1 /* Tags file maker to go with GNU Emacs -*- coding: utf-8 -*-
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, 1998-2015 Free Software
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.
71 * Francesco Potortì maintained and improved it for many years
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";
88 # define NDEBUG /* disable assert */
94 # define _GNU_SOURCE 1 /* enables some compiler checks on GNU */
97 /* WIN32_NATIVE is for XEmacs.
98 MSDOS, WINDOWSNT, DOS_NT are for Emacs. */
103 #endif /* WIN32_NATIVE */
108 # include <sys/param.h>
115 # define MAXPATHLEN _MAX_PATH
119 # define O_CLOEXEC O_NOINHERIT
120 #endif /* WINDOWSNT */
127 #include <sysstdio.h>
130 #include <sys/types.h>
131 #include <sys/stat.h>
132 #include <binary-io.h>
134 #include <c-strcase.h>
138 # undef assert /* some systems have a buggy assert.h */
139 # define assert(x) ((void) 0)
145 /* Define CTAGS to make the program "ctags" compatible with the usual one.
146 Leave it undefined to make the program "etags", which makes emacs-style
147 tag tables and tags typedefs, #defines and struct/union/enum by default. */
155 #define streq(s,t) (assert ((s)!=NULL || (t)!=NULL), !strcmp (s, t))
156 #define strcaseeq(s,t) (assert ((s)!=NULL && (t)!=NULL), !c_strcasecmp (s, t))
157 #define strneq(s,t,n) (assert ((s)!=NULL || (t)!=NULL), !strncmp (s, t, n))
158 #define strncaseeq(s,t,n) (assert ((s)!=NULL && (t)!=NULL), !c_strncasecmp (s, t, n))
160 /* C is not in a name. */
162 notinname (unsigned char c
)
164 /* Look at make_tag before modifying! */
165 static bool const table
[UCHAR_MAX
+ 1] = {
166 ['\0']=1, ['\t']=1, ['\n']=1, ['\f']=1, ['\r']=1, [' ']=1,
167 ['(']=1, [')']=1, [',']=1, [';']=1, ['=']=1
172 /* C can start a token. */
174 begtoken (unsigned char c
)
176 static bool const table
[UCHAR_MAX
+ 1] = {
178 ['A']=1, ['B']=1, ['C']=1, ['D']=1, ['E']=1, ['F']=1, ['G']=1, ['H']=1,
179 ['I']=1, ['J']=1, ['K']=1, ['L']=1, ['M']=1, ['N']=1, ['O']=1, ['P']=1,
180 ['Q']=1, ['R']=1, ['S']=1, ['T']=1, ['U']=1, ['V']=1, ['W']=1, ['X']=1,
183 ['a']=1, ['b']=1, ['c']=1, ['d']=1, ['e']=1, ['f']=1, ['g']=1, ['h']=1,
184 ['i']=1, ['j']=1, ['k']=1, ['l']=1, ['m']=1, ['n']=1, ['o']=1, ['p']=1,
185 ['q']=1, ['r']=1, ['s']=1, ['t']=1, ['u']=1, ['v']=1, ['w']=1, ['x']=1,
192 /* C can be in the middle of a token. */
194 intoken (unsigned char c
)
196 static bool const table
[UCHAR_MAX
+ 1] = {
198 ['0']=1, ['1']=1, ['2']=1, ['3']=1, ['4']=1,
199 ['5']=1, ['6']=1, ['7']=1, ['8']=1, ['9']=1,
200 ['A']=1, ['B']=1, ['C']=1, ['D']=1, ['E']=1, ['F']=1, ['G']=1, ['H']=1,
201 ['I']=1, ['J']=1, ['K']=1, ['L']=1, ['M']=1, ['N']=1, ['O']=1, ['P']=1,
202 ['Q']=1, ['R']=1, ['S']=1, ['T']=1, ['U']=1, ['V']=1, ['W']=1, ['X']=1,
205 ['a']=1, ['b']=1, ['c']=1, ['d']=1, ['e']=1, ['f']=1, ['g']=1, ['h']=1,
206 ['i']=1, ['j']=1, ['k']=1, ['l']=1, ['m']=1, ['n']=1, ['o']=1, ['p']=1,
207 ['q']=1, ['r']=1, ['s']=1, ['t']=1, ['u']=1, ['v']=1, ['w']=1, ['x']=1,
213 /* C can end a token. */
215 endtoken (unsigned char c
)
217 static bool const table
[UCHAR_MAX
+ 1] = {
218 ['\0']=1, ['\t']=1, ['\n']=1, ['\r']=1, [' ']=1,
219 ['!']=1, ['"']=1, ['#']=1, ['%']=1, ['&']=1, ['\'']=1, ['(']=1, [')']=1,
220 ['*']=1, ['+']=1, [',']=1, ['-']=1, ['.']=1, ['/']=1, [':']=1, [';']=1,
221 ['<']=1, ['=']=1, ['>']=1, ['?']=1, ['[']=1, [']']=1, ['^']=1,
222 ['{']=1, ['|']=1, ['}']=1, ['~']=1
228 * xnew, xrnew -- allocate, reallocate storage
230 * SYNOPSIS: Type *xnew (int n, Type);
231 * void xrnew (OldPointer, int n, Type);
233 #define xnew(n, Type) ((Type *) xmalloc ((n) * sizeof (Type)))
234 #define xrnew(op, n, Type) ((op) = (Type *) xrealloc (op, (n) * sizeof (Type)))
236 typedef void Lang_function (FILE *);
240 const char *suffix
; /* file name suffix for this compressor */
241 const char *command
; /* takes one arg and decompresses to stdout */
246 const char *name
; /* language name */
247 const char *help
; /* detailed help for the language */
248 Lang_function
*function
; /* parse function */
249 const char **suffixes
; /* name suffixes of this language's files */
250 const char **filenames
; /* names of this language's files */
251 const char **interpreters
; /* interpreters for this language */
252 bool metasource
; /* source used to generate other sources */
257 struct fdesc
*next
; /* for the linked list */
258 char *infname
; /* uncompressed input file name */
259 char *infabsname
; /* absolute uncompressed input file name */
260 char *infabsdir
; /* absolute dir of input file */
261 char *taggedfname
; /* file name to write in tagfile */
262 language
*lang
; /* language of file */
263 char *prop
; /* file properties to write in tagfile */
264 bool usecharno
; /* etags tags shall contain char number */
265 bool written
; /* entry written in the tags file */
268 typedef struct node_st
269 { /* sorting structure */
270 struct node_st
*left
, *right
; /* left and right sons */
271 fdesc
*fdp
; /* description of file to whom tag belongs */
272 char *name
; /* tag name */
273 char *regex
; /* search regexp */
274 bool valid
; /* write this tag on the tag file */
275 bool is_func
; /* function tag: use regexp in CTAGS mode */
276 bool been_warned
; /* warning already given for duplicated tag */
277 int lno
; /* line number tag is on */
278 long cno
; /* character number line starts on */
282 * A `linebuffer' is a structure which holds a line of text.
283 * `readline_internal' reads a line from a stream into a linebuffer
284 * and works regardless of the length of the line.
285 * SIZE is the size of BUFFER, LEN is the length of the string in
286 * BUFFER after readline reads it.
295 /* Used to support mixing of --lang and file names. */
299 at_language
, /* a language specification */
300 at_regexp
, /* a regular expression */
301 at_filename
, /* a file name */
302 at_stdin
, /* read from stdin here */
303 at_end
/* stop parsing the list */
304 } arg_type
; /* argument type */
305 language
*lang
; /* language associated with the argument */
306 char *what
; /* the argument itself */
309 /* Structure defining a regular expression. */
310 typedef struct regexp
312 struct regexp
*p_next
; /* pointer to next in list */
313 language
*lang
; /* if set, use only for this language */
314 char *pattern
; /* the regexp pattern */
315 char *name
; /* tag name */
316 struct re_pattern_buffer
*pat
; /* the compiled pattern */
317 struct re_registers regs
; /* re registers */
318 bool error_signaled
; /* already signaled for this regexp */
319 bool force_explicit_name
; /* do not allow implicit tag name */
320 bool ignore_case
; /* ignore case when matching */
321 bool multi_line
; /* do a multi-line match on the whole file */
325 /* Many compilers barf on this:
326 Lang_function Ada_funcs;
327 so let's write it this way */
328 static void Ada_funcs (FILE *);
329 static void Asm_labels (FILE *);
330 static void C_entries (int c_ext
, FILE *);
331 static void default_C_entries (FILE *);
332 static void plain_C_entries (FILE *);
333 static void Cjava_entries (FILE *);
334 static void Cobol_paragraphs (FILE *);
335 static void Cplusplus_entries (FILE *);
336 static void Cstar_entries (FILE *);
337 static void Erlang_functions (FILE *);
338 static void Forth_words (FILE *);
339 static void Fortran_functions (FILE *);
340 static void HTML_labels (FILE *);
341 static void Lisp_functions (FILE *);
342 static void Lua_functions (FILE *);
343 static void Makefile_targets (FILE *);
344 static void Pascal_functions (FILE *);
345 static void Perl_functions (FILE *);
346 static void PHP_functions (FILE *);
347 static void PS_functions (FILE *);
348 static void Prolog_functions (FILE *);
349 static void Python_functions (FILE *);
350 static void Scheme_functions (FILE *);
351 static void TeX_commands (FILE *);
352 static void Texinfo_nodes (FILE *);
353 static void Yacc_entries (FILE *);
354 static void just_read_file (FILE *);
356 static language
*get_language_from_langname (const char *);
357 static void readline (linebuffer
*, FILE *);
358 static long readline_internal (linebuffer
*, FILE *);
359 static bool nocase_tail (const char *);
360 static void get_tag (char *, char **);
362 static void analyze_regex (char *);
363 static void free_regexps (void);
364 static void regex_tag_multiline (void);
365 static void error (const char *, ...) ATTRIBUTE_FORMAT_PRINTF (1, 2);
366 static _Noreturn
void suggest_asking_for_help (void);
367 _Noreturn
void fatal (const char *, const char *);
368 static _Noreturn
void pfatal (const char *);
369 static void add_node (node
*, node
**);
371 static void process_file_name (char *, language
*);
372 static void process_file (FILE *, char *, language
*);
373 static void find_entries (FILE *);
374 static void free_tree (node
*);
375 static void free_fdesc (fdesc
*);
376 static void pfnote (char *, bool, char *, int, int, long);
377 static void invalidate_nodes (fdesc
*, node
**);
378 static void put_entries (node
*);
380 static char *concat (const char *, const char *, const char *);
381 static char *skip_spaces (char *);
382 static char *skip_non_spaces (char *);
383 static char *skip_name (char *);
384 static char *savenstr (const char *, int);
385 static char *savestr (const char *);
386 static char *etags_getcwd (void);
387 static char *relative_filename (char *, char *);
388 static char *absolute_filename (char *, char *);
389 static char *absolute_dirname (char *, char *);
390 static bool filename_is_absolute (char *f
);
391 static void canonicalize_filename (char *);
392 static char *etags_mktmp (void);
393 static void linebuffer_init (linebuffer
*);
394 static void linebuffer_setlen (linebuffer
*, int);
395 static void *xmalloc (size_t);
396 static void *xrealloc (void *, size_t);
399 static char searchar
= '/'; /* use /.../ searches */
401 static char *tagfile
; /* output file */
402 static char *progname
; /* name this program was invoked with */
403 static char *cwd
; /* current working directory */
404 static char *tagfiledir
; /* directory of tagfile */
405 static FILE *tagf
; /* ioptr for tags file */
406 static ptrdiff_t whatlen_max
; /* maximum length of any 'what' member */
408 static fdesc
*fdhead
; /* head of file description list */
409 static fdesc
*curfdp
; /* current file description */
410 static int lineno
; /* line number of current line */
411 static long charno
; /* current character number */
412 static long linecharno
; /* charno of start of current line */
413 static char *dbp
; /* pointer to start of current tag */
415 static const int invalidcharno
= -1;
417 static node
*nodehead
; /* the head of the binary tree of tags */
418 static node
*last_node
; /* the last node created */
420 static linebuffer lb
; /* the current line */
421 static linebuffer filebuf
; /* a buffer containing the whole file */
422 static linebuffer token_name
; /* a buffer containing a tag name */
424 static bool append_to_tagfile
; /* -a: append to tags */
425 /* The next five default to true in C and derived languages. */
426 static bool typedefs
; /* -t: create tags for C and Ada typedefs */
427 static bool typedefs_or_cplusplus
; /* -T: create tags for C typedefs, level */
428 /* 0 struct/enum/union decls, and C++ */
429 /* member functions. */
430 static bool constantypedefs
; /* -d: create tags for C #define, enum */
431 /* constants and variables. */
432 /* -D: opposite of -d. Default under ctags. */
433 static int globals
; /* create tags for global variables */
434 static int members
; /* create tags for C member variables */
435 static int declarations
; /* --declarations: tag them and extern in C&Co*/
436 static int no_line_directive
; /* ignore #line directives (undocumented) */
437 static int no_duplicates
; /* no duplicate tags for ctags (undocumented) */
438 static bool update
; /* -u: update tags */
439 static bool vgrind_style
; /* -v: create vgrind style index output */
440 static bool no_warnings
; /* -w: suppress warnings (undocumented) */
441 static bool cxref_style
; /* -x: create cxref style output */
442 static bool cplusplus
; /* .[hc] means C++, not C (undocumented) */
443 static bool ignoreindent
; /* -I: ignore indentation in C */
444 static int packages_only
; /* --packages-only: in Ada, only tag packages*/
446 /* STDIN is defined in LynxOS system headers */
451 #define STDIN 0x1001 /* returned by getopt_long on --parse-stdin */
452 static bool parsing_stdin
; /* --parse-stdin used */
454 static regexp
*p_head
; /* list of all regexps */
455 static bool need_filebuf
; /* some regexes are multi-line */
457 static struct option longopts
[] =
459 { "append", no_argument
, NULL
, 'a' },
460 { "packages-only", no_argument
, &packages_only
, 1 },
461 { "c++", no_argument
, NULL
, 'C' },
462 { "declarations", no_argument
, &declarations
, 1 },
463 { "no-line-directive", no_argument
, &no_line_directive
, 1 },
464 { "no-duplicates", no_argument
, &no_duplicates
, 1 },
465 { "help", no_argument
, NULL
, 'h' },
466 { "help", no_argument
, NULL
, 'H' },
467 { "ignore-indentation", no_argument
, NULL
, 'I' },
468 { "language", required_argument
, NULL
, 'l' },
469 { "members", no_argument
, &members
, 1 },
470 { "no-members", no_argument
, &members
, 0 },
471 { "output", required_argument
, NULL
, 'o' },
472 { "regex", required_argument
, NULL
, 'r' },
473 { "no-regex", no_argument
, NULL
, 'R' },
474 { "ignore-case-regex", required_argument
, NULL
, 'c' },
475 { "parse-stdin", required_argument
, NULL
, STDIN
},
476 { "version", no_argument
, NULL
, 'V' },
478 #if CTAGS /* Ctags options */
479 { "backward-search", no_argument
, NULL
, 'B' },
480 { "cxref", no_argument
, NULL
, 'x' },
481 { "defines", no_argument
, NULL
, 'd' },
482 { "globals", no_argument
, &globals
, 1 },
483 { "typedefs", no_argument
, NULL
, 't' },
484 { "typedefs-and-c++", no_argument
, NULL
, 'T' },
485 { "update", no_argument
, NULL
, 'u' },
486 { "vgrind", no_argument
, NULL
, 'v' },
487 { "no-warn", no_argument
, NULL
, 'w' },
489 #else /* Etags options */
490 { "no-defines", no_argument
, NULL
, 'D' },
491 { "no-globals", no_argument
, &globals
, 0 },
492 { "include", required_argument
, NULL
, 'i' },
497 static compressor compressors
[] =
499 { "z", "gzip -d -c"},
500 { "Z", "gzip -d -c"},
501 { "gz", "gzip -d -c"},
502 { "GZ", "gzip -d -c"},
503 { "bz2", "bzip2 -d -c" },
504 { "xz", "xz -d -c" },
513 static const char *Ada_suffixes
[] =
514 { "ads", "adb", "ada", NULL
};
515 static const char Ada_help
[] =
516 "In Ada code, functions, procedures, packages, tasks and types are\n\
517 tags. Use the `--packages-only' option to create tags for\n\
519 Ada tag names have suffixes indicating the type of entity:\n\
520 Entity type: Qualifier:\n\
521 ------------ ----------\n\
528 Thus, `M-x find-tag <RET> bidule/b <RET>' will go directly to the\n\
529 body of the package `bidule', while `M-x find-tag <RET> bidule <RET>'\n\
530 will just search for any tag `bidule'.";
533 static const char *Asm_suffixes
[] =
534 { "a", /* Unix assembler */
535 "asm", /* Microcontroller assembly */
536 "def", /* BSO/Tasking definition includes */
537 "inc", /* Microcontroller include files */
538 "ins", /* Microcontroller include files */
539 "s", "sa", /* Unix assembler */
540 "S", /* cpp-processed Unix assembler */
541 "src", /* BSO/Tasking C compiler output */
544 static const char Asm_help
[] =
545 "In assembler code, labels appearing at the beginning of a line,\n\
546 followed by a colon, are tags.";
549 /* Note that .c and .h can be considered C++, if the --c++ flag was
550 given, or if the `class' or `template' keywords are met inside the file.
