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 _Noreturn
void suggest_asking_for_help (void);
365 _Noreturn
void fatal (const char *, const char *);
366 static _Noreturn
void pfatal (const char *);
367 static void add_node (node
*, node
**);
369 static void process_file_name (char *, language
*);
370 static void process_file (FILE *, char *, language
*);
371 static void find_entries (FILE *);
372 static void free_tree (node
*);
373 static void free_fdesc (fdesc
*);
374 static void pfnote (char *, bool, char *, int, int, long);
375 static void invalidate_nodes (fdesc
*, node
**);
376 static void put_entries (node
*);
378 static char *concat (const char *, const char *, const char *);
379 static char *skip_spaces (char *);
380 static char *skip_non_spaces (char *);
381 static char *skip_name (char *);
382 static char *savenstr (const char *, int);
383 static char *savestr (const char *);
384 static char *etags_getcwd (void);
385 static char *relative_filename (char *, char *);
386 static char *absolute_filename (char *, char *);
387 static char *absolute_dirname (char *, char *);
388 static bool filename_is_absolute (char *f
);
389 static void canonicalize_filename (char *);
390 static char *etags_mktmp (void);
391 static void linebuffer_init (linebuffer
*);
392 static void linebuffer_setlen (linebuffer
*, int);
393 static void *xmalloc (size_t);
394 static void *xrealloc (void *, size_t);
397 static char searchar
= '/'; /* use /.../ searches */
399 static char *tagfile
; /* output file */
400 static char *progname
; /* name this program was invoked with */
401 static char *cwd
; /* current working directory */
402 static char *tagfiledir
; /* directory of tagfile */
403 static FILE *tagf
; /* ioptr for tags file */
404 static ptrdiff_t whatlen_max
; /* maximum length of any 'what' member */
406 static fdesc
*fdhead
; /* head of file description list */
407 static fdesc
*curfdp
; /* current file description */
408 static char *infilename
; /* current input file name */
409 static int lineno
; /* line number of current line */
410 static long charno
; /* current character number */
411 static long linecharno
; /* charno of start of current line */
412 static char *dbp
; /* pointer to start of current tag */
414 static const int invalidcharno
= -1;
416 static node
*nodehead
; /* the head of the binary tree of tags */
417 static node
*last_node
; /* the last node created */
419 static linebuffer lb
; /* the current line */
420 static linebuffer filebuf
; /* a buffer containing the whole file */
421 static linebuffer token_name
; /* a buffer containing a tag name */
423 static bool append_to_tagfile
; /* -a: append to tags */
424 /* The next five default to true in C and derived languages. */
425 static bool typedefs
; /* -t: create tags for C and Ada typedefs */
426 static bool typedefs_or_cplusplus
; /* -T: create tags for C typedefs, level */
427 /* 0 struct/enum/union decls, and C++ */
428 /* member functions. */
429 static bool constantypedefs
; /* -d: create tags for C #define, enum */
430 /* constants and variables. */
431 /* -D: opposite of -d. Default under ctags. */
432 static int globals
; /* create tags for global variables */
433 static int members
; /* create tags for C member variables */
434 static int declarations
; /* --declarations: tag them and extern in C&Co*/
435 static int no_line_directive
; /* ignore #line directives (undocumented) */
436 static int no_duplicates
; /* no duplicate tags for ctags (undocumented) */
437 static bool update
; /* -u: update tags */
438 static bool vgrind_style
; /* -v: create vgrind style index output */
439 static bool no_warnings
; /* -w: suppress warnings (undocumented) */
440 static bool cxref_style
; /* -x: create cxref style output */
441 static bool cplusplus
; /* .[hc] means C++, not C (undocumented) */
442 static bool ignoreindent
; /* -I: ignore indentation in C */
443 static int packages_only
; /* --packages-only: in Ada, only tag packages*/
445 /* STDIN is defined in LynxOS system headers */
450 #define STDIN 0x1001 /* returned by getopt_long on --parse-stdin */
451 static bool parsing_stdin
; /* --parse-stdin used */
453 static regexp
*p_head
; /* list of all regexps */
454 static bool need_filebuf
; /* some regexes are multi-line */
456 static struct option longopts
[] =
458 { "append", no_argument
, NULL
, 'a' },
459 { "packages-only", no_argument
, &packages_only
, 1 },
460 { "c++", no_argument
, NULL
, 'C' },
461 { "declarations", no_argument
, &declarations
, 1 },
462 { "no-line-directive", no_argument
, &no_line_directive
, 1 },
463 { "no-duplicates", no_argument
, &no_duplicates
, 1 },
464 { "help", no_argument
, NULL
, 'h' },
465 { "help", no_argument
, NULL
, 'H' },
466 { "ignore-indentation", no_argument
, NULL
, 'I' },
467 { "language", required_argument
, NULL
, 'l' },
468 { "members", no_argument
, &members
, 1 },
469 { "no-members", no_argument
, &members
, 0 },
470 { "output", required_argument
, NULL
, 'o' },
471 { "regex", required_argument
, NULL
, 'r' },
472 { "no-regex", no_argument
, NULL
, 'R' },
473 { "ignore-case-regex", required_argument
, NULL
, 'c' },
474 { "parse-stdin", required_argument
, NULL
, STDIN
},
475 { "version", no_argument
, NULL
, 'V' },
477 #if CTAGS /* Ctags options */
478 { "backward-search", no_argument
, NULL
, 'B' },
479 { "cxref", no_argument
, NULL
, 'x' },
480 { "defines", no_argument
, NULL
, 'd' },
481 { "globals", no_argument
, &globals
, 1 },
482 { "typedefs", no_argument
, NULL
, 't' },
483 { "typedefs-and-c++", no_argument
, NULL
, 'T' },
484 { "update", no_argument
, NULL
, 'u' },
485 { "vgrind", no_argument
, NULL
, 'v' },
486 { "no-warn", no_argument
, NULL
, 'w' },
488 #else /* Etags options */
489 { "no-defines", no_argument
, NULL
, 'D' },
490 { "no-globals", no_argument
, &globals
, 0 },
491 { "include", required_argument
, NULL
, 'i' },
496 static compressor compressors
[] =
498 { "z", "gzip -d -c"},
499 { "Z", "gzip -d -c"},
500 { "gz", "gzip -d -c"},
501 { "GZ", "gzip -d -c"},
502 { "bz2", "bzip2 -d -c" },
503 { "xz", "xz -d -c" },
512 static const char *Ada_suffixes
[] =
513 { "ads", "adb", "ada", NULL
};
514 static const char Ada_help
[] =
515 "In Ada code, functions, procedures, packages, tasks and types are\n\
516 tags. Use the `--packages-only' option to create tags for\n\
518 Ada tag names have suffixes indicating the type of entity:\n\
519 Entity type: Qualifier:\n\
520 ------------ ----------\n\
527 Thus, `M-x find-tag <RET> bidule/b <RET>' will go directly to the\n\
528 body of the package `bidule', while `M-x find-tag <RET> bidule <RET>'\n\
529 will just search for any tag `bidule'.";
532 static const char *Asm_suffixes
[] =
533 { "a", /* Unix assembler */
534 "asm", /* Microcontroller assembly */
535 "def", /* BSO/Tasking definition includes */
536 "inc", /* Microcontroller include files */
537 "ins", /* Microcontroller include files */
538 "s", "sa", /* Unix assembler */
539 "S", /* cpp-processed Unix assembler */
540 "src", /* BSO/Tasking C compiler output */
543 static const char Asm_help
[] =
544 "In assembler code, labels appearing at the beginning of a line,\n\
545 followed by a colon, are tags.";
548 /* Note that .c and .h can be considered C++, if the --c++ flag was
549 given, or if the `class' or `template' keywords are met inside the file.
550 That is why default_C_entries is called for these. */
551 static const char *default_C_suffixes
[] =
553 #if CTAGS /* C help for Ctags */
554 static const char default_C_help
[] =
555 "In C code, any C function is a tag. Use -t to tag typedefs.\n\
556 Use -T to tag definitions of `struct', `union' and `enum'.\n\
557 Use -d to tag `#define' macro definitions and `enum' constants.\n\
558 Use --globals to tag global variables.\n\
559 You can tag function declarations and external variables by\n\
560 using `--declarations', and struct members by using `--members'.";
561 #else /* C help for Etags */
562 static const char default_C_help
[] =
563 "In C code, any C function or typedef is a tag, and so are\n\
564 definitions of `struct', `union' and `enum'. `#define' macro\n\
565 definitions and `enum' constants are tags unless you specify\n\
566 `--no-defines'. Global variables are tags unless you specify\n\
567 `--no-globals' and so are struct members unless you specify\n\
568 `--no-members'. Use of `--no-globals', `--no-defines' and\n\
569 `--no-members' can make the tags table file much smaller.\n\
570 You can tag function declarations and external variables by\n\
571 using `--declarations'.";
572 #endif /* C help for Ctags and Etags */
574 static const char *Cplusplus_suffixes
[] =
575 { "C", "c++", "cc", "cpp", "cxx", "H", "h++", "hh", "hpp", "hxx",
576 "M", /* Objective C++ */
577 "pdb", /* PostScript with C syntax */
579 static const char Cplusplus_help
[] =
580 "In C++ code, all the tag constructs of C code are tagged. (Use\n\
581 --help --lang=c --lang=c++ for full help.)\n\
582 In addition to C tags, member functions are also recognized. Member\n\
583 variables are recognized unless you use the `--no-members' option.\n\
584 Tags for variables and functions in classes are named `CLASS::VARIABLE'\n\
585 and `CLASS::FUNCTION'. `operator' definitions have tag names like\n\
588 static const char *Cjava_suffixes
[] =
590 static char Cjava_help
[] =
591 "In Java code, all the tags constructs of C and C++ code are\n\
592 tagged. (Use --help --lang=c --lang=c++ --lang=java for full help.)";
595 static const char *Cobol_suffixes
[] =
596 { "COB", "cob", NULL
};
597 static char Cobol_help
[] =
598 "In Cobol code, tags are paragraph names; that is, any word\n\
599 starting in column 8 and followed by a period.";
601 static const char *Cstar_suffixes
[] =
602 { "cs", "hs", NULL
};
604 static const char *Erlang_suffixes
[] =
605 { "erl", "hrl", NULL
};
606 static const char Erlang_help
[] =
607 "In Erlang code, the tags are the functions, records and macros\n\
608 defined in the file.";
610 const char *Forth_suffixes
[] =
611 { "fth", "tok", NULL
};
612 static const char Forth_help
[] =
613 "In Forth code, tags are words defined by `:',\n\
614 constant, code, create, defer, value, variable, buffer:, field.";
616 static const char *Fortran_suffixes
[] =
617 { "F", "f", "f90", "for", NULL
};
618 static const char Fortran_help
[] =
619 "In Fortran code, functions, subroutines and block data are tags.";
621 static const char *HTML_suffixes
[] =
622 { "htm", "html", "shtml", NULL
};
623 static const char HTML_help
[] =
624 "In HTML input files, the tags are the `title' and the `h1', `h2',\n\
625 `h3' headers. Also, tags are `name=' in anchors and all\n\
626 occurrences of `id='.";
628 static const char *Lisp_suffixes
[] =
629 { "cl", "clisp", "el", "l", "lisp", "LSP", "lsp", "ml", NULL
};
630 static const char Lisp_help
[] =
631 "In Lisp code, any function defined with `defun', any variable\n\
632 defined with `defvar' or `defconst', and in general the first\n\
633 argument of any expression that starts with `(def' in column zero\n\
635 The `--declarations' option tags \"(defvar foo)\" constructs too.";
637 static const char *Lua_suffixes
[] =
638 { "lua", "LUA", NULL
};
639 static const char Lua_help
[] =
640 "In Lua scripts, all functions are tags.";
642 static const char *Makefile_filenames
[] =
643 { "Makefile", "makefile", "GNUMakefile", "Makefile.in", "Makefile.am", NULL
};
644 static const char Makefile_help
[] =
645 "In makefiles, targets are tags; additionally, variables are tags\n\
646 unless you specify `--no-globals'.";
648 static const char *Objc_suffixes
[] =
649 { "lm", /* Objective lex file */
650 "m", /* Objective C file */
652 static const char Objc_help
[] =
653 "In Objective C code, tags include Objective C definitions for classes,\n\
654 class categories, methods and protocols. Tags for variables and\n\
655 functions in classes are named `CLASS::VARIABLE' and `CLASS::FUNCTION'.\n\
656 (Use --help --lang=c --lang=objc --lang=java for full help.)";
658 static const char *Pascal_suffixes
[] =
659 { "p", "pas", NULL
};
660 static const char Pascal_help
[] =
661 "In Pascal code, the tags are the functions and procedures defined\n\
663 /* " // this is for working around an Emacs highlighting bug... */
665 static const char *Perl_suffixes
[] =
666 { "pl", "pm", NULL
};
667 static const char *Perl_interpreters
[] =
668 { "perl", "@PERL@", NULL
};
669 static const char Perl_help
[] =
670 "In Perl code, the tags are the packages, subroutines and variables\n\
671 defined by the `package', `sub', `my' and `local' keywords. Use\n\
672 `--globals' if you want to tag global variables. Tags for\n\
673 subroutines are named `PACKAGE::SUB'. The name for subroutines\n\
674 defined in the default package is `main::SUB'.";
676 static const char *PHP_suffixes
[] =
677 { "php", "php3", "php4", NULL
};
678 static const char PHP_help
[] =
679 "In PHP code, tags are functions, classes and defines. Unless you use\n\
680 the `--no-members' option, vars are tags too.";
682 static const char *plain_C_suffixes
[] =
683 { "pc", /* Pro*C file */
686 static const char *PS_suffixes
[] =
687 { "ps", "psw", NULL
}; /* .psw is for PSWrap */
688 static const char PS_help
[] =
689 "In PostScript code, the tags are the functions.";
691 static const char *Prolog_suffixes
[] =
693 static const char Prolog_help
[] =
694 "In Prolog code, tags are predicates and rules at the beginning of\n\
697 static const char *Python_suffixes
[] =
699 static const char Python_help
[] =
700 "In Python code, `def' or `class' at the beginning of a line\n\
703 /* Can't do the `SCM' or `scm' prefix with a version number. */
704 static const char *Scheme_suffixes
[] =
705 { "oak", "sch", "scheme", "SCM", "scm", "SM", "sm", "ss", "t", NULL
};
706 static const char Scheme_help
[] =
707 "In Scheme code, tags include anything defined with `def' or with a\n\
708 construct whose name starts with `def'. They also include\n\
709 variables set with `set!' at top level in the file.";
711 static const char *TeX_suffixes
[] =
712 { "bib", "clo", "cls", "ltx", "sty", "TeX", "tex", NULL
};
713 static const char TeX_help
[] =
714 "In LaTeX text, the argument of any of the commands `\\chapter',\n\
715 `\\section', `\\subsection', `\\subsubsection', `\\eqno', `\\label',\n\
716 `\\ref', `\\cite', `\\bibitem', `\\part', `\\appendix', `\\entry',\n\
717 `\\index', `\\def', `\\newcommand', `\\renewcommand',\n\
718 `\\newenvironment' or `\\renewenvironment' is a tag.\n\
720 Other commands can be specified by setting the environment variable\n\
721 `TEXTAGS' to a colon-separated list like, for example,\n\
722 TEXTAGS=\"mycommand:myothercommand\".";
725 static const char *Texinfo_suffixes
[] =
726 { "texi", "texinfo", "txi", NULL
};
727 static const char Texinfo_help
[] =
728 "for texinfo files, lines starting with @node are tagged.";
730 static const char *Yacc_suffixes
[] =
731 { "y", "y++", "ym", "yxx", "yy", NULL
}; /* .ym is Objective yacc file */
732 static const char Yacc_help
[] =
733 "In Bison or Yacc input files, each rule defines as a tag the\n\
734 nonterminal it constructs. The portions of the file that contain\n\
735 C code are parsed as C code (use --help --lang=c --lang=yacc\n\
738 static const char auto_help
[] =
739 "`auto' is not a real language, it indicates to use\n\
740 a default language for files base on file name suffix and file contents.";
742 static const char none_help
[] =
743 "`none' is not a real language, it indicates to only do\n\
744 regexp processing on files.";
746 static const char no_lang_help
[] =
747 "No detailed help available for this language.";
751 * Table of languages.
753 * It is ok for a given function to be listed under more than one
754 * name. I just didn't.
757 static language lang_names
[] =
759 { "ada", Ada_help
, Ada_funcs
, Ada_suffixes
},
760 { "asm", Asm_help
, Asm_labels
, Asm_suffixes
},
761 { "c", default_C_help
, default_C_entries
, default_C_suffixes
},
762 { "c++", Cplusplus_help
, Cplusplus_entries
, Cplusplus_suffixes
},
763 { "c*", no_lang_help
, Cstar_entries
, Cstar_suffixes
},
764 { "cobol", Cobol_help
, Cobol_paragraphs
, Cobol_suffixes
},
765 { "erlang", Erlang_help
, Erlang_functions
, Erlang_suffixes
},
766 { "forth", Forth_help
, Forth_words
, Forth_suffixes
},
767 { "fortran", Fortran_help
, Fortran_functions
, Fortran_suffixes
},
768 { "html", HTML_help
, HTML_labels
, HTML_suffixes
},
769 { "java", Cjava_help
, Cjava_entries
, Cjava_suffixes
},
770 { "lisp", Lisp_help
, Lisp_functions
, Lisp_suffixes
},
771 { "lua", Lua_help
, Lua_functions
, Lua_suffixes
},
772 { "makefile", Makefile_help
,Makefile_targets
,NULL
,Makefile_filenames
},
773 { "objc", Objc_help
, plain_C_entries
, Objc_suffixes
},
774 { "pascal", Pascal_help
, Pascal_functions
, Pascal_suffixes
},
775 { "perl",Perl_help
,Perl_functions
,Perl_suffixes
,NULL
,Perl_interpreters
},
776 { "php", PHP_help
, PHP_functions
, PHP_suffixes
},
777 { "postscript",PS_help
, PS_functions
, PS_suffixes
},
778 { "proc", no_lang_help
, plain_C_entries
, plain_C_suffixes
},
779 { "prolog", Prolog_help
, Prolog_functions
, Prolog_suffixes
},
780 { "python", Python_help
, Python_functions
, Python_suffixes
},
781 { "scheme", Scheme_help
, Scheme_functions
, Scheme_suffixes
},
782 { "tex", TeX_help
, TeX_commands
, TeX_suffixes
},
783 { "texinfo", Texinfo_help
, Texinfo_nodes
, Texinfo_suffixes
},
784 { "yacc", Yacc_help
,Yacc_entries
,Yacc_suffixes
,NULL
,NULL
,true},
785 { "auto", auto_help
}, /* default guessing scheme */
786 { "none", none_help
, just_read_file
}, /* regexp matching only */
787 { NULL
} /* end of list */
792 print_language_names (void)
795 const char **name
, **ext
;
797 puts ("\nThese are the currently supported languages, along with the\n\
798 default file names and dot suffixes:");
799 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
801 printf (" %-*s", 10, lang
->name
);
802 if (lang
->filenames
!= NULL
)
803 for (name
= lang
->filenames
; *name
!= NULL
; name
++)
804 printf (" %s", *name
);
805 if (lang
->suffixes
!= NULL
)
806 for (ext
= lang
->suffixes
; *ext
!= NULL
; ext
++)
807 printf (" .%s", *ext
);
810 puts ("where `auto' means use default language for files based on file\n\
811 name suffix, and `none' means only do regexp processing on files.\n\
812 If no language is specified and no matching suffix is found,\n\
813 the first line of the file is read for a sharp-bang (#!) sequence\n\
814 followed by the name of an interpreter. If no such sequence is found,\n\
815 Fortran is tried first; if no tags are found, C is tried next.\n\
816 When parsing any C file, a \"class\" or \"template\" keyword\n\
818 puts ("Compressed files are supported using gzip, bzip2, and xz.\n\
820 For detailed help on a given language use, for example,\n\
821 etags --help --lang=ada.");
825 # define EMACS_NAME "standalone"
828 # define VERSION "17.38.1.4"
830 static _Noreturn
void
833 char emacs_copyright
[] = COPYRIGHT
;
835 printf ("%s (%s %s)\n", (CTAGS
) ? "ctags" : "etags", EMACS_NAME
, VERSION
);
836 puts (emacs_copyright
);
837 puts ("This program is distributed under the terms in ETAGS.README");
842 #ifndef PRINT_UNDOCUMENTED_OPTIONS_HELP
843 # define PRINT_UNDOCUMENTED_OPTIONS_HELP false
846 static _Noreturn
void
847 print_help (argument
*argbuffer
)
849 bool help_for_lang
= false;
851 for (; argbuffer
->arg_type
!= at_end
; argbuffer
++)
852 if (argbuffer
->arg_type
== at_language
)
856 puts (argbuffer
->lang
->help
);
857 help_for_lang
= true;
863 printf ("Usage: %s [options] [[regex-option ...] file-name] ...\n\
865 These are the options accepted by %s.\n", progname
, progname
);
866 puts ("You may use unambiguous abbreviations for the long option names.");
867 puts (" A - as file name means read names from stdin (one per line).\n\
868 Absolute names are stored in the output file as they are.\n\
869 Relative ones are stored relative to the output file's directory.\n");
871 puts ("-a, --append\n\
872 Append tag entries to existing tags file.");
874 puts ("--packages-only\n\
875 For Ada files, only generate tags for packages.");
878 puts ("-B, --backward-search\n\
879 Write the search commands for the tag entries using '?', the\n\
880 backward-search command instead of '/', the forward-search command.");
882 /* This option is mostly obsolete, because etags can now automatically
883 detect C++. Retained for backward compatibility and for debugging and
884 experimentation. In principle, we could want to tag as C++ even
885 before any "class" or "template" keyword.
