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, 1998-2016 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 (at
39 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>
118 # define O_CLOEXEC O_NOINHERIT
119 #endif /* WINDOWSNT */
126 #include <sysstdio.h>
129 #include <binary-io.h>
131 #include <c-strcase.h>
135 # undef assert /* some systems have a buggy assert.h */
136 # define assert(x) ((void) 0)
142 /* Define CTAGS to make the program "ctags" compatible with the usual one.
143 Leave it undefined to make the program "etags", which makes emacs-style
144 tag tables and tags typedefs, #defines and struct/union/enum by default. */
153 streq (char const *s
, char const *t
)
155 return strcmp (s
, t
) == 0;
159 strcaseeq (char const *s
, char const *t
)
161 return c_strcasecmp (s
, t
) == 0;
165 strneq (char const *s
, char const *t
, size_t n
)
167 return strncmp (s
, t
, n
) == 0;
171 strncaseeq (char const *s
, char const *t
, size_t n
)
173 return c_strncasecmp (s
, t
, n
) == 0;
176 /* C is not in a name. */
178 notinname (unsigned char c
)
180 /* Look at make_tag before modifying! */
181 static bool const table
[UCHAR_MAX
+ 1] = {
182 ['\0']=1, ['\t']=1, ['\n']=1, ['\f']=1, ['\r']=1, [' ']=1,
183 ['(']=1, [')']=1, [',']=1, [';']=1, ['=']=1
188 /* C can start a token. */
190 begtoken (unsigned char c
)
192 static bool const table
[UCHAR_MAX
+ 1] = {
194 ['A']=1, ['B']=1, ['C']=1, ['D']=1, ['E']=1, ['F']=1, ['G']=1, ['H']=1,
195 ['I']=1, ['J']=1, ['K']=1, ['L']=1, ['M']=1, ['N']=1, ['O']=1, ['P']=1,
196 ['Q']=1, ['R']=1, ['S']=1, ['T']=1, ['U']=1, ['V']=1, ['W']=1, ['X']=1,
199 ['a']=1, ['b']=1, ['c']=1, ['d']=1, ['e']=1, ['f']=1, ['g']=1, ['h']=1,
200 ['i']=1, ['j']=1, ['k']=1, ['l']=1, ['m']=1, ['n']=1, ['o']=1, ['p']=1,
201 ['q']=1, ['r']=1, ['s']=1, ['t']=1, ['u']=1, ['v']=1, ['w']=1, ['x']=1,
208 /* C can be in the middle of a token. */
210 intoken (unsigned char c
)
212 static bool const table
[UCHAR_MAX
+ 1] = {
214 ['0']=1, ['1']=1, ['2']=1, ['3']=1, ['4']=1,
215 ['5']=1, ['6']=1, ['7']=1, ['8']=1, ['9']=1,
216 ['A']=1, ['B']=1, ['C']=1, ['D']=1, ['E']=1, ['F']=1, ['G']=1, ['H']=1,
217 ['I']=1, ['J']=1, ['K']=1, ['L']=1, ['M']=1, ['N']=1, ['O']=1, ['P']=1,
218 ['Q']=1, ['R']=1, ['S']=1, ['T']=1, ['U']=1, ['V']=1, ['W']=1, ['X']=1,
221 ['a']=1, ['b']=1, ['c']=1, ['d']=1, ['e']=1, ['f']=1, ['g']=1, ['h']=1,
222 ['i']=1, ['j']=1, ['k']=1, ['l']=1, ['m']=1, ['n']=1, ['o']=1, ['p']=1,
223 ['q']=1, ['r']=1, ['s']=1, ['t']=1, ['u']=1, ['v']=1, ['w']=1, ['x']=1,
229 /* C can end a token. */
231 endtoken (unsigned char c
)
233 static bool const table
[UCHAR_MAX
+ 1] = {
234 ['\0']=1, ['\t']=1, ['\n']=1, ['\r']=1, [' ']=1,
235 ['!']=1, ['"']=1, ['#']=1, ['%']=1, ['&']=1, ['\'']=1, ['(']=1, [')']=1,
236 ['*']=1, ['+']=1, [',']=1, ['-']=1, ['.']=1, ['/']=1, [':']=1, [';']=1,
237 ['<']=1, ['=']=1, ['>']=1, ['?']=1, ['[']=1, [']']=1, ['^']=1,
238 ['{']=1, ['|']=1, ['}']=1, ['~']=1
244 * xnew, xrnew -- allocate, reallocate storage
246 * SYNOPSIS: Type *xnew (int n, Type);
247 * void xrnew (OldPointer, int n, Type);
249 #define xnew(n, Type) ((Type *) xmalloc ((n) * sizeof (Type)))
250 #define xrnew(op, n, Type) ((op) = (Type *) xrealloc (op, (n) * sizeof (Type)))
252 typedef void Lang_function (FILE *);
256 const char *suffix
; /* file name suffix for this compressor */
257 const char *command
; /* takes one arg and decompresses to stdout */
262 const char *name
; /* language name */
263 const char *help
; /* detailed help for the language */
264 Lang_function
*function
; /* parse function */
265 const char **suffixes
; /* name suffixes of this language's files */
266 const char **filenames
; /* names of this language's files */
267 const char **interpreters
; /* interpreters for this language */
268 bool metasource
; /* source used to generate other sources */
273 struct fdesc
*next
; /* for the linked list */
274 char *infname
; /* uncompressed input file name */
275 char *infabsname
; /* absolute uncompressed input file name */
276 char *infabsdir
; /* absolute dir of input file */
277 char *taggedfname
; /* file name to write in tagfile */
278 language
*lang
; /* language of file */
279 char *prop
; /* file properties to write in tagfile */
280 bool usecharno
; /* etags tags shall contain char number */
281 bool written
; /* entry written in the tags file */
284 typedef struct node_st
285 { /* sorting structure */
286 struct node_st
*left
, *right
; /* left and right sons */
287 fdesc
*fdp
; /* description of file to whom tag belongs */
288 char *name
; /* tag name */
289 char *regex
; /* search regexp */
290 bool valid
; /* write this tag on the tag file */
291 bool is_func
; /* function tag: use regexp in CTAGS mode */
292 bool been_warned
; /* warning already given for duplicated tag */
293 int lno
; /* line number tag is on */
294 long cno
; /* character number line starts on */
298 * A `linebuffer' is a structure which holds a line of text.
299 * `readline_internal' reads a line from a stream into a linebuffer
300 * and works regardless of the length of the line.
301 * SIZE is the size of BUFFER, LEN is the length of the string in
302 * BUFFER after readline reads it.