551 That is why default_C_entries is called for these. */
552 static const char *default_C_suffixes
[] =
554 #if CTAGS /* C help for Ctags */
555 static const char default_C_help
[] =
556 "In C code, any C function is a tag. Use -t to tag typedefs.\n\
557 Use -T to tag definitions of `struct', `union' and `enum'.\n\
558 Use -d to tag `#define' macro definitions and `enum' constants.\n\
559 Use --globals to tag global variables.\n\
560 You can tag function declarations and external variables by\n\
561 using `--declarations', and struct members by using `--members'.";
562 #else /* C help for Etags */
563 static const char default_C_help
[] =
564 "In C code, any C function or typedef is a tag, and so are\n\
565 definitions of `struct', `union' and `enum'. `#define' macro\n\
566 definitions and `enum' constants are tags unless you specify\n\
567 `--no-defines'. Global variables are tags unless you specify\n\
568 `--no-globals' and so are struct members unless you specify\n\
569 `--no-members'. Use of `--no-globals', `--no-defines' and\n\
570 `--no-members' can make the tags table file much smaller.\n\
571 You can tag function declarations and external variables by\n\
572 using `--declarations'.";
573 #endif /* C help for Ctags and Etags */
575 static const char *Cplusplus_suffixes
[] =
576 { "C", "c++", "cc", "cpp", "cxx", "H", "h++", "hh", "hpp", "hxx",
577 "M", /* Objective C++ */
578 "pdb", /* PostScript with C syntax */
580 static const char Cplusplus_help
[] =
581 "In C++ code, all the tag constructs of C code are tagged. (Use\n\
582 --help --lang=c --lang=c++ for full help.)\n\
583 In addition to C tags, member functions are also recognized. Member\n\
584 variables are recognized unless you use the `--no-members' option.\n\
585 Tags for variables and functions in classes are named `CLASS::VARIABLE'\n\
586 and `CLASS::FUNCTION'. `operator' definitions have tag names like\n\
589 static const char *Cjava_suffixes
[] =
591 static char Cjava_help
[] =
592 "In Java code, all the tags constructs of C and C++ code are\n\
593 tagged. (Use --help --lang=c --lang=c++ --lang=java for full help.)";
596 static const char *Cobol_suffixes
[] =
597 { "COB", "cob", NULL
};
598 static char Cobol_help
[] =
599 "In Cobol code, tags are paragraph names; that is, any word\n\
600 starting in column 8 and followed by a period.";
602 static const char *Cstar_suffixes
[] =
603 { "cs", "hs", NULL
};
605 static const char *Erlang_suffixes
[] =
606 { "erl", "hrl", NULL
};
607 static const char Erlang_help
[] =
608 "In Erlang code, the tags are the functions, records and macros\n\
609 defined in the file.";
611 const char *Forth_suffixes
[] =
612 { "fth", "tok", NULL
};
613 static const char Forth_help
[] =
614 "In Forth code, tags are words defined by `:',\n\
615 constant, code, create, defer, value, variable, buffer:, field.";
617 static const char *Fortran_suffixes
[] =
618 { "F", "f", "f90", "for", NULL
};
619 static const char Fortran_help
[] =
620 "In Fortran code, functions, subroutines and block data are tags.";
622 static const char *HTML_suffixes
[] =
623 { "htm", "html", "shtml", NULL
};
624 static const char HTML_help
[] =
625 "In HTML input files, the tags are the `title' and the `h1', `h2',\n\
626 `h3' headers. Also, tags are `name=' in anchors and all\n\
627 occurrences of `id='.";
629 static const char *Lisp_suffixes
[] =
630 { "cl", "clisp", "el", "l", "lisp", "LSP", "lsp", "ml", NULL
};
631 static const char Lisp_help
[] =
632 "In Lisp code, any function defined with `defun', any variable\n\
633 defined with `defvar' or `defconst', and in general the first\n\
634 argument of any expression that starts with `(def' in column zero\n\
636 The `--declarations' option tags \"(defvar foo)\" constructs too.";
638 static const char *Lua_suffixes
[] =
639 { "lua", "LUA", NULL
};
640 static const char Lua_help
[] =
641 "In Lua scripts, all functions are tags.";
643 static const char *Makefile_filenames
[] =
644 { "Makefile", "makefile", "GNUMakefile", "Makefile.in", "Makefile.am", NULL
};
645 static const char Makefile_help
[] =
646 "In makefiles, targets are tags; additionally, variables are tags\n\
647 unless you specify `--no-globals'.";
649 static const char *Objc_suffixes
[] =
650 { "lm", /* Objective lex file */
651 "m", /* Objective C file */
653 static const char Objc_help
[] =
654 "In Objective C code, tags include Objective C definitions for classes,\n\
655 class categories, methods and protocols. Tags for variables and\n\
656 functions in classes are named `CLASS::VARIABLE' and `CLASS::FUNCTION'.\n\
657 (Use --help --lang=c --lang=objc --lang=java for full help.)";
659 static const char *Pascal_suffixes
[] =
660 { "p", "pas", NULL
};
661 static const char Pascal_help
[] =
662 "In Pascal code, the tags are the functions and procedures defined\n\
664 /* " // this is for working around an Emacs highlighting bug... */
666 static const char *Perl_suffixes
[] =
667 { "pl", "pm", NULL
};
668 static const char *Perl_interpreters
[] =
669 { "perl", "@PERL@", NULL
};
670 static const char Perl_help
[] =
671 "In Perl code, the tags are the packages, subroutines and variables\n\
672 defined by the `package', `sub', `my' and `local' keywords. Use\n\
673 `--globals' if you want to tag global variables. Tags for\n\
674 subroutines are named `PACKAGE::SUB'. The name for subroutines\n\
675 defined in the default package is `main::SUB'.";
677 static const char *PHP_suffixes
[] =
678 { "php", "php3", "php4", NULL
};
679 static const char PHP_help
[] =
680 "In PHP code, tags are functions, classes and defines. Unless you use\n\
681 the `--no-members' option, vars are tags too.";
683 static const char *plain_C_suffixes
[] =
684 { "pc", /* Pro*C file */
687 static const char *PS_suffixes
[] =
688 { "ps", "psw", NULL
}; /* .psw is for PSWrap */
689 static const char PS_help
[] =
690 "In PostScript code, the tags are the functions.";
692 static const char *Prolog_suffixes
[] =
694 static const char Prolog_help
[] =
695 "In Prolog code, tags are predicates and rules at the beginning of\n\
698 static const char *Python_suffixes
[] =
700 static const char Python_help
[] =
701 "In Python code, `def' or `class' at the beginning of a line\n\
704 /* Can't do the `SCM' or `scm' prefix with a version number. */
705 static const char *Scheme_suffixes
[] =
706 { "oak", "sch", "scheme", "SCM", "scm", "SM", "sm", "ss", "t", NULL
};
707 static const char Scheme_help
[] =
708 "In Scheme code, tags include anything defined with `def' or with a\n\
709 construct whose name starts with `def'. They also include\n\
710 variables set with `set!' at top level in the file.";
712 static const char *TeX_suffixes
[] =
713 { "bib", "clo", "cls", "ltx", "sty", "TeX", "tex", NULL
};
714 static const char TeX_help
[] =
715 "In LaTeX text, the argument of any of the commands `\\chapter',\n\
716 `\\section', `\\subsection', `\\subsubsection', `\\eqno', `\\label',\n\
717 `\\ref', `\\cite', `\\bibitem', `\\part', `\\appendix', `\\entry',\n\
718 `\\index', `\\def', `\\newcommand', `\\renewcommand',\n\
719 `\\newenvironment' or `\\renewenvironment' is a tag.\n\
721 Other commands can be specified by setting the environment variable\n\
722 `TEXTAGS' to a colon-separated list like, for example,\n\
723 TEXTAGS=\"mycommand:myothercommand\".";
726 static const char *Texinfo_suffixes
[] =
727 { "texi", "texinfo", "txi", NULL
};
728 static const char Texinfo_help
[] =
729 "for texinfo files, lines starting with @node are tagged.";
731 static const char *Yacc_suffixes
[] =
732 { "y", "y++", "ym", "yxx", "yy", NULL
}; /* .ym is Objective yacc file */
733 static const char Yacc_help
[] =
734 "In Bison or Yacc input files, each rule defines as a tag the\n\
735 nonterminal it constructs. The portions of the file that contain\n\
736 C code are parsed as C code (use --help --lang=c --lang=yacc\n\
739 static const char auto_help
[] =
740 "`auto' is not a real language, it indicates to use\n\
741 a default language for files base on file name suffix and file contents.";
743 static const char none_help
[] =
744 "`none' is not a real language, it indicates to only do\n\
745 regexp processing on files.";
747 static const char no_lang_help
[] =
748 "No detailed help available for this language.";
752 * Table of languages.
754 * It is ok for a given function to be listed under more than one
755 * name. I just didn't.
758 static language lang_names
[] =
760 { "ada", Ada_help
, Ada_funcs
, Ada_suffixes
},
761 { "asm", Asm_help
, Asm_labels
, Asm_suffixes
},
762 { "c", default_C_help
, default_C_entries
, default_C_suffixes
},
763 { "c++", Cplusplus_help
, Cplusplus_entries
, Cplusplus_suffixes
},
764 { "c*", no_lang_help
, Cstar_entries
, Cstar_suffixes
},
765 { "cobol", Cobol_help
, Cobol_paragraphs
, Cobol_suffixes
},
766 { "erlang", Erlang_help
, Erlang_functions
, Erlang_suffixes
},
767 { "forth", Forth_help
, Forth_words
, Forth_suffixes
},
768 { "fortran", Fortran_help
, Fortran_functions
, Fortran_suffixes
},
769 { "html", HTML_help
, HTML_labels
, HTML_suffixes
},
770 { "java", Cjava_help
, Cjava_entries
, Cjava_suffixes
},
771 { "lisp", Lisp_help
, Lisp_functions
, Lisp_suffixes
},
772 { "lua", Lua_help
, Lua_functions
, Lua_suffixes
},
773 { "makefile", Makefile_help
,Makefile_targets
,NULL
,Makefile_filenames
},
774 { "objc", Objc_help
, plain_C_entries
, Objc_suffixes
},
775 { "pascal", Pascal_help
, Pascal_functions
, Pascal_suffixes
},
776 { "perl",Perl_help
,Perl_functions
,Perl_suffixes
,NULL
,Perl_interpreters
},
777 { "php", PHP_help
, PHP_functions
, PHP_suffixes
},
778 { "postscript",PS_help
, PS_functions
, PS_suffixes
},
779 { "proc", no_lang_help
, plain_C_entries
, plain_C_suffixes
},
780 { "prolog", Prolog_help
, Prolog_functions
, Prolog_suffixes
},
781 { "python", Python_help
, Python_functions
, Python_suffixes
},
782 { "scheme", Scheme_help
, Scheme_functions
, Scheme_suffixes
},
783 { "tex", TeX_help
, TeX_commands
, TeX_suffixes
},
784 { "texinfo", Texinfo_help
, Texinfo_nodes
, Texinfo_suffixes
},
785 { "yacc", Yacc_help
,Yacc_entries
,Yacc_suffixes
,NULL
,NULL
,true},
786 { "auto", auto_help
}, /* default guessing scheme */
787 { "none", none_help
, just_read_file
}, /* regexp matching only */
788 { NULL
} /* end of list */
793 print_language_names (void)
796 const char **name
, **ext
;
798 puts ("\nThese are the currently supported languages, along with the\n\
799 default file names and dot suffixes:");
800 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
802 printf (" %-*s", 10, lang
->name
);
803 if (lang
->filenames
!= NULL
)
804 for (name
= lang
->filenames
; *name
!= NULL
; name
++)
805 printf (" %s", *name
);
806 if (lang
->suffixes
!= NULL
)
807 for (ext
= lang
->suffixes
; *ext
!= NULL
; ext
++)
808 printf (" .%s", *ext
);
811 puts ("where `auto' means use default language for files based on file\n\
812 name suffix, and `none' means only do regexp processing on files.\n\
813 If no language is specified and no matching suffix is found,\n\
814 the first line of the file is read for a sharp-bang (#!) sequence\n\
815 followed by the name of an interpreter. If no such sequence is found,\n\
816 Fortran is tried first; if no tags are found, C is tried next.\n\
817 When parsing any C file, a \"class\" or \"template\" keyword\n\
819 puts ("Compressed files are supported using gzip, bzip2, and xz.\n\
821 For detailed help on a given language use, for example,\n\
822 etags --help --lang=ada.");
826 # define EMACS_NAME "standalone"
829 # define VERSION "17.38.1.4"
831 static _Noreturn
void
834 char emacs_copyright
[] = COPYRIGHT
;
836 printf ("%s (%s %s)\n", (CTAGS
) ? "ctags" : "etags", EMACS_NAME
, VERSION
);
837 puts (emacs_copyright
);
838 puts ("This program is distributed under the terms in ETAGS.README");
843 #ifndef PRINT_UNDOCUMENTED_OPTIONS_HELP
844 # define PRINT_UNDOCUMENTED_OPTIONS_HELP false
847 static _Noreturn
void
848 print_help (argument
*argbuffer
)
850 bool help_for_lang
= false;
852 for (; argbuffer
->arg_type
!= at_end
; argbuffer
++)
853 if (argbuffer
->arg_type
== at_language
)
857 puts (argbuffer
->lang
->help
);
858 help_for_lang
= true;
864 printf ("Usage: %s [options] [[regex-option ...] file-name] ...\n\
866 These are the options accepted by %s.\n", progname
, progname
);
867 puts ("You may use unambiguous abbreviations for the long option names.");
868 puts (" A - as file name means read names from stdin (one per line).\n\
869 Absolute names are stored in the output file as they are.\n\
870 Relative ones are stored relative to the output file's directory.\n");
872 puts ("-a, --append\n\
873 Append tag entries to existing tags file.");
875 puts ("--packages-only\n\
876 For Ada files, only generate tags for packages.");
879 puts ("-B, --backward-search\n\
880 Write the search commands for the tag entries using '?', the\n\
881 backward-search command instead of '/', the forward-search command.");
883 /* This option is mostly obsolete, because etags can now automatically
884 detect C++. Retained for backward compatibility and for debugging and
885 experimentation. In principle, we could want to tag as C++ even
886 before any "class" or "template" keyword.
888 Treat files whose name suffix defaults to C language as C++ files.");
891 puts ("--declarations\n\
892 In C and derived languages, create tags for function declarations,");
894 puts ("\tand create tags for extern variables if --globals is used.");
897 ("\tand create tags for extern variables unless --no-globals is used.");
900 puts ("-d, --defines\n\
901 Create tag entries for C #define constants and enum constants, too.");
903 puts ("-D, --no-defines\n\
904 Don't create tag entries for C #define constants and enum constants.\n\
905 This makes the tags file smaller.");
908 puts ("-i FILE, --include=FILE\n\
909 Include a note in tag file indicating that, when searching for\n\
910 a tag, one should also consult the tags file FILE after\n\
911 checking the current file.");
913 puts ("-l LANG, --language=LANG\n\
914 Force the following files to be considered as written in the\n\
915 named language up to the next --language=LANG option.");
919 Create tag entries for global variables in some languages.");
921 puts ("--no-globals\n\
922 Do not create tag entries for global variables in some\n\
923 languages. This makes the tags file smaller.");
925 if (PRINT_UNDOCUMENTED_OPTIONS_HELP
)
926 puts ("--no-line-directive\n\
927 Ignore #line preprocessor directives in C and derived languages.");
931 Create tag entries for members of structures in some languages.");
933 puts ("--no-members\n\
934 Do not create tag entries for members of structures\n\
935 in some languages.");
937 puts ("-r REGEXP, --regex=REGEXP or --regex=@regexfile\n\
938 Make a tag for each line matching a regular expression pattern\n\
939 in the following files. {LANGUAGE}REGEXP uses REGEXP for LANGUAGE\n\
940 files only. REGEXFILE is a file containing one REGEXP per line.\n\
941 REGEXP takes the form /TAGREGEXP/TAGNAME/MODS, where TAGNAME/ is\n\
942 optional. The TAGREGEXP pattern is anchored (as if preceded by ^).");
943 puts (" If TAGNAME/ is present, the tags created are named.\n\
944 For example Tcl named tags can be created with:\n\
945 --regex=\"/proc[ \\t]+\\([^ \\t]+\\)/\\1/.\".\n\
946 MODS are optional one-letter modifiers: `i' means to ignore case,\n\
947 `m' means to allow multi-line matches, `s' implies `m' and\n\
948 causes dot to match any character, including newline.");
950 puts ("-R, --no-regex\n\
951 Don't create tags from regexps for the following files.");
953 puts ("-I, --ignore-indentation\n\
954 In C and C++ do not assume that a closing brace in the first\n\
955 column is the final brace of a function or structure definition.");
957 puts ("-o FILE, --output=FILE\n\
958 Write the tags to FILE.");
960 puts ("--parse-stdin=NAME\n\
961 Read from standard input and record tags as belonging to file NAME.");
965 puts ("-t, --typedefs\n\
966 Generate tag entries for C and Ada typedefs.");
967 puts ("-T, --typedefs-and-c++\n\
968 Generate tag entries for C typedefs, C struct/enum/union tags,\n\
969 and C++ member functions.");
973 puts ("-u, --update\n\
974 Update the tag entries for the given files, leaving tag\n\
975 entries for other files in place. Currently, this is\n\
976 implemented by deleting the existing entries for the given\n\
977 files and then rewriting the new entries at the end of the\n\
978 tags file. It is often faster to simply rebuild the entire\n\
979 tag file than to use this.");
983 puts ("-v, --vgrind\n\
984 Print on the standard output an index of items intended for\n\
985 human consumption, similar to the output of vgrind. The index\n\
986 is sorted, and gives the page number of each item.");
988 if (PRINT_UNDOCUMENTED_OPTIONS_HELP
)
989 puts ("-w, --no-duplicates\n\
990 Do not create duplicate tag entries, for compatibility with\n\
991 traditional ctags.");
993 if (PRINT_UNDOCUMENTED_OPTIONS_HELP
)
994 puts ("-w, --no-warn\n\
995 Suppress warning messages about duplicate tag entries.");
997 puts ("-x, --cxref\n\
998 Like --vgrind, but in the style of cxref, rather than vgrind.\n\
999 The output uses line numbers instead of page numbers, but\n\
1000 beyond that the differences are cosmetic; try both to see\n\
1004 puts ("-V, --version\n\
1005 Print the version of the program.\n\
1007 Print this help message.\n\
1008 Followed by one or more `--language' options prints detailed\n\
1009 help about tag generation for the specified languages.");
1011 print_language_names ();
1014 puts ("Report bugs to bug-gnu-emacs@gnu.org");
1016 exit (EXIT_SUCCESS
);
1021 main (int argc
, char **argv
)
1024 unsigned int nincluded_files
;
1025 char **included_files
;
1026 argument
*argbuffer
;
1027 int current_arg
, file_count
;
1028 linebuffer filename_lb
;
1029 bool help_asked
= false;
1035 nincluded_files
= 0;
1036 included_files
= xnew (argc
, char *);
1040 /* Allocate enough no matter what happens. Overkill, but each one
1042 argbuffer
= xnew (argc
, argument
);
1045 * Always find typedefs and structure tags.
1046 * Also default to find macro constants, enum constants, struct
1047 * members and global variables. Do it for both etags and ctags.
1049 typedefs
= typedefs_or_cplusplus
= constantypedefs
= true;
1050 globals
= members
= true;
1052 /* When the optstring begins with a '-' getopt_long does not rearrange the
1053 non-options arguments to be at the end, but leaves them alone. */
1054 optstring
= concat ("-ac:Cf:Il:o:r:RSVhH",
1055 (CTAGS
) ? "BxdtTuvw" : "Di:",
1058 while ((opt
= getopt_long (argc
, argv
, optstring
, longopts
, NULL
)) != EOF
)
1062 /* If getopt returns 0, then it has already processed a
1063 long-named option. We should do nothing. */
1067 /* This means that a file name has been seen. Record it. */
1068 argbuffer
[current_arg
].arg_type
= at_filename
;
1069 argbuffer
[current_arg
].what
= optarg
;
1070 len
= strlen (optarg
);
1071 if (whatlen_max
< len
)
1078 /* Parse standard input. Idea by Vivek <vivek@etla.org>. */
1079 argbuffer
[current_arg
].arg_type
= at_stdin
;
1080 argbuffer
[current_arg
].what
= optarg
;
1081 len
= strlen (optarg
);
1082 if (whatlen_max
< len
)
1087 fatal ("cannot parse standard input more than once", (char *)NULL
);
1088 parsing_stdin
= true;
1091 /* Common options. */
1092 case 'a': append_to_tagfile
= true; break;
1093 case 'C': cplusplus
= true; break;
1094 case 'f': /* for compatibility with old makefiles */
1098 error ("-o option may only be given once.");
1099 suggest_asking_for_help ();
1105 case 'S': /* for backward compatibility */
1106 ignoreindent
= true;
1110 language
*lang
= get_language_from_langname (optarg
);
1113 argbuffer
[current_arg
].lang
= lang
;
1114 argbuffer
[current_arg
].arg_type
= at_language
;
1120 /* Backward compatibility: support obsolete --ignore-case-regexp. */
1121 optarg
= concat (optarg
, "i", ""); /* memory leak here */
1124 argbuffer
[current_arg
].arg_type
= at_regexp
;
1125 argbuffer
[current_arg
].what
= optarg
;
1126 len
= strlen (optarg
);
1127 if (whatlen_max
< len
)
1132 argbuffer
[current_arg
].arg_type
= at_regexp
;
1133 argbuffer
[current_arg
].what
= NULL
;
1145 case 'D': constantypedefs
= false; break;
1146 case 'i': included_files
[nincluded_files
++] = optarg
; break;
1148 /* Ctags options. */
1149 case 'B': searchar
= '?'; break;
1150 case 'd': constantypedefs
= true; break;
1151 case 't': typedefs
= true; break;
1152 case 'T': typedefs
= typedefs_or_cplusplus
= true; break;
1153 case 'u': update
= true; break;
1154 case 'v': vgrind_style
= true; /*FALLTHRU*/
1155 case 'x': cxref_style
= true; break;
1156 case 'w': no_warnings
= true; break;
1158 suggest_asking_for_help ();
1162 /* No more options. Store the rest of arguments. */
1163 for (; optind
< argc
; optind
++)
1165 argbuffer
[current_arg
].arg_type
= at_filename
;
1166 argbuffer
[current_arg
].what
= argv
[optind
];
1167 len
= strlen (argv
[optind
]);
1168 if (whatlen_max
< len
)
1174 argbuffer
[current_arg
].arg_type
= at_end
;
1177 print_help (argbuffer
);
1180 if (nincluded_files
== 0 && file_count
== 0)
1182 error ("no input files specified.");
1183 suggest_asking_for_help ();
1187 if (tagfile
== NULL
)
1188 tagfile
= savestr (CTAGS
? "tags" : "TAGS");
1189 cwd
= etags_getcwd (); /* the current working directory */
1190 if (cwd
[strlen (cwd
) - 1] != '/')
1193 cwd
= concat (oldcwd
, "/", "");
1197 /* Compute base directory for relative file names. */
1198 if (streq (tagfile
, "-")
1199 || strneq (tagfile
, "/dev/", 5))
1200 tagfiledir
= cwd
; /* relative file names are relative to cwd */
1203 canonicalize_filename (tagfile
);
1204 tagfiledir
= absolute_dirname (tagfile
, cwd
);
1207 linebuffer_init (&lb
);
1208 linebuffer_init (&filename_lb
);
1209 linebuffer_init (&filebuf
);
1210 linebuffer_init (&token_name
);
1214 if (streq (tagfile
, "-"))
1217 SET_BINARY (fileno (stdout
));
1220 tagf
= fopen (tagfile
, append_to_tagfile
? "ab" : "wb");
1226 * Loop through files finding functions.
