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 <binary-io.h>
132 #include <c-strcase.h>
136 # undef assert /* some systems have a buggy assert.h */
137 # define assert(x) ((void) 0)
143 /* Define CTAGS to make the program "ctags" compatible with the usual one.
144 Leave it undefined to make the program "etags", which makes emacs-style
145 tag tables and tags typedefs, #defines and struct/union/enum by default. */
153 #define streq(s,t) (assert ((s)!=NULL || (t)!=NULL), !strcmp (s, t))
154 #define strcaseeq(s,t) (assert ((s)!=NULL && (t)!=NULL), !c_strcasecmp (s, t))
155 #define strneq(s,t,n) (assert ((s)!=NULL || (t)!=NULL), !strncmp (s, t, n))
156 #define strncaseeq(s,t,n) (assert ((s)!=NULL && (t)!=NULL), !c_strncasecmp (s, t, n))
158 /* C is not in a name. */
160 notinname (unsigned char c
)
162 /* Look at make_tag before modifying! */
163 static bool const table
[UCHAR_MAX
+ 1] = {
164 ['\0']=1, ['\t']=1, ['\n']=1, ['\f']=1, ['\r']=1, [' ']=1,
165 ['(']=1, [')']=1, [',']=1, [';']=1, ['=']=1
170 /* C can start a token. */
172 begtoken (unsigned char c
)
174 static bool const table
[UCHAR_MAX
+ 1] = {
176 ['A']=1, ['B']=1, ['C']=1, ['D']=1, ['E']=1, ['F']=1, ['G']=1, ['H']=1,
177 ['I']=1, ['J']=1, ['K']=1, ['L']=1, ['M']=1, ['N']=1, ['O']=1, ['P']=1,
178 ['Q']=1, ['R']=1, ['S']=1, ['T']=1, ['U']=1, ['V']=1, ['W']=1, ['X']=1,
181 ['a']=1, ['b']=1, ['c']=1, ['d']=1, ['e']=1, ['f']=1, ['g']=1, ['h']=1,
182 ['i']=1, ['j']=1, ['k']=1, ['l']=1, ['m']=1, ['n']=1, ['o']=1, ['p']=1,
183 ['q']=1, ['r']=1, ['s']=1, ['t']=1, ['u']=1, ['v']=1, ['w']=1, ['x']=1,
190 /* C can be in the middle of a token. */
192 intoken (unsigned char c
)
194 static bool const table
[UCHAR_MAX
+ 1] = {
196 ['0']=1, ['1']=1, ['2']=1, ['3']=1, ['4']=1,
197 ['5']=1, ['6']=1, ['7']=1, ['8']=1, ['9']=1,
198 ['A']=1, ['B']=1, ['C']=1, ['D']=1, ['E']=1, ['F']=1, ['G']=1, ['H']=1,
199 ['I']=1, ['J']=1, ['K']=1, ['L']=1, ['M']=1, ['N']=1, ['O']=1, ['P']=1,
200 ['Q']=1, ['R']=1, ['S']=1, ['T']=1, ['U']=1, ['V']=1, ['W']=1, ['X']=1,
203 ['a']=1, ['b']=1, ['c']=1, ['d']=1, ['e']=1, ['f']=1, ['g']=1, ['h']=1,
204 ['i']=1, ['j']=1, ['k']=1, ['l']=1, ['m']=1, ['n']=1, ['o']=1, ['p']=1,
205 ['q']=1, ['r']=1, ['s']=1, ['t']=1, ['u']=1, ['v']=1, ['w']=1, ['x']=1,
211 /* C can end a token. */
213 endtoken (unsigned char c
)
215 static bool const table
[UCHAR_MAX
+ 1] = {
216 ['\0']=1, ['\t']=1, ['\n']=1, ['\r']=1, [' ']=1,
217 ['!']=1, ['"']=1, ['#']=1, ['%']=1, ['&']=1, ['\'']=1, ['(']=1, [')']=1,
218 ['*']=1, ['+']=1, [',']=1, ['-']=1, ['.']=1, ['/']=1, [':']=1, [';']=1,
219 ['<']=1, ['=']=1, ['>']=1, ['?']=1, ['[']=1, [']']=1, ['^']=1,
220 ['{']=1, ['|']=1, ['}']=1, ['~']=1
226 * xnew, xrnew -- allocate, reallocate storage
228 * SYNOPSIS: Type *xnew (int n, Type);
229 * void xrnew (OldPointer, int n, Type);
231 #define xnew(n, Type) ((Type *) xmalloc ((n) * sizeof (Type)))
232 #define xrnew(op, n, Type) ((op) = (Type *) xrealloc (op, (n) * sizeof (Type)))
234 typedef void Lang_function (FILE *);
238 const char *suffix
; /* file name suffix for this compressor */
239 const char *command
; /* takes one arg and decompresses to stdout */
244 const char *name
; /* language name */
245 const char *help
; /* detailed help for the language */
246 Lang_function
*function
; /* parse function */
247 const char **suffixes
; /* name suffixes of this language's files */
248 const char **filenames
; /* names of this language's files */
249 const char **interpreters
; /* interpreters for this language */
250 bool metasource
; /* source used to generate other sources */
255 struct fdesc
*next
; /* for the linked list */
256 char *infname
; /* uncompressed input file name */
257 char *infabsname
; /* absolute uncompressed input file name */
258 char *infabsdir
; /* absolute dir of input file */
259 char *taggedfname
; /* file name to write in tagfile */
260 language
*lang
; /* language of file */
261 char *prop
; /* file properties to write in tagfile */
262 bool usecharno
; /* etags tags shall contain char number */
263 bool written
; /* entry written in the tags file */
266 typedef struct node_st
267 { /* sorting structure */
268 struct node_st
*left
, *right
; /* left and right sons */
269 fdesc
*fdp
; /* description of file to whom tag belongs */
270 char *name
; /* tag name */
271 char *regex
; /* search regexp */
272 bool valid
; /* write this tag on the tag file */
273 bool is_func
; /* function tag: use regexp in CTAGS mode */
274 bool been_warned
; /* warning already given for duplicated tag */
275 int lno
; /* line number tag is on */
276 long cno
; /* character number line starts on */
280 * A `linebuffer' is a structure which holds a line of text.
281 * `readline_internal' reads a line from a stream into a linebuffer
282 * and works regardless of the length of the line.
283 * SIZE is the size of BUFFER, LEN is the length of the string in
284 * BUFFER after readline reads it.
293 /* Used to support mixing of --lang and file names. */
297 at_language
, /* a language specification */
298 at_regexp
, /* a regular expression */
299 at_filename
, /* a file name */
300 at_stdin
, /* read from stdin here */
301 at_end
/* stop parsing the list */
302 } arg_type
; /* argument type */
303 language
*lang
; /* language associated with the argument */
304 char *what
; /* the argument itself */
307 /* Structure defining a regular expression. */
308 typedef struct regexp
310 struct regexp
*p_next
; /* pointer to next in list */
311 language
*lang
; /* if set, use only for this language */
312 char *pattern
; /* the regexp pattern */
313 char *name
; /* tag name */
314 struct re_pattern_buffer
*pat
; /* the compiled pattern */
315 struct re_registers regs
; /* re registers */
316 bool error_signaled
; /* already signaled for this regexp */
317 bool force_explicit_name
; /* do not allow implicit tag name */
318 bool ignore_case
; /* ignore case when matching */
319 bool multi_line
; /* do a multi-line match on the whole file */
323 /* Many compilers barf on this:
324 Lang_function Ada_funcs;
325 so let's write it this way */
326 static void Ada_funcs (FILE *);
327 static void Asm_labels (FILE *);
328 static void C_entries (int c_ext
, FILE *);
329 static void default_C_entries (FILE *);
330 static void plain_C_entries (FILE *);
331 static void Cjava_entries (FILE *);
332 static void Cobol_paragraphs (FILE *);
333 static void Cplusplus_entries (FILE *);
334 static void Cstar_entries (FILE *);
335 static void Erlang_functions (FILE *);
336 static void Forth_words (FILE *);
337 static void Fortran_functions (FILE *);
338 static void HTML_labels (FILE *);
339 static void Lisp_functions (FILE *);
340 static void Lua_functions (FILE *);
341 static void Makefile_targets (FILE *);
342 static void Pascal_functions (FILE *);
343 static void Perl_functions (FILE *);
344 static void PHP_functions (FILE *);
345 static void PS_functions (FILE *);
346 static void Prolog_functions (FILE *);
347 static void Python_functions (FILE *);
348 static void Scheme_functions (FILE *);
349 static void TeX_commands (FILE *);
350 static void Texinfo_nodes (FILE *);
351 static void Yacc_entries (FILE *);
352 static void just_read_file (FILE *);
354 static language
*get_language_from_langname (const char *);
355 static void readline (linebuffer
*, FILE *);
356 static long readline_internal (linebuffer
*, FILE *, char const *);
357 static bool nocase_tail (const char *);
358 static void get_tag (char *, char **);
360 static void analyze_regex (char *);
361 static void free_regexps (void);
362 static void regex_tag_multiline (void);
363 static void error (const char *, ...) ATTRIBUTE_FORMAT_PRINTF (1, 2);
364 static void verror (char const *, va_list) ATTRIBUTE_FORMAT_PRINTF (1, 0);
365 static _Noreturn
void suggest_asking_for_help (void);
366 static _Noreturn
void fatal (char const *, ...) ATTRIBUTE_FORMAT_PRINTF (1, 2);
367 static _Noreturn
void pfatal (const char *);
368 static void add_node (node
*, node
**);
370 static void process_file_name (char *, language
*);
371 static void process_file (FILE *, char *, language
*);
372 static void find_entries (FILE *);
373 static void free_tree (node
*);
374 static void free_fdesc (fdesc
*);
375 static void pfnote (char *, bool, char *, int, int, long);
376 static void invalidate_nodes (fdesc
*, node
**);
377 static void put_entries (node
*);
379 static char *concat (const char *, const char *, const char *);
380 static char *skip_spaces (char *);
381 static char *skip_non_spaces (char *);
382 static char *skip_name (char *);
383 static char *savenstr (const char *, int);
384 static char *savestr (const char *);
385 static char *etags_getcwd (void);
386 static char *relative_filename (char *, char *);
387 static char *absolute_filename (char *, char *);
388 static char *absolute_dirname (char *, char *);
389 static bool filename_is_absolute (char *f
);
390 static void canonicalize_filename (char *);
391 static char *etags_mktmp (void);
392 static void linebuffer_init (linebuffer
*);
393 static void linebuffer_setlen (linebuffer
*, int);
394 static void *xmalloc (size_t);
395 static void *xrealloc (void *, size_t);
398 static char searchar
= '/'; /* use /.../ searches */
400 static char *tagfile
; /* output file */
401 static char *progname
; /* name this program was invoked with */
402 static char *cwd
; /* current working directory */
403 static char *tagfiledir
; /* directory of tagfile */
404 static FILE *tagf
; /* ioptr for tags file */
405 static ptrdiff_t whatlen_max
; /* maximum length of any 'what' member */
407 static fdesc
*fdhead
; /* head of file description list */
408 static fdesc
*curfdp
; /* current file description */
409 static char *infilename
; /* current input file name */
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*/
445 static int class_qualify
; /* -Q: produce class-qualified tags in C++/Java */
447 /* STDIN is defined in LynxOS system headers */
452 #define STDIN 0x1001 /* returned by getopt_long on --parse-stdin */
453 static bool parsing_stdin
; /* --parse-stdin used */
455 static regexp
*p_head
; /* list of all regexps */
456 static bool need_filebuf
; /* some regexes are multi-line */
458 static struct option longopts
[] =
460 { "append", no_argument
, NULL
, 'a' },
461 { "packages-only", no_argument
, &packages_only
, 1 },
462 { "c++", no_argument
, NULL
, 'C' },
463 { "declarations", no_argument
, &declarations
, 1 },
464 { "no-line-directive", no_argument
, &no_line_directive
, 1 },
465 { "no-duplicates", no_argument
, &no_duplicates
, 1 },
466 { "help", no_argument
, NULL
, 'h' },
467 { "help", no_argument
, NULL
, 'H' },
468 { "ignore-indentation", no_argument
, NULL
, 'I' },
469 { "language", required_argument
, NULL
, 'l' },
470 { "members", no_argument
, &members
, 1 },
471 { "no-members", no_argument
, &members
, 0 },
472 { "output", required_argument
, NULL
, 'o' },
473 { "class-qualify", no_argument
, &class_qualify
, 'Q' },
474 { "regex", required_argument
, NULL
, 'r' },
475 { "no-regex", no_argument
, NULL
, 'R' },
476 { "ignore-case-regex", required_argument
, NULL
, 'c' },
477 { "parse-stdin", required_argument
, NULL
, STDIN
},
478 { "version", no_argument
, NULL
, 'V' },
480 #if CTAGS /* Ctags options */
481 { "backward-search", no_argument
, NULL
, 'B' },
482 { "cxref", no_argument
, NULL
, 'x' },
483 { "defines", no_argument
, NULL
, 'd' },
484 { "globals", no_argument
, &globals
, 1 },
485 { "typedefs", no_argument
, NULL
, 't' },
486 { "typedefs-and-c++", no_argument
, NULL
, 'T' },
487 { "update", no_argument
, NULL
, 'u' },
488 { "vgrind", no_argument
, NULL
, 'v' },
489 { "no-warn", no_argument
, NULL
, 'w' },
491 #else /* Etags options */
492 { "no-defines", no_argument
, NULL
, 'D' },
493 { "no-globals", no_argument
, &globals
, 0 },
494 { "include", required_argument
, NULL
, 'i' },
499 static compressor compressors
[] =
501 { "z", "gzip -d -c"},
502 { "Z", "gzip -d -c"},
503 { "gz", "gzip -d -c"},
504 { "GZ", "gzip -d -c"},
505 { "bz2", "bzip2 -d -c" },
506 { "xz", "xz -d -c" },
515 static const char *Ada_suffixes
[] =
516 { "ads", "adb", "ada", NULL
};
517 static const char Ada_help
[] =
518 "In Ada code, functions, procedures, packages, tasks and types are\n\
519 tags. Use the '--packages-only' option to create tags for\n\
521 Ada tag names have suffixes indicating the type of entity:\n\
522 Entity type: Qualifier:\n\
523 ------------ ----------\n\
530 Thus, 'M-x find-tag <RET> bidule/b <RET>' will go directly to the\n\
531 body of the package 'bidule', while 'M-x find-tag <RET> bidule <RET>'\n\
532 will just search for any tag 'bidule'.";
535 static const char *Asm_suffixes
[] =
536 { "a", /* Unix assembler */
537 "asm", /* Microcontroller assembly */
538 "def", /* BSO/Tasking definition includes */
539 "inc", /* Microcontroller include files */
540 "ins", /* Microcontroller include files */
541 "s", "sa", /* Unix assembler */
542 "S", /* cpp-processed Unix assembler */
543 "src", /* BSO/Tasking C compiler output */
546 static const char Asm_help
[] =
547 "In assembler code, labels appearing at the beginning of a line,\n\
548 followed by a colon, are tags.";
551 /* Note that .c and .h can be considered C++, if the --c++ flag was
552 given, or if the `class' or `template' keywords are met inside the file.
553 That is why default_C_entries is called for these. */
554 static const char *default_C_suffixes
[] =
556 #if CTAGS /* C help for Ctags */
557 static const char default_C_help
[] =
558 "In C code, any C function is a tag. Use -t to tag typedefs.\n\
559 Use -T to tag definitions of 'struct', 'union' and 'enum'.\n\
560 Use -d to tag '#define' macro definitions and 'enum' constants.\n\
561 Use --globals to tag global variables.\n\
562 You can tag function declarations and external variables by\n\
563 using '--declarations', and struct members by using '--members'.";
564 #else /* C help for Etags */
565 static const char default_C_help
[] =
566 "In C code, any C function or typedef is a tag, and so are\n\
567 definitions of 'struct', 'union' and 'enum'. '#define' macro\n\
568 definitions and 'enum' constants are tags unless you specify\n\
569 '--no-defines'. Global variables are tags unless you specify\n\
570 '--no-globals' and so are struct members unless you specify\n\
571 '--no-members'. Use of '--no-globals', '--no-defines' and\n\
572 '--no-members' can make the tags table file much smaller.\n\
573 You can tag function declarations and external variables by\n\
574 using '--declarations'.";
575 #endif /* C help for Ctags and Etags */
577 static const char *Cplusplus_suffixes
[] =
578 { "C", "c++", "cc", "cpp", "cxx", "H", "h++", "hh", "hpp", "hxx",
579 "M", /* Objective C++ */
580 "pdb", /* PostScript with C syntax */
582 static const char Cplusplus_help
[] =
583 "In C++ code, all the tag constructs of C code are tagged. (Use\n\
584 --help --lang=c --lang=c++ for full help.)\n\
585 In addition to C tags, member functions are also recognized. Member\n\
586 variables are recognized unless you use the '--no-members' option.\n\
587 Tags for variables and functions in classes are named 'CLASS::VARIABLE'\n\
588 and 'CLASS::FUNCTION'. 'operator' definitions have tag names like\n\
591 static const char *Cjava_suffixes
[] =
593 static char Cjava_help
[] =
594 "In Java code, all the tags constructs of C and C++ code are\n\
595 tagged. (Use --help --lang=c --lang=c++ --lang=java for full help.)";
598 static const char *Cobol_suffixes
[] =
599 { "COB", "cob", NULL
};
600 static char Cobol_help
[] =
601 "In Cobol code, tags are paragraph names; that is, any word\n\
602 starting in column 8 and followed by a period.";
604 static const char *Cstar_suffixes
[] =
605 { "cs", "hs", NULL
};
607 static const char *Erlang_suffixes
[] =
608 { "erl", "hrl", NULL
};
609 static const char Erlang_help
[] =
610 "In Erlang code, the tags are the functions, records and macros\n\
611 defined in the file.";
613 const char *Forth_suffixes
[] =
614 { "fth", "tok", NULL
};
615 static const char Forth_help
[] =
616 "In Forth code, tags are words defined by ':',\n\
617 constant, code, create, defer, value, variable, buffer:, field.";
619 static const char *Fortran_suffixes
[] =
620 { "F", "f", "f90", "for", NULL
};
621 static const char Fortran_help
[] =
622 "In Fortran code, functions, subroutines and block data are tags.";
624 static const char *HTML_suffixes
[] =
625 { "htm", "html", "shtml", NULL
};
626 static const char HTML_help
[] =
627 "In HTML input files, the tags are the 'title' and the 'h1', 'h2',\n\
628 'h3' headers. Also, tags are 'name=' in anchors and all\n\
629 occurrences of 'id='.";
631 static const char *Lisp_suffixes
[] =
632 { "cl", "clisp", "el", "l", "lisp", "LSP", "lsp", "ml", NULL
};
633 static const char Lisp_help
[] =
634 "In Lisp code, any function defined with 'defun', any variable\n\
635 defined with 'defvar' or 'defconst', and in general the first\n\
636 argument of any expression that starts with '(def' in column zero\n\
638 The '--declarations' option tags \"(defvar foo)\" constructs too.";
640 static const char *Lua_suffixes
[] =
641 { "lua", "LUA", NULL
};
642 static const char Lua_help
[] =
643 "In Lua scripts, all functions are tags.";
645 static const char *Makefile_filenames
[] =
646 { "Makefile", "makefile", "GNUMakefile", "Makefile.in", "Makefile.am", NULL
};
647 static const char Makefile_help
[] =
648 "In makefiles, targets are tags; additionally, variables are tags\n\
649 unless you specify '--no-globals'.";
651 static const char *Objc_suffixes
[] =
652 { "lm", /* Objective lex file */
653 "m", /* Objective C file */
655 static const char Objc_help
[] =
656 "In Objective C code, tags include Objective C definitions for classes,\n\
657 class categories, methods and protocols. Tags for variables and\n\
658 functions in classes are named 'CLASS::VARIABLE' and 'CLASS::FUNCTION'.\n\
659 (Use --help --lang=c --lang=objc --lang=java for full help.)";
661 static const char *Pascal_suffixes
[] =
662 { "p", "pas", NULL
};
663 static const char Pascal_help
[] =
664 "In Pascal code, the tags are the functions and procedures defined\n\
666 /* " // this is for working around an Emacs highlighting bug... */
668 static const char *Perl_suffixes
[] =
669 { "pl", "pm", NULL
};
670 static const char *Perl_interpreters
[] =
671 { "perl", "@PERL@", NULL
};
672 static const char Perl_help
[] =
673 "In Perl code, the tags are the packages, subroutines and variables\n\
674 defined by the 'package', 'sub', 'my' and 'local' keywords. Use\n\
675 '--globals' if you want to tag global variables. Tags for\n\
676 subroutines are named 'PACKAGE::SUB'. The name for subroutines\n\
677 defined in the default package is 'main::SUB'.";
679 static const char *PHP_suffixes
[] =
680 { "php", "php3", "php4", NULL
};
681 static const char PHP_help
[] =
682 "In PHP code, tags are functions, classes and defines. Unless you use\n\
683 the '--no-members' option, vars are tags too.";
685 static const char *plain_C_suffixes
[] =
686 { "pc", /* Pro*C file */
689 static const char *PS_suffixes
[] =
690 { "ps", "psw", NULL
}; /* .psw is for PSWrap */
691 static const char PS_help
[] =
692 "In PostScript code, the tags are the functions.";
694 static const char *Prolog_suffixes
[] =
696 static const char Prolog_help
[] =
697 "In Prolog code, tags are predicates and rules at the beginning of\n\
700 static const char *Python_suffixes
[] =
702 static const char Python_help
[] =
703 "In Python code, 'def' or 'class' at the beginning of a line\n\
706 /* Can't do the `SCM' or `scm' prefix with a version number. */
707 static const char *Scheme_suffixes
[] =
708 { "oak", "sch", "scheme", "SCM", "scm", "SM", "sm", "ss", "t", NULL
};
709 static const char Scheme_help
[] =
710 "In Scheme code, tags include anything defined with 'def' or with a\n\
711 construct whose name starts with 'def'. They also include\n\
712 variables set with 'set!' at top level in the file.";
714 static const char *TeX_suffixes
[] =
715 { "bib", "clo", "cls", "ltx", "sty", "TeX", "tex", NULL
};
716 static const char TeX_help
[] =
717 "In LaTeX text, the argument of any of the commands '\\chapter',\n\
718 '\\section', '\\subsection', '\\subsubsection', '\\eqno', '\\label',\n\
719 '\\ref', '\\cite', '\\bibitem', '\\part', '\\appendix', '\\entry',\n\
720 '\\index', '\\def', '\\newcommand', '\\renewcommand',\n\
721 '\\newenvironment' or '\\renewenvironment' is a tag.\n\
723 Other commands can be specified by setting the environment variable\n\
724 'TEXTAGS' to a colon-separated list like, for example,\n\
725 TEXTAGS=\"mycommand:myothercommand\".";
728 static const char *Texinfo_suffixes
[] =
729 { "texi", "texinfo", "txi", NULL
};
730 static const char Texinfo_help
[] =
731 "for texinfo files, lines starting with @node are tagged.";
733 static const char *Yacc_suffixes
[] =
734 { "y", "y++", "ym", "yxx", "yy", NULL
}; /* .ym is Objective yacc file */
735 static const char Yacc_help
[] =
736 "In Bison or Yacc input files, each rule defines as a tag the\n\
737 nonterminal it constructs. The portions of the file that contain\n\
738 C code are parsed as C code (use --help --lang=c --lang=yacc\n\
741 static const char auto_help
[] =
742 "'auto' is not a real language, it indicates to use\n\
743 a default language for files base on file name suffix and file contents.";
745 static const char none_help
[] =
746 "'none' is not a real language, it indicates to only do\n\
747 regexp processing on files.";
749 static const char no_lang_help
[] =
750 "No detailed help available for this language.";
754 * Table of languages.