311 /* Used to support mixing of --lang and file names. */
315 at_language
, /* a language specification */
316 at_regexp
, /* a regular expression */
317 at_filename
, /* a file name */
318 at_stdin
, /* read from stdin here */
319 at_end
/* stop parsing the list */
320 } arg_type
; /* argument type */
321 language
*lang
; /* language associated with the argument */
322 char *what
; /* the argument itself */
325 /* Structure defining a regular expression. */
326 typedef struct regexp
328 struct regexp
*p_next
; /* pointer to next in list */
329 language
*lang
; /* if set, use only for this language */
330 char *pattern
; /* the regexp pattern */
331 char *name
; /* tag name */
332 struct re_pattern_buffer
*pat
; /* the compiled pattern */
333 struct re_registers regs
; /* re registers */
334 bool error_signaled
; /* already signaled for this regexp */
335 bool force_explicit_name
; /* do not allow implicit tag name */
336 bool ignore_case
; /* ignore case when matching */
337 bool multi_line
; /* do a multi-line match on the whole file */
341 /* Many compilers barf on this:
342 Lang_function Ada_funcs;
343 so let's write it this way */
344 static void Ada_funcs (FILE *);
345 static void Asm_labels (FILE *);
346 static void C_entries (int c_ext
, FILE *);
347 static void default_C_entries (FILE *);
348 static void plain_C_entries (FILE *);
349 static void Cjava_entries (FILE *);
350 static void Cobol_paragraphs (FILE *);
351 static void Cplusplus_entries (FILE *);
352 static void Cstar_entries (FILE *);
353 static void Erlang_functions (FILE *);
354 static void Forth_words (FILE *);
355 static void Fortran_functions (FILE *);
356 static void Go_functions (FILE *);
357 static void HTML_labels (FILE *);
358 static void Lisp_functions (FILE *);
359 static void Lua_functions (FILE *);
360 static void Makefile_targets (FILE *);
361 static void Pascal_functions (FILE *);
362 static void Perl_functions (FILE *);
363 static void PHP_functions (FILE *);
364 static void PS_functions (FILE *);
365 static void Prolog_functions (FILE *);
366 static void Python_functions (FILE *);
367 static void Ruby_functions (FILE *);
368 static void Scheme_functions (FILE *);
369 static void TeX_commands (FILE *);
370 static void Texinfo_nodes (FILE *);
371 static void Yacc_entries (FILE *);
372 static void just_read_file (FILE *);
374 static language
*get_language_from_langname (const char *);
375 static void readline (linebuffer
*, FILE *);
376 static long readline_internal (linebuffer
*, FILE *, char const *);
377 static bool nocase_tail (const char *);
378 static void get_tag (char *, char **);
380 static void analyze_regex (char *);
381 static void free_regexps (void);
382 static void regex_tag_multiline (void);
383 static void error (const char *, ...) ATTRIBUTE_FORMAT_PRINTF (1, 2);
384 static void verror (char const *, va_list) ATTRIBUTE_FORMAT_PRINTF (1, 0);
385 static _Noreturn
void suggest_asking_for_help (void);
386 static _Noreturn
void fatal (char const *, ...) ATTRIBUTE_FORMAT_PRINTF (1, 2);
387 static _Noreturn
void pfatal (const char *);
388 static void add_node (node
*, node
**);
390 static void process_file_name (char *, language
*);
391 static void process_file (FILE *, char *, language
*);
392 static void find_entries (FILE *);
393 static void free_tree (node
*);
394 static void free_fdesc (fdesc
*);
395 static void pfnote (char *, bool, char *, int, int, long);
396 static void invalidate_nodes (fdesc
*, node
**);
397 static void put_entries (node
*);
399 static char *concat (const char *, const char *, const char *);
400 static char *skip_spaces (char *);
401 static char *skip_non_spaces (char *);
402 static char *skip_name (char *);
403 static char *savenstr (const char *, int);
404 static char *savestr (const char *);
405 static char *etags_getcwd (void);
406 static char *relative_filename (char *, char *);
407 static char *absolute_filename (char *, char *);
408 static char *absolute_dirname (char *, char *);
409 static bool filename_is_absolute (char *f
);
410 static void canonicalize_filename (char *);
411 static char *etags_mktmp (void);
412 static void linebuffer_init (linebuffer
*);
413 static void linebuffer_setlen (linebuffer
*, int);
414 static void *xmalloc (size_t);
415 static void *xrealloc (void *, size_t);
418 static char searchar
= '/'; /* use /.../ searches */
420 static char *tagfile
; /* output file */
421 static char *progname
; /* name this program was invoked with */
422 static char *cwd
; /* current working directory */
423 static char *tagfiledir
; /* directory of tagfile */
424 static FILE *tagf
; /* ioptr for tags file */
425 static ptrdiff_t whatlen_max
; /* maximum length of any 'what' member */
427 static fdesc
*fdhead
; /* head of file description list */
428 static fdesc
*curfdp
; /* current file description */
429 static char *infilename
; /* current input file name */
430 static int lineno
; /* line number of current line */
431 static long charno
; /* current character number */
432 static long linecharno
; /* charno of start of current line */
433 static char *dbp
; /* pointer to start of current tag */
435 static const int invalidcharno
= -1;
437 static node
*nodehead
; /* the head of the binary tree of tags */
438 static node
*last_node
; /* the last node created */
440 static linebuffer lb
; /* the current line */
441 static linebuffer filebuf
; /* a buffer containing the whole file */
442 static linebuffer token_name
; /* a buffer containing a tag name */
444 static bool append_to_tagfile
; /* -a: append to tags */
445 /* The next five default to true in C and derived languages. */
446 static bool typedefs
; /* -t: create tags for C and Ada typedefs */
447 static bool typedefs_or_cplusplus
; /* -T: create tags for C typedefs, level */
448 /* 0 struct/enum/union decls, and C++ */
449 /* member functions. */
450 static bool constantypedefs
; /* -d: create tags for C #define, enum */
451 /* constants and variables. */
452 /* -D: opposite of -d. Default under ctags. */
453 static int globals
; /* create tags for global variables */
454 static int members
; /* create tags for C member variables */
455 static int declarations
; /* --declarations: tag them and extern in C&Co*/
456 static int no_line_directive
; /* ignore #line directives (undocumented) */
457 static int no_duplicates
; /* no duplicate tags for ctags (undocumented) */
458 static bool update