1228 for (i
= 0; i
< current_arg
; i
++)
1230 static language
*lang
; /* non-NULL if language is forced */
1233 switch (argbuffer
[i
].arg_type
)
1236 lang
= argbuffer
[i
].lang
;
1239 analyze_regex (argbuffer
[i
].what
);
1242 this_file
= argbuffer
[i
].what
;
1243 /* Input file named "-" means read file names from stdin
1244 (one per line) and use them. */
1245 if (streq (this_file
, "-"))
1248 fatal ("cannot parse standard input AND read file names from it",
1250 while (readline_internal (&filename_lb
, stdin
) > 0)
1251 process_file_name (filename_lb
.buffer
, lang
);
1254 process_file_name (this_file
, lang
);
1257 this_file
= argbuffer
[i
].what
;
1258 process_file (stdin
, this_file
, lang
);
1265 free (filebuf
.buffer
);
1266 free (token_name
.buffer
);
1268 if (!CTAGS
|| cxref_style
)
1270 /* Write the remaining tags to tagf (ETAGS) or stdout (CXREF). */
1271 put_entries (nodehead
);
1272 free_tree (nodehead
);
1278 /* Output file entries that have no tags. */
1279 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
1281 fprintf (tagf
, "\f\n%s,0\n", fdp
->taggedfname
);
1283 while (nincluded_files
-- > 0)
1284 fprintf (tagf
, "\f\n%s,include\n", *included_files
++);
1286 if (fclose (tagf
) == EOF
)
1290 exit (EXIT_SUCCESS
);
1293 /* From here on, we are in (CTAGS && !cxref_style) */
1297 xmalloc (strlen (tagfile
) + whatlen_max
+
1298 sizeof "mv..OTAGS;fgrep -v '\t\t' OTAGS >;rm OTAGS");
1299 for (i
= 0; i
< current_arg
; ++i
)
1301 switch (argbuffer
[i
].arg_type
)
1307 continue; /* the for loop */
1309 char *z
= stpcpy (cmd
, "mv ");
1310 z
= stpcpy (z
, tagfile
);
1311 z
= stpcpy (z
, " OTAGS;fgrep -v '\t");
1312 z
= stpcpy (z
, argbuffer
[i
].what
);
1313 z
= stpcpy (z
, "\t' OTAGS >");
1314 z
= stpcpy (z
, tagfile
);
1315 strcpy (z
, ";rm OTAGS");
1316 if (system (cmd
) != EXIT_SUCCESS
)
1317 fatal ("failed to execute shell command", (char *)NULL
);
1320 append_to_tagfile
= true;
1323 tagf
= fopen (tagfile
, append_to_tagfile
? "ab" : "wb");
1326 put_entries (nodehead
); /* write all the tags (CTAGS) */
1327 free_tree (nodehead
);
1329 if (fclose (tagf
) == EOF
)
1333 if (append_to_tagfile
|| update
)
1335 char *cmd
= xmalloc (2 * strlen (tagfile
) + sizeof "sort -u -o..");
1336 /* Maybe these should be used:
1337 setenv ("LC_COLLATE", "C", 1);
1338 setenv ("LC_ALL", "C", 1); */
1339 char *z
= stpcpy (cmd
, "sort -u -o ");
1340 z
= stpcpy (z
, tagfile
);
1342 strcpy (z
, tagfile
);
1343 exit (system (cmd
));
1345 return EXIT_SUCCESS
;
1350 * Return a compressor given the file name. If EXTPTR is non-zero,
1351 * return a pointer into FILE where the compressor-specific
1352 * extension begins. If no compressor is found, NULL is returned
1353 * and EXTPTR is not significant.
1354 * Idea by Vladimir Alexiev <vladimir@cs.ualberta.ca> (1998)
1357 get_compressor_from_suffix (char *file
, char **extptr
)
1360 char *slash
, *suffix
;
1362 /* File has been processed by canonicalize_filename,
1363 so we don't need to consider backslashes on DOS_NT. */
1364 slash
= strrchr (file
, '/');
1365 suffix
= strrchr (file
, '.');
1366 if (suffix
== NULL
|| suffix
< slash
)
1371 /* Let those poor souls who live with DOS 8+3 file name limits get
1372 some solace by treating foo.cgz as if it were foo.c.gz, etc.
1373 Only the first do loop is run if not MSDOS */
1376 for (compr
= compressors
; compr
->suffix
!= NULL
; compr
++)
1377 if (streq (compr
->suffix
, suffix
))
1380 break; /* do it only once: not really a loop */
1383 } while (*suffix
!= '\0');
1390 * Return a language given the name.
1393 get_language_from_langname (const char *name
)
1398 error ("empty language name");
1401 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1402 if (streq (name
, lang
->name
))
1404 error ("unknown language \"%s\"", name
);
1412 * Return a language given the interpreter name.
1415 get_language_from_interpreter (char *interpreter
)
1420 if (interpreter
== NULL
)
1422 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1423 if (lang
->interpreters
!= NULL
)
1424 for (iname
= lang
->interpreters
; *iname
!= NULL
; iname
++)
1425 if (streq (*iname
, interpreter
))
1434 * Return a language given the file name.
1437 get_language_from_filename (char *file
, int case_sensitive
)
1440 const char **name
, **ext
, *suffix
;
1442 /* Try whole file name first. */
1443 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1444 if (lang
->filenames
!= NULL
)
1445 for (name
= lang
->filenames
; *name
!= NULL
; name
++)
1446 if ((case_sensitive
)
1447 ? streq (*name
, file
)
1448 : strcaseeq (*name
, file
))
1451 /* If not found, try suffix after last dot. */
1452 suffix
= strrchr (file
, '.');
1456 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1457 if (lang
->suffixes
!= NULL
)
1458 for (ext
= lang
->suffixes
; *ext
!= NULL
; ext
++)
1459 if ((case_sensitive
)
1460 ? streq (*ext
, suffix
)
1461 : strcaseeq (*ext
, suffix
))
1468 * This routine is called on each file argument.
1471 process_file_name (char *file
, language
*lang
)
1473 struct stat stat_buf
;
1477 char *compressed_name
, *uncompressed_name
;
1478 char *ext
, *real_name
, *tmp_name
;
1481 canonicalize_filename (file
);
1482 if (streq (file
, tagfile
) && !streq (tagfile
, "-"))
1484 error ("skipping inclusion of %s in self.", file
);
1487 compr
= get_compressor_from_suffix (file
, &ext
);
1490 real_name
= compressed_name
= savestr (file
);
1491 uncompressed_name
= savenstr (file
, ext
- file
);
1495 compressed_name
= NULL
;
1496 real_name
= uncompressed_name
= savestr (file
);
1499 /* If the canonicalized uncompressed name
1500 has already been dealt with, skip it silently. */
1501 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
1503 assert (fdp
->infname
!= NULL
);
1504 if (streq (uncompressed_name
, fdp
->infname
))
1508 if (stat (real_name
, &stat_buf
) != 0)
1510 /* Reset real_name and try with a different name. */
1512 if (compressed_name
!= NULL
) /* try with the given suffix */
1514 if (stat (uncompressed_name
, &stat_buf
) == 0)
1515 real_name
= uncompressed_name
;
1517 else /* try all possible suffixes */
1519 for (compr
= compressors
; compr
->suffix
!= NULL
; compr
++)
1521 compressed_name
= concat (file
, ".", compr
->suffix
);
1522 if (stat (compressed_name
, &stat_buf
) != 0)
1526 char *suf
= compressed_name
+ strlen (file
);
1527 size_t suflen
= strlen (compr
->suffix
) + 1;
1528 for ( ; suf
[1]; suf
++, suflen
--)
1530 memmove (suf
, suf
+ 1, suflen
);
1531 if (stat (compressed_name
, &stat_buf
) == 0)
1533 real_name
= compressed_name
;
1537 if (real_name
!= NULL
)
1540 free (compressed_name
);
1541 compressed_name
= NULL
;
1545 real_name
= compressed_name
;
1550 if (real_name
== NULL
)
1555 } /* try with a different name */
1557 if (!S_ISREG (stat_buf
.st_mode
))
1559 error ("skipping %s: it is not a regular file.", real_name
);
1562 if (real_name
== compressed_name
)
1564 tmp_name
= etags_mktmp ();
1569 char *cmd1
= concat (compr
->command
, " ", real_name
);
1570 char *cmd
= concat (cmd1
, " > ", tmp_name
);
1572 if (system (cmd
) == -1)
1575 inf
= fopen (tmp_name
, "r" FOPEN_BINARY
);
1580 inf
= fopen (real_name
, "r" FOPEN_BINARY
);
1587 process_file (inf
, uncompressed_name
, lang
);
1589 retval
= fclose (inf
);
1590 if (real_name
== compressed_name
)
1599 free (compressed_name
);
1600 free (uncompressed_name
);
1607 process_file (FILE *fh
, char *fn
, language
*lang
)
1609 static const fdesc emptyfdesc
;
1612 /* Create a new input file description entry. */
1613 fdp
= xnew (1, fdesc
);
1616 fdp
->infname
= savestr (fn
);
1618 fdp
->infabsname
= absolute_filename (fn
, cwd
);
1619 fdp
->infabsdir
= absolute_dirname (fn
, cwd
);
1620 if (filename_is_absolute (fn
))
1622 /* An absolute file name. Canonicalize it. */
1623 fdp
->taggedfname
= absolute_filename (fn
, NULL
);
1627 /* A file name relative to cwd. Make it relative
1628 to the directory of the tags file. */
1629 fdp
->taggedfname
= relative_filename (fn
, tagfiledir
);
1631 fdp
->usecharno
= true; /* use char position when making tags */
1633 fdp
->written
= false; /* not written on tags file yet */
1636 curfdp
= fdhead
; /* the current file description */
1640 /* If not Ctags, and if this is not metasource and if it contained no #line
1641 directives, we can write the tags and free all nodes pointing to
1644 && curfdp
->usecharno
/* no #line directives in this file */
1645 && !curfdp
->lang
->metasource
)
1649 /* Look for the head of the sublist relative to this file. See add_node
1650 for the structure of the node tree. */
1652 for (np
= nodehead
; np
!= NULL
; prev
= np
, np
= np
->left
)
1653 if (np
->fdp
== curfdp
)
1656 /* If we generated tags for this file, write and delete them. */
1659 /* This is the head of the last sublist, if any. The following
1660 instructions depend on this being true. */
1661 assert (np
->left
== NULL
);
1663 assert (fdhead
== curfdp
);
1664 assert (last_node
->fdp
== curfdp
);
1665 put_entries (np
); /* write tags for file curfdp->taggedfname */
1666 free_tree (np
); /* remove the written nodes */
1668 nodehead
= NULL
; /* no nodes left */
1670 prev
->left
= NULL
; /* delete the pointer to the sublist */
1676 * This routine opens the specified file and calls the function
1677 * which finds the function and type definitions.
1680 find_entries (FILE *inf
)
1683 language
*lang
= curfdp
->lang
;
1684 Lang_function
*parser
= NULL
;
1686 /* If user specified a language, use it. */
1687 if (lang
!= NULL
&& lang
->function
!= NULL
)
1689 parser
= lang
->function
;
1692 /* Else try to guess the language given the file name. */
1695 lang
= get_language_from_filename (curfdp
->infname
, true);
1696 if (lang
!= NULL
&& lang
->function
!= NULL
)
1698 curfdp
->lang
= lang
;
1699 parser
= lang
->function
;
1703 /* Else look for sharp-bang as the first two characters. */
1705 && readline_internal (&lb
, inf
) > 0
1707 && lb
.buffer
[0] == '#'
1708 && lb
.buffer
[1] == '!')
1712 /* Set lp to point at the first char after the last slash in the
1713 line or, if no slashes, at the first nonblank. Then set cp to
1714 the first successive blank and terminate the string. */
1715 lp
= strrchr (lb
.buffer
+2, '/');
1719 lp
= skip_spaces (lb
.buffer
+ 2);
1720 cp
= skip_non_spaces (lp
);
1723 if (strlen (lp
) > 0)
1725 lang
= get_language_from_interpreter (lp
);
1726 if (lang
!= NULL
&& lang
->function
!= NULL
)
1728 curfdp
->lang
= lang
;
1729 parser
= lang
->function
;
1736 /* Else try to guess the language given the case insensitive file name. */
1739 lang
= get_language_from_filename (curfdp
->infname
, false);
1740 if (lang
!= NULL
&& lang
->function
!= NULL
)
1742 curfdp
->lang
= lang
;
1743 parser
= lang
->function
;
1747 /* Else try Fortran or C. */
1750 node
*old_last_node
= last_node
;
1752 curfdp
->lang
= get_language_from_langname ("fortran");
1755 if (old_last_node
== last_node
)
1756 /* No Fortran entries found. Try C. */
1759 curfdp
->lang
= get_language_from_langname (cplusplus
? "c++" : "c");
1765 if (!no_line_directive
1766 && curfdp
->lang
!= NULL
&& curfdp
->lang
->metasource
)
1767 /* It may be that this is a bingo.y file, and we already parsed a bingo.c
1768 file, or anyway we parsed a file that is automatically generated from
1769 this one. If this is the case, the bingo.c file contained #line
1770 directives that generated tags pointing to this file. Let's delete
1771 them all before parsing this file, which is the real source. */
1773 fdesc
**fdpp
= &fdhead
;
1774 while (*fdpp
!= NULL
)
1776 && streq ((*fdpp
)->taggedfname
, curfdp
->taggedfname
))
1777 /* We found one of those! We must delete both the file description
1778 and all tags referring to it. */
1780 fdesc
*badfdp
= *fdpp
;
1782 /* Delete the tags referring to badfdp->taggedfname
1783 that were obtained from badfdp->infname. */
1784 invalidate_nodes (badfdp
, &nodehead
);
1786 *fdpp
= badfdp
->next
; /* remove the bad description from the list */
1787 free_fdesc (badfdp
);
1790 fdpp
= &(*fdpp
)->next
; /* advance the list pointer */
1793 assert (parser
!= NULL
);
1795 /* Generic initializations before reading from file. */
1796 linebuffer_setlen (&filebuf
, 0); /* reset the file buffer */
1798 /* Generic initializations before parsing file with readline. */
1799 lineno
= 0; /* reset global line number */
1800 charno
= 0; /* reset global char number */
1801 linecharno
= 0; /* reset global char number of line start */
1805 regex_tag_multiline ();
1810 * Check whether an implicitly named tag should be created,
1811 * then call `pfnote'.
1812 * NAME is a string that is internally copied by this function.
1814 * TAGS format specification
1815 * Idea by Sam Kendall <kendall@mv.mv.com> (1997)
1816 * The following is explained in some more detail in etc/ETAGS.EBNF.
1818 * make_tag creates tags with "implicit tag names" (unnamed tags)
1819 * if the following are all true, assuming NONAM=" \f\t\n\r()=,;":
1820 * 1. NAME does not contain any of the characters in NONAM;
1821 * 2. LINESTART contains name as either a rightmost, or rightmost but
1822 * one character, substring;
1823 * 3. the character, if any, immediately before NAME in LINESTART must
1824 * be a character in NONAM;
1825 * 4. the character, if any, immediately after NAME in LINESTART must
1826 * also be a character in NONAM.
1828 * The implementation uses the notinname() macro, which recognizes the
1829 * characters stored in the string `nonam'.
1830 * etags.el needs to use the same characters that are in NONAM.
1833 make_tag (const char *name
, /* tag name, or NULL if unnamed */
1834 int namelen
, /* tag length */
1835 bool is_func
, /* tag is a function */
1836 char *linestart
, /* start of the line where tag is */
1837 int linelen
, /* length of the line where tag is */
1838 int lno
, /* line number */
1839 long int cno
) /* character number */
1841 bool named
= (name
!= NULL
&& namelen
> 0);
1844 if (!CTAGS
&& named
) /* maybe set named to false */
1845 /* Let's try to make an implicit tag name, that is, create an unnamed tag
1846 such that etags.el can guess a name from it. */
1849 register const char *cp
= name
;
1851 for (i
= 0; i
< namelen
; i
++)
1852 if (notinname (*cp
++))
1854 if (i
== namelen
) /* rule #1 */
1856 cp
= linestart
+ linelen
- namelen
;
1857 if (notinname (linestart
[linelen
-1]))
1858 cp
-= 1; /* rule #4 */
1859 if (cp
>= linestart
/* rule #2 */
1861 || notinname (cp
[-1])) /* rule #3 */
1862 && strneq (name
, cp
, namelen
)) /* rule #2 */
1863 named
= false; /* use implicit tag name */
1868 nname
= savenstr (name
, namelen
);
1870 pfnote (nname
, is_func
, linestart
, linelen
, lno
, cno
);
1875 pfnote (char *name
, bool is_func
, char *linestart
, int linelen
, int lno
,
1877 /* tag name, or NULL if unnamed */
1878 /* tag is a function */
1879 /* start of the line where tag is */
1880 /* length of the line where tag is */
1882 /* character number */
1886 assert (name
== NULL
|| name
[0] != '\0');
1887 if (CTAGS
&& name
== NULL
)
1890 np
= xnew (1, node
);
1892 /* If ctags mode, change name "main" to M<thisfilename>. */
1893 if (CTAGS
&& !cxref_style
&& streq (name
, "main"))
1895 char *fp
= strrchr (curfdp
->taggedfname
, '/');
1896 np
->name
= concat ("M", fp
== NULL
? curfdp
->taggedfname
: fp
+ 1, "");
1897 fp
= strrchr (np
->name
, '.');
1898 if (fp
!= NULL
&& fp
[1] != '\0' && fp
[2] == '\0')
1904 np
->been_warned
= false;
1906 np
->is_func
= is_func
;
1908 if (np
->fdp
->usecharno
)
1909 /* Our char numbers are 0-base, because of C language tradition?
1910 ctags compatibility? old versions compatibility? I don't know.
1911 Anyway, since emacs's are 1-base we expect etags.el to take care
1912 of the difference. If we wanted to have 1-based numbers, we would
1913 uncomment the +1 below. */
1914 np
->cno
= cno
/* + 1 */ ;
1916 np
->cno
= invalidcharno
;
1917 np
->left
= np
->right
= NULL
;
1918 if (CTAGS
&& !cxref_style
)
1920 if (strlen (linestart
) < 50)
1921 np
->regex
= concat (linestart
, "$", "");
1923 np
->regex
= savenstr (linestart
, 50);
1926 np
->regex
= savenstr (linestart
, linelen
);
1928 add_node (np
, &nodehead
);
1933 * recurse on left children, iterate on right children.
1936 free_tree (register node
*np
)
1940 register node
*node_right
= np
->right
;
1941 free_tree (np
->left
);
1951 * delete a file description
1954 free_fdesc (register fdesc
*fdp
)
1956 free (fdp
->infname
);
1957 free (fdp
->infabsname
);
1958 free (fdp
->infabsdir
);
1959 free (fdp
->taggedfname
);
1966 * Adds a node to the tree of nodes. In etags mode, sort by file
1967 * name. In ctags mode, sort by tag name. Make no attempt at
1970 * add_node is the only function allowed to add nodes, so it can
1974 add_node (node
*np
, node
**cur_node_p
)
1977 register node
*cur_node
= *cur_node_p
;
1979 if (cur_node
== NULL
)
1989 /* For each file name, tags are in a linked sublist on the right
1990 pointer. The first tags of different files are a linked list
1991 on the left pointer. last_node points to the end of the last
1993 if (last_node
!= NULL
&& last_node
->fdp
== np
->fdp
)
1995 /* Let's use the same sublist as the last added node. */
1996 assert (last_node
->right
== NULL
);
1997 last_node
->right
= np
;
2000 else if (cur_node
->fdp
== np
->fdp
)
2002 /* Scanning the list we found the head of a sublist which is
2003 good for us. Let's scan this sublist. */
2004 add_node (np
, &cur_node
->right
);
2007 /* The head of this sublist is not good for us. Let's try the
2009 add_node (np
, &cur_node
->left
);
2010 } /* if ETAGS mode */
2015 dif
= strcmp (np
->name
, cur_node
->name
);
2018 * If this tag name matches an existing one, then
2019 * do not add the node, but maybe print a warning.
