let's scan this sublist. */
2073 add_node (np
, &cur_node
->right
);
2076 /* The head of this sublist is not good for us. Let's try the
2078 add_node (np
, &cur_node
->left
);
2079 } /* if ETAGS mode */
2084 dif
= strcmp (np
->name
, cur_node
->name
);
2087 * If this tag name matches an existing one, then
2088 * do not add the node, but maybe print a warning.
2090 if (no_duplicates
&& !dif
)
2092 if (np
->fdp
== cur_node
->fdp
)
2096 fprintf (stderr
, "Duplicate entry in file %s, line %d: %s\n",
2097 np
->fdp
->infname
, lineno
, np
->name
);
2098 fprintf (stderr
, "Second entry ignored\n");
2101 else if (!cur_node
->been_warned
&& !no_warnings
)
2105 "Duplicate entry in files %s and %s: %s (Warning only)\n",
2106 np
->fdp
->infname
, cur_node
->fdp
->infname
, np
->name
);
2107 cur_node
->been_warned
= TRUE
;
2112 /* Actually add the node */
2113 add_node (np
, dif
< 0 ? &cur_node
->left
: &cur_node
->right
);
2114 } /* if CTAGS mode */
2118 * invalidate_nodes ()
2119 * Scan the node tree and invalidate all nodes pointing to the
2120 * given file description (CTAGS case) or free them (ETAGS case).
2123 invalidate_nodes (fdesc
*badfdp
, node
**npp
)
2132 if (np
->left
!= NULL
)
2133 invalidate_nodes (badfdp
, &np
->left
);
2134 if (np
->fdp
== badfdp
)
2136 if (np
->right
!= NULL
)
2137 invalidate_nodes (badfdp
, &np
->right
);
2141 assert (np
->fdp
!= NULL
);
2142 if (np
->fdp
== badfdp
)
2144 *npp
= np
->left
; /* detach the sublist from the list */
2145 np
->left
= NULL
; /* isolate it */
2146 free_tree (np
); /* free it */
2147 invalidate_nodes (badfdp
, npp
);
2150 invalidate_nodes (badfdp
, &np
->left
);
2155 static int total_size_of_entries (node
*);
2156 static int number_len (long);
2158 /* Length of a non-negative number's decimal representation. */
2160 number_len (long int num
)
2163 while ((num
/= 10) > 0)
2169 * Return total number of characters that put_entries will output for
2170 * the nodes in the linked list at the right of the specified node.
2171 * This count is irrelevant with etags.el since emacs 19.34 at least,
2172 * but is still supplied for backward compatibility.
2175 total_size_of_entries (register node
*np
)
2177 register int total
= 0;
2179 for (; np
!= NULL
; np
= np
->right
)
2182 total
+= strlen (np
->regex
) + 1; /* pat\177 */
2183 if (np
->name
!= NULL
)
2184 total
+= strlen (np
->name
) + 1; /* name\001 */
2185 total
+= number_len ((long) np
->lno
) + 1; /* lno, */
2186 if (np
->cno
!= invalidcharno
) /* cno */
2187 total
+= number_len (np
->cno
);
2188 total
+= 1; /* newline */
2195 put_entries (register node
*np
)
2198 static fdesc
*fdp
= NULL
;
2203 /* Output subentries that precede this one */
2205 put_entries (np
->left
);
2207 /* Output this entry */
2216 fprintf (tagf
, "\f\n%s,%d\n",
2217 fdp
->taggedfname
, total_size_of_entries (np
));
2218 fdp
->written
= TRUE
;
2220 fputs (np
->regex
, tagf
);
2221 fputc ('\177', tagf
);
2222 if (np
->name
!= NULL
)
2224 fputs (np
->name
, tagf
);
2225 fputc ('\001', tagf
);
2227 fprintf (tagf
, "%d,", np
->lno
);
2228 if (np
->cno
!= invalidcharno
)
2229 fprintf (tagf
, "%ld", np
->cno
);
2235 if (np
->name
== NULL
)
2236 error ("internal error: NULL name in ctags mode.", (char *)NULL
);
2241 fprintf (stdout
, "%s %s %d\n",
2242 np
->name
, np
->fdp
->taggedfname
, (np
->lno
+ 63) / 64);
2244 fprintf (stdout
, "%-16s %3d %-16s %s\n",
2245 np
->name
, np
->lno
, np
->fdp
->taggedfname
, np
->regex
);
2249 fprintf (tagf
, "%s\t%s\t", np
->name
, np
->fdp
->taggedfname
);
2252 { /* function or #define macro with args */
2253 putc (searchar
, tagf
);
2256 for (sp
= np
->regex
; *sp
; sp
++)
2258 if (*sp
== '\\' || *sp
== searchar
)
2262 putc (searchar
, tagf
);
2265 { /* anything else; text pattern inadequate */
2266 fprintf (tagf
, "%d", np
->lno
);
2271 } /* if this node contains a valid tag */
2273 /* Output subentries that follow this one */
2274 put_entries (np
->right
);
2276 put_entries (np
->left
);
2281 #define C_EXT 0x00fff /* C extensions */
2282 #define C_PLAIN 0x00000 /* C */
2283 #define C_PLPL 0x00001 /* C++ */
2284 #define C_STAR 0x00003 /* C* */
2285 #define C_JAVA 0x00005 /* JAVA */
2286 #define C_AUTO 0x01000 /* C, but switch to C++ if `class' is met */
2287 #define YACC 0x10000 /* yacc file */
2290 * The C symbol tables.
2295 st_C_objprot
, st_C_objimpl
, st_C_objend
,
2297 st_C_ignore
, st_C_attribute
,
2300 st_C_class
, st_C_template
,
2301 st_C_struct
, st_C_extern
, st_C_enum
, st_C_define
, st_C_typedef
2304 static unsigned int hash (const char *, unsigned int);
2305 static struct C_stab_entry
* in_word_set (const char *, unsigned int);
2306 static enum sym_type
C_symtype (char *, int, int);
2308 /* Feed stuff between (but not including) %[ and %] lines to:
2314 struct C_stab_entry { char *name; int c_ext; enum sym_type type; }
2318 while, 0, st_C_ignore
2319 switch, 0, st_C_ignore
2320 return, 0, st_C_ignore
2321 __attribute__, 0, st_C_attribute
2322 GTY, 0, st_C_attribute
2323 @interface, 0, st_C_objprot
2324 @protocol, 0, st_C_objprot
2325 @implementation,0, st_C_objimpl
2326 @end, 0, st_C_objend
2327 import, (C_JAVA & ~C_PLPL), st_C_ignore
2328 package, (C_JAVA & ~C_PLPL), st_C_ignore
2329 friend, C_PLPL, st_C_ignore
2330 extends, (C_JAVA & ~C_PLPL), st_C_javastruct
2331 implements, (C_JAVA & ~C_PLPL), st_C_javastruct
2332 interface, (C_JAVA & ~C_PLPL), st_C_struct
2333 class, 0, st_C_class
2334 namespace, C_PLPL, st_C_struct
2335 domain, C_STAR, st_C_struct
2336 union, 0, st_C_struct
2337 struct, 0, st_C_struct
2338 extern, 0, st_C_extern
2340 typedef, 0, st_C_typedef
2341 define, 0, st_C_define
2342 undef, 0, st_C_define
2343 operator, C_PLPL, st_C_operator
2344 template, 0, st_C_template
2345 # DEFUN used in emacs, the next three used in glibc (SYSCALL only for mach).
2346 DEFUN, 0, st_C_gnumacro
2347 SYSCALL, 0, st_C_gnumacro
2348 ENTRY, 0, st_C_gnumacro
2349 PSEUDO, 0, st_C_gnumacro
2350 # These are defined inside C functions, so currently they are not met.
2351 # EXFUN used in glibc, DEFVAR_* in emacs.
2352 #EXFUN, 0, st_C_gnumacro
2353 #DEFVAR_, 0, st_C_gnumacro
2355 and replace lines between %< and %> with its output, then:
2356 - remove the #if characterset check
2357 - make in_word_set static and not inline. */
2359 /* C code produced by gperf version 3.0.1 */
2360 /* Command-line: gperf -m 5 */
2361 /* Computed positions: -k'2-3' */
2363 struct C_stab_entry
{ const char *name
; int c_ext
; enum sym_type type
; };
2364 /* maximum key range = 33, duplicates = 0 */
2366 static inline unsigned int
2367 hash (register const char *str
, register unsigned int len
)
2369 static unsigned char asso_values
[] =
2371 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2372 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2373 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2374 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2375 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2376 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2377 35, 35, 35, 35, 35, 35, 35, 35, 35, 3,
2378 26, 35, 35, 35, 35, 35, 35, 35, 27, 35,
2379 35, 35, 35, 24, 0, 35, 35, 35, 35, 0,
2380 35, 35, 35, 35, 35, 1, 35, 16, 35, 6,
2381 23, 0, 0, 35, 22, 0, 35, 35, 5, 0,
2382 0, 15, 1, 35, 6, 35, 8, 19, 35, 16,
2383 4, 5, 35, 35, 35, 35, 35, 35, 35, 35,
2384 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2385 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2386 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2387 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2388 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2389 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2390 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2391 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2392 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2393 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2394 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2395 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2396 35, 35, 35, 35, 35, 35
2398 register int hval
= len
;
2403 hval
+= asso_values
[(unsigned char)str
[2]];
2406 hval
+= asso_values
[(unsigned char)str
[1]];
2412 static struct C_stab_entry
*
2413 in_word_set (register const char *str
, register unsigned int len
)
2417 TOTAL_KEYWORDS
= 33,
2418 MIN_WORD_LENGTH
= 2,
2419 MAX_WORD_LENGTH
= 15,
2424 static struct C_stab_entry wordlist
[] =
2427 {"if", 0, st_C_ignore
},
2428 {"GTY", 0, st_C_attribute
},
2429 {"@end", 0, st_C_objend
},
2430 {"union", 0, st_C_struct
},
2431 {"define", 0, st_C_define
},
2432 {"import", (C_JAVA
& ~C_PLPL
), st_C_ignore
},
2433 {"template", 0, st_C_template
},
2434 {"operator", C_PLPL
, st_C_operator
},
2435 {"@interface", 0, st_C_objprot
},
2436 {"implements", (C_JAVA
& ~C_PLPL
), st_C_javastruct
},
2437 {"friend", C_PLPL
, st_C_ignore
},
2438 {"typedef", 0, st_C_typedef
},
2439 {"return", 0, st_C_ignore
},
2440 {"@implementation",0, st_C_objimpl
},
2441 {"@protocol", 0, st_C_objprot
},
2442 {"interface", (C_JAVA
& ~C_PLPL
), st_C_struct
},
2443 {"extern", 0, st_C_extern
},
2444 {"extends", (C_JAVA
& ~C_PLPL
), st_C_javastruct
},
2445 {"struct", 0, st_C_struct
},
2446 {"domain", C_STAR
, st_C_struct
},
2447 {"switch", 0, st_C_ignore
},
2448 {"enum", 0, st_C_enum
},
2449 {"for", 0, st_C_ignore
},
2450 {"namespace", C_PLPL
, st_C_struct
},
2451 {"class", 0, st_C_class
},
2452 {"while", 0, st_C_ignore
},
2453 {"undef", 0, st_C_define
},
2454 {"package", (C_JAVA
& ~C_PLPL
), st_C_ignore
},
2455 {"__attribute__", 0, st_C_attribute
},
2456 {"SYSCALL", 0, st_C_gnumacro
},
2457 {"ENTRY", 0, st_C_gnumacro
},
2458 {"PSEUDO", 0, st_C_gnumacro
},
2459 {"DEFUN", 0, st_C_gnumacro
}
2462 if (len
<= MAX_WORD_LENGTH
&& len
>= MIN_WORD_LENGTH
)
2464 register int key
= hash (str
, len
);
2466 if (key
<= MAX_HASH_VALUE
&& key
>= 0)
2468 register const char *s
= wordlist
[key
].name
;
2470 if (*str
== *s
&& !strncmp (str
+ 1, s
+ 1, len
- 1) && s
[len
] == '\0')
2471 return &wordlist
[key
];
2478 static enum sym_type
2479 C_symtype (char *str
, int len
, int c_ext
)
2481 register struct C_stab_entry
*se
= in_word_set (str
, len
);
2483 if (se
== NULL
|| (se
->c_ext
&& !(c_ext
& se
->c_ext
)))
2490 * Ignoring __attribute__ ((list))
2492 static bool inattribute
; /* looking at an __attribute__ construct */
2495 * C functions and variables are recognized using a simple
2496 * finite automaton. fvdef is its state variable.
2500 fvnone
, /* nothing seen */
2501 fdefunkey
, /* Emacs DEFUN keyword seen */
2502 fdefunname
, /* Emacs DEFUN name seen */
2503 foperator
, /* func: operator keyword seen (cplpl) */
2504 fvnameseen
, /* function or variable name seen */
2505 fstartlist
, /* func: just after open parenthesis */
2506 finlist
, /* func: in parameter list */
2507 flistseen
, /* func: after parameter list */
2508 fignore
, /* func: before open brace */
2509 vignore
/* var-like: ignore until ';' */
2512 static bool fvextern
; /* func or var: extern keyword seen; */
2515 * typedefs are recognized using a simple finite automaton.