887 Treat files whose name suffix defaults to C language as C++ files.");
890 puts ("--declarations\n\
891 In C and derived languages, create tags for function declarations,");
893 puts ("\tand create tags for extern variables if --globals is used.");
896 ("\tand create tags for extern variables unless --no-globals is used.");
899 puts ("-d, --defines\n\
900 Create tag entries for C #define constants and enum constants, too.");
902 puts ("-D, --no-defines\n\
903 Don't create tag entries for C #define constants and enum constants.\n\
904 This makes the tags file smaller.");
907 puts ("-i FILE, --include=FILE\n\
908 Include a note in tag file indicating that, when searching for\n\
909 a tag, one should also consult the tags file FILE after\n\
910 checking the current file.");
912 puts ("-l LANG, --language=LANG\n\
913 Force the following files to be considered as written in the\n\
914 named language up to the next --language=LANG option.");
918 Create tag entries for global variables in some languages.");
920 puts ("--no-globals\n\
921 Do not create tag entries for global variables in some\n\
922 languages. This makes the tags file smaller.");
924 if (PRINT_UNDOCUMENTED_OPTIONS_HELP
)
925 puts ("--no-line-directive\n\
926 Ignore #line preprocessor directives in C and derived languages.");
930 Create tag entries for members of structures in some languages.");
932 puts ("--no-members\n\
933 Do not create tag entries for members of structures\n\
934 in some languages.");
936 puts ("-r REGEXP, --regex=REGEXP or --regex=@regexfile\n\
937 Make a tag for each line matching a regular expression pattern\n\
938 in the following files. {LANGUAGE}REGEXP uses REGEXP for LANGUAGE\n\
939 files only. REGEXFILE is a file containing one REGEXP per line.\n\
940 REGEXP takes the form /TAGREGEXP/TAGNAME/MODS, where TAGNAME/ is\n\
941 optional. The TAGREGEXP pattern is anchored (as if preceded by ^).");
942 puts (" If TAGNAME/ is present, the tags created are named.\n\
943 For example Tcl named tags can be created with:\n\
944 --regex=\"/proc[ \\t]+\\([^ \\t]+\\)/\\1/.\".\n\
945 MODS are optional one-letter modifiers: `i' means to ignore case,\n\
946 `m' means to allow multi-line matches, `s' implies `m' and\n\
947 causes dot to match any character, including newline.");
949 puts ("-R, --no-regex\n\
950 Don't create tags from regexps for the following files.");
952 puts ("-I, --ignore-indentation\n\
953 In C and C++ do not assume that a closing brace in the first\n\
954 column is the final brace of a function or structure definition.");
956 puts ("-o FILE, --output=FILE\n\
957 Write the tags to FILE.");
959 puts ("--parse-stdin=NAME\n\
960 Read from standard input and record tags as belonging to file NAME.");
964 puts ("-t, --typedefs\n\
965 Generate tag entries for C and Ada typedefs.");
966 puts ("-T, --typedefs-and-c++\n\
967 Generate tag entries for C typedefs, C struct/enum/union tags,\n\
968 and C++ member functions.");
972 puts ("-u, --update\n\
973 Update the tag entries for the given files, leaving tag\n\
974 entries for other files in place. Currently, this is\n\
975 implemented by deleting the existing entries for the given\n\
976 files and then rewriting the new entries at the end of the\n\
977 tags file. It is often faster to simply rebuild the entire\n\
978 tag file than to use this.");
982 puts ("-v, --vgrind\n\
983 Print on the standard output an index of items intended for\n\
984 human consumption, similar to the output of vgrind. The index\n\
985 is sorted, and gives the page number of each item.");
987 if (PRINT_UNDOCUMENTED_OPTIONS_HELP
)
988 puts ("-w, --no-duplicates\n\
989 Do not create duplicate tag entries, for compatibility with\n\
990 traditional ctags.");
992 if (PRINT_UNDOCUMENTED_OPTIONS_HELP
)
993 puts ("-w, --no-warn\n\
994 Suppress warning messages about duplicate tag entries.");
996 puts ("-x, --cxref\n\
997 Like --vgrind, but in the style of cxref, rather than vgrind.\n\
998 The output uses line numbers instead of page numbers, but\n\
999 beyond that the differences are cosmetic; try both to see\n\
1003 puts ("-V, --version\n\
1004 Print the version of the program.\n\
1006 Print this help message.\n\
1007 Followed by one or more `--language' options prints detailed\n\
1008 help about tag generation for the specified languages.");
1010 print_language_names ();
1013 puts ("Report bugs to bug-gnu-emacs@gnu.org");
1015 exit (EXIT_SUCCESS
);
1020 main (int argc
, char **argv
)
1023 unsigned int nincluded_files
;
1024 char **included_files
;
1025 argument
*argbuffer
;
1026 int current_arg
, file_count
;
1027 linebuffer filename_lb
;
1028 bool help_asked
= false;
1034 nincluded_files
= 0;
1035 included_files
= xnew (argc
, char *);
1039 /* Allocate enough no matter what happens. Overkill, but each one
1041 argbuffer
= xnew (argc
, argument
);
1044 * Always find typedefs and structure tags.
1045 * Also default to find macro constants, enum constants, struct
1046 * members and global variables. Do it for both etags and ctags.
1048 typedefs
= typedefs_or_cplusplus
= constantypedefs
= true;
1049 globals
= members
= true;
1051 /* When the optstring begins with a '-' getopt_long does not rearrange the
1052 non-options arguments to be at the end, but leaves them alone. */
1053 optstring
= concat ("-ac:Cf:Il:o:r:RSVhH",
1054 (CTAGS
) ? "BxdtTuvw" : "Di:",
1057 while ((opt
= getopt_long (argc
, argv
, optstring
, longopts
, NULL
)) != EOF
)
1061 /* If getopt returns 0, then it has already processed a
1062 long-named option. We should do nothing. */
1066 /* This means that a file name has been seen. Record it. */
1067 argbuffer
[current_arg
].arg_type
= at_filename
;
1068 argbuffer
[current_arg
].what
= optarg
;
1069 len
= strlen (optarg
);
1070 if (whatlen_max
< len
)
1077 /* Parse standard input. Idea by Vivek <vivek@etla.org>. */
1078 argbuffer
[current_arg
].arg_type
= at_stdin
;
1079 argbuffer
[current_arg
].what
= optarg
;
1080 len
= strlen (optarg
);
1081 if (whatlen_max
< len
)
1086 fatal ("cannot parse standard input more than once", (char *)NULL
);
1087 parsing_stdin
= true;
1090 /* Common options. */
1091 case 'a': append_to_tagfile
= true; break;
1092 case 'C': cplusplus
= true; break;
1093 case 'f': /* for compatibility with old makefiles */
1097 error ("-o option may only be given once.");
1098 suggest_asking_for_help ();
1104 case 'S': /* for backward compatibility */
1105 ignoreindent
= true;
1109 language
*lang
= get_language_from_langname (optarg
);
1112 argbuffer
[current_arg
].lang
= lang
;
1113 argbuffer
[current_arg
].arg_type
= at_language
;
1119 /* Backward compatibility: support obsolete --ignore-case-regexp. */
1120 optarg
= concat (optarg
, "i", ""); /* memory leak here */
1123 argbuffer
[current_arg
].arg_type
= at_regexp
;
1124 argbuffer
[current_arg
].what
= optarg
;
1125 len
= strlen (optarg
);
1126 if (whatlen_max
< len
)
1131 argbuffer
[current_arg
].arg_type
= at_regexp
;
1132 argbuffer
[current_arg
].what
= NULL
;
1144 case 'D': constantypedefs
= false; break;
1145 case 'i': included_files
[nincluded_files
++] = optarg
; break;
1147 /* Ctags options. */
1148 case 'B': searchar
= '?'; break;
1149 case 'd': constantypedefs
= true; break;
1150 case 't': typedefs
= true; break;
1151 case 'T': typedefs
= typedefs_or_cplusplus
= true; break;
1152 case 'u': update
= true; break;
1153 case 'v': vgrind_style
= true; /*FALLTHRU*/
1154 case 'x': cxref_style
= true; break;
1155 case 'w': no_warnings
= true; break;
1157 suggest_asking_for_help ();
1161 /* No more options. Store the rest of arguments. */
1162 for (; optind
< argc
; optind
++)
1164 argbuffer
[current_arg
].arg_type
= at_filename
;
1165 argbuffer
[current_arg
].what
= argv
[optind
];
1166 len
= strlen (argv
[optind
]);
1167 if (whatlen_max
< len
)
1173 argbuffer
[current_arg
].arg_type
= at_end
;
1176 print_help (argbuffer
);
1179 if (nincluded_files
== 0 && file_count
== 0)
1181 error ("no input files specified.");
1182 suggest_asking_for_help ();
1186 if (tagfile
== NULL
)
1187 tagfile
= savestr (CTAGS
? "tags" : "TAGS");
1188 cwd
= etags_getcwd (); /* the current working directory */
1189 if (cwd
[strlen (cwd
) - 1] != '/')
1192 cwd
= concat (oldcwd
, "/", "");
1196 /* Compute base directory for relative file names. */
1197 if (streq (tagfile
, "-")
1198 || strneq (tagfile
, "/dev/", 5))
1199 tagfiledir
= cwd
; /* relative file names are relative to cwd */
1202 canonicalize_filename (tagfile
);
1203 tagfiledir
= absolute_dirname (tagfile
, cwd
);
1206 linebuffer_init (&lb
);
1207 linebuffer_init (&filename_lb
);
1208 linebuffer_init (&filebuf
);
1209 linebuffer_init (&token_name
);
1213 if (streq (tagfile
, "-"))
1216 SET_BINARY (fileno (stdout
));
1219 tagf
= fopen (tagfile
, append_to_tagfile
? "ab" : "wb");
1225 * Loop through files finding functions.
1227 for (i
= 0; i
< current_arg
; i
++)
1229 static language
*lang
; /* non-NULL if language is forced */
1232 switch (argbuffer
[i
].arg_type
)
1235 lang
= argbuffer
[i
].lang
;
1238 analyze_regex (argbuffer
[i
].what
);
1241 this_file
= argbuffer
[i
].what
;
1242 /* Input file named "-" means read file names from stdin
1243 (one per line) and use them. */
1244 if (streq (this_file
, "-"))
1247 fatal ("cannot parse standard input AND read file names from it",
1249 while (readline_internal (&filename_lb
, stdin
, "-") > 0)
1250 process_file_name (filename_lb
.buffer
, lang
);
1253 process_file_name (this_file
, lang
);
1256 this_file
= argbuffer
[i
].what
;
1257 process_file (stdin
, this_file
, lang
);
1264 free (filebuf
.buffer
);
1265 free (token_name
.buffer
);
1267 if (!CTAGS
|| cxref_style
)
1269 /* Write the remaining tags to tagf (ETAGS) or stdout (CXREF). */
1270 put_entries (nodehead
);
1271 free_tree (nodehead
);
1277 /* Output file entries that have no tags. */
1278 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
1280 fprintf (tagf
, "\f\n%s,0\n", fdp
->taggedfname
);
1282 while (nincluded_files
-- > 0)
1283 fprintf (tagf
, "\f\n%s,include\n", *included_files
++);
1285 if (fclose (tagf
) == EOF
)
1289 exit (EXIT_SUCCESS
);
1292 /* From here on, we are in (CTAGS && !cxref_style) */
1296 xmalloc (strlen (tagfile
) + whatlen_max
+
1297 sizeof "mv..OTAGS;fgrep -v '\t\t' OTAGS >;rm OTAGS");
1298 for (i
= 0; i
< current_arg
; ++i
)
1300 switch (argbuffer
[i
].arg_type
)
1306 continue; /* the for loop */
1308 char *z
= stpcpy (cmd
, "mv ");
1309 z
= stpcpy (z
, tagfile
);
1310 z
= stpcpy (z
, " OTAGS;fgrep -v '\t");
1311 z
= stpcpy (z
, argbuffer
[i
].what
);
1312 z
= stpcpy (z
, "\t' OTAGS >");
1313 z
= stpcpy (z
, tagfile
);
1314 strcpy (z
, ";rm OTAGS");
1315 if (system (cmd
) != EXIT_SUCCESS
)
1316 fatal ("failed to execute shell command", (char *)NULL
);
1319 append_to_tagfile
= true;
1322 tagf
= fopen (tagfile
, append_to_tagfile
? "ab" : "wb");
1325 put_entries (nodehead
); /* write all the tags (CTAGS) */
1326 free_tree (nodehead
);
1328 if (fclose (tagf
) == EOF
)
1332 if (append_to_tagfile
|| update
)
1334 char *cmd
= xmalloc (2 * strlen (tagfile
) + sizeof "sort -u -o..");
1335 /* Maybe these should be used:
1336 setenv ("LC_COLLATE", "C", 1);
1337 setenv ("LC_ALL", "C", 1); */
1338 char *z
= stpcpy (cmd
, "sort -u -o ");
1339 z
= stpcpy (z
, tagfile
);
1341 strcpy (z
, tagfile
);
1342 exit (system (cmd
));
1344 return EXIT_SUCCESS
;
1349 * Return a compressor given the file name. If EXTPTR is non-zero,
1350 * return a pointer into FILE where the compressor-specific
1351 * extension begins. If no compressor is found, NULL is returned
1352 * and EXTPTR is not significant.
1353 * Idea by Vladimir Alexiev <vladimir@cs.ualberta.ca> (1998)
1356 get_compressor_from_suffix (char *file
, char **extptr
)
1359 char *slash
, *suffix
;
1361 /* File has been processed by canonicalize_filename,
1362 so we don't need to consider backslashes on DOS_NT. */
1363 slash
= strrchr (file
, '/');
1364 suffix
= strrchr (file
, '.');
1365 if (suffix
== NULL
|| suffix
< slash
)
1370 /* Let those poor souls who live with DOS 8+3 file name limits get
1371 some solace by treating foo.cgz as if it were foo.c.gz, etc.
1372 Only the first do loop is run if not MSDOS */
1375 for (compr
= compressors
; compr
->suffix
!= NULL
; compr
++)
1376 if (streq (compr
->suffix
, suffix
))
1379 break; /* do it only once: not really a loop */
1382 } while (*suffix
!= '\0');
1389 * Return a language given the name.
1392 get_language_from_langname (const char *name
)
1397 error ("empty language name");
1400 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1401 if (streq (name
, lang
->name
))
1403 error ("unknown language \"%s\"", name
);
1411 * Return a language given the interpreter name.
1414 get_language_from_interpreter (char *interpreter
)
1419 if (interpreter
== NULL
)
1421 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1422 if (lang
->interpreters
!= NULL
)
1423 for (iname
= lang
->interpreters
; *iname
!= NULL
; iname
++)
1424 if (streq (*iname
, interpreter
))
1433 * Return a language given the file name.
1436 get_language_from_filename (char *file
, int case_sensitive
)
1439 const char **name
, **ext
, *suffix
;
1441 /* Try whole file name first. */
1442 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1443 if (lang
->filenames
!= NULL
)
1444 for (name
= lang
->filenames
; *name
!= NULL
; name
++)
1445 if ((case_sensitive
)
1446 ? streq (*name
, file
)
1447 : strcaseeq (*name
, file
))
1450 /* If not found, try suffix after last dot. */
1451 suffix
= strrchr (file
, '.');
1455 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1456 if (lang
->suffixes
!= NULL
)
1457 for (ext
= lang
->suffixes
; *ext
!= NULL
; ext
++)
1458 if ((case_sensitive
)
1459 ? streq (*ext
, suffix
)
1460 : strcaseeq (*ext
, suffix
))
1467 * This routine is called on each file argument.