2021 if (no_duplicates
&& !dif
)
2023 if (np
->fdp
== cur_node
->fdp
)
2027 fprintf (stderr
, "Duplicate entry in file %s, line %d: %s\n",
2028 np
->fdp
->infname
, lineno
, np
->name
);
2029 fprintf (stderr
, "Second entry ignored\n");
2032 else if (!cur_node
->been_warned
&& !no_warnings
)
2036 "Duplicate entry in files %s and %s: %s (Warning only)\n",
2037 np
->fdp
->infname
, cur_node
->fdp
->infname
, np
->name
);
2038 cur_node
->been_warned
= true;
2043 /* Actually add the node */
2044 add_node (np
, dif
< 0 ? &cur_node
->left
: &cur_node
->right
);
2045 } /* if CTAGS mode */
2049 * invalidate_nodes ()
2050 * Scan the node tree and invalidate all nodes pointing to the
2051 * given file description (CTAGS case) or free them (ETAGS case).
2054 invalidate_nodes (fdesc
*badfdp
, node
**npp
)
2063 if (np
->left
!= NULL
)
2064 invalidate_nodes (badfdp
, &np
->left
);
2065 if (np
->fdp
== badfdp
)
2067 if (np
->right
!= NULL
)
2068 invalidate_nodes (badfdp
, &np
->right
);
2072 assert (np
->fdp
!= NULL
);
2073 if (np
->fdp
== badfdp
)
2075 *npp
= np
->left
; /* detach the sublist from the list */
2076 np
->left
= NULL
; /* isolate it */
2077 free_tree (np
); /* free it */
2078 invalidate_nodes (badfdp
, npp
);
2081 invalidate_nodes (badfdp
, &np
->left
);
2086 static int total_size_of_entries (node
*);
2087 static int number_len (long) ATTRIBUTE_CONST
;
2089 /* Length of a non-negative number's decimal representation. */
2091 number_len (long int num
)
2094 while ((num
/= 10) > 0)
2100 * Return total number of characters that put_entries will output for
2101 * the nodes in the linked list at the right of the specified node.
2102 * This count is irrelevant with etags.el since emacs 19.34 at least,
2103 * but is still supplied for backward compatibility.
2106 total_size_of_entries (register node
*np
)
2108 register int total
= 0;
2110 for (; np
!= NULL
; np
= np
->right
)
2113 total
+= strlen (np
->regex
) + 1; /* pat\177 */
2114 if (np
->name
!= NULL
)
2115 total
+= strlen (np
->name
) + 1; /* name\001 */
2116 total
+= number_len ((long) np
->lno
) + 1; /* lno, */
2117 if (np
->cno
!= invalidcharno
) /* cno */
2118 total
+= number_len (np
->cno
);
2119 total
+= 1; /* newline */
2126 put_entries (register node
*np
)
2129 static fdesc
*fdp
= NULL
;
2134 /* Output subentries that precede this one */
2136 put_entries (np
->left
);
2138 /* Output this entry */
2147 fprintf (tagf
, "\f\n%s,%d\n",
2148 fdp
->taggedfname
, total_size_of_entries (np
));
2149 fdp
->written
= true;
2151 fputs (np
->regex
, tagf
);
2152 fputc ('\177', tagf
);
2153 if (np
->name
!= NULL
)
2155 fputs (np
->name
, tagf
);
2156 fputc ('\001', tagf
);
2158 fprintf (tagf
, "%d,", np
->lno
);
2159 if (np
->cno
!= invalidcharno
)
2160 fprintf (tagf
, "%ld", np
->cno
);
2166 if (np
->name
== NULL
)
2167 error ("internal error: NULL name in ctags mode.");
2172 fprintf (stdout
, "%s %s %d\n",
2173 np
->name
, np
->fdp
->taggedfname
, (np
->lno
+ 63) / 64);
2175 fprintf (stdout
, "%-16s %3d %-16s %s\n",
2176 np
->name
, np
->lno
, np
->fdp
->taggedfname
, np
->regex
);
2180 fprintf (tagf
, "%s\t%s\t", np
->name
, np
->fdp
->taggedfname
);
2183 { /* function or #define macro with args */
2184 putc (searchar
, tagf
);
2187 for (sp
= np
->regex
; *sp
; sp
++)
2189 if (*sp
== '\\' || *sp
== searchar
)
2193 putc (searchar
, tagf
);
2196 { /* anything else; text pattern inadequate */
2197 fprintf (tagf
, "%d", np
->lno
);
2202 } /* if this node contains a valid tag */
2204 /* Output subentries that follow this one */
2205 put_entries (np
->right
);
2207 put_entries (np
->left
);
2212 #define C_EXT 0x00fff /* C extensions */
2213 #define C_PLAIN 0x00000 /* C */
2214 #define C_PLPL 0x00001 /* C++ */
2215 #define C_STAR 0x00003 /* C* */
2216 #define C_JAVA 0x00005 /* JAVA */
2217 #define C_AUTO 0x01000 /* C, but switch to C++ if `class' is met */
2218 #define YACC 0x10000 /* yacc file */
2221 * The C symbol tables.
2226 st_C_objprot
, st_C_objimpl
, st_C_objend
,
2228 st_C_ignore
, st_C_attribute
,
2231 st_C_class
, st_C_template
,
2232 st_C_struct
, st_C_extern
, st_C_enum
, st_C_define
, st_C_typedef
2235 /* Feed stuff between (but not including) %[ and %] lines to:
2241 struct C_stab_entry { char *name; int c_ext; enum sym_type type; }
2245 while, 0, st_C_ignore
2246 switch, 0, st_C_ignore
2247 return, 0, st_C_ignore
2248 __attribute__, 0, st_C_attribute
2249 GTY, 0, st_C_attribute
2250 @interface, 0, st_C_objprot
2251 @protocol, 0, st_C_objprot
2252 @implementation,0, st_C_objimpl
2253 @end, 0, st_C_objend
2254 import, (C_JAVA & ~C_PLPL), st_C_ignore
2255 package, (C_JAVA & ~C_PLPL), st_C_ignore
2256 friend, C_PLPL, st_C_ignore
2257 extends, (C_JAVA & ~C_PLPL), st_C_javastruct
2258 implements, (C_JAVA & ~C_PLPL), st_C_javastruct
2259 interface, (C_JAVA & ~C_PLPL), st_C_struct
2260 class, 0, st_C_class
2261 namespace, C_PLPL, st_C_struct
2262 domain, C_STAR, st_C_struct
2263 union, 0, st_C_struct
2264 struct, 0, st_C_struct
2265 extern, 0, st_C_extern
2267 typedef, 0, st_C_typedef
2268 define, 0, st_C_define
2269 undef, 0, st_C_define
2270 operator, C_PLPL, st_C_operator
2271 template, 0, st_C_template
2272 # DEFUN used in emacs, the next three used in glibc (SYSCALL only for mach).
2273 DEFUN, 0, st_C_gnumacro
2274 SYSCALL, 0, st_C_gnumacro
2275 ENTRY, 0, st_C_gnumacro
2276 PSEUDO, 0, st_C_gnumacro
2277 # These are defined inside C functions, so currently they are not met.
2278 # EXFUN used in glibc, DEFVAR_* in emacs.
2279 #EXFUN, 0, st_C_gnumacro
2280 #DEFVAR_, 0, st_C_gnumacro
2282 and replace lines between %< and %> with its output, then:
2283 - remove the #if characterset check
2284 - make in_word_set static and not inline. */
2286 /* C code produced by gperf version 3.0.1 */
2287 /* Command-line: gperf -m 5 */
2288 /* Computed positions: -k'2-3' */
2290 struct C_stab_entry
{ const char *name
; int c_ext
; enum sym_type type
; };
2291 /* maximum key range = 33, duplicates = 0 */
2294 hash (const char *str
, int len
)
2296 static char const asso_values
[] =
2298 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2299 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2300 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2301 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2302 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2303 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2304 35, 35, 35, 35, 35, 35, 35, 35, 35, 3,
2305 26, 35, 35, 35, 35, 35, 35, 35, 27, 35,
2306 35, 35, 35, 24, 0, 35, 35, 35, 35, 0,
2307 35, 35, 35, 35, 35, 1, 35, 16, 35, 6,
2308 23, 0, 0, 35, 22, 0, 35, 35, 5, 0,
2309 0, 15, 1, 35, 6, 35, 8, 19, 35, 16,
2310 4, 5, 35, 35, 35, 35, 35, 35, 35, 35,
2311 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2312 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2313 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2314 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2315 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2316 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2317 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2318 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2319 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2320 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2321 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2322 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2323 35, 35, 35, 35, 35, 35
2330 hval
+= asso_values
[(unsigned char) str
[2]];
2333 hval
+= asso_values
[(unsigned char) str
[1]];
2339 static struct C_stab_entry
*
2340 in_word_set (register const char *str
, register unsigned int len
)
2344 TOTAL_KEYWORDS
= 33,
2345 MIN_WORD_LENGTH
= 2,
2346 MAX_WORD_LENGTH
= 15,
2351 static struct C_stab_entry wordlist
[] =
2354 {"if", 0, st_C_ignore
},
2355 {"GTY", 0, st_C_attribute
},
2356 {"@end", 0, st_C_objend
},
2357 {"union", 0, st_C_struct
},
2358 {"define", 0, st_C_define
},
2359 {"import", (C_JAVA
& ~C_PLPL
), st_C_ignore
},
2360 {"template", 0, st_C_template
},
2361 {"operator", C_PLPL
, st_C_operator
},
2362 {"@interface", 0, st_C_objprot
},
2363 {"implements", (C_JAVA
& ~C_PLPL
), st_C_javastruct
},
2364 {"friend", C_PLPL
, st_C_ignore
},
2365 {"typedef", 0, st_C_typedef
},
2366 {"return", 0, st_C_ignore
},
2367 {"@implementation",0, st_C_objimpl
},
2368 {"@protocol", 0, st_C_objprot
},
2369 {"interface", (C_JAVA
& ~C_PLPL
), st_C_struct
},
2370 {"extern", 0, st_C_extern
},
2371 {"extends", (C_JAVA
& ~C_PLPL
), st_C_javastruct
},
2372 {"struct", 0, st_C_struct
},
2373 {"domain", C_STAR
, st_C_struct
},
2374 {"switch", 0, st_C_ignore
},
2375 {"enum", 0, st_C_enum
},
2376 {"for", 0, st_C_ignore
},
2377 {"namespace", C_PLPL
, st_C_struct
},
2378 {"class", 0, st_C_class
},
2379 {"while", 0, st_C_ignore
},
2380 {"undef", 0, st_C_define
},
2381 {"package", (C_JAVA
& ~C_PLPL
), st_C_ignore
},
2382 {"__attribute__", 0, st_C_attribute
},
2383 {"SYSCALL", 0, st_C_gnumacro
},
2384 {"ENTRY", 0, st_C_gnumacro
},
2385 {"PSEUDO", 0, st_C_gnumacro
},
2386 {"DEFUN", 0, st_C_gnumacro
}
2389 if (len
<= MAX_WORD_LENGTH
&& len
>= MIN_WORD_LENGTH
)
2391 int key
= hash (str
, len
);
2393 if (key
<= MAX_HASH_VALUE
&& key
>= 0)
2395 const char *s
= wordlist
[key
].name
;
2397 if (*str
== *s
&& !strncmp (str
+ 1, s
+ 1, len
- 1) && s
[len
] == '\0')
2398 return &wordlist
[key
];
2405 static enum sym_type
2406 C_symtype (char *str
, int len
, int c_ext
)
2408 register struct C_stab_entry
*se
= in_word_set (str
, len
);
2410 if (se
== NULL
|| (se
->c_ext
&& !(c_ext
& se
->c_ext
)))
2417 * Ignoring __attribute__ ((list))
2419 static bool inattribute
; /* looking at an __attribute__ construct */
2422 * C functions and variables are recognized using a simple
2423 * finite automaton. fvdef is its state variable.
2427 fvnone
, /* nothing seen */
2428 fdefunkey
, /* Emacs DEFUN keyword seen */
2429 fdefunname
, /* Emacs DEFUN name seen */
2430 foperator
, /* func: operator keyword seen (cplpl) */
2431 fvnameseen
, /* function or variable name seen */
2432 fstartlist
, /* func: just after open parenthesis */
2433 finlist
, /* func: in parameter list */
2434 flistseen
, /* func: after parameter list */
2435 fignore
, /* func: before open brace */
2436 vignore
/* var-like: ignore until ';' */
2439 static bool fvextern
; /* func or var: extern keyword seen; */
2442 * typedefs are recognized using a simple finite automaton.
2443 * typdef is its state variable.
2447 tnone
, /* nothing seen */
2448 tkeyseen
, /* typedef keyword seen */
2449 ttypeseen
, /* defined type seen */
2450 tinbody
, /* inside typedef body */
2451 tend
, /* just before typedef tag */
2452 tignore
/* junk after typedef tag */
2456 * struct-like structures (enum, struct and union) are recognized
2457 * using another simple finite automaton. `structdef' is its state
2462 snone
, /* nothing seen yet,
2463 or in struct body if bracelev > 0 */
2464 skeyseen
, /* struct-like keyword seen */
2465 stagseen
, /* struct-like tag seen */
2466 scolonseen
/* colon seen after struct-like tag */
2470 * When objdef is different from onone, objtag is the name of the class.
2472 static const char *objtag
= "<uninited>";
2475 * Yet another little state machine to deal with preprocessor lines.
2479 dnone
, /* nothing seen */
2480 dsharpseen
, /* '#' seen as first char on line */
2481 ddefineseen
, /* '#' and 'define' seen */
2482 dignorerest
/* ignore rest of line */
2486 * State machine for Objective C protocols and implementations.
2487 * Idea by Tom R.Hageman <tom@basil.icce.rug.nl> (1995)
2491 onone
, /* nothing seen */
2492 oprotocol
, /* @interface or @protocol seen */
2493 oimplementation
, /* @implementations seen */
2494 otagseen
, /* class name seen */
2495 oparenseen
, /* parenthesis before category seen */
2496 ocatseen
, /* category name seen */
2497 oinbody
, /* in @implementation body */
2498 omethodsign
, /* in @implementation body, after +/- */
2499 omethodtag
, /* after method name */
2500 omethodcolon
, /* after method colon */
2501 omethodparm
, /* after method parameter */
2502 oignore
/* wait for @end */
2507 * Use this structure to keep info about the token read, and how it
2508 * should be tagged. Used by the make_C_tag function to build a tag.
2512 char *line
; /* string containing the token */
2513 int offset
; /* where the token starts in LINE */
2514 int length
; /* token length */
2516 The previous members can be used to pass strings around for generic
2517 purposes. The following ones specifically refer to creating tags. In this
2518 case the token contained here is the pattern that will be used to create a
2521 bool valid
; /* do not create a tag; the token should be
2522 invalidated whenever a state machine is
2523 reset prematurely */
2524 bool named
; /* create a named tag */
2525 int lineno
; /* source line number of tag */
2526 long linepos
; /* source char number of tag */
2527 } token
; /* latest token read */
2530 * Variables and functions for dealing with nested structures.
2531 * Idea by Mykola Dzyuba <mdzyuba@yahoo.com> (2001)
2533 static void pushclass_above (int, char *, int);
2534 static void popclass_above (int);
2535 static void write_classname (linebuffer
*, const char *qualifier
);
2538 char **cname
; /* nested class names */
2539 int *bracelev
; /* nested class brace level */
2540 int nl
; /* class nesting level (elements used) */
2541 int size
; /* length of the array */
2542 } cstack
; /* stack for nested declaration tags */
2543 /* Current struct nesting depth (namespace, class, struct, union, enum). */
2544 #define nestlev (cstack.nl)
2545 /* After struct keyword or in struct body, not inside a nested function. */
2546 #define instruct (structdef == snone && nestlev > 0 \
2547 && bracelev == cstack.bracelev[nestlev-1] + 1)
2550 pushclass_above (int bracelev
, char *str
, int len
)
2554 popclass_above (bracelev
);
2556 if (nl
>= cstack
.size
)
2558 int size
= cstack
.size
*= 2;
2559 xrnew (cstack
.cname
, size
, char *);
2560 xrnew (cstack
.bracelev
, size
, int);
2562 assert (nl
== 0 || cstack
.bracelev
[nl
-1] < bracelev
);
2563 cstack
.cname
[nl
] = (str
== NULL
) ? NULL
: savenstr (str
, len
);
2564 cstack
.bracelev
[nl
] = bracelev
;
2569 popclass_above (int bracelev
)
2573 for (nl
= cstack
.nl
- 1;
2574 nl
>= 0 && cstack
.bracelev
[nl
] >= bracelev
;
2577 free (cstack
.cname
[nl
]);
2583 write_classname (linebuffer
*cn
, const char *qualifier
)
2586 int qlen
= strlen (qualifier
);
2588 if (cstack
.nl
== 0 || cstack
.cname
[0] == NULL
)
2592 cn
->buffer
[0] = '\0';
2596 len
= strlen (cstack
.cname
[0]);
2597 linebuffer_setlen (cn
, len
);
2598 strcpy (cn
->buffer
, cstack
.cname
[0]);
2600 for (i
= 1; i
< cstack
.nl
; i
++)
2602 char *s
= cstack
.cname
[i
];
2605 linebuffer_setlen (cn
, len
+ qlen
+ strlen (s
));
2606 len
+= sprintf (cn
->buffer
+ len
, "%s%s", qualifier
, s
);
2611 static bool consider_token (char *, int, int, int *, int, int, bool *);
2612 static void make_C_tag (bool);
2616 * checks to see if the current token is at the start of a
2617 * function or variable, or corresponds to a typedef, or
2618 * is a struct/union/enum tag, or #define, or an enum constant.
2620 * *IS_FUNC_OR_VAR gets true if the token is a function or #define macro
2621 * with args. C_EXTP points to which language we are looking at.
2632 consider_token (char *str
, int len
, int c
, int *c_extp
,
2633 int bracelev
, int parlev
, bool *is_func_or_var
)
2634 /* IN: token pointer */
2635 /* IN: token length */
2636 /* IN: first char after the token */
2637 /* IN, OUT: C extensions mask */
2638 /* IN: brace level */
2639 /* IN: parenthesis level */
2640 /* OUT: function or variable found */
2642 /* When structdef is stagseen, scolonseen, or snone with bracelev > 0,
2643 structtype is the type of the preceding struct-like keyword, and
2644 structbracelev is the brace level where it has been seen. */
2645 static enum sym_type structtype
;
2646 static int structbracelev
;
2647 static enum sym_type toktype
;
2650 toktype
= C_symtype (str
, len
, *c_extp
);
2653 * Skip __attribute__
2655 if (toktype
== st_C_attribute
)
2662 * Advance the definedef state machine.
2667 /* We're not on a preprocessor line. */
2668 if (toktype
== st_C_gnumacro
)
2675 if (toktype
== st_C_define
)
2677 definedef
= ddefineseen
;
2681 definedef
= dignorerest
;
2686 * Make a tag for any macro, unless it is a constant
2687 * and constantypedefs is false.