2516 * typdef is its state variable.
2520 tnone
, /* nothing seen */
2521 tkeyseen
, /* typedef keyword seen */
2522 ttypeseen
, /* defined type seen */
2523 tinbody
, /* inside typedef body */
2524 tend
, /* just before typedef tag */
2525 tignore
/* junk after typedef tag */
2529 * struct-like structures (enum, struct and union) are recognized
2530 * using another simple finite automaton. `structdef' is its state
2535 snone
, /* nothing seen yet,
2536 or in struct body if bracelev > 0 */
2537 skeyseen
, /* struct-like keyword seen */
2538 stagseen
, /* struct-like tag seen */
2539 scolonseen
/* colon seen after struct-like tag */
2543 * When objdef is different from onone, objtag is the name of the class.
2545 static const char *objtag
= "<uninited>";
2548 * Yet another little state machine to deal with preprocessor lines.
2552 dnone
, /* nothing seen */
2553 dsharpseen
, /* '#' seen as first char on line */
2554 ddefineseen
, /* '#' and 'define' seen */
2555 dignorerest
/* ignore rest of line */
2559 * State machine for Objective C protocols and implementations.
2560 * Idea by Tom R.Hageman <tom@basil.icce.rug.nl> (1995)
2564 onone
, /* nothing seen */
2565 oprotocol
, /* @interface or @protocol seen */
2566 oimplementation
, /* @implementations seen */
2567 otagseen
, /* class name seen */
2568 oparenseen
, /* parenthesis before category seen */
2569 ocatseen
, /* category name seen */
2570 oinbody
, /* in @implementation body */
2571 omethodsign
, /* in @implementation body, after +/- */
2572 omethodtag
, /* after method name */
2573 omethodcolon
, /* after method colon */
2574 omethodparm
, /* after method parameter */
2575 oignore
/* wait for @end */
2580 * Use this structure to keep info about the token read, and how it
2581 * should be tagged. Used by the make_C_tag function to build a tag.
2585 char *line
; /* string containing the token */
2586 int offset
; /* where the token starts in LINE */
2587 int length
; /* token length */
2589 The previous members can be used to pass strings around for generic
2590 purposes. The following ones specifically refer to creating tags. In this
2591 case the token contained here is the pattern that will be used to create a
2594 bool valid
; /* do not create a tag; the token should be
2595 invalidated whenever a state machine is
2596 reset prematurely */
2597 bool named
; /* create a named tag */
2598 int lineno
; /* source line number of tag */
2599 long linepos
; /* source char number of tag */
2600 } token
; /* latest token read */
2603 * Variables and functions for dealing with nested structures.
2604 * Idea by Mykola Dzyuba <mdzyuba@yahoo.com> (2001)
2606 static void pushclass_above (int, char *, int);
2607 static void popclass_above (int);
2608 static void write_classname (linebuffer
*, const char *qualifier
);
2611 char **cname
; /* nested class names */
2612 int *bracelev
; /* nested class brace level */
2613 int nl
; /* class nesting level (elements used) */
2614 int size
; /* length of the array */
2615 } cstack
; /* stack for nested declaration tags */
2616 /* Current struct nesting depth (namespace, class, struct, union, enum). */
2617 #define nestlev (cstack.nl)
2618 /* After struct keyword or in struct body, not inside a nested function. */
2619 #define instruct (structdef == snone && nestlev > 0 \
2620 && bracelev == cstack.bracelev[nestlev-1] + 1)
2623 pushclass_above (int bracelev
, char *str
, int len
)
2627 popclass_above (bracelev
);
2629 if (nl
>= cstack
.size
)
2631 int size
= cstack
.size
*= 2;
2632 xrnew (cstack
.cname
, size
, char *);
2633 xrnew (cstack
.bracelev
, size
, int);
2635 assert (nl
== 0 || cstack
.bracelev
[nl
-1] < bracelev
);
2636 cstack
.cname
[nl
] = (str
== NULL
) ? NULL
: savenstr (str
, len
);
2637 cstack
.bracelev
[nl
] = bracelev
;
2642 popclass_above (int bracelev
)
2646 for (nl
= cstack
.nl
- 1;
2647 nl
>= 0 && cstack
.bracelev
[nl
] >= bracelev
;
2650 free (cstack
.cname
[nl
]);
2656 write_classname (linebuffer
*cn
, const char *qualifier
)
2659 int qlen
= strlen (qualifier
);
2661 if (cstack
.nl
== 0 || cstack
.cname
[0] == NULL
)
2665 cn
->buffer
[0] = '\0';
2669 len
= strlen (cstack
.cname
[0]);
2670 linebuffer_setlen (cn
, len
);
2671 strcpy (cn
->buffer
, cstack
.cname
[0]);
2673 for (i
= 1; i
< cstack
.nl
; i
++)
2678 s
= cstack
.cname
[i
];
2683 linebuffer_setlen (cn
, len
);
2684 strncat (cn
->buffer
, qualifier
, qlen
);
2685 strncat (cn
->buffer
, s
, slen
);
2690 static bool consider_token (char *, int, int, int *, int, int, bool *);
2691 static void make_C_tag (bool);
2695 * checks to see if the current token is at the start of a
2696 * function or variable, or corresponds to a typedef, or
2697 * is a struct/union/enum tag, or #define, or an enum constant.
2699 * *IS_FUNC gets TRUE if the token is a function or #define macro
2700 * with args. C_EXTP points to which language we are looking at.
2711 consider_token (register char *str
, register int len
, register int c
, int *c_extp
, int bracelev
, int parlev
, int *is_func_or_var
)
2712 /* IN: token pointer */
2713 /* IN: token length */
2714 /* IN: first char after the token */
2715 /* IN, OUT: C extensions mask */
2716 /* IN: brace level */
2717 /* IN: parenthesis level */
2718 /* OUT: function or variable found */
2720 /* When structdef is stagseen, scolonseen, or snone with bracelev > 0,
2721 structtype is the type of the preceding struct-like keyword, and
2722 structbracelev is the brace level where it has been seen. */
2723 static enum sym_type structtype
;
2724 static int structbracelev
;
2725 static enum sym_type toktype
;
2728 toktype
= C_symtype (str
, len
, *c_extp
);
2731 * Skip __attribute__
2733 if (toktype
== st_C_attribute
)
2740 * Advance the definedef state machine.
2745 /* We're not on a preprocessor line. */
2746 if (toktype
== st_C_gnumacro
)
2753 if (toktype
== st_C_define
)
2755 definedef
= ddefineseen
;
2759 definedef
= dignorerest
;
2764 * Make a tag for any macro, unless it is a constant
2765 * and constantypedefs is FALSE.
2767 definedef
= dignorerest
;
2768 *is_func_or_var
= (c
== '(');
2769 if (!*is_func_or_var
&& !constantypedefs
)
2776 error ("internal error: definedef value.", (char *)NULL
);
2785 if (toktype
== st_C_typedef
)
2805 if (structdef
== snone
&& fvdef
== fvnone
)
2824 case st_C_javastruct
:
2825 if (structdef
== stagseen
)
2826 structdef
= scolonseen
;
2830 if ((*c_extp
& C_AUTO
) /* automatic detection of C++ language */
2832 && definedef
== dnone
&& structdef
== snone
2833 && typdef
== tnone
&& fvdef
== fvnone
)
2834 *c_extp
= (*c_extp
| C_PLPL
) & ~C_AUTO
;
2835 if (toktype
== st_C_template
)
2842 && (typdef
== tkeyseen
2843 || (typedefs_or_cplusplus
&& structdef
== snone
)))
2845 structdef
= skeyseen
;
2846 structtype
= toktype
;
2847 structbracelev
= bracelev
;
2848 if (fvdef
== fvnameseen
)
2854 if (structdef
== skeyseen
)
2856 structdef
= stagseen
;
2860 if (typdef
!= tnone
)
2863 /* Detect Objective C constructs. */
2873 objdef
= oimplementation
;
2877 case oimplementation
:
2878 /* Save the class tag for functions or variables defined inside. */
2879 objtag
= savenstr (str
, len
);
2883 /* Save the class tag for categories. */
2884 objtag
= savenstr (str
, len
);
2886 *is_func_or_var
= TRUE
;
2890 *is_func_or_var
= TRUE
;
2898 objdef
= omethodtag
;
2899 linebuffer_setlen (&token_name
, len
);
2900 strncpy (token_name
.buffer
, str
, len
);
2901 token_name
.buffer
[len
] = '\0';
2907 objdef
= omethodparm
;
2913 objdef
= omethodtag
;
2914 linebuffer_setlen (&token_name
, token_name
.len
+ len
);
2915 strncat (token_name
.buffer
, str
, len
);
2920 if (toktype
== st_C_objend
)
2922 /* Memory leakage here: the string pointed by objtag is
2923 never released, because many tests would be needed to
2924 avoid breaking on incorrect input code. The amount of
2925 memory leaked here is the sum of the lengths of the
2933 /* A function, variable or enum constant? */
2955 *is_func_or_var
= TRUE
;
2959 && structdef
== snone
2960 && structtype
== st_C_enum
&& bracelev
> structbracelev
)
2961 return TRUE
; /* enum constant */
2967 fvdef
= fdefunname
; /* GNU macro */
2968 *is_func_or_var
= TRUE
;
2976 if ((strneq (str
, "asm", 3) && endtoken (str
[3]))
2977 || (strneq (str
, "__asm__", 7) && endtoken (str
[7])))
2986 if (len
>= 10 && strneq (str
+len
-10, "::operator", 10))
2988 if (*c_extp
& C_AUTO
) /* automatic detection of C++ */
2989 *c_extp
= (*c_extp
| C_PLPL
) & ~C_AUTO
;
2991 *is_func_or_var
= TRUE
;
2994 if (bracelev
> 0 && !instruct
)
2996 fvdef
= fvnameseen
; /* function or variable */
2997 *is_func_or_var
= TRUE
;
3008 * C_entries often keeps pointers to tokens or lines which are older than
3009 * the line currently read. By keeping two line buffers, and switching
3010 * them at end of line, it is possible to use those pointers.
3018 #define current_lb_is_new (newndx == curndx)
3019 #define switch_line_buffers() (curndx = 1 - curndx)
3021 #define curlb (lbs[curndx].lb)
3022 #define newlb (lbs[newndx].lb)
3023 #define curlinepos (lbs[curndx].linepos)
3024 #define newlinepos (lbs[newndx].linepos)
3026 #define plainc ((c_ext & C_EXT) == C_PLAIN)
3027 #define cplpl (c_ext & C_PLPL)
3028 #define cjava ((c_ext & C_JAVA) == C_JAVA)
3030 #define CNL_SAVE_DEFINEDEF() \
3032 curlinepos = charno; \
3033 readline (&curlb, inf); \
3034 lp = curlb.buffer; \
3041 CNL_SAVE_DEFINEDEF(); \
3042 if (savetoken.valid) \
3044 token = savetoken; \
3045 savetoken.valid = FALSE; \
3047 definedef = dnone; \
3052 make_C_tag (int isfun
)
3054 /* This function is never called when token.valid is FALSE, but
3055 we must protect against invalid input or internal errors. */
3057 make_tag (token_name
.buffer
, token_name
.len
, isfun
, token
.line
,
3058 token
.offset
+token
.length
+1, token
.lineno
, token
.linepos
);
3060 { /* this branch is optimised away if !DEBUG */
3061 make_tag (concat ("INVALID TOKEN:-->", token_name
.buffer
, ""),
3062 token_name
.len
+ 17, isfun
, token
.line
,
3063 token
.offset
+token
.length
+1, token
.lineno
, token
.linepos
);
3064 error ("INVALID TOKEN", NULL
);
3067 token
.valid
= FALSE
;
3073 * This routine finds functions, variables, typedefs,
3074 * #define's, enum constants and struct/union/enum definitions in
3075 * C syntax and adds them to the list.