1470 process_file_name (char *file
, language
*lang
)
1475 char *compressed_name
, *uncompressed_name
;
1476 char *ext
, *real_name
, *tmp_name
;
1479 canonicalize_filename (file
);
1480 if (streq (file
, tagfile
) && !streq (tagfile
, "-"))
1482 error ("skipping inclusion of %s in self.", file
);
1485 compr
= get_compressor_from_suffix (file
, &ext
);
1488 compressed_name
= file
;
1489 uncompressed_name
= savenstr (file
, ext
- file
);
1493 compressed_name
= NULL
;
1494 uncompressed_name
= file
;
1497 /* If the canonicalized uncompressed name
1498 has already been dealt with, skip it silently. */
1499 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
1501 assert (fdp
->infname
!= NULL
);
1502 if (streq (uncompressed_name
, fdp
->infname
))
1506 inf
= fopen (file
, "r" FOPEN_BINARY
);
1511 int file_errno
= errno
;
1512 if (compressed_name
)
1514 /* Try with the given suffix. */
1515 inf
= fopen (uncompressed_name
, "r" FOPEN_BINARY
);
1517 real_name
= uncompressed_name
;
1521 /* Try all possible suffixes. */
1522 for (compr
= compressors
; compr
->suffix
!= NULL
; compr
++)
1524 compressed_name
= concat (file
, ".", compr
->suffix
);
1525 inf
= fopen (compressed_name
, "r" FOPEN_BINARY
);
1528 real_name
= compressed_name
;
1533 char *suf
= compressed_name
+ strlen (file
);
1534 size_t suflen
= strlen (compr
->suffix
) + 1;
1535 for ( ; suf
[1]; suf
++, suflen
--)
1537 memmove (suf
, suf
+ 1, suflen
);
1538 inf
= fopen (compressed_name
, "r" FOPEN_BINARY
);
1541 real_name
= compressed_name
;
1548 free (compressed_name
);
1549 compressed_name
= NULL
;
1560 if (real_name
== compressed_name
)
1563 tmp_name
= etags_mktmp ();
1568 char *cmd1
= concat (compr
->command
, " '", real_name
);
1569 char *cmd
= concat (cmd1
, "' > ", tmp_name
);
1572 if (system (cmd
) == -1)
1579 inf
= fopen (tmp_name
, "r" FOPEN_BINARY
);
1593 process_file (inf
, uncompressed_name
, lang
);
1595 retval
= fclose (inf
);
1596 if (real_name
== compressed_name
)
1605 if (compressed_name
!= file
)
1606 free (compressed_name
);
1607 if (uncompressed_name
!= file
)
1608 free (uncompressed_name
);
1615 process_file (FILE *fh
, char *fn
, language
*lang
)
1617 static const fdesc emptyfdesc
;
1621 /* Create a new input file description entry. */
1622 fdp
= xnew (1, fdesc
);
1625 fdp
->infname
= savestr (fn
);
1627 fdp
->infabsname
= absolute_filename (fn
, cwd
);
1628 fdp
->infabsdir
= absolute_dirname (fn
, cwd
);
1629 if (filename_is_absolute (fn
))
1631 /* An absolute file name. Canonicalize it. */
1632 fdp
->taggedfname
= absolute_filename (fn
, NULL
);
1636 /* A file name relative to cwd. Make it relative
1637 to the directory of the tags file. */
1638 fdp
->taggedfname
= relative_filename (fn
, tagfiledir
);
1640 fdp
->usecharno
= true; /* use char position when making tags */
1642 fdp
->written
= false; /* not written on tags file yet */
1645 curfdp
= fdhead
; /* the current file description */
1649 /* If not Ctags, and if this is not metasource and if it contained no #line
1650 directives, we can write the tags and free all nodes pointing to
1653 && curfdp
->usecharno
/* no #line directives in this file */
1654 && !curfdp
->lang
->metasource
)
1658 /* Look for the head of the sublist relative to this file. See add_node
1659 for the structure of the node tree. */
1661 for (np
= nodehead
; np
!= NULL
; prev
= np
, np
= np
->left
)
1662 if (np
->fdp
== curfdp
)
1665 /* If we generated tags for this file, write and delete them. */
1668 /* This is the head of the last sublist, if any. The following
1669 instructions depend on this being true. */
1670 assert (np
->left
== NULL
);
1672 assert (fdhead
== curfdp
);
1673 assert (last_node
->fdp
== curfdp
);
1674 put_entries (np
); /* write tags for file curfdp->taggedfname */
1675 free_tree (np
); /* remove the written nodes */
1677 nodehead
= NULL
; /* no nodes left */
1679 prev
->left
= NULL
; /* delete the pointer to the sublist */
1685 reset_input (FILE *inf
)
1687 if (fseek (inf
, 0, SEEK_SET
) != 0)
1688 perror (infilename
);
1692 * This routine opens the specified file and calls the function
1693 * which finds the function and type definitions.
1696 find_entries (FILE *inf
)
1699 language
*lang
= curfdp
->lang
;
1700 Lang_function
*parser
= NULL
;
1702 /* If user specified a language, use it. */
1703 if (lang
!= NULL
&& lang
->function
!= NULL
)
1705 parser
= lang
->function
;
1708 /* Else try to guess the language given the file name. */
1711 lang
= get_language_from_filename (curfdp
->infname
, true);
1712 if (lang
!= NULL
&& lang
->function
!= NULL
)
1714 curfdp
->lang
= lang
;
1715 parser
= lang
->function
;
1719 /* Else look for sharp-bang as the first two characters. */
1721 && readline_internal (&lb
, inf
, infilename
) > 0
1723 && lb
.buffer
[0] == '#'
1724 && lb
.buffer
[1] == '!')
1728 /* Set lp to point at the first char after the last slash in the
1729 line or, if no slashes, at the first nonblank. Then set cp to
1730 the first successive blank and terminate the string. */
1731 lp
= strrchr (lb
.buffer
+2, '/');
1735 lp
= skip_spaces (lb
.buffer
+ 2);
1736 cp
= skip_non_spaces (lp
);
1739 if (strlen (lp
) > 0)
1741 lang
= get_language_from_interpreter (lp
);
1742 if (lang
!= NULL
&& lang
->function
!= NULL
)
1744 curfdp
->lang
= lang
;
1745 parser
= lang
->function
;
1752 /* Else try to guess the language given the case insensitive file name. */
1755 lang
= get_language_from_filename (curfdp
->infname
, false);
1756 if (lang
!= NULL
&& lang
->function
!= NULL
)
1758 curfdp
->lang
= lang
;
1759 parser
= lang
->function
;
1763 /* Else try Fortran or C. */
1766 node
*old_last_node
= last_node
;
1768 curfdp
->lang
= get_language_from_langname ("fortran");
1771 if (old_last_node
== last_node
)
1772 /* No Fortran entries found. Try C. */
1775 curfdp
->lang
= get_language_from_langname (cplusplus
? "c++" : "c");
1781 if (!no_line_directive
1782 && curfdp
->lang
!= NULL
&& curfdp
->lang
->metasource
)
1783 /* It may be that this is a bingo.y file, and we already parsed a bingo.c
1784 file, or anyway we parsed a file that is automatically generated from
1785 this one. If this is the case, the bingo.c file contained #line
1786 directives that generated tags pointing to this file. Let's delete
1787 them all before parsing this file, which is the real source. */
1789 fdesc
**fdpp
= &fdhead
;
1790 while (*fdpp
!= NULL
)
1792 && streq ((*fdpp
)->taggedfname
, curfdp
->taggedfname
))
1793 /* We found one of those! We must delete both the file description
1794 and all tags referring to it. */
1796 fdesc
*badfdp
= *fdpp
;
1798 /* Delete the tags referring to badfdp->taggedfname
1799 that were obtained from badfdp->infname. */
1800 invalidate_nodes (badfdp
, &nodehead
);
1802 *fdpp
= badfdp
->next
; /* remove the bad description from the list */
1803 free_fdesc (badfdp
);
1806 fdpp
= &(*fdpp
)->next
; /* advance the list pointer */
1809 assert (parser
!= NULL
);
1811 /* Generic initializations before reading from file. */
1812 linebuffer_setlen (&filebuf
, 0); /* reset the file buffer */
1814 /* Generic initializations before parsing file with readline. */
1815 lineno
= 0; /* reset global line number */
1816 charno
= 0; /* reset global char number */
1817 linecharno
= 0; /* reset global char number of line start */
1821 regex_tag_multiline ();
1826 * Check whether an implicitly named tag should be created,
1827 * then call `pfnote'.
1828 * NAME is a string that is internally copied by this function.
1830 * TAGS format specification
1831 * Idea by Sam Kendall <kendall@mv.mv.com> (1997)
1832 * The following is explained in some more detail in etc/ETAGS.EBNF.
1834 * make_tag creates tags with "implicit tag names" (unnamed tags)
1835 * if the following are all true, assuming NONAM=" \f\t\n\r()=,;":
1836 * 1. NAME does not contain any of the characters in NONAM;
1837 * 2. LINESTART contains name as either a rightmost, or rightmost but
1838 * one character, substring;
1839 * 3. the character, if any, immediately before NAME in LINESTART must
1840 * be a character in NONAM;
1841 * 4. the character, if any, immediately after NAME in LINESTART must
1842 * also be a character in NONAM.
1844 * The implementation uses the notinname() macro, which recognizes the
1845 * characters stored in the string `nonam'.
1846 * etags.el needs to use the same characters that are in NONAM.
1849 make_tag (const char *name
, /* tag name, or NULL if unnamed */
1850 int namelen
, /* tag length */
1851 bool is_func
, /* tag is a function */
1852 char *linestart
, /* start of the line where tag is */
1853 int linelen
, /* length of the line where tag is */
1854 int lno
, /* line number */
1855 long int cno
) /* character number */
1857 bool named
= (name
!= NULL
&& namelen
> 0);
1860 if (!CTAGS
&& named
) /* maybe set named to false */
1861 /* Let's try to make an implicit tag name, that is, create an unnamed tag
1862 such that etags.el can guess a name from it. */
1865 register const char *cp
= name
;
1867 for (i
= 0; i
< namelen
; i
++)
1868 if (notinname (*cp
++))
1870 if (i
== namelen
) /* rule #1 */
1872 cp
= linestart
+ linelen
- namelen
;
1873 if (notinname (linestart
[linelen
-1]))
1874 cp
-= 1; /* rule #4 */
1875 if (cp
>= linestart
/* rule #2 */
1877 || notinname (cp
[-1])) /* rule #3 */
1878 && strneq (name
, cp
, namelen
)) /* rule #2 */
1879 named
= false; /* use implicit tag name */
1884 nname
= savenstr (name
, namelen
);
1886 pfnote (nname
, is_func
, linestart
, linelen
, lno
, cno
);
1891 pfnote (char *name
, bool is_func
, char *linestart
, int linelen
, int lno
,
1893 /* tag name, or NULL if unnamed */
1894 /* tag is a function */
1895 /* start of the line where tag is */
1896 /* length of the line where tag is */
1898 /* character number */
1902 assert (name
== NULL
|| name
[0] != '\0');
1903 if (CTAGS
&& name
== NULL
)
1906 np
= xnew (1, node
);
1908 /* If ctags mode, change name "main" to M<thisfilename>. */
1909 if (CTAGS
&& !cxref_style
&& streq (name
, "main"))
1911 char *fp
= strrchr (curfdp
->taggedfname
, '/');
1912 np
->name
= concat ("M", fp
== NULL
? curfdp
->taggedfname
: fp
+ 1, "");
1913 fp
= strrchr (np
->name
, '.');
1914 if (fp
!= NULL
&& fp
[1] != '\0' && fp
[2] == '\0')
1920 np
->been_warned
= false;
1922 np
->is_func
= is_func
;
1924 if (np
->fdp
->usecharno
)
1925 /* Our char numbers are 0-base, because of C language tradition?
1926 ctags compatibility? old versions compatibility? I don't know.
1927 Anyway, since emacs's are 1-base we expect etags.el to take care
1928 of the difference. If we wanted to have 1-based numbers, we would
1929 uncomment the +1 below. */
1930 np
->cno
= cno
/* + 1 */ ;
1932 np
->cno
= invalidcharno
;
1933 np
->left
= np
->right
= NULL
;
1934 if (CTAGS
&& !cxref_style
)
1936 if (strlen (linestart
) < 50)
1937 np
->regex
= concat (linestart
, "$", "");
1939 np
->regex
= savenstr (linestart
, 50);
1942 np
->regex
= savenstr (linestart
, linelen
);
1944 add_node (np
, &nodehead
);
1949 * recurse on left children, iterate on right children.
1952 free_tree (register node
*np
)
1956 register node
*node_right
= np
->right
;
1957 free_tree (np
->left
);
1967 * delete a file description
1970 free_fdesc (register fdesc
*fdp
)
1972 free (fdp
->infname
);
1973 free (fdp
->infabsname
);
1974 free (fdp
->infabsdir
);
1975 free (fdp
->taggedfname
);
1982 * Adds a node to the tree of nodes. In etags mode, sort by file
1983 * name. In ctags mode, sort by tag name. Make no attempt at
1986 * add_node is the only function allowed to add nodes, so it can
1990 add_node (node
*np
, node
**cur_node_p
)
1993 register node
*cur_node
= *cur_node_p
;
1995 if (cur_node
== NULL
)
2005 /* For each file name, tags are in a linked sublist on the right
2006 pointer. The first tags of different files are a linked list
2007 on the left pointer. last_node points to the end of the last
2009 if (last_node
!= NULL
&& last_node
->fdp
== np
->fdp
)
2011 /* Let's use the same sublist as the last added node. */
2012 assert (last_node
->right
== NULL
);
2013 last_node
->right
= np
;
2016 else if (cur_node
->fdp
== np
->fdp
)
2018 /* Scanning the list we found the head of a sublist which is
2019 good for us. Let's scan this sublist. */
2020 add_node (np
, &cur_node
->right
);
2023 /* The head of this sublist is not good for us. Let's try the
2025 add_node (np
, &cur_node
->left
);
2026 } /* if ETAGS mode */
2031 dif
= strcmp (np
->name
, cur_node
->name
);
2034 * If this tag name matches an existing one, then
2035 * do not add the node, but maybe print a warning.
2037 if (no_duplicates
&& !dif
)
2039 if (np
->fdp
== cur_node
->fdp
)
2043 fprintf (stderr
, "Duplicate entry in file %s, line %d: %s\n",
2044 np
->fdp
->infname
, lineno
, np
->name
);
2045 fprintf (stderr
, "Second entry ignored\n");
2048 else if (!cur_node
->been_warned
&& !no_warnings
)
2052 "Duplicate entry in files %s and %s: %s (Warning only)\n",
2053 np
->fdp
->infname
, cur_node
->fdp
->infname
, np
->name
);
2054 cur_node
->been_warned
= true;
2059 /* Actually add the node */
2060 add_node (np
, dif
< 0 ? &cur_node
->left
: &cur_node
->right
);
2061 } /* if CTAGS mode */
2065 * invalidate_nodes ()
2066 * Scan the node tree and invalidate all nodes pointing to the
2067 * given file description (CTAGS case) or free them (ETAGS case).
2070 invalidate_nodes (fdesc
*badfdp
, node
**npp
)
2079 if (np
->left
!= NULL
)
2080 invalidate_nodes (badfdp
, &np
->left
);
2081 if (np
->fdp
== badfdp
)
2083 if (np
->right
!= NULL
)
2084 invalidate_nodes (badfdp
, &np
->right
);
2088 assert (np
->fdp
!= NULL
);
2089 if (np
->fdp
== badfdp
)
2091 *npp
= np
->left
; /* detach the sublist from the list */
2092 np
->left
= NULL
; /* isolate it */
2093 free_tree (np
); /* free it */
2094 invalidate_nodes (badfdp
, npp
);
2097 invalidate_nodes (badfdp
, &np
->left
);
2102 static int total_size_of_entries (node
*);
2103 static int number_len (long) ATTRIBUTE_CONST
;
2105 /* Length of a non-negative number's decimal representation. */
2107 number_len (long int num
)
2110 while ((num
/= 10) > 0)
2116 * Return total number of characters that put_entries will output for
2117 * the nodes in the linked list at the right of the specified node.
2118 * This count is irrelevant with etags.el since emacs 19.34 at least,
2119 * but is still supplied for backward compatibility.
2122 total_size_of_entries (register node
*np
)
2124 register int total
= 0;
2126 for (; np
!= NULL
; np
= np
->right
)
2129 total
+= strlen (np
->regex
) + 1; /* pat\177 */
2130 if (np
->name
!= NULL
)
2131 total
+= strlen (np
->name
) + 1; /* name\001 */
2132 total
+= number_len ((long) np
->lno
) + 1; /* lno, */
2133 if (np
->cno
!= invalidcharno
) /* cno */
2134 total
+= number_len (np
->cno
);
2135 total
+= 1; /* newline */
2142 put_entries (register node
*np
)
2145 static fdesc
*fdp
= NULL
;
2150 /* Output subentries that precede this one */
2152 put_entries (np
->left
);
2154 /* Output this entry */
2163 fprintf (tagf
, "\f\n%s,%d\n",
2164 fdp
->taggedfname
, total_size_of_entries (np
));
2165 fdp
->written
= true;
2167 fputs (np
->regex
, tagf
);
2168 fputc ('\177', tagf
);
2169 if (np
->name
!= NULL
)
2171 fputs (np
->name
, tagf
);
2172 fputc ('\001', tagf
);
2174 fprintf (tagf
, "%d,", np
->lno
);
2175 if (np
->cno
!= invalidcharno
)
2176 fprintf (tagf
, "%ld", np
->cno
);
2182 if (np
->name
== NULL
)
2183 error ("internal error: NULL name in ctags mode.");
2188 fprintf (stdout
, "%s %s %d\n",
2189 np
->name
, np
->fdp
->taggedfname
, (np
->lno
+ 63) / 64);
2191 fprintf (stdout
, "%-16s %3d %-16s %s\n",
2192 np
->name
, np
->lno
, np
->fdp
->taggedfname
, np
->regex
);
2196 fprintf (tagf
, "%s\t%s\t", np
->name
, np
->fdp
->taggedfname
);
2199 { /* function or #define macro with args */
2200 putc (searchar
, tagf
);
2203 for (sp
= np
->regex
; *sp
; sp
++)
2205 if (*sp
== '\\' || *sp
== searchar
)
2209 putc (searchar
, tagf
);
2212 { /* anything else; text pattern inadequate */
2213 fprintf (tagf
, "%d", np
->lno
);
2218 } /* if this node contains a valid tag */
2220 /* Output subentries that follow this one */
2221 put_entries (np
->right
);
2223 put_entries (np
->left
);
2228 #define C_EXT 0x00fff /* C extensions */
2229 #define C_PLAIN 0x00000 /* C */
2230 #define C_PLPL 0x00001 /* C++ */
2231 #define C_STAR 0x00003 /* C* */
2232 #define C_JAVA 0x00005 /* JAVA */
2233 #define C_AUTO 0x01000 /* C, but switch to C++ if `class' is met */
2234 #define YACC 0x10000 /* yacc file */
2237 * The C symbol tables.
2242 st_C_objprot
, st_C_objimpl
, st_C_objend
,
2244 st_C_ignore
, st_C_attribute
,
2247 st_C_class
, st_C_template
,
2248 st_C_struct
, st_C_extern
, st_C_enum
, st_C_define
, st_C_typedef
2251 /* Feed stuff between (but not including) %[ and %] lines to:
2257 struct C_stab_entry { char *name; int c_ext; enum sym_type type; }
2261 while, 0, st_C_ignore
2262 switch, 0, st_C_ignore
2263 return, 0, st_C_ignore
2264 __attribute__, 0, st_C_attribute
2265 GTY, 0, st_C_attribute
2266 @interface, 0, st_C_objprot
2267 @protocol, 0, st_C_objprot
2268 @implementation,0, st_C_objimpl
2269 @end, 0, st_C_objend
2270 import, (C_JAVA & ~C_PLPL), st_C_ignore
2271 package, (C_JAVA & ~C_PLPL), st_C_ignore
2272 friend, C_PLPL, st_C_ignore
2273 extends, (C_JAVA & ~C_PLPL), st_C_javastruct
2274 implements, (C_JAVA & ~C_PLPL), st_C_javastruct
2275 interface, (C_JAVA & ~C_PLPL), st_C_struct
2276 class, 0, st_C_class
2277 namespace, C_PLPL, st_C_struct
2278 domain, C_STAR, st_C_struct
2279 union, 0, st_C_struct
2280 struct, 0, st_C_struct
2281 extern, 0, st_C_extern
2283 typedef, 0, st_C_typedef
2284 define, 0, st_C_define
2285 undef, 0, st_C_define
2286 operator, C_PLPL, st_C_operator
2287 template, 0, st_C_template
2288 # DEFUN used in emacs, the next three used in glibc (SYSCALL only for mach).