; /* -u: update tags */
459 static bool vgrind_style
; /* -v: create vgrind style index output */
460 static bool no_warnings
; /* -w: suppress warnings (undocumented) */
461 static bool cxref_style
; /* -x: create cxref style output */
462 static bool cplusplus
; /* .[hc] means C++, not C (undocumented) */
463 static bool ignoreindent
; /* -I: ignore indentation in C */
464 static int packages_only
; /* --packages-only: in Ada, only tag packages*/
465 static int class_qualify
; /* -Q: produce class-qualified tags in C++/Java */
467 /* STDIN is defined in LynxOS system headers */
472 #define STDIN 0x1001 /* returned by getopt_long on --parse-stdin */
473 static bool parsing_stdin
; /* --parse-stdin used */
475 static regexp
*p_head
; /* list of all regexps */
476 static bool need_filebuf
; /* some regexes are multi-line */
478 static struct option longopts
[] =
480 { "append", no_argument
, NULL
, 'a' },
481 { "packages-only", no_argument
, &packages_only
, 1 },
482 { "c++", no_argument
, NULL
, 'C' },
483 { "declarations", no_argument
, &declarations
, 1 },
484 { "no-line-directive", no_argument
, &no_line_directive
, 1 },
485 { "no-duplicates", no_argument
, &no_duplicates
, 1 },
486 { "help", no_argument
, NULL
, 'h' },
487 { "help", no_argument
, NULL
, 'H' },
488 { "ignore-indentation", no_argument
, NULL
, 'I' },
489 { "language", required_argument
, NULL
, 'l' },
490 { "members", no_argument
, &members
, 1 },
491 { "no-members", no_argument
, &members
, 0 },
492 { "output", required_argument
, NULL
, 'o' },
493 { "class-qualify", no_argument
, &class_qualify
, 'Q' },
494 { "regex", required_argument
, NULL
, 'r' },
495 { "no-regex", no_argument
, NULL
, 'R' },
496 { "ignore-case-regex", required_argument
, NULL
, 'c' },
497 { "parse-stdin", required_argument
, NULL
, STDIN
},
498 { "version", no_argument
, NULL
, 'V' },
500 #if CTAGS /* Ctags options */
501 { "backward-search", no_argument
, NULL
, 'B' },
502 { "cxref", no_argument
, NULL
, 'x' },
503 { "defines", no_argument
, NULL
, 'd' },
504 { "globals", no_argument
, &globals
, 1 },
505 { "typedefs", no_argument
, NULL
, 't' },
506 { "typedefs-and-c++", no_argument
, NULL
, 'T' },
507 { "update", no_argument
, NULL
, 'u' },
508 { "vgrind", no_argument
, NULL
, 'v' },
509 { "no-warn", no_argument
, NULL
, 'w' },
511 #else /* Etags options */
512 { "no-defines", no_argument
, NULL
, 'D' },
513 { "no-globals", no_argument
, &globals
, 0 },
514 { "include", required_argument
, NULL
, 'i' },
519 static compressor compressors
[] =
521 { "z", "gzip -d -c"},
522 { "Z", "gzip -d -c"},
523 { "gz", "gzip -d -c"},
524 { "GZ", "gzip -d -c"},
525 { "bz2", "bzip2 -d -c" },
526 { "xz", "xz -d -c" },
535 static const char *Ada_suffixes
[] =
536 { "ads", "adb", "ada", NULL
};
537 static const char Ada_help
[] =
538 "In Ada code, functions, procedures, packages, tasks and types are\n\
539 tags. Use the '--packages-only' option to create tags for\n\
541 Ada tag names have suffixes indicating the type of entity:\n\
542 Entity type: Qualifier:\n\
543 ------------ ----------\n\
550 Thus, 'M-x find-tag <RET> bidule/b <RET>' will go directly to the\n\
551 body of the package 'bidule', while 'M-x find-tag <RET> bidule <RET>'\n\
552 will just search for any tag 'bidule'.";
555 static const char *Asm_suffixes
[] =
556 { "a", /* Unix assembler */
557 "asm", /* Microcontroller assembly */
558 "def", /* BSO/Tasking definition includes */
559 "inc", /* Microcontroller include files */
560 "ins", /* Microcontroller include files */
561 "s", "sa", /* Unix assembler */
562 "S", /* cpp-processed Unix assembler */
563 "src", /* BSO/Tasking C compiler output */
566 static const char Asm_help
[] =
567 "In assembler code, labels appearing at the beginning of a line,\n\
568 followed by a colon, are tags.";
571 /* Note that .c and .h can be considered C++, if the --c++ flag was
572 given, or if the `class' or `template' keywords are met inside the file.
573 That is why default_C_entries is called for these. */
574 static const char *default_C_suffixes
[] =
576 #if CTAGS /* C help for Ctags */
577 static const char default_C_help
[] =
578 "In C code, any C function is a tag. Use -t to tag typedefs.\n\
579 Use -T to tag definitions of 'struct', 'union' and 'enum'.\n\
580 Use -d to tag '#define' macro definitions and 'enum' constants.\n\
581 Use --globals to tag global variables.\n\
582 You can tag function declarations and external variables by\n\
583 using '--declarations', and struct members by using '--members'.";
584 #else /* C help for Etags */
585 static const char default_C_help
[] =
586 "In C code, any C function or typedef is a tag, and so are\n\
587 definitions of 'struct', 'union' and 'enum'. '#define' macro\n\
588 definitions and 'enum' constants are tags unless you specify\n\
589 '--no-defines'. Global variables are tags unless you specify\n\
590 '--no-globals' and so are struct members unless you specify\n\
591 '--no-members'. Use of '--no-globals', '--no-defines' and\n\
592 '--no-members' can make the tags table file much smaller.\n\
593 You can tag function declarations and external variables by\n\
594 using '--declarations'.";
595 #endif /* C help for Ctags and Etags */
597 static const char *Cplusplus_suffixes
[] =
598 { "C", "c++", "cc", "cpp", "cxx", "H", "h++", "hh", "hpp", "hxx",
599 "M", /* Objective C++ */
600 "pdb", /* PostScript with C syntax */
602 static const char Cplusplus_help
[] =
603 "In C++ code, all the tag constructs of C code are tagged. (Use\n\
604 --help --lang=c --lang=c++ for full help.)\n\
605 In addition to C tags, member functions are also recognized. Member\n\
606 variables are recognized unless you use the '--no-members' option.\n\
607 Tags for variables and functions in classes are named 'CLASS::VARIABLE'\n\
608 and 'CLASS::FUNCTION'. 'operator' definitions have tag names like\n\
611 static const char *Cjava_suffixes
[] =
613 static char Cjava_help
[] =
614 "In Java code, all the tags constructs of C and C++ code are\n\
615 tagged. (Use --help --lang=c --lang=c++ --lang=java for full help.)";
618 static const char *Cobol_suffixes
[] =
619 { "COB", "cob", NULL
};
620 static char Cobol_help
[] =
621 "In Cobol code, tags are paragraph names; that is, any word\n\
622 starting in column 8 and followed by a period.";
624 static const char *Cstar_suffixes
[] =
625 { "cs", "hs", NULL
};
627 static const char *Erlang_suffixes
[] =