3078 C_entries (int c_ext
, FILE *inf
)
3079 /* extension of C */
3082 register char c
; /* latest char read; '\0' for end of line */
3083 register char *lp
; /* pointer one beyond the character `c' */
3084 int curndx
, newndx
; /* indices for current and new lb */
3085 register int tokoff
; /* offset in line of start of current token */
3086 register int toklen
; /* length of current token */
3087 const char *qualifier
; /* string used to qualify names */
3088 int qlen
; /* length of qualifier */
3089 int bracelev
; /* current brace level */
3090 int bracketlev
; /* current bracket level */
3091 int parlev
; /* current parenthesis level */
3092 int attrparlev
; /* __attribute__ parenthesis level */
3093 int templatelev
; /* current template level */
3094 int typdefbracelev
; /* bracelev where a typedef struct body begun */
3095 bool incomm
, inquote
, inchar
, quotednl
, midtoken
;
3096 bool yacc_rules
; /* in the rules part of a yacc file */
3097 struct tok savetoken
= {0}; /* token saved during preprocessor handling */
3100 linebuffer_init (&lbs
[0].lb
);
3101 linebuffer_init (&lbs
[1].lb
);
3102 if (cstack
.size
== 0)
3104 cstack
.size
= (DEBUG
) ? 1 : 4;
3106 cstack
.cname
= xnew (cstack
.size
, char *);
3107 cstack
.bracelev
= xnew (cstack
.size
, int);
3110 tokoff
= toklen
= typdefbracelev
= 0; /* keep compiler quiet */
3111 curndx
= newndx
= 0;
3115 fvdef
= fvnone
; fvextern
= FALSE
; typdef
= tnone
;
3116 structdef
= snone
; definedef
= dnone
; objdef
= onone
;
3118 midtoken
= inquote
= inchar
= incomm
= quotednl
= FALSE
;
3119 token
.valid
= savetoken
.valid
= FALSE
;
3120 bracelev
= bracketlev
= parlev
= attrparlev
= templatelev
= 0;
3122 { qualifier
= "."; qlen
= 1; }
3124 { qualifier
= "::"; qlen
= 2; }
3132 /* If we are at the end of the line, the next character is a
3133 '\0'; do not skip it, because it is what tells us
3134 to read the next line. */
3155 /* Newlines inside comments do not end macro definitions in
3157 CNL_SAVE_DEFINEDEF ();
3170 /* Newlines inside strings do not end macro definitions
3171 in traditional cpp, even though compilers don't
3172 usually accept them. */
3173 CNL_SAVE_DEFINEDEF ();
3183 /* Hmmm, something went wrong. */
3192 else if (bracketlev
> 0)
3197 if (--bracketlev
> 0)
3201 CNL_SAVE_DEFINEDEF ();
3229 if (fvdef
!= finlist
&& fvdef
!= fignore
&& fvdef
!=vignore
)
3242 else if (/* cplpl && */ *lp
== '/')
3248 if ((c_ext
& YACC
) && *lp
== '%')
3250 /* Entering or exiting rules section in yacc file. */
3252 definedef
= dnone
; fvdef
= fvnone
; fvextern
= FALSE
;
3253 typdef
= tnone
; structdef
= snone
;
3254 midtoken
= inquote
= inchar
= incomm
= quotednl
= FALSE
;
3256 yacc_rules
= !yacc_rules
;
3262 if (definedef
== dnone
)
3265 bool cpptoken
= TRUE
;
3267 /* Look back on this line. If all blanks, or nonblanks
3268 followed by an end of comment, this is a preprocessor
3270 for (cp
= newlb
.buffer
; cp
< lp
-1; cp
++)
3273 if (*cp
== '*' && *(cp
+1) == '/')
3282 definedef
= dsharpseen
;
3283 } /* if (definedef == dnone) */
3291 /* Consider token only if some involved conditions are satisfied. */
3292 if (typdef
!= tignore
3293 && definedef
!= dignorerest
3296 && (definedef
!= dnone
3297 || structdef
!= scolonseen
)
3304 if (c
== ':' && *lp
== ':' && begtoken (lp
[1]))
3305 /* This handles :: in the middle,
3306 but not at the beginning of an identifier.
3307 Also, space-separated :: is not recognised. */
3309 if (c_ext
& C_AUTO
) /* automatic detection of C++ */
3310 c_ext
= (c_ext
| C_PLPL
) & ~C_AUTO
;
3314 goto still_in_token
;
3318 bool funorvar
= FALSE
;
3321 || consider_token (newlb
.buffer
+ tokoff
, toklen
, c
,
3322 &c_ext
, bracelev
, parlev
,
3325 if (fvdef
== foperator
)
3328 lp
= skip_spaces (lp
-1);
3332 && !iswhite (*lp
) && *lp
!= '(')
3335 toklen
+= lp
- oldlp
;
3337 token
.named
= FALSE
;
3339 && nestlev
> 0 && definedef
== dnone
)
3340 /* in struct body */
3342 write_classname (&token_name
, qualifier
);
3343 linebuffer_setlen (&token_name
,
3344 token_name
.len
+qlen
+toklen
);
3345 strcat (token_name
.buffer
, qualifier
);
3346 strncat (token_name
.buffer
,
3347 newlb
.buffer
+ tokoff
, toklen
);
3350 else if (objdef
== ocatseen
)
3351 /* Objective C category */
3353 int len
= strlen (objtag
) + 2 + toklen
;
3354 linebuffer_setlen (&token_name
, len
);
3355 strcpy (token_name
.buffer
, objtag
);
3356 strcat (token_name
.buffer
, "(");
3357 strncat (token_name
.buffer
,
3358 newlb
.buffer
+ tokoff
, toklen
);
3359 strcat (token_name
.buffer
, ")");
3362 else if (objdef
== omethodtag
3363 || objdef
== omethodparm
)
3364 /* Objective C method */
3368 else if (fvdef
== fdefunname
)
3369 /* GNU DEFUN and similar macros */
3371 bool defun
= (newlb
.buffer
[tokoff
] == 'F');
3375 /* Rewrite the tag so that emacs lisp DEFUNs
3376 can be found by their elisp name */
3382 linebuffer_setlen (&token_name
, len
);
3383 strncpy (token_name
.buffer
,
3384 newlb
.buffer
+ off
, len
);
3385 token_name
.buffer
[len
] = '\0';
3388 if (token_name
.buffer
[len
] == '_')
3389 token_name
.buffer
[len
] = '-';
3390 token
.named
= defun
;
3394 linebuffer_setlen (&token_name
, toklen
);
3395 strncpy (token_name
.buffer
,
3396 newlb
.buffer
+ tokoff
, toklen
);
3397 token_name
.buffer
[toklen
] = '\0';
3398 /* Name macros and members. */
3399 token
.named
= (structdef
== stagseen
3400 || typdef
== ttypeseen
3403 && definedef
== dignorerest
)
3405 && definedef
== dnone
3406 && structdef
== snone
3409 token
.lineno
= lineno
;
3410 token
.offset
= tokoff
;
3411 token
.length
= toklen
;
3412 token
.line
= newlb
.buffer
;
3413 token
.linepos
= newlinepos
;
3416 if (definedef
== dnone
3417 && (fvdef
== fvnameseen
3418 || fvdef
== foperator
3419 || structdef
== stagseen
3421 || typdef
== ttypeseen
3422 || objdef
!= onone
))
3424 if (current_lb_is_new
)
3425 switch_line_buffers ();
3427 else if (definedef
!= dnone
3428 || fvdef
== fdefunname
3430 make_C_tag (funorvar
);
3432 else /* not yacc and consider_token failed */
3434 if (inattribute
&& fvdef
== fignore
)
3436 /* We have just met __attribute__ after a
3437 function parameter list: do not tag the
3444 } /* if (endtoken (c)) */
3445 else if (intoken (c
))
3451 } /* if (midtoken) */
3452 else if (begtoken (c
))
3460 /* This prevents tagging fb in
3461 void (__attribute__((noreturn)) *fb) (void);
3462 Fixing this is not easy and not very important. */
3466 if (plainc
|| declarations
)
3468 make_C_tag (TRUE
); /* a function */
3473 if (structdef
== stagseen
&& !cjava
)
3475 popclass_above (bracelev
);
3483 if (!yacc_rules
|| lp
== newlb
.buffer
+ 1)
3485 tokoff
= lp
- 1 - newlb
.buffer
;
3490 } /* if (begtoken) */
3491 } /* if must look at token */
3494 /* Detect end of line, colon, comma, semicolon and various braces
3495 after having handled a token.*/
3501 if (yacc_rules
&& token
.offset
== 0 && token
.valid
)
3503 make_C_tag (FALSE
); /* a yacc function */
3506 if (definedef
!= dnone
)
3512 make_C_tag (TRUE
); /* an Objective C class */
3516 objdef
= omethodcolon
;
3517 linebuffer_setlen (&token_name
, token_name
.len
+ 1);
3518 strcat (token_name
.buffer
, ":");
3521 if (structdef
== stagseen
)
3523 structdef
= scolonseen
;
3526 /* Should be useless, but may be work as a safety net. */
3527 if (cplpl
&& fvdef
== flistseen
)
3529 make_C_tag (TRUE
); /* a function */
3535 if (definedef
!= dnone
|| inattribute
)
3541 make_C_tag (FALSE
); /* a typedef */
3551 if (typdef
== tignore
|| cplpl
)
3555 if ((globals
&& bracelev
== 0 && (!fvextern
|| declarations
))
3556 || (members
&& instruct
))
3557 make_C_tag (FALSE
); /* a variable */
3560 token
.valid
= FALSE
;
3564 && (cplpl
|| !instruct
)
3565 && (typdef
== tnone
|| (typdef
!= tignore
&& instruct
)))
3567 && plainc
&& instruct
))
3568 make_C_tag (TRUE
); /* a function */
3574 && cplpl
&& structdef
== stagseen
)
3575 make_C_tag (FALSE
); /* forward declaration */
3577 token
.valid
= FALSE
;
3578 } /* switch (fvdef) */
3584 if (structdef
== stagseen
)
3588 if (definedef
!= dnone
|| inattribute
)
3594 make_C_tag (TRUE
); /* an Objective C method */
3615 && (!fvextern
|| declarations
))
3616 || (members
&& instruct
)))
3617 make_C_tag (FALSE
); /* a variable */
3620 if ((declarations
&& typdef
== tnone
&& !instruct
)
3621 || (members
&& typdef
!= tignore
&& instruct
))
3623 make_C_tag (TRUE
); /* a function */
3626 else if (!declarations
)
3628 token
.valid
= FALSE
;
3633 if (structdef
== stagseen
)
3637 if (definedef
!= dnone
|| inattribute
)
3639 if (structdef
== stagseen
)
3646 make_C_tag (FALSE
); /* a typedef */
3658 if ((members
&& bracelev
== 1)
3659 || (globals
&& bracelev
== 0
3660 && (!fvextern
|| declarations
)))
3661 make_C_tag (FALSE
); /* a variable */
3675 if (definedef
!= dnone
)
3677 if (objdef
== otagseen
&& parlev
== 0)
3678 objdef
= oparenseen
;
3682 if (typdef
== ttypeseen
3686 /* This handles constructs like:
3687 typedef void OperatorFun (int fun); */
3706 if (--attrparlev
== 0)
3707 inattribute
= FALSE
;
3710 if (definedef
!= dnone
)
3712 if (objdef
== ocatseen
&& parlev
== 1)
3714 make_C_tag (TRUE
); /* an Objective C category */
3728 || typdef
== ttypeseen
))
3731 make_C_tag (FALSE
); /* a typedef */
3734 else if (parlev
< 0) /* can happen due to ill-conceived #if's. */
3738 if (definedef
!= dnone
)
3740 if (typdef
== ttypeseen
)
3742 /* Whenever typdef is set to tinbody (currently only
3743 here), typdefbracelev should be set to bracelev. */
3745 typdefbracelev
= bracelev
;
3750 make_C_tag (TRUE
); /* a function */
3759 make_C_tag (TRUE
); /* an Objective C class */
3764 make_C_tag (TRUE
); /* an Objective C method */
3768 /* Neutralize `extern "C" {' grot. */
3769 if (bracelev
== 0 && structdef
== snone
&& nestlev
== 0
3777 case skeyseen
: /* unnamed struct */
3778 pushclass_above (bracelev
, NULL
, 0);
3781 case stagseen
: /* named struct or enum */
3782 case scolonseen
: /* a class */
3783 pushclass_above (bracelev
,token
.line
+token
.offset
, token
.length
);
3785 make_C_tag (FALSE
); /* a struct or enum */
3791 if (definedef
!= dnone
)
3793 if (fvdef
== fstartlist
)
3795 fvdef
= fvnone
; /* avoid tagging `foo' in `foo (*bar()) ()' */
3796 token
.valid
= FALSE
;
3800 if (definedef
!= dnone
)
3803 if (!ignoreindent
&& lp
== newlb
.buffer
+ 1)
3806 token
.valid
= FALSE
; /* unexpected value, token unreliable */
3807 bracelev
= 0; /* reset brace level if first column */
3808 parlev
= 0; /* also reset paren level, just in case... */
3810 else if (bracelev
< 0)
3812 token
.valid
= FALSE
; /* something gone amiss, token unreliable */
3815 if (bracelev
== 0 && fvdef
== vignore
)
3816 fvdef
= fvnone
; /* end of function */
3817 popclass_above (bracelev
);
3819 /* Only if typdef == tinbody is typdefbracelev significant. */
3820 if (typdef
== tinbody
&& bracelev
<= typdefbracelev
)
3822 assert (bracelev
== typdefbracelev
);
3827 if (definedef
!= dnone
)
3837 if ((members
&& bracelev
== 1)
3838 || (globals
&& bracelev
== 0 && (!fvextern
|| declarations
)))
3839 make_C_tag (FALSE
); /* a variable */
3847 && (structdef
== stagseen
|| fvdef
== fvnameseen
))
3854 if (templatelev
> 0)
3862 if (objdef
== oinbody
&& bracelev
== 0)
3864 objdef
= omethodsign
;
3869 case '#': case '~': case '&': case '%': case '/':
3870 case '|': case '^': case '!': case '.': case '?':
3871 if (definedef
!= dnone
)
3873 /* These surely cannot follow a function tag in C. */
3886 if (objdef
== otagseen
)
3888 make_C_tag (TRUE
); /* an Objective C class */
3891 /* If a macro spans multiple lines don't reset its state. */
3893 CNL_SAVE_DEFINEDEF ();
3899 } /* while not eof */
3901 free (lbs
[0].lb
.buffer
);
3902 free (lbs
[1].lb
.buffer
);
3906 * Process either a C++ file or a C file depending on the setting
3910 default_C_entries (FILE *inf
)
3912 C_entries (cplusplus
? C_PLPL
: C_AUTO
, inf
);
3915 /* Always do plain C. */
3917 plain_C_entries (FILE *inf
)
3922 /* Always do C++. */
3924 Cplusplus_entries (FILE *inf
)
3926 C_entries (C_PLPL
, inf
);
3929 /* Always do Java. */
3931 Cjava_entries (FILE *inf
)
3933 C_entries (C_JAVA
, inf
);
3938 Cstar_entries (FILE *inf
)
3940 C_entries (C_STAR
, inf
);
3943 /* Always do Yacc. */
3945 Yacc_entries (FILE *inf
)
3947 C_entries (YACC
, inf
);
3951 /* Useful macros. */
3952 #define LOOP_ON_INPUT_LINES(file_pointer, line_buffer, char_pointer) \
3953 for (; /* loop initialization */ \
3954 !feof (file_pointer) /* loop test */ \
3955 && /* instructions at start of loop */ \
3956 (readline (&line_buffer, file_pointer), \
3957 char_pointer = line_buffer.buffer, \
3961 #define LOOKING_AT(cp, kw) /* kw is the keyword, a literal string */ \
3962 ((assert("" kw), TRUE) /* syntax error if not a literal string */ \
3963 && strneq ((cp), kw, sizeof(kw)-1) /* cp points at kw */ \
3964 && notinname ((cp)[sizeof(kw)-1]) /* end of kw */ \
3965 && ((cp) = skip_spaces((cp)+sizeof(kw)-1))) /* skip spaces */
3967 /* Similar to LOOKING_AT but does not use notinname, does not skip */
3968 #define LOOKING_AT_NOCASE(cp, kw) /* the keyword is a literal string */ \
3969 ((assert("" kw), TRUE) /* syntax error if not a literal string */ \
3970 && strncaseeq ((cp), kw, sizeof(kw)-1) /* cp points at kw */ \
3971 && ((cp) += sizeof(kw)-1)) /* skip spaces */
3974 * Read a file, but do no processing. This is used to do regexp
3975 * matching on files that have no language defined.