2289 DEFUN, 0, st_C_gnumacro
2290 SYSCALL, 0, st_C_gnumacro
2291 ENTRY, 0, st_C_gnumacro
2292 PSEUDO, 0, st_C_gnumacro
2293 # These are defined inside C functions, so currently they are not met.
2294 # EXFUN used in glibc, DEFVAR_* in emacs.
2295 #EXFUN, 0, st_C_gnumacro
2296 #DEFVAR_, 0, st_C_gnumacro
2298 and replace lines between %< and %> with its output, then:
2299 - remove the #if characterset check
2300 - make in_word_set static and not inline. */
2302 /* C code produced by gperf version 3.0.1 */
2303 /* Command-line: gperf -m 5 */
2304 /* Computed positions: -k'2-3' */
2306 struct C_stab_entry
{ const char *name
; int c_ext
; enum sym_type type
; };
2307 /* maximum key range = 33, duplicates = 0 */
2310 hash (const char *str
, int len
)
2312 static char const asso_values
[] =
2314 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2315 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2316 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2317 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2318 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2319 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2320 35, 35, 35, 35, 35, 35, 35, 35, 35, 3,
2321 26, 35, 35, 35, 35, 35, 35, 35, 27, 35,
2322 35, 35, 35, 24, 0, 35, 35, 35, 35, 0,
2323 35, 35, 35, 35, 35, 1, 35, 16, 35, 6,
2324 23, 0, 0, 35, 22, 0, 35, 35, 5, 0,
2325 0, 15, 1, 35, 6, 35, 8, 19, 35, 16,
2326 4, 5, 35, 35, 35, 35, 35, 35, 35, 35,
2327 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2328 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2329 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2330 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2331 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2332 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
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
2346 hval
+= asso_values
[(unsigned char) str
[2]];
2349 hval
+= asso_values
[(unsigned char) str
[1]];
2355 static struct C_stab_entry
*
2356 in_word_set (register const char *str
, register unsigned int len
)
2360 TOTAL_KEYWORDS
= 33,
2361 MIN_WORD_LENGTH
= 2,
2362 MAX_WORD_LENGTH
= 15,
2367 static struct C_stab_entry wordlist
[] =
2370 {"if", 0, st_C_ignore
},
2371 {"GTY", 0, st_C_attribute
},
2372 {"@end", 0, st_C_objend
},
2373 {"union", 0, st_C_struct
},
2374 {"define", 0, st_C_define
},
2375 {"import", (C_JAVA
& ~C_PLPL
), st_C_ignore
},
2376 {"template", 0, st_C_template
},
2377 {"operator", C_PLPL
, st_C_operator
},
2378 {"@interface", 0, st_C_objprot
},
2379 {"implements", (C_JAVA
& ~C_PLPL
), st_C_javastruct
},
2380 {"friend", C_PLPL
, st_C_ignore
},
2381 {"typedef", 0, st_C_typedef
},
2382 {"return", 0, st_C_ignore
},
2383 {"@implementation",0, st_C_objimpl
},
2384 {"@protocol", 0, st_C_objprot
},
2385 {"interface", (C_JAVA
& ~C_PLPL
), st_C_struct
},
2386 {"extern", 0, st_C_extern
},
2387 {"extends", (C_JAVA
& ~C_PLPL
), st_C_javastruct
},
2388 {"struct", 0, st_C_struct
},
2389 {"domain", C_STAR
, st_C_struct
},
2390 {"switch", 0, st_C_ignore
},
2391 {"enum", 0, st_C_enum
},
2392 {"for", 0, st_C_ignore
},
2393 {"namespace", C_PLPL
, st_C_struct
},
2394 {"class", 0, st_C_class
},
2395 {"while", 0, st_C_ignore
},
2396 {"undef", 0, st_C_define
},
2397 {"package", (C_JAVA
& ~C_PLPL
), st_C_ignore
},
2398 {"__attribute__", 0, st_C_attribute
},
2399 {"SYSCALL", 0, st_C_gnumacro
},
2400 {"ENTRY", 0, st_C_gnumacro
},
2401 {"PSEUDO", 0, st_C_gnumacro
},
2402 {"DEFUN", 0, st_C_gnumacro
}
2405 if (len
<= MAX_WORD_LENGTH
&& len
>= MIN_WORD_LENGTH
)
2407 int key
= hash (str
, len
);
2409 if (key
<= MAX_HASH_VALUE
&& key
>= 0)
2411 const char *s
= wordlist
[key
].name
;
2413 if (*str
== *s
&& !strncmp (str
+ 1, s
+ 1, len
- 1) && s
[len
] == '\0')
2414 return &wordlist
[key
];
2421 static enum sym_type
2422 C_symtype (char *str
, int len
, int c_ext
)
2424 register struct C_stab_entry
*se
= in_word_set (str
, len
);
2426 if (se
== NULL
|| (se
->c_ext
&& !(c_ext
& se
->c_ext
)))
2433 * Ignoring __attribute__ ((list))
2435 static bool inattribute
; /* looking at an __attribute__ construct */
2438 * C functions and variables are recognized using a simple
2439 * finite automaton. fvdef is its state variable.
2443 fvnone
, /* nothing seen */
2444 fdefunkey
, /* Emacs DEFUN keyword seen */
2445 fdefunname
, /* Emacs DEFUN name seen */
2446 foperator
, /* func: operator keyword seen (cplpl) */
2447 fvnameseen
, /* function or variable name seen */
2448 fstartlist
, /* func: just after open parenthesis */
2449 finlist
, /* func: in parameter list */
2450 flistseen
, /* func: after parameter list */
2451 fignore
, /* func: before open brace */
2452 vignore
/* var-like: ignore until ';' */
2455 static bool fvextern
; /* func or var: extern keyword seen; */
2458 * typedefs are recognized using a simple finite automaton.
2459 * typdef is its state variable.
2463 tnone
, /* nothing seen */
2464 tkeyseen
, /* typedef keyword seen */
2465 ttypeseen
, /* defined type seen */
2466 tinbody
, /* inside typedef body */
2467 tend
, /* just before typedef tag */
2468 tignore
/* junk after typedef tag */
2472 * struct-like structures (enum, struct and union) are recognized
2473 * using another simple finite automaton. `structdef' is its state
2478 snone
, /* nothing seen yet,
2479 or in struct body if bracelev > 0 */
2480 skeyseen
, /* struct-like keyword seen */
2481 stagseen
, /* struct-like tag seen */
2482 scolonseen
/* colon seen after struct-like tag */
2486 * When objdef is different from onone, objtag is the name of the class.
2488 static const char *objtag
= "<uninited>";
2491 * Yet another little state machine to deal with preprocessor lines.
2495 dnone
, /* nothing seen */
2496 dsharpseen
, /* '#' seen as first char on line */
2497 ddefineseen
, /* '#' and 'define' seen */
2498 dignorerest
/* ignore rest of line */
2502 * State machine for Objective C protocols and implementations.
2503 * Idea by Tom R.Hageman <tom@basil.icce.rug.nl> (1995)
2507 onone
, /* nothing seen */
2508 oprotocol
, /* @interface or @protocol seen */
2509 oimplementation
, /* @implementations seen */
2510 otagseen
, /* class name seen */
2511 oparenseen
, /* parenthesis before category seen */
2512 ocatseen
, /* category name seen */
2513 oinbody
, /* in @implementation body */
2514 omethodsign
, /* in @implementation body, after +/- */
2515 omethodtag
, /* after method name */
2516 omethodcolon
, /* after method colon */
2517 omethodparm
, /* after method parameter */
2518 oignore
/* wait for @end */
2523 * Use this structure to keep info about the token read, and how it
2524 * should be tagged. Used by the make_C_tag function to build a tag.
2528 char *line
; /* string containing the token */
2529 int offset
; /* where the token starts in LINE */
2530 int length
; /* token length */
2532 The previous members can be used to pass strings around for generic
2533 purposes. The following ones specifically refer to creating tags. In this
2534 case the token contained here is the pattern that will be used to create a
2537 bool valid
; /* do not create a tag; the token should be
2538 invalidated whenever a state machine is
2539 reset prematurely */
2540 bool named
; /* create a named tag */
2541 int lineno
; /* source line number of tag */
2542 long linepos
; /* source char number of tag */
2543 } token
; /* latest token read */
2546 * Variables and functions for dealing with nested structures.
2547 * Idea by Mykola Dzyuba <mdzyuba@yahoo.com> (2001)
2549 static void pushclass_above (int, char *, int);
2550 static void popclass_above (int);
2551 static void write_classname (linebuffer
*, const char *qualifier
);
2554 char **cname
; /* nested class names */
2555 int *bracelev
; /* nested class brace level */
2556 int nl
; /* class nesting level (elements used) */
2557 int size
; /* length of the array */
2558 } cstack
; /* stack for nested declaration tags */
2559 /* Current struct nesting depth (namespace, class, struct, union, enum). */
2560 #define nestlev (cstack.nl)
2561 /* After struct keyword or in struct body, not inside a nested function. */
2562 #define instruct (structdef == snone && nestlev > 0 \
2563 && bracelev == cstack.bracelev[nestlev-1] + 1)
2566 pushclass_above (int bracelev
, char *str
, int len
)
2570 popclass_above (bracelev
);
2572 if (nl
>= cstack
.size
)
2574 int size
= cstack
.size
*= 2;
2575 xrnew (cstack
.cname
, size
, char *);
2576 xrnew (cstack
.bracelev
, size
, int);
2578 assert (nl
== 0 || cstack
.bracelev
[nl
-1] < bracelev
);
2579 cstack
.cname
[nl
] = (str
== NULL
) ? NULL
: savenstr (str
, len
);
2580 cstack
.bracelev
[nl
] = bracelev
;
2585 popclass_above (int bracelev
)
2589 for (nl
= cstack
.nl
- 1;
2590 nl
>= 0 && cstack
.bracelev
[nl
] >= bracelev
;
2593 free (cstack
.cname
[nl
]);
2599 write_classname (linebuffer
*cn
, const char *qualifier
)
2602 int qlen
= strlen (qualifier
);
2604 if (cstack
.nl
== 0 || cstack
.cname
[0] == NULL
)
2608 cn
->buffer
[0] = '\0';
2612 len
= strlen (cstack
.cname
[0]);
2613 linebuffer_setlen (cn
, len
);
2614 strcpy (cn
->buffer
, cstack
.cname
[0]);
2616 for (i
= 1; i
< cstack
.nl
; i
++)
2618 char *s
= cstack
.cname
[i
];
2621 linebuffer_setlen (cn
, len
+ qlen
+ strlen (s
));
2622 len
+= sprintf (cn
->buffer
+ len
, "%s%s", qualifier
, s
);
2627 static bool consider_token (char *, int, int, int *, int, int, bool *);
2628 static void make_C_tag (bool);
2632 * checks to see if the current token is at the start of a
2633 * function or variable, or corresponds to a typedef, or
2634 * is a struct/union/enum tag, or #define, or an enum constant.
2636 * *IS_FUNC_OR_VAR gets true if the token is a function or #define macro
2637 * with args. C_EXTP points to which language we are looking at.
2648 consider_token (char *str
, int len
, int c
, int *c_extp
,
2649 int bracelev
, int parlev
, bool *is_func_or_var
)
2650 /* IN: token pointer */
2651 /* IN: token length */
2652 /* IN: first char after the token */
2653 /* IN, OUT: C extensions mask */
2654 /* IN: brace level */
2655 /* IN: parenthesis level */
2656 /* OUT: function or variable found */
2658 /* When structdef is stagseen, scolonseen, or snone with bracelev > 0,
2659 structtype is the type of the preceding struct-like keyword, and
2660 structbracelev is the brace level where it has been seen. */
2661 static enum sym_type structtype
;
2662 static int structbracelev
;
2663 static enum sym_type toktype
;
2666 toktype
= C_symtype (str
, len
, *c_extp
);
2669 * Skip __attribute__
2671 if (toktype
== st_C_attribute
)
2678 * Advance the definedef state machine.
2683 /* We're not on a preprocessor line. */
2684 if (toktype
== st_C_gnumacro
)
2691 if (toktype
== st_C_define
)
2693 definedef
= ddefineseen
;
2697 definedef
= dignorerest
;
2702 * Make a tag for any macro, unless it is a constant
2703 * and constantypedefs is false.
2705 definedef
= dignorerest
;
2706 *is_func_or_var
= (c
== '(');
2707 if (!*is_func_or_var
&& !constantypedefs
)
2714 error ("internal error: definedef value.");
2723 if (toktype
== st_C_typedef
)
2743 if (structdef
== snone
&& fvdef
== fvnone
)
2762 case st_C_javastruct
:
2763 if (structdef
== stagseen
)
2764 structdef
= scolonseen
;
2768 if ((*c_extp
& C_AUTO
) /* automatic detection of C++ language */
2770 && definedef
== dnone
&& structdef
== snone
2771 && typdef
== tnone
&& fvdef
== fvnone
)
2772 *c_extp
= (*c_extp
| C_PLPL
) & ~C_AUTO
;
2773 if (toktype
== st_C_template
)
2780 && (typdef
== tkeyseen
2781 || (typedefs_or_cplusplus
&& structdef
== snone
)))
2783 structdef
= skeyseen
;
2784 structtype
= toktype
;
2785 structbracelev
= bracelev
;
2786 if (fvdef
== fvnameseen
)
2792 if (structdef
== skeyseen
)
2794 structdef
= stagseen
;
2798 if (typdef
!= tnone
)
2801 /* Detect Objective C constructs. */
2811 objdef
= oimplementation
;
2815 case oimplementation
:
2816 /* Save the class tag for functions or variables defined inside. */
2817 objtag
= savenstr (str
, len
);
2821 /* Save the class tag for categories. */
2822 objtag
= savenstr (str
, len
);
2824 *is_func_or_var
= true;
2828 *is_func_or_var
= true;
2836 objdef
= omethodtag
;
2837 linebuffer_setlen (&token_name
, len
);
2838 memcpy (token_name
.buffer
, str
, len
);
2839 token_name
.buffer
[len
] = '\0';
2845 objdef
= omethodparm
;
2850 int oldlen
= token_name
.len
;
2852 objdef
= omethodtag
;
2853 linebuffer_setlen (&token_name
, oldlen
+ len
);
2854 memcpy (token_name
.buffer
+ oldlen
, str
, len
);
2855 token_name
.buffer
[oldlen
+ len
] = '\0';
2860 if (toktype
== st_C_objend
)
2862 /* Memory leakage here: the string pointed by objtag is
2863 never released, because many tests would be needed to
2864 avoid breaking on incorrect input code. The amount of
2865 memory leaked here is the sum of the lengths of the
2873 /* A function, variable or enum constant? */
2895 *is_func_or_var
= true;
2899 && structdef
== snone
2900 && structtype
== st_C_enum
&& bracelev
> structbracelev
2901 /* Don't tag tokens in expressions that assign values to enum
2903 && fvdef
!= vignore
)
2904 return true; /* enum constant */
2910 fvdef
= fdefunname
; /* GNU macro */
2911 *is_func_or_var
= true;
2919 if ((strneq (str
, "asm", 3) && endtoken (str
[3]))
2920 || (strneq (str
, "__asm__", 7) && endtoken (str
[7])))
2929 if (len
>= 10 && strneq (str
+len
-10, "::operator", 10))
2931 if (*c_extp
& C_AUTO
) /* automatic detection of C++ */
2932 *c_extp
= (*c_extp
| C_PLPL
) & ~C_AUTO
;
2934 *is_func_or_var
= true;
2937 if (bracelev
> 0 && !instruct
)
2939 fvdef
= fvnameseen
; /* function or variable */
2940 *is_func_or_var
= true;
2951 * C_entries often keeps pointers to tokens or lines which are older than
2952 * the line currently read. By keeping two line buffers, and switching
2953 * them at end of line, it is possible to use those pointers.
2961 #define current_lb_is_new (newndx == curndx)
2962 #define switch_line_buffers() (curndx = 1 - curndx)
2964 #define curlb (lbs[curndx].lb)
2965 #define newlb (lbs[newndx].lb)
2966 #define curlinepos (lbs[curndx].linepos)
2967 #define newlinepos (lbs[newndx].linepos)
2969 #define plainc ((c_ext & C_EXT) == C_PLAIN)
2970 #define cplpl (c_ext & C_PLPL)
2971 #define cjava ((c_ext & C_JAVA) == C_JAVA)
2973 #define CNL_SAVE_DEFINEDEF() \
2975 curlinepos = charno; \
2976 readline (&curlb, inf); \
2977 lp = curlb.buffer; \
2984 CNL_SAVE_DEFINEDEF (); \
2985 if (savetoken.valid) \
2987 token = savetoken; \
2988 savetoken.valid = false; \
2990 definedef = dnone; \
2995 make_C_tag (bool isfun
)
2997 /* This function is never called when token.valid is false, but
2998 we must protect against invalid input or internal errors. */
3000 make_tag (token_name
.buffer
, token_name
.len
, isfun
, token
.line
,
3001 token
.offset
+token
.length
+1, token
.lineno
, token
.linepos
);
3003 { /* this branch is optimized away if !DEBUG */
3004 make_tag (concat ("INVALID TOKEN:-->", token_name
.buffer
, ""),
3005 token_name
.len
+ 17, isfun
, token
.line
,
3006 token
.offset
+token
.length
+1, token
.lineno
, token
.linepos
);
3007 error ("INVALID TOKEN");
3010 token
.valid
= false;
3014 perhaps_more_input (FILE *inf
)
3016 return !feof (inf
) && !ferror (inf
);
3022 * This routine finds functions, variables, typedefs,
3023 * #define's, enum constants and struct/union/enum definitions in
3024 * C syntax and adds them to the list.