628 { "erl", "hrl", NULL
};
629 static const char Erlang_help
[] =
630 "In Erlang code, the tags are the functions, records and macros\n\
631 defined in the file.";
633 const char *Forth_suffixes
[] =
634 { "fth", "tok", NULL
};
635 static const char Forth_help
[] =
636 "In Forth code, tags are words defined by ':',\n\
637 constant, code, create, defer, value, variable, buffer:, field.";
639 static const char *Fortran_suffixes
[] =
640 { "F", "f", "f90", "for", NULL
};
641 static const char Fortran_help
[] =
642 "In Fortran code, functions, subroutines and block data are tags.";
644 static const char *Go_suffixes
[] = {"go", NULL
};
645 static const char Go_help
[] =
646 "In Go code, functions, interfaces and packages are tags.";
648 static const char *HTML_suffixes
[] =
649 { "htm", "html", "shtml", NULL
};
650 static const char HTML_help
[] =
651 "In HTML input files, the tags are the 'title' and the 'h1', 'h2',\n\
652 'h3' headers. Also, tags are 'name=' in anchors and all\n\
653 occurrences of 'id='.";
655 static const char *Lisp_suffixes
[] =
656 { "cl", "clisp", "el", "l", "lisp", "LSP", "lsp", "ml", NULL
};
657 static const char Lisp_help
[] =
658 "In Lisp code, any function defined with 'defun', any variable\n\
659 defined with 'defvar' or 'defconst', and in general the first\n\
660 argument of any expression that starts with '(def' in column zero\n\
662 The '--declarations' option tags \"(defvar foo)\" constructs too.";
664 static const char *Lua_suffixes
[] =
665 { "lua", "LUA", NULL
};
666 static const char Lua_help
[] =
667 "In Lua scripts, all functions are tags.";
669 static const char *Makefile_filenames
[] =
670 { "Makefile", "makefile", "GNUMakefile", "Makefile.in", "Makefile.am", NULL
};
671 static const char Makefile_help
[] =
672 "In makefiles, targets are tags; additionally, variables are tags\n\
673 unless you specify '--no-globals'.";
675 static const char *Objc_suffixes
[] =
676 { "lm", /* Objective lex file */
677 "m", /* Objective C file */
679 static const char Objc_help
[] =
680 "In Objective C code, tags include Objective C definitions for classes,\n\
681 class categories, methods and protocols. Tags for variables and\n\
682 functions in classes are named 'CLASS::VARIABLE' and 'CLASS::FUNCTION'.\
683 \n(Use --help --lang=c --lang=objc --lang=java for full help.)";
685 static const char *Pascal_suffixes
[] =
686 { "p", "pas", NULL
};
687 static const char Pascal_help
[] =
688 "In Pascal code, the tags are the functions and procedures defined\n\
690 /* " // this is for working around an Emacs highlighting bug... */
692 static const char *Perl_suffixes
[] =
693 { "pl", "pm", NULL
};
694 static const char *Perl_interpreters
[] =
695 { "perl", "@PERL@", NULL
};
696 static const char Perl_help
[] =
697 "In Perl code, the tags are the packages, subroutines and variables\n\
698 defined by the 'package', 'sub', 'my' and 'local' keywords. Use\n\
699 '--globals' if you want to tag global variables. Tags for\n\
700 subroutines are named 'PACKAGE::SUB'. The name for subroutines\n\
701 defined in the default package is 'main::SUB'.";
703 static const char *PHP_suffixes
[] =
704 { "php", "php3", "php4", NULL
};
705 static const char PHP_help
[] =
706 "In PHP code, tags are functions, classes and defines. Unless you use\n\
707 the '--no-members' option, vars are tags too.";
709 static const char *plain_C_suffixes
[] =
710 { "pc", /* Pro*C file */
713 static const char *PS_suffixes
[] =
714 { "ps", "psw", NULL
}; /* .psw is for PSWrap */
715 static const char PS_help
[] =
716 "In PostScript code, the tags are the functions.";
718 static const char *Prolog_suffixes
[] =
720 static const char Prolog_help
[] =
721 "In Prolog code, tags are predicates and rules at the beginning of\n\
724 static const char *Python_suffixes
[] =
726 static const char Python_help
[] =
727 "In Python code, 'def' or 'class' at the beginning of a line\n\
730 static const char *Ruby_suffixes
[] =
731 { "rb", "ru", "rbw", NULL
};
732 static const char *Ruby_filenames
[] =
733 { "Rakefile", "Thorfile", NULL
};
734 static const char Ruby_help
[] =
735 "In Ruby code, 'def' or 'class' or 'module' at the beginning of\n\
736 a line generate a tag. Constants also generate a tag.";
738 /* Can't do the `SCM' or `scm' prefix with a version number. */
739 static const char *Scheme_suffixes
[] =
740 { "oak", "sch", "scheme", "SCM", "scm", "SM", "sm", "ss", "t", NULL
};
741 static const char Scheme_help
[] =
742 "In Scheme code, tags include anything defined with 'def' or with a\n\
743 construct whose name starts with 'def'. They also include\n\
744 variables set with 'set!' at top level in the file.";
746 static const char *TeX_suffixes
[] =
747 { "bib", "clo", "cls", "ltx", "sty", "TeX", "tex", NULL
};
748 static const char TeX_help
[] =
749 "In LaTeX text, the argument of any of the commands '\\chapter',\n\
750 '\\section', '\\subsection', '\\subsubsection', '\\eqno', '\\label',\n\
751 '\\ref', '\\cite', '\\bibitem', '\\part', '\\appendix', '\\entry',\n\
752 '\\index', '\\def', '\\newcommand', '\\renewcommand',\n\
753 '\\newenvironment' or '\\renewenvironment' is a tag.\n\
755 Other commands can be specified by setting the environment variable\n\
756 'TEXTAGS' to a colon-separated list like, for example,\n\
757 TEXTAGS=\"mycommand:myothercommand\".";
760 static const char *Texinfo_suffixes
[] =
761 { "texi", "texinfo", "txi", NULL
};
762 static const char Texinfo_help
[] =
763 "for texinfo files, lines starting with @node are tagged.";
765 static const char *Yacc_suffixes
[] =
766 { "y", "y++", "ym", "yxx", "yy", NULL
}; /* .ym is Objective yacc file */
767 static const char Yacc_help
[] =
768 "In Bison or Yacc input files, each rule defines as a tag the\n\
769 nonterminal it constructs. The portions of the file that contain\n\
770 C code are parsed as C code (use --help --lang=c --lang=yacc\n\
773 static const char auto_help
[] =
774 "'auto' is not a real language, it indicates to use\n\
775 a default language for files base on file name suffix and file contents.";
777 static const char none_help
[] =
778 "'none' is not a real language, it indicates to only do\n\
779 regexp processing on files.";
781 static const char no_lang_help
[] =
782 "No detailed help available for this language.";
786 * Table of languages.
788 * It is ok for a given function to be listed under more than one
789 * name. I just didn't.