3978 just_read_file (FILE *inf
)
3981 readline (&lb
, inf
);
3985 /* Fortran parsing */
3987 static void F_takeprec (void);
3988 static void F_getit (FILE *);
3993 dbp
= skip_spaces (dbp
);
3997 dbp
= skip_spaces (dbp
);
3998 if (strneq (dbp
, "(*)", 3))
4003 if (!ISDIGIT (*dbp
))
4005 --dbp
; /* force failure */
4010 while (ISDIGIT (*dbp
));
4018 dbp
= skip_spaces (dbp
);
4021 readline (&lb
, inf
);
4026 dbp
= skip_spaces (dbp
);
4028 if (!ISALPHA (*dbp
) && *dbp
!= '_' && *dbp
!= '$')
4030 for (cp
= dbp
+ 1; *cp
!= '\0' && intoken (*cp
); cp
++)
4032 make_tag (dbp
, cp
-dbp
, TRUE
,
4033 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4038 Fortran_functions (FILE *inf
)
4040 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
4043 dbp
++; /* Ratfor escape to fortran */
4044 dbp
= skip_spaces (dbp
);
4048 if (LOOKING_AT_NOCASE (dbp
, "recursive"))
4049 dbp
= skip_spaces (dbp
);
4051 if (LOOKING_AT_NOCASE (dbp
, "pure"))
4052 dbp
= skip_spaces (dbp
);
4054 switch (lowcase (*dbp
))
4057 if (nocase_tail ("integer"))
4061 if (nocase_tail ("real"))
4065 if (nocase_tail ("logical"))
4069 if (nocase_tail ("complex") || nocase_tail ("character"))
4073 if (nocase_tail ("double"))
4075 dbp
= skip_spaces (dbp
);
4078 if (nocase_tail ("precision"))
4084 dbp
= skip_spaces (dbp
);
4087 switch (lowcase (*dbp
))
4090 if (nocase_tail ("function"))
4094 if (nocase_tail ("subroutine"))
4098 if (nocase_tail ("entry"))
4102 if (nocase_tail ("blockdata") || nocase_tail ("block data"))
4104 dbp
= skip_spaces (dbp
);
4105 if (*dbp
== '\0') /* assume un-named */
4106 make_tag ("blockdata", 9, TRUE
,
4107 lb
.buffer
, dbp
- lb
.buffer
, lineno
, linecharno
);
4109 F_getit (inf
); /* look for name */
4120 * Philippe Waroquiers (1998)
4123 /* Once we are positioned after an "interesting" keyword, let's get
4124 the real tag value necessary. */
4126 Ada_getit (FILE *inf
, const char *name_qualifier
)
4134 dbp
= skip_spaces (dbp
);
4136 || (dbp
[0] == '-' && dbp
[1] == '-'))
4138 readline (&lb
, inf
);
4141 switch (lowcase(*dbp
))
4144 if (nocase_tail ("body"))
4146 /* Skipping body of procedure body or package body or ....
4147 resetting qualifier to body instead of spec. */
4148 name_qualifier
= "/b";
4153 /* Skipping type of task type or protected type ... */
4154 if (nocase_tail ("type"))
4161 for (cp
= dbp
; *cp
!= '\0' && *cp
!= '"'; cp
++)
4166 dbp
= skip_spaces (dbp
);
4169 && (ISALPHA (*cp
) || ISDIGIT (*cp
) || *cp
== '_' || *cp
== '.'));
4177 name
= concat (dbp
, name_qualifier
, "");
4179 make_tag (name
, strlen (name
), TRUE
,
4180 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4189 Ada_funcs (FILE *inf
)
4191 bool inquote
= FALSE
;
4192 bool skip_till_semicolumn
= FALSE
;
4194 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
4196 while (*dbp
!= '\0')
4198 /* Skip a string i.e. "abcd". */
4199 if (inquote
|| (*dbp
== '"'))
4201 dbp
= etags_strchr (dbp
+ !inquote
, '"');
4206 continue; /* advance char */
4211 break; /* advance line */
4215 /* Skip comments. */
4216 if (dbp
[0] == '-' && dbp
[1] == '-')
4217 break; /* advance line */
4219 /* Skip character enclosed in single quote i.e. 'a'
4220 and skip single quote starting an attribute i.e. 'Image. */
4229 if (skip_till_semicolumn
)
4232 skip_till_semicolumn
= FALSE
;
4234 continue; /* advance char */
4237 /* Search for beginning of a token. */
4238 if (!begtoken (*dbp
))
4241 continue; /* advance char */
4244 /* We are at the beginning of a token. */
4245 switch (lowcase(*dbp
))
4248 if (!packages_only
&& nocase_tail ("function"))
4249 Ada_getit (inf
, "/f");
4251 break; /* from switch */
4252 continue; /* advance char */
4254 if (!packages_only
&& nocase_tail ("procedure"))
4255 Ada_getit (inf
, "/p");
4256 else if (nocase_tail ("package"))
4257 Ada_getit (inf
, "/s");
4258 else if (nocase_tail ("protected")) /* protected type */
4259 Ada_getit (inf
, "/t");
4261 break; /* from switch */
4262 continue; /* advance char */
4265 if (typedefs
&& !packages_only
&& nocase_tail ("use"))
4267 /* when tagging types, avoid tagging use type Pack.Typename;
4268 for this, we will skip everything till a ; */
4269 skip_till_semicolumn
= TRUE
;
4270 continue; /* advance char */
4274 if (!packages_only
&& nocase_tail ("task"))
4275 Ada_getit (inf
, "/k");
4276 else if (typedefs
&& !packages_only
&& nocase_tail ("type"))
4278 Ada_getit (inf
, "/t");
4279 while (*dbp
!= '\0')
4283 break; /* from switch */
4284 continue; /* advance char */
4287 /* Look for the end of the token. */
4288 while (!endtoken (*dbp
))
4291 } /* advance char */
4292 } /* advance line */
4297 * Unix and microcontroller assembly tag handling
4298 * Labels: /^[a-zA-Z_.$][a-zA_Z0-9_.$]*[: ^I^J]/
4299 * Idea by Bob Weiner, Motorola Inc. (1994)
4302 Asm_labels (FILE *inf
)
4306 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4308 /* If first char is alphabetic or one of [_.$], test for colon
4309 following identifier. */
4310 if (ISALPHA (*cp
) || *cp
== '_' || *cp
== '.' || *cp
== '$')
4312 /* Read past label. */
4314 while (ISALNUM (*cp
) || *cp
== '_' || *cp
== '.' || *cp
== '$')
4316 if (*cp
== ':' || iswhite (*cp
))
4317 /* Found end of label, so copy it and add it to the table. */
4318 make_tag (lb
.buffer
, cp
- lb
.buffer
, TRUE
,
4319 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4327 * Perl sub names: /^sub[ \t\n]+[^ \t\n{]+/
4328 * Perl variable names: /^(my|local).../
4329 * Original code by Bart Robinson <lomew@cs.utah.edu> (1995)
4330 * Additions by Michael Ernst <mernst@alum.mit.edu> (1997)
4331 * Ideas by Kai Großjohann <Kai.Grossjohann@CS.Uni-Dortmund.DE> (2001)
4334 Perl_functions (FILE *inf
)
4336 char *package
= savestr ("main"); /* current package name */
4339 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4341 cp
= skip_spaces (cp
);
4343 if (LOOKING_AT (cp
, "package"))
4346 get_tag (cp
, &package
);
4348 else if (LOOKING_AT (cp
, "sub"))
4353 while (!notinname (*cp
))
4356 continue; /* nothing found */
4357 if ((pos
= etags_strchr (sp
, ':')) != NULL
4358 && pos
< cp
&& pos
[1] == ':')
4359 /* The name is already qualified. */
4360 make_tag (sp
, cp
- sp
, TRUE
,
4361 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4365 char savechar
, *name
;
4369 name
= concat (package
, "::", sp
);
4371 make_tag (name
, strlen(name
), TRUE
,
4372 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4376 else if (globals
) /* only if we are tagging global vars */
4378 /* Skip a qualifier, if any. */
4379 bool qual
= LOOKING_AT (cp
, "my") || LOOKING_AT (cp
, "local");
4380 /* After "my" or "local", but before any following paren or space. */
4381 char *varstart
= cp
;
4383 if (qual
/* should this be removed? If yes, how? */
4384 && (*cp
== '$' || *cp
== '@' || *cp
== '%'))
4389 while (ISALNUM (*cp
) || *cp
== '_');
4393 /* Should be examining a variable list at this point;
4394 could insist on seeing an open parenthesis. */
4395 while (*cp
!= '\0' && *cp
!= ';' && *cp
!= '=' && *cp
!= ')')
4401 make_tag (varstart
, cp
- varstart
, FALSE
,
4402 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4411 * Look for /^[\t]*def[ \t\n]+[^ \t\n(:]+/ or /^class[ \t\n]+[^ \t\n(:]+/
4412 * Idea by Eric S. Raymond <esr@thyrsus.com> (1997)
4413 * More ideas by seb bacon <seb@jamkit.com> (2002)
4416 Python_functions (FILE *inf
)
4420 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4422 cp
= skip_spaces (cp
);
4423 if (LOOKING_AT (cp
, "def") || LOOKING_AT (cp
, "class"))
4426 while (!notinname (*cp
) && *cp
!= ':')
4428 make_tag (name
, cp
- name
, TRUE
,
4429 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4438 * - /^[ \t]*function[ \t\n]+[^ \t\n(]+/
4439 * - /^[ \t]*class[ \t\n]+[^ \t\n]+/
4440 * - /^[ \t]*define\(\"[^\"]+/
4441 * Only with --members:
4442 * - /^[ \t]*var[ \t\n]+\$[^ \t\n=;]/
4443 * Idea by Diez B. Roggisch (2001)
4446 PHP_functions (FILE *inf
)
4448 register char *cp
, *name
;
4449 bool search_identifier
= FALSE
;
4451 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4453 cp
= skip_spaces (cp
);
4455 if (search_identifier
4458 while (!notinname (*cp
))
4460 make_tag (name
, cp
- name
, TRUE
,
4461 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4462 search_identifier
= FALSE
;
4464 else if (LOOKING_AT (cp
, "function"))
4467 cp
= skip_spaces (cp
+1);
4471 while (!notinname (*cp
))
4473 make_tag (name
, cp
- name
, TRUE
,
4474 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4477 search_identifier
= TRUE
;
4479 else if (LOOKING_AT (cp
, "class"))
4484 while (*cp
!= '\0' && !iswhite (*cp
))
4486 make_tag (name
, cp
- name
, FALSE
,
4487 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4490 search_identifier
= TRUE
;
4492 else if (strneq (cp
, "define", 6)
4493 && (cp
= skip_spaces (cp
+6))
4495 && (*cp
== '"' || *cp
== '\''))
4499 while (*cp
!= quote
&& *cp
!= '\0')
4501 make_tag (name
, cp
- name
, FALSE
,
4502 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4505 && LOOKING_AT (cp
, "var")
4509 while (!notinname(*cp
))
4511 make_tag (name
, cp
- name
, FALSE
,
4512 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4519 * Cobol tag functions
4520 * We could look for anything that could be a paragraph name.