756 * It is ok for a given function to be listed under more than one
757 * name. I just didn't.
760 static language lang_names
[] =
762 { "ada", Ada_help
, Ada_funcs
, Ada_suffixes
},
763 { "asm", Asm_help
, Asm_labels
, Asm_suffixes
},
764 { "c", default_C_help
, default_C_entries
, default_C_suffixes
},
765 { "c++", Cplusplus_help
, Cplusplus_entries
, Cplusplus_suffixes
},
766 { "c*", no_lang_help
, Cstar_entries
, Cstar_suffixes
},
767 { "cobol", Cobol_help
, Cobol_paragraphs
, Cobol_suffixes
},
768 { "erlang", Erlang_help
, Erlang_functions
, Erlang_suffixes
},
769 { "forth", Forth_help
, Forth_words
, Forth_suffixes
},
770 { "fortran", Fortran_help
, Fortran_functions
, Fortran_suffixes
},
771 { "html", HTML_help
, HTML_labels
, HTML_suffixes
},
772 { "java", Cjava_help
, Cjava_entries
, Cjava_suffixes
},
773 { "lisp", Lisp_help
, Lisp_functions
, Lisp_suffixes
},
774 { "lua", Lua_help
, Lua_functions
, Lua_suffixes
},
775 { "makefile", Makefile_help
,Makefile_targets
,NULL
,Makefile_filenames
},
776 { "objc", Objc_help
, plain_C_entries
, Objc_suffixes
},
777 { "pascal", Pascal_help
, Pascal_functions
, Pascal_suffixes
},
778 { "perl",Perl_help
,Perl_functions
,Perl_suffixes
,NULL
,Perl_interpreters
},
779 { "php", PHP_help
, PHP_functions
, PHP_suffixes
},
780 { "postscript",PS_help
, PS_functions
, PS_suffixes
},
781 { "proc", no_lang_help
, plain_C_entries
, plain_C_suffixes
},
782 { "prolog", Prolog_help
, Prolog_functions
, Prolog_suffixes
},
783 { "python", Python_help
, Python_functions
, Python_suffixes
},
784 { "scheme", Scheme_help
, Scheme_functions
, Scheme_suffixes
},
785 { "tex", TeX_help
, TeX_commands
, TeX_suffixes
},
786 { "texinfo", Texinfo_help
, Texinfo_nodes
, Texinfo_suffixes
},
787 { "yacc", Yacc_help
,Yacc_entries
,Yacc_suffixes
,NULL
,NULL
,true},
788 { "auto", auto_help
}, /* default guessing scheme */
789 { "none", none_help
, just_read_file
}, /* regexp matching only */
790 { NULL
} /* end of list */
795 print_language_names (void)
798 const char **name
, **ext
;
800 puts ("\nThese are the currently supported languages, along with the\n\
801 default file names and dot suffixes:");
802 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
804 printf (" %-*s", 10, lang
->name
);
805 if (lang
->filenames
!= NULL
)
806 for (name
= lang
->filenames
; *name
!= NULL
; name
++)
807 printf (" %s", *name
);
808 if (lang
->suffixes
!= NULL
)
809 for (ext
= lang
->suffixes
; *ext
!= NULL
; ext
++)
810 printf (" .%s", *ext
);
813 puts ("where 'auto' means use default language for files based on file\n\
814 name suffix, and 'none' means only do regexp processing on files.\n\
815 If no language is specified and no matching suffix is found,\n\
816 the first line of the file is read for a sharp-bang (#!) sequence\n\
817 followed by the name of an interpreter. If no such sequence is found,\n\
818 Fortran is tried first; if no tags are found, C is tried next.\n\
819 When parsing any C file, a \"class\" or \"template\" keyword\n\
821 puts ("Compressed files are supported using gzip, bzip2, and xz.\n\
823 For detailed help on a given language use, for example,\n\
824 etags --help --lang=ada.");
828 # define EMACS_NAME "standalone"
831 # define VERSION "17.38.1.4"
833 static _Noreturn
void
836 char emacs_copyright
[] = COPYRIGHT
;
838 printf ("%s (%s %s)\n", (CTAGS
) ? "ctags" : "etags", EMACS_NAME
, VERSION
);
839 puts (emacs_copyright
);
840 puts ("This program is distributed under the terms in ETAGS.README");
845 #ifndef PRINT_UNDOCUMENTED_OPTIONS_HELP
846 # define PRINT_UNDOCUMENTED_OPTIONS_HELP false
849 static _Noreturn
void
850 print_help (argument
*argbuffer
)
852 bool help_for_lang
= false;
854 for (; argbuffer
->arg_type
!= at_end
; argbuffer
++)
855 if (argbuffer
->arg_type
== at_language
)
859 puts (argbuffer
->lang
->help
);
860 help_for_lang
= true;
866 printf ("Usage: %s [options] [[regex-option ...] file-name] ...\n\
868 These are the options accepted by %s.\n", progname
, progname
);
869 puts ("You may use unambiguous abbreviations for the long option names.");
870 puts (" A - as file name means read names from stdin (one per line).\n\
871 Absolute names are stored in the output file as they are.\n\
872 Relative ones are stored relative to the output file's directory.\n");
874 puts ("-a, --append\n\
875 Append tag entries to existing tags file.");
877 puts ("--packages-only\n\
878 For Ada files, only generate tags for packages.");
881 puts ("-B, --backward-search\n\
882 Write the search commands for the tag entries using '?', the\n\
883 backward-search command instead of '/', the forward-search command.");
885 /* This option is mostly obsolete, because etags can now automatically
886 detect C++. Retained for backward compatibility and for debugging and
887 experimentation. In principle, we could want to tag as C++ even
888 before any "class" or "template" keyword.
890 Treat files whose name suffix defaults to C language as C++ files.");
893 puts ("--declarations\n\
894 In C and derived languages, create tags for function declarations,");
896 puts ("\tand create tags for extern variables if --globals is used.");
899 ("\tand create tags for extern variables unless --no-globals is used.");
902 puts ("-d, --defines\n\
903 Create tag entries for C #define constants and enum constants, too.");
905 puts ("-D, --no-defines\n\
906 Don't create tag entries for C #define constants and enum constants.\n\
907 This makes the tags file smaller.");
910 puts ("-i FILE, --include=FILE\n\
911 Include a note in tag file indicating that, when searching for\n\
912 a tag, one should also consult the tags file FILE after\n\
913 checking the current file.");
915 puts ("-l LANG, --language=LANG\n\
916 Force the following files to be considered as written in the\n\
917 named language up to the next --language=LANG option.");
921 Create tag entries for global variables in some languages.");
923 puts ("--no-globals\n\
924 Do not create tag entries for global variables in some\n\
925 languages. This makes the tags file smaller.");
927 if (PRINT_UNDOCUMENTED_OPTIONS_HELP
)
928 puts ("--no-line-directive\n\
929 Ignore #line preprocessor directives in C and derived languages.");
933 Create tag entries for members of structures in some languages.");
935 puts ("--no-members\n\
936 Do not create tag entries for members of structures\n\
937 in some languages.");
939 puts ("-Q, --class-qualify\n\
940 Qualify tag names with their class name in C++, ObjC, and Java.\n\
941 This produces tag names of the form \"class::member\" for C++,\n\
942 \"class(category)\" for Objective C, and \"class.member\" for Java.\n\
943 For Objective C, this also produces class methods qualified with\n\
944 their arguments, as in \"foo:bar:baz:more\".");
945 puts ("-r REGEXP, --regex=REGEXP or --regex=@regexfile\n\
946 Make a tag for each line matching a regular expression pattern\n\
947 in the following files. {LANGUAGE}REGEXP uses REGEXP for LANGUAGE\n\
948 files only. REGEXFILE is a file containing one REGEXP per line.\n\
949 REGEXP takes the form /TAGREGEXP/TAGNAME/MODS, where TAGNAME/ is\n\
950 optional. The TAGREGEXP pattern is anchored (as if preceded by ^).");
951 puts (" If TAGNAME/ is present, the tags created are named.\n\
952 For example Tcl named tags can be created with:\n\
953 --regex=\"/proc[ \\t]+\\([^ \\t]+\\)/\\1/.\".\n\
954 MODS are optional one-letter modifiers: 'i' means to ignore case,\n\
955 'm' means to allow multi-line matches, 's' implies 'm' and\n\
956 causes dot to match any character, including newline.");
958 puts ("-R, --no-regex\n\
959 Don't create tags from regexps for the following files.");
961 puts ("-I, --ignore-indentation\n\
962 In C and C++ do not assume that a closing brace in the first\n\
963 column is the final brace of a function or structure definition.");
965 puts ("-o FILE, --output=FILE\n\
966 Write the tags to FILE.");
968 puts ("--parse-stdin=NAME\n\
969 Read from standard input and record tags as belonging to file NAME.");
973 puts ("-t, --typedefs\n\
974 Generate tag entries for C and Ada typedefs.");
975 puts ("-T, --typedefs-and-c++\n\
976 Generate tag entries for C typedefs, C struct/enum/union tags,\n\
977 and C++ member functions.");
981 puts ("-u, --update\n\
982 Update the tag entries for the given files, leaving tag\n\
983 entries for other files in place. Currently, this is\n\
984 implemented by deleting the existing entries for the given\n\
985 files and then rewriting the new entries at the end of the\n\
986 tags file. It is often faster to simply rebuild the entire\n\
987 tag file than to use this.");
991 puts ("-v, --vgrind\n\
992 Print on the standard output an index of items intended for\n\
993 human consumption, similar to the output of vgrind. The index\n\
994 is sorted, and gives the page number of each item.");
996 if (PRINT_UNDOCUMENTED_OPTIONS_HELP
)
997 puts ("-w, --no-duplicates\n\
998 Do not create duplicate tag entries, for compatibility with\n\
999 traditional ctags.");
1001 if (PRINT_UNDOCUMENTED_OPTIONS_HELP
)
1002 puts ("-w, --no-warn\n\
1003 Suppress warning messages about duplicate tag entries.");
1005 puts ("-x, --cxref\n\
1006 Like --vgrind, but in the style of cxref, rather than vgrind.\n\
1007 The output uses line numbers instead of page numbers, but\n\
1008 beyond that the differences are cosmetic; try both to see\n\
1012 puts ("-V, --version\n\
1013 Print the version of the program.\n\
1015 Print this help message.\n\
1016 Followed by one or more '--language' options prints detailed\n\
1017 help about tag generation for the specified languages.");
1019 print_language_names ();
1022 puts ("Report bugs to bug-gnu-emacs@gnu.org");
1024 exit (EXIT_SUCCESS
);
1029 main (int argc
, char **argv
)
1032 unsigned int nincluded_files
;
1033 char **included_files
;
1034 argument
*argbuffer
;
1035 int current_arg
, file_count
;
1036 linebuffer filename_lb
;
1037 bool help_asked
= false;
1043 nincluded_files
= 0;
1044 included_files
= xnew (argc
, char *);
1048 /* Allocate enough no matter what happens. Overkill, but each one
1050 argbuffer
= xnew (argc
, argument
);
1053 * Always find typedefs and structure tags.
1054 * Also default to find macro constants, enum constants, struct
1055 * members and global variables. Do it for both etags and ctags.
1057 typedefs
= typedefs_or_cplusplus
= constantypedefs
= true;
1058 globals
= members
= true;
1060 /* When the optstring begins with a '-' getopt_long does not rearrange the
1061 non-options arguments to be at the end, but leaves them alone. */
1062 optstring
= concat ("-ac:Cf:Il:o:Qr:RSVhH",
1063 (CTAGS
) ? "BxdtTuvw" : "Di:",
1066 while ((opt
= getopt_long (argc
, argv
, optstring
, longopts
, NULL
)) != EOF
)
1070 /* If getopt returns 0, then it has already processed a
1071 long-named option. We should do nothing. */
1075 /* This means that a file name has been seen. Record it. */
1076 argbuffer
[current_arg
].arg_type
= at_filename
;
1077 argbuffer
[current_arg
].what
= optarg
;
1078 len
= strlen (optarg
);
1079 if (whatlen_max
< len
)
1086 /* Parse standard input. Idea by Vivek <vivek@etla.org>. */
1087 argbuffer
[current_arg
].arg_type
= at_stdin
;
1088 argbuffer
[current_arg
].what
= optarg
;
1089 len
= strlen (optarg
);
1090 if (whatlen_max
< len
)
1095 fatal ("cannot parse standard input more than once");
1096 parsing_stdin
= true;
1099 /* Common options. */
1100 case 'a': append_to_tagfile
= true; break;
1101 case 'C': cplusplus
= true; break;
1102 case 'f': /* for compatibility with old makefiles */
1106 error ("-o option may only be given once.");
1107 suggest_asking_for_help ();
1113 case 'S': /* for backward compatibility */
1114 ignoreindent
= true;
1118 language
*lang
= get_language_from_langname (optarg
);
1121 argbuffer
[current_arg
].lang
= lang
;
1122 argbuffer
[current_arg
].arg_type
= at_language
;
1128 /* Backward compatibility: support obsolete --ignore-case-regexp. */
1129 optarg
= concat (optarg
, "i", ""); /* memory leak here */
1132 argbuffer
[current_arg
].arg_type
= at_regexp
;
1133 argbuffer
[current_arg
].what
= optarg
;
1134 len
= strlen (optarg
);
1135 if (whatlen_max
< len
)
1140 argbuffer
[current_arg
].arg_type
= at_regexp
;
1141 argbuffer
[current_arg
].what
= NULL
;
1156 case 'D': constantypedefs
= false; break;
1157 case 'i': included_files
[nincluded_files
++] = optarg
; break;
1159 /* Ctags options. */
1160 case 'B': searchar
= '?'; break;
1161 case 'd': constantypedefs
= true; break;
1162 case 't': typedefs
= true; break;
1163 case 'T': typedefs
= typedefs_or_cplusplus
= true; break;
1164 case 'u': update
= true; break;
1165 case 'v': vgrind_style
= true; /*FALLTHRU*/
1166 case 'x': cxref_style
= true; break;
1167 case 'w': no_warnings
= true; break;
1169 suggest_asking_for_help ();
1173 /* No more options. Store the rest of arguments. */
1174 for (; optind
< argc
; optind
++)
1176 argbuffer
[current_arg
].arg_type
= at_filename
;
1177 argbuffer
[current_arg
].what
= argv
[optind
];
1178 len
= strlen (argv
[optind
]);
1179 if (whatlen_max
< len
)
1185 argbuffer
[current_arg
].arg_type
= at_end
;
1188 print_help (argbuffer
);
1191 if (nincluded_files
== 0 && file_count
== 0)
1193 error ("no input files specified.");
1194 suggest_asking_for_help ();
1198 if (tagfile
== NULL
)
1199 tagfile
= savestr (CTAGS
? "tags" : "TAGS");
1200 cwd
= etags_getcwd (); /* the current working directory */
1201 if (cwd
[strlen (cwd
) - 1] != '/')
1204 cwd
= concat (oldcwd
, "/", "");
1208 /* Compute base directory for relative file names. */
1209 if (streq (tagfile
, "-")
1210 || strneq (tagfile
, "/dev/", 5))
1211 tagfiledir
= cwd
; /* relative file names are relative to cwd */
1214 canonicalize_filename (tagfile
);
1215 tagfiledir
= absolute_dirname (tagfile
, cwd
);
1218 linebuffer_init (&lb
);
1219 linebuffer_init (&filename_lb
);
1220 linebuffer_init (&filebuf
);
1221 linebuffer_init (&token_name
);
1225 if (streq (tagfile
, "-"))
1228 SET_BINARY (fileno (stdout
));
1231 tagf
= fopen (tagfile
, append_to_tagfile
? "ab" : "wb");
1237 * Loop through files finding functions.