792 static language lang_names
[] =
794 { "ada", Ada_help
, Ada_funcs
, Ada_suffixes
},
795 { "asm", Asm_help
, Asm_labels
, Asm_suffixes
},
796 { "c", default_C_help
, default_C_entries
, default_C_suffixes
},
797 { "c++", Cplusplus_help
, Cplusplus_entries
, Cplusplus_suffixes
},
798 { "c*", no_lang_help
, Cstar_entries
, Cstar_suffixes
},
799 { "cobol", Cobol_help
, Cobol_paragraphs
, Cobol_suffixes
},
800 { "erlang", Erlang_help
, Erlang_functions
, Erlang_suffixes
},
801 { "forth", Forth_help
, Forth_words
, Forth_suffixes
},
802 { "fortran", Fortran_help
, Fortran_functions
, Fortran_suffixes
},
803 { "go", Go_help
, Go_functions
, Go_suffixes
},
804 { "html", HTML_help
, HTML_labels
, HTML_suffixes
},
805 { "java", Cjava_help
, Cjava_entries
, Cjava_suffixes
},
806 { "lisp", Lisp_help
, Lisp_functions
, Lisp_suffixes
},
807 { "lua", Lua_help
, Lua_functions
, Lua_suffixes
},
808 { "makefile", Makefile_help
,Makefile_targets
,NULL
,Makefile_filenames
},
809 { "objc", Objc_help
, plain_C_entries
, Objc_suffixes
},
810 { "pascal", Pascal_help
, Pascal_functions
, Pascal_suffixes
},
811 { "perl",Perl_help
,Perl_functions
,Perl_suffixes
,NULL
,Perl_interpreters
},
812 { "php", PHP_help
, PHP_functions
, PHP_suffixes
},
813 { "postscript",PS_help
, PS_functions
, PS_suffixes
},
814 { "proc", no_lang_help
, plain_C_entries
, plain_C_suffixes
},
815 { "prolog", Prolog_help
, Prolog_functions
, Prolog_suffixes
},
816 { "python", Python_help
, Python_functions
, Python_suffixes
},
817 { "ruby", Ruby_help
,Ruby_functions
,Ruby_suffixes
,Ruby_filenames
},
818 { "scheme", Scheme_help
, Scheme_functions
, Scheme_suffixes
},
819 { "tex", TeX_help
, TeX_commands
, TeX_suffixes
},
820 { "texinfo", Texinfo_help
, Texinfo_nodes
, Texinfo_suffixes
},
821 { "yacc", Yacc_help
,Yacc_entries
,Yacc_suffixes
,NULL
,NULL
,true},
822 { "auto", auto_help
}, /* default guessing scheme */
823 { "none", none_help
, just_read_file
}, /* regexp matching only */
824 { NULL
} /* end of list */
829 print_language_names (void)
832 const char **name
, **ext
;
834 puts ("\nThese are the currently supported languages, along with the\n\
835 default file names and dot suffixes:");
836 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
838 printf (" %-*s", 10, lang
->name
);
839 if (lang
->filenames
!= NULL
)
840 for (name
= lang
->filenames
; *name
!= NULL
; name
++)
841 printf (" %s", *name
);
842 if (lang
->suffixes
!= NULL
)
843 for (ext
= lang
->suffixes
; *ext
!= NULL
; ext
++)
844 printf (" .%s", *ext
);
847 puts ("where 'auto' means use default language for files based on file\n\
848 name suffix, and 'none' means only do regexp processing on files.\n\
849 If no language is specified and no matching suffix is found,\n\
850 the first line of the file is read for a sharp-bang (#!) sequence\n\
851 followed by the name of an interpreter. If no such sequence is found,\n\
852 Fortran is tried first; if no tags are found, C is tried next.\n\
853 When parsing any C file, a \"class\" or \"template\" keyword\n\
855 puts ("Compressed files are supported using gzip, bzip2, and xz.\n\
857 For detailed help on a given language use, for example,\n\
858 etags --help --lang=ada.");
862 # define EMACS_NAME "standalone"
865 # define VERSION "17.38.1.4"
867 static _Noreturn
void
870 char emacs_copyright
[] = COPYRIGHT
;
872 printf ("%s (%s %s)\n", (CTAGS
) ? "ctags" : "etags", EMACS_NAME
, VERSION
);
873 puts (emacs_copyright
);
874 puts ("This program is distributed under the terms in ETAGS.README");
879 #ifndef PRINT_UNDOCUMENTED_OPTIONS_HELP
880 # define PRINT_UNDOCUMENTED_OPTIONS_HELP false
883 static _Noreturn
void
884 print_help (argument
*argbuffer
)
886 bool help_for_lang
= false;
888 for (; argbuffer
->arg_type
!= at_end
; argbuffer
++)
889 if (argbuffer
->arg_type
== at_language
)
893 puts (argbuffer
->lang
->help
);
894 help_for_lang
= true;
900 printf ("Usage: %s [options] [[regex-option ...] file-name] ...\n\
902 These are the options accepted by %s.\n", progname
, progname
);
903 puts ("You may use unambiguous abbreviations for the long option names.");
904 puts (" A - as file name means read names from stdin (one per line).\n\
905 Absolute names are stored in the output file as they are.\n\
906 Relative ones are stored relative to the output file's directory.\n");
908 puts ("-a, --append\n\
909 Append tag entries to existing tags file.");
911 puts ("--packages-only\n\
912 For Ada files, only generate tags for packages.");
915 puts ("-B, --backward-search\n\
916 Write the search commands for the tag entries using '?', the\n\
917 backward-search command instead of '/', the forward-search command.");
919 /* This option is mostly obsolete, because etags can now automatically
920 detect C++. Retained for backward compatibility and for debugging and
921 experimentation. In principle, we could want to tag as C++ even
922 before any "class" or "template" keyword.