3027 C_entries (int c_ext
, FILE *inf
)
3028 /* extension of C */
3031 register char c
; /* latest char read; '\0' for end of line */
3032 register char *lp
; /* pointer one beyond the character `c' */
3033 int curndx
, newndx
; /* indices for current and new lb */
3034 register int tokoff
; /* offset in line of start of current token */
3035 register int toklen
; /* length of current token */
3036 const char *qualifier
; /* string used to qualify names */
3037 int qlen
; /* length of qualifier */
3038 int bracelev
; /* current brace level */
3039 int bracketlev
; /* current bracket level */
3040 int parlev
; /* current parenthesis level */
3041 int attrparlev
; /* __attribute__ parenthesis level */
3042 int templatelev
; /* current template level */
3043 int typdefbracelev
; /* bracelev where a typedef struct body begun */
3044 bool incomm
, inquote
, inchar
, quotednl
, midtoken
;
3045 bool yacc_rules
; /* in the rules part of a yacc file */
3046 struct tok savetoken
= {0}; /* token saved during preprocessor handling */
3049 linebuffer_init (&lbs
[0].lb
);
3050 linebuffer_init (&lbs
[1].lb
);
3051 if (cstack
.size
== 0)
3053 cstack
.size
= (DEBUG
) ? 1 : 4;
3055 cstack
.cname
= xnew (cstack
.size
, char *);
3056 cstack
.bracelev
= xnew (cstack
.size
, int);
3059 tokoff
= toklen
= typdefbracelev
= 0; /* keep compiler quiet */
3060 curndx
= newndx
= 0;
3064 fvdef
= fvnone
; fvextern
= false; typdef
= tnone
;
3065 structdef
= snone
; definedef
= dnone
; objdef
= onone
;
3067 midtoken
= inquote
= inchar
= incomm
= quotednl
= false;
3068 token
.valid
= savetoken
.valid
= false;
3069 bracelev
= bracketlev
= parlev
= attrparlev
= templatelev
= 0;
3071 { qualifier
= "."; qlen
= 1; }
3073 { qualifier
= "::"; qlen
= 2; }
3076 while (perhaps_more_input (inf
))
3081 /* If we are at the end of the line, the next character is a
3082 '\0'; do not skip it, because it is what tells us
3083 to read the next line. */
3104 /* Newlines inside comments do not end macro definitions in
3106 CNL_SAVE_DEFINEDEF ();
3119 /* Newlines inside strings do not end macro definitions
3120 in traditional cpp, even though compilers don't
3121 usually accept them. */
3122 CNL_SAVE_DEFINEDEF ();
3132 /* Hmmm, something went wrong. */
3168 if (fvdef
!= finlist
&& fvdef
!= fignore
&& fvdef
!= vignore
)
3183 else if (/* cplpl && */ *lp
== '/')
3189 if ((c_ext
& YACC
) && *lp
== '%')
3191 /* Entering or exiting rules section in yacc file. */
3193 definedef
= dnone
; fvdef
= fvnone
; fvextern
= false;
3194 typdef
= tnone
; structdef
= snone
;
3195 midtoken
= inquote
= inchar
= incomm
= quotednl
= false;
3197 yacc_rules
= !yacc_rules
;
3203 if (definedef
== dnone
)
3206 bool cpptoken
= true;
3208 /* Look back on this line. If all blanks, or nonblanks
3209 followed by an end of comment, this is a preprocessor
3211 for (cp
= newlb
.buffer
; cp
< lp
-1; cp
++)
3212 if (!c_isspace (*cp
))
3214 if (*cp
== '*' && cp
[1] == '/')
3224 definedef
= dsharpseen
;
3225 /* This is needed for tagging enum values: when there are
3226 preprocessor conditionals inside the enum, we need to
3227 reset the value of fvdef so that the next enum value is
3228 tagged even though the one before it did not end in a
3230 if (fvdef
== vignore
&& instruct
&& parlev
== 0)
3232 if (strneq (cp
, "#if", 3) || strneq (cp
, "#el", 3))
3236 } /* if (definedef == dnone) */
3247 CNL_SAVE_DEFINEDEF ();
3254 /* Consider token only if some involved conditions are satisfied. */
3255 if (typdef
!= tignore
3256 && definedef
!= dignorerest
3259 && (definedef
!= dnone
3260 || structdef
!= scolonseen
)
3267 if (c
== ':' && *lp
== ':' && begtoken (lp
[1]))
3268 /* This handles :: in the middle,
3269 but not at the beginning of an identifier.
3270 Also, space-separated :: is not recognized. */
3272 if (c_ext
& C_AUTO
) /* automatic detection of C++ */
3273 c_ext
= (c_ext
| C_PLPL
) & ~C_AUTO
;
3277 goto still_in_token
;
3281 bool funorvar
= false;
3284 || consider_token (newlb
.buffer
+ tokoff
, toklen
, c
,
3285 &c_ext
, bracelev
, parlev
,
3288 if (fvdef
== foperator
)
3291 lp
= skip_spaces (lp
-1);
3295 && !c_isspace (*lp
) && *lp
!= '(')
3298 toklen
+= lp
- oldlp
;
3300 token
.named
= false;
3302 && nestlev
> 0 && definedef
== dnone
)
3303 /* in struct body */
3306 write_classname (&token_name
, qualifier
);
3307 len
= token_name
.len
;
3308 linebuffer_setlen (&token_name
, len
+qlen
+toklen
);
3309 sprintf (token_name
.buffer
+ len
, "%s%.*s",
3310 qualifier
, toklen
, newlb
.buffer
+ tokoff
);
3313 else if (objdef
== ocatseen
)
3314 /* Objective C category */
3316 int len
= strlen (objtag
) + 2 + toklen
;
3317 linebuffer_setlen (&token_name
, len
);
3318 sprintf (token_name
.buffer
, "%s(%.*s)",
3319 objtag
, toklen
, newlb
.buffer
+ tokoff
);
3322 else if (objdef
== omethodtag
3323 || objdef
== omethodparm
)
3324 /* Objective C method */
3328 else if (fvdef
== fdefunname
)
3329 /* GNU DEFUN and similar macros */
3331 bool defun
= (newlb
.buffer
[tokoff
] == 'F');
3335 /* Rewrite the tag so that emacs lisp DEFUNs
3336 can be found by their elisp name */
3342 linebuffer_setlen (&token_name
, len
);
3343 memcpy (token_name
.buffer
,
3344 newlb
.buffer
+ off
, len
);
3345 token_name
.buffer
[len
] = '\0';
3348 if (token_name
.buffer
[len
] == '_')
3349 token_name
.buffer
[len
] = '-';
3350 token
.named
= defun
;
3354 linebuffer_setlen (&token_name
, toklen
);
3355 memcpy (token_name
.buffer
,
3356 newlb
.buffer
+ tokoff
, toklen
);
3357 token_name
.buffer
[toklen
] = '\0';
3358 /* Name macros and members. */
3359 token
.named
= (structdef
== stagseen
3360 || typdef
== ttypeseen
3363 && definedef
== dignorerest
)
3365 && definedef
== dnone
3366 && structdef
== snone
3369 token
.lineno
= lineno
;
3370 token
.offset
= tokoff
;
3371 token
.length
= toklen
;
3372 token
.line
= newlb
.buffer
;
3373 token
.linepos
= newlinepos
;
3376 if (definedef
== dnone
3377 && (fvdef
== fvnameseen
3378 || fvdef
== foperator
3379 || structdef
== stagseen
3381 || typdef
== ttypeseen
3382 || objdef
!= onone
))
3384 if (current_lb_is_new
)
3385 switch_line_buffers ();
3387 else if (definedef
!= dnone
3388 || fvdef
== fdefunname
3390 make_C_tag (funorvar
);
3392 else /* not yacc and consider_token failed */
3394 if (inattribute
&& fvdef
== fignore
)
3396 /* We have just met __attribute__ after a
3397 function parameter list: do not tag the
3404 } /* if (endtoken (c)) */
3405 else if (intoken (c
))
3411 } /* if (midtoken) */
3412 else if (begtoken (c
))
3420 /* This prevents tagging fb in
3421 void (__attribute__((noreturn)) *fb) (void);
3422 Fixing this is not easy and not very important. */
3426 if (plainc
|| declarations
)
3428 make_C_tag (true); /* a function */
3433 if (structdef
== stagseen
&& !cjava
)
3435 popclass_above (bracelev
);
3443 if (!yacc_rules
|| lp
== newlb
.buffer
+ 1)
3445 tokoff
= lp
- 1 - newlb
.buffer
;
3450 } /* if (begtoken) */
3451 } /* if must look at token */
3454 /* Detect end of line, colon, comma, semicolon and various braces
3455 after having handled a token.*/
3461 if (yacc_rules
&& token
.offset
== 0 && token
.valid
)
3463 make_C_tag (false); /* a yacc function */
3466 if (definedef
!= dnone
)
3472 make_C_tag (true); /* an Objective C class */
3476 objdef
= omethodcolon
;
3477 int toklen
= token_name
.len
;
3478 linebuffer_setlen (&token_name
, toklen
+ 1);
3479 strcpy (token_name
.buffer
+ toklen
, ":");
3482 if (structdef
== stagseen
)
3484 structdef
= scolonseen
;
3487 /* Should be useless, but may be work as a safety net. */
3488 if (cplpl
&& fvdef
== flistseen
)
3490 make_C_tag (true); /* a function */
3496 if (definedef
!= dnone
|| inattribute
)
3502 make_C_tag (false); /* a typedef */
3512 if (typdef
== tignore
|| cplpl
)
3516 if ((globals
&& bracelev
== 0 && (!fvextern
|| declarations
))
3517 || (members
&& instruct
))
3518 make_C_tag (false); /* a variable */
3521 token
.valid
= false;
3525 && (cplpl
|| !instruct
)
3526 && (typdef
== tnone
|| (typdef
!= tignore
&& instruct
)))
3528 && plainc
&& instruct
))
3529 make_C_tag (true); /* a function */
3535 && cplpl
&& structdef
== stagseen
)
3536 make_C_tag (false); /* forward declaration */
3538 token
.valid
= false;
3539 } /* switch (fvdef) */
3545 if (structdef
== stagseen
)
3549 if (definedef
!= dnone
|| inattribute
)
3555 make_C_tag (true); /* an Objective C method */
3568 if (instruct
&& parlev
== 0)
3579 && (!fvextern
|| declarations
))
3580 || (members
&& instruct
)))
3581 make_C_tag (false); /* a variable */
3584 if ((declarations
&& typdef
== tnone
&& !instruct
)
3585 || (members
&& typdef
!= tignore
&& instruct
))
3587 make_C_tag (true); /* a function */
3590 else if (!declarations
)
3592 token
.valid
= false;
3597 if (structdef
== stagseen
)
3601 if (definedef
!= dnone
|| inattribute
)
3603 if (structdef
== stagseen
)
3610 make_C_tag (false); /* a typedef */
3622 if ((members
&& bracelev
== 1)
3623 || (globals
&& bracelev
== 0
3624 && (!fvextern
|| declarations
)))
3625 make_C_tag (false); /* a variable */
3639 if (definedef
!= dnone
)
3641 if (objdef
== otagseen
&& parlev
== 0)
3642 objdef
= oparenseen
;
3646 if (typdef
== ttypeseen
3650 /* This handles constructs like:
3651 typedef void OperatorFun (int fun); */
3670 if (--attrparlev
== 0)
3671 inattribute
= false;
3674 if (definedef
!= dnone
)
3676 if (objdef
== ocatseen
&& parlev
== 1)
3678 make_C_tag (true); /* an Objective C category */
3692 || typdef
== ttypeseen
))
3695 make_C_tag (false); /* a typedef */
3698 else if (parlev
< 0) /* can happen due to ill-conceived #if's. */
3702 if (definedef
!= dnone
)
3704 if (typdef
== ttypeseen
)
3706 /* Whenever typdef is set to tinbody (currently only
3707 here), typdefbracelev should be set to bracelev. */
3709 typdefbracelev
= bracelev
;
3714 make_C_tag (true); /* a function */
3723 make_C_tag (true); /* an Objective C class */
3728 make_C_tag (true); /* an Objective C method */
3732 /* Neutralize `extern "C" {' grot. */
3733 if (bracelev
== 0 && structdef
== snone
&& nestlev
== 0
3741 case skeyseen
: /* unnamed struct */
3742 pushclass_above (bracelev
, NULL
, 0);
3745 case stagseen
: /* named struct or enum */
3746 case scolonseen
: /* a class */
3747 pushclass_above (bracelev
,token
.line
+token
.offset
, token
.length
);
3749 make_C_tag (false); /* a struct or enum */
3755 if (definedef
!= dnone
)
3757 if (fvdef
== fstartlist
)
3759 fvdef
= fvnone
; /* avoid tagging `foo' in `foo (*bar()) ()' */
3760 token
.valid
= false;
3764 if (definedef
!= dnone
)
3767 if (!ignoreindent
&& lp
== newlb
.buffer
+ 1)
3770 token
.valid
= false; /* unexpected value, token unreliable */
3771 bracelev
= 0; /* reset brace level if first column */
3772 parlev
= 0; /* also reset paren level, just in case... */
3774 else if (bracelev
< 0)
3776 token
.valid
= false; /* something gone amiss, token unreliable */
3779 if (bracelev
== 0 && fvdef
== vignore
)
3780 fvdef
= fvnone
; /* end of function */
3781 popclass_above (bracelev
);
3783 /* Only if typdef == tinbody is typdefbracelev significant. */
3784 if (typdef
== tinbody
&& bracelev
<= typdefbracelev
)
3786 assert (bracelev
== typdefbracelev
);
3791 if (definedef
!= dnone
)
3801 if ((members
&& bracelev
== 1)
3802 || (globals
&& bracelev
== 0 && (!fvextern
|| declarations
)))
3803 make_C_tag (false); /* a variable */
3811 && (structdef
== stagseen
|| fvdef
== fvnameseen
))
3818 if (templatelev
> 0)
3826 if (objdef
== oinbody
&& bracelev
== 0)
3828 objdef
= omethodsign
;
3833 case '#': case '~': case '&': case '%': case '/':
3834 case '|': case '^': case '!': case '.': case '?':
3835 if (definedef
!= dnone
)
3837 /* These surely cannot follow a function tag in C. */
3850 if (objdef
== otagseen
)
3852 make_C_tag (true); /* an Objective C class */
3855 /* If a macro spans multiple lines don't reset its state. */
3857 CNL_SAVE_DEFINEDEF ();
3863 } /* while not eof */
3865 free (lbs
[0].lb
.buffer
);
3866 free (lbs
[1].lb
.buffer
);
3870 * Process either a C++ file or a C file depending on the setting
3874 default_C_entries (FILE *inf
)
3876 C_entries (cplusplus
? C_PLPL
: C_AUTO
, inf
);
3879 /* Always do plain C. */
3881 plain_C_entries (FILE *inf
)
3886 /* Always do C++. */
3888 Cplusplus_entries (FILE *inf
)
3890 C_entries (C_PLPL
, inf
);
3893 /* Always do Java. */
3895 Cjava_entries (FILE *inf
)
3897 C_entries (C_JAVA
, inf
);
3902 Cstar_entries (FILE *inf
)
3904 C_entries (C_STAR
, inf
);
3907 /* Always do Yacc. */
3909 Yacc_entries (FILE *inf
)
3911 C_entries (YACC
, inf
);
3915 /* Useful macros. */
3916 #define LOOP_ON_INPUT_LINES(file_pointer, line_buffer, char_pointer) \
3917 while (perhaps_more_input (file_pointer) \
3918 && (readline (&(line_buffer), file_pointer), \
3919 (char_pointer) = (line_buffer).buffer, \
3922 #define LOOKING_AT(cp, kw) /* kw is the keyword, a literal string */ \
3923 ((assert ("" kw), true) /* syntax error if not a literal string */ \
3924 && strneq ((cp), kw, sizeof (kw)-1) /* cp points at kw */ \
3925 && notinname ((cp)[sizeof (kw)-1]) /* end of kw */ \
3926 && ((cp) = skip_spaces ((cp)+sizeof (kw)-1))) /* skip spaces */
3928 /* Similar to LOOKING_AT but does not use notinname, does not skip */
3929 #define LOOKING_AT_NOCASE(cp, kw) /* the keyword is a literal string */ \
3930 ((assert ("" kw), true) /* syntax error if not a literal string */ \
3931 && strncaseeq ((cp), kw, sizeof (kw)-1) /* cp points at kw */ \
3932 && ((cp) += sizeof (kw)-1)) /* skip spaces */
3935 * Read a file, but do no processing. This is used to do regexp
3936 * matching on files that have no language defined.
3939 just_read_file (FILE *inf
)
3941 while (perhaps_more_input (inf
))
3942 readline (&lb
, inf
);
3946 /* Fortran parsing */
3948 static void F_takeprec (void);
3949 static void F_getit (FILE *);
3954 dbp
= skip_spaces (dbp
);
3958 dbp
= skip_spaces (dbp
);
3959 if (strneq (dbp
, "(*)", 3))
3964 if (!c_isdigit (*dbp
))
3966 --dbp
; /* force failure */
3971 while (c_isdigit (*dbp
));
3979 dbp
= skip_spaces (dbp
);
3982 readline (&lb
, inf
);
3987 dbp
= skip_spaces (dbp
);
3989 if (!c_isalpha (*dbp
) && *dbp
!= '_' && *dbp
!= '$')
3991 for (cp
= dbp
+ 1; *cp
!= '\0' && intoken (*cp
); cp
++)
3993 make_tag (dbp
, cp
-dbp
, true,
3994 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
3999 Fortran_functions (FILE *inf
)
4001 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
4004 dbp
++; /* Ratfor escape to fortran */
4005 dbp
= skip_spaces (dbp
);
4009 if (LOOKING_AT_NOCASE (dbp
, "recursive"))
4010 dbp
= skip_spaces (dbp
);
4012 if (LOOKING_AT_NOCASE (dbp
, "pure"))
4013 dbp
= skip_spaces (dbp
);
4015 if (LOOKING_AT_NOCASE (dbp
, "elemental"))
4016 dbp
= skip_spaces (dbp
);
4018 switch (c_tolower (*dbp
))
4021 if (nocase_tail ("integer"))
4025 if (nocase_tail ("real"))
4029 if (nocase_tail ("logical"))
4033 if (nocase_tail ("complex") || nocase_tail ("character"))
4037 if (nocase_tail ("double"))
4039 dbp
= skip_spaces (dbp
);
4042 if (nocase_tail ("precision"))
4048 dbp
= skip_spaces (dbp
);
4051 switch (c_tolower (*dbp
))
4054 if (nocase_tail ("function"))
4058 if (nocase_tail ("subroutine"))
4062 if (nocase_tail ("entry"))
4066 if (nocase_tail ("blockdata") || nocase_tail ("block data"))
4068 dbp
= skip_spaces (dbp
);
4069 if (*dbp
== '\0') /* assume un-named */
4070 make_tag ("blockdata", 9, true,
4071 lb
.buffer
, dbp
- lb
.buffer
, lineno
, linecharno
);
4073 F_getit (inf
); /* look for name */
4084 * Philippe Waroquiers (1998)
4087 /* Once we are positioned after an "interesting" keyword, let's get
4088 the real tag value necessary. */
4090 Ada_getit (FILE *inf
, const char *name_qualifier
)
4096 while (perhaps_more_input (inf
))
4098 dbp
= skip_spaces (dbp
);
4100 || (dbp
[0] == '-' && dbp
[1] == '-'))
4102 readline (&lb
, inf
);
4105 switch (c_tolower (*dbp
))
4108 if (nocase_tail ("body"))
4110 /* Skipping body of procedure body or package body or ....