2689 definedef
= dignorerest
;
2690 *is_func_or_var
= (c
== '(');
2691 if (!*is_func_or_var
&& !constantypedefs
)
2698 error ("internal error: definedef value.");
2707 if (toktype
== st_C_typedef
)
2727 if (structdef
== snone
&& fvdef
== fvnone
)
2746 case st_C_javastruct
:
2747 if (structdef
== stagseen
)
2748 structdef
= scolonseen
;
2752 if ((*c_extp
& C_AUTO
) /* automatic detection of C++ language */
2754 && definedef
== dnone
&& structdef
== snone
2755 && typdef
== tnone
&& fvdef
== fvnone
)
2756 *c_extp
= (*c_extp
| C_PLPL
) & ~C_AUTO
;
2757 if (toktype
== st_C_template
)
2764 && (typdef
== tkeyseen
2765 || (typedefs_or_cplusplus
&& structdef
== snone
)))
2767 structdef
= skeyseen
;
2768 structtype
= toktype
;
2769 structbracelev
= bracelev
;
2770 if (fvdef
== fvnameseen
)
2776 if (structdef
== skeyseen
)
2778 structdef
= stagseen
;
2782 if (typdef
!= tnone
)
2785 /* Detect Objective C constructs. */
2795 objdef
= oimplementation
;
2799 case oimplementation
:
2800 /* Save the class tag for functions or variables defined inside. */
2801 objtag
= savenstr (str
, len
);
2805 /* Save the class tag for categories. */
2806 objtag
= savenstr (str
, len
);
2808 *is_func_or_var
= true;
2812 *is_func_or_var
= true;
2820 objdef
= omethodtag
;
2821 linebuffer_setlen (&token_name
, len
);
2822 memcpy (token_name
.buffer
, str
, len
);
2823 token_name
.buffer
[len
] = '\0';
2829 objdef
= omethodparm
;
2834 int oldlen
= token_name
.len
;
2836 objdef
= omethodtag
;
2837 linebuffer_setlen (&token_name
, oldlen
+ len
);
2838 memcpy (token_name
.buffer
+ oldlen
, str
, len
);
2839 token_name
.buffer
[oldlen
+ len
] = '\0';
2844 if (toktype
== st_C_objend
)
2846 /* Memory leakage here: the string pointed by objtag is
2847 never released, because many tests would be needed to
2848 avoid breaking on incorrect input code. The amount of
2849 memory leaked here is the sum of the lengths of the
2857 /* A function, variable or enum constant? */
2879 *is_func_or_var
= true;
2883 && structdef
== snone
2884 && structtype
== st_C_enum
&& bracelev
> structbracelev
2885 /* Don't tag tokens in expressions that assign values to enum
2887 && fvdef
!= vignore
)
2888 return true; /* enum constant */
2894 fvdef
= fdefunname
; /* GNU macro */
2895 *is_func_or_var
= true;
2903 if ((strneq (str
, "asm", 3) && endtoken (str
[3]))
2904 || (strneq (str
, "__asm__", 7) && endtoken (str
[7])))
2913 if (len
>= 10 && strneq (str
+len
-10, "::operator", 10))
2915 if (*c_extp
& C_AUTO
) /* automatic detection of C++ */
2916 *c_extp
= (*c_extp
| C_PLPL
) & ~C_AUTO
;
2918 *is_func_or_var
= true;
2921 if (bracelev
> 0 && !instruct
)
2923 fvdef
= fvnameseen
; /* function or variable */
2924 *is_func_or_var
= true;
2935 * C_entries often keeps pointers to tokens or lines which are older than
2936 * the line currently read. By keeping two line buffers, and switching
2937 * them at end of line, it is possible to use those pointers.
2945 #define current_lb_is_new (newndx == curndx)
2946 #define switch_line_buffers() (curndx = 1 - curndx)
2948 #define curlb (lbs[curndx].lb)
2949 #define newlb (lbs[newndx].lb)
2950 #define curlinepos (lbs[curndx].linepos)
2951 #define newlinepos (lbs[newndx].linepos)
2953 #define plainc ((c_ext & C_EXT) == C_PLAIN)
2954 #define cplpl (c_ext & C_PLPL)
2955 #define cjava ((c_ext & C_JAVA) == C_JAVA)
2957 #define CNL_SAVE_DEFINEDEF() \
2959 curlinepos = charno; \
2960 readline (&curlb, inf); \
2961 lp = curlb.buffer; \
2968 CNL_SAVE_DEFINEDEF(); \
2969 if (savetoken.valid) \
2971 token = savetoken; \
2972 savetoken.valid = false; \
2974 definedef = dnone; \
2979 make_C_tag (bool isfun
)
2981 /* This function is never called when token.valid is false, but
2982 we must protect against invalid input or internal errors. */
2984 make_tag (token_name
.buffer
, token_name
.len
, isfun
, token
.line
,
2985 token
.offset
+token
.length
+1, token
.lineno
, token
.linepos
);
2987 { /* this branch is optimized away if !DEBUG */
2988 make_tag (concat ("INVALID TOKEN:-->", token_name
.buffer
, ""),
2989 token_name
.len
+ 17, isfun
, token
.line
,
2990 token
.offset
+token
.length
+1, token
.lineno
, token
.linepos
);
2991 error ("INVALID TOKEN");
2994 token
.valid
= false;
3000 * This routine finds functions, variables, typedefs,
3001 * #define's, enum constants and struct/union/enum definitions in
3002 * C syntax and adds them to the list.
3005 C_entries (int c_ext
, FILE *inf
)
3006 /* extension of C */
3009 register char c
; /* latest char read; '\0' for end of line */
3010 register char *lp
; /* pointer one beyond the character `c' */
3011 int curndx
, newndx
; /* indices for current and new lb */
3012 register int tokoff
; /* offset in line of start of current token */
3013 register int toklen
; /* length of current token */
3014 const char *qualifier
; /* string used to qualify names */
3015 int qlen
; /* length of qualifier */
3016 int bracelev
; /* current brace level */
3017 int bracketlev
; /* current bracket level */
3018 int parlev
; /* current parenthesis level */
3019 int attrparlev
; /* __attribute__ parenthesis level */
3020 int templatelev
; /* current template level */
3021 int typdefbracelev
; /* bracelev where a typedef struct body begun */
3022 bool incomm
, inquote
, inchar
, quotednl
, midtoken
;
3023 bool yacc_rules
; /* in the rules part of a yacc file */
3024 struct tok savetoken
= {0}; /* token saved during preprocessor handling */
3027 linebuffer_init (&lbs
[0].lb
);
3028 linebuffer_init (&lbs
[1].lb
);
3029 if (cstack
.size
== 0)
3031 cstack
.size
= (DEBUG
) ? 1 : 4;
3033 cstack
.cname
= xnew (cstack
.size
, char *);
3034 cstack
.bracelev
= xnew (cstack
.size
, int);
3037 tokoff
= toklen
= typdefbracelev
= 0; /* keep compiler quiet */
3038 curndx
= newndx
= 0;
3042 fvdef
= fvnone
; fvextern
= false; typdef
= tnone
;
3043 structdef
= snone
; definedef
= dnone
; objdef
= onone
;
3045 midtoken
= inquote
= inchar
= incomm
= quotednl
= false;
3046 token
.valid
= savetoken
.valid
= false;
3047 bracelev
= bracketlev
= parlev
= attrparlev
= templatelev
= 0;
3049 { qualifier
= "."; qlen
= 1; }
3051 { qualifier
= "::"; qlen
= 2; }
3059 /* If we are at the end of the line, the next character is a
3060 '\0'; do not skip it, because it is what tells us
3061 to read the next line. */
3082 /* Newlines inside comments do not end macro definitions in
3084 CNL_SAVE_DEFINEDEF ();
3097 /* Newlines inside strings do not end macro definitions
3098 in traditional cpp, even though compilers don't
3099 usually accept them. */
3100 CNL_SAVE_DEFINEDEF ();
3110 /* Hmmm, something went wrong. */
3146 if (fvdef
!= finlist
&& fvdef
!= fignore
&& fvdef
!= vignore
)
3161 else if (/* cplpl && */ *lp
== '/')
3167 if ((c_ext
& YACC
) && *lp
== '%')
3169 /* Entering or exiting rules section in yacc file. */
3171 definedef
= dnone
; fvdef
= fvnone
; fvextern
= false;
3172 typdef
= tnone
; structdef
= snone
;
3173 midtoken
= inquote
= inchar
= incomm
= quotednl
= false;
3175 yacc_rules
= !yacc_rules
;
3181 if (definedef
== dnone
)
3184 bool cpptoken
= true;
3186 /* Look back on this line. If all blanks, or nonblanks
3187 followed by an end of comment, this is a preprocessor
3189 for (cp
= newlb
.buffer
; cp
< lp
-1; cp
++)
3190 if (!c_isspace (*cp
))
3192 if (*cp
== '*' && cp
[1] == '/')
3202 definedef
= dsharpseen
;
3203 /* This is needed for tagging enum values: when there are
3204 preprocessor conditionals inside the enum, we need to
3205 reset the value of fvdef so that the next enum value is
3206 tagged even though the one before it did not end in a
3208 if (fvdef
== vignore
&& instruct
&& parlev
== 0)
3210 if (strneq (cp
, "#if", 3) || strneq (cp
, "#el", 3))
3214 } /* if (definedef == dnone) */
3225 CNL_SAVE_DEFINEDEF ();
3232 /* Consider token only if some involved conditions are satisfied. */
3233 if (typdef
!= tignore
3234 && definedef
!= dignorerest
3237 && (definedef
!= dnone
3238 || structdef
!= scolonseen
)
3245 if (c
== ':' && *lp
== ':' && begtoken (lp
[1]))
3246 /* This handles :: in the middle,
3247 but not at the beginning of an identifier.
3248 Also, space-separated :: is not recognized. */
3250 if (c_ext
& C_AUTO
) /* automatic detection of C++ */
3251 c_ext
= (c_ext
| C_PLPL
) & ~C_AUTO
;
3255 goto still_in_token
;
3259 bool funorvar
= false;
3262 || consider_token (newlb
.buffer
+ tokoff
, toklen
, c
,
3263 &c_ext
, bracelev
, parlev
,
3266 if (fvdef
== foperator
)
3269 lp
= skip_spaces (lp
-1);
3273 && !c_isspace (*lp
) && *lp
!= '(')
3276 toklen
+= lp
- oldlp
;
3278 token
.named
= false;
3280 && nestlev
> 0 && definedef
== dnone
)
3281 /* in struct body */
3284 write_classname (&token_name
, qualifier
);
3285 len
= token_name
.len
;
3286 linebuffer_setlen (&token_name
, len
+qlen
+toklen
);
3287 sprintf (token_name
.buffer
+ len
, "%s%.*s",
3288 qualifier
, toklen
, newlb
.buffer
+ tokoff
);
3291 else if (objdef
== ocatseen
)
3292 /* Objective C category */
3294 int len
= strlen (objtag
) + 2 + toklen
;
3295 linebuffer_setlen (&token_name
, len
);
3296 sprintf (token_name
.buffer
, "%s(%.*s)",
3297 objtag
, toklen
, newlb
.buffer
+ tokoff
);
3300 else if (objdef
== omethodtag
3301 || objdef
== omethodparm
)
3302 /* Objective C method */
3306 else if (fvdef
== fdefunname
)
3307 /* GNU DEFUN and similar macros */
3309 bool defun
= (newlb
.buffer
[tokoff
] == 'F');
3313 /* Rewrite the tag so that emacs lisp DEFUNs
3314 can be found by their elisp name */
3320 linebuffer_setlen (&token_name
, len
);
3321 memcpy (token_name
.buffer
,
3322 newlb
.buffer
+ off
, len
);
3323 token_name
.buffer
[len
] = '\0';
3326 if (token_name
.buffer
[len
] == '_')
3327 token_name
.buffer
[len
] = '-';
3328 token
.named
= defun
;
3332 linebuffer_setlen (&token_name
, toklen
);
3333 memcpy (token_name
.buffer
,
3334 newlb
.buffer
+ tokoff
, toklen
);
3335 token_name
.buffer
[toklen
] = '\0';
3336 /* Name macros and members. */
3337 token
.named
= (structdef
== stagseen
3338 || typdef
== ttypeseen
3341 && definedef
== dignorerest
)
3343 && definedef
== dnone
3344 && structdef
== snone
3347 token
.lineno
= lineno
;
3348 token
.offset
= tokoff
;
3349 token
.length
= toklen
;
3350 token
.line
= newlb
.buffer
;
3351 token
.linepos
= newlinepos
;
3354 if (definedef
== dnone
3355 && (fvdef
== fvnameseen
3356 || fvdef
== foperator
3357 || structdef
== stagseen
3359 || typdef
== ttypeseen
3360 || objdef
!= onone
))
3362 if (current_lb_is_new
)
3363 switch_line_buffers ();
3365 else if (definedef
!= dnone
3366 || fvdef
== fdefunname
3368 make_C_tag (funorvar
);
3370 else /* not yacc and consider_token failed */
3372 if (inattribute
&& fvdef
== fignore
)
3374 /* We have just met __attribute__ after a
3375 function parameter list: do not tag the
3382 } /* if (endtoken (c)) */
3383 else if (intoken (c
))
3389 } /* if (midtoken) */
3390 else if (begtoken (c
))
3398 /* This prevents tagging fb in
3399 void (__attribute__((noreturn)) *fb) (void);
3400 Fixing this is not easy and not very important. */
3404 if (plainc
|| declarations
)
3406 make_C_tag (true); /* a function */
3411 if (structdef
== stagseen
&& !cjava
)
3413 popclass_above (bracelev
);
3421 if (!yacc_rules
|| lp
== newlb
.buffer
+ 1)
3423 tokoff
= lp
- 1 - newlb
.buffer
;
3428 } /* if (begtoken) */
3429 } /* if must look at token */
3432 /* Detect end of line, colon, comma, semicolon and various braces
3433 after having handled a token.*/
3439 if (yacc_rules
&& token
.offset
== 0 && token
.valid
)
3441 make_C_tag (false); /* a yacc function */
3444 if (definedef
!= dnone
)
3450 make_C_tag (true); /* an Objective C class */
3454 objdef
= omethodcolon
;
3455 int toklen
= token_name
.len
;
3456 linebuffer_setlen (&token_name
, toklen
+ 1);
3457 strcpy (token_name
.buffer
+ toklen
, ":");
3460 if (structdef
== stagseen
)
3462 structdef
= scolonseen
;
3465 /* Should be useless, but may be work as a safety net. */
3466 if (cplpl
&& fvdef
== flistseen
)
3468 make_C_tag (true); /* a function */
3474 if (definedef
!= dnone
|| inattribute
)
3480 make_C_tag (false); /* a typedef */
3490 if (typdef
== tignore
|| cplpl
)
3494 if ((globals
&& bracelev
== 0 && (!fvextern
|| declarations
))
3495 || (members
&& instruct
))
3496 make_C_tag (false); /* a variable */
3499 token
.valid
= false;
3503 && (cplpl
|| !instruct
)
3504 && (typdef
== tnone
|| (typdef
!= tignore
&& instruct
)))
3506 && plainc
&& instruct
))
3507 make_C_tag (true); /* a function */
3513 && cplpl
&& structdef
== stagseen
)
3514 make_C_tag (false); /* forward declaration */
3516 token
.valid
= false;
3517 } /* switch (fvdef) */
3523 if (structdef
== stagseen
)
3527 if (definedef
!= dnone
|| inattribute
)
3533 make_C_tag (true); /* an Objective C method */
3546 if (instruct
&& parlev
== 0)
3557 && (!fvextern
|| declarations
))
3558 || (members
&& instruct
)))
3559 make_C_tag (false); /* a variable */
3562 if ((declarations
&& typdef
== tnone
&& !instruct
)
3563 || (members
&& typdef
!= tignore
&& instruct
))
3565 make_C_tag (true); /* a function */
3568 else if (!declarations
)
3570 token
.valid
= false;
3575 if (structdef
== stagseen
)
3579 if (definedef
!= dnone
|| inattribute
)
3581 if (structdef
== stagseen
)
3588 make_C_tag (false); /* a typedef */
3600 if ((members
&& bracelev
== 1)
3601 || (globals
&& bracelev
== 0
3602 && (!fvextern
|| declarations
)))
3603 make_C_tag (false); /* a variable */
3617 if (definedef
!= dnone
)
3619 if (objdef
== otagseen
&& parlev
== 0)
3620 objdef
= oparenseen
;
3624 if (typdef
== ttypeseen
3628 /* This handles constructs like:
3629 typedef void OperatorFun (int fun); */
3648 if (--attrparlev
== 0)
3649 inattribute
= false;
3652 if (definedef
!= dnone
)
3654 if (objdef
== ocatseen
&& parlev
== 1)
3656 make_C_tag (true); /* an Objective C category */
3670 || typdef
== ttypeseen
))
3673 make_C_tag (false); /* a typedef */
3676 else if (parlev
< 0) /* can happen due to ill-conceived #if's. */
3680 if (definedef
!= dnone
)
3682 if (typdef
== ttypeseen
)
3684 /* Whenever typdef is set to tinbody (currently only
3685 here), typdefbracelev should be set to bracelev. */
3687 typdefbracelev
= bracelev
;
3692 make_C_tag (true); /* a function */
3701 make_C_tag (true); /* an Objective C class */
3706 make_C_tag (true); /* an Objective C method */
3710 /* Neutralize `extern "C" {' grot. */
3711 if (bracelev
== 0 && structdef
== snone
&& nestlev
== 0
3719 case skeyseen
: /* unnamed struct */
3720 pushclass_above (bracelev
, NULL
, 0);
3723 case stagseen
: /* named struct or enum */
3724 case scolonseen
: /* a class */
3725 pushclass_above (bracelev
,token
.line
+token
.offset
, token
.length
);
3727 make_C_tag (false); /* a struct or enum */
3733 if (definedef
!= dnone
)
3735 if (fvdef
== fstartlist
)
3737 fvdef
= fvnone
; /* avoid tagging `foo' in `foo (*bar()) ()' */
3738 token
.valid
= false;
3742 if (definedef
!= dnone
)
3745 if (!ignoreindent
&& lp
== newlb
.buffer
+ 1)
3748 token
.valid
= false; /* unexpected value, token unreliable */
3749 bracelev
= 0; /* reset brace level if first column */
3750 parlev
= 0; /* also reset paren level, just in case... */
3752 else if (bracelev
< 0)
3754 token
.valid
= false; /* something gone amiss, token unreliable */
3757 if (bracelev
== 0 && fvdef
== vignore
)
3758 fvdef
= fvnone
; /* end of function */
3759 popclass_above (bracelev
);
3761 /* Only if typdef == tinbody is typdefbracelev significant. */
3762 if (typdef
== tinbody
&& bracelev
<= typdefbracelev
)
3764 assert (bracelev
== typdefbracelev
);
3769 if (definedef
!= dnone
)
3779 if ((members
&& bracelev
== 1)
3780 || (globals
&& bracelev
== 0 && (!fvextern
|| declarations
)))
3781 make_C_tag (false); /* a variable */
3789 && (structdef
== stagseen
|| fvdef
== fvnameseen
))
3796 if (templatelev
> 0)
3804 if (objdef
== oinbody
&& bracelev
== 0)
3806 objdef
= omethodsign
;
3811 case '#': case '~': case '&': case '%': case '/':
3812 case '|': case '^': case '!': case '.': case '?':
3813 if (definedef
!= dnone
)
3815 /* These surely cannot follow a function tag in C. */
3828 if (objdef
== otagseen
)
3830 make_C_tag (true); /* an Objective C class */
3833 /* If a macro spans multiple lines don't reset its state. */
3835 CNL_SAVE_DEFINEDEF ();
3841 } /* while not eof */
3843 free (lbs
[0].lb
.buffer
);
3844 free (lbs
[1].lb
.buffer
);
3848 * Process either a C++ file or a C file depending on the setting
3852 default_C_entries (FILE *inf
)
3854 C_entries (cplusplus
? C_PLPL
: C_AUTO
, inf
);
3857 /* Always do plain C. */
3859 plain_C_entries (FILE *inf
)
3864 /* Always do C++. */
3866 Cplusplus_entries (FILE *inf
)
3868 C_entries (C_PLPL
, inf
);
3871 /* Always do Java. */
3873 Cjava_entries (FILE *inf
)
3875 C_entries (C_JAVA
, inf
);
3880 Cstar_entries (FILE *inf
)
3882 C_entries (C_STAR
, inf
);
3885 /* Always do Yacc. */
3887 Yacc_entries (FILE *inf
)
3889 C_entries (YACC
, inf
);
3893 /* Useful macros. */
3894 #define LOOP_ON_INPUT_LINES(file_pointer, line_buffer, char_pointer) \
3895 for (; /* loop initialization */ \
3896 !feof (file_pointer) /* loop test */ \
3897 && /* instructions at start of loop */ \
3898 (readline (&line_buffer, file_pointer), \
3899 char_pointer = line_buffer.buffer, \
3903 #define LOOKING_AT(cp, kw) /* kw is the keyword, a literal string */ \
3904 ((assert ("" kw), true) /* syntax error if not a literal string */ \
3905 && strneq ((cp), kw, sizeof (kw)-1) /* cp points at kw */ \
3906 && notinname ((cp)[sizeof (kw)-1]) /* end of kw */ \
3907 && ((cp) = skip_spaces ((cp)+sizeof (kw)-1))) /* skip spaces */
3909 /* Similar to LOOKING_AT but does not use notinname, does not skip */
3910 #define LOOKING_AT_NOCASE(cp, kw) /* the keyword is a literal string */ \
3911 ((assert ("" kw), true) /* syntax error if not a literal string */ \
3912 && strncaseeq ((cp), kw, sizeof (kw)-1) /* cp points at kw */ \
3913 && ((cp) += sizeof (kw)-1)) /* skip spaces */
3916 * Read a file, but do no processing. This is used to do regexp
3917 * matching on files that have no language defined.