924 Treat files whose name suffix defaults to C language as C++ files.");
927 puts ("--declarations\n\
928 In C and derived languages, create tags for function declarations,");
930 puts ("\tand create tags for extern variables if --globals is used.");
933 ("\tand create tags for extern variables unless --no-globals is used.");
936 puts ("-d, --defines\n\
937 Create tag entries for C #define constants and enum constants, too.");
939 puts ("-D, --no-defines\n\
940 Don't create tag entries for C #define constants and enum constants.\n\
941 This makes the tags file smaller.");
944 puts ("-i FILE, --include=FILE\n\
945 Include a note in tag file indicating that, when searching for\n\
946 a tag, one should also consult the tags file FILE after\n\
947 checking the current file.");
949 puts ("-l LANG, --language=LANG\n\
950 Force the following files to be considered as written in the\n\
951 named language up to the next --language=LANG option.");
955 Create tag entries for global variables in some languages.");
957 puts ("--no-globals\n\
958 Do not create tag entries for global variables in some\n\
959 languages. This makes the tags file smaller.");
961 puts ("--no-line-directive\n\
962 Ignore #line preprocessor directives in C and derived languages.");
966 Create tag entries for members of structures in some languages.");
968 puts ("--no-members\n\
969 Do not create tag entries for members of structures\n\
970 in some languages.");
972 puts ("-Q, --class-qualify\n\
973 Qualify tag names with their class name in C++, ObjC, Java, and Perl.\n\
974 This produces tag names of the form \"class::member\" for C++,\n\
975 \"class(category)\" for Objective C, and \"class.member\" for Java.\n\
976 For Objective C, this also produces class methods qualified with\n\
977 their arguments, as in \"foo:bar:baz:more\".\n\
978 For Perl, this produces \"package::member\".");
979 puts ("-r REGEXP, --regex=REGEXP or --regex=@regexfile\n\
980 Make a tag for each line matching a regular expression pattern\n\
981 in the following files. {LANGUAGE}REGEXP uses REGEXP for LANGUAGE\n\
982 files only. REGEXFILE is a file containing one REGEXP per line.\n\
983 REGEXP takes the form /TAGREGEXP/TAGNAME/MODS, where TAGNAME/ is\n\
984 optional. The TAGREGEXP pattern is anchored (as if preceded by ^).");
985 puts (" If TAGNAME/ is present, the tags created are named.\n\
986 For example Tcl named tags can be created with:\n\
987 --regex=\"/proc[ \\t]+\\([^ \\t]+\\)/\\1/.\".\n\
988 MODS are optional one-letter modifiers: 'i' means to ignore case,\n\
989 'm' means to allow multi-line matches, 's' implies 'm' and\n\
990 causes dot to match any character, including newline.");
992 puts ("-R, --no-regex\n\
993 Don't create tags from regexps for the following files.");
995 puts ("-I, --ignore-indentation\n\
996 In C and C++ do not assume that a closing brace in the first\n\
997 column is the final brace of a function or structure definition.");
999 puts ("-o FILE, --output=FILE\n\
1000 Write the tags to FILE.");
1002 puts ("--parse-stdin=NAME\n\
1003 Read from standard input and record tags as belonging to file NAME.");
1007 puts ("-t, --typedefs\n\
1008 Generate tag entries for C and Ada typedefs.");
1009 puts ("-T, --typedefs-and-c++\n\
1010 Generate tag entries for C typedefs, C struct/enum/union tags,\n\
1011 and C++ member functions.");
1015 puts ("-u, --update\n\
1016 Update the tag entries for the given files, leaving tag\n\
1017 entries for other files in place. Currently, this is\n\
1018 implemented by deleting the existing entries for the given\n\
1019 files and then rewriting the new entries at the end of the\n\
1020 tags file. It is often faster to simply rebuild the entire\n\
1021 tag file than to use this.");
1025 puts ("-v, --vgrind\n\
1026 Print on the standard output an index of items intended for\n\
1027 human consumption, similar to the output of vgrind. The index\n\
1028 is sorted, and gives the page number of each item.");
1030 if (PRINT_UNDOCUMENTED_OPTIONS_HELP
)
1031 puts ("-w, --no-duplicates\n\
1032 Do not create duplicate tag entries, for compatibility with\n\
1033 traditional ctags.");
1035 if (PRINT_UNDOCUMENTED_OPTIONS_HELP
)
1036 puts ("-w, --no-warn\n\
1037 Suppress warning messages about duplicate tag entries.");
1039 puts ("-x, --cxref\n\
1040 Like --vgrind, but in the style of cxref, rather than vgrind.\n\
1041 The output uses line numbers instead of page numbers, but\n\
1042 beyond that the differences are cosmetic; try both to see\n\
1046 puts ("-V, --version\n\
1047 Print the version of the program.\n\
1049 Print this help message.\n\
1050 Followed by one or more '--language' options prints detailed\n\
1051 help about tag generation for the specified languages.");
1053 print_language_names ();
1056 puts ("Report bugs to bug-gnu-emacs@gnu.org");
1058 exit (EXIT_SUCCESS
);
1063 main (int argc
, char **argv
)
1066 unsigned int nincluded_files
;
1067 char **included_files
;
1068 argument
*argbuffer
;
1069 int current_arg
, file_count
;
1070 linebuffer filename_lb
;
1071 bool help_asked
= false;
1077 nincluded_files
= 0;
1078 included_files
= xnew (argc
, char *);
1082 /* Allocate enough no matter what happens. Overkill, but each one
1084 argbuffer
= xnew (argc
, argument
);
1087 * Always find typedefs and structure tags.
1088 * Also default to find macro constants, enum constants, struct
1089 * members and global variables. Do it for both etags and ctags.
1091 typedefs
= typedefs_or_cplusplus
= constantypedefs
= true;
1092 globals
= members
= true;
1094 /* When the optstring begins with a '-' getopt_long does not rearrange the
1095 non-options arguments to be at the end, but leaves them alone. */
1096 optstring
= concat ("-ac:Cf:Il:o:Qr:RSVhH",
1097 (CTAGS
) ? "BxdtTuvw" : "Di:",
1100 while ((opt
= getopt_long (argc
, argv
, optstring
, longopts
, NULL
)) != EOF
)
1104 /* If getopt returns 0, then it has already processed a
1105 long-named option. We should do nothing. */
1109 /* This means that a file name has been seen. Record it. */
1110 argbuffer
[current_arg
].arg_type
= at_filename
;
1111 argbuffer
[current_arg
].what
= optarg
;
1112 len
= strlen (optarg
);
1113 if (whatlen_max
< len
)
1120 /* Parse standard input. Idea by Vivek <vivek@etla.org>. */
1121 argbuffer
[current_arg
].arg_type
= at_stdin
;
1122 argbuffer
[current_arg
].what
= optarg
;
1123 len
= strlen (optarg
);
1124 if (whatlen_max
< len
)
1129 fatal ("cannot parse standard input more than once");
1130 parsing_stdin
= true;
1133 /* Common options. */
1134 case 'a': append_to_tagfile
= true; break;
1135 case 'C': cplusplus
= true; break;
1136 case 'f': /* for compatibility with old makefiles */
1140 error ("-o option may only be given once.");
1141 suggest_asking_for_help ();
1147 case 'S': /* for backward compatibility */
1148 ignoreindent
= true;
1152 language
*lang
= get_language_from_langname (optarg
);
1155 argbuffer
[current_arg
].lang
= lang
;
1156 argbuffer
[current_arg
].arg_type
= at_language
;
1162 /* Backward compatibility: support obsolete --ignore-case-regexp. */
1163 optarg
= concat (optarg
, "i", ""); /* memory leak here */
1166 argbuffer
[current_arg
].arg_type
= at_regexp
;
1167 argbuffer
[current_arg
].what
= optarg
;
1168 len
= strlen (optarg
);
1169 if (whatlen_max
< len
)
1174 argbuffer
[current_arg
].arg_type
= at_regexp
;
1175 argbuffer
[current_arg
].what
= NULL
;
1190 case 'D': constantypedefs
= false; break;
1191 case 'i': included_files
[nincluded_files
++] = optarg
; break;
1193 /* Ctags options. */
1194 case 'B': searchar
= '?'; break;
1195 case 'd': constantypedefs
= true; break;
1196 case 't': typedefs
= true; break;
1197 case 'T': typedefs
= typedefs_or_cplusplus
= true; break;
1198 case 'u': update
= true; break;
1199 case 'v': vgrind_style
= true; /*FALLTHRU*/
1200 case 'x': cxref_style
= true; break;
1201 case 'w': no_warnings
= true; break;
1203 suggest_asking_for_help ();
1207 /* No more options. Store the rest of arguments. */
1208 for (; optind
< argc
; optind
++)
1210 argbuffer
[current_arg
].arg_type
= at_filename
;
1211 argbuffer
[current_arg
].what
= argv
[optind
];
1212 len
= strlen (argv
[optind
]);
1213 if (whatlen_max
< len
)
1219 argbuffer
[current_arg
].arg_type
= at_end
;
1222 print_help (argbuffer
);
1225 if (nincluded_files
== 0 && file_count
== 0)
1227 error ("no input files specified.");
1228 suggest_asking_for_help ();
1232 if (tagfile
== NULL
)
1233 tagfile
= savestr (CTAGS
? "tags" : "TAGS");
1234 cwd
= etags_getcwd (); /* the current working directory */
1235 if (cwd
[strlen (cwd
) - 1] != '/')
1238 cwd
= concat (oldcwd
, "/", "");
1242 /* Compute base directory for relative file names. */
1243 if (streq (tagfile
, "-")
1244 || strneq (tagfile
, "/dev/", 5))
1245 tagfiledir
= cwd
; /* relative file names are relative to cwd */
1248 canonicalize_filename (tagfile
);
1249 tagfiledir
= absolute_dirname (tagfile
, cwd
);
1252 linebuffer_init (&lb
);
1253 linebuffer_init (&filename_lb
);
1254 linebuffer_init (&filebuf
);
1255 linebuffer_init (&token_name
);
1259 if (streq (tagfile
, "-"))
1262 SET_BINARY (fileno (stdout
));
1265 tagf
= fopen (tagfile
, append_to_tagfile
? "ab" : "wb");
1271 * Loop through files finding functions.