4111 resetting qualifier to body instead of spec. */
4112 name_qualifier
= "/b";
4117 /* Skipping type of task type or protected type ... */
4118 if (nocase_tail ("type"))
4125 for (cp
= dbp
; *cp
!= '\0' && *cp
!= '"'; cp
++)
4130 dbp
= skip_spaces (dbp
);
4132 c_isalnum (*cp
) || *cp
== '_' || *cp
== '.';
4140 name
= concat (dbp
, name_qualifier
, "");
4142 make_tag (name
, strlen (name
), true,
4143 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4152 Ada_funcs (FILE *inf
)
4154 bool inquote
= false;
4155 bool skip_till_semicolumn
= false;
4157 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
4159 while (*dbp
!= '\0')
4161 /* Skip a string i.e. "abcd". */
4162 if (inquote
|| (*dbp
== '"'))
4164 dbp
= strchr (dbp
+ !inquote
, '"');
4169 continue; /* advance char */
4174 break; /* advance line */
4178 /* Skip comments. */
4179 if (dbp
[0] == '-' && dbp
[1] == '-')
4180 break; /* advance line */
4182 /* Skip character enclosed in single quote i.e. 'a'
4183 and skip single quote starting an attribute i.e. 'Image. */
4192 if (skip_till_semicolumn
)
4195 skip_till_semicolumn
= false;
4197 continue; /* advance char */
4200 /* Search for beginning of a token. */
4201 if (!begtoken (*dbp
))
4204 continue; /* advance char */
4207 /* We are at the beginning of a token. */
4208 switch (c_tolower (*dbp
))
4211 if (!packages_only
&& nocase_tail ("function"))
4212 Ada_getit (inf
, "/f");
4214 break; /* from switch */
4215 continue; /* advance char */
4217 if (!packages_only
&& nocase_tail ("procedure"))
4218 Ada_getit (inf
, "/p");
4219 else if (nocase_tail ("package"))
4220 Ada_getit (inf
, "/s");
4221 else if (nocase_tail ("protected")) /* protected type */
4222 Ada_getit (inf
, "/t");
4224 break; /* from switch */
4225 continue; /* advance char */
4228 if (typedefs
&& !packages_only
&& nocase_tail ("use"))
4230 /* when tagging types, avoid tagging use type Pack.Typename;
4231 for this, we will skip everything till a ; */
4232 skip_till_semicolumn
= true;
4233 continue; /* advance char */
4237 if (!packages_only
&& nocase_tail ("task"))
4238 Ada_getit (inf
, "/k");
4239 else if (typedefs
&& !packages_only
&& nocase_tail ("type"))
4241 Ada_getit (inf
, "/t");
4242 while (*dbp
!= '\0')
4246 break; /* from switch */
4247 continue; /* advance char */
4250 /* Look for the end of the token. */
4251 while (!endtoken (*dbp
))
4254 } /* advance char */
4255 } /* advance line */
4260 * Unix and microcontroller assembly tag handling
4261 * Labels: /^[a-zA-Z_.$][a-zA_Z0-9_.$]*[: ^I^J]/
4262 * Idea by Bob Weiner, Motorola Inc. (1994)
4265 Asm_labels (FILE *inf
)
4269 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4271 /* If first char is alphabetic or one of [_.$], test for colon
4272 following identifier. */
4273 if (c_isalpha (*cp
) || *cp
== '_' || *cp
== '.' || *cp
== '$')
4275 /* Read past label. */
4277 while (c_isalnum (*cp
) || *cp
== '_' || *cp
== '.' || *cp
== '$')
4279 if (*cp
== ':' || c_isspace (*cp
))
4280 /* Found end of label, so copy it and add it to the table. */
4281 make_tag (lb
.buffer
, cp
- lb
.buffer
, true,
4282 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4290 * Perl sub names: /^sub[ \t\n]+[^ \t\n{]+/
4291 * /^use constant[ \t\n]+[^ \t\n{=,;]+/
4292 * Perl variable names: /^(my|local).../
4293 * Original code by Bart Robinson <lomew@cs.utah.edu> (1995)
4294 * Additions by Michael Ernst <mernst@alum.mit.edu> (1997)
4295 * Ideas by Kai Großjohann <Kai.Grossjohann@CS.Uni-Dortmund.DE> (2001)
4298 Perl_functions (FILE *inf
)
4300 char *package
= savestr ("main"); /* current package name */
4303 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4305 cp
= skip_spaces (cp
);
4307 if (LOOKING_AT (cp
, "package"))
4310 get_tag (cp
, &package
);
4312 else if (LOOKING_AT (cp
, "sub"))
4318 while (!notinname (*cp
))
4321 continue; /* nothing found */
4322 pos
= strchr (sp
, ':');
4323 if (pos
&& pos
< cp
&& pos
[1] == ':')
4324 /* The name is already qualified. */
4325 make_tag (sp
, cp
- sp
, true,
4326 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4330 char savechar
, *name
;
4334 name
= concat (package
, "::", sp
);
4336 make_tag (name
, strlen (name
), true,
4337 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4341 else if (LOOKING_AT (cp
, "use constant")
4342 || LOOKING_AT (cp
, "use constant::defer"))
4344 /* For hash style multi-constant like
4345 use constant { FOO => 123,
4347 only the first FOO is picked up. Parsing across the value
4348 expressions would be difficult in general, due to possible nested
4349 hashes, here-documents, etc. */
4351 cp
= skip_spaces (cp
+1);
4354 else if (globals
) /* only if we are tagging global vars */
4356 /* Skip a qualifier, if any. */
4357 bool qual
= LOOKING_AT (cp
, "my") || LOOKING_AT (cp
, "local");
4358 /* After "my" or "local", but before any following paren or space. */
4359 char *varstart
= cp
;
4361 if (qual
/* should this be removed? If yes, how? */
4362 && (*cp
== '$' || *cp
== '@' || *cp
== '%'))
4367 while (c_isalnum (*cp
) || *cp
== '_');
4371 /* Should be examining a variable list at this point;
4372 could insist on seeing an open parenthesis. */
4373 while (*cp
!= '\0' && *cp
!= ';' && *cp
!= '=' && *cp
!= ')')
4379 make_tag (varstart
, cp
- varstart
, false,
4380 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4389 * Look for /^[\t]*def[ \t\n]+[^ \t\n(:]+/ or /^class[ \t\n]+[^ \t\n(:]+/
4390 * Idea by Eric S. Raymond <esr@thyrsus.com> (1997)
4391 * More ideas by seb bacon <seb@jamkit.com> (2002)
4394 Python_functions (FILE *inf
)
4398 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4400 cp
= skip_spaces (cp
);
4401 if (LOOKING_AT (cp
, "def") || LOOKING_AT (cp
, "class"))
4404 while (!notinname (*cp
) && *cp
!= ':')
4406 make_tag (name
, cp
- name
, true,
4407 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4416 * - /^[ \t]*function[ \t\n]+[^ \t\n(]+/
4417 * - /^[ \t]*class[ \t\n]+[^ \t\n]+/
4418 * - /^[ \t]*define\(\"[^\"]+/
4419 * Only with --members:
4420 * - /^[ \t]*var[ \t\n]+\$[^ \t\n=;]/
4421 * Idea by Diez B. Roggisch (2001)
4424 PHP_functions (FILE *inf
)
4427 bool search_identifier
= false;
4429 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4431 cp
= skip_spaces (cp
);
4433 if (search_identifier
4436 while (!notinname (*cp
))
4438 make_tag (name
, cp
- name
, true,
4439 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4440 search_identifier
= false;
4442 else if (LOOKING_AT (cp
, "function"))
4445 cp
= skip_spaces (cp
+1);
4449 while (!notinname (*cp
))
4451 make_tag (name
, cp
- name
, true,
4452 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4455 search_identifier
= true;
4457 else if (LOOKING_AT (cp
, "class"))
4462 while (*cp
!= '\0' && !c_isspace (*cp
))
4464 make_tag (name
, cp
- name
, false,
4465 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4468 search_identifier
= true;
4470 else if (strneq (cp
, "define", 6)
4471 && (cp
= skip_spaces (cp
+6))
4473 && (*cp
== '"' || *cp
== '\''))
4477 while (*cp
!= quote
&& *cp
!= '\0')
4479 make_tag (name
, cp
- name
, false,
4480 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4483 && LOOKING_AT (cp
, "var")
4487 while (!notinname (*cp
))
4489 make_tag (name
, cp
- name
, false,
4490 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4497 * Cobol tag functions
4498 * We could look for anything that could be a paragraph name.
4499 * i.e. anything that starts in column 8 is one word and ends in a full stop.
4500 * Idea by Corny de Souza (1993)
4503 Cobol_paragraphs (FILE *inf
)
4505 register char *bp
, *ep
;
4507 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4513 /* If eoln, compiler option or comment ignore whole line. */
4514 if (bp
[-1] != ' ' || !c_isalnum (bp
[0]))
4517 for (ep
= bp
; c_isalnum (*ep
) || *ep
== '-'; ep
++)
4520 make_tag (bp
, ep
- bp
, true,
4521 lb
.buffer
, ep
- lb
.buffer
+ 1, lineno
, linecharno
);
4528 * Ideas by Assar Westerlund <assar@sics.se> (2001)
4531 Makefile_targets (FILE *inf
)
4535 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4537 if (*bp
== '\t' || *bp
== '#')
4539 while (*bp
!= '\0' && *bp
!= '=' && *bp
!= ':')
4541 if (*bp
== ':' || (globals
&& *bp
== '='))
4543 /* We should detect if there is more than one tag, but we do not.
4544 We just skip initial and final spaces. */
4545 char * namestart
= skip_spaces (lb
.buffer
);
4546 while (--bp
> namestart
)
4547 if (!notinname (*bp
))
4549 make_tag (namestart
, bp
- namestart
+ 1, true,
4550 lb
.buffer
, bp
- lb
.buffer
+ 2, lineno
, linecharno
);
4558 * Original code by Mosur K. Mohan (1989)
4560 * Locates tags for procedures & functions. Doesn't do any type- or
4561 * var-definitions. It does look for the keyword "extern" or
4562 * "forward" immediately following the procedure statement; if found,
4563 * the tag is skipped.
4566 Pascal_functions (FILE *inf
)
4568 linebuffer tline
; /* mostly copied from C_entries */
4570 int save_lineno
, namelen
, taglen
;
4573 bool /* each of these flags is true if: */
4574 incomment
, /* point is inside a comment */
4575 inquote
, /* point is inside '..' string */
4576 get_tagname
, /* point is after PROCEDURE/FUNCTION
4577 keyword, so next item = potential tag */
4578 found_tag
, /* point is after a potential tag */
4579 inparms
, /* point is within parameter-list */
4580 verify_tag
; /* point has passed the parm-list, so the
4581 next token will determine whether this
4582 is a FORWARD/EXTERN to be ignored, or
4583 whether it is a real tag */
4585 save_lcno
= save_lineno
= namelen
= taglen
= 0; /* keep compiler quiet */
4586 name
= NULL
; /* keep compiler quiet */
4589 linebuffer_init (&tline
);
4591 incomment
= inquote
= false;
4592 found_tag
= false; /* have a proc name; check if extern */
4593 get_tagname
= false; /* found "procedure" keyword */
4594 inparms
= false; /* found '(' after "proc" */
4595 verify_tag
= false; /* check if "extern" is ahead */
4598 while (perhaps_more_input (inf
)) /* long main loop to get next char */
4601 if (c
== '\0') /* if end of line */
4603 readline (&lb
, inf
);
4607 if (!((found_tag
&& verify_tag
)
4609 c
= *dbp
++; /* only if don't need *dbp pointing
4610 to the beginning of the name of
4611 the procedure or function */
4615 if (c
== '}') /* within { } comments */
4617 else if (c
== '*' && *dbp
== ')') /* within (* *) comments */
4634 inquote
= true; /* found first quote */
4636 case '{': /* found open { comment */
4640 if (*dbp
== '*') /* found open (* comment */
4645 else if (found_tag
) /* found '(' after tag, i.e., parm-list */
4648 case ')': /* end of parms list */
4653 if (found_tag
&& !inparms
) /* end of proc or fn stmt */
4660 if (found_tag
&& verify_tag
&& (*dbp
!= ' '))
4662 /* Check if this is an "extern" declaration. */
4665 if (c_tolower (*dbp
) == 'e')
4667 if (nocase_tail ("extern")) /* superfluous, really! */
4673 else if (c_tolower (*dbp
) == 'f')
4675 if (nocase_tail ("forward")) /* check for forward reference */
4681 if (found_tag
&& verify_tag
) /* not external proc, so make tag */
4685 make_tag (name
, namelen
, true,
4686 tline
.buffer
, taglen
, save_lineno
, save_lcno
);
4690 if (get_tagname
) /* grab name of proc or fn */
4697 /* Find block name. */
4698 for (cp
= dbp
+ 1; *cp
!= '\0' && !endtoken (*cp
); cp
++)
4701 /* Save all values for later tagging. */
4702 linebuffer_setlen (&tline
, lb
.len
);
4703 strcpy (tline
.buffer
, lb
.buffer
);
4704 save_lineno
= lineno
;
4705 save_lcno
= linecharno
;
4706 name
= tline
.buffer
+ (dbp
- lb
.buffer
);
4708 taglen
= cp
- lb
.buffer
+ 1;
4710 dbp
= cp
; /* set dbp to e-o-token */
4711 get_tagname
= false;
4715 /* And proceed to check for "extern". */
4717 else if (!incomment
&& !inquote
&& !found_tag
)
4719 /* Check for proc/fn keywords. */
4720 switch (c_tolower (c
))
4723 if (nocase_tail ("rocedure")) /* c = 'p', dbp has advanced */
4727 if (nocase_tail ("unction"))
4732 } /* while not eof */
4734 free (tline
.buffer
);
4739 * Lisp tag functions
4740 * look for (def or (DEF, quote or QUOTE
4743 static void L_getit (void);
4748 if (*dbp
== '\'') /* Skip prefix quote */
4750 else if (*dbp
== '(')
4753 /* Try to skip "(quote " */
4754 if (!LOOKING_AT (dbp
, "quote") && !LOOKING_AT (dbp
, "QUOTE"))
4755 /* Ok, then skip "(" before name in (defstruct (foo)) */
4756 dbp
= skip_spaces (dbp
);
4758 get_tag (dbp
, NULL
);
4762 Lisp_functions (FILE *inf
)
4764 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
4769 /* "(defvar foo)" is a declaration rather than a definition. */
4773 if (LOOKING_AT (p
, "defvar"))
4775 p
= skip_name (p
); /* past var name */
4776 p
= skip_spaces (p
);
4782 if (strneq (dbp
+ 1, "cl-", 3) || strneq (dbp
+ 1, "CL-", 3))
4785 if (strneq (dbp
+1, "def", 3) || strneq (dbp
+1, "DEF", 3))
4787 dbp
= skip_non_spaces (dbp
);
4788 dbp
= skip_spaces (dbp
);
4793 /* Check for (foo::defmumble name-defined ... */
4796 while (!notinname (*dbp
) && *dbp
!= ':');
4801 while (*dbp
== ':');
4803 if (strneq (dbp
, "def", 3) || strneq (dbp
, "DEF", 3))
4805 dbp
= skip_non_spaces (dbp
);
4806 dbp
= skip_spaces (dbp
);
4816 * Lua script language parsing
4817 * Original code by David A. Capello <dacap@users.sourceforge.net> (2004)
4819 * "function" and "local function" are tags if they start at column 1.
4822 Lua_functions (FILE *inf
)
4826 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4828 if (bp
[0] != 'f' && bp
[0] != 'l')
4831 (void)LOOKING_AT (bp
, "local"); /* skip possible "local" */
4833 if (LOOKING_AT (bp
, "function"))
4841 * Just look for lines where the first character is '/'
4842 * Also look at "defineps" for PSWrap
4844 * Richard Mlynarik <mly@adoc.xerox.com> (1997)
4845 * Masatake Yamato <masata-y@is.aist-nara.ac.jp> (1999)
4848 PS_functions (FILE *inf
)
4850 register char *bp
, *ep
;
4852 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4857 *ep
!= '\0' && *ep
!= ' ' && *ep
!= '{';
4860 make_tag (bp
, ep
- bp
, true,
4861 lb
.buffer
, ep
- lb
.buffer
+ 1, lineno
, linecharno
);
4863 else if (LOOKING_AT (bp
, "defineps"))
4871 * Ignore anything after \ followed by space or in ( )
4872 * Look for words defined by :
4873 * Look for constant, code, create, defer, value, and variable
4874 * OBP extensions: Look for buffer:, field,
4875 * Ideas by Eduardo Horvath <eeh@netbsd.org> (2004)
4878 Forth_words (FILE *inf
)
4882 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4883 while ((bp
= skip_spaces (bp
))[0] != '\0')
4884 if (bp
[0] == '\\' && c_isspace (bp
[1]))
4885 break; /* read next line */
4886 else if (bp
[0] == '(' && c_isspace (bp
[1]))
4887 do /* skip to ) or eol */
4889 while (*bp
!= ')' && *bp
!= '\0');
4890 else if ((bp
[0] == ':' && c_isspace (bp
[1]) && bp
++)
4891 || LOOKING_AT_NOCASE (bp
, "constant")
4892 || LOOKING_AT_NOCASE (bp
, "code")
4893 || LOOKING_AT_NOCASE (bp
, "create")
4894 || LOOKING_AT_NOCASE (bp
, "defer")
4895 || LOOKING_AT_NOCASE (bp
, "value")
4896 || LOOKING_AT_NOCASE (bp
, "variable")
4897 || LOOKING_AT_NOCASE (bp
, "buffer:")
4898 || LOOKING_AT_NOCASE (bp
, "field"))
4899 get_tag (skip_spaces (bp
), NULL
); /* Yay! A definition! */
4901 bp
= skip_non_spaces (bp
);
4906 * Scheme tag functions
4907 * look for (def... xyzzy
4909 * (def ... ((...(xyzzy ....
4911 * Original code by Ken Haase (1985?)
4914 Scheme_functions (FILE *inf
)
4918 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4920 if (strneq (bp
, "(def", 4) || strneq (bp
, "(DEF", 4))
4922 bp
= skip_non_spaces (bp
+4);
4923 /* Skip over open parens and white space. Don't continue past
4925 while (*bp
&& notinname (*bp
))
4929 if (LOOKING_AT (bp
, "(SET!") || LOOKING_AT (bp
, "(set!"))
4935 /* Find tags in TeX and LaTeX input files. */
4937 /* TEX_toktab is a table of TeX control sequences that define tags.
4938 * Each entry records one such control sequence.
4940 * Original code from who knows whom.
4942 * Stefan Monnier (2002)
4945 static linebuffer
*TEX_toktab
= NULL
; /* Table with tag tokens */
4947 /* Default set of control sequences to put into TEX_toktab.
4948 The value of environment var TEXTAGS is prepended to this. */
4949 static const char *TEX_defenv
= "\
4950 :chapter:section:subsection:subsubsection:eqno:label:ref:cite:bibitem\
4951 :part:appendix:entry:index:def\
4952 :newcommand:renewcommand:newenvironment:renewenvironment";
4954 static void TEX_decode_env (const char *, const char *);
4957 * TeX/LaTeX scanning loop.