3920 just_read_file (FILE *inf
)
3923 readline (&lb
, inf
);
3927 /* Fortran parsing */
3929 static void F_takeprec (void);
3930 static void F_getit (FILE *);
3935 dbp
= skip_spaces (dbp
);
3939 dbp
= skip_spaces (dbp
);
3940 if (strneq (dbp
, "(*)", 3))
3945 if (!c_isdigit (*dbp
))
3947 --dbp
; /* force failure */
3952 while (c_isdigit (*dbp
));
3960 dbp
= skip_spaces (dbp
);
3963 readline (&lb
, inf
);
3968 dbp
= skip_spaces (dbp
);
3970 if (!c_isalpha (*dbp
) && *dbp
!= '_' && *dbp
!= '$')
3972 for (cp
= dbp
+ 1; *cp
!= '\0' && intoken (*cp
); cp
++)
3974 make_tag (dbp
, cp
-dbp
, true,
3975 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
3980 Fortran_functions (FILE *inf
)
3982 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
3985 dbp
++; /* Ratfor escape to fortran */
3986 dbp
= skip_spaces (dbp
);
3990 if (LOOKING_AT_NOCASE (dbp
, "recursive"))
3991 dbp
= skip_spaces (dbp
);
3993 if (LOOKING_AT_NOCASE (dbp
, "pure"))
3994 dbp
= skip_spaces (dbp
);
3996 if (LOOKING_AT_NOCASE (dbp
, "elemental"))
3997 dbp
= skip_spaces (dbp
);
3999 switch (c_tolower (*dbp
))
4002 if (nocase_tail ("integer"))
4006 if (nocase_tail ("real"))
4010 if (nocase_tail ("logical"))
4014 if (nocase_tail ("complex") || nocase_tail ("character"))
4018 if (nocase_tail ("double"))
4020 dbp
= skip_spaces (dbp
);
4023 if (nocase_tail ("precision"))
4029 dbp
= skip_spaces (dbp
);
4032 switch (c_tolower (*dbp
))
4035 if (nocase_tail ("function"))
4039 if (nocase_tail ("subroutine"))
4043 if (nocase_tail ("entry"))
4047 if (nocase_tail ("blockdata") || nocase_tail ("block data"))
4049 dbp
= skip_spaces (dbp
);
4050 if (*dbp
== '\0') /* assume un-named */
4051 make_tag ("blockdata", 9, true,
4052 lb
.buffer
, dbp
- lb
.buffer
, lineno
, linecharno
);
4054 F_getit (inf
); /* look for name */
4065 * Philippe Waroquiers (1998)
4068 /* Once we are positioned after an "interesting" keyword, let's get
4069 the real tag value necessary. */
4071 Ada_getit (FILE *inf
, const char *name_qualifier
)
4079 dbp
= skip_spaces (dbp
);
4081 || (dbp
[0] == '-' && dbp
[1] == '-'))
4083 readline (&lb
, inf
);
4086 switch (c_tolower (*dbp
))
4089 if (nocase_tail ("body"))
4091 /* Skipping body of procedure body or package body or ....
4092 resetting qualifier to body instead of spec. */
4093 name_qualifier
= "/b";
4098 /* Skipping type of task type or protected type ... */
4099 if (nocase_tail ("type"))
4106 for (cp
= dbp
; *cp
!= '\0' && *cp
!= '"'; cp
++)
4111 dbp
= skip_spaces (dbp
);
4113 c_isalnum (*cp
) || *cp
== '_' || *cp
== '.';
4121 name
= concat (dbp
, name_qualifier
, "");
4123 make_tag (name
, strlen (name
), true,
4124 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4133 Ada_funcs (FILE *inf
)
4135 bool inquote
= false;
4136 bool skip_till_semicolumn
= false;
4138 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
4140 while (*dbp
!= '\0')
4142 /* Skip a string i.e. "abcd". */
4143 if (inquote
|| (*dbp
== '"'))
4145 dbp
= strchr (dbp
+ !inquote
, '"');
4150 continue; /* advance char */
4155 break; /* advance line */
4159 /* Skip comments. */
4160 if (dbp
[0] == '-' && dbp
[1] == '-')
4161 break; /* advance line */
4163 /* Skip character enclosed in single quote i.e. 'a'
4164 and skip single quote starting an attribute i.e. 'Image. */
4173 if (skip_till_semicolumn
)
4176 skip_till_semicolumn
= false;
4178 continue; /* advance char */
4181 /* Search for beginning of a token. */
4182 if (!begtoken (*dbp
))
4185 continue; /* advance char */
4188 /* We are at the beginning of a token. */
4189 switch (c_tolower (*dbp
))
4192 if (!packages_only
&& nocase_tail ("function"))
4193 Ada_getit (inf
, "/f");
4195 break; /* from switch */
4196 continue; /* advance char */
4198 if (!packages_only
&& nocase_tail ("procedure"))
4199 Ada_getit (inf
, "/p");
4200 else if (nocase_tail ("package"))
4201 Ada_getit (inf
, "/s");
4202 else if (nocase_tail ("protected")) /* protected type */
4203 Ada_getit (inf
, "/t");
4205 break; /* from switch */
4206 continue; /* advance char */
4209 if (typedefs
&& !packages_only
&& nocase_tail ("use"))
4211 /* when tagging types, avoid tagging use type Pack.Typename;
4212 for this, we will skip everything till a ; */
4213 skip_till_semicolumn
= true;
4214 continue; /* advance char */
4218 if (!packages_only
&& nocase_tail ("task"))
4219 Ada_getit (inf
, "/k");
4220 else if (typedefs
&& !packages_only
&& nocase_tail ("type"))
4222 Ada_getit (inf
, "/t");
4223 while (*dbp
!= '\0')
4227 break; /* from switch */
4228 continue; /* advance char */
4231 /* Look for the end of the token. */
4232 while (!endtoken (*dbp
))
4235 } /* advance char */
4236 } /* advance line */
4241 * Unix and microcontroller assembly tag handling
4242 * Labels: /^[a-zA-Z_.$][a-zA_Z0-9_.$]*[: ^I^J]/
4243 * Idea by Bob Weiner, Motorola Inc. (1994)
4246 Asm_labels (FILE *inf
)
4250 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4252 /* If first char is alphabetic or one of [_.$], test for colon
4253 following identifier. */
4254 if (c_isalpha (*cp
) || *cp
== '_' || *cp
== '.' || *cp
== '$')
4256 /* Read past label. */
4258 while (c_isalnum (*cp
) || *cp
== '_' || *cp
== '.' || *cp
== '$')
4260 if (*cp
== ':' || c_isspace (*cp
))
4261 /* Found end of label, so copy it and add it to the table. */
4262 make_tag (lb
.buffer
, cp
- lb
.buffer
, true,
4263 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4271 * Perl sub names: /^sub[ \t\n]+[^ \t\n{]+/
4272 * /^use constant[ \t\n]+[^ \t\n{=,;]+/
4273 * Perl variable names: /^(my|local).../
4274 * Original code by Bart Robinson <lomew@cs.utah.edu> (1995)
4275 * Additions by Michael Ernst <mernst@alum.mit.edu> (1997)
4276 * Ideas by Kai Großjohann <Kai.Grossjohann@CS.Uni-Dortmund.DE> (2001)
4279 Perl_functions (FILE *inf
)
4281 char *package
= savestr ("main"); /* current package name */
4284 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4286 cp
= skip_spaces (cp
);
4288 if (LOOKING_AT (cp
, "package"))
4291 get_tag (cp
, &package
);
4293 else if (LOOKING_AT (cp
, "sub"))
4299 while (!notinname (*cp
))
4302 continue; /* nothing found */
4303 pos
= strchr (sp
, ':');
4304 if (pos
&& pos
< cp
&& pos
[1] == ':')
4305 /* The name is already qualified. */
4306 make_tag (sp
, cp
- sp
, true,
4307 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4311 char savechar
, *name
;
4315 name
= concat (package
, "::", sp
);
4317 make_tag (name
, strlen (name
), true,
4318 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4322 else if (LOOKING_AT (cp
, "use constant")
4323 || LOOKING_AT (cp
, "use constant::defer"))
4325 /* For hash style multi-constant like
4326 use constant { FOO => 123,
4328 only the first FOO is picked up. Parsing across the value
4329 expressions would be difficult in general, due to possible nested
4330 hashes, here-documents, etc. */
4332 cp
= skip_spaces (cp
+1);
4335 else if (globals
) /* only if we are tagging global vars */
4337 /* Skip a qualifier, if any. */
4338 bool qual
= LOOKING_AT (cp
, "my") || LOOKING_AT (cp
, "local");
4339 /* After "my" or "local", but before any following paren or space. */
4340 char *varstart
= cp
;
4342 if (qual
/* should this be removed? If yes, how? */
4343 && (*cp
== '$' || *cp
== '@' || *cp
== '%'))
4348 while (c_isalnum (*cp
) || *cp
== '_');
4352 /* Should be examining a variable list at this point;
4353 could insist on seeing an open parenthesis. */
4354 while (*cp
!= '\0' && *cp
!= ';' && *cp
!= '=' && *cp
!= ')')
4360 make_tag (varstart
, cp
- varstart
, false,
4361 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4370 * Look for /^[\t]*def[ \t\n]+[^ \t\n(:]+/ or /^class[ \t\n]+[^ \t\n(:]+/
4371 * Idea by Eric S. Raymond <esr@thyrsus.com> (1997)
4372 * More ideas by seb bacon <seb@jamkit.com> (2002)
4375 Python_functions (FILE *inf
)
4379 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4381 cp
= skip_spaces (cp
);
4382 if (LOOKING_AT (cp
, "def") || LOOKING_AT (cp
, "class"))
4385 while (!notinname (*cp
) && *cp
!= ':')
4387 make_tag (name
, cp
- name
, true,
4388 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4397 * - /^[ \t]*function[ \t\n]+[^ \t\n(]+/
4398 * - /^[ \t]*class[ \t\n]+[^ \t\n]+/
4399 * - /^[ \t]*define\(\"[^\"]+/
4400 * Only with --members:
4401 * - /^[ \t]*var[ \t\n]+\$[^ \t\n=;]/
4402 * Idea by Diez B. Roggisch (2001)
4405 PHP_functions (FILE *inf
)
4408 bool search_identifier
= false;
4410 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4412 cp
= skip_spaces (cp
);
4414 if (search_identifier
4417 while (!notinname (*cp
))
4419 make_tag (name
, cp
- name
, true,
4420 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4421 search_identifier
= false;
4423 else if (LOOKING_AT (cp
, "function"))
4426 cp
= skip_spaces (cp
+1);
4430 while (!notinname (*cp
))
4432 make_tag (name
, cp
- name
, true,
4433 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4436 search_identifier
= true;
4438 else if (LOOKING_AT (cp
, "class"))
4443 while (*cp
!= '\0' && !c_isspace (*cp
))
4445 make_tag (name
, cp
- name
, false,
4446 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4449 search_identifier
= true;
4451 else if (strneq (cp
, "define", 6)
4452 && (cp
= skip_spaces (cp
+6))
4454 && (*cp
== '"' || *cp
== '\''))
4458 while (*cp
!= quote
&& *cp
!= '\0')
4460 make_tag (name
, cp
- name
, false,
4461 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4464 && LOOKING_AT (cp
, "var")
4468 while (!notinname (*cp
))
4470 make_tag (name
, cp
- name
, false,
4471 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4478 * Cobol tag functions
4479 * We could look for anything that could be a paragraph name.
4480 * i.e. anything that starts in column 8 is one word and ends in a full stop.
4481 * Idea by Corny de Souza (1993)
4484 Cobol_paragraphs (FILE *inf
)
4486 register char *bp
, *ep
;
4488 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4494 /* If eoln, compiler option or comment ignore whole line. */
4495 if (bp
[-1] != ' ' || !c_isalnum (bp
[0]))
4498 for (ep
= bp
; c_isalnum (*ep
) || *ep
== '-'; ep
++)
4501 make_tag (bp
, ep
- bp
, true,
4502 lb
.buffer
, ep
- lb
.buffer
+ 1, lineno
, linecharno
);
4509 * Ideas by Assar Westerlund <assar@sics.se> (2001)
4512 Makefile_targets (FILE *inf
)
4516 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4518 if (*bp
== '\t' || *bp
== '#')
4520 while (*bp
!= '\0' && *bp
!= '=' && *bp
!= ':')
4522 if (*bp
== ':' || (globals
&& *bp
== '='))
4524 /* We should detect if there is more than one tag, but we do not.
4525 We just skip initial and final spaces. */
4526 char * namestart
= skip_spaces (lb
.buffer
);
4527 while (--bp
> namestart
)
4528 if (!notinname (*bp
))
4530 make_tag (namestart
, bp
- namestart
+ 1, true,
4531 lb
.buffer
, bp
- lb
.buffer
+ 2, lineno
, linecharno
);
4539 * Original code by Mosur K. Mohan (1989)
4541 * Locates tags for procedures & functions. Doesn't do any type- or
4542 * var-definitions. It does look for the keyword "extern" or
4543 * "forward" immediately following the procedure statement; if found,
4544 * the tag is skipped.
4547 Pascal_functions (FILE *inf
)
4549 linebuffer tline
; /* mostly copied from C_entries */
4551 int save_lineno
, namelen
, taglen
;
4554 bool /* each of these flags is true if: */
4555 incomment
, /* point is inside a comment */
4556 inquote
, /* point is inside '..' string */
4557 get_tagname
, /* point is after PROCEDURE/FUNCTION
4558 keyword, so next item = potential tag */
4559 found_tag
, /* point is after a potential tag */
4560 inparms
, /* point is within parameter-list */
4561 verify_tag
; /* point has passed the parm-list, so the
4562 next token will determine whether this
4563 is a FORWARD/EXTERN to be ignored, or
4564 whether it is a real tag */
4566 save_lcno
= save_lineno
= namelen
= taglen
= 0; /* keep compiler quiet */
4567 name
= NULL
; /* keep compiler quiet */
4570 linebuffer_init (&tline
);
4572 incomment
= inquote
= false;
4573 found_tag
= false; /* have a proc name; check if extern */
4574 get_tagname
= false; /* found "procedure" keyword */
4575 inparms
= false; /* found '(' after "proc" */
4576 verify_tag
= false; /* check if "extern" is ahead */
4579 while (!feof (inf
)) /* long main loop to get next char */
4582 if (c
== '\0') /* if end of line */
4584 readline (&lb
, inf
);
4588 if (!((found_tag
&& verify_tag
)
4590 c
= *dbp
++; /* only if don't need *dbp pointing
4591 to the beginning of the name of
4592 the procedure or function */
4596 if (c
== '}') /* within { } comments */
4598 else if (c
== '*' && *dbp
== ')') /* within (* *) comments */
4615 inquote
= true; /* found first quote */
4617 case '{': /* found open { comment */
4621 if (*dbp
== '*') /* found open (* comment */
4626 else if (found_tag
) /* found '(' after tag, i.e., parm-list */
4629 case ')': /* end of parms list */
4634 if (found_tag
&& !inparms
) /* end of proc or fn stmt */
4641 if (found_tag
&& verify_tag
&& (*dbp
!= ' '))
4643 /* Check if this is an "extern" declaration. */
4646 if (c_tolower (*dbp
) == 'e')
4648 if (nocase_tail ("extern")) /* superfluous, really! */
4654 else if (c_tolower (*dbp
) == 'f')
4656 if (nocase_tail ("forward")) /* check for forward reference */
4662 if (found_tag
&& verify_tag
) /* not external proc, so make tag */
4666 make_tag (name
, namelen
, true,
4667 tline
.buffer
, taglen
, save_lineno
, save_lcno
);
4671 if (get_tagname
) /* grab name of proc or fn */
4678 /* Find block name. */
4679 for (cp
= dbp
+ 1; *cp
!= '\0' && !endtoken (*cp
); cp
++)
4682 /* Save all values for later tagging. */
4683 linebuffer_setlen (&tline
, lb
.len
);
4684 strcpy (tline
.buffer
, lb
.buffer
);
4685 save_lineno
= lineno
;
4686 save_lcno
= linecharno
;
4687 name
= tline
.buffer
+ (dbp
- lb
.buffer
);
4689 taglen
= cp
- lb
.buffer
+ 1;
4691 dbp
= cp
; /* set dbp to e-o-token */
4692 get_tagname
= false;
4696 /* And proceed to check for "extern". */
4698 else if (!incomment
&& !inquote
&& !found_tag
)
4700 /* Check for proc/fn keywords. */
4701 switch (c_tolower (c
))
4704 if (nocase_tail ("rocedure")) /* c = 'p', dbp has advanced */
4708 if (nocase_tail ("unction"))
4713 } /* while not eof */
4715 free (tline
.buffer
);
4720 * Lisp tag functions
4721 * look for (def or (DEF, quote or QUOTE
4724 static void L_getit (void);
4729 if (*dbp
== '\'') /* Skip prefix quote */
4731 else if (*dbp
== '(')
4734 /* Try to skip "(quote " */
4735 if (!LOOKING_AT (dbp
, "quote") && !LOOKING_AT (dbp
, "QUOTE"))
4736 /* Ok, then skip "(" before name in (defstruct (foo)) */
4737 dbp
= skip_spaces (dbp
);
4739 get_tag (dbp
, NULL
);
4743 Lisp_functions (FILE *inf
)
4745 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
4750 /* "(defvar foo)" is a declaration rather than a definition. */
4754 if (LOOKING_AT (p
, "defvar"))
4756 p
= skip_name (p
); /* past var name */
4757 p
= skip_spaces (p
);
4763 if (strneq (dbp
+ 1, "cl-", 3) || strneq (dbp
+ 1, "CL-", 3))
4766 if (strneq (dbp
+1, "def", 3) || strneq (dbp
+1, "DEF", 3))
4768 dbp
= skip_non_spaces (dbp
);
4769 dbp
= skip_spaces (dbp
);
4774 /* Check for (foo::defmumble name-defined ... */
4777 while (!notinname (*dbp
) && *dbp
!= ':');
4782 while (*dbp
== ':');
4784 if (strneq (dbp
, "def", 3) || strneq (dbp
, "DEF", 3))
4786 dbp
= skip_non_spaces (dbp
);
4787 dbp
= skip_spaces (dbp
);
4797 * Lua script language parsing
4798 * Original code by David A. Capello <dacap@users.sourceforge.net> (2004)
4800 * "function" and "local function" are tags if they start at column 1.
4803 Lua_functions (FILE *inf
)
4807 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4809 if (bp
[0] != 'f' && bp
[0] != 'l')
4812 (void)LOOKING_AT (bp
, "local"); /* skip possible "local" */
4814 if (LOOKING_AT (bp
, "function"))
4822 * Just look for lines where the first character is '/'
4823 * Also look at "defineps" for PSWrap
4825 * Richard Mlynarik <mly@adoc.xerox.com> (1997)
4826 * Masatake Yamato <masata-y@is.aist-nara.ac.jp> (1999)
4829 PS_functions (FILE *inf
)
4831 register char *bp
, *ep
;
4833 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4838 *ep
!= '\0' && *ep
!= ' ' && *ep
!= '{';
4841 make_tag (bp
, ep
- bp
, true,
4842 lb
.buffer
, ep
- lb
.buffer
+ 1, lineno
, linecharno
);
4844 else if (LOOKING_AT (bp
, "defineps"))
4852 * Ignore anything after \ followed by space or in ( )
4853 * Look for words defined by :
4854 * Look for constant, code, create, defer, value, and variable
4855 * OBP extensions: Look for buffer:, field,
4856 * Ideas by Eduardo Horvath <eeh@netbsd.org> (2004)
4859 Forth_words (FILE *inf
)
4863 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4864 while ((bp
= skip_spaces (bp
))[0] != '\0')
4865 if (bp
[0] == '\\' && c_isspace (bp
[1]))
4866 break; /* read next line */
4867 else if (bp
[0] == '(' && c_isspace (bp
[1]))
4868 do /* skip to ) or eol */
4870 while (*bp
!= ')' && *bp
!= '\0');
4871 else if ((bp
[0] == ':' && c_isspace (bp
[1]) && bp
++)
4872 || LOOKING_AT_NOCASE (bp
, "constant")
4873 || LOOKING_AT_NOCASE (bp
, "code")
4874 || LOOKING_AT_NOCASE (bp
, "create")
4875 || LOOKING_AT_NOCASE (bp
, "defer")
4876 || LOOKING_AT_NOCASE (bp
, "value")
4877 || LOOKING_AT_NOCASE (bp
, "variable")
4878 || LOOKING_AT_NOCASE (bp
, "buffer:")
4879 || LOOKING_AT_NOCASE (bp
, "field"))
4880 get_tag (skip_spaces (bp
), NULL
); /* Yay! A definition! */
4882 bp
= skip_non_spaces (bp
);
4887 * Scheme tag functions
4888 * look for (def... xyzzy
4890 * (def ... ((...(xyzzy ....