1239 for (i
= 0; i
< current_arg
; i
++)
1241 static language
*lang
; /* non-NULL if language is forced */
1244 switch (argbuffer
[i
].arg_type
)
1247 lang
= argbuffer
[i
].lang
;
1250 analyze_regex (argbuffer
[i
].what
);
1253 this_file
= argbuffer
[i
].what
;
1254 /* Input file named "-" means read file names from stdin
1255 (one per line) and use them. */
1256 if (streq (this_file
, "-"))
1259 fatal ("cannot parse standard input "
1260 "AND read file names from it");
1261 while (readline_internal (&filename_lb
, stdin
, "-") > 0)
1262 process_file_name (filename_lb
.buffer
, lang
);
1265 process_file_name (this_file
, lang
);
1268 this_file
= argbuffer
[i
].what
;
1269 process_file (stdin
, this_file
, lang
);
1272 error ("internal error: arg_type");
1278 free (filebuf
.buffer
);
1279 free (token_name
.buffer
);
1281 if (!CTAGS
|| cxref_style
)
1283 /* Write the remaining tags to tagf (ETAGS) or stdout (CXREF). */
1284 put_entries (nodehead
);
1285 free_tree (nodehead
);
1291 /* Output file entries that have no tags. */
1292 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
1294 fprintf (tagf
, "\f\n%s,0\n", fdp
->taggedfname
);
1296 while (nincluded_files
-- > 0)
1297 fprintf (tagf
, "\f\n%s,include\n", *included_files
++);
1299 if (fclose (tagf
) == EOF
)
1303 exit (EXIT_SUCCESS
);
1306 /* From here on, we are in (CTAGS && !cxref_style) */
1310 xmalloc (strlen (tagfile
) + whatlen_max
+
1311 sizeof "mv..OTAGS;fgrep -v '\t\t' OTAGS >;rm OTAGS");
1312 for (i
= 0; i
< current_arg
; ++i
)
1314 switch (argbuffer
[i
].arg_type
)
1320 continue; /* the for loop */
1322 char *z
= stpcpy (cmd
, "mv ");
1323 z
= stpcpy (z
, tagfile
);
1324 z
= stpcpy (z
, " OTAGS;fgrep -v '\t");
1325 z
= stpcpy (z
, argbuffer
[i
].what
);
1326 z
= stpcpy (z
, "\t' OTAGS >");
1327 z
= stpcpy (z
, tagfile
);
1328 strcpy (z
, ";rm OTAGS");
1329 if (system (cmd
) != EXIT_SUCCESS
)
1330 fatal ("failed to execute shell command");
1333 append_to_tagfile
= true;
1336 tagf
= fopen (tagfile
, append_to_tagfile
? "ab" : "wb");
1339 put_entries (nodehead
); /* write all the tags (CTAGS) */
1340 free_tree (nodehead
);
1342 if (fclose (tagf
) == EOF
)
1346 if (append_to_tagfile
|| update
)
1348 char *cmd
= xmalloc (2 * strlen (tagfile
) + sizeof "sort -u -o..");
1349 /* Maybe these should be used:
1350 setenv ("LC_COLLATE", "C", 1);
1351 setenv ("LC_ALL", "C", 1); */
1352 char *z
= stpcpy (cmd
, "sort -u -o ");
1353 z
= stpcpy (z
, tagfile
);
1355 strcpy (z
, tagfile
);
1356 exit (system (cmd
));
1358 return EXIT_SUCCESS
;
1363 * Return a compressor given the file name. If EXTPTR is non-zero,
1364 * return a pointer into FILE where the compressor-specific
1365 * extension begins. If no compressor is found, NULL is returned
1366 * and EXTPTR is not significant.
1367 * Idea by Vladimir Alexiev <vladimir@cs.ualberta.ca> (1998)
1370 get_compressor_from_suffix (char *file
, char **extptr
)
1373 char *slash
, *suffix
;
1375 /* File has been processed by canonicalize_filename,
1376 so we don't need to consider backslashes on DOS_NT. */
1377 slash
= strrchr (file
, '/');
1378 suffix
= strrchr (file
, '.');
1379 if (suffix
== NULL
|| suffix
< slash
)
1384 /* Let those poor souls who live with DOS 8+3 file name limits get
1385 some solace by treating foo.cgz as if it were foo.c.gz, etc.
1386 Only the first do loop is run if not MSDOS */
1389 for (compr
= compressors
; compr
->suffix
!= NULL
; compr
++)
1390 if (streq (compr
->suffix
, suffix
))
1393 break; /* do it only once: not really a loop */
1396 } while (*suffix
!= '\0');
1403 * Return a language given the name.
1406 get_language_from_langname (const char *name
)
1411 error ("empty language name");
1414 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1415 if (streq (name
, lang
->name
))
1417 error ("unknown language \"%s\"", name
);
1425 * Return a language given the interpreter name.
1428 get_language_from_interpreter (char *interpreter
)
1433 if (interpreter
== NULL
)
1435 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1436 if (lang
->interpreters
!= NULL
)
1437 for (iname
= lang
->interpreters
; *iname
!= NULL
; iname
++)
1438 if (streq (*iname
, interpreter
))
1447 * Return a language given the file name.
1450 get_language_from_filename (char *file
, int case_sensitive
)
1453 const char **name
, **ext
, *suffix
;
1455 /* Try whole file name first. */
1456 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1457 if (lang
->filenames
!= NULL
)
1458 for (name
= lang
->filenames
; *name
!= NULL
; name
++)
1459 if ((case_sensitive
)
1460 ? streq (*name
, file
)
1461 : strcaseeq (*name
, file
))
1464 /* If not found, try suffix after last dot. */
1465 suffix
= strrchr (file
, '.');
1469 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1470 if (lang
->suffixes
!= NULL
)
1471 for (ext
= lang
->suffixes
; *ext
!= NULL
; ext
++)
1472 if ((case_sensitive
)
1473 ? streq (*ext
, suffix
)
1474 : strcaseeq (*ext
, suffix
))
1481 * This routine is called on each file argument.
1484 process_file_name (char *file
, language
*lang
)
1489 char *compressed_name
, *uncompressed_name
;
1490 char *ext
, *real_name
, *tmp_name
;
1493 canonicalize_filename (file
);
1494 if (streq (file
, tagfile
) && !streq (tagfile
, "-"))
1496 error ("skipping inclusion of %s in self.", file
);
1499 compr
= get_compressor_from_suffix (file
, &ext
);
1502 compressed_name
= file
;
1503 uncompressed_name
= savenstr (file
, ext
- file
);
1507 compressed_name
= NULL
;
1508 uncompressed_name
= file
;
1511 /* If the canonicalized uncompressed name
1512 has already been dealt with, skip it silently. */
1513 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
1515 assert (fdp
->infname
!= NULL
);
1516 if (streq (uncompressed_name
, fdp
->infname
))
1520 inf
= fopen (file
, "r" FOPEN_BINARY
);
1525 int file_errno
= errno
;
1526 if (compressed_name
)
1528 /* Try with the given suffix. */
1529 inf
= fopen (uncompressed_name
, "r" FOPEN_BINARY
);
1531 real_name
= uncompressed_name
;
1535 /* Try all possible suffixes. */
1536 for (compr
= compressors
; compr
->suffix
!= NULL
; compr
++)
1538 compressed_name
= concat (file
, ".", compr
->suffix
);
1539 inf
= fopen (compressed_name
, "r" FOPEN_BINARY
);
1542 real_name
= compressed_name
;
1547 char *suf
= compressed_name
+ strlen (file
);
1548 size_t suflen
= strlen (compr
->suffix
) + 1;
1549 for ( ; suf
[1]; suf
++, suflen
--)
1551 memmove (suf
, suf
+ 1, suflen
);
1552 inf
= fopen (compressed_name
, "r" FOPEN_BINARY
);
1555 real_name
= compressed_name
;
1562 free (compressed_name
);
1563 compressed_name
= NULL
;
1574 if (real_name
== compressed_name
)
1577 tmp_name
= etags_mktmp ();
1582 #if MSDOS || defined (DOS_NT)
1583 char *cmd1
= concat (compr
->command
, " \"", real_name
);
1584 char *cmd
= concat (cmd1
, "\" > ", tmp_name
);
1586 char *cmd1
= concat (compr
->command
, " '", real_name
);
1587 char *cmd
= concat (cmd1
, "' > ", tmp_name
);
1591 if (system (cmd
) == -1)
1598 inf
= fopen (tmp_name
, "r" FOPEN_BINARY
);
1612 process_file (inf
, uncompressed_name
, lang
);
1614 retval
= fclose (inf
);
1615 if (real_name
== compressed_name
)
1624 if (compressed_name
!= file
)
1625 free (compressed_name
);
1626 if (uncompressed_name
!= file
)
1627 free (uncompressed_name
);
1634 process_file (FILE *fh
, char *fn
, language
*lang
)
1636 static const fdesc emptyfdesc
;
1640 /* Create a new input file description entry. */
1641 fdp
= xnew (1, fdesc
);
1644 fdp
->infname
= savestr (fn
);
1646 fdp
->infabsname
= absolute_filename (fn
, cwd
);
1647 fdp
->infabsdir
= absolute_dirname (fn
, cwd
);
1648 if (filename_is_absolute (fn
))
1650 /* An absolute file name. Canonicalize it. */
1651 fdp
->taggedfname
= absolute_filename (fn
, NULL
);
1655 /* A file name relative to cwd. Make it relative
1656 to the directory of the tags file. */
1657 fdp
->taggedfname
= relative_filename (fn
, tagfiledir
);
1659 fdp
->usecharno
= true; /* use char position when making tags */
1661 fdp
->written
= false; /* not written on tags file yet */
1664 curfdp
= fdhead
; /* the current file description */
1668 /* If not Ctags, and if this is not metasource and if it contained no #line
1669 directives, we can write the tags and free all nodes pointing to
1672 && curfdp
->usecharno
/* no #line directives in this file */
1673 && !curfdp
->lang
->metasource
)
1677 /* Look for the head of the sublist relative to this file. See add_node
1678 for the structure of the node tree. */
1680 for (np
= nodehead
; np
!= NULL
; prev
= np
, np
= np
->left
)
1681 if (np
->fdp
== curfdp
)
1684 /* If we generated tags for this file, write and delete them. */
1687 /* This is the head of the last sublist, if any. The following
1688 instructions depend on this being true. */
1689 assert (np
->left
== NULL
);
1691 assert (fdhead
== curfdp
);
1692 assert (last_node
->fdp
== curfdp
);
1693 put_entries (np
); /* write tags for file curfdp->taggedfname */
1694 free_tree (np
); /* remove the written nodes */
1696 nodehead
= NULL
; /* no nodes left */
1698 prev
->left
= NULL
; /* delete the pointer to the sublist */
1704 reset_input (FILE *inf
)
1706 if (fseek (inf
, 0, SEEK_SET
) != 0)
1707 perror (infilename
);
1711 * This routine opens the specified file and calls the function
1712 * which finds the function and type definitions.
1715 find_entries (FILE *inf
)
1718 language
*lang
= curfdp
->lang
;
1719 Lang_function
*parser
= NULL
;
1721 /* If user specified a language, use it. */
1722 if (lang
!= NULL
&& lang
->function
!= NULL
)
1724 parser
= lang
->function
;
1727 /* Else try to guess the language given the file name. */
1730 lang
= get_language_from_filename (curfdp
->infname
, true);
1731 if (lang
!= NULL
&& lang
->function
!= NULL
)
1733 curfdp
->lang
= lang
;
1734 parser
= lang
->function
;
1738 /* Else look for sharp-bang as the first two characters. */
1740 && readline_internal (&lb
, inf
, infilename
) > 0
1742 && lb
.buffer
[0] == '#'
1743 && lb
.buffer
[1] == '!')
1747 /* Set lp to point at the first char after the last slash in the
1748 line or, if no slashes, at the first nonblank. Then set cp to
1749 the first successive blank and terminate the string. */
1750 lp
= strrchr (lb
.buffer
+2, '/');
1754 lp
= skip_spaces (lb
.buffer
+ 2);
1755 cp
= skip_non_spaces (lp
);
1758 if (strlen (lp
) > 0)
1760 lang
= get_language_from_interpreter (lp
);
1761 if (lang
!= NULL
&& lang
->function
!= NULL
)
1763 curfdp
->lang
= lang
;
1764 parser
= lang
->function
;
1771 /* Else try to guess the language given the case insensitive file name. */
1774 lang
= get_language_from_filename (curfdp
->infname
, false);
1775 if (lang
!= NULL
&& lang
->function
!= NULL
)
1777 curfdp
->lang
= lang
;
1778 parser
= lang
->function
;
1782 /* Else try Fortran or C. */
1785 node
*old_last_node
= last_node
;
1787 curfdp
->lang
= get_language_from_langname ("fortran");
1790 if (old_last_node
== last_node
)
1791 /* No Fortran entries found. Try C. */
1794 curfdp
->lang
= get_language_from_langname (cplusplus
? "c++" : "c");
1800 if (!no_line_directive
1801 && curfdp
->lang
!= NULL
&& curfdp
->lang
->metasource
)
1802 /* It may be that this is a bingo.y file, and we already parsed a bingo.c
1803 file, or anyway we parsed a file that is automatically generated from
1804 this one. If this is the case, the bingo.c file contained #line
1805 directives that generated tags pointing to this file. Let's delete
1806 them all before parsing this file, which is the real source. */
1808 fdesc
**fdpp
= &fdhead
;
1809 while (*fdpp
!= NULL
)
1811 && streq ((*fdpp
)->taggedfname
, curfdp
->taggedfname
))
1812 /* We found one of those! We must delete both the file description
1813 and all tags referring to it. */
1815 fdesc
*badfdp
= *fdpp
;
1817 /* Delete the tags referring to badfdp->taggedfname
1818 that were obtained from badfdp->infname. */
1819 invalidate_nodes (badfdp
, &nodehead
);
1821 *fdpp
= badfdp
->next
; /* remove the bad description from the list */
1822 free_fdesc (badfdp
);
1825 fdpp
= &(*fdpp
)->next
; /* advance the list pointer */
1828 assert (parser
!= NULL
);
1830 /* Generic initializations before reading from file. */
1831 linebuffer_setlen (&filebuf
, 0); /* reset the file buffer */
1833 /* Generic initializations before parsing file with readline. */
1834 lineno
= 0; /* reset global line number */
1835 charno
= 0; /* reset global char number */
1836 linecharno
= 0; /* reset global char number of line start */
1840 regex_tag_multiline ();
1845 * Check whether an implicitly named tag should be created,
1846 * then call `pfnote'.
1847 * NAME is a string that is internally copied by this function.
1849 * TAGS format specification
1850 * Idea by Sam Kendall <kendall@mv.mv.com> (1997)
1851 * The following is explained in some more detail in etc/ETAGS.EBNF.
1853 * make_tag creates tags with "implicit tag names" (unnamed tags)
1854 * if the following are all true, assuming NONAM=" \f\t\n\r()=,;":
1855 * 1. NAME does not contain any of the characters in NONAM;
1856 * 2. LINESTART contains name as either a rightmost, or rightmost but
1857 * one character, substring;
1858 * 3. the character, if any, immediately before NAME in LINESTART must
1859 * be a character in NONAM;
1860 * 4. the character, if any, immediately after NAME in LINESTART must
1861 * also be a character in NONAM.
1863 * The implementation uses the notinname() macro, which recognizes the
1864 * characters stored in the string `nonam'.
1865 * etags.el needs to use the same characters that are in NONAM.
1868 make_tag (const char *name
, /* tag name, or NULL if unnamed */
1869 int namelen
, /* tag length */
1870 bool is_func
, /* tag is a function */
1871 char *linestart
, /* start of the line where tag is */
1872 int linelen
, /* length of the line where tag is */
1873 int lno
, /* line number */
1874 long int cno
) /* character number */
1876 bool named
= (name
!= NULL
&& namelen
> 0);
1879 if (!CTAGS
&& named
) /* maybe set named to false */
1880 /* Let's try to make an implicit tag name, that is, create an unnamed tag
1881 such that etags.el can guess a name from it. */
1884 register const char *cp
= name
;
1886 for (i
= 0; i
< namelen
; i
++)
1887 if (notinname (*cp
++))
1889 if (i
== namelen
) /* rule #1 */
1891 cp
= linestart
+ linelen
- namelen
;
1892 if (notinname (linestart
[linelen
-1]))
1893 cp
-= 1; /* rule #4 */
1894 if (cp
>= linestart
/* rule #2 */
1896 || notinname (cp
[-1])) /* rule #3 */
1897 && strneq (name
, cp
, namelen
)) /* rule #2 */
1898 named
= false; /* use implicit tag name */
1903 nname
= savenstr (name
, namelen
);
1905 pfnote (nname
, is_func
, linestart
, linelen
, lno
, cno
);
1910 pfnote (char *name
, bool is_func
, char *linestart
, int linelen
, int lno
,
1912 /* tag name, or NULL if unnamed */
1913 /* tag is a function */
1914 /* start of the line where tag is */
1915 /* length of the line where tag is */
1917 /* character number */
1921 assert (name
== NULL
|| name
[0] != '\0');
1922 if (CTAGS
&& name
== NULL
)
1925 np
= xnew (1, node
);
1927 /* If ctags mode, change name "main" to M<thisfilename>. */
1928 if (CTAGS
&& !cxref_style
&& streq (name
, "main"))
1930 char *fp
= strrchr (curfdp
->taggedfname
, '/');
1931 np
->name
= concat ("M", fp
== NULL
? curfdp
->taggedfname
: fp
+ 1, "");
1932 fp
= strrchr (np
->name
, '.');
1933 if (fp
!= NULL
&& fp
[1] != '\0' && fp
[2] == '\0')
1939 np
->been_warned
= false;
1941 np
->is_func
= is_func
;
1943 if (np
->fdp
->usecharno
)
1944 /* Our char numbers are 0-base, because of C language tradition?
1945 ctags compatibility? old versions compatibility? I don't know.
1946 Anyway, since emacs's are 1-base we expect etags.el to take care
1947 of the difference. If we wanted to have 1-based numbers, we would
1948 uncomment the +1 below. */
1949 np
->cno
= cno
/* + 1 */ ;
1951 np
->cno
= invalidcharno
;
1952 np
->left
= np
->right
= NULL
;
1953 if (CTAGS
&& !cxref_style
)
1955 if (strlen (linestart
) < 50)
1956 np
->regex
= concat (linestart
, "$", "");
1958 np
->regex
= savenstr (linestart
, 50);
1961 np
->regex
= savenstr (linestart
, linelen
);
1963 add_node (np
, &nodehead
);
1968 * recurse on left children, iterate on right children.
1971 free_tree (register node
*np
)
1975 register node
*node_right
= np
->right
;
1976 free_tree (np
->left
);
1986 * delete a file description
1989 free_fdesc (register fdesc
*fdp
)
1991 free (fdp
->infname
);
1992 free (fdp
->infabsname
);
1993 free (fdp
->infabsdir
);
1994 free (fdp
->taggedfname
);
2001 * Adds a node to the tree of nodes. In etags mode, sort by file
2002 * name. In ctags mode, sort by tag name. Make no attempt at
2005 * add_node is the only function allowed to add nodes, so it can
2009 add_node (node
*np
, node
**cur_node_p
)
2012 register node
*cur_node
= *cur_node_p
;
2014 if (cur_node
== NULL
)
2024 /* For each file name, tags are in a linked sublist on the right
2025 pointer. The first tags of different files are a linked list
2026 on the left pointer. last_node points to the end of the last
2028 if (last_node
!= NULL
&& last_node
->fdp
== np
->fdp
)
2030 /* Let's use the same sublist as the last added node. */
2031 assert (last_node
->right
== NULL
);
2032 last_node
->right
= np
;
2035 else if (cur_node
->fdp
== np
->fdp
)
2037 /* Scanning the list we found the head of a sublist which is
2038 good for us. Let's scan this sublist. */
2039 add_node (np
, &cur_node
->right
);
2042 /* The head of this sublist is not good for us. Let's try the
2044 add_node (np
, &cur_node
->left
);
2045 } /* if ETAGS mode */
2050 dif
= strcmp (np
->name
, cur_node
->name
);
2053 * If this tag name matches an existing one, then
2054 * do not add the node, but maybe print a warning.
2056 if (no_duplicates
&& !dif
)
2058 if (np
->fdp
== cur_node
->fdp
)
2062 fprintf (stderr
, "Duplicate entry in file %s, line %d: %s\n",
2063 np
->fdp
->infname
, lineno
, np
->name
);
2064 fprintf (stderr
, "Second entry ignored\n");
2067 else if (!cur_node
->been_warned
&& !no_warnings
)
2071 "Duplicate entry in files %s and %s: %s (Warning only)\n",
2072 np
->fdp
->infname
, cur_node
->fdp
->infname
, np
->name
);
2073 cur_node
->been_warned
= true;
2078 /* Actually add the node */
2079 add_node (np
, dif
< 0 ? &cur_node
->left
: &cur_node
->right
);
2080 } /* if CTAGS mode */
2084 * invalidate_nodes ()
2085 * Scan the node tree and invalidate all nodes pointing to the
2086 * given file description (CTAGS case) or free them (ETAGS case).