4960 TeX_commands (FILE *inf
)
4965 char TEX_esc
= '\0';
4966 char TEX_opgrp
, TEX_clgrp
;
4968 /* Initialize token table once from environment. */
4969 if (TEX_toktab
== NULL
)
4970 TEX_decode_env ("TEXTAGS", TEX_defenv
);
4972 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4974 /* Look at each TEX keyword in line. */
4977 /* Look for a TEX escape. */
4981 if (c
== '\0' || c
== '%')
4984 /* Select either \ or ! as escape character, whichever comes
4985 first outside a comment. */
5006 for (key
= TEX_toktab
; key
->buffer
!= NULL
; key
++)
5007 if (strneq (cp
, key
->buffer
, key
->len
))
5010 int namelen
, linelen
;
5013 cp
= skip_spaces (cp
+ key
->len
);
5014 if (*cp
== TEX_opgrp
)
5020 (!c_isspace (*p
) && *p
!= '#' &&
5021 *p
!= TEX_opgrp
&& *p
!= TEX_clgrp
);
5026 if (!opgrp
|| *p
== TEX_clgrp
)
5028 while (*p
!= '\0' && *p
!= TEX_opgrp
&& *p
!= TEX_clgrp
)
5030 linelen
= p
- lb
.buffer
+ 1;
5032 make_tag (cp
, namelen
, true,
5033 lb
.buffer
, linelen
, lineno
, linecharno
);
5034 goto tex_next_line
; /* We only tag a line once */
5042 /* Read environment and prepend it to the default string.
5043 Build token table. */
5045 TEX_decode_env (const char *evarname
, const char *defenv
)
5047 register const char *env
, *p
;
5050 /* Append default string to environment. */
5051 env
= getenv (evarname
);
5055 env
= concat (env
, defenv
, "");
5057 /* Allocate a token table */
5058 for (len
= 1, p
= env
; (p
= strchr (p
, ':')); )
5061 TEX_toktab
= xnew (len
, linebuffer
);
5063 /* Unpack environment string into token table. Be careful about */
5064 /* zero-length strings (leading ':', "::" and trailing ':') */
5065 for (i
= 0; *env
!= '\0';)
5067 p
= strchr (env
, ':');
5068 if (!p
) /* End of environment string. */
5069 p
= env
+ strlen (env
);
5071 { /* Only non-zero strings. */
5072 TEX_toktab
[i
].buffer
= savenstr (env
, p
- env
);
5073 TEX_toktab
[i
].len
= p
- env
;
5080 TEX_toktab
[i
].buffer
= NULL
; /* Mark end of table. */
5081 TEX_toktab
[i
].len
= 0;
5088 /* Texinfo support. Dave Love, Mar. 2000. */
5090 Texinfo_nodes (FILE *inf
)
5093 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5094 if (LOOKING_AT (cp
, "@node"))
5097 while (*cp
!= '\0' && *cp
!= ',')
5099 make_tag (start
, cp
- start
, true,
5100 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
5107 * Contents of <title>, <h1>, <h2>, <h3> are tags.
5108 * Contents of <a name=xxx> are tags with name xxx.
5110 * Francesco Potortì, 2002.
5113 HTML_labels (FILE *inf
)
5115 bool getnext
= false; /* next text outside of HTML tags is a tag */
5116 bool skiptag
= false; /* skip to the end of the current HTML tag */
5117 bool intag
= false; /* inside an html tag, looking for ID= */
5118 bool inanchor
= false; /* when INTAG, is an anchor, look for NAME= */
5122 linebuffer_setlen (&token_name
, 0); /* no name in buffer */
5124 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
5125 for (;;) /* loop on the same line */
5127 if (skiptag
) /* skip HTML tag */
5129 while (*dbp
!= '\0' && *dbp
!= '>')
5135 continue; /* look on the same line */
5137 break; /* go to next line */
5140 else if (intag
) /* look for "name=" or "id=" */
5142 while (*dbp
!= '\0' && *dbp
!= '>'
5143 && c_tolower (*dbp
) != 'n' && c_tolower (*dbp
) != 'i')
5146 break; /* go to next line */
5151 continue; /* look on the same line */
5153 if ((inanchor
&& LOOKING_AT_NOCASE (dbp
, "name="))
5154 || LOOKING_AT_NOCASE (dbp
, "id="))
5156 bool quoted
= (dbp
[0] == '"');
5159 for (end
= ++dbp
; *end
!= '\0' && *end
!= '"'; end
++)
5162 for (end
= dbp
; *end
!= '\0' && intoken (*end
); end
++)
5164 linebuffer_setlen (&token_name
, end
- dbp
);
5165 memcpy (token_name
.buffer
, dbp
, end
- dbp
);
5166 token_name
.buffer
[end
- dbp
] = '\0';
5169 intag
= false; /* we found what we looked for */
5170 skiptag
= true; /* skip to the end of the tag */
5171 getnext
= true; /* then grab the text */
5172 continue; /* look on the same line */
5177 else if (getnext
) /* grab next tokens and tag them */
5179 dbp
= skip_spaces (dbp
);
5181 break; /* go to next line */
5185 inanchor
= (c_tolower (dbp
[1]) == 'a' && !intoken (dbp
[2]));
5186 continue; /* look on the same line */
5189 for (end
= dbp
+ 1; *end
!= '\0' && *end
!= '<'; end
++)
5191 make_tag (token_name
.buffer
, token_name
.len
, true,
5192 dbp
, end
- dbp
, lineno
, linecharno
);
5193 linebuffer_setlen (&token_name
, 0); /* no name in buffer */
5195 break; /* go to next line */
5198 else /* look for an interesting HTML tag */
5200 while (*dbp
!= '\0' && *dbp
!= '<')
5203 break; /* go to next line */
5205 if (c_tolower (dbp
[1]) == 'a' && !intoken (dbp
[2]))
5208 continue; /* look on the same line */
5210 else if (LOOKING_AT_NOCASE (dbp
, "<title>")
5211 || LOOKING_AT_NOCASE (dbp
, "<h1>")
5212 || LOOKING_AT_NOCASE (dbp
, "<h2>")
5213 || LOOKING_AT_NOCASE (dbp
, "<h3>"))
5217 continue; /* look on the same line */
5228 * Assumes that the predicate or rule starts at column 0.
5229 * Only the first clause of a predicate or rule is added.
5230 * Original code by Sunichirou Sugou (1989)
5231 * Rewritten by Anders Lindgren (1996)
5233 static size_t prolog_pr (char *, char *);
5234 static void prolog_skip_comment (linebuffer
*, FILE *);
5235 static size_t prolog_atom (char *, size_t);
5238 Prolog_functions (FILE *inf
)
5248 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5250 if (cp
[0] == '\0') /* Empty line */
5252 else if (c_isspace (cp
[0])) /* Not a predicate */
5254 else if (cp
[0] == '/' && cp
[1] == '*') /* comment. */
5255 prolog_skip_comment (&lb
, inf
);
5256 else if ((len
= prolog_pr (cp
, last
)) > 0)
5258 /* Predicate or rule. Store the function name so that we
5259 only generate a tag for the first clause. */
5261 last
= xnew (len
+ 1, char);
5262 else if (len
+ 1 > allocated
)
5263 xrnew (last
, len
+ 1, char);
5264 allocated
= len
+ 1;
5265 memcpy (last
, cp
, len
);
5274 prolog_skip_comment (linebuffer
*plb
, FILE *inf
)
5280 for (cp
= plb
->buffer
; *cp
!= '\0'; cp
++)
5281 if (cp
[0] == '*' && cp
[1] == '/')
5283 readline (plb
, inf
);
5285 while (perhaps_more_input (inf
));
5289 * A predicate or rule definition is added if it matches:
5290 * <beginning of line><Prolog Atom><whitespace>(
5291 * or <beginning of line><Prolog Atom><whitespace>:-
5293 * It is added to the tags database if it doesn't match the
5294 * name of the previous clause header.
5296 * Return the size of the name of the predicate or rule, or 0 if no
5300 prolog_pr (char *s
, char *last
)
5302 /* Name of last clause. */
5307 pos
= prolog_atom (s
, 0);
5312 pos
= skip_spaces (s
+ pos
) - s
;
5315 || (s
[pos
] == '(' && (pos
+= 1))
5316 || (s
[pos
] == ':' && s
[pos
+ 1] == '-' && (pos
+= 2)))
5317 && (last
== NULL
/* save only the first clause */
5318 || len
!= strlen (last
)
5319 || !strneq (s
, last
, len
)))
5321 make_tag (s
, len
, true, s
, pos
, lineno
, linecharno
);
5329 * Consume a Prolog atom.
5330 * Return the number of bytes consumed, or 0 if there was an error.
5332 * A prolog atom, in this context, could be one of:
5333 * - An alphanumeric sequence, starting with a lower case letter.
5334 * - A quoted arbitrary string. Single quotes can escape themselves.
5335 * Backslash quotes everything.
5338 prolog_atom (char *s
, size_t pos
)
5344 if (c_islower (s
[pos
]) || s
[pos
] == '_')
5346 /* The atom is unquoted. */
5348 while (c_isalnum (s
[pos
]) || s
[pos
] == '_')
5352 return pos
- origpos
;
5354 else if (s
[pos
] == '\'')
5365 pos
++; /* A double quote */
5367 else if (s
[pos
] == '\0')
5368 /* Multiline quoted atoms are ignored. */
5370 else if (s
[pos
] == '\\')
5372 if (s
[pos
+1] == '\0')
5379 return pos
- origpos
;
5387 * Support for Erlang
5389 * Generates tags for functions, defines, and records.
5390 * Assumes that Erlang functions start at column 0.
5391 * Original code by Anders Lindgren (1996)
5393 static int erlang_func (char *, char *);
5394 static void erlang_attribute (char *);
5395 static int erlang_atom (char *);
5398 Erlang_functions (FILE *inf
)
5408 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5410 if (cp
[0] == '\0') /* Empty line */
5412 else if (c_isspace (cp
[0])) /* Not function nor attribute */
5414 else if (cp
[0] == '%') /* comment */
5416 else if (cp
[0] == '"') /* Sometimes, strings start in column one */
5418 else if (cp
[0] == '-') /* attribute, e.g. "-define" */
5420 erlang_attribute (cp
);
5427 else if ((len
= erlang_func (cp
, last
)) > 0)
5430 * Function. Store the function name so that we only
5431 * generates a tag for the first clause.
5434 last
= xnew (len
+ 1, char);
5435 else if (len
+ 1 > allocated
)
5436 xrnew (last
, len
+ 1, char);
5437 allocated
= len
+ 1;
5438 memcpy (last
, cp
, len
);
5447 * A function definition is added if it matches:
5448 * <beginning of line><Erlang Atom><whitespace>(
5450 * It is added to the tags database if it doesn't match the
5451 * name of the previous clause header.
5453 * Return the size of the name of the function, or 0 if no function
5457 erlang_func (char *s
, char *last
)
5459 /* Name of last clause. */
5464 pos
= erlang_atom (s
);
5469 pos
= skip_spaces (s
+ pos
) - s
;
5471 /* Save only the first clause. */
5474 || len
!= (int)strlen (last
)
5475 || !strneq (s
, last
, len
)))
5477 make_tag (s
, len
, true, s
, pos
, lineno
, linecharno
);
5486 * Handle attributes. Currently, tags are generated for defines
5489 * They are on the form:
5490 * -define(foo, bar).
5491 * -define(Foo(M, N), M+N).
5492 * -record(graph, {vtab = notable, cyclic = true}).
5495 erlang_attribute (char *s
)
5499 if ((LOOKING_AT (cp
, "-define") || LOOKING_AT (cp
, "-record"))
5502 int len
= erlang_atom (skip_spaces (cp
));
5504 make_tag (cp
, len
, true, s
, cp
+ len
- s
, lineno
, linecharno
);
5511 * Consume an Erlang atom (or variable).
5512 * Return the number of bytes consumed, or -1 if there was an error.
5515 erlang_atom (char *s
)
5519 if (c_isalpha (s
[pos
]) || s
[pos
] == '_')
5521 /* The atom is unquoted. */
5524 while (c_isalnum (s
[pos
]) || s
[pos
] == '_');
5526 else if (s
[pos
] == '\'')
5528 for (pos
++; s
[pos
] != '\''; pos
++)
5529 if (s
[pos
] == '\0' /* multiline quoted atoms are ignored */
5530 || (s
[pos
] == '\\' && s
[++pos
] == '\0'))
5539 static char *scan_separators (char *);
5540 static void add_regex (char *, language
*);
5541 static char *substitute (char *, char *, struct re_registers
*);
5544 * Take a string like "/blah/" and turn it into "blah", verifying
5545 * that the first and last characters are the same, and handling
5546 * quoted separator characters. Actually, stops on the occurrence of
5547 * an unquoted separator. Also process \t, \n, etc. and turn into
5548 * appropriate characters. Works in place. Null terminates name string.
5549 * Returns pointer to terminating separator, or NULL for
5550 * unterminated regexps.
5553 scan_separators (char *name
)
5556 char *copyto
= name
;
5557 bool quoted
= false;
5559 for (++name
; *name
!= '\0'; ++name
)
5565 case 'a': *copyto
++ = '\007'; break; /* BEL (bell) */
5566 case 'b': *copyto
++ = '\b'; break; /* BS (back space) */
5567 case 'd': *copyto
++ = 0177; break; /* DEL (delete) */
5568 case 'e': *copyto
++ = 033; break; /* ESC (delete) */
5569 case 'f': *copyto
++ = '\f'; break; /* FF (form feed) */
5570 case 'n': *copyto
++ = '\n'; break; /* NL (new line) */
5571 case 'r': *copyto
++ = '\r'; break; /* CR (carriage return) */
5572 case 't': *copyto
++ = '\t'; break; /* TAB (horizontal tab) */
5573 case 'v': *copyto
++ = '\v'; break; /* VT (vertical tab) */
5579 /* Something else is quoted, so preserve the quote. */
5587 else if (*name
== '\\')
5589 else if (*name
== sep
)
5595 name
= NULL
; /* signal unterminated regexp */
5597 /* Terminate copied string. */
5602 /* Look at the argument of --regex or --no-regex and do the right
5603 thing. Same for each line of a regexp file. */
5605 analyze_regex (char *regex_arg
)
5607 if (regex_arg
== NULL
)
5609 free_regexps (); /* --no-regex: remove existing regexps */
5613 /* A real --regexp option or a line in a regexp file. */
5614 switch (regex_arg
[0])
5616 /* Comments in regexp file or null arg to --regex. */
5622 /* Read a regex file. This is recursive and may result in a
5623 loop, which will stop when the file descriptors are exhausted. */
5627 linebuffer regexbuf
;
5628 char *regexfile
= regex_arg
+ 1;
5630 /* regexfile is a file containing regexps, one per line. */
5631 regexfp
= fopen (regexfile
, "r" FOPEN_BINARY
);
5632 if (regexfp
== NULL
)
5634 linebuffer_init (®exbuf
);
5635 while (readline_internal (®exbuf
, regexfp
, regexfile
) > 0)
5636 analyze_regex (regexbuf
.buffer
);
5637 free (regexbuf
.buffer
);
5638 if (fclose (regexfp
) != 0)
5643 /* Regexp to be used for a specific language only. */
5647 char *lang_name
= regex_arg
+ 1;
5650 for (cp
= lang_name
; *cp
!= '}'; cp
++)
5653 error ("unterminated language name in regex: %s", regex_arg
);
5657 lang
= get_language_from_langname (lang_name
);
5660 add_regex (cp
, lang
);
5664 /* Regexp to be used for any language. */
5666 add_regex (regex_arg
, NULL
);
5671 /* Separate the regexp pattern, compile it,
5672 and care for optional name and modifiers. */
5674 add_regex (char *regexp_pattern
, language
*lang
)
5676 static struct re_pattern_buffer zeropattern
;
5677 char sep
, *pat
, *name
, *modifiers
;
5680 struct re_pattern_buffer
*patbuf
;
5683 force_explicit_name
= true, /* do not use implicit tag names */
5684 ignore_case
= false, /* case is significant */
5685 multi_line
= false, /* matches are done one line at a time */
5686 single_line
= false; /* dot does not match newline */
5689 if (strlen (regexp_pattern
) < 3)
5691 error ("null regexp");
5694 sep
= regexp_pattern
[0];
5695 name
= scan_separators (regexp_pattern
);
5698 error ("%s: unterminated regexp", regexp_pattern
);
5703 error ("null name for regexp \"%s\"", regexp_pattern
);
5706 modifiers
= scan_separators (name
);
5707 if (modifiers
== NULL
) /* no terminating separator --> no name */
5713 modifiers
+= 1; /* skip separator */
5715 /* Parse regex modifiers. */
5716 for (; modifiers
[0] != '\0'; modifiers
++)
5717 switch (modifiers
[0])
5720 if (modifiers
== name
)
5721 error ("forcing explicit tag name but no name, ignoring");
5722 force_explicit_name
= true;
5732 need_filebuf
= true;
5735 error ("invalid regexp modifier `%c', ignoring", modifiers
[0]);
5739 patbuf
= xnew (1, struct re_pattern_buffer
);
5740 *patbuf
= zeropattern
;
5743 static char lc_trans
[UCHAR_MAX
+ 1];
5745 for (i
= 0; i
< UCHAR_MAX
+ 1; i
++)
5746 lc_trans
[i
] = c_tolower (i
);
5747 patbuf
->translate
= lc_trans
; /* translation table to fold case */
5751 pat
= concat ("^", regexp_pattern
, ""); /* anchor to beginning of line */
5753 pat
= regexp_pattern
;
5756 re_set_syntax (RE_SYNTAX_EMACS
| RE_DOT_NEWLINE
);
5758 re_set_syntax (RE_SYNTAX_EMACS
);
5760 err
= re_compile_pattern (pat
, strlen (pat
), patbuf
);
5765 error ("%s while compiling pattern", err
);
5770 p_head
= xnew (1, regexp
);
5771 p_head
->pattern
= savestr (regexp_pattern
);
5772 p_head
->p_next
= rp
;
5773 p_head
->lang
= lang
;
5774 p_head
->pat
= patbuf
;
5775 p_head
->name
= savestr (name
);
5776 p_head
->error_signaled
= false;
5777 p_head
->force_explicit_name
= force_explicit_name
;
5778 p_head
->ignore_case
= ignore_case
;
5779 p_head
->multi_line
= multi_line
;
5783 * Do the substitutions indicated by the regular expression and
5787 substitute (char *in
, char *out
, struct re_registers
*regs
)
5790 int size
, dig
, diglen
;
5793 size
= strlen (out
);
5795 /* Pass 1: figure out how much to allocate by finding all \N strings. */
5796 if (out
[size
- 1] == '\\')
5797 fatal ("pattern error in \"%s\"", out
);
5798 for (t
= strchr (out
, '\\');
5800 t
= strchr (t
+ 2, '\\'))
5801 if (c_isdigit (t
[1]))
5804 diglen
= regs
->end
[dig
] - regs
->start
[dig
];
5810 /* Allocate space and do the substitutions. */
5812 result
= xnew (size
+ 1, char);
5814 for (t
= result
; *out
!= '\0'; out
++)
5815 if (*out
== '\\' && c_isdigit (*++out
))
5818 diglen
= regs
->end
[dig
] - regs
->start
[dig
];
5819 memcpy (t
, in
+ regs
->start
[dig
], diglen
);
5826 assert (t
<= result
+ size
);
5827 assert (t
- result
== (int)strlen (result
));
5832 /* Deallocate all regexps. */
5837 while (p_head
!= NULL
)
5839 rp
= p_head
->p_next
;
5840 free (p_head
->pattern
);
5841 free (p_head
->name
);
5849 * Reads the whole file as a single string from `filebuf' and looks for
5850 * multi-line regular expressions, creating tags on matches.