4892 * Original code by Ken Haase (1985?)
4895 Scheme_functions (FILE *inf
)
4899 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4901 if (strneq (bp
, "(def", 4) || strneq (bp
, "(DEF", 4))
4903 bp
= skip_non_spaces (bp
+4);
4904 /* Skip over open parens and white space. Don't continue past
4906 while (*bp
&& notinname (*bp
))
4910 if (LOOKING_AT (bp
, "(SET!") || LOOKING_AT (bp
, "(set!"))
4916 /* Find tags in TeX and LaTeX input files. */
4918 /* TEX_toktab is a table of TeX control sequences that define tags.
4919 * Each entry records one such control sequence.
4921 * Original code from who knows whom.
4923 * Stefan Monnier (2002)
4926 static linebuffer
*TEX_toktab
= NULL
; /* Table with tag tokens */
4928 /* Default set of control sequences to put into TEX_toktab.
4929 The value of environment var TEXTAGS is prepended to this. */
4930 static const char *TEX_defenv
= "\
4931 :chapter:section:subsection:subsubsection:eqno:label:ref:cite:bibitem\
4932 :part:appendix:entry:index:def\
4933 :newcommand:renewcommand:newenvironment:renewenvironment";
4935 static void TEX_mode (FILE *);
4936 static void TEX_decode_env (const char *, const char *);
4938 static char TEX_esc
= '\\';
4939 static char TEX_opgrp
= '{';
4940 static char TEX_clgrp
= '}';
4943 * TeX/LaTeX scanning loop.
4946 TeX_commands (FILE *inf
)
4951 /* Select either \ or ! as escape character. */
4954 /* Initialize token table once from environment. */
4955 if (TEX_toktab
== NULL
)
4956 TEX_decode_env ("TEXTAGS", TEX_defenv
);
4958 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4960 /* Look at each TEX keyword in line. */
4963 /* Look for a TEX escape. */
4964 while (*cp
++ != TEX_esc
)
4965 if (cp
[-1] == '\0' || cp
[-1] == '%')
4968 for (key
= TEX_toktab
; key
->buffer
!= NULL
; key
++)
4969 if (strneq (cp
, key
->buffer
, key
->len
))
4972 int namelen
, linelen
;
4975 cp
= skip_spaces (cp
+ key
->len
);
4976 if (*cp
== TEX_opgrp
)
4982 (!c_isspace (*p
) && *p
!= '#' &&
4983 *p
!= TEX_opgrp
&& *p
!= TEX_clgrp
);
4988 if (!opgrp
|| *p
== TEX_clgrp
)
4990 while (*p
!= '\0' && *p
!= TEX_opgrp
&& *p
!= TEX_clgrp
)
4992 linelen
= p
- lb
.buffer
+ 1;
4994 make_tag (cp
, namelen
, true,
4995 lb
.buffer
, linelen
, lineno
, linecharno
);
4996 goto tex_next_line
; /* We only tag a line once */
5004 #define TEX_LESC '\\'
5005 #define TEX_SESC '!'
5007 /* Figure out whether TeX's escapechar is '\\' or '!' and set grouping
5008 chars accordingly. */
5010 TEX_mode (FILE *inf
)
5014 while ((c
= getc (inf
)) != EOF
)
5016 /* Skip to next line if we hit the TeX comment char. */
5018 while (c
!= '\n' && c
!= EOF
)
5020 else if (c
== TEX_LESC
|| c
== TEX_SESC
)
5039 /* Read environment and prepend it to the default string.
5040 Build token table. */
5042 TEX_decode_env (const char *evarname
, const char *defenv
)
5044 register const char *env
, *p
;
5047 /* Append default string to environment. */
5048 env
= getenv (evarname
);
5052 env
= concat (env
, defenv
, "");
5054 /* Allocate a token table */
5055 for (len
= 1, p
= env
; (p
= strchr (p
, ':')); )
5058 TEX_toktab
= xnew (len
, linebuffer
);
5060 /* Unpack environment string into token table. Be careful about */
5061 /* zero-length strings (leading ':', "::" and trailing ':') */
5062 for (i
= 0; *env
!= '\0';)
5064 p
= strchr (env
, ':');
5065 if (!p
) /* End of environment string. */
5066 p
= env
+ strlen (env
);
5068 { /* Only non-zero strings. */
5069 TEX_toktab
[i
].buffer
= savenstr (env
, p
- env
);
5070 TEX_toktab
[i
].len
= p
- env
;
5077 TEX_toktab
[i
].buffer
= NULL
; /* Mark end of table. */
5078 TEX_toktab
[i
].len
= 0;
5085 /* Texinfo support. Dave Love, Mar. 2000. */
5087 Texinfo_nodes (FILE *inf
)
5090 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5091 if (LOOKING_AT (cp
, "@node"))
5094 while (*cp
!= '\0' && *cp
!= ',')
5096 make_tag (start
, cp
- start
, true,
5097 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
5104 * Contents of <title>, <h1>, <h2>, <h3> are tags.
5105 * Contents of <a name=xxx> are tags with name xxx.
5107 * Francesco Potortì, 2002.
5110 HTML_labels (FILE *inf
)
5112 bool getnext
= false; /* next text outside of HTML tags is a tag */
5113 bool skiptag
= false; /* skip to the end of the current HTML tag */
5114 bool intag
= false; /* inside an html tag, looking for ID= */
5115 bool inanchor
= false; /* when INTAG, is an anchor, look for NAME= */
5119 linebuffer_setlen (&token_name
, 0); /* no name in buffer */
5121 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
5122 for (;;) /* loop on the same line */
5124 if (skiptag
) /* skip HTML tag */
5126 while (*dbp
!= '\0' && *dbp
!= '>')
5132 continue; /* look on the same line */
5134 break; /* go to next line */
5137 else if (intag
) /* look for "name=" or "id=" */
5139 while (*dbp
!= '\0' && *dbp
!= '>'
5140 && c_tolower (*dbp
) != 'n' && c_tolower (*dbp
) != 'i')
5143 break; /* go to next line */
5148 continue; /* look on the same line */
5150 if ((inanchor
&& LOOKING_AT_NOCASE (dbp
, "name="))
5151 || LOOKING_AT_NOCASE (dbp
, "id="))
5153 bool quoted
= (dbp
[0] == '"');
5156 for (end
= ++dbp
; *end
!= '\0' && *end
!= '"'; end
++)
5159 for (end
= dbp
; *end
!= '\0' && intoken (*end
); end
++)
5161 linebuffer_setlen (&token_name
, end
- dbp
);
5162 memcpy (token_name
.buffer
, dbp
, end
- dbp
);
5163 token_name
.buffer
[end
- dbp
] = '\0';
5166 intag
= false; /* we found what we looked for */
5167 skiptag
= true; /* skip to the end of the tag */
5168 getnext
= true; /* then grab the text */
5169 continue; /* look on the same line */
5174 else if (getnext
) /* grab next tokens and tag them */
5176 dbp
= skip_spaces (dbp
);
5178 break; /* go to next line */
5182 inanchor
= (c_tolower (dbp
[1]) == 'a' && !intoken (dbp
[2]));
5183 continue; /* look on the same line */
5186 for (end
= dbp
+ 1; *end
!= '\0' && *end
!= '<'; end
++)
5188 make_tag (token_name
.buffer
, token_name
.len
, true,
5189 dbp
, end
- dbp
, lineno
, linecharno
);
5190 linebuffer_setlen (&token_name
, 0); /* no name in buffer */
5192 break; /* go to next line */
5195 else /* look for an interesting HTML tag */
5197 while (*dbp
!= '\0' && *dbp
!= '<')
5200 break; /* go to next line */
5202 if (c_tolower (dbp
[1]) == 'a' && !intoken (dbp
[2]))
5205 continue; /* look on the same line */
5207 else if (LOOKING_AT_NOCASE (dbp
, "<title>")
5208 || LOOKING_AT_NOCASE (dbp
, "<h1>")
5209 || LOOKING_AT_NOCASE (dbp
, "<h2>")
5210 || LOOKING_AT_NOCASE (dbp
, "<h3>"))
5214 continue; /* look on the same line */
5225 * Assumes that the predicate or rule starts at column 0.
5226 * Only the first clause of a predicate or rule is added.
5227 * Original code by Sunichirou Sugou (1989)
5228 * Rewritten by Anders Lindgren (1996)
5230 static size_t prolog_pr (char *, char *);
5231 static void prolog_skip_comment (linebuffer
*, FILE *);
5232 static size_t prolog_atom (char *, size_t);
5235 Prolog_functions (FILE *inf
)
5245 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5247 if (cp
[0] == '\0') /* Empty line */
5249 else if (c_isspace (cp
[0])) /* Not a predicate */
5251 else if (cp
[0] == '/' && cp
[1] == '*') /* comment. */
5252 prolog_skip_comment (&lb
, inf
);
5253 else if ((len
= prolog_pr (cp
, last
)) > 0)
5255 /* Predicate or rule. Store the function name so that we
5256 only generate a tag for the first clause. */
5258 last
= xnew (len
+ 1, char);
5259 else if (len
+ 1 > allocated
)
5260 xrnew (last
, len
+ 1, char);
5261 allocated
= len
+ 1;
5262 memcpy (last
, cp
, len
);
5271 prolog_skip_comment (linebuffer
*plb
, FILE *inf
)
5277 for (cp
= plb
->buffer
; *cp
!= '\0'; cp
++)
5278 if (cp
[0] == '*' && cp
[1] == '/')
5280 readline (plb
, inf
);
5282 while (!feof (inf
));
5286 * A predicate or rule definition is added if it matches:
5287 * <beginning of line><Prolog Atom><whitespace>(
5288 * or <beginning of line><Prolog Atom><whitespace>:-
5290 * It is added to the tags database if it doesn't match the
5291 * name of the previous clause header.
5293 * Return the size of the name of the predicate or rule, or 0 if no
5297 prolog_pr (char *s
, char *last
)
5299 /* Name of last clause. */
5304 pos
= prolog_atom (s
, 0);
5309 pos
= skip_spaces (s
+ pos
) - s
;
5312 || (s
[pos
] == '(' && (pos
+= 1))
5313 || (s
[pos
] == ':' && s
[pos
+ 1] == '-' && (pos
+= 2)))
5314 && (last
== NULL
/* save only the first clause */
5315 || len
!= strlen (last
)
5316 || !strneq (s
, last
, len
)))
5318 make_tag (s
, len
, true, s
, pos
, lineno
, linecharno
);
5326 * Consume a Prolog atom.
5327 * Return the number of bytes consumed, or 0 if there was an error.
5329 * A prolog atom, in this context, could be one of:
5330 * - An alphanumeric sequence, starting with a lower case letter.
5331 * - A quoted arbitrary string. Single quotes can escape themselves.
5332 * Backslash quotes everything.
5335 prolog_atom (char *s
, size_t pos
)
5341 if (c_islower (s
[pos
]) || s
[pos
] == '_')
5343 /* The atom is unquoted. */
5345 while (c_isalnum (s
[pos
]) || s
[pos
] == '_')
5349 return pos
- origpos
;
5351 else if (s
[pos
] == '\'')
5362 pos
++; /* A double quote */
5364 else if (s
[pos
] == '\0')
5365 /* Multiline quoted atoms are ignored. */
5367 else if (s
[pos
] == '\\')
5369 if (s
[pos
+1] == '\0')
5376 return pos
- origpos
;
5384 * Support for Erlang
5386 * Generates tags for functions, defines, and records.
5387 * Assumes that Erlang functions start at column 0.
5388 * Original code by Anders Lindgren (1996)
5390 static int erlang_func (char *, char *);
5391 static void erlang_attribute (char *);
5392 static int erlang_atom (char *);
5395 Erlang_functions (FILE *inf
)
5405 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5407 if (cp
[0] == '\0') /* Empty line */
5409 else if (c_isspace (cp
[0])) /* Not function nor attribute */
5411 else if (cp
[0] == '%') /* comment */
5413 else if (cp
[0] == '"') /* Sometimes, strings start in column one */
5415 else if (cp
[0] == '-') /* attribute, e.g. "-define" */
5417 erlang_attribute (cp
);
5424 else if ((len
= erlang_func (cp
, last
)) > 0)
5427 * Function. Store the function name so that we only
5428 * generates a tag for the first clause.
5431 last
= xnew (len
+ 1, char);
5432 else if (len
+ 1 > allocated
)
5433 xrnew (last
, len
+ 1, char);
5434 allocated
= len
+ 1;
5435 memcpy (last
, cp
, len
);
5444 * A function definition is added if it matches:
5445 * <beginning of line><Erlang Atom><whitespace>(
5447 * It is added to the tags database if it doesn't match the
5448 * name of the previous clause header.
5450 * Return the size of the name of the function, or 0 if no function
5454 erlang_func (char *s
, char *last
)
5456 /* Name of last clause. */
5461 pos
= erlang_atom (s
);
5466 pos
= skip_spaces (s
+ pos
) - s
;
5468 /* Save only the first clause. */
5471 || len
!= (int)strlen (last
)
5472 || !strneq (s
, last
, len
)))
5474 make_tag (s
, len
, true, s
, pos
, lineno
, linecharno
);
5483 * Handle attributes. Currently, tags are generated for defines
5486 * They are on the form:
5487 * -define(foo, bar).
5488 * -define(Foo(M, N), M+N).
5489 * -record(graph, {vtab = notable, cyclic = true}).
5492 erlang_attribute (char *s
)
5496 if ((LOOKING_AT (cp
, "-define") || LOOKING_AT (cp
, "-record"))
5499 int len
= erlang_atom (skip_spaces (cp
));
5501 make_tag (cp
, len
, true, s
, cp
+ len
- s
, lineno
, linecharno
);
5508 * Consume an Erlang atom (or variable).
5509 * Return the number of bytes consumed, or -1 if there was an error.
5512 erlang_atom (char *s
)
5516 if (c_isalpha (s
[pos
]) || s
[pos
] == '_')
5518 /* The atom is unquoted. */
5521 while (c_isalnum (s
[pos
]) || s
[pos
] == '_');
5523 else if (s
[pos
] == '\'')
5525 for (pos
++; s
[pos
] != '\''; pos
++)
5526 if (s
[pos
] == '\0' /* multiline quoted atoms are ignored */
5527 || (s
[pos
] == '\\' && s
[++pos
] == '\0'))
5536 static char *scan_separators (char *);
5537 static void add_regex (char *, language
*);
5538 static char *substitute (char *, char *, struct re_registers
*);
5541 * Take a string like "/blah/" and turn it into "blah", verifying
5542 * that the first and last characters are the same, and handling
5543 * quoted separator characters. Actually, stops on the occurrence of
5544 * an unquoted separator. Also process \t, \n, etc. and turn into
5545 * appropriate characters. Works in place. Null terminates name string.
5546 * Returns pointer to terminating separator, or NULL for
5547 * unterminated regexps.
5550 scan_separators (char *name
)
5553 char *copyto
= name
;
5554 bool quoted
= false;
5556 for (++name
; *name
!= '\0'; ++name
)
5562 case 'a': *copyto
++ = '\007'; break; /* BEL (bell) */
5563 case 'b': *copyto
++ = '\b'; break; /* BS (back space) */
5564 case 'd': *copyto
++ = 0177; break; /* DEL (delete) */
5565 case 'e': *copyto
++ = 033; break; /* ESC (delete) */
5566 case 'f': *copyto
++ = '\f'; break; /* FF (form feed) */
5567 case 'n': *copyto
++ = '\n'; break; /* NL (new line) */
5568 case 'r': *copyto
++ = '\r'; break; /* CR (carriage return) */
5569 case 't': *copyto
++ = '\t'; break; /* TAB (horizontal tab) */
5570 case 'v': *copyto
++ = '\v'; break; /* VT (vertical tab) */
5576 /* Something else is quoted, so preserve the quote. */
5584 else if (*name
== '\\')
5586 else if (*name
== sep
)
5592 name
= NULL
; /* signal unterminated regexp */
5594 /* Terminate copied string. */
5599 /* Look at the argument of --regex or --no-regex and do the right
5600 thing. Same for each line of a regexp file. */
5602 analyze_regex (char *regex_arg
)
5604 if (regex_arg
== NULL
)
5606 free_regexps (); /* --no-regex: remove existing regexps */
5610 /* A real --regexp option or a line in a regexp file. */
5611 switch (regex_arg
[0])
5613 /* Comments in regexp file or null arg to --regex. */
5619 /* Read a regex file. This is recursive and may result in a
5620 loop, which will stop when the file descriptors are exhausted. */
5624 linebuffer regexbuf
;
5625 char *regexfile
= regex_arg
+ 1;
5627 /* regexfile is a file containing regexps, one per line. */
5628 regexfp
= fopen (regexfile
, "r" FOPEN_BINARY
);
5629 if (regexfp
== NULL
)
5631 linebuffer_init (®exbuf
);
5632 while (readline_internal (®exbuf
, regexfp
) > 0)
5633 analyze_regex (regexbuf
.buffer
);
5634 free (regexbuf
.buffer
);
5639 /* Regexp to be used for a specific language only. */
5643 char *lang_name
= regex_arg
+ 1;
5646 for (cp
= lang_name
; *cp
!= '}'; cp
++)
5649 error ("unterminated language name in regex: %s", regex_arg
);
5653 lang
= get_language_from_langname (lang_name
);
5656 add_regex (cp
, lang
);
5660 /* Regexp to be used for any language. */
5662 add_regex (regex_arg
, NULL
);
5667 /* Separate the regexp pattern, compile it,
5668 and care for optional name and modifiers. */
5670 add_regex (char *regexp_pattern
, language
*lang
)
5672 static struct re_pattern_buffer zeropattern
;
5673 char sep
, *pat
, *name
, *modifiers
;
5676 struct re_pattern_buffer
*patbuf
;
5679 force_explicit_name
= true, /* do not use implicit tag names */
5680 ignore_case
= false, /* case is significant */
5681 multi_line
= false, /* matches are done one line at a time */
5682 single_line
= false; /* dot does not match newline */
5685 if (strlen (regexp_pattern
) < 3)
5687 error ("null regexp");
5690 sep
= regexp_pattern
[0];
5691 name
= scan_separators (regexp_pattern
);
5694 error ("%s: unterminated regexp", regexp_pattern
);
5699 error ("null name for regexp \"%s\"", regexp_pattern
);
5702 modifiers
= scan_separators (name
);
5703 if (modifiers
== NULL
) /* no terminating separator --> no name */
5709 modifiers
+= 1; /* skip separator */
5711 /* Parse regex modifiers. */
5712 for (; modifiers
[0] != '\0'; modifiers
++)
5713 switch (modifiers
[0])
5716 if (modifiers
== name
)
5717 error ("forcing explicit tag name but no name, ignoring");
5718 force_explicit_name
= true;
5728 need_filebuf
= true;
5731 error ("invalid regexp modifier `%c', ignoring", modifiers
[0]);
5735 patbuf
= xnew (1, struct re_pattern_buffer
);
5736 *patbuf
= zeropattern
;
5739 static char lc_trans
[UCHAR_MAX
+ 1];
5741 for (i
= 0; i
< UCHAR_MAX
+ 1; i
++)
5742 lc_trans
[i
] = c_tolower (i
);
5743 patbuf
->translate
= lc_trans
; /* translation table to fold case */
5747 pat
= concat ("^", regexp_pattern
, ""); /* anchor to beginning of line */
5749 pat
= regexp_pattern
;
5752 re_set_syntax (RE_SYNTAX_EMACS
| RE_DOT_NEWLINE
);
5754 re_set_syntax (RE_SYNTAX_EMACS
);
5756 err
= re_compile_pattern (pat
, strlen (pat
), patbuf
);
5761 error ("%s while compiling pattern", err
);
5766 p_head
= xnew (1, regexp
);
5767 p_head
->pattern
= savestr (regexp_pattern
);
5768 p_head
->p_next
= rp
;
5769 p_head
->lang
= lang
;
5770 p_head
->pat
= patbuf
;
5771 p_head
->name
= savestr (name
);
5772 p_head
->error_signaled
= false;
5773 p_head
->force_explicit_name
= force_explicit_name
;
5774 p_head
->ignore_case
= ignore_case
;
5775 p_head
->multi_line
= multi_line
;
5779 * Do the substitutions indicated by the regular expression and
5783 substitute (char *in
, char *out
, struct re_registers
*regs
)
5786 int size
, dig
, diglen
;
5789 size
= strlen (out
);
5791 /* Pass 1: figure out how much to allocate by finding all \N strings. */
5792 if (out
[size
- 1] == '\\')
5793 fatal ("pattern error in \"%s\"", out
);
5794 for (t
= strchr (out
, '\\');
5796 t
= strchr (t
+ 2, '\\'))
5797 if (c_isdigit (t
[1]))
5800 diglen
= regs
->end
[dig
] - regs
->start
[dig
];
5806 /* Allocate space and do the substitutions. */
5808 result
= xnew (size
+ 1, char);
5810 for (t
= result
; *out
!= '\0'; out
++)
5811 if (*out
== '\\' && c_isdigit (*++out
))
5814 diglen
= regs
->end
[dig
] - regs
->start
[dig
];
5815 memcpy (t
, in
+ regs
->start
[dig
], diglen
);
5822 assert (t
<= result
+ size
);
5823 assert (t
- result
== (int)strlen (result
));
5828 /* Deallocate all regexps. */
5833 while (p_head
!= NULL
)
5835 rp
= p_head
->p_next
;
5836 free (p_head
->pattern
);
5837 free (p_head
->name
);
5845 * Reads the whole file as a single string from `filebuf' and looks for
5846 * multi-line regular expressions, creating tags on matches.