2089 invalidate_nodes (fdesc
*badfdp
, node
**npp
)
2098 if (np
->left
!= NULL
)
2099 invalidate_nodes (badfdp
, &np
->left
);
2100 if (np
->fdp
== badfdp
)
2102 if (np
->right
!= NULL
)
2103 invalidate_nodes (badfdp
, &np
->right
);
2107 assert (np
->fdp
!= NULL
);
2108 if (np
->fdp
== badfdp
)
2110 *npp
= np
->left
; /* detach the sublist from the list */
2111 np
->left
= NULL
; /* isolate it */
2112 free_tree (np
); /* free it */
2113 invalidate_nodes (badfdp
, npp
);
2116 invalidate_nodes (badfdp
, &np
->left
);
2121 static int total_size_of_entries (node
*);
2122 static int number_len (long) ATTRIBUTE_CONST
;
2124 /* Length of a non-negative number's decimal representation. */
2126 number_len (long int num
)
2129 while ((num
/= 10) > 0)
2135 * Return total number of characters that put_entries will output for
2136 * the nodes in the linked list at the right of the specified node.
2137 * This count is irrelevant with etags.el since emacs 19.34 at least,
2138 * but is still supplied for backward compatibility.
2141 total_size_of_entries (register node
*np
)
2143 register int total
= 0;
2145 for (; np
!= NULL
; np
= np
->right
)
2148 total
+= strlen (np
->regex
) + 1; /* pat\177 */
2149 if (np
->name
!= NULL
)
2150 total
+= strlen (np
->name
) + 1; /* name\001 */
2151 total
+= number_len ((long) np
->lno
) + 1; /* lno, */
2152 if (np
->cno
!= invalidcharno
) /* cno */
2153 total
+= number_len (np
->cno
);
2154 total
+= 1; /* newline */
2161 put_entries (register node
*np
)
2164 static fdesc
*fdp
= NULL
;
2169 /* Output subentries that precede this one */
2171 put_entries (np
->left
);
2173 /* Output this entry */
2182 fprintf (tagf
, "\f\n%s,%d\n",
2183 fdp
->taggedfname
, total_size_of_entries (np
));
2184 fdp
->written
= true;
2186 fputs (np
->regex
, tagf
);
2187 fputc ('\177', tagf
);
2188 if (np
->name
!= NULL
)
2190 fputs (np
->name
, tagf
);
2191 fputc ('\001', tagf
);
2193 fprintf (tagf
, "%d,", np
->lno
);
2194 if (np
->cno
!= invalidcharno
)
2195 fprintf (tagf
, "%ld", np
->cno
);
2201 if (np
->name
== NULL
)
2202 error ("internal error: NULL name in ctags mode.");
2207 fprintf (stdout
, "%s %s %d\n",
2208 np
->name
, np
->fdp
->taggedfname
, (np
->lno
+ 63) / 64);
2210 fprintf (stdout
, "%-16s %3d %-16s %s\n",
2211 np
->name
, np
->lno
, np
->fdp
->taggedfname
, np
->regex
);
2215 fprintf (tagf
, "%s\t%s\t", np
->name
, np
->fdp
->taggedfname
);
2218 { /* function or #define macro with args */
2219 putc (searchar
, tagf
);
2222 for (sp
= np
->regex
; *sp
; sp
++)
2224 if (*sp
== '\\' || *sp
== searchar
)
2228 putc (searchar
, tagf
);
2231 { /* anything else; text pattern inadequate */
2232 fprintf (tagf
, "%d", np
->lno
);
2237 } /* if this node contains a valid tag */
2239 /* Output subentries that follow this one */
2240 put_entries (np
->right
);
2242 put_entries (np
->left
);
2247 #define C_EXT 0x00fff /* C extensions */
2248 #define C_PLAIN 0x00000 /* C */
2249 #define C_PLPL 0x00001 /* C++ */
2250 #define C_STAR 0x00003 /* C* */
2251 #define C_JAVA 0x00005 /* JAVA */
2252 #define C_AUTO 0x01000 /* C, but switch to C++ if `class' is met */
2253 #define YACC 0x10000 /* yacc file */
2256 * The C symbol tables.
2261 st_C_objprot
, st_C_objimpl
, st_C_objend
,
2263 st_C_ignore
, st_C_attribute
,
2266 st_C_class
, st_C_template
,
2267 st_C_struct
, st_C_extern
, st_C_enum
, st_C_define
, st_C_typedef
2270 /* Feed stuff between (but not including) %[ and %] lines to:
2276 struct C_stab_entry { char *name; int c_ext; enum sym_type type; }
2280 while, 0, st_C_ignore
2281 switch, 0, st_C_ignore
2282 return, 0, st_C_ignore
2283 __attribute__, 0, st_C_attribute
2284 GTY, 0, st_C_attribute
2285 @interface, 0, st_C_objprot
2286 @protocol, 0, st_C_objprot
2287 @implementation,0, st_C_objimpl
2288 @end, 0, st_C_objend
2289 import, (C_JAVA & ~C_PLPL), st_C_ignore
2290 package, (C_JAVA & ~C_PLPL), st_C_ignore
2291 friend, C_PLPL, st_C_ignore
2292 extends, (C_JAVA & ~C_PLPL), st_C_javastruct
2293 implements, (C_JAVA & ~C_PLPL), st_C_javastruct
2294 interface, (C_JAVA & ~C_PLPL), st_C_struct
2295 class, 0, st_C_class
2296 namespace, C_PLPL, st_C_struct
2297 domain, C_STAR, st_C_struct
2298 union, 0, st_C_struct
2299 struct, 0, st_C_struct
2300 extern, 0, st_C_extern
2302 typedef, 0, st_C_typedef
2303 define, 0, st_C_define
2304 undef, 0, st_C_define
2305 operator, C_PLPL, st_C_operator
2306 template, 0, st_C_template
2307 # DEFUN used in emacs, the next three used in glibc (SYSCALL only for mach).
2308 DEFUN, 0, st_C_gnumacro
2309 SYSCALL, 0, st_C_gnumacro
2310 ENTRY, 0, st_C_gnumacro
2311 PSEUDO, 0, st_C_gnumacro
2312 # These are defined inside C functions, so currently they are not met.
2313 # EXFUN used in glibc, DEFVAR_* in emacs.
2314 #EXFUN, 0, st_C_gnumacro
2315 #DEFVAR_, 0, st_C_gnumacro
2317 and replace lines between %< and %> with its output, then:
2318 - remove the #if characterset check
2319 - make in_word_set static and not inline. */
2321 /* C code produced by gperf version 3.0.1 */
2322 /* Command-line: gperf -m 5 */
2323 /* Computed positions: -k'2-3' */
2325 struct C_stab_entry
{ const char *name
; int c_ext
; enum sym_type type
; };
2326 /* maximum key range = 33, duplicates = 0 */
2329 hash (const char *str
, int len
)
2331 static char const asso_values
[] =
2333 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2334 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2335 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2336 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2337 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2338 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2339 35, 35, 35, 35, 35, 35, 35, 35, 35, 3,
2340 26, 35, 35, 35, 35, 35, 35, 35, 27, 35,
2341 35, 35, 35, 24, 0, 35, 35, 35, 35, 0,
2342 35, 35, 35, 35, 35, 1, 35, 16, 35, 6,
2343 23, 0, 0, 35, 22, 0, 35, 35, 5, 0,
2344 0, 15, 1, 35, 6, 35, 8, 19, 35, 16,
2345 4, 5, 35, 35, 35, 35, 35, 35, 35, 35,
2346 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2347 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2348 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2349 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2350 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2351 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2352 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2353 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2354 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2355 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2356 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2357 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2358 35, 35, 35, 35, 35, 35
2365 hval
+= asso_values
[(unsigned char) str
[2]];
2368 hval
+= asso_values
[(unsigned char) str
[1]];
2374 static struct C_stab_entry
*
2375 in_word_set (register const char *str
, register unsigned int len
)
2379 TOTAL_KEYWORDS
= 33,
2380 MIN_WORD_LENGTH
= 2,
2381 MAX_WORD_LENGTH
= 15,
2386 static struct C_stab_entry wordlist
[] =
2389 {"if", 0, st_C_ignore
},
2390 {"GTY", 0, st_C_attribute
},
2391 {"@end", 0, st_C_objend
},
2392 {"union", 0, st_C_struct
},
2393 {"define", 0, st_C_define
},
2394 {"import", (C_JAVA
& ~C_PLPL
), st_C_ignore
},
2395 {"template", 0, st_C_template
},
2396 {"operator", C_PLPL
, st_C_operator
},
2397 {"@interface", 0, st_C_objprot
},
2398 {"implements", (C_JAVA
& ~C_PLPL
), st_C_javastruct
},
2399 {"friend", C_PLPL
, st_C_ignore
},
2400 {"typedef", 0, st_C_typedef
},
2401 {"return", 0, st_C_ignore
},
2402 {"@implementation",0, st_C_objimpl
},
2403 {"@protocol", 0, st_C_objprot
},
2404 {"interface", (C_JAVA
& ~C_PLPL
), st_C_struct
},
2405 {"extern", 0, st_C_extern
},
2406 {"extends", (C_JAVA
& ~C_PLPL
), st_C_javastruct
},
2407 {"struct", 0, st_C_struct
},
2408 {"domain", C_STAR
, st_C_struct
},
2409 {"switch", 0, st_C_ignore
},
2410 {"enum", 0, st_C_enum
},
2411 {"for", 0, st_C_ignore
},
2412 {"namespace", C_PLPL
, st_C_struct
},
2413 {"class", 0, st_C_class
},
2414 {"while", 0, st_C_ignore
},
2415 {"undef", 0, st_C_define
},
2416 {"package", (C_JAVA
& ~C_PLPL
), st_C_ignore
},
2417 {"__attribute__", 0, st_C_attribute
},
2418 {"SYSCALL", 0, st_C_gnumacro
},
2419 {"ENTRY", 0, st_C_gnumacro
},
2420 {"PSEUDO", 0, st_C_gnumacro
},
2421 {"DEFUN", 0, st_C_gnumacro
}
2424 if (len
<= MAX_WORD_LENGTH
&& len
>= MIN_WORD_LENGTH
)
2426 int key
= hash (str
, len
);
2428 if (key
<= MAX_HASH_VALUE
&& key
>= 0)
2430 const char *s
= wordlist
[key
].name
;
2432 if (*str
== *s
&& !strncmp (str
+ 1, s
+ 1, len
- 1) && s
[len
] == '\0')
2433 return &wordlist
[key
];
2440 static enum sym_type
2441 C_symtype (char *str
, int len
, int c_ext
)
2443 register struct C_stab_entry
*se
= in_word_set (str
, len
);
2445 if (se
== NULL
|| (se
->c_ext
&& !(c_ext
& se
->c_ext
)))
2452 * Ignoring __attribute__ ((list))
2454 static bool inattribute
; /* looking at an __attribute__ construct */
2457 * C functions and variables are recognized using a simple
2458 * finite automaton. fvdef is its state variable.
2462 fvnone
, /* nothing seen */
2463 fdefunkey
, /* Emacs DEFUN keyword seen */
2464 fdefunname
, /* Emacs DEFUN name seen */
2465 foperator
, /* func: operator keyword seen (cplpl) */
2466 fvnameseen
, /* function or variable name seen */
2467 fstartlist
, /* func: just after open parenthesis */
2468 finlist
, /* func: in parameter list */
2469 flistseen
, /* func: after parameter list */
2470 fignore
, /* func: before open brace */
2471 vignore
/* var-like: ignore until ';' */
2474 static bool fvextern
; /* func or var: extern keyword seen; */
2477 * typedefs are recognized using a simple finite automaton.
2478 * typdef is its state variable.
2482 tnone
, /* nothing seen */
2483 tkeyseen
, /* typedef keyword seen */
2484 ttypeseen
, /* defined type seen */
2485 tinbody
, /* inside typedef body */
2486 tend
, /* just before typedef tag */
2487 tignore
/* junk after typedef tag */
2491 * struct-like structures (enum, struct and union) are recognized
2492 * using another simple finite automaton. `structdef' is its state
2497 snone
, /* nothing seen yet,
2498 or in struct body if bracelev > 0 */
2499 skeyseen
, /* struct-like keyword seen */
2500 stagseen
, /* struct-like tag seen */
2501 scolonseen
/* colon seen after struct-like tag */
2505 * When objdef is different from onone, objtag is the name of the class.
2507 static const char *objtag
= "<uninited>";
2510 * Yet another little state machine to deal with preprocessor lines.
2514 dnone
, /* nothing seen */
2515 dsharpseen
, /* '#' seen as first char on line */
2516 ddefineseen
, /* '#' and 'define' seen */
2517 dignorerest
/* ignore rest of line */
2521 * State machine for Objective C protocols and implementations.
2522 * Idea by Tom R.Hageman <tom@basil.icce.rug.nl> (1995)
2526 onone
, /* nothing seen */
2527 oprotocol
, /* @interface or @protocol seen */
2528 oimplementation
, /* @implementations seen */
2529 otagseen
, /* class name seen */
2530 oparenseen
, /* parenthesis before category seen */
2531 ocatseen
, /* category name seen */
2532 oinbody
, /* in @implementation body */
2533 omethodsign
, /* in @implementation body, after +/- */
2534 omethodtag
, /* after method name */
2535 omethodcolon
, /* after method colon */
2536 omethodparm
, /* after method parameter */
2537 oignore
/* wait for @end */
2542 * Use this structure to keep info about the token read, and how it
2543 * should be tagged. Used by the make_C_tag function to build a tag.
2547 char *line
; /* string containing the token */
2548 int offset
; /* where the token starts in LINE */
2549 int length
; /* token length */
2551 The previous members can be used to pass strings around for generic
2552 purposes. The following ones specifically refer to creating tags. In this
2553 case the token contained here is the pattern that will be used to create a
2556 bool valid
; /* do not create a tag; the token should be
2557 invalidated whenever a state machine is
2558 reset prematurely */
2559 bool named
; /* create a named tag */
2560 int lineno
; /* source line number of tag */
2561 long linepos
; /* source char number of tag */
2562 } token
; /* latest token read */
2565 * Variables and functions for dealing with nested structures.
2566 * Idea by Mykola Dzyuba <mdzyuba@yahoo.com> (2001)
2568 static void pushclass_above (int, char *, int);
2569 static void popclass_above (int);
2570 static void write_classname (linebuffer
*, const char *qualifier
);
2573 char **cname
; /* nested class names */
2574 int *bracelev
; /* nested class brace level */
2575 int nl
; /* class nesting level (elements used) */
2576 int size
; /* length of the array */
2577 } cstack
; /* stack for nested declaration tags */
2578 /* Current struct nesting depth (namespace, class, struct, union, enum). */
2579 #define nestlev (cstack.nl)
2580 /* After struct keyword or in struct body, not inside a nested function. */
2581 #define instruct (structdef == snone && nestlev > 0 \
2582 && bracelev == cstack.bracelev[nestlev-1] + 1)
2585 pushclass_above (int bracelev
, char *str
, int len
)
2589 popclass_above (bracelev
);
2591 if (nl
>= cstack
.size
)
2593 int size
= cstack
.size
*= 2;
2594 xrnew (cstack
.cname
, size
, char *);
2595 xrnew (cstack
.bracelev
, size
, int);
2597 assert (nl
== 0 || cstack
.bracelev
[nl
-1] < bracelev
);
2598 cstack
.cname
[nl
] = (str
== NULL
) ? NULL
: savenstr (str
, len
);
2599 cstack
.bracelev
[nl
] = bracelev
;
2604 popclass_above (int bracelev
)
2608 for (nl
= cstack
.nl
- 1;
2609 nl
>= 0 && cstack
.bracelev
[nl
] >= bracelev
;
2612 free (cstack
.cname
[nl
]);
2618 write_classname (linebuffer
*cn
, const char *qualifier
)
2621 int qlen
= strlen (qualifier
);
2623 if (cstack
.nl
== 0 || cstack
.cname
[0] == NULL
)
2627 cn
->buffer
[0] = '\0';
2631 len
= strlen (cstack
.cname
[0]);
2632 linebuffer_setlen (cn
, len
);
2633 strcpy (cn
->buffer
, cstack
.cname
[0]);
2635 for (i
= 1; i
< cstack
.nl
; i
++)
2637 char *s
= cstack
.cname
[i
];
2640 linebuffer_setlen (cn
, len
+ qlen
+ strlen (s
));
2641 len
+= sprintf (cn
->buffer
+ len
, "%s%s", qualifier
, s
);
2646 static bool consider_token (char *, int, int, int *, int, int, bool *);
2647 static void make_C_tag (bool);
2651 * checks to see if the current token is at the start of a
2652 * function or variable, or corresponds to a typedef, or
2653 * is a struct/union/enum tag, or #define, or an enum constant.
2655 * *IS_FUNC_OR_VAR gets true if the token is a function or #define macro
2656 * with args. C_EXTP points to which language we are looking at.
2667 consider_token (char *str
, int len
, int c
, int *c_extp
,
2668 int bracelev
, int parlev
, bool *is_func_or_var
)
2669 /* IN: token pointer */
2670 /* IN: token length */
2671 /* IN: first char after the token */
2672 /* IN, OUT: C extensions mask */
2673 /* IN: brace level */
2674 /* IN: parenthesis level */
2675 /* OUT: function or variable found */
2677 /* When structdef is stagseen, scolonseen, or snone with bracelev > 0,
2678 structtype is the type of the preceding struct-like keyword, and
2679 structbracelev is the brace level where it has been seen. */
2680 static enum sym_type structtype
;
2681 static int structbracelev
;
2682 static enum sym_type toktype
;
2685 toktype
= C_symtype (str
, len
, *c_extp
);
2688 * Skip __attribute__
2690 if (toktype
== st_C_attribute
)
2697 * Advance the definedef state machine.
2702 /* We're not on a preprocessor line. */
2703 if (toktype
== st_C_gnumacro
)
2710 if (toktype
== st_C_define
)
2712 definedef
= ddefineseen
;
2716 definedef
= dignorerest
;
2721 * Make a tag for any macro, unless it is a constant
2722 * and constantypedefs is false.
2724 definedef
= dignorerest
;
2725 *is_func_or_var
= (c
== '(');
2726 if (!*is_func_or_var
&& !constantypedefs
)
2733 error ("internal error: definedef value.");
2742 if (toktype
== st_C_typedef
)
2765 if (structdef
== snone
&& fvdef
== fvnone
)
2787 case st_C_javastruct
:
2788 if (structdef
== stagseen
)
2789 structdef
= scolonseen
;
2793 if ((*c_extp
& C_AUTO
) /* automatic detection of C++ language */
2795 && definedef
== dnone
&& structdef
== snone
2796 && typdef
== tnone
&& fvdef
== fvnone
)
2797 *c_extp
= (*c_extp
| C_PLPL
) & ~C_AUTO
;
2798 if (toktype
== st_C_template
)
2805 && (typdef
== tkeyseen
2806 || (typedefs_or_cplusplus
&& structdef
== snone
)))
2808 structdef
= skeyseen
;
2809 structtype
= toktype
;
2810 structbracelev
= bracelev
;
2811 if (fvdef
== fvnameseen
)
2819 if (structdef
== skeyseen
)
2821 structdef
= stagseen
;
2825 if (typdef
!= tnone
)
2828 /* Detect Objective C constructs. */
2838 objdef
= oimplementation
;
2844 case oimplementation
:
2845 /* Save the class tag for functions or variables defined inside. */
2846 objtag
= savenstr (str
, len
);
2850 /* Save the class tag for categories. */
2851 objtag
= savenstr (str
, len
);
2853 *is_func_or_var
= true;
2857 *is_func_or_var
= true;
2865 objdef
= omethodtag
;
2866 linebuffer_setlen (&token_name
, len
);
2867 memcpy (token_name
.buffer
, str
, len
);
2868 token_name
.buffer
[len
] = '\0';
2874 objdef
= omethodparm
;
2879 objdef
= omethodtag
;
2882 int oldlen
= token_name
.len
;
2884 linebuffer_setlen (&token_name
, oldlen
+ len
);
2885 memcpy (token_name
.buffer
+ oldlen
, str
, len
);
2886 token_name
.buffer
[oldlen
+ len
] = '\0';
2892 if (toktype
== st_C_objend
)
2894 /* Memory leakage here: the string pointed by objtag is
2895 never released, because many tests would be needed to
2896 avoid breaking on incorrect input code. The amount of
2897 memory leaked here is the sum of the lengths of the
2907 /* A function, variable or enum constant? */
2929 *is_func_or_var
= true;
2933 && structdef
== snone
2934 && structtype
== st_C_enum
&& bracelev
> structbracelev
2935 /* Don't tag tokens in expressions that assign values to enum
2937 && fvdef
!= vignore
)
2938 return true; /* enum constant */
2944 fvdef
= fdefunname
; /* GNU macro */
2945 *is_func_or_var
= true;
2953 if ((strneq (str
, "asm", 3) && endtoken (str
[3]))
2954 || (strneq (str
, "__asm__", 7) && endtoken (str
[7])))
2965 if (len
>= 10 && strneq (str
+len
-10, "::operator", 10))
2967 if (*c_extp
& C_AUTO
) /* automatic detection of C++ */
2968 *c_extp
= (*c_extp
| C_PLPL
) & ~C_AUTO
;
2970 *is_func_or_var
= true;
2973 if (bracelev
> 0 && !instruct
)
2975 fvdef
= fvnameseen
; /* function or variable */
2976 *is_func_or_var
= true;
2991 * C_entries often keeps pointers to tokens or lines which are older than
2992 * the line currently read. By keeping two line buffers, and switching
2993 * them at end of line, it is possible to use those pointers.