5851 * readline already dealt with normal regexps.
5853 * Idea by Ben Wing <ben@666.com> (2002).
5856 regex_tag_multiline (void)
5858 char *buffer
= filebuf
.buffer
;
5862 for (rp
= p_head
; rp
!= NULL
; rp
= rp
->p_next
)
5866 if (!rp
->multi_line
)
5867 continue; /* skip normal regexps */
5869 /* Generic initializations before parsing file from memory. */
5870 lineno
= 1; /* reset global line number */
5871 charno
= 0; /* reset global char number */
5872 linecharno
= 0; /* reset global char number of line start */
5874 /* Only use generic regexps or those for the current language. */
5875 if (rp
->lang
!= NULL
&& rp
->lang
!= curfdp
->lang
)
5878 while (match
>= 0 && match
< filebuf
.len
)
5880 match
= re_search (rp
->pat
, buffer
, filebuf
.len
, charno
,
5881 filebuf
.len
- match
, &rp
->regs
);
5886 if (!rp
->error_signaled
)
5888 error ("regexp stack overflow while matching \"%s\"",
5890 rp
->error_signaled
= true;
5897 if (match
== rp
->regs
.end
[0])
5899 if (!rp
->error_signaled
)
5901 error ("regexp matches the empty string: \"%s\"",
5903 rp
->error_signaled
= true;
5905 match
= -3; /* exit from while loop */
5909 /* Match occurred. Construct a tag. */
5910 while (charno
< rp
->regs
.end
[0])
5911 if (buffer
[charno
++] == '\n')
5912 lineno
++, linecharno
= charno
;
5914 if (name
[0] == '\0')
5916 else /* make a named tag */
5917 name
= substitute (buffer
, rp
->name
, &rp
->regs
);
5918 if (rp
->force_explicit_name
)
5919 /* Force explicit tag name, if a name is there. */
5920 pfnote (name
, true, buffer
+ linecharno
,
5921 charno
- linecharno
+ 1, lineno
, linecharno
);
5923 make_tag (name
, strlen (name
), true, buffer
+ linecharno
,
5924 charno
- linecharno
+ 1, lineno
, linecharno
);
5933 nocase_tail (const char *cp
)
5937 while (*cp
!= '\0' && c_tolower (*cp
) == c_tolower (dbp
[len
]))
5939 if (*cp
== '\0' && !intoken (dbp
[len
]))
5948 get_tag (register char *bp
, char **namepp
)
5950 register char *cp
= bp
;
5954 /* Go till you get to white space or a syntactic break */
5955 for (cp
= bp
+ 1; !notinname (*cp
); cp
++)
5957 make_tag (bp
, cp
- bp
, true,
5958 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
5962 *namepp
= savenstr (bp
, cp
- bp
);
5966 * Read a line of text from `stream' into `lbp', excluding the
5967 * newline or CR-NL, if any. Return the number of characters read from
5968 * `stream', which is the length of the line including the newline.
5970 * On DOS or Windows we do not count the CR character, if any before the
5971 * NL, in the returned length; this mirrors the behavior of Emacs on those
5972 * platforms (for text files, it translates CR-NL to NL as it reads in the
5975 * If multi-line regular expressions are requested, each line read is
5976 * appended to `filebuf'.
5979 readline_internal (linebuffer
*lbp
, FILE *stream
, char const *filename
)
5981 char *buffer
= lbp
->buffer
;
5982 char *p
= lbp
->buffer
;
5986 pend
= p
+ lbp
->size
; /* Separate to avoid 386/IX compiler bug. */
5990 register int c
= getc (stream
);
5993 /* We're at the end of linebuffer: expand it. */
5995 xrnew (buffer
, lbp
->size
, char);
5996 p
+= buffer
- lbp
->buffer
;
5997 pend
= buffer
+ lbp
->size
;
5998 lbp
->buffer
= buffer
;
6002 if (ferror (stream
))
6010 if (p
> buffer
&& p
[-1] == '\r')
6014 /* Assume CRLF->LF translation will be performed by Emacs
6015 when loading this file, so CRs won't appear in the buffer.
6016 It would be cleaner to compensate within Emacs;
6017 however, Emacs does not know how many CRs were deleted
6018 before any given point in the file. */
6033 lbp
->len
= p
- buffer
;
6035 if (need_filebuf
/* we need filebuf for multi-line regexps */
6036 && chars_deleted
> 0) /* not at EOF */
6038 while (filebuf
.size
<= filebuf
.len
+ lbp
->len
+ 1) /* +1 for \n */
6040 /* Expand filebuf. */
6042 xrnew (filebuf
.buffer
, filebuf
.size
, char);
6044 memcpy (filebuf
.buffer
+ filebuf
.len
, lbp
->buffer
, lbp
->len
);
6045 filebuf
.len
+= lbp
->len
;
6046 filebuf
.buffer
[filebuf
.len
++] = '\n';
6047 filebuf
.buffer
[filebuf
.len
] = '\0';
6050 return lbp
->len
+ chars_deleted
;
6054 * Like readline_internal, above, but in addition try to match the
6055 * input line against relevant regular expressions and manage #line
6059 readline (linebuffer
*lbp
, FILE *stream
)
6063 linecharno
= charno
; /* update global char number of line start */
6064 result
= readline_internal (lbp
, stream
, infilename
); /* read line */
6065 lineno
+= 1; /* increment global line number */
6066 charno
+= result
; /* increment global char number */
6068 /* Honor #line directives. */
6069 if (!no_line_directive
)
6071 static bool discard_until_line_directive
;
6073 /* Check whether this is a #line directive. */
6074 if (result
> 12 && strneq (lbp
->buffer
, "#line ", 6))
6079 if (sscanf (lbp
->buffer
, "#line %u \"%n", &lno
, &start
) >= 1
6080 && start
> 0) /* double quote character found */
6082 char *endp
= lbp
->buffer
+ start
;
6084 while ((endp
= strchr (endp
, '"')) != NULL
6085 && endp
[-1] == '\\')
6088 /* Ok, this is a real #line directive. Let's deal with it. */
6090 char *taggedabsname
; /* absolute name of original file */
6091 char *taggedfname
; /* name of original file as given */
6092 char *name
; /* temp var */
6094 discard_until_line_directive
= false; /* found it */
6095 name
= lbp
->buffer
+ start
;
6097 canonicalize_filename (name
);
6098 taggedabsname
= absolute_filename (name
, tagfiledir
);
6099 if (filename_is_absolute (name
)
6100 || filename_is_absolute (curfdp
->infname
))
6101 taggedfname
= savestr (taggedabsname
);
6103 taggedfname
= relative_filename (taggedabsname
,tagfiledir
);
6105 if (streq (curfdp
->taggedfname
, taggedfname
))
6106 /* The #line directive is only a line number change. We
6107 deal with this afterwards. */
6110 /* The tags following this #line directive should be
6111 attributed to taggedfname. In order to do this, set
6112 curfdp accordingly. */
6114 fdesc
*fdp
; /* file description pointer */
6116 /* Go look for a file description already set up for the
6117 file indicated in the #line directive. If there is
6118 one, use it from now until the next #line
6120 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
6121 if (streq (fdp
->infname
, curfdp
->infname
)
6122 && streq (fdp
->taggedfname
, taggedfname
))
6123 /* If we remove the second test above (after the &&)
6124 then all entries pertaining to the same file are
6125 coalesced in the tags file. If we use it, then
6126 entries pertaining to the same file but generated
6127 from different files (via #line directives) will
6128 go into separate sections in the tags file. These
6129 alternatives look equivalent. The first one
6130 destroys some apparently useless information. */
6136 /* Else, if we already tagged the real file, skip all
6137 input lines until the next #line directive. */
6138 if (fdp
== NULL
) /* not found */
6139 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
6140 if (streq (fdp
->infabsname
, taggedabsname
))
6142 discard_until_line_directive
= true;
6146 /* Else create a new file description and use that from
6147 now on, until the next #line directive. */
6148 if (fdp
== NULL
) /* not found */
6151 fdhead
= xnew (1, fdesc
);
6152 *fdhead
= *curfdp
; /* copy curr. file description */
6154 fdhead
->infname
= savestr (curfdp
->infname
);
6155 fdhead
->infabsname
= savestr (curfdp
->infabsname
);
6156 fdhead
->infabsdir
= savestr (curfdp
->infabsdir
);
6157 fdhead
->taggedfname
= taggedfname
;
6158 fdhead
->usecharno
= false;
6159 fdhead
->prop
= NULL
;
6160 fdhead
->written
= false;
6164 free (taggedabsname
);
6166 readline (lbp
, stream
);
6168 } /* if a real #line directive */
6169 } /* if #line is followed by a number */
6170 } /* if line begins with "#line " */
6172 /* If we are here, no #line directive was found. */
6173 if (discard_until_line_directive
)
6177 /* Do a tail recursion on ourselves, thus discarding the contents
6178 of the line buffer. */
6179 readline (lbp
, stream
);
6183 discard_until_line_directive
= false;
6186 } /* if #line directives should be considered */
6193 /* Match against relevant regexps. */
6195 for (rp
= p_head
; rp
!= NULL
; rp
= rp
->p_next
)
6197 /* Only use generic regexps or those for the current language.
6198 Also do not use multiline regexps, which is the job of
6199 regex_tag_multiline. */
6200 if ((rp
->lang
!= NULL
&& rp
->lang
!= fdhead
->lang
)
6204 match
= re_match (rp
->pat
, lbp
->buffer
, lbp
->len
, 0, &rp
->regs
);
6209 if (!rp
->error_signaled
)
6211 error ("regexp stack overflow while matching \"%s\"",
6213 rp
->error_signaled
= true;
6220 /* Empty string matched. */
6221 if (!rp
->error_signaled
)
6223 error ("regexp matches the empty string: \"%s\"", rp
->pattern
);
6224 rp
->error_signaled
= true;
6228 /* Match occurred. Construct a tag. */
6230 if (name
[0] == '\0')
6232 else /* make a named tag */
6233 name
= substitute (lbp
->buffer
, rp
->name
, &rp
->regs
);
6234 if (rp
->force_explicit_name
)
6235 /* Force explicit tag name, if a name is there. */
6236 pfnote (name
, true, lbp
->buffer
, match
, lineno
, linecharno
);
6238 make_tag (name
, strlen (name
), true,
6239 lbp
->buffer
, match
, lineno
, linecharno
);
6248 * Return a pointer to a space of size strlen(cp)+1 allocated
6249 * with xnew where the string CP has been copied.
6252 savestr (const char *cp
)
6254 return savenstr (cp
, strlen (cp
));
6258 * Return a pointer to a space of size LEN+1 allocated with xnew where
6259 * the string CP has been copied for at most the first LEN characters.
6262 savenstr (const char *cp
, int len
)
6264 char *dp
= xnew (len
+ 1, char);
6266 return memcpy (dp
, cp
, len
);
6269 /* Skip spaces (end of string is not space), return new pointer. */
6271 skip_spaces (char *cp
)
6273 while (c_isspace (*cp
))
6278 /* Skip non spaces, except end of string, return new pointer. */
6280 skip_non_spaces (char *cp
)
6282 while (*cp
!= '\0' && !c_isspace (*cp
))
6287 /* Skip any chars in the "name" class.*/
6289 skip_name (char *cp
)
6291 /* '\0' is a notinname() so loop stops there too */
6292 while (! notinname (*cp
))
6297 /* Print error message and exit. */
6299 fatal (const char *s1
, const char *s2
)
6302 exit (EXIT_FAILURE
);
6306 pfatal (const char *s1
)
6309 exit (EXIT_FAILURE
);
6313 suggest_asking_for_help (void)
6315 fprintf (stderr
, "\tTry `%s --help' for a complete list of options.\n",
6317 exit (EXIT_FAILURE
);
6320 /* Output a diagnostic with printf-style FORMAT and args. */
6322 error (const char *format
, ...)
6325 va_start (ap
, format
);
6326 fprintf (stderr
, "%s: ", progname
);
6327 vfprintf (stderr
, format
, ap
);
6328 fprintf (stderr
, "\n");
6332 /* Return a newly-allocated string whose contents
6333 concatenate those of s1, s2, s3. */
6335 concat (const char *s1
, const char *s2
, const char *s3
)
6337 int len1
= strlen (s1
), len2
= strlen (s2
), len3
= strlen (s3
);
6338 char *result
= xnew (len1
+ len2
+ len3
+ 1, char);
6340 strcpy (result
, s1
);
6341 strcpy (result
+ len1
, s2
);
6342 strcpy (result
+ len1
+ len2
, s3
);
6348 /* Does the same work as the system V getcwd, but does not need to
6349 guess the buffer size in advance. */
6354 char *path
= xnew (bufsize
, char);
6356 while (getcwd (path
, bufsize
) == NULL
)
6358 if (errno
!= ERANGE
)
6362 path
= xnew (bufsize
, char);
6365 canonicalize_filename (path
);
6369 /* Return a newly allocated string containing a name of a temporary file. */
6373 const char *tmpdir
= getenv ("TMPDIR");
6374 const char *slash
= "/";
6376 #if MSDOS || defined (DOS_NT)
6378 tmpdir
= getenv ("TEMP");
6380 tmpdir
= getenv ("TMP");
6383 if (tmpdir
[strlen (tmpdir
) - 1] == '/'
6384 || tmpdir
[strlen (tmpdir
) - 1] == '\\')
6389 if (tmpdir
[strlen (tmpdir
) - 1] == '/')
6393 char *templt
= concat (tmpdir
, slash
, "etXXXXXX");
6394 int fd
= mkostemp (templt
, O_CLOEXEC
);
6395 if (fd
< 0 || close (fd
) != 0)
6397 int temp_errno
= errno
;
6403 #if defined (DOS_NT)
6404 /* The file name will be used in shell redirection, so it needs to have
6405 DOS-style backslashes, or else the Windows shell will barf. */
6407 for (p
= templt
; *p
; p
++)
6415 /* Return a newly allocated string containing the file name of FILE
6416 relative to the absolute directory DIR (which should end with a slash). */
6418 relative_filename (char *file
, char *dir
)
6420 char *fp
, *dp
, *afn
, *res
;
6423 /* Find the common root of file and dir (with a trailing slash). */
6424 afn
= absolute_filename (file
, cwd
);
6427 while (*fp
++ == *dp
++)
6429 fp
--, dp
--; /* back to the first differing char */
6431 if (fp
== afn
&& afn
[0] != '/') /* cannot build a relative name */
6434 do /* look at the equal chars until '/' */
6438 /* Build a sequence of "../" strings for the resulting relative file name. */
6440 while ((dp
= strchr (dp
+ 1, '/')) != NULL
)
6442 res
= xnew (3*i
+ strlen (fp
+ 1) + 1, char);
6445 z
= stpcpy (z
, "../");
6447 /* Add the file name relative to the common root of file and dir. */
6454 /* Return a newly allocated string containing the absolute file name
6455 of FILE given DIR (which should end with a slash). */
6457 absolute_filename (char *file
, char *dir
)
6459 char *slashp
, *cp
, *res
;
6461 if (filename_is_absolute (file
))
6462 res
= savestr (file
);
6464 /* We don't support non-absolute file names with a drive
6465 letter, like `d:NAME' (it's too much hassle). */
6466 else if (file
[1] == ':')
6467 fatal ("%s: relative file names with drive letters not supported", file
);
6470 res
= concat (dir
, file
, "");
6472 /* Delete the "/dirname/.." and "/." substrings. */
6473 slashp
= strchr (res
, '/');
6474 while (slashp
!= NULL
&& slashp
[0] != '\0')
6476 if (slashp
[1] == '.')
6478 if (slashp
[2] == '.'
6479 && (slashp
[3] == '/' || slashp
[3] == '\0'))
6484 while (cp
>= res
&& !filename_is_absolute (cp
));
6486 cp
= slashp
; /* the absolute name begins with "/.." */
6488 /* Under MSDOS and NT we get `d:/NAME' as absolute
6489 file name, so the luser could say `d:/../NAME'.
6490 We silently treat this as `d:/NAME'. */
6491 else if (cp
[0] != '/')
6494 memmove (cp
, slashp
+ 3, strlen (slashp
+ 2));
6498 else if (slashp
[2] == '/' || slashp
[2] == '\0')
6500 memmove (slashp
, slashp
+ 2, strlen (slashp
+ 1));
6505 slashp
= strchr (slashp
+ 1, '/');
6508 if (res
[0] == '\0') /* just a safety net: should never happen */
6511 return savestr ("/");
6517 /* Return a newly allocated string containing the absolute
6518 file name of dir where FILE resides given DIR (which should
6519 end with a slash). */
6521 absolute_dirname (char *file
, char *dir
)
6526 slashp
= strrchr (file
, '/');
6528 return savestr (dir
);
6531 res
= absolute_filename (file
, dir
);
6537 /* Whether the argument string is an absolute file name. The argument
6538 string must have been canonicalized with canonicalize_filename. */
6540 filename_is_absolute (char *fn
)
6542 return (fn
[0] == '/'
6544 || (c_isalpha (fn
[0]) && fn
[1] == ':' && fn
[2] == '/')
6549 /* Downcase DOS drive letter and collapse separators into single slashes.
6552 canonicalize_filename (register char *fn
)
6557 /* Canonicalize drive letter case. */
6558 if (c_isupper (fn
[0]) && fn
[1] == ':')
6559 fn
[0] = c_tolower (fn
[0]);
6561 /* Collapse multiple forward- and back-slashes into a single forward
6563 for (cp
= fn
; *cp
!= '\0'; cp
++, fn
++)
6564 if (*cp
== '/' || *cp
== '\\')
6567 while (cp
[1] == '/' || cp
[1] == '\\')
6575 /* Collapse multiple slashes into a single slash. */
6576 for (cp
= fn
; *cp
!= '\0'; cp
++, fn
++)
6580 while (cp
[1] == '/')
6586 #endif /* !DOS_NT */
6592 /* Initialize a linebuffer for use. */
6594 linebuffer_init (linebuffer
*lbp
)
6596 lbp
->size
= (DEBUG
) ? 3 : 200;
6597 lbp
->buffer
= xnew (lbp
->size
, char);
6598 lbp
->buffer
[0] = '\0';
6602 /* Set the minimum size of a string contained in a linebuffer. */
6604 linebuffer_setlen (linebuffer
*lbp
, int toksize
)
6606 while (lbp
->size
<= toksize
)
6609 xrnew (lbp
->buffer
, lbp
->size
, char);
6614 /* Like malloc but get fatal error if memory is exhausted. */
6616 xmalloc (size_t size
)
6618 void *result
= malloc (size
);
6620 fatal ("virtual memory exhausted", (char *)NULL
);
6625 xrealloc (void *ptr
, size_t size
)
6627 void *result
= realloc (ptr
, size
);
6629 fatal ("virtual memory exhausted", (char *)NULL
);
6635 * indent-tabs-mode: t
6638 * c-font-lock-extra-types: ("FILE" "bool" "language" "linebuffer" "fdesc" "node" "regexp")
6639 * c-file-style: "gnu"
6643 /* etags.c ends here */