4521 * i.e. anything that starts in column 8 is one word and ends in a full stop.
4522 * Idea by Corny de Souza (1993)
4525 Cobol_paragraphs (FILE *inf
)
4527 register char *bp
, *ep
;
4529 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4535 /* If eoln, compiler option or comment ignore whole line. */
4536 if (bp
[-1] != ' ' || !ISALNUM (bp
[0]))
4539 for (ep
= bp
; ISALNUM (*ep
) || *ep
== '-'; ep
++)
4542 make_tag (bp
, ep
- bp
, TRUE
,
4543 lb
.buffer
, ep
- lb
.buffer
+ 1, lineno
, linecharno
);
4550 * Ideas by Assar Westerlund <assar@sics.se> (2001)
4553 Makefile_targets (FILE *inf
)
4557 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4559 if (*bp
== '\t' || *bp
== '#')
4561 while (*bp
!= '\0' && *bp
!= '=' && *bp
!= ':')
4563 if (*bp
== ':' || (globals
&& *bp
== '='))
4565 /* We should detect if there is more than one tag, but we do not.
4566 We just skip initial and final spaces. */
4567 char * namestart
= skip_spaces (lb
.buffer
);
4568 while (--bp
> namestart
)
4569 if (!notinname (*bp
))
4571 make_tag (namestart
, bp
- namestart
+ 1, TRUE
,
4572 lb
.buffer
, bp
- lb
.buffer
+ 2, lineno
, linecharno
);
4580 * Original code by Mosur K. Mohan (1989)
4582 * Locates tags for procedures & functions. Doesn't do any type- or
4583 * var-definitions. It does look for the keyword "extern" or
4584 * "forward" immediately following the procedure statement; if found,
4585 * the tag is skipped.
4588 Pascal_functions (FILE *inf
)
4590 linebuffer tline
; /* mostly copied from C_entries */
4592 int save_lineno
, namelen
, taglen
;
4595 bool /* each of these flags is TRUE if: */
4596 incomment
, /* point is inside a comment */
4597 inquote
, /* point is inside '..' string */
4598 get_tagname
, /* point is after PROCEDURE/FUNCTION
4599 keyword, so next item = potential tag */
4600 found_tag
, /* point is after a potential tag */
4601 inparms
, /* point is within parameter-list */
4602 verify_tag
; /* point has passed the parm-list, so the
4603 next token will determine whether this
4604 is a FORWARD/EXTERN to be ignored, or
4605 whether it is a real tag */
4607 save_lcno
= save_lineno
= namelen
= taglen
= 0; /* keep compiler quiet */
4608 name
= NULL
; /* keep compiler quiet */
4611 linebuffer_init (&tline
);
4613 incomment
= inquote
= FALSE
;
4614 found_tag
= FALSE
; /* have a proc name; check if extern */
4615 get_tagname
= FALSE
; /* found "procedure" keyword */
4616 inparms
= FALSE
; /* found '(' after "proc" */
4617 verify_tag
= FALSE
; /* check if "extern" is ahead */
4620 while (!feof (inf
)) /* long main loop to get next char */
4623 if (c
== '\0') /* if end of line */
4625 readline (&lb
, inf
);
4629 if (!((found_tag
&& verify_tag
)
4631 c
= *dbp
++; /* only if don't need *dbp pointing
4632 to the beginning of the name of
4633 the procedure or function */
4637 if (c
== '}') /* within { } comments */
4639 else if (c
== '*' && *dbp
== ')') /* within (* *) comments */
4656 inquote
= TRUE
; /* found first quote */
4658 case '{': /* found open { comment */
4662 if (*dbp
== '*') /* found open (* comment */
4667 else if (found_tag
) /* found '(' after tag, i.e., parm-list */
4670 case ')': /* end of parms list */
4675 if (found_tag
&& !inparms
) /* end of proc or fn stmt */
4682 if (found_tag
&& verify_tag
&& (*dbp
!= ' '))
4684 /* Check if this is an "extern" declaration. */
4687 if (lowcase (*dbp
== 'e'))
4689 if (nocase_tail ("extern")) /* superfluous, really! */
4695 else if (lowcase (*dbp
) == 'f')
4697 if (nocase_tail ("forward")) /* check for forward reference */
4703 if (found_tag
&& verify_tag
) /* not external proc, so make tag */
4707 make_tag (name
, namelen
, TRUE
,
4708 tline
.buffer
, taglen
, save_lineno
, save_lcno
);
4712 if (get_tagname
) /* grab name of proc or fn */
4719 /* Find block name. */
4720 for (cp
= dbp
+ 1; *cp
!= '\0' && !endtoken (*cp
); cp
++)
4723 /* Save all values for later tagging. */
4724 linebuffer_setlen (&tline
, lb
.len
);
4725 strcpy (tline
.buffer
, lb
.buffer
);
4726 save_lineno
= lineno
;
4727 save_lcno
= linecharno
;
4728 name
= tline
.buffer
+ (dbp
- lb
.buffer
);
4730 taglen
= cp
- lb
.buffer
+ 1;
4732 dbp
= cp
; /* set dbp to e-o-token */
4733 get_tagname
= FALSE
;
4737 /* And proceed to check for "extern". */
4739 else if (!incomment
&& !inquote
&& !found_tag
)
4741 /* Check for proc/fn keywords. */
4742 switch (lowcase (c
))
4745 if (nocase_tail ("rocedure")) /* c = 'p', dbp has advanced */
4749 if (nocase_tail ("unction"))
4754 } /* while not eof */
4756 free (tline
.buffer
);
4761 * Lisp tag functions
4762 * look for (def or (DEF, quote or QUOTE
4765 static void L_getit (void);
4770 if (*dbp
== '\'') /* Skip prefix quote */
4772 else if (*dbp
== '(')
4775 /* Try to skip "(quote " */
4776 if (!LOOKING_AT (dbp
, "quote") && !LOOKING_AT (dbp
, "QUOTE"))
4777 /* Ok, then skip "(" before name in (defstruct (foo)) */
4778 dbp
= skip_spaces (dbp
);
4780 get_tag (dbp
, NULL
);
4784 Lisp_functions (FILE *inf
)
4786 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
4791 if (strneq (dbp
+1, "def", 3) || strneq (dbp
+1, "DEF", 3))
4793 dbp
= skip_non_spaces (dbp
);
4794 dbp
= skip_spaces (dbp
);
4799 /* Check for (foo::defmumble name-defined ... */
4802 while (!notinname (*dbp
) && *dbp
!= ':');
4807 while (*dbp
== ':');
4809 if (strneq (dbp
, "def", 3) || strneq (dbp
, "DEF", 3))
4811 dbp
= skip_non_spaces (dbp
);
4812 dbp
= skip_spaces (dbp
);
4822 * Lua script language parsing
4823 * Original code by David A. Capello <dacap@users.sourceforge.net> (2004)
4825 * "function" and "local function" are tags if they start at column 1.
4828 Lua_functions (FILE *inf
)
4832 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4834 if (bp
[0] != 'f' && bp
[0] != 'l')
4837 (void)LOOKING_AT (bp
, "local"); /* skip possible "local" */
4839 if (LOOKING_AT (bp
, "function"))
4847 * Just look for lines where the first character is '/'
4848 * Also look at "defineps" for PSWrap
4850 * Richard Mlynarik <mly@adoc.xerox.com> (1997)
4851 * Masatake Yamato <masata-y@is.aist-nara.ac.jp> (1999)
4854 PS_functions (FILE *inf
)
4856 register char *bp
, *ep
;
4858 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4863 *ep
!= '\0' && *ep
!= ' ' && *ep
!= '{';
4866 make_tag (bp
, ep
- bp
, TRUE
,
4867 lb
.buffer
, ep
- lb
.buffer
+ 1, lineno
, linecharno
);
4869 else if (LOOKING_AT (bp
, "defineps"))
4877 * Ignore anything after \ followed by space or in ( )
4878 * Look for words defined by :
4879 * Look for constant, code, create, defer, value, and variable
4880 * OBP extensions: Look for buffer:, field,
4881 * Ideas by Eduardo Horvath <eeh@netbsd.org> (2004)
4884 Forth_words (FILE *inf
)
4888 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4889 while ((bp
= skip_spaces (bp
))[0] != '\0')
4890 if (bp
[0] == '\\' && iswhite(bp
[1]))
4891 break; /* read next line */
4892 else if (bp
[0] == '(' && iswhite(bp
[1]))
4893 do /* skip to ) or eol */
4895 while (*bp
!= ')' && *bp
!= '\0');
4896 else if ((bp
[0] == ':' && iswhite(bp
[1]) && bp
++)
4897 || LOOKING_AT_NOCASE (bp
, "constant")
4898 || LOOKING_AT_NOCASE (bp
, "code")
4899 || LOOKING_AT_NOCASE (bp
, "create")
4900 || LOOKING_AT_NOCASE (bp
, "defer")
4901 || LOOKING_AT_NOCASE (bp
, "value")
4902 || LOOKING_AT_NOCASE (bp
, "variable")
4903 || LOOKING_AT_NOCASE (bp
, "buffer:")
4904 || LOOKING_AT_NOCASE (bp
, "field"))
4905 get_tag (skip_spaces (bp
), NULL
); /* Yay! A definition! */
4907 bp
= skip_non_spaces (bp
);
4912 * Scheme tag functions
4913 * look for (def... xyzzy
4915 * (def ... ((...(xyzzy ....
4917 * Original code by Ken Haase (1985?)
4920 Scheme_functions (FILE *inf
)
4924 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4926 if (strneq (bp
, "(def", 4) || strneq (bp
, "(DEF", 4))
4928 bp
= skip_non_spaces (bp
+4);
4929 /* Skip over open parens and white space. Don't continue past
4931 while (*bp
&& notinname (*bp
))
4935 if (LOOKING_AT (bp
, "(SET!") || LOOKING_AT (bp
, "(set!"))
4941 /* Find tags in TeX and LaTeX input files. */
4943 /* TEX_toktab is a table of TeX control sequences that define tags.
4944 * Each entry records one such control sequence.
4946 * Original code from who knows whom.
4948 * Stefan Monnier (2002)
4951 static linebuffer
*TEX_toktab
= NULL
; /* Table with tag tokens */
4953 /* Default set of control sequences to put into TEX_toktab.
4954 The value of environment var TEXTAGS is prepended to this. */
4955 static const char *TEX_defenv
= "\
4956 :chapter:section:subsection:subsubsection:eqno:label:ref:cite:bibitem\
4957 :part:appendix:entry:index:def\
4958 :newcommand:renewcommand:newenvironment:renewenvironment";
4960 static void TEX_mode (FILE *);
4961 static void TEX_decode_env (const char *, const char *);
4963 static char TEX_esc
= '\\';
4964 static char TEX_opgrp
= '{';
4965 static char TEX_clgrp
= '}';
4968 * TeX/LaTeX scanning loop.
4971 TeX_commands (FILE *inf
)
4976 /* Select either \ or ! as escape character. */
4979 /* Initialize token table once from environment. */
4980 if (TEX_toktab
== NULL
)
4981 TEX_decode_env ("TEXTAGS", TEX_defenv
);
4983 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4985 /* Look at each TEX keyword in line. */
4988 /* Look for a TEX escape. */
4989 while (*cp
++ != TEX_esc
)
4990 if (cp
[-1] == '\0' || cp
[-1] == '%')
4993 for (key
= TEX_toktab
; key
->buffer
!= NULL
; key
++)
4994 if (strneq (cp
, key
->buffer
, key
->len
))
4997 int namelen
, linelen
;
5000 cp
= skip_spaces (cp
+ key
->len
);
5001 if (*cp
== TEX_opgrp
)
5007 (!iswhite (*p
) && *p
!= '#' &&
5008 *p
!= TEX_opgrp
&& *p
!= TEX_clgrp
);
5013 if (!opgrp
|| *p
== TEX_clgrp
)
5015 while (*p
!= '\0' && *p
!= TEX_opgrp
&& *p
!= TEX_clgrp
)
5017 linelen
= p
- lb
.buffer
+ 1;
5019 make_tag (cp
, namelen
, TRUE
,
5020 lb
.buffer
, linelen
, lineno
, linecharno
);
5021 goto tex_next_line
; /* We only tag a line once */
5029 #define TEX_LESC '\\'
5030 #define TEX_SESC '!'