3001 #define current_lb_is_new (newndx == curndx)
3002 #define switch_line_buffers() (curndx = 1 - curndx)
3004 #define curlb (lbs[curndx].lb)
3005 #define newlb (lbs[newndx].lb)
3006 #define curlinepos (lbs[curndx].linepos)
3007 #define newlinepos (lbs[newndx].linepos)
3009 #define plainc ((c_ext & C_EXT) == C_PLAIN)
3010 #define cplpl (c_ext & C_PLPL)
3011 #define cjava ((c_ext & C_JAVA) == C_JAVA)
3013 #define CNL_SAVE_DEFINEDEF() \
3015 curlinepos = charno; \
3016 readline (&curlb, inf); \
3017 lp = curlb.buffer; \
3024 CNL_SAVE_DEFINEDEF (); \
3025 if (savetoken.valid) \
3027 token = savetoken; \
3028 savetoken.valid = false; \
3030 definedef = dnone; \
3035 make_C_tag (bool isfun
)
3037 /* This function is never called when token.valid is false, but
3038 we must protect against invalid input or internal errors. */
3040 make_tag (token_name
.buffer
, token_name
.len
, isfun
, token
.line
,
3041 token
.offset
+token
.length
+1, token
.lineno
, token
.linepos
);
3043 { /* this branch is optimized away if !DEBUG */
3044 make_tag (concat ("INVALID TOKEN:-->", token_name
.buffer
, ""),
3045 token_name
.len
+ 17, isfun
, token
.line
,
3046 token
.offset
+token
.length
+1, token
.lineno
, token
.linepos
);
3047 error ("INVALID TOKEN");
3050 token
.valid
= false;
3054 perhaps_more_input (FILE *inf
)
3056 return !feof (inf
) && !ferror (inf
);
3062 * This routine finds functions, variables, typedefs,
3063 * #define's, enum constants and struct/union/enum definitions in
3064 * C syntax and adds them to the list.
3067 C_entries (int c_ext
, FILE *inf
)
3068 /* extension of C */
3071 register char c
; /* latest char read; '\0' for end of line */
3072 register char *lp
; /* pointer one beyond the character `c' */
3073 int curndx
, newndx
; /* indices for current and new lb */
3074 register int tokoff
; /* offset in line of start of current token */
3075 register int toklen
; /* length of current token */
3076 const char *qualifier
; /* string used to qualify names */
3077 int qlen
; /* length of qualifier */
3078 int bracelev
; /* current brace level */
3079 int bracketlev
; /* current bracket level */
3080 int parlev
; /* current parenthesis level */
3081 int attrparlev
; /* __attribute__ parenthesis level */
3082 int templatelev
; /* current template level */
3083 int typdefbracelev
; /* bracelev where a typedef struct body begun */
3084 bool incomm
, inquote
, inchar
, quotednl
, midtoken
;
3085 bool yacc_rules
; /* in the rules part of a yacc file */
3086 struct tok savetoken
= {0}; /* token saved during preprocessor handling */
3089 linebuffer_init (&lbs
[0].lb
);
3090 linebuffer_init (&lbs
[1].lb
);
3091 if (cstack
.size
== 0)
3093 cstack
.size
= (DEBUG
) ? 1 : 4;
3095 cstack
.cname
= xnew (cstack
.size
, char *);
3096 cstack
.bracelev
= xnew (cstack
.size
, int);
3099 tokoff
= toklen
= typdefbracelev
= 0; /* keep compiler quiet */
3100 curndx
= newndx
= 0;
3104 fvdef
= fvnone
; fvextern
= false; typdef
= tnone
;
3105 structdef
= snone
; definedef
= dnone
; objdef
= onone
;
3107 midtoken
= inquote
= inchar
= incomm
= quotednl
= false;
3108 token
.valid
= savetoken
.valid
= false;
3109 bracelev
= bracketlev
= parlev
= attrparlev
= templatelev
= 0;
3111 { qualifier
= "."; qlen
= 1; }
3113 { qualifier
= "::"; qlen
= 2; }
3116 while (perhaps_more_input (inf
))
3121 /* If we are at the end of the line, the next character is a
3122 '\0'; do not skip it, because it is what tells us
3123 to read the next line. */
3144 /* Newlines inside comments do not end macro definitions in
3146 CNL_SAVE_DEFINEDEF ();
3159 /* Newlines inside strings do not end macro definitions
3160 in traditional cpp, even though compilers don't
3161 usually accept them. */
3162 CNL_SAVE_DEFINEDEF ();
3172 /* Hmmm, something went wrong. */
3208 if (fvdef
!= finlist
&& fvdef
!= fignore
&& fvdef
!= vignore
)
3223 else if (/* cplpl && */ *lp
== '/')
3229 if ((c_ext
& YACC
) && *lp
== '%')
3231 /* Entering or exiting rules section in yacc file. */
3233 definedef
= dnone
; fvdef
= fvnone
; fvextern
= false;
3234 typdef
= tnone
; structdef
= snone
;
3235 midtoken
= inquote
= inchar
= incomm
= quotednl
= false;
3237 yacc_rules
= !yacc_rules
;
3243 if (definedef
== dnone
)
3246 bool cpptoken
= true;
3248 /* Look back on this line. If all blanks, or nonblanks
3249 followed by an end of comment, this is a preprocessor
3251 for (cp
= newlb
.buffer
; cp
< lp
-1; cp
++)
3252 if (!c_isspace (*cp
))
3254 if (*cp
== '*' && cp
[1] == '/')
3264 definedef
= dsharpseen
;
3265 /* This is needed for tagging enum values: when there are
3266 preprocessor conditionals inside the enum, we need to
3267 reset the value of fvdef so that the next enum value is
3268 tagged even though the one before it did not end in a
3270 if (fvdef
== vignore
&& instruct
&& parlev
== 0)
3272 if (strneq (cp
, "#if", 3) || strneq (cp
, "#el", 3))
3276 } /* if (definedef == dnone) */
3287 CNL_SAVE_DEFINEDEF ();
3294 /* Consider token only if some involved conditions are satisfied. */
3295 if (typdef
!= tignore
3296 && definedef
!= dignorerest
3299 && (definedef
!= dnone
3300 || structdef
!= scolonseen
)
3307 if (c
== ':' && *lp
== ':' && begtoken (lp
[1]))
3308 /* This handles :: in the middle,
3309 but not at the beginning of an identifier.
3310 Also, space-separated :: is not recognized. */
3312 if (c_ext
& C_AUTO
) /* automatic detection of C++ */
3313 c_ext
= (c_ext
| C_PLPL
) & ~C_AUTO
;
3317 goto still_in_token
;
3321 bool funorvar
= false;
3324 || consider_token (newlb
.buffer
+ tokoff
, toklen
, c
,
3325 &c_ext
, bracelev
, parlev
,
3328 if (fvdef
== foperator
)
3331 lp
= skip_spaces (lp
-1);
3335 && !c_isspace (*lp
) && *lp
!= '(')
3338 toklen
+= lp
- oldlp
;
3340 token
.named
= false;
3342 && nestlev
> 0 && definedef
== dnone
)
3343 /* in struct body */
3348 write_classname (&token_name
, qualifier
);
3349 len
= token_name
.len
;
3350 linebuffer_setlen (&token_name
,
3351 len
+ qlen
+ toklen
);
3352 sprintf (token_name
.buffer
+ len
, "%s%.*s",
3354 newlb
.buffer
+ tokoff
);
3358 linebuffer_setlen (&token_name
, toklen
);
3359 sprintf (token_name
.buffer
, "%.*s",
3360 toklen
, newlb
.buffer
+ tokoff
);
3364 else if (objdef
== ocatseen
)
3365 /* Objective C category */
3369 int len
= strlen (objtag
) + 2 + toklen
;
3370 linebuffer_setlen (&token_name
, len
);
3371 sprintf (token_name
.buffer
, "%s(%.*s)",
3373 newlb
.buffer
+ tokoff
);
3377 linebuffer_setlen (&token_name
, toklen
);
3378 sprintf (token_name
.buffer
, "%.*s",
3379 toklen
, newlb
.buffer
+ tokoff
);
3383 else if (objdef
== omethodtag
3384 || objdef
== omethodparm
)
3385 /* Objective C method */
3389 else if (fvdef
== fdefunname
)
3390 /* GNU DEFUN and similar macros */
3392 bool defun
= (newlb
.buffer
[tokoff
] == 'F');
3396 /* Rewrite the tag so that emacs lisp DEFUNs
3397 can be found by their elisp name */
3403 linebuffer_setlen (&token_name
, len
);
3404 memcpy (token_name
.buffer
,
3405 newlb
.buffer
+ off
, len
);
3406 token_name
.buffer
[len
] = '\0';
3409 if (token_name
.buffer
[len
] == '_')
3410 token_name
.buffer
[len
] = '-';
3411 token
.named
= defun
;
3415 linebuffer_setlen (&token_name
, toklen
);
3416 memcpy (token_name
.buffer
,
3417 newlb
.buffer
+ tokoff
, toklen
);
3418 token_name
.buffer
[toklen
] = '\0';
3419 /* Name macros and members. */
3420 token
.named
= (structdef
== stagseen
3421 || typdef
== ttypeseen
3424 && definedef
== dignorerest
)
3426 && definedef
== dnone
3427 && structdef
== snone
3430 token
.lineno
= lineno
;
3431 token
.offset
= tokoff
;
3432 token
.length
= toklen
;
3433 token
.line
= newlb
.buffer
;
3434 token
.linepos
= newlinepos
;
3437 if (definedef
== dnone
3438 && (fvdef
== fvnameseen
3439 || fvdef
== foperator
3440 || structdef
== stagseen
3442 || typdef
== ttypeseen
3443 || objdef
!= onone
))
3445 if (current_lb_is_new
)
3446 switch_line_buffers ();
3448 else if (definedef
!= dnone
3449 || fvdef
== fdefunname
3451 make_C_tag (funorvar
);
3453 else /* not yacc and consider_token failed */
3455 if (inattribute
&& fvdef
== fignore
)
3457 /* We have just met __attribute__ after a
3458 function parameter list: do not tag the
3465 } /* if (endtoken (c)) */
3466 else if (intoken (c
))
3472 } /* if (midtoken) */
3473 else if (begtoken (c
))
3481 /* This prevents tagging fb in
3482 void (__attribute__((noreturn)) *fb) (void);
3483 Fixing this is not easy and not very important. */
3487 if (plainc
|| declarations
)
3489 make_C_tag (true); /* a function */
3496 if (structdef
== stagseen
&& !cjava
)
3498 popclass_above (bracelev
);
3508 if (!yacc_rules
|| lp
== newlb
.buffer
+ 1)
3510 tokoff
= lp
- 1 - newlb
.buffer
;
3515 } /* if (begtoken) */
3516 } /* if must look at token */
3519 /* Detect end of line, colon, comma, semicolon and various braces
3520 after having handled a token.*/
3526 if (yacc_rules
&& token
.offset
== 0 && token
.valid
)
3528 make_C_tag (false); /* a yacc function */
3531 if (definedef
!= dnone
)
3537 make_C_tag (true); /* an Objective C class */
3541 objdef
= omethodcolon
;
3544 int toklen
= token_name
.len
;
3545 linebuffer_setlen (&token_name
, toklen
+ 1);
3546 strcpy (token_name
.buffer
+ toklen
, ":");
3552 if (structdef
== stagseen
)
3554 structdef
= scolonseen
;
3557 /* Should be useless, but may be work as a safety net. */
3558 if (cplpl
&& fvdef
== flistseen
)
3560 make_C_tag (true); /* a function */
3566 if (definedef
!= dnone
|| inattribute
)
3572 make_C_tag (false); /* a typedef */
3582 if (typdef
== tignore
|| cplpl
)
3586 if ((globals
&& bracelev
== 0 && (!fvextern
|| declarations
))
3587 || (members
&& instruct
))
3588 make_C_tag (false); /* a variable */
3591 token
.valid
= false;
3595 && (cplpl
|| !instruct
)
3596 && (typdef
== tnone
|| (typdef
!= tignore
&& instruct
)))
3598 && plainc
&& instruct
))
3599 make_C_tag (true); /* a function */
3605 && cplpl
&& structdef
== stagseen
)
3606 make_C_tag (false); /* forward declaration */
3608 token
.valid
= false;
3609 } /* switch (fvdef) */
3615 if (structdef
== stagseen
)
3619 if (definedef
!= dnone
|| inattribute
)
3625 make_C_tag (true); /* an Objective C method */
3640 if (instruct
&& parlev
== 0)
3651 && (!fvextern
|| declarations
))
3652 || (members
&& instruct
)))
3653 make_C_tag (false); /* a variable */
3656 if ((declarations
&& typdef
== tnone
&& !instruct
)
3657 || (members
&& typdef
!= tignore
&& instruct
))
3659 make_C_tag (true); /* a function */
3662 else if (!declarations
)
3664 token
.valid
= false;
3669 if (structdef
== stagseen
)
3673 if (definedef
!= dnone
|| inattribute
)
3675 if (structdef
== stagseen
)
3682 make_C_tag (false); /* a typedef */
3694 if ((members
&& bracelev
== 1)
3695 || (globals
&& bracelev
== 0
3696 && (!fvextern
|| declarations
)))
3697 make_C_tag (false); /* a variable */
3713 if (definedef
!= dnone
)
3715 if (objdef
== otagseen
&& parlev
== 0)
3716 objdef
= oparenseen
;
3720 if (typdef
== ttypeseen
3724 /* This handles constructs like:
3725 typedef void OperatorFun (int fun); */
3746 if (--attrparlev
== 0)
3747 inattribute
= false;
3750 if (definedef
!= dnone
)
3752 if (objdef
== ocatseen
&& parlev
== 1)
3754 make_C_tag (true); /* an Objective C category */
3770 || typdef
== ttypeseen
))
3773 make_C_tag (false); /* a typedef */
3776 else if (parlev
< 0) /* can happen due to ill-conceived #if's. */
3780 if (definedef
!= dnone
)
3782 if (typdef
== ttypeseen
)
3784 /* Whenever typdef is set to tinbody (currently only
3785 here), typdefbracelev should be set to bracelev. */
3787 typdefbracelev
= bracelev
;
3792 if (cplpl
&& !class_qualify
)
3794 /* Remove class and namespace qualifiers from the token,
3795 leaving only the method/member name. */
3796 char *cc
, *uqname
= token_name
.buffer
;
3797 char *tok_end
= token_name
.buffer
+ token_name
.len
;
3799 for (cc
= token_name
.buffer
; cc
< tok_end
; cc
++)
3801 if (*cc
== ':' && cc
[1] == ':')
3807 if (uqname
> token_name
.buffer
)
3809 int uqlen
= strlen (uqname
);
3810 linebuffer_setlen (&token_name
, uqlen
);
3811 memmove (token_name
.buffer
, uqname
, uqlen
+ 1);
3814 make_C_tag (true); /* a function */
3823 make_C_tag (true); /* an Objective C class */
3828 make_C_tag (true); /* an Objective C method */
3832 /* Neutralize `extern "C" {' grot. */
3833 if (bracelev
== 0 && structdef
== snone
&& nestlev
== 0
3843 case skeyseen
: /* unnamed struct */
3844 pushclass_above (bracelev
, NULL
, 0);
3847 case stagseen
: /* named struct or enum */
3848 case scolonseen
: /* a class */
3849 pushclass_above (bracelev
,token
.line
+token
.offset
, token
.length
);
3851 make_C_tag (false); /* a struct or enum */
3859 if (definedef
!= dnone
)
3861 if (fvdef
== fstartlist
)
3863 fvdef
= fvnone
; /* avoid tagging `foo' in `foo (*bar()) ()' */
3864 token
.valid
= false;
3868 if (definedef
!= dnone
)
3871 if (!ignoreindent
&& lp
== newlb
.buffer
+ 1)
3874 token
.valid
= false; /* unexpected value, token unreliable */
3875 bracelev
= 0; /* reset brace level if first column */
3876 parlev
= 0; /* also reset paren level, just in case... */
3878 else if (bracelev
< 0)
3880 token
.valid
= false; /* something gone amiss, token unreliable */
3883 if (bracelev
== 0 && fvdef
== vignore
)
3884 fvdef
= fvnone
; /* end of function */
3885 popclass_above (bracelev
);
3887 /* Only if typdef == tinbody is typdefbracelev significant. */
3888 if (typdef
== tinbody
&& bracelev
<= typdefbracelev
)
3890 assert (bracelev
== typdefbracelev
);
3895 if (definedef
!= dnone
)
3905 if ((members
&& bracelev
== 1)
3906 || (globals
&& bracelev
== 0 && (!fvextern
|| declarations
)))
3907 make_C_tag (false); /* a variable */
3915 && (structdef
== stagseen
|| fvdef
== fvnameseen
))
3922 if (templatelev
> 0)
3930 if (objdef
== oinbody
&& bracelev
== 0)
3932 objdef
= omethodsign
;
3937 case '#': case '~': case '&': case '%': case '/':
3938 case '|': case '^': case '!': case '.': case '?':
3939 if (definedef
!= dnone
)
3941 /* These surely cannot follow a function tag in C. */
3954 if (objdef
== otagseen
)
3956 make_C_tag (true); /* an Objective C class */
3959 /* If a macro spans multiple lines don't reset its state. */
3961 CNL_SAVE_DEFINEDEF ();
3967 } /* while not eof */
3969 free (lbs
[0].lb
.buffer
);
3970 free (lbs
[1].lb
.buffer
);
3974 * Process either a C++ file or a C file depending on the setting
3978 default_C_entries (FILE *inf
)
3980 C_entries (cplusplus
? C_PLPL
: C_AUTO
, inf
);
3983 /* Always do plain C. */
3985 plain_C_entries (FILE *inf
)
3990 /* Always do C++. */
3992 Cplusplus_entries (FILE *inf
)
3994 C_entries (C_PLPL
, inf
);
3997 /* Always do Java. */
3999 Cjava_entries (FILE *inf
)
4001 C_entries (C_JAVA
, inf
);
4006 Cstar_entries (FILE *inf
)
4008 C_entries (C_STAR
, inf
);
4011 /* Always do Yacc. */
4013 Yacc_entries (FILE *inf
)
4015 C_entries (YACC
, inf
);
4019 /* Useful macros. */
4020 #define LOOP_ON_INPUT_LINES(file_pointer, line_buffer, char_pointer) \
4021 while (perhaps_more_input (file_pointer) \
4022 && (readline (&(line_buffer), file_pointer), \
4023 (char_pointer) = (line_buffer).buffer, \
4026 #define LOOKING_AT(cp, kw) /* kw is the keyword, a literal string */ \
4027 ((assert ("" kw), true) /* syntax error if not a literal string */ \
4028 && strneq ((cp), kw, sizeof (kw)-1) /* cp points at kw */ \
4029 && notinname ((cp)[sizeof (kw)-1]) /* end of kw */ \
4030 && ((cp) = skip_spaces ((cp)+sizeof (kw)-1))) /* skip spaces */
4032 /* Similar to LOOKING_AT but does not use notinname, does not skip */
4033 #define LOOKING_AT_NOCASE(cp, kw) /* the keyword is a literal string */ \
4034 ((assert ("" kw), true) /* syntax error if not a literal string */ \
4035 && strncaseeq ((cp), kw, sizeof (kw)-1) /* cp points at kw */ \
4036 && ((cp) += sizeof (kw)-1)) /* skip spaces */
4039 * Read a file, but do no processing. This is used to do regexp
4040 * matching on files that have no language defined.