1273 for (i
= 0; i
< current_arg
; i
++)
1275 static language
*lang
; /* non-NULL if language is forced */
1278 switch (argbuffer
[i
].arg_type
)
1281 lang
= argbuffer
[i
].lang
;
1284 analyze_regex (argbuffer
[i
].what
);
1287 this_file
= argbuffer
[i
].what
;
1288 /* Input file named "-" means read file names from stdin
1289 (one per line) and use them. */
1290 if (streq (this_file
, "-"))
1293 fatal ("cannot parse standard input "
1294 "AND read file names from it");
1295 while (readline_internal (&filename_lb
, stdin
, "-") > 0)
1296 process_file_name (filename_lb
.buffer
, lang
);
1299 process_file_name (this_file
, lang
);
1302 this_file
= argbuffer
[i
].what
;
1303 process_file (stdin
, this_file
, lang
);
1306 error ("internal error: arg_type");
1312 free (filebuf
.buffer
);
1313 free (token_name
.buffer
);
1315 if (!CTAGS
|| cxref_style
)
1317 /* Write the remaining tags to tagf (ETAGS) or stdout (CXREF). */
1318 put_entries (nodehead
);
1319 free_tree (nodehead
);
1325 /* Output file entries that have no tags. */
1326 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
1328 fprintf (tagf
, "\f\n%s,0\n", fdp
->taggedfname
);
1330 while (nincluded_files
-- > 0)
1331 fprintf (tagf
, "\f\n%s,include\n", *included_files
++);
1333 if (fclose (tagf
) == EOF
)
1337 exit (EXIT_SUCCESS
);
1340 /* From here on, we are in (CTAGS && !cxref_style) */
1344 xmalloc (strlen (tagfile
) + whatlen_max
+
1345 sizeof "mv..OTAGS;grep -Fv '\t\t' OTAGS >;rm OTAGS");
1346 for (i
= 0; i
< current_arg
; ++i
)
1348 switch (argbuffer
[i
].arg_type
)
1354 continue; /* the for loop */
1356 char *z
= stpcpy (cmd
, "mv ");
1357 z
= stpcpy (z
, tagfile
);
1358 z
= stpcpy (z
, " OTAGS;grep -Fv '\t");
1359 z
= stpcpy (z
, argbuffer
[i
].what
);
1360 z
= stpcpy (z
, "\t' OTAGS >");
1361 z
= stpcpy (z
, tagfile
);
1362 strcpy (z
, ";rm OTAGS");
1363 if (system (cmd
) != EXIT_SUCCESS
)
1364 fatal ("failed to execute shell command");
1367 append_to_tagfile
= true;
1370 tagf
= fopen (tagfile
, append_to_tagfile
? "ab" : "wb");
1373 put_entries (nodehead
); /* write all the tags (CTAGS) */
1374 free_tree (nodehead
);
1376 if (fclose (tagf
) == EOF
)
1380 if (append_to_tagfile
|| update
)
1382 char *cmd
= xmalloc (2 * strlen (tagfile
) + sizeof "sort -u -o..");
1383 /* Maybe these should be used:
1384 setenv ("LC_COLLATE", "C", 1);
1385 setenv ("LC_ALL", "C", 1); */
1386 char *z
= stpcpy (cmd
, "sort -u -o ");
1387 z
= stpcpy (z
, tagfile
);
1389 strcpy (z
, tagfile
);
1390 exit (system (cmd
));
1392 return EXIT_SUCCESS
;
1397 * Return a compressor given the file name. If EXTPTR is non-zero,
1398 * return a pointer into FILE where the compressor-specific
1399 * extension begins. If no compressor is found, NULL is returned
1400 * and EXTPTR is not significant.
1401 * Idea by Vladimir Alexiev <vladimir@cs.ualberta.ca> (1998)
1404 get_compressor_from_suffix (char *file
, char **extptr
)
1407 char *slash
, *suffix
;
1409 /* File has been processed by canonicalize_filename,
1410 so we don't need to consider backslashes on DOS_NT. */
1411 slash
= strrchr (file
, '/');
1412 suffix
= strrchr (file
, '.');
1413 if (suffix
== NULL
|| suffix
< slash
)
1418 /* Let those poor souls who live with DOS 8+3 file name limits get
1419 some solace by treating foo.cgz as if it were foo.c.gz, etc.
1420 Only the first do loop is run if not MSDOS */
1423 for (compr
= compressors
; compr
->suffix
!= NULL
; compr
++)
1424 if (streq (compr
->suffix
, suffix
))
1427 break; /* do it only once: not really a loop */
1430 } while (*suffix
!= '\0');
1437 * Return a language given the name.
1440 get_language_from_langname (const char *name
)
1445 error ("empty language name");
1448 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1449 if (streq (name
, lang
->name
))
1451 error ("unknown language \"%s\"", name
);
1459 * Return a language given the interpreter name.
1462 get_language_from_interpreter (char *interpreter
)
1467 if (interpreter
== NULL
)
1469 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1470 if (lang
->interpreters
!= NULL
)
1471 for (iname
= lang
->interpreters
; *iname
!= NULL
; iname
++)
1472 if (streq (*iname
, interpreter
))
1481 * Return a language given the file name.