5032 /* Figure out whether TeX's escapechar is '\\' or '!' and set grouping
5033 chars accordingly. */
5035 TEX_mode (FILE *inf
)
5039 while ((c
= getc (inf
)) != EOF
)
5041 /* Skip to next line if we hit the TeX comment char. */
5043 while (c
!= '\n' && c
!= EOF
)
5045 else if (c
== TEX_LESC
|| c
== TEX_SESC
)
5061 /* If the input file is compressed, inf is a pipe, and rewind may fail.
5062 No attempt is made to correct the situation. */
5066 /* Read environment and prepend it to the default string.
5067 Build token table. */
5069 TEX_decode_env (const char *evarname
, const char *defenv
)
5071 register const char *env
, *p
;
5074 /* Append default string to environment. */
5075 env
= getenv (evarname
);
5079 env
= concat (env
, defenv
, "");
5081 /* Allocate a token table */
5082 for (len
= 1, p
= env
; p
;)
5083 if ((p
= etags_strchr (p
, ':')) && *++p
!= '\0')
5085 TEX_toktab
= xnew (len
, linebuffer
);
5087 /* Unpack environment string into token table. Be careful about */
5088 /* zero-length strings (leading ':', "::" and trailing ':') */
5089 for (i
= 0; *env
!= '\0';)
5091 p
= etags_strchr (env
, ':');
5092 if (!p
) /* End of environment string. */
5093 p
= env
+ strlen (env
);
5095 { /* Only non-zero strings. */
5096 TEX_toktab
[i
].buffer
= savenstr (env
, p
- env
);
5097 TEX_toktab
[i
].len
= p
- env
;
5104 TEX_toktab
[i
].buffer
= NULL
; /* Mark end of table. */
5105 TEX_toktab
[i
].len
= 0;
5112 /* Texinfo support. Dave Love, Mar. 2000. */
5114 Texinfo_nodes (FILE *inf
)
5117 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5118 if (LOOKING_AT (cp
, "@node"))
5121 while (*cp
!= '\0' && *cp
!= ',')
5123 make_tag (start
, cp
- start
, TRUE
,
5124 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
5131 * Contents of <title>, <h1>, <h2>, <h3> are tags.
5132 * Contents of <a name=xxx> are tags with name xxx.
5134 * Francesco Potortì, 2002.
5137 HTML_labels (FILE *inf
)
5139 bool getnext
= FALSE
; /* next text outside of HTML tags is a tag */
5140 bool skiptag
= FALSE
; /* skip to the end of the current HTML tag */
5141 bool intag
= FALSE
; /* inside an html tag, looking for ID= */
5142 bool inanchor
= FALSE
; /* when INTAG, is an anchor, look for NAME= */
5146 linebuffer_setlen (&token_name
, 0); /* no name in buffer */
5148 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
5149 for (;;) /* loop on the same line */
5151 if (skiptag
) /* skip HTML tag */
5153 while (*dbp
!= '\0' && *dbp
!= '>')
5159 continue; /* look on the same line */
5161 break; /* go to next line */
5164 else if (intag
) /* look for "name=" or "id=" */
5166 while (*dbp
!= '\0' && *dbp
!= '>'
5167 && lowcase (*dbp
) != 'n' && lowcase (*dbp
) != 'i')
5170 break; /* go to next line */
5175 continue; /* look on the same line */
5177 if ((inanchor
&& LOOKING_AT_NOCASE (dbp
, "name="))
5178 || LOOKING_AT_NOCASE (dbp
, "id="))
5180 bool quoted
= (dbp
[0] == '"');
5183 for (end
= ++dbp
; *end
!= '\0' && *end
!= '"'; end
++)
5186 for (end
= dbp
; *end
!= '\0' && intoken (*end
); end
++)
5188 linebuffer_setlen (&token_name
, end
- dbp
);
5189 strncpy (token_name
.buffer
, dbp
, end
- dbp
);
5190 token_name
.buffer
[end
- dbp
] = '\0';
5193 intag
= FALSE
; /* we found what we looked for */
5194 skiptag
= TRUE
; /* skip to the end of the tag */
5195 getnext
= TRUE
; /* then grab the text */
5196 continue; /* look on the same line */
5201 else if (getnext
) /* grab next tokens and tag them */
5203 dbp
= skip_spaces (dbp
);
5205 break; /* go to next line */
5209 inanchor
= (lowcase (dbp
[1]) == 'a' && !intoken (dbp
[2]));
5210 continue; /* look on the same line */
5213 for (end
= dbp
+ 1; *end
!= '\0' && *end
!= '<'; end
++)
5215 make_tag (token_name
.buffer
, token_name
.len
, TRUE
,
5216 dbp
, end
- dbp
, lineno
, linecharno
);
5217 linebuffer_setlen (&token_name
, 0); /* no name in buffer */
5219 break; /* go to next line */
5222 else /* look for an interesting HTML tag */
5224 while (*dbp
!= '\0' && *dbp
!= '<')
5227 break; /* go to next line */
5229 if (lowcase (dbp
[1]) == 'a' && !intoken (dbp
[2]))
5232 continue; /* look on the same line */
5234 else if (LOOKING_AT_NOCASE (dbp
, "<title>")
5235 || LOOKING_AT_NOCASE (dbp
, "<h1>")
5236 || LOOKING_AT_NOCASE (dbp
, "<h2>")
5237 || LOOKING_AT_NOCASE (dbp
, "<h3>"))
5241 continue; /* look on the same line */
5252 * Assumes that the predicate or rule starts at column 0.
5253 * Only the first clause of a predicate or rule is added.
5254 * Original code by Sunichirou Sugou (1989)
5255 * Rewritten by Anders Lindgren (1996)
5257 static size_t prolog_pr (char *, char *);
5258 static void prolog_skip_comment (linebuffer
*, FILE *);
5259 static size_t prolog_atom (char *, size_t);
5262 Prolog_functions (FILE *inf
)
5272 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5274 if (cp
[0] == '\0') /* Empty line */
5276 else if (iswhite (cp
[0])) /* Not a predicate */
5278 else if (cp
[0] == '/' && cp
[1] == '*') /* comment. */
5279 prolog_skip_comment (&lb
, inf
);
5280 else if ((len
= prolog_pr (cp
, last
)) > 0)
5282 /* Predicate or rule. Store the function name so that we
5283 only generate a tag for the first clause. */
5285 last
= xnew(len
+ 1, char);
5286 else if (len
+ 1 > allocated
)
5287 xrnew (last
, len
+ 1, char);
5288 allocated
= len
+ 1;
5289 strncpy (last
, cp
, len
);
5298 prolog_skip_comment (linebuffer
*plb
, FILE *inf
)
5304 for (cp
= plb
->buffer
; *cp
!= '\0'; cp
++)
5305 if (cp
[0] == '*' && cp
[1] == '/')
5307 readline (plb
, inf
);
5313 * A predicate or rule definition is added if it matches:
5314 * <beginning of line><Prolog Atom><whitespace>(
5315 * or <beginning of line><Prolog Atom><whitespace>:-
5317 * It is added to the tags database if it doesn't match the
5318 * name of the previous clause header.
5320 * Return the size of the name of the predicate or rule, or 0 if no
5324 prolog_pr (char *s
, char *last
)
5326 /* Name of last clause. */
5331 pos
= prolog_atom (s
, 0);
5336 pos
= skip_spaces (s
+ pos
) - s
;
5339 || (s
[pos
] == '(' && (pos
+= 1))
5340 || (s
[pos
] == ':' && s
[pos
+ 1] == '-' && (pos
+= 2)))
5341 && (last
== NULL
/* save only the first clause */
5342 || len
!= strlen (last
)
5343 || !strneq (s
, last
, len
)))
5345 make_tag (s
, len
, TRUE
, s
, pos
, lineno
, linecharno
);
5353 * Consume a Prolog atom.
5354 * Return the number of bytes consumed, or 0 if there was an error.
5356 * A prolog atom, in this context, could be one of:
5357 * - An alphanumeric sequence, starting with a lower case letter.
5358 * - A quoted arbitrary string. Single quotes can escape themselves.
5359 * Backslash quotes everything.
5362 prolog_atom (char *s
, size_t pos
)
5368 if (ISLOWER(s
[pos
]) || (s
[pos
] == '_'))
5370 /* The atom is unquoted. */
5372 while (ISALNUM(s
[pos
]) || (s
[pos
] == '_'))
5376 return pos
- origpos
;
5378 else if (s
[pos
] == '\'')
5389 pos
++; /* A double quote */
5391 else if (s
[pos
] == '\0')
5392 /* Multiline quoted atoms are ignored. */
5394 else if (s
[pos
] == '\\')
5396 if (s
[pos
+1] == '\0')
5403 return pos
- origpos
;
5411 * Support for Erlang
5413 * Generates tags for functions, defines, and records.
5414 * Assumes that Erlang functions start at column 0.
5415 * Original code by Anders Lindgren (1996)
5417 static int erlang_func (char *, char *);
5418 static void erlang_attribute (char *);
5419 static int erlang_atom (char *);
5422 Erlang_functions (FILE *inf
)
5432 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5434 if (cp
[0] == '\0') /* Empty line */
5436 else if (iswhite (cp
[0])) /* Not function nor attribute */
5438 else if (cp
[0] == '%') /* comment */
5440 else if (cp
[0] == '"') /* Sometimes, strings start in column one */
5442 else if (cp
[0] == '-') /* attribute, e.g. "-define" */
5444 erlang_attribute (cp
);
5451 else if ((len
= erlang_func (cp
, last
)) > 0)
5454 * Function. Store the function name so that we only
5455 * generates a tag for the first clause.
5458 last
= xnew (len
+ 1, char);
5459 else if (len
+ 1 > allocated
)
5460 xrnew (last
, len
+ 1, char);
5461 allocated
= len
+ 1;
5462 strncpy (last
, cp
, len
);
5471 * A function definition is added if it matches:
5472 * <beginning of line><Erlang Atom><whitespace>(
5474 * It is added to the tags database if it doesn't match the
5475 * name of the previous clause header.
5477 * Return the size of the name of the function, or 0 if no function
5481 erlang_func (char *s
, char *last
)
5483 /* Name of last clause. */
5488 pos
= erlang_atom (s
);
5493 pos
= skip_spaces (s
+ pos
) - s
;
5495 /* Save only the first clause. */
5498 || len
!= (int)strlen (last
)
5499 || !strneq (s
, last
, len
)))
5501 make_tag (s
, len
, TRUE
, s
, pos
, lineno
, linecharno
);
5510 * Handle attributes. Currently, tags are generated for defines
5513 * They are on the form:
5514 * -define(foo, bar).
5515 * -define(Foo(M, N), M+N).
5516 * -record(graph, {vtab = notable, cyclic = true}).
5519 erlang_attribute (char *s
)
5523 if ((LOOKING_AT (cp
, "-define") || LOOKING_AT (cp
, "-record"))
5526 int len
= erlang_atom (skip_spaces (cp
));
5528 make_tag (cp
, len
, TRUE
, s
, cp
+ len
- s
, lineno
, linecharno
);
5535 * Consume an Erlang atom (or variable).
5536 * Return the number of bytes consumed, or -1 if there was an error.
5539 erlang_atom (char *s
)
5543 if (ISALPHA (s
[pos
]) || s
[pos
] == '_')
5545 /* The atom is unquoted. */
5548 while (ISALNUM (s
[pos
]) || s
[pos
] == '_');
5550 else if (s
[pos
] == '\'')
5552 for (pos
++; s
[pos
] != '\''; pos
++)
5553 if (s
[pos
] == '\0' /* multiline quoted atoms are ignored */
5554 || (s
[pos
] == '\\' && s
[++pos
] == '\0'))
5563 static char *scan_separators (char *);
5564 static void add_regex (char *, language
*);
5565 static char *substitute (char *, char *, struct re_registers
*);
5568 * Take a string like "/blah/" and turn it into "blah", verifying
5569 * that the first and last characters are the same, and handling
5570 * quoted separator characters. Actually, stops on the occurrence of
5571 * an unquoted separator. Also process \t, \n, etc. and turn into
5572 * appropriate characters. Works in place. Null terminates name string.
5573 * Returns pointer to terminating separator, or NULL for
5574 * unterminated regexps.