4043 just_read_file (FILE *inf
)
4045 while (perhaps_more_input (inf
))
4046 readline (&lb
, inf
);
4050 /* Fortran parsing */
4052 static void F_takeprec (void);
4053 static void F_getit (FILE *);
4058 dbp
= skip_spaces (dbp
);
4062 dbp
= skip_spaces (dbp
);
4063 if (strneq (dbp
, "(*)", 3))
4068 if (!c_isdigit (*dbp
))
4070 --dbp
; /* force failure */
4075 while (c_isdigit (*dbp
));
4083 dbp
= skip_spaces (dbp
);
4086 readline (&lb
, inf
);
4091 dbp
= skip_spaces (dbp
);
4093 if (!c_isalpha (*dbp
) && *dbp
!= '_' && *dbp
!= '$')
4095 for (cp
= dbp
+ 1; *cp
!= '\0' && intoken (*cp
); cp
++)
4097 make_tag (dbp
, cp
-dbp
, true,
4098 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4103 Fortran_functions (FILE *inf
)
4105 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
4108 dbp
++; /* Ratfor escape to fortran */
4109 dbp
= skip_spaces (dbp
);
4113 if (LOOKING_AT_NOCASE (dbp
, "recursive"))
4114 dbp
= skip_spaces (dbp
);
4116 if (LOOKING_AT_NOCASE (dbp
, "pure"))
4117 dbp
= skip_spaces (dbp
);
4119 if (LOOKING_AT_NOCASE (dbp
, "elemental"))
4120 dbp
= skip_spaces (dbp
);
4122 switch (c_tolower (*dbp
))
4125 if (nocase_tail ("integer"))
4129 if (nocase_tail ("real"))
4133 if (nocase_tail ("logical"))
4137 if (nocase_tail ("complex") || nocase_tail ("character"))
4141 if (nocase_tail ("double"))
4143 dbp
= skip_spaces (dbp
);
4146 if (nocase_tail ("precision"))
4152 dbp
= skip_spaces (dbp
);
4155 switch (c_tolower (*dbp
))
4158 if (nocase_tail ("function"))
4162 if (nocase_tail ("subroutine"))
4166 if (nocase_tail ("entry"))
4170 if (nocase_tail ("blockdata") || nocase_tail ("block data"))
4172 dbp
= skip_spaces (dbp
);
4173 if (*dbp
== '\0') /* assume un-named */
4174 make_tag ("blockdata", 9, true,
4175 lb
.buffer
, dbp
- lb
.buffer
, lineno
, linecharno
);
4177 F_getit (inf
); /* look for name */
4188 * Philippe Waroquiers (1998)
4191 /* Once we are positioned after an "interesting" keyword, let's get
4192 the real tag value necessary. */
4194 Ada_getit (FILE *inf
, const char *name_qualifier
)
4200 while (perhaps_more_input (inf
))
4202 dbp
= skip_spaces (dbp
);
4204 || (dbp
[0] == '-' && dbp
[1] == '-'))
4206 readline (&lb
, inf
);
4209 switch (c_tolower (*dbp
))
4212 if (nocase_tail ("body"))
4214 /* Skipping body of procedure body or package body or ....
4215 resetting qualifier to body instead of spec. */
4216 name_qualifier
= "/b";
4221 /* Skipping type of task type or protected type ... */
4222 if (nocase_tail ("type"))
4229 for (cp
= dbp
; *cp
!= '\0' && *cp
!= '"'; cp
++)
4234 dbp
= skip_spaces (dbp
);
4236 c_isalnum (*cp
) || *cp
== '_' || *cp
== '.';
4244 name
= concat (dbp
, name_qualifier
, "");
4246 make_tag (name
, strlen (name
), true,
4247 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4256 Ada_funcs (FILE *inf
)
4258 bool inquote
= false;
4259 bool skip_till_semicolumn
= false;
4261 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
4263 while (*dbp
!= '\0')
4265 /* Skip a string i.e. "abcd". */
4266 if (inquote
|| (*dbp
== '"'))
4268 dbp
= strchr (dbp
+ !inquote
, '"');
4273 continue; /* advance char */
4278 break; /* advance line */
4282 /* Skip comments. */
4283 if (dbp
[0] == '-' && dbp
[1] == '-')
4284 break; /* advance line */
4286 /* Skip character enclosed in single quote i.e. 'a'
4287 and skip single quote starting an attribute i.e. 'Image. */
4296 if (skip_till_semicolumn
)
4299 skip_till_semicolumn
= false;
4301 continue; /* advance char */
4304 /* Search for beginning of a token. */
4305 if (!begtoken (*dbp
))
4308 continue; /* advance char */
4311 /* We are at the beginning of a token. */
4312 switch (c_tolower (*dbp
))
4315 if (!packages_only
&& nocase_tail ("function"))
4316 Ada_getit (inf
, "/f");
4318 break; /* from switch */
4319 continue; /* advance char */
4321 if (!packages_only
&& nocase_tail ("procedure"))
4322 Ada_getit (inf
, "/p");
4323 else if (nocase_tail ("package"))
4324 Ada_getit (inf
, "/s");
4325 else if (nocase_tail ("protected")) /* protected type */
4326 Ada_getit (inf
, "/t");
4328 break; /* from switch */
4329 continue; /* advance char */
4332 if (typedefs
&& !packages_only
&& nocase_tail ("use"))
4334 /* when tagging types, avoid tagging use type Pack.Typename;
4335 for this, we will skip everything till a ; */
4336 skip_till_semicolumn
= true;
4337 continue; /* advance char */
4341 if (!packages_only
&& nocase_tail ("task"))
4342 Ada_getit (inf
, "/k");
4343 else if (typedefs
&& !packages_only
&& nocase_tail ("type"))
4345 Ada_getit (inf
, "/t");
4346 while (*dbp
!= '\0')
4350 break; /* from switch */
4351 continue; /* advance char */
4354 /* Look for the end of the token. */
4355 while (!endtoken (*dbp
))
4358 } /* advance char */
4359 } /* advance line */
4364 * Unix and microcontroller assembly tag handling
4365 * Labels: /^[a-zA-Z_.$][a-zA_Z0-9_.$]*[: ^I^J]/
4366 * Idea by Bob Weiner, Motorola Inc. (1994)
4369 Asm_labels (FILE *inf
)
4373 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4375 /* If first char is alphabetic or one of [_.$], test for colon
4376 following identifier. */
4377 if (c_isalpha (*cp
) || *cp
== '_' || *cp
== '.' || *cp
== '$')
4379 /* Read past label. */
4381 while (c_isalnum (*cp
) || *cp
== '_' || *cp
== '.' || *cp
== '$')
4383 if (*cp
== ':' || c_isspace (*cp
))
4384 /* Found end of label, so copy it and add it to the table. */
4385 make_tag (lb
.buffer
, cp
- lb
.buffer
, true,
4386 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4394 * Perl sub names: /^sub[ \t\n]+[^ \t\n{]+/
4395 * /^use constant[ \t\n]+[^ \t\n{=,;]+/
4396 * Perl variable names: /^(my|local).../
4397 * Original code by Bart Robinson <lomew@cs.utah.edu> (1995)
4398 * Additions by Michael Ernst <mernst@alum.mit.edu> (1997)
4399 * Ideas by Kai Großjohann <Kai.Grossjohann@CS.Uni-Dortmund.DE> (2001)
4402 Perl_functions (FILE *inf
)
4404 char *package
= savestr ("main"); /* current package name */
4407 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4409 cp
= skip_spaces (cp
);
4411 if (LOOKING_AT (cp
, "package"))
4414 get_tag (cp
, &package
);
4416 else if (LOOKING_AT (cp
, "sub"))
4422 while (!notinname (*cp
))
4425 continue; /* nothing found */
4426 pos
= strchr (sp
, ':');
4427 if (pos
&& pos
< cp
&& pos
[1] == ':')
4428 /* The name is already qualified. */
4429 make_tag (sp
, cp
- sp
, true,
4430 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4434 char savechar
, *name
;
4438 name
= concat (package
, "::", sp
);
4440 make_tag (name
, strlen (name
), true,
4441 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4445 else if (LOOKING_AT (cp
, "use constant")
4446 || LOOKING_AT (cp
, "use constant::defer"))
4448 /* For hash style multi-constant like
4449 use constant { FOO => 123,
4451 only the first FOO is picked up. Parsing across the value
4452 expressions would be difficult in general, due to possible nested
4453 hashes, here-documents, etc. */
4455 cp
= skip_spaces (cp
+1);
4458 else if (globals
) /* only if we are tagging global vars */
4460 /* Skip a qualifier, if any. */
4461 bool qual
= LOOKING_AT (cp
, "my") || LOOKING_AT (cp
, "local");
4462 /* After "my" or "local", but before any following paren or space. */
4463 char *varstart
= cp
;
4465 if (qual
/* should this be removed? If yes, how? */
4466 && (*cp
== '$' || *cp
== '@' || *cp
== '%'))
4471 while (c_isalnum (*cp
) || *cp
== '_');
4475 /* Should be examining a variable list at this point;
4476 could insist on seeing an open parenthesis. */
4477 while (*cp
!= '\0' && *cp
!= ';' && *cp
!= '=' && *cp
!= ')')
4483 make_tag (varstart
, cp
- varstart
, false,
4484 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4493 * Look for /^[\t]*def[ \t\n]+[^ \t\n(:]+/ or /^class[ \t\n]+[^ \t\n(:]+/
4494 * Idea by Eric S. Raymond <esr@thyrsus.com> (1997)
4495 * More ideas by seb bacon <seb@jamkit.com> (2002)
4498 Python_functions (FILE *inf
)
4502 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4504 cp
= skip_spaces (cp
);
4505 if (LOOKING_AT (cp
, "def") || LOOKING_AT (cp
, "class"))
4508 while (!notinname (*cp
) && *cp
!= ':')
4510 make_tag (name
, cp
- name
, true,
4511 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4520 * - /^[ \t]*function[ \t\n]+[^ \t\n(]+/
4521 * - /^[ \t]*class[ \t\n]+[^ \t\n]+/
4522 * - /^[ \t]*define\(\"[^\"]+/
4523 * Only with --members:
4524 * - /^[ \t]*var[ \t\n]+\$[^ \t\n=;]/
4525 * Idea by Diez B. Roggisch (2001)
4528 PHP_functions (FILE *inf
)
4531 bool search_identifier
= false;
4533 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4535 cp
= skip_spaces (cp
);
4537 if (search_identifier
4540 while (!notinname (*cp
))
4542 make_tag (name
, cp
- name
, true,
4543 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4544 search_identifier
= false;
4546 else if (LOOKING_AT (cp
, "function"))
4549 cp
= skip_spaces (cp
+1);
4553 while (!notinname (*cp
))
4555 make_tag (name
, cp
- name
, true,
4556 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4559 search_identifier
= true;
4561 else if (LOOKING_AT (cp
, "class"))
4566 while (*cp
!= '\0' && !c_isspace (*cp
))
4568 make_tag (name
, cp
- name
, false,
4569 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4572 search_identifier
= true;
4574 else if (strneq (cp
, "define", 6)
4575 && (cp
= skip_spaces (cp
+6))
4577 && (*cp
== '"' || *cp
== '\''))
4581 while (*cp
!= quote
&& *cp
!= '\0')
4583 make_tag (name
, cp
- name
, false,
4584 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4587 && LOOKING_AT (cp
, "var")
4591 while (!notinname (*cp
))
4593 make_tag (name
, cp
- name
, false,
4594 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4601 * Cobol tag functions
4602 * We could look for anything that could be a paragraph name.
4603 * i.e. anything that starts in column 8 is one word and ends in a full stop.
4604 * Idea by Corny de Souza (1993)
4607 Cobol_paragraphs (FILE *inf
)
4609 register char *bp
, *ep
;
4611 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4617 /* If eoln, compiler option or comment ignore whole line. */
4618 if (bp
[-1] != ' ' || !c_isalnum (bp
[0]))
4621 for (ep
= bp
; c_isalnum (*ep
) || *ep
== '-'; ep
++)
4624 make_tag (bp
, ep
- bp
, true,
4625 lb
.buffer
, ep
- lb
.buffer
+ 1, lineno
, linecharno
);
4632 * Ideas by Assar Westerlund <assar@sics.se> (2001)
4635 Makefile_targets (FILE *inf
)
4639 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4641 if (*bp
== '\t' || *bp
== '#')
4643 while (*bp
!= '\0' && *bp
!= '=' && *bp
!= ':')
4645 if (*bp
== ':' || (globals
&& *bp
== '='))
4647 /* We should detect if there is more than one tag, but we do not.
4648 We just skip initial and final spaces. */
4649 char * namestart
= skip_spaces (lb
.buffer
);
4650 while (--bp
> namestart
)
4651 if (!notinname (*bp
))
4653 make_tag (namestart
, bp
- namestart
+ 1, true,
4654 lb
.buffer
, bp
- lb
.buffer
+ 2, lineno
, linecharno
);
4662 * Original code by Mosur K. Mohan (1989)
4664 * Locates tags for procedures & functions. Doesn't do any type- or
4665 * var-definitions. It does look for the keyword "extern" or
4666 * "forward" immediately following the procedure statement; if found,
4667 * the tag is skipped.
4670 Pascal_functions (FILE *inf
)
4672 linebuffer tline
; /* mostly copied from C_entries */
4674 int save_lineno
, namelen
, taglen
;
4677 bool /* each of these flags is true if: */
4678 incomment
, /* point is inside a comment */
4679 inquote
, /* point is inside '..' string */
4680 get_tagname
, /* point is after PROCEDURE/FUNCTION
4681 keyword, so next item = potential tag */
4682 found_tag
, /* point is after a potential tag */
4683 inparms
, /* point is within parameter-list */
4684 verify_tag
; /* point has passed the parm-list, so the
4685 next token will determine whether this
4686 is a FORWARD/EXTERN to be ignored, or
4687 whether it is a real tag */
4689 save_lcno
= save_lineno
= namelen
= taglen
= 0; /* keep compiler quiet */
4690 name
= NULL
; /* keep compiler quiet */
4693 linebuffer_init (&tline
);
4695 incomment
= inquote
= false;
4696 found_tag
= false; /* have a proc name; check if extern */
4697 get_tagname
= false; /* found "procedure" keyword */
4698 inparms
= false; /* found '(' after "proc" */
4699 verify_tag
= false; /* check if "extern" is ahead */
4702 while (perhaps_more_input (inf
)) /* long main loop to get next char */
4705 if (c
== '\0') /* if end of line */
4707 readline (&lb
, inf
);
4711 if (!((found_tag
&& verify_tag
)
4713 c
= *dbp
++; /* only if don't need *dbp pointing
4714 to the beginning of the name of
4715 the procedure or function */
4719 if (c
== '}') /* within { } comments */
4721 else if (c
== '*' && *dbp
== ')') /* within (* *) comments */
4738 inquote
= true; /* found first quote */
4740 case '{': /* found open { comment */
4744 if (*dbp
== '*') /* found open (* comment */
4749 else if (found_tag
) /* found '(' after tag, i.e., parm-list */
4752 case ')': /* end of parms list */
4757 if (found_tag
&& !inparms
) /* end of proc or fn stmt */
4764 if (found_tag
&& verify_tag
&& (*dbp
!= ' '))
4766 /* Check if this is an "extern" declaration. */
4769 if (c_tolower (*dbp
) == 'e')
4771 if (nocase_tail ("extern")) /* superfluous, really! */
4777 else if (c_tolower (*dbp
) == 'f')
4779 if (nocase_tail ("forward")) /* check for forward reference */
4785 if (found_tag
&& verify_tag
) /* not external proc, so make tag */
4789 make_tag (name
, namelen
, true,
4790 tline
.buffer
, taglen
, save_lineno
, save_lcno
);
4794 if (get_tagname
) /* grab name of proc or fn */
4801 /* Find block name. */
4802 for (cp
= dbp
+ 1; *cp
!= '\0' && !endtoken (*cp
); cp
++)
4805 /* Save all values for later tagging. */
4806 linebuffer_setlen (&tline
, lb
.len
);
4807 strcpy (tline
.buffer
, lb
.buffer
);
4808 save_lineno
= lineno
;
4809 save_lcno
= linecharno
;
4810 name
= tline
.buffer
+ (dbp
- lb
.buffer
);
4812 taglen
= cp
- lb
.buffer
+ 1;
4814 dbp
= cp
; /* set dbp to e-o-token */
4815 get_tagname
= false;
4819 /* And proceed to check for "extern". */
4821 else if (!incomment
&& !inquote
&& !found_tag
)
4823 /* Check for proc/fn keywords. */
4824 switch (c_tolower (c
))
4827 if (nocase_tail ("rocedure")) /* c = 'p', dbp has advanced */
4831 if (nocase_tail ("unction"))
4836 } /* while not eof */
4838 free (tline
.buffer
);
4843 * Lisp tag functions
4844 * look for (def or (DEF, quote or QUOTE
4847 static void L_getit (void);
4852 if (*dbp
== '\'') /* Skip prefix quote */
4854 else if (*dbp
== '(')
4857 /* Try to skip "(quote " */
4858 if (!LOOKING_AT (dbp
, "quote") && !LOOKING_AT (dbp
, "QUOTE"))
4859 /* Ok, then skip "(" before name in (defstruct (foo)) */
4860 dbp
= skip_spaces (dbp
);
4862 get_tag (dbp
, NULL
);
4866 Lisp_functions (FILE *inf
)
4868 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
4873 /* "(defvar foo)" is a declaration rather than a definition. */
4877 if (LOOKING_AT (p
, "defvar"))
4879 p
= skip_name (p
); /* past var name */
4880 p
= skip_spaces (p
);
4886 if (strneq (dbp
+ 1, "cl-", 3) || strneq (dbp
+ 1, "CL-", 3))
4889 if (strneq (dbp
+1, "def", 3) || strneq (dbp
+1, "DEF", 3))
4891 dbp
= skip_non_spaces (dbp
);
4892 dbp
= skip_spaces (dbp
);
4897 /* Check for (foo::defmumble name-defined ... */
4900 while (!notinname (*dbp
) && *dbp
!= ':');
4905 while (*dbp
== ':');
4907 if (strneq (dbp
, "def", 3) || strneq (dbp
, "DEF", 3))
4909 dbp
= skip_non_spaces (dbp
);
4910 dbp
= skip_spaces (dbp
);
4920 * Lua script language parsing
4921 * Original code by David A. Capello <dacap@users.sourceforge.net> (2004)
4923 * "function" and "local function" are tags if they start at column 1.
4926 Lua_functions (FILE *inf
)
4930 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4932 if (bp
[0] != 'f' && bp
[0] != 'l')
4935 (void)LOOKING_AT (bp
, "local"); /* skip possible "local" */
4937 if (LOOKING_AT (bp
, "function"))
4945 * Just look for lines where the first character is '/'
4946 * Also look at "defineps" for PSWrap
4948 * Richard Mlynarik <mly@adoc.xerox.com> (1997)
4949 * Masatake Yamato <masata-y@is.aist-nara.ac.jp> (1999)
4952 PS_functions (FILE *inf
)
4954 register char *bp
, *ep
;
4956 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4961 *ep
!= '\0' && *ep
!= ' ' && *ep
!= '{';
4964 make_tag (bp
, ep
- bp
, true,
4965 lb
.buffer
, ep
- lb
.buffer
+ 1, lineno
, linecharno
);
4967 else if (LOOKING_AT (bp
, "defineps"))
4975 * Ignore anything after \ followed by space or in ( )
4976 * Look for words defined by :
4977 * Look for constant, code, create, defer, value, and variable
4978 * OBP extensions: Look for buffer:, field,
4979 * Ideas by Eduardo Horvath <eeh@netbsd.org> (2004)
4982 Forth_words (FILE *inf
)
4986 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4987 while ((bp
= skip_spaces (bp
))[0] != '\0')
4988 if (bp
[0] == '\\' && c_isspace (bp
[1]))
4989 break; /* read next line */
4990 else if (bp
[0] == '(' && c_isspace (bp
[1]))
4991 do /* skip to ) or eol */
4993 while (*bp
!= ')' && *bp
!= '\0');
4994 else if ((bp
[0] == ':' && c_isspace (bp
[1]) && bp
++)
4995 || LOOKING_AT_NOCASE (bp
, "constant")
4996 || LOOKING_AT_NOCASE (bp
, "code")
4997 || LOOKING_AT_NOCASE (bp
, "create")
4998 || LOOKING_AT_NOCASE (bp
, "defer")
4999 || LOOKING_AT_NOCASE (bp
, "value")
5000 || LOOKING_AT_NOCASE (bp
, "variable")
5001 || LOOKING_AT_NOCASE (bp
, "buffer:")
5002 || LOOKING_AT_NOCASE (bp
, "field"))
5003 get_tag (skip_spaces (bp
), NULL
); /* Yay! A definition! */
5005 bp
= skip_non_spaces (bp
);
5010 * Scheme tag functions
5011 * look for (def... xyzzy
5013 * (def ... ((...(xyzzy ....