5847 * readline already dealt with normal regexps.
5849 * Idea by Ben Wing <ben@666.com> (2002).
5852 regex_tag_multiline (void)
5854 char *buffer
= filebuf
.buffer
;
5858 for (rp
= p_head
; rp
!= NULL
; rp
= rp
->p_next
)
5862 if (!rp
->multi_line
)
5863 continue; /* skip normal regexps */
5865 /* Generic initializations before parsing file from memory. */
5866 lineno
= 1; /* reset global line number */
5867 charno
= 0; /* reset global char number */
5868 linecharno
= 0; /* reset global char number of line start */
5870 /* Only use generic regexps or those for the current language. */
5871 if (rp
->lang
!= NULL
&& rp
->lang
!= curfdp
->lang
)
5874 while (match
>= 0 && match
< filebuf
.len
)
5876 match
= re_search (rp
->pat
, buffer
, filebuf
.len
, charno
,
5877 filebuf
.len
- match
, &rp
->regs
);
5882 if (!rp
->error_signaled
)
5884 error ("regexp stack overflow while matching \"%s\"",
5886 rp
->error_signaled
= true;
5893 if (match
== rp
->regs
.end
[0])
5895 if (!rp
->error_signaled
)
5897 error ("regexp matches the empty string: \"%s\"",
5899 rp
->error_signaled
= true;
5901 match
= -3; /* exit from while loop */
5905 /* Match occurred. Construct a tag. */
5906 while (charno
< rp
->regs
.end
[0])
5907 if (buffer
[charno
++] == '\n')
5908 lineno
++, linecharno
= charno
;
5910 if (name
[0] == '\0')
5912 else /* make a named tag */
5913 name
= substitute (buffer
, rp
->name
, &rp
->regs
);
5914 if (rp
->force_explicit_name
)
5915 /* Force explicit tag name, if a name is there. */
5916 pfnote (name
, true, buffer
+ linecharno
,
5917 charno
- linecharno
+ 1, lineno
, linecharno
);
5919 make_tag (name
, strlen (name
), true, buffer
+ linecharno
,
5920 charno
- linecharno
+ 1, lineno
, linecharno
);
5929 nocase_tail (const char *cp
)
5933 while (*cp
!= '\0' && c_tolower (*cp
) == c_tolower (dbp
[len
]))
5935 if (*cp
== '\0' && !intoken (dbp
[len
]))
5944 get_tag (register char *bp
, char **namepp
)
5946 register char *cp
= bp
;
5950 /* Go till you get to white space or a syntactic break */
5951 for (cp
= bp
+ 1; !notinname (*cp
); cp
++)
5953 make_tag (bp
, cp
- bp
, true,
5954 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
5958 *namepp
= savenstr (bp
, cp
- bp
);
5962 * Read a line of text from `stream' into `lbp', excluding the
5963 * newline or CR-NL, if any. Return the number of characters read from
5964 * `stream', which is the length of the line including the newline.
5966 * On DOS or Windows we do not count the CR character, if any before the
5967 * NL, in the returned length; this mirrors the behavior of Emacs on those
5968 * platforms (for text files, it translates CR-NL to NL as it reads in the
5971 * If multi-line regular expressions are requested, each line read is
5972 * appended to `filebuf'.
5975 readline_internal (linebuffer
*lbp
, register FILE *stream
)
5977 char *buffer
= lbp
->buffer
;
5978 register char *p
= lbp
->buffer
;
5979 register char *pend
;
5982 pend
= p
+ lbp
->size
; /* Separate to avoid 386/IX compiler bug. */
5986 register int c
= getc (stream
);
5989 /* We're at the end of linebuffer: expand it. */
5991 xrnew (buffer
, lbp
->size
, char);
5992 p
+= buffer
- lbp
->buffer
;
5993 pend
= buffer
+ lbp
->size
;
5994 lbp
->buffer
= buffer
;
6004 if (p
> buffer
&& p
[-1] == '\r')
6008 /* Assume CRLF->LF translation will be performed by Emacs
6009 when loading this file, so CRs won't appear in the buffer.
6010 It would be cleaner to compensate within Emacs;
6011 however, Emacs does not know how many CRs were deleted
6012 before any given point in the file. */
6027 lbp
->len
= p
- buffer
;
6029 if (need_filebuf
/* we need filebuf for multi-line regexps */
6030 && chars_deleted
> 0) /* not at EOF */
6032 while (filebuf
.size
<= filebuf
.len
+ lbp
->len
+ 1) /* +1 for \n */
6034 /* Expand filebuf. */
6036 xrnew (filebuf
.buffer
, filebuf
.size
, char);
6038 memcpy (filebuf
.buffer
+ filebuf
.len
, lbp
->buffer
, lbp
->len
);
6039 filebuf
.len
+= lbp
->len
;
6040 filebuf
.buffer
[filebuf
.len
++] = '\n';
6041 filebuf
.buffer
[filebuf
.len
] = '\0';
6044 return lbp
->len
+ chars_deleted
;
6048 * Like readline_internal, above, but in addition try to match the
6049 * input line against relevant regular expressions and manage #line
6053 readline (linebuffer
*lbp
, FILE *stream
)
6057 linecharno
= charno
; /* update global char number of line start */
6058 result
= readline_internal (lbp
, stream
); /* read line */
6059 lineno
+= 1; /* increment global line number */
6060 charno
+= result
; /* increment global char number */
6062 /* Honor #line directives. */
6063 if (!no_line_directive
)
6065 static bool discard_until_line_directive
;
6067 /* Check whether this is a #line directive. */
6068 if (result
> 12 && strneq (lbp
->buffer
, "#line ", 6))
6073 if (sscanf (lbp
->buffer
, "#line %u \"%n", &lno
, &start
) >= 1
6074 && start
> 0) /* double quote character found */
6076 char *endp
= lbp
->buffer
+ start
;
6078 while ((endp
= strchr (endp
, '"')) != NULL
6079 && endp
[-1] == '\\')
6082 /* Ok, this is a real #line directive. Let's deal with it. */
6084 char *taggedabsname
; /* absolute name of original file */
6085 char *taggedfname
; /* name of original file as given */
6086 char *name
; /* temp var */
6088 discard_until_line_directive
= false; /* found it */
6089 name
= lbp
->buffer
+ start
;
6091 canonicalize_filename (name
);
6092 taggedabsname
= absolute_filename (name
, tagfiledir
);
6093 if (filename_is_absolute (name
)
6094 || filename_is_absolute (curfdp
->infname
))
6095 taggedfname
= savestr (taggedabsname
);
6097 taggedfname
= relative_filename (taggedabsname
,tagfiledir
);
6099 if (streq (curfdp
->taggedfname
, taggedfname
))
6100 /* The #line directive is only a line number change. We
6101 deal with this afterwards. */
6104 /* The tags following this #line directive should be
6105 attributed to taggedfname. In order to do this, set
6106 curfdp accordingly. */
6108 fdesc
*fdp
; /* file description pointer */
6110 /* Go look for a file description already set up for the
6111 file indicated in the #line directive. If there is
6112 one, use it from now until the next #line
6114 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
6115 if (streq (fdp
->infname
, curfdp
->infname
)
6116 && streq (fdp
->taggedfname
, taggedfname
))
6117 /* If we remove the second test above (after the &&)
6118 then all entries pertaining to the same file are
6119 coalesced in the tags file. If we use it, then
6120 entries pertaining to the same file but generated
6121 from different files (via #line directives) will
6122 go into separate sections in the tags file. These
6123 alternatives look equivalent. The first one
6124 destroys some apparently useless information. */
6130 /* Else, if we already tagged the real file, skip all
6131 input lines until the next #line directive. */
6132 if (fdp
== NULL
) /* not found */
6133 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
6134 if (streq (fdp
->infabsname
, taggedabsname
))
6136 discard_until_line_directive
= true;
6140 /* Else create a new file description and use that from
6141 now on, until the next #line directive. */
6142 if (fdp
== NULL
) /* not found */
6145 fdhead
= xnew (1, fdesc
);
6146 *fdhead
= *curfdp
; /* copy curr. file description */
6148 fdhead
->infname
= savestr (curfdp
->infname
);
6149 fdhead
->infabsname
= savestr (curfdp
->infabsname
);
6150 fdhead
->infabsdir
= savestr (curfdp
->infabsdir
);
6151 fdhead
->taggedfname
= taggedfname
;
6152 fdhead
->usecharno
= false;
6153 fdhead
->prop
= NULL
;
6154 fdhead
->written
= false;
6158 free (taggedabsname
);
6160 readline (lbp
, stream
);
6162 } /* if a real #line directive */
6163 } /* if #line is followed by a number */
6164 } /* if line begins with "#line " */
6166 /* If we are here, no #line directive was found. */
6167 if (discard_until_line_directive
)
6171 /* Do a tail recursion on ourselves, thus discarding the contents
6172 of the line buffer. */
6173 readline (lbp
, stream
);
6177 discard_until_line_directive
= false;
6180 } /* if #line directives should be considered */
6187 /* Match against relevant regexps. */
6189 for (rp
= p_head
; rp
!= NULL
; rp
= rp
->p_next
)
6191 /* Only use generic regexps or those for the current language.
6192 Also do not use multiline regexps, which is the job of
6193 regex_tag_multiline. */
6194 if ((rp
->lang
!= NULL
&& rp
->lang
!= fdhead
->lang
)
6198 match
= re_match (rp
->pat
, lbp
->buffer
, lbp
->len
, 0, &rp
->regs
);
6203 if (!rp
->error_signaled
)
6205 error ("regexp stack overflow while matching \"%s\"",
6207 rp
->error_signaled
= true;
6214 /* Empty string matched. */
6215 if (!rp
->error_signaled
)
6217 error ("regexp matches the empty string: \"%s\"", rp
->pattern
);
6218 rp
->error_signaled
= true;
6222 /* Match occurred. Construct a tag. */
6224 if (name
[0] == '\0')
6226 else /* make a named tag */
6227 name
= substitute (lbp
->buffer
, rp
->name
, &rp
->regs
);
6228 if (rp
->force_explicit_name
)
6229 /* Force explicit tag name, if a name is there. */
6230 pfnote (name
, true, lbp
->buffer
, match
, lineno
, linecharno
);
6232 make_tag (name
, strlen (name
), true,
6233 lbp
->buffer
, match
, lineno
, linecharno
);
6242 * Return a pointer to a space of size strlen(cp)+1 allocated
6243 * with xnew where the string CP has been copied.
6246 savestr (const char *cp
)
6248 return savenstr (cp
, strlen (cp
));
6252 * Return a pointer to a space of size LEN+1 allocated with xnew where
6253 * the string CP has been copied for at most the first LEN characters.
6256 savenstr (const char *cp
, int len
)
6258 char *dp
= xnew (len
+ 1, char);
6260 return memcpy (dp
, cp
, len
);
6263 /* Skip spaces (end of string is not space), return new pointer. */
6265 skip_spaces (char *cp
)
6267 while (c_isspace (*cp
))
6272 /* Skip non spaces, except end of string, return new pointer. */
6274 skip_non_spaces (char *cp
)
6276 while (*cp
!= '\0' && !c_isspace (*cp
))
6281 /* Skip any chars in the "name" class.*/
6283 skip_name (char *cp
)
6285 /* '\0' is a notinname() so loop stops there too */
6286 while (! notinname (*cp
))
6291 /* Print error message and exit. */
6293 fatal (const char *s1
, const char *s2
)
6296 exit (EXIT_FAILURE
);
6300 pfatal (const char *s1
)
6303 exit (EXIT_FAILURE
);
6307 suggest_asking_for_help (void)
6309 fprintf (stderr
, "\tTry `%s --help' for a complete list of options.\n",
6311 exit (EXIT_FAILURE
);
6314 /* Output a diagnostic with printf-style FORMAT and args. */
6316 error (const char *format
, ...)
6319 va_start (ap
, format
);
6320 fprintf (stderr
, "%s: ", progname
);
6321 vfprintf (stderr
, format
, ap
);
6322 fprintf (stderr
, "\n");
6326 /* Return a newly-allocated string whose contents
6327 concatenate those of s1, s2, s3. */
6329 concat (const char *s1
, const char *s2
, const char *s3
)
6331 int len1
= strlen (s1
), len2
= strlen (s2
), len3
= strlen (s3
);
6332 char *result
= xnew (len1
+ len2
+ len3
+ 1, char);
6334 strcpy (result
, s1
);
6335 strcpy (result
+ len1
, s2
);
6336 strcpy (result
+ len1
+ len2
, s3
);
6342 /* Does the same work as the system V getcwd, but does not need to
6343 guess the buffer size in advance. */
6348 char *path
= xnew (bufsize
, char);
6350 while (getcwd (path
, bufsize
) == NULL
)
6352 if (errno
!= ERANGE
)
6356 path
= xnew (bufsize
, char);
6359 canonicalize_filename (path
);
6363 /* Return a newly allocated string containing a name of a temporary file. */
6367 const char *tmpdir
= getenv ("TMPDIR");
6368 const char *slash
= "/";
6370 #if MSDOS || defined (DOS_NT)
6372 tmpdir
= getenv ("TEMP");
6374 tmpdir
= getenv ("TMP");
6377 if (tmpdir
[strlen (tmpdir
) - 1] == '/'
6378 || tmpdir
[strlen (tmpdir
) - 1] == '\\')
6383 if (tmpdir
[strlen (tmpdir
) - 1] == '/')
6387 char *templt
= concat (tmpdir
, slash
, "etXXXXXX");
6388 int fd
= mkostemp (templt
, O_CLOEXEC
);
6397 #if defined (DOS_NT)
6398 /* The file name will be used in shell redirection, so it needs to have
6399 DOS-style backslashes, or else the Windows shell will barf. */
6401 for (p
= templt
; *p
; p
++)
6408 /* Return a newly allocated string containing the file name of FILE
6409 relative to the absolute directory DIR (which should end with a slash). */
6411 relative_filename (char *file
, char *dir
)
6413 char *fp
, *dp
, *afn
, *res
;
6416 /* Find the common root of file and dir (with a trailing slash). */
6417 afn
= absolute_filename (file
, cwd
);
6420 while (*fp
++ == *dp
++)
6422 fp
--, dp
--; /* back to the first differing char */
6424 if (fp
== afn
&& afn
[0] != '/') /* cannot build a relative name */
6427 do /* look at the equal chars until '/' */
6431 /* Build a sequence of "../" strings for the resulting relative file name. */
6433 while ((dp
= strchr (dp
+ 1, '/')) != NULL
)
6435 res
= xnew (3*i
+ strlen (fp
+ 1) + 1, char);
6438 z
= stpcpy (z
, "../");
6440 /* Add the file name relative to the common root of file and dir. */
6447 /* Return a newly allocated string containing the absolute file name
6448 of FILE given DIR (which should end with a slash). */
6450 absolute_filename (char *file
, char *dir
)
6452 char *slashp
, *cp
, *res
;
6454 if (filename_is_absolute (file
))
6455 res
= savestr (file
);
6457 /* We don't support non-absolute file names with a drive
6458 letter, like `d:NAME' (it's too much hassle). */
6459 else if (file
[1] == ':')
6460 fatal ("%s: relative file names with drive letters not supported", file
);
6463 res
= concat (dir
, file
, "");
6465 /* Delete the "/dirname/.." and "/." substrings. */
6466 slashp
= strchr (res
, '/');
6467 while (slashp
!= NULL
&& slashp
[0] != '\0')
6469 if (slashp
[1] == '.')
6471 if (slashp
[2] == '.'
6472 && (slashp
[3] == '/' || slashp
[3] == '\0'))
6477 while (cp
>= res
&& !filename_is_absolute (cp
));
6479 cp
= slashp
; /* the absolute name begins with "/.." */
6481 /* Under MSDOS and NT we get `d:/NAME' as absolute
6482 file name, so the luser could say `d:/../NAME'.
6483 We silently treat this as `d:/NAME'. */
6484 else if (cp
[0] != '/')
6487 memmove (cp
, slashp
+ 3, strlen (slashp
+ 2));
6491 else if (slashp
[2] == '/' || slashp
[2] == '\0')
6493 memmove (slashp
, slashp
+ 2, strlen (slashp
+ 1));
6498 slashp
= strchr (slashp
+ 1, '/');
6501 if (res
[0] == '\0') /* just a safety net: should never happen */
6504 return savestr ("/");
6510 /* Return a newly allocated string containing the absolute
6511 file name of dir where FILE resides given DIR (which should
6512 end with a slash). */
6514 absolute_dirname (char *file
, char *dir
)
6519 slashp
= strrchr (file
, '/');
6521 return savestr (dir
);
6524 res
= absolute_filename (file
, dir
);
6530 /* Whether the argument string is an absolute file name. The argument
6531 string must have been canonicalized with canonicalize_filename. */
6533 filename_is_absolute (char *fn
)
6535 return (fn
[0] == '/'
6537 || (c_isalpha (fn
[0]) && fn
[1] == ':' && fn
[2] == '/')
6542 /* Downcase DOS drive letter and collapse separators into single slashes.
6545 canonicalize_filename (register char *fn
)
6550 /* Canonicalize drive letter case. */
6551 if (c_isupper (fn
[0]) && fn
[1] == ':')
6552 fn
[0] = c_tolower (fn
[0]);
6554 /* Collapse multiple forward- and back-slashes into a single forward
6556 for (cp
= fn
; *cp
!= '\0'; cp
++, fn
++)
6557 if (*cp
== '/' || *cp
== '\\')
6560 while (cp
[1] == '/' || cp
[1] == '\\')
6568 /* Collapse multiple slashes into a single slash. */
6569 for (cp
= fn
; *cp
!= '\0'; cp
++, fn
++)
6573 while (cp
[1] == '/')
6579 #endif /* !DOS_NT */
6585 /* Initialize a linebuffer for use. */
6587 linebuffer_init (linebuffer
*lbp
)
6589 lbp
->size
= (DEBUG
) ? 3 : 200;
6590 lbp
->buffer
= xnew (lbp
->size
, char);
6591 lbp
->buffer
[0] = '\0';
6595 /* Set the minimum size of a string contained in a linebuffer. */
6597 linebuffer_setlen (linebuffer
*lbp
, int toksize
)
6599 while (lbp
->size
<= toksize
)
6602 xrnew (lbp
->buffer
, lbp
->size
, char);
6607 /* Like malloc but get fatal error if memory is exhausted. */
6609 xmalloc (size_t size
)
6611 void *result
= malloc (size
);
6613 fatal ("virtual memory exhausted", (char *)NULL
);
6618 xrealloc (void *ptr
, size_t size
)
6620 void *result
= realloc (ptr
, size
);
6622 fatal ("virtual memory exhausted", (char *)NULL
);
6628 * indent-tabs-mode: t
6631 * c-font-lock-extra-types: ("FILE" "bool" "language" "linebuffer" "fdesc" "node" "regexp")
6632 * c-file-style: "gnu"
6636 /* etags.c ends here */