5577 scan_separators (char *name
)
5580 char *copyto
= name
;
5581 bool quoted
= FALSE
;
5583 for (++name
; *name
!= '\0'; ++name
)
5589 case 'a': *copyto
++ = '\007'; break; /* BEL (bell) */
5590 case 'b': *copyto
++ = '\b'; break; /* BS (back space) */
5591 case 'd': *copyto
++ = 0177; break; /* DEL (delete) */
5592 case 'e': *copyto
++ = 033; break; /* ESC (delete) */
5593 case 'f': *copyto
++ = '\f'; break; /* FF (form feed) */
5594 case 'n': *copyto
++ = '\n'; break; /* NL (new line) */
5595 case 'r': *copyto
++ = '\r'; break; /* CR (carriage return) */
5596 case 't': *copyto
++ = '\t'; break; /* TAB (horizontal tab) */
5597 case 'v': *copyto
++ = '\v'; break; /* VT (vertical tab) */
5603 /* Something else is quoted, so preserve the quote. */
5611 else if (*name
== '\\')
5613 else if (*name
== sep
)
5619 name
= NULL
; /* signal unterminated regexp */
5621 /* Terminate copied string. */
5626 /* Look at the argument of --regex or --no-regex and do the right
5627 thing. Same for each line of a regexp file. */
5629 analyse_regex (char *regex_arg
)
5631 if (regex_arg
== NULL
)
5633 free_regexps (); /* --no-regex: remove existing regexps */
5637 /* A real --regexp option or a line in a regexp file. */
5638 switch (regex_arg
[0])
5640 /* Comments in regexp file or null arg to --regex. */
5646 /* Read a regex file. This is recursive and may result in a
5647 loop, which will stop when the file descriptors are exhausted. */
5651 linebuffer regexbuf
;
5652 char *regexfile
= regex_arg
+ 1;
5654 /* regexfile is a file containing regexps, one per line. */
5655 regexfp
= fopen (regexfile
, "r");
5656 if (regexfp
== NULL
)
5661 linebuffer_init (®exbuf
);
5662 while (readline_internal (®exbuf
, regexfp
) > 0)
5663 analyse_regex (regexbuf
.buffer
);
5664 free (regexbuf
.buffer
);
5669 /* Regexp to be used for a specific language only. */
5673 char *lang_name
= regex_arg
+ 1;
5676 for (cp
= lang_name
; *cp
!= '}'; cp
++)
5679 error ("unterminated language name in regex: %s", regex_arg
);
5683 lang
= get_language_from_langname (lang_name
);
5686 add_regex (cp
, lang
);
5690 /* Regexp to be used for any language. */
5692 add_regex (regex_arg
, NULL
);
5697 /* Separate the regexp pattern, compile it,
5698 and care for optional name and modifiers. */
5700 add_regex (char *regexp_pattern
, language
*lang
)
5702 static struct re_pattern_buffer zeropattern
;
5703 char sep
, *pat
, *name
, *modifiers
;
5706 struct re_pattern_buffer
*patbuf
;
5709 force_explicit_name
= TRUE
, /* do not use implicit tag names */
5710 ignore_case
= FALSE
, /* case is significant */
5711 multi_line
= FALSE
, /* matches are done one line at a time */
5712 single_line
= FALSE
; /* dot does not match newline */
5715 if (strlen(regexp_pattern
) < 3)
5717 error ("null regexp", (char *)NULL
);
5720 sep
= regexp_pattern
[0];
5721 name
= scan_separators (regexp_pattern
);
5724 error ("%s: unterminated regexp", regexp_pattern
);
5729 error ("null name for regexp \"%s\"", regexp_pattern
);
5732 modifiers
= scan_separators (name
);
5733 if (modifiers
== NULL
) /* no terminating separator --> no name */
5739 modifiers
+= 1; /* skip separator */
5741 /* Parse regex modifiers. */
5742 for (; modifiers
[0] != '\0'; modifiers
++)
5743 switch (modifiers
[0])
5746 if (modifiers
== name
)
5747 error ("forcing explicit tag name but no name, ignoring", NULL
);
5748 force_explicit_name
= TRUE
;
5758 need_filebuf
= TRUE
;
5763 wrongmod
[0] = modifiers
[0];
5765 error ("invalid regexp modifier `%s', ignoring", wrongmod
);
5770 patbuf
= xnew (1, struct re_pattern_buffer
);
5771 *patbuf
= zeropattern
;
5774 static char lc_trans
[CHARS
];
5776 for (i
= 0; i
< CHARS
; i
++)
5777 lc_trans
[i
] = lowcase (i
);
5778 patbuf
->translate
= lc_trans
; /* translation table to fold case */
5782 pat
= concat ("^", regexp_pattern
, ""); /* anchor to beginning of line */
5784 pat
= regexp_pattern
;
5787 re_set_syntax (RE_SYNTAX_EMACS
| RE_DOT_NEWLINE
);
5789 re_set_syntax (RE_SYNTAX_EMACS
);
5791 err
= re_compile_pattern (pat
, strlen (pat
), patbuf
);
5796 error ("%s while compiling pattern", err
);
5801 p_head
= xnew (1, regexp
);
5802 p_head
->pattern
= savestr (regexp_pattern
);
5803 p_head
->p_next
= rp
;
5804 p_head
->lang
= lang
;
5805 p_head
->pat
= patbuf
;
5806 p_head
->name
= savestr (name
);
5807 p_head
->error_signaled
= FALSE
;
5808 p_head
->force_explicit_name
= force_explicit_name
;
5809 p_head
->ignore_case
= ignore_case
;
5810 p_head
->multi_line
= multi_line
;
5814 * Do the substitutions indicated by the regular expression and
5818 substitute (char *in
, char *out
, struct re_registers
*regs
)
5821 int size
, dig
, diglen
;
5824 size
= strlen (out
);
5826 /* Pass 1: figure out how much to allocate by finding all \N strings. */
5827 if (out
[size
- 1] == '\\')
5828 fatal ("pattern error in \"%s\"", out
);
5829 for (t
= etags_strchr (out
, '\\');
5831 t
= etags_strchr (t
+ 2, '\\'))
5835 diglen
= regs
->end
[dig
] - regs
->start
[dig
];
5841 /* Allocate space and do the substitutions. */
5843 result
= xnew (size
+ 1, char);
5845 for (t
= result
; *out
!= '\0'; out
++)
5846 if (*out
== '\\' && ISDIGIT (*++out
))
5849 diglen
= regs
->end
[dig
] - regs
->start
[dig
];
5850 strncpy (t
, in
+ regs
->start
[dig
], diglen
);
5857 assert (t
<= result
+ size
);
5858 assert (t
- result
== (int)strlen (result
));
5863 /* Deallocate all regexps. */
5868 while (p_head
!= NULL
)
5870 rp
= p_head
->p_next
;
5871 free (p_head
->pattern
);
5872 free (p_head
->name
);
5880 * Reads the whole file as a single string from `filebuf' and looks for
5881 * multi-line regular expressions, creating tags on matches.
5882 * readline already dealt with normal regexps.
5884 * Idea by Ben Wing <ben@666.com> (2002).
5887 regex_tag_multiline (void)
5889 char *buffer
= filebuf
.buffer
;
5893 for (rp
= p_head
; rp
!= NULL
; rp
= rp
->p_next
)
5897 if (!rp
->multi_line
)
5898 continue; /* skip normal regexps */
5900 /* Generic initialisations before parsing file from memory. */
5901 lineno
= 1; /* reset global line number */
5902 charno
= 0; /* reset global char number */
5903 linecharno
= 0; /* reset global char number of line start */
5905 /* Only use generic regexps or those for the current language. */
5906 if (rp
->lang
!= NULL
&& rp
->lang
!= curfdp
->lang
)
5909 while (match
>= 0 && match
< filebuf
.len
)
5911 match
= re_search (rp
->pat
, buffer
, filebuf
.len
, charno
,
5912 filebuf
.len
- match
, &rp
->regs
);
5917 if (!rp
->error_signaled
)
5919 error ("regexp stack overflow while matching \"%s\"",
5921 rp
->error_signaled
= TRUE
;
5928 if (match
== rp
->regs
.end
[0])
5930 if (!rp
->error_signaled
)
5932 error ("regexp matches the empty string: \"%s\"",
5934 rp
->error_signaled
= TRUE
;
5936 match
= -3; /* exit from while loop */
5940 /* Match occurred. Construct a tag. */
5941 while (charno
< rp
->regs
.end
[0])
5942 if (buffer
[charno
++] == '\n')
5943 lineno
++, linecharno
= charno
;
5945 if (name
[0] == '\0')
5947 else /* make a named tag */
5948 name
= substitute (buffer
, rp
->name
, &rp
->regs
);
5949 if (rp
->force_explicit_name
)
5950 /* Force explicit tag name, if a name is there. */
5951 pfnote (name
, TRUE
, buffer
+ linecharno
,
5952 charno
- linecharno
+ 1, lineno
, linecharno
);
5954 make_tag (name
, strlen (name
), TRUE
, buffer
+ linecharno
,
5955 charno
- linecharno
+ 1, lineno
, linecharno
);
5964 nocase_tail (const char *cp
)
5966 register int len
= 0;
5968 while (*cp
!= '\0' && lowcase (*cp
) == lowcase (dbp
[len
]))
5970 if (*cp
== '\0' && !intoken (dbp
[len
]))
5979 get_tag (register char *bp
, char **namepp
)
5981 register char *cp
= bp
;
5985 /* Go till you get to white space or a syntactic break */
5986 for (cp
= bp
+ 1; !notinname (*cp
); cp
++)
5988 make_tag (bp
, cp
- bp
, TRUE
,
5989 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
5993 *namepp
= savenstr (bp
, cp
- bp
);
5997 * Read a line of text from `stream' into `lbp', excluding the
5998 * newline or CR-NL, if any. Return the number of characters read from
5999 * `stream', which is the length of the line including the newline.
6001 * On DOS or Windows we do not count the CR character, if any before the
6002 * NL, in the returned length; this mirrors the behavior of Emacs on those
6003 * platforms (for text files, it translates CR-NL to NL as it reads in the
6006 * If multi-line regular expressions are requested, each line read is
6007 * appended to `filebuf'.
6010 readline_internal (linebuffer
*lbp
, register FILE *stream
)
6012 char *buffer
= lbp
->buffer
;
6013 register char *p
= lbp
->buffer
;
6014 register char *pend
;
6017 pend
= p
+ lbp
->size
; /* Separate to avoid 386/IX compiler bug. */
6021 register int c
= getc (stream
);
6024 /* We're at the end of linebuffer: expand it. */
6026 xrnew (buffer
, lbp
->size
, char);
6027 p
+= buffer
- lbp
->buffer
;
6028 pend
= buffer
+ lbp
->size
;
6029 lbp
->buffer
= buffer
;
6039 if (p
> buffer
&& p
[-1] == '\r')
6043 /* Assume CRLF->LF translation will be performed by Emacs
6044 when loading this file, so CRs won't appear in the buffer.
6045 It would be cleaner to compensate within Emacs;
6046 however, Emacs does not know how many CRs were deleted
6047 before any given point in the file. */
6062 lbp
->len
= p
- buffer
;
6064 if (need_filebuf
/* we need filebuf for multi-line regexps */
6065 && chars_deleted
> 0) /* not at EOF */
6067 while (filebuf
.size
<= filebuf
.len
+ lbp
->len
+ 1) /* +1 for \n */
6069 /* Expand filebuf. */
6071 xrnew (filebuf
.buffer
, filebuf
.size
, char);
6073 strncpy (filebuf
.buffer
+ filebuf
.len
, lbp
->buffer
, lbp
->len
);
6074 filebuf
.len
+= lbp
->len
;
6075 filebuf
.buffer
[filebuf
.len
++] = '\n';
6076 filebuf
.buffer
[filebuf
.len
] = '\0';
6079 return lbp
->len
+ chars_deleted
;
6083 * Like readline_internal, above, but in addition try to match the
6084 * input line against relevant regular expressions and manage #line
6088 readline (linebuffer
*lbp
, FILE *stream
)
6092 linecharno
= charno
; /* update global char number of line start */
6093 result
= readline_internal (lbp
, stream
); /* read line */
6094 lineno
+= 1; /* increment global line number */
6095 charno
+= result
; /* increment global char number */
6097 /* Honour #line directives. */
6098 if (!no_line_directive
)
6100 static bool discard_until_line_directive
;
6102 /* Check whether this is a #line directive. */
6103 if (result
> 12 && strneq (lbp
->buffer
, "#line ", 6))
6108 if (sscanf (lbp
->buffer
, "#line %u \"%n", &lno
, &start
) >= 1
6109 && start
> 0) /* double quote character found */
6111 char *endp
= lbp
->buffer
+ start
;
6113 while ((endp
= etags_strchr (endp
, '"')) != NULL
6114 && endp
[-1] == '\\')
6117 /* Ok, this is a real #line directive. Let's deal with it. */
6119 char *taggedabsname
; /* absolute name of original file */
6120 char *taggedfname
; /* name of original file as given */
6121 char *name
; /* temp var */
6123 discard_until_line_directive
= FALSE
; /* found it */
6124 name
= lbp
->buffer
+ start
;
6126 canonicalize_filename (name
);
6127 taggedabsname
= absolute_filename (name
, tagfiledir
);
6128 if (filename_is_absolute (name
)
6129 || filename_is_absolute (curfdp
->infname
))
6130 taggedfname
= savestr (taggedabsname
);
6132 taggedfname
= relative_filename (taggedabsname
,tagfiledir
);
6134 if (streq (curfdp
->taggedfname
, taggedfname
))
6135 /* The #line directive is only a line number change. We
6136 deal with this afterwards. */
6139 /* The tags following this #line directive should be
6140 attributed to taggedfname. In order to do this, set
6141 curfdp accordingly. */
6143 fdesc
*fdp
; /* file description pointer */
6145 /* Go look for a file description already set up for the
6146 file indicated in the #line directive. If there is
6147 one, use it from now until the next #line
6149 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
6150 if (streq (fdp
->infname
, curfdp
->infname
)
6151 && streq (fdp
->taggedfname
, taggedfname
))
6152 /* If we remove the second test above (after the &&)
6153 then all entries pertaining to the same file are
6154 coalesced in the tags file. If we use it, then
6155 entries pertaining to the same file but generated
6156 from different files (via #line directives) will
6157 go into separate sections in the tags file. These
6158 alternatives look equivalent. The first one
6159 destroys some apparently useless information. */
6165 /* Else, if we already tagged the real file, skip all
6166 input lines until the next #line directive. */
6167 if (fdp
== NULL
) /* not found */
6168 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
6169 if (streq (fdp
->infabsname
, taggedabsname
))
6171 discard_until_line_directive
= TRUE
;
6175 /* Else create a new file description and use that from
6176 now on, until the next #line directive. */
6177 if (fdp
== NULL
) /* not found */
6180 fdhead
= xnew (1, fdesc
);
6181 *fdhead
= *curfdp
; /* copy curr. file description */
6183 fdhead
->infname
= savestr (curfdp
->infname
);
6184 fdhead
->infabsname
= savestr (curfdp
->infabsname
);
6185 fdhead
->infabsdir
= savestr (curfdp
->infabsdir
);
6186 fdhead
->taggedfname
= taggedfname
;
6187 fdhead
->usecharno
= FALSE
;
6188 fdhead
->prop
= NULL
;
6189 fdhead
->written
= FALSE
;
6193 free (taggedabsname
);
6195 readline (lbp
, stream
);
6197 } /* if a real #line directive */
6198 } /* if #line is followed by a number */
6199 } /* if line begins with "#line " */
6201 /* If we are here, no #line directive was found. */
6202 if (discard_until_line_directive
)
6206 /* Do a tail recursion on ourselves, thus discarding the contents
6207 of the line buffer. */
6208 readline (lbp
, stream
);
6212 discard_until_line_directive
= FALSE
;
6215 } /* if #line directives should be considered */
6222 /* Match against relevant regexps. */
6224 for (rp
= p_head
; rp
!= NULL
; rp
= rp
->p_next
)
6226 /* Only use generic regexps or those for the current language.