5015 * Original code by Ken Haase (1985?)
5018 Scheme_functions (FILE *inf
)
5022 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
5024 if (strneq (bp
, "(def", 4) || strneq (bp
, "(DEF", 4))
5026 bp
= skip_non_spaces (bp
+4);
5027 /* Skip over open parens and white space. Don't continue past
5029 while (*bp
&& notinname (*bp
))
5033 if (LOOKING_AT (bp
, "(SET!") || LOOKING_AT (bp
, "(set!"))
5039 /* Find tags in TeX and LaTeX input files. */
5041 /* TEX_toktab is a table of TeX control sequences that define tags.
5042 * Each entry records one such control sequence.
5044 * Original code from who knows whom.
5046 * Stefan Monnier (2002)
5049 static linebuffer
*TEX_toktab
= NULL
; /* Table with tag tokens */
5051 /* Default set of control sequences to put into TEX_toktab.
5052 The value of environment var TEXTAGS is prepended to this. */
5053 static const char *TEX_defenv
= "\
5054 :chapter:section:subsection:subsubsection:eqno:label:ref:cite:bibitem\
5055 :part:appendix:entry:index:def\
5056 :newcommand:renewcommand:newenvironment:renewenvironment";
5058 static void TEX_decode_env (const char *, const char *);
5061 * TeX/LaTeX scanning loop.
5064 TeX_commands (FILE *inf
)
5069 char TEX_esc
= '\0';
5070 char TEX_opgrp
, TEX_clgrp
;
5072 /* Initialize token table once from environment. */
5073 if (TEX_toktab
== NULL
)
5074 TEX_decode_env ("TEXTAGS", TEX_defenv
);
5076 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5078 /* Look at each TEX keyword in line. */
5081 /* Look for a TEX escape. */
5085 if (c
== '\0' || c
== '%')
5088 /* Select either \ or ! as escape character, whichever comes
5089 first outside a comment. */
5110 for (key
= TEX_toktab
; key
->buffer
!= NULL
; key
++)
5111 if (strneq (cp
, key
->buffer
, key
->len
))
5114 int namelen
, linelen
;
5117 cp
= skip_spaces (cp
+ key
->len
);
5118 if (*cp
== TEX_opgrp
)
5124 (!c_isspace (*p
) && *p
!= '#' &&
5125 *p
!= TEX_opgrp
&& *p
!= TEX_clgrp
);
5130 if (!opgrp
|| *p
== TEX_clgrp
)
5132 while (*p
!= '\0' && *p
!= TEX_opgrp
&& *p
!= TEX_clgrp
)
5134 linelen
= p
- lb
.buffer
+ 1;
5136 make_tag (cp
, namelen
, true,
5137 lb
.buffer
, linelen
, lineno
, linecharno
);
5138 goto tex_next_line
; /* We only tag a line once */
5146 /* Read environment and prepend it to the default string.
5147 Build token table. */
5149 TEX_decode_env (const char *evarname
, const char *defenv
)
5151 register const char *env
, *p
;
5154 /* Append default string to environment. */
5155 env
= getenv (evarname
);
5159 env
= concat (env
, defenv
, "");
5161 /* Allocate a token table */
5162 for (len
= 1, p
= env
; (p
= strchr (p
, ':')); )
5165 TEX_toktab
= xnew (len
, linebuffer
);
5167 /* Unpack environment string into token table. Be careful about */
5168 /* zero-length strings (leading ':', "::" and trailing ':') */
5169 for (i
= 0; *env
!= '\0';)
5171 p
= strchr (env
, ':');
5172 if (!p
) /* End of environment string. */
5173 p
= env
+ strlen (env
);
5175 { /* Only non-zero strings. */
5176 TEX_toktab
[i
].buffer
= savenstr (env
, p
- env
);
5177 TEX_toktab
[i
].len
= p
- env
;
5184 TEX_toktab
[i
].buffer
= NULL
; /* Mark end of table. */
5185 TEX_toktab
[i
].len
= 0;
5192 /* Texinfo support. Dave Love, Mar. 2000. */
5194 Texinfo_nodes (FILE *inf
)
5197 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5198 if (LOOKING_AT (cp
, "@node"))
5201 while (*cp
!= '\0' && *cp
!= ',')
5203 make_tag (start
, cp
- start
, true,
5204 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
5211 * Contents of <title>, <h1>, <h2>, <h3> are tags.
5212 * Contents of <a name=xxx> are tags with name xxx.
5214 * Francesco Potortì, 2002.
5217 HTML_labels (FILE *inf
)
5219 bool getnext
= false; /* next text outside of HTML tags is a tag */
5220 bool skiptag
= false; /* skip to the end of the current HTML tag */
5221 bool intag
= false; /* inside an html tag, looking for ID= */
5222 bool inanchor
= false; /* when INTAG, is an anchor, look for NAME= */
5226 linebuffer_setlen (&token_name
, 0); /* no name in buffer */
5228 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
5229 for (;;) /* loop on the same line */
5231 if (skiptag
) /* skip HTML tag */
5233 while (*dbp
!= '\0' && *dbp
!= '>')
5239 continue; /* look on the same line */
5241 break; /* go to next line */
5244 else if (intag
) /* look for "name=" or "id=" */
5246 while (*dbp
!= '\0' && *dbp
!= '>'
5247 && c_tolower (*dbp
) != 'n' && c_tolower (*dbp
) != 'i')
5250 break; /* go to next line */
5255 continue; /* look on the same line */
5257 if ((inanchor
&& LOOKING_AT_NOCASE (dbp
, "name="))
5258 || LOOKING_AT_NOCASE (dbp
, "id="))
5260 bool quoted
= (dbp
[0] == '"');
5263 for (end
= ++dbp
; *end
!= '\0' && *end
!= '"'; end
++)
5266 for (end
= dbp
; *end
!= '\0' && intoken (*end
); end
++)
5268 linebuffer_setlen (&token_name
, end
- dbp
);
5269 memcpy (token_name
.buffer
, dbp
, end
- dbp
);
5270 token_name
.buffer
[end
- dbp
] = '\0';
5273 intag
= false; /* we found what we looked for */
5274 skiptag
= true; /* skip to the end of the tag */
5275 getnext
= true; /* then grab the text */
5276 continue; /* look on the same line */
5281 else if (getnext
) /* grab next tokens and tag them */
5283 dbp
= skip_spaces (dbp
);
5285 break; /* go to next line */
5289 inanchor
= (c_tolower (dbp
[1]) == 'a' && !intoken (dbp
[2]));
5290 continue; /* look on the same line */
5293 for (end
= dbp
+ 1; *end
!= '\0' && *end
!= '<'; end
++)
5295 make_tag (token_name
.buffer
, token_name
.len
, true,
5296 dbp
, end
- dbp
, lineno
, linecharno
);
5297 linebuffer_setlen (&token_name
, 0); /* no name in buffer */
5299 break; /* go to next line */
5302 else /* look for an interesting HTML tag */
5304 while (*dbp
!= '\0' && *dbp
!= '<')
5307 break; /* go to next line */
5309 if (c_tolower (dbp
[1]) == 'a' && !intoken (dbp
[2]))
5312 continue; /* look on the same line */
5314 else if (LOOKING_AT_NOCASE (dbp
, "<title>")
5315 || LOOKING_AT_NOCASE (dbp
, "<h1>")
5316 || LOOKING_AT_NOCASE (dbp
, "<h2>")
5317 || LOOKING_AT_NOCASE (dbp
, "<h3>"))
5321 continue; /* look on the same line */
5332 * Assumes that the predicate or rule starts at column 0.
5333 * Only the first clause of a predicate or rule is added.
5334 * Original code by Sunichirou Sugou (1989)
5335 * Rewritten by Anders Lindgren (1996)
5337 static size_t prolog_pr (char *, char *);
5338 static void prolog_skip_comment (linebuffer
*, FILE *);
5339 static size_t prolog_atom (char *, size_t);
5342 Prolog_functions (FILE *inf
)
5352 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5354 if (cp
[0] == '\0') /* Empty line */
5356 else if (c_isspace (cp
[0])) /* Not a predicate */
5358 else if (cp
[0] == '/' && cp
[1] == '*') /* comment. */
5359 prolog_skip_comment (&lb
, inf
);
5360 else if ((len
= prolog_pr (cp
, last
)) > 0)
5362 /* Predicate or rule. Store the function name so that we
5363 only generate a tag for the first clause. */
5365 last
= xnew (len
+ 1, char);
5366 else if (len
+ 1 > allocated
)
5367 xrnew (last
, len
+ 1, char);
5368 allocated
= len
+ 1;
5369 memcpy (last
, cp
, len
);
5378 prolog_skip_comment (linebuffer
*plb
, FILE *inf
)
5384 for (cp
= plb
->buffer
; *cp
!= '\0'; cp
++)
5385 if (cp
[0] == '*' && cp
[1] == '/')
5387 readline (plb
, inf
);
5389 while (perhaps_more_input (inf
));
5393 * A predicate or rule definition is added if it matches:
5394 * <beginning of line><Prolog Atom><whitespace>(
5395 * or <beginning of line><Prolog Atom><whitespace>:-
5397 * It is added to the tags database if it doesn't match the
5398 * name of the previous clause header.
5400 * Return the size of the name of the predicate or rule, or 0 if no
5404 prolog_pr (char *s
, char *last
)
5406 /* Name of last clause. */
5411 pos
= prolog_atom (s
, 0);
5416 pos
= skip_spaces (s
+ pos
) - s
;
5419 || (s
[pos
] == '(' && (pos
+= 1))
5420 || (s
[pos
] == ':' && s
[pos
+ 1] == '-' && (pos
+= 2)))
5421 && (last
== NULL
/* save only the first clause */
5422 || len
!= strlen (last
)
5423 || !strneq (s
, last
, len
)))
5425 make_tag (s
, len
, true, s
, pos
, lineno
, linecharno
);
5433 * Consume a Prolog atom.
5434 * Return the number of bytes consumed, or 0 if there was an error.
5436 * A prolog atom, in this context, could be one of:
5437 * - An alphanumeric sequence, starting with a lower case letter.
5438 * - A quoted arbitrary string. Single quotes can escape themselves.
5439 * Backslash quotes everything.
5442 prolog_atom (char *s
, size_t pos
)
5448 if (c_islower (s
[pos
]) || s
[pos
] == '_')
5450 /* The atom is unquoted. */
5452 while (c_isalnum (s
[pos
]) || s
[pos
] == '_')
5456 return pos
- origpos
;
5458 else if (s
[pos
] == '\'')
5469 pos
++; /* A double quote */
5471 else if (s
[pos
] == '\0')
5472 /* Multiline quoted atoms are ignored. */
5474 else if (s
[pos
] == '\\')
5476 if (s
[pos
+1] == '\0')
5483 return pos
- origpos
;
5491 * Support for Erlang
5493 * Generates tags for functions, defines, and records.
5494 * Assumes that Erlang functions start at column 0.
5495 * Original code by Anders Lindgren (1996)
5497 static int erlang_func (char *, char *);
5498 static void erlang_attribute (char *);
5499 static int erlang_atom (char *);
5502 Erlang_functions (FILE *inf
)
5512 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5514 if (cp
[0] == '\0') /* Empty line */
5516 else if (c_isspace (cp
[0])) /* Not function nor attribute */
5518 else if (cp
[0] == '%') /* comment */
5520 else if (cp
[0] == '"') /* Sometimes, strings start in column one */
5522 else if (cp
[0] == '-') /* attribute, e.g. "-define" */
5524 erlang_attribute (cp
);
5531 else if ((len
= erlang_func (cp
, last
)) > 0)
5534 * Function. Store the function name so that we only
5535 * generates a tag for the first clause.
5538 last
= xnew (len
+ 1, char);
5539 else if (len
+ 1 > allocated
)
5540 xrnew (last
, len
+ 1, char);
5541 allocated
= len
+ 1;
5542 memcpy (last
, cp
, len
);
5551 * A function definition is added if it matches:
5552 * <beginning of line><Erlang Atom><whitespace>(
5554 * It is added to the tags database if it doesn't match the
5555 * name of the previous clause header.
5557 * Return the size of the name of the function, or 0 if no function
5561 erlang_func (char *s
, char *last
)
5563 /* Name of last clause. */
5568 pos
= erlang_atom (s
);
5573 pos
= skip_spaces (s
+ pos
) - s
;
5575 /* Save only the first clause. */
5578 || len
!= (int)strlen (last
)
5579 || !strneq (s
, last
, len
)))
5581 make_tag (s
, len
, true, s
, pos
, lineno
, linecharno
);
5590 * Handle attributes. Currently, tags are generated for defines
5593 * They are on the form:
5594 * -define(foo, bar).
5595 * -define(Foo(M, N), M+N).
5596 * -record(graph, {vtab = notable, cyclic = true}).
5599 erlang_attribute (char *s
)
5603 if ((LOOKING_AT (cp
, "-define") || LOOKING_AT (cp
, "-record"))
5606 int len
= erlang_atom (skip_spaces (cp
));
5608 make_tag (cp
, len
, true, s
, cp
+ len
- s
, lineno
, linecharno
);
5615 * Consume an Erlang atom (or variable).
5616 * Return the number of bytes consumed, or -1 if there was an error.
5619 erlang_atom (char *s
)
5623 if (c_isalpha (s
[pos
]) || s
[pos
] == '_')
5625 /* The atom is unquoted. */
5628 while (c_isalnum (s
[pos
]) || s
[pos
] == '_');
5630 else if (s
[pos
] == '\'')
5632 for (pos
++; s
[pos
] != '\''; pos
++)
5633 if (s
[pos
] == '\0' /* multiline quoted atoms are ignored */
5634 || (s
[pos
] == '\\' && s
[++pos
] == '\0'))
5643 static char *scan_separators (char *);
5644 static void add_regex (char *, language
*);
5645 static char *substitute (char *, char *, struct re_registers
*);
5648 * Take a string like "/blah/" and turn it into "blah", verifying
5649 * that the first and last characters are the same, and handling
5650 * quoted separator characters. Actually, stops on the occurrence of
5651 * an unquoted separator. Also process \t, \n, etc. and turn into
5652 * appropriate characters. Works in place. Null terminates name string.
5653 * Returns pointer to terminating separator, or NULL for
5654 * unterminated regexps.
5657 scan_separators (char *name
)
5660 char *copyto
= name
;
5661 bool quoted
= false;
5663 for (++name
; *name
!= '\0'; ++name
)
5669 case 'a': *copyto
++ = '\007'; break; /* BEL (bell) */
5670 case 'b': *copyto
++ = '\b'; break; /* BS (back space) */
5671 case 'd': *copyto
++ = 0177; break; /* DEL (delete) */
5672 case 'e': *copyto
++ = 033; break; /* ESC (delete) */
5673 case 'f': *copyto
++ = '\f'; break; /* FF (form feed) */
5674 case 'n': *copyto
++ = '\n'; break; /* NL (new line) */
5675 case 'r': *copyto
++ = '\r'; break; /* CR (carriage return) */
5676 case 't': *copyto
++ = '\t'; break; /* TAB (horizontal tab) */
5677 case 'v': *copyto
++ = '\v'; break; /* VT (vertical tab) */
5683 /* Something else is quoted, so preserve the quote. */
5691 else if (*name
== '\\')
5693 else if (*name
== sep
)
5699 name
= NULL
; /* signal unterminated regexp */
5701 /* Terminate copied string. */
5706 /* Look at the argument of --regex or --no-regex and do the right
5707 thing. Same for each line of a regexp file. */
5709 analyze_regex (char *regex_arg
)
5711 if (regex_arg
== NULL
)
5713 free_regexps (); /* --no-regex: remove existing regexps */
5717 /* A real --regexp option or a line in a regexp file. */
5718 switch (regex_arg
[0])
5720 /* Comments in regexp file or null arg to --regex. */
5726 /* Read a regex file. This is recursive and may result in a
5727 loop, which will stop when the file descriptors are exhausted. */
5731 linebuffer regexbuf
;
5732 char *regexfile
= regex_arg
+ 1;
5734 /* regexfile is a file containing regexps, one per line. */
5735 regexfp
= fopen (regexfile
, "r" FOPEN_BINARY
);
5736 if (regexfp
== NULL
)
5738 linebuffer_init (®exbuf
);
5739 while (readline_internal (®exbuf
, regexfp
, regexfile
) > 0)
5740 analyze_regex (regexbuf
.buffer
);
5741 free (regexbuf
.buffer
);
5742 if (fclose (regexfp
) != 0)
5747 /* Regexp to be used for a specific language only. */
5751 char *lang_name
= regex_arg
+ 1;
5754 for (cp
= lang_name
; *cp
!= '}'; cp
++)
5757 error ("unterminated language name in regex: %s", regex_arg
);
5761 lang
= get_language_from_langname (lang_name
);
5764 add_regex (cp
, lang
);
5768 /* Regexp to be used for any language. */
5770 add_regex (regex_arg
, NULL
);
5775 /* Separate the regexp pattern, compile it,
5776 and care for optional name and modifiers. */
5778 add_regex (char *regexp_pattern
, language
*lang
)
5780 static struct re_pattern_buffer zeropattern
;
5781 char sep
, *pat
, *name
, *modifiers
;
5784 struct re_pattern_buffer
*patbuf
;
5787 force_explicit_name
= true, /* do not use implicit tag names */
5788 ignore_case
= false, /* case is significant */
5789 multi_line
= false, /* matches are done one line at a time */
5790 single_line
= false; /* dot does not match newline */
5793 if (strlen (regexp_pattern
) < 3)
5795 error ("null regexp");
5798 sep
= regexp_pattern
[0];
5799 name
= scan_separators (regexp_pattern
);
5802 error ("%s: unterminated regexp", regexp_pattern
);
5807 error ("null name for regexp \"%s\"", regexp_pattern
);
5810 modifiers
= scan_separators (name
);
5811 if (modifiers
== NULL
) /* no terminating separator --> no name */
5817 modifiers
+= 1; /* skip separator */
5819 /* Parse regex modifiers. */
5820 for (; modifiers
[0] != '\0'; modifiers
++)
5821 switch (modifiers
[0])
5824 if (modifiers
== name
)
5825 error ("forcing explicit tag name but no name, ignoring");
5826 force_explicit_name
= true;
5836 need_filebuf
= true;
5839 error ("invalid regexp modifier '%c', ignoring", modifiers
[0]);
5843 patbuf
= xnew (1, struct re_pattern_buffer
);
5844 *patbuf
= zeropattern
;
5847 static char lc_trans
[UCHAR_MAX
+ 1];
5849 for (i
= 0; i
< UCHAR_MAX
+ 1; i
++)
5850 lc_trans
[i
] = c_tolower (i
);
5851 patbuf
->translate
= lc_trans
; /* translation table to fold case */
5855 pat
= concat ("^", regexp_pattern
, ""); /* anchor to beginning of line */
5857 pat
= regexp_pattern
;
5860 re_set_syntax (RE_SYNTAX_EMACS
| RE_DOT_NEWLINE
);
5862 re_set_syntax (RE_SYNTAX_EMACS
);
5864 err
= re_compile_pattern (pat
, strlen (pat
), patbuf
);
5869 error ("%s while compiling pattern", err
);
5874 p_head
= xnew (1, regexp
);
5875 p_head
->pattern
= savestr (regexp_pattern
);
5876 p_head
->p_next
= rp
;
5877 p_head
->lang
= lang
;
5878 p_head
->pat
= patbuf
;
5879 p_head
->name
= savestr (name
);
5880 p_head
->error_signaled
= false;
5881 p_head
->force_explicit_name
= force_explicit_name
;
5882 p_head
->ignore_case
= ignore_case
;
5883 p_head
->multi_line
= multi_line
;
5887 * Do the substitutions indicated by the regular expression and
5891 substitute (char *in
, char *out
, struct re_registers
*regs
)
5894 int size
, dig
, diglen
;
5897 size
= strlen (out
);
5899 /* Pass 1: figure out how much to allocate by finding all \N strings. */
5900 if (out
[size
- 1] == '\\')
5901 fatal ("pattern error in \"%s\"", out
);
5902 for (t
= strchr (out
, '\\');
5904 t
= strchr (t
+ 2, '\\'))
5905 if (c_isdigit (t
[1]))
5908 diglen
= regs
->end
[dig
] - regs
->start
[dig
];
5914 /* Allocate space and do the substitutions. */
5916 result
= xnew (size
+ 1, char);
5918 for (t
= result
; *out
!= '\0'; out
++)
5919 if (*out
== '\\' && c_isdigit (*++out
))
5922 diglen
= regs
->end
[dig
] - regs
->start
[dig
];
5923 memcpy (t
, in
+ regs
->start
[dig
], diglen
);
5930 assert (t
<= result
+ size
);
5931 assert (t
- result
== (int)strlen (result
));
5936 /* Deallocate all regexps. */
5941 while (p_head
!= NULL
)
5943 rp
= p_head
->p_next
;
5944 free (p_head
->pattern
);
5945 free (p_head
->name
);
5953 * Reads the whole file as a single string from `filebuf' and looks for
5954 * multi-line regular expressions, creating tags on matches.
5955 * readline already dealt with normal regexps.
5957 * Idea by Ben Wing <ben@666.com> (2002).
5960 regex_tag_multiline (void)
5962 char *buffer
= filebuf
.buffer
;
5966 for (rp
= p_head
; rp
!= NULL
; rp
= rp
->p_next
)
5970 if (!rp
->multi_line
)
5971 continue; /* skip normal regexps */
5973 /* Generic initializations before parsing file from memory. */
5974 lineno
= 1; /* reset global line number */
5975 charno
= 0; /* reset global char number */
5976 linecharno
= 0; /* reset global char number of line start */
5978 /* Only use generic regexps or those for the current language. */
5979 if (rp
->lang
!= NULL
&& rp
->lang
!= curfdp
->lang
)
5982 while (match
>= 0 && match
< filebuf
.len
)
5984 match
= re_search (rp
->pat
, buffer
, filebuf
.len
, charno
,
5985 filebuf
.len
- match
, &rp
->regs
);
5990 if (!rp
->error_signaled
)
5992 error ("regexp stack overflow while matching \"%s\"",
5994 rp
->error_signaled
= true;
6001 if (match
== rp
->regs
.end
[0])
6003 if (!rp
->error_signaled
)
6005 error ("regexp matches the empty string: \"%s\"",
6007 rp
->error_signaled
= true;
6009 match
= -3; /* exit from while loop */
6013 /* Match occurred. Construct a tag. */
6014 while (charno
< rp
->regs
.end
[0])
6015 if (buffer
[charno
++] == '\n')
6016 lineno
++, linecharno
= charno
;
6018 if (name
[0] == '\0')
6020 else /* make a named tag */
6021 name
= substitute (buffer
, rp
->name
, &rp
->regs
);
6022 if (rp
->force_explicit_name
)
6023 /* Force explicit tag name, if a name is there. */
6024 pfnote (name
, true, buffer
+ linecharno
,
6025 charno
- linecharno
+ 1, lineno
, linecharno
);
6027 make_tag (name
, strlen (name
), true, buffer
+ linecharno
,
6028 charno
- linecharno
+ 1, lineno
, linecharno
);
6037 nocase_tail (const char *cp
)
6041 while (*cp
!= '\0' && c_tolower (*cp
) == c_tolower (dbp
[len
]))
6043 if (*cp
== '\0' && !intoken (dbp
[len
]))
6052 get_tag (register char *bp
, char **namepp
)
6054 register char *cp
= bp
;
6058 /* Go till you get to white space or a syntactic break */
6059 for (cp
= bp
+ 1; !notinname (*cp
); cp
++)
6061 make_tag (bp
, cp
- bp
, true,
6062 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
6066 *namepp
= savenstr (bp
, cp
- bp
);
6070 * Read a line of text from `stream' into `lbp', excluding the
6071 * newline or CR-NL, if any. Return the number of characters read from
6072 * `stream', which is the length of the line including the newline.
6074 * On DOS or Windows we do not count the CR character, if any before the
6075 * NL, in the returned length; this mirrors the behavior of Emacs on those
6076 * platforms (for text files, it translates CR-NL to NL as it reads in the
6079 * If multi-line regular expressions are requested, each line read is
6080 * appended to `filebuf'.
6083 readline_internal (linebuffer
*lbp
, FILE *stream
, char const *filename
)
6085 char *buffer
= lbp
->buffer
;
6086 char *p
= lbp
->buffer
;
6090 pend
= p
+ lbp
->size
; /* Separate to avoid 386/IX compiler bug. */
6094 register int c
= getc (stream
);
6097 /* We're at the end of linebuffer: expand it. */
6099 xrnew (buffer
, lbp
->size
, char);
6100 p
+= buffer
- lbp
->buffer
;
6101 pend
= buffer
+ lbp
->size
;
6102 lbp
->buffer
= buffer
;
6106 if (ferror (stream
))
6114 if (p
> buffer
&& p
[-1] == '\r')
6128 lbp
->len
= p
- buffer
;
6130 if (need_filebuf
/* we need filebuf for multi-line regexps */
6131 && chars_deleted
> 0) /* not at EOF */
6133 while (filebuf
.size
<= filebuf
.len
+ lbp
->len
+ 1) /* +1 for \n */
6135 /* Expand filebuf. */
6137 xrnew (filebuf
.buffer
, filebuf
.size
, char);
6139 memcpy (filebuf
.buffer
+ filebuf
.len
, lbp
->buffer
, lbp
->len
);
6140 filebuf
.len
+= lbp
->len
;
6141 filebuf
.buffer
[filebuf
.len
++] = '\n';
6142 filebuf
.buffer
[filebuf
.len
] = '\0';
6145 return lbp
->len
+ chars_deleted
;
6149 * Like readline_internal, above, but in addition try to match the
6150 * input line against relevant regular expressions and manage #line
6154 readline (linebuffer
*lbp
, FILE *stream
)
6158 linecharno
= charno
; /* update global char number of line start */
6159 result
= readline_internal (lbp
, stream
, infilename
); /* read line */
6160 lineno
+= 1; /* increment global line number */
6161 charno
+= result
; /* increment global char number */
6163 /* Honor #line directives. */
6164 if (!no_line_directive
)
6166 static bool discard_until_line_directive
;
6168 /* Check whether this is a #line directive. */
6169 if (result
> 12 && strneq (lbp
->buffer
, "#line ", 6))
6174 if (sscanf (lbp
->buffer
, "#line %u \"%n", &lno
, &start
) >= 1
6175 && start
> 0) /* double quote character found */
6177 char *endp
= lbp
->buffer
+ start
;
6179 while ((endp
= strchr (endp
, '"')) != NULL
6180 && endp
[-1] == '\\')
6183 /* Ok, this is a real #line directive. Let's deal with it. */
6185 char *taggedabsname
; /* absolute name of original file */
6186 char *taggedfname
; /* name of original file as given */
6187 char *name
; /* temp var */
6189 discard_until_line_directive
= false; /* found it */
6190 name
= lbp
->buffer
+ start
;
6192 canonicalize_filename (name
);
6193 taggedabsname
= absolute_filename (name
, tagfiledir
);
6194 if (filename_is_absolute (name
)
6195 || filename_is_absolute (curfdp
->infname
))
6196 taggedfname
= savestr (taggedabsname
);
6198 taggedfname
= relative_filename (taggedabsname
,tagfiledir
);
6200 if (streq (curfdp
->taggedfname
, taggedfname
))
6201 /* The #line directive is only a line number change. We
6202 deal with this afterwards. */
6205 /* The tags following this #line directive should be
6206 attributed to taggedfname. In order to do this, set
6207 curfdp accordingly. */
6209 fdesc
*fdp
; /* file description pointer */
6211 /* Go look for a file description already set up for the
6212 file indicated in the #line directive. If there is
6213 one, use it from now until the next #line
6215 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
6216 if (streq (fdp
->infname
, curfdp
->infname
)
6217 && streq (fdp
->taggedfname
, taggedfname
))
6218 /* If we remove the second test above (after the &&)
6219 then all entries pertaining to the same file are
6220 coalesced in the tags file. If we use it, then
6221 entries pertaining to the same file but generated
6222 from different files (via #line directives) will
6223 go into separate sections in the tags file. These
6224 alternatives look equivalent. The first one
6225 destroys some apparently useless information. */
6231 /* Else, if we already tagged the real file, skip all
6232 input lines until the next #line directive. */
6233 if (fdp
== NULL
) /* not found */
6234 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
6235 if (streq (fdp
->infabsname
, taggedabsname
))
6237 discard_until_line_directive
= true;
6241 /* Else create a new file description and use that from
6242 now on, until the next #line directive. */
6243 if (fdp
== NULL
) /* not found */
6246 fdhead
= xnew (1, fdesc
);
6247 *fdhead
= *curfdp
; /* copy curr. file description */
6249 fdhead
->infname
= savestr (curfdp
->infname
);
6250 fdhead
->infabsname
= savestr (curfdp
->infabsname
);
6251 fdhead
->infabsdir
= savestr (curfdp
->infabsdir
);
6252 fdhead
->taggedfname
= taggedfname
;
6253 fdhead
->usecharno
= false;
6254 fdhead
->prop
= NULL
;
6255 fdhead
->written
= false;
6259 free (taggedabsname
);
6261 readline (lbp
, stream
);
6263 } /* if a real #line directive */
6264 } /* if #line is followed by a number */
6265 } /* if line begins with "#line " */
6267 /* If we are here, no #line directive was found. */
6268 if (discard_until_line_directive
)
6272 /* Do a tail recursion on ourselves, thus discarding the contents
6273 of the line buffer. */
6274 readline (lbp
, stream
);
6278 discard_until_line_directive
= false;
6281 } /* if #line directives should be considered */
6288 /* Match against relevant regexps. */
6290 for (rp
= p_head
; rp
!= NULL
; rp
= rp
->p_next
)
6292 /* Only use generic regexps or those for the current language.
6293 Also do not use multiline regexps, which is the job of
6294 regex_tag_multiline. */
6295 if ((rp
->lang
!= NULL
&& rp
->lang
!= fdhead
->lang
)
6299 match
= re_match (rp
->pat
, lbp
->buffer
, lbp
->len
, 0, &rp
->regs
);
6304 if (!rp
->error_signaled
)
6306 error ("regexp stack overflow while matching \"%s\"",
6308 rp
->error_signaled
= true;
6315 /* Empty string matched. */
6316 if (!rp
->error_signaled
)
6318 error ("regexp matches the empty string: \"%s\"", rp
->pattern
);
6319 rp
->error_signaled
= true;
6323 /* Match occurred. Construct a tag. */
6325 if (name
[0] == '\0')
6327 else /* make a named tag */
6328 name
= substitute (lbp
->buffer
, rp
->name
, &rp
->regs
);
6329 if (rp
->force_explicit_name
)
6330 /* Force explicit tag name, if a name is there. */
6331 pfnote (name
, true, lbp
->buffer
, match
, lineno
, linecharno
);
6333 make_tag (name
, strlen (name
), true,
6334 lbp
->buffer
, match
, lineno
, linecharno
);
6343 * Return a pointer to a space of size strlen(cp)+1 allocated
6344 * with xnew where the string CP has been copied.
6347 savestr (const char *cp
)
6349 return savenstr (cp
, strlen (cp
));
6353 * Return a pointer to a space of size LEN+1 allocated with xnew where
6354 * the string CP has been copied for at most the first LEN characters.
6357 savenstr (const char *cp
, int len
)
6359 char *dp
= xnew (len
+ 1, char);
6361 return memcpy (dp
, cp
, len
);
6364 /* Skip spaces (end of string is not space), return new pointer. */
6366 skip_spaces (char *cp
)
6368 while (c_isspace (*cp
))
6373 /* Skip non spaces, except end of string, return new pointer. */
6375 skip_non_spaces (char *cp
)
6377 while (*cp
!= '\0' && !c_isspace (*cp
))
6382 /* Skip any chars in the "name" class.*/
6384 skip_name (char *cp
)
6386 /* '\0' is a notinname() so loop stops there too */
6387 while (! notinname (*cp
))
6392 /* Print error message and exit. */
6394 fatal (char const *format
, ...)
6397 va_start (ap
, format
);
6398 verror (format
, ap
);
6400 exit (EXIT_FAILURE
);
6404 pfatal (const char *s1
)
6407 exit (EXIT_FAILURE
);
6411 suggest_asking_for_help (void)
6413 fprintf (stderr
, "\tTry '%s --help' for a complete list of options.\n",
6415 exit (EXIT_FAILURE
);
6418 /* Output a diagnostic with printf-style FORMAT and args. */
6420 error (const char *format
, ...)
6423 va_start (ap
, format
);
6424 verror (format
, ap
);
6429 verror (char const *format
, va_list ap
)
6431 fprintf (stderr
, "%s: ", progname
);
6432 vfprintf (stderr
, format
, ap
);
6433 fprintf (stderr
, "\n");
6436 /* Return a newly-allocated string whose contents
6437 concatenate those of s1, s2, s3. */
6439 concat (const char *s1
, const char *s2
, const char *s3
)
6441 int len1
= strlen (s1
), len2
= strlen (s2
), len3
= strlen (s3
);
6442 char *result
= xnew (len1
+ len2
+ len3
+ 1, char);
6444 strcpy (result
, s1
);
6445 strcpy (result
+ len1
, s2
);
6446 strcpy (result
+ len1
+ len2
, s3
);
6452 /* Does the same work as the system V getcwd, but does not need to
6453 guess the buffer size in advance. */
6458 char *path
= xnew (bufsize
, char);
6460 while (getcwd (path
, bufsize
) == NULL
)
6462 if (errno
!= ERANGE
)
6466 path
= xnew (bufsize
, char);
6469 canonicalize_filename (path
);
6473 /* Return a newly allocated string containing a name of a temporary file. */
6477 const char *tmpdir
= getenv ("TMPDIR");
6478 const char *slash
= "/";
6480 #if MSDOS || defined (DOS_NT)
6482 tmpdir
= getenv ("TEMP");
6484 tmpdir
= getenv ("TMP");
6487 if (tmpdir
[strlen (tmpdir
) - 1] == '/'
6488 || tmpdir
[strlen (tmpdir
) - 1] == '\\')
6493 if (tmpdir
[strlen (tmpdir
) - 1] == '/')
6497 char *templt
= concat (tmpdir
, slash
, "etXXXXXX");
6498 int fd
= mkostemp (templt
, O_CLOEXEC
);
6499 if (fd
< 0 || close (fd
) != 0)
6501 int temp_errno
= errno
;
6507 #if defined (DOS_NT)
6508 /* The file name will be used in shell redirection, so it needs to have
6509 DOS-style backslashes, or else the Windows shell will barf. */
6511 for (p
= templt
; *p
; p
++)
6519 /* Return a newly allocated string containing the file name of FILE
6520 relative to the absolute directory DIR (which should end with a slash). */
6522 relative_filename (char *file
, char *dir
)
6524 char *fp
, *dp
, *afn
, *res
;
6527 /* Find the common root of file and dir (with a trailing slash). */
6528 afn
= absolute_filename (file
, cwd
);
6531 while (*fp
++ == *dp
++)
6533 fp
--, dp
--; /* back to the first differing char */
6535 if (fp
== afn
&& afn
[0] != '/') /* cannot build a relative name */
6538 do /* look at the equal chars until '/' */
6542 /* Build a sequence of "../" strings for the resulting relative file name. */
6544 while ((dp
= strchr (dp
+ 1, '/')) != NULL
)
6546 res
= xnew (3*i
+ strlen (fp
+ 1) + 1, char);
6549 z
= stpcpy (z
, "../");
6551 /* Add the file name relative to the common root of file and dir. */
6558 /* Return a newly allocated string containing the absolute file name
6559 of FILE given DIR (which should end with a slash). */
6561 absolute_filename (char *file
, char *dir
)
6563 char *slashp
, *cp
, *res
;
6565 if (filename_is_absolute (file
))
6566 res
= savestr (file
);
6568 /* We don't support non-absolute file names with a drive
6569 letter, like `d:NAME' (it's too much hassle). */
6570 else if (file
[1] == ':')
6571 fatal ("%s: relative file names with drive letters not supported", file
);
6574 res
= concat (dir
, file
, "");
6576 /* Delete the "/dirname/.." and "/." substrings. */
6577 slashp
= strchr (res
, '/');
6578 while (slashp
!= NULL
&& slashp
[0] != '\0')
6580 if (slashp
[1] == '.')
6582 if (slashp
[2] == '.'
6583 && (slashp
[3] == '/' || slashp
[3] == '\0'))
6588 while (cp
>= res
&& !filename_is_absolute (cp
));
6590 cp
= slashp
; /* the absolute name begins with "/.." */
6592 /* Under MSDOS and NT we get `d:/NAME' as absolute
6593 file name, so the luser could say `d:/../NAME'.
6594 We silently treat this as `d:/NAME'. */
6595 else if (cp
[0] != '/')
6598 memmove (cp
, slashp
+ 3, strlen (slashp
+ 2));
6602 else if (slashp
[2] == '/' || slashp
[2] == '\0')
6604 memmove (slashp
, slashp
+ 2, strlen (slashp
+ 1));
6609 slashp
= strchr (slashp
+ 1, '/');
6612 if (res
[0] == '\0') /* just a safety net: should never happen */
6615 return savestr ("/");
6621 /* Return a newly allocated string containing the absolute
6622 file name of dir where FILE resides given DIR (which should
6623 end with a slash). */
6625 absolute_dirname (char *file
, char *dir
)
6630 slashp
= strrchr (file
, '/');
6632 return savestr (dir
);
6635 res
= absolute_filename (file
, dir
);
6641 /* Whether the argument string is an absolute file name. The argument
6642 string must have been canonicalized with canonicalize_filename. */
6644 filename_is_absolute (char *fn
)
6646 return (fn
[0] == '/'
6648 || (c_isalpha (fn
[0]) && fn
[1] == ':' && fn
[2] == '/')
6653 /* Downcase DOS drive letter and collapse separators into single slashes.
6656 canonicalize_filename (register char *fn
)
6661 /* Canonicalize drive letter case. */
6662 if (c_isupper (fn
[0]) && fn
[1] == ':')
6663 fn
[0] = c_tolower (fn
[0]);
6665 /* Collapse multiple forward- and back-slashes into a single forward
6667 for (cp
= fn
; *cp
!= '\0'; cp
++, fn
++)
6668 if (*cp
== '/' || *cp
== '\\')
6671 while (cp
[1] == '/' || cp
[1] == '\\')
6679 /* Collapse multiple slashes into a single slash. */
6680 for (cp
= fn
; *cp
!= '\0'; cp
++, fn
++)
6684 while (cp
[1] == '/')
6690 #endif /* !DOS_NT */
6696 /* Initialize a linebuffer for use. */
6698 linebuffer_init (linebuffer
*lbp
)
6700 lbp
->size
= (DEBUG
) ? 3 : 200;
6701 lbp
->buffer
= xnew (lbp
->size
, char);
6702 lbp
->buffer
[0] = '\0';
6706 /* Set the minimum size of a string contained in a linebuffer. */
6708 linebuffer_setlen (linebuffer
*lbp
, int toksize
)
6710 while (lbp
->size
<= toksize
)
6713 xrnew (lbp
->buffer
, lbp
->size
, char);
6718 /* Like malloc but get fatal error if memory is exhausted. */
6720 xmalloc (size_t size
)
6722 void *result
= malloc (size
);
6724 fatal ("virtual memory exhausted");
6729 xrealloc (void *ptr
, size_t size
)
6731 void *result
= realloc (ptr
, size
);
6733 fatal ("virtual memory exhausted");
6739 * indent-tabs-mode: t
6742 * c-font-lock-extra-types: ("FILE" "bool" "language" "linebuffer" "fdesc" "node" "regexp")
6743 * c-file-style: "gnu"
6747 /* etags.c ends here */