6227 Also do not use multiline regexps, which is the job of
6228 regex_tag_multiline. */
6229 if ((rp
->lang
!= NULL
&& rp
->lang
!= fdhead
->lang
)
6233 match
= re_match (rp
->pat
, lbp
->buffer
, lbp
->len
, 0, &rp
->regs
);
6238 if (!rp
->error_signaled
)
6240 error ("regexp stack overflow while matching \"%s\"",
6242 rp
->error_signaled
= TRUE
;
6249 /* Empty string matched. */
6250 if (!rp
->error_signaled
)
6252 error ("regexp matches the empty string: \"%s\"", rp
->pattern
);
6253 rp
->error_signaled
= TRUE
;
6257 /* Match occurred. Construct a tag. */
6259 if (name
[0] == '\0')
6261 else /* make a named tag */
6262 name
= substitute (lbp
->buffer
, rp
->name
, &rp
->regs
);
6263 if (rp
->force_explicit_name
)
6264 /* Force explicit tag name, if a name is there. */
6265 pfnote (name
, TRUE
, lbp
->buffer
, match
, lineno
, linecharno
);
6267 make_tag (name
, strlen (name
), TRUE
,
6268 lbp
->buffer
, match
, lineno
, linecharno
);
6277 * Return a pointer to a space of size strlen(cp)+1 allocated
6278 * with xnew where the string CP has been copied.
6281 savestr (const char *cp
)
6283 return savenstr (cp
, strlen (cp
));
6287 * Return a pointer to a space of size LEN+1 allocated with xnew where
6288 * the string CP has been copied for at most the first LEN characters.
6291 savenstr (const char *cp
, int len
)
6295 dp
= xnew (len
+ 1, char);
6296 strncpy (dp
, cp
, len
);
6302 * Return the ptr in sp at which the character c last
6303 * appears; NULL if not found
6305 * Identical to POSIX strrchr, included for portability.
6308 etags_strrchr (register const char *sp
, register int c
)
6310 register const char *r
;
6322 * Return the ptr in sp at which the character c first
6323 * appears; NULL if not found
6325 * Identical to POSIX strchr, included for portability.
6328 etags_strchr (register const char *sp
, register int c
)
6339 * Compare two strings, ignoring case for alphabetic characters.
6341 * Same as BSD's strcasecmp, included for portability.
6344 etags_strcasecmp (register const char *s1
, register const char *s2
)
6347 && (ISALPHA (*s1
) && ISALPHA (*s2
)
6348 ? lowcase (*s1
) == lowcase (*s2
)
6352 return (ISALPHA (*s1
) && ISALPHA (*s2
)
6353 ? lowcase (*s1
) - lowcase (*s2
)
6358 * Compare two strings, ignoring case for alphabetic characters.
6359 * Stop after a given number of characters
6361 * Same as BSD's strncasecmp, included for portability.
6364 etags_strncasecmp (register const char *s1
, register const char *s2
, register int n
)
6366 while (*s1
!= '\0' && n
-- > 0
6367 && (ISALPHA (*s1
) && ISALPHA (*s2
)
6368 ? lowcase (*s1
) == lowcase (*s2
)
6375 return (ISALPHA (*s1
) && ISALPHA (*s2
)
6376 ? lowcase (*s1
) - lowcase (*s2
)
6380 /* Skip spaces (end of string is not space), return new pointer. */
6382 skip_spaces (char *cp
)
6384 while (iswhite (*cp
))
6389 /* Skip non spaces, except end of string, return new pointer. */
6391 skip_non_spaces (char *cp
)
6393 while (*cp
!= '\0' && !iswhite (*cp
))
6398 /* Print error message and exit. */
6400 fatal (const char *s1
, const char *s2
)
6403 exit (EXIT_FAILURE
);
6407 pfatal (const char *s1
)
6410 exit (EXIT_FAILURE
);
6414 suggest_asking_for_help (void)
6416 fprintf (stderr
, "\tTry `%s %s' for a complete list of options.\n",
6417 progname
, NO_LONG_OPTIONS
? "-h" : "--help");
6418 exit (EXIT_FAILURE
);
6421 /* Print error message. `s1' is printf control string, `s2' is arg for it. */
6423 error (const char *s1
, const char *s2
)
6425 fprintf (stderr
, "%s: ", progname
);
6426 fprintf (stderr
, s1
, s2
);
6427 fprintf (stderr
, "\n");
6430 /* Return a newly-allocated string whose contents
6431 concatenate those of s1, s2, s3. */
6433 concat (const char *s1
, const char *s2
, const char *s3
)
6435 int len1
= strlen (s1
), len2
= strlen (s2
), len3
= strlen (s3
);
6436 char *result
= xnew (len1
+ len2
+ len3
+ 1, char);
6438 strcpy (result
, s1
);
6439 strcpy (result
+ len1
, s2
);
6440 strcpy (result
+ len1
+ len2
, s3
);
6441 result
[len1
+ len2
+ len3
] = '\0';
6447 /* Does the same work as the system V getcwd, but does not need to
6448 guess the buffer size in advance. */
6454 char *path
= xnew (bufsize
, char);
6456 while (getcwd (path
, bufsize
) == NULL
)
6458 if (errno
!= ERANGE
)
6462 path
= xnew (bufsize
, char);
6465 canonicalize_filename (path
);
6468 #else /* not HAVE_GETCWD */
6471 char *p
, path
[MAXPATHLEN
+ 1]; /* Fixed size is safe on MSDOS. */
6475 for (p
= path
; *p
!= '\0'; p
++)
6481 return strdup (path
);
6482 #else /* not MSDOS */
6486 linebuffer_init (&path
);
6487 pipe
= (FILE *) popen ("pwd 2>/dev/null", "r");
6488 if (pipe
== NULL
|| readline_internal (&path
, pipe
) == 0)
6493 #endif /* not MSDOS */
6494 #endif /* not HAVE_GETCWD */
6497 /* Return a newly allocated string containing the file name of FILE
6498 relative to the absolute directory DIR (which should end with a slash). */
6500 relative_filename (char *file
, char *dir
)
6502 char *fp
, *dp
, *afn
, *res
;
6505 /* Find the common root of file and dir (with a trailing slash). */
6506 afn
= absolute_filename (file
, cwd
);
6509 while (*fp
++ == *dp
++)
6511 fp
--, dp
--; /* back to the first differing char */
6513 if (fp
== afn
&& afn
[0] != '/') /* cannot build a relative name */
6516 do /* look at the equal chars until '/' */
6520 /* Build a sequence of "../" strings for the resulting relative file name. */
6522 while ((dp
= etags_strchr (dp
+ 1, '/')) != NULL
)
6524 res
= xnew (3*i
+ strlen (fp
+ 1) + 1, char);
6527 strcat (res
, "../");
6529 /* Add the file name relative to the common root of file and dir. */
6530 strcat (res
, fp
+ 1);
6536 /* Return a newly allocated string containing the absolute file name
6537 of FILE given DIR (which should end with a slash). */
6539 absolute_filename (char *file
, char *dir
)
6541 char *slashp
, *cp
, *res
;
6543 if (filename_is_absolute (file
))
6544 res
= savestr (file
);
6546 /* We don't support non-absolute file names with a drive
6547 letter, like `d:NAME' (it's too much hassle). */
6548 else if (file
[1] == ':')
6549 fatal ("%s: relative file names with drive letters not supported", file
);
6552 res
= concat (dir
, file
, "");
6554 /* Delete the "/dirname/.." and "/." substrings. */
6555 slashp
= etags_strchr (res
, '/');
6556 while (slashp
!= NULL
&& slashp
[0] != '\0')
6558 if (slashp
[1] == '.')
6560 if (slashp
[2] == '.'
6561 && (slashp
[3] == '/' || slashp
[3] == '\0'))
6566 while (cp
>= res
&& !filename_is_absolute (cp
));
6568 cp
= slashp
; /* the absolute name begins with "/.." */
6570 /* Under MSDOS and NT we get `d:/NAME' as absolute
6571 file name, so the luser could say `d:/../NAME'.
6572 We silently treat this as `d:/NAME'. */
6573 else if (cp
[0] != '/')
6576 memmove (cp
, slashp
+ 3, strlen (slashp
+ 2));
6580 else if (slashp
[2] == '/' || slashp
[2] == '\0')
6582 memmove (slashp
, slashp
+ 2, strlen (slashp
+ 1));
6587 slashp
= etags_strchr (slashp
+ 1, '/');
6590 if (res
[0] == '\0') /* just a safety net: should never happen */
6593 return savestr ("/");
6599 /* Return a newly allocated string containing the absolute
6600 file name of dir where FILE resides given DIR (which should
6601 end with a slash). */
6603 absolute_dirname (char *file
, char *dir
)
6608 slashp
= etags_strrchr (file
, '/');
6610 return savestr (dir
);
6613 res
= absolute_filename (file
, dir
);
6619 /* Whether the argument string is an absolute file name. The argument
6620 string must have been canonicalized with canonicalize_filename. */
6622 filename_is_absolute (char *fn
)
6624 return (fn
[0] == '/'
6626 || (ISALPHA(fn
[0]) && fn
[1] == ':' && fn
[2] == '/')
6631 /* Downcase DOS drive letter and collapse separators into single slashes.
6634 canonicalize_filename (register char *fn
)
6640 /* Canonicalize drive letter case. */
6641 # define ISUPPER(c) isupper (CHAR(c))
6642 if (fn
[0] != '\0' && fn
[1] == ':' && ISUPPER (fn
[0]))
6643 fn
[0] = lowcase (fn
[0]);
6648 /* Collapse multiple separators into a single slash. */
6649 for (cp
= fn
; *cp
!= '\0'; cp
++, fn
++)
6653 while (cp
[1] == sep
)
6662 /* Initialize a linebuffer for use. */
6664 linebuffer_init (linebuffer
*lbp
)
6666 lbp
->size
= (DEBUG
) ? 3 : 200;
6667 lbp
->buffer
= xnew (lbp
->size
, char);
6668 lbp
->buffer
[0] = '\0';
6672 /* Set the minimum size of a string contained in a linebuffer. */
6674 linebuffer_setlen (linebuffer
*lbp
, int toksize
)
6676 while (lbp
->size
<= toksize
)
6679 xrnew (lbp
->buffer
, lbp
->size
, char);
6684 /* Like malloc but get fatal error if memory is exhausted. */
6686 xmalloc (size_t size
)
6688 PTR result
= (PTR
) malloc (size
);
6690 fatal ("virtual memory exhausted", (char *)NULL
);
6695 xrealloc (char *ptr
, size_t size
)
6697 PTR result
= (PTR
) realloc (ptr
, size
);
6699 fatal ("virtual memory exhausted", (char *)NULL
);
6705 * indent-tabs-mode: t
6708 * c-font-lock-extra-types: ("FILE" "bool" "language" "linebuffer" "fdesc" "node" "regexp")
6709 * c-file-style: "gnu"
6713 /* etags.c ends here */