1484 get_language_from_filename (char *file
, int case_sensitive
)
1487 const char **name
, **ext
, *suffix
;
1490 /* Try whole file name first. */
1491 slash
= strrchr (file
, '/');
1495 else if (file
[0] && file
[1] == ':')
1498 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1499 if (lang
->filenames
!= NULL
)
1500 for (name
= lang
->filenames
; *name
!= NULL
; name
++)
1501 if ((case_sensitive
)
1502 ? streq (*name
, file
)
1503 : strcaseeq (*name
, file
))
1506 /* If not found, try suffix after last dot. */
1507 suffix
= strrchr (file
, '.');
1511 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1512 if (lang
->suffixes
!= NULL
)
1513 for (ext
= lang
->suffixes
; *ext
!= NULL
; ext
++)
1514 if ((case_sensitive
)
1515 ? streq (*ext
, suffix
)
1516 : strcaseeq (*ext
, suffix
))
1523 * This routine is called on each file argument.
1526 process_file_name (char *file
, language
*lang
)
1531 char *compressed_name
, *uncompressed_name
;
1532 char *ext
, *real_name
, *tmp_name
;
1535 canonicalize_filename (file
);
1536 if (streq (file
, tagfile
) && !streq (tagfile
, "-"))
1538 error ("skipping inclusion of %s in self.", file
);
1541 compr
= get_compressor_from_suffix (file
, &ext
);
1544 compressed_name
= file
;
1545 uncompressed_name
= savenstr (file
, ext
- file
);
1549 compressed_name
= NULL
;
1550 uncompressed_name
= file
;
1553 /* If the canonicalized uncompressed name
1554 has already been dealt with, skip it silently. */
1555 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
1557 assert (fdp
->infname
!= NULL
);
1558 if (streq (uncompressed_name
, fdp
->infname
))
1562 inf
= fopen (file
, "r" FOPEN_BINARY
);
1567 int file_errno
= errno
;
1568 if (compressed_name
)
1570 /* Try with the given suffix. */
1571 inf
= fopen (uncompressed_name
, "r" FOPEN_BINARY
);
1573 real_name
= uncompressed_name
;
1577 /* Try all possible suffixes. */
1578 for (compr
= compressors
; compr
->suffix
!= NULL
; compr
++)
1580 compressed_name
= concat (file
, ".", compr
->suffix
);
1581 inf
= fopen (compressed_name
, "r" FOPEN_BINARY
);
1584 real_name
= compressed_name
;
1589 char *suf
= compressed_name
+ strlen (file
);
1590 size_t suflen
= strlen (compr
->suffix
) + 1;
1591 for ( ; suf
[1]; suf
++, suflen
--)
1593 memmove (suf
, suf
+ 1, suflen
);
1594 inf
= fopen (compressed_name
, "r" FOPEN_BINARY
);
1597 real_name
= compressed_name
;
1604 free (compressed_name
);
1605 compressed_name
= NULL
;
1616 if (real_name
== compressed_name
)
1619 tmp_name
= etags_mktmp ();
1624 #if MSDOS || defined (DOS_NT)
1625 char *cmd1
= concat (compr
->command
, " \"", real_name
);
1626 char *cmd
= concat (cmd1
, "\" > ", tmp_name
);
1628 char *cmd1
= concat (compr
->command
, " '", real_name
);
1629 char *cmd
= concat (cmd1
, "' > ", tmp_name
);
1633 if (system (cmd
) == -1)
1640 inf
= fopen (tmp_name
, "r" FOPEN_BINARY
);
1654 process_file (inf
, uncompressed_name
, lang
);
1656 retval
= fclose (inf
);
1657 if (real_name
== compressed_name
)
1666 if (compressed_name
!= file
)
1667 free (compressed_name
);
1668 if (uncompressed_name
!= file
)
1669 free (uncompressed_name
);
1676 process_file (FILE *fh
, char *fn
, language
*lang
)
1678 static const fdesc emptyfdesc
;
1682 /* Create a new input file description entry. */
1683 fdp
= xnew (1, fdesc
);
1686 fdp
->infname
= savestr (fn
);
1688 fdp
->infabsname
= absolute_filename (fn
, cwd
);
1689 fdp
->infabsdir
= absolute_dirname (fn
, cwd
);
1690 if (filename_is_absolute (fn
))
1692 /* An absolute file name. Canonicalize it. */
1693 fdp
->taggedfname
= absolute_filename (fn
, NULL
);
1697 /* A file name relative to cwd. Make it relative
1698 to the directory of the tags file. */
1699 fdp
->taggedfname
= relative_filename (fn
, tagfiledir
);
1701 fdp
->usecharno
= true; /* use char position when making tags */
1703 fdp
->written
= false; /* not written on tags file yet */
1706 curfdp
= fdhead
; /* the current file description */
1710 /* If not Ctags, and if this is not metasource and if it contained no #line
1711 directives, we can write the tags and free all nodes pointing to
1714 && curfdp
->usecharno
/* no #line directives in this file */
1715 && !curfdp
->lang
->metasource
)
1719 /* Look for the head of the sublist relative to this file. See add_node
1720 for the structure of the node tree. */
1722 for (np
= nodehead
; np
!= NULL
; prev
= np
, np
= np
->left
)
1723 if (np
->fdp
== curfdp
)
1726 /* If we generated tags for this file, write and delete them. */
1729 /* This is the head of the last sublist, if any. The following
1730 instructions depend on this being true. */
1731 assert (np
->left
== NULL
);
1733 assert (fdhead
== curfdp
);
1734 assert (last_node
->fdp
== curfdp
);
1735 put_entries (np
); /* write tags for file curfdp->taggedfname */
1736 free_tree (np
); /* remove the written nodes */
1738 nodehead
= NULL
; /* no nodes left */
1740 prev
->left
= NULL
; /* delete the pointer to the sublist */
1746 reset_input (FILE *inf
)
1748 if (fseek (inf
, 0, SEEK_SET
) != 0)
1749 perror (infilename
);
1753 * This routine opens the specified file and calls the function
1754 * which finds the function and type definitions.
1757 find_entries (FILE *inf
)
1760 language
*lang
= curfdp
->lang
;
1761 Lang_function
*parser
= NULL
;
1763 /* If user specified a language, use it. */
1764 if (lang
!= NULL
&& lang
->function
!= NULL
)
1766 parser
= lang
->function
;
1769 /* Else try to guess the language given the file name. */
1772 lang
= get_language_from_filename (curfdp
->infname
, true);
1773 if (lang
!= NULL
&& lang
->function
!= NULL
)
1775 curfdp
->lang
= lang
;
1776 parser
= lang
->function
;
1780 /* Else look for sharp-bang as the first two characters. */
1782 && readline_internal (&lb
, inf
, infilename
) > 0
1784 && lb
.buffer
[0] == '#'
1785 && lb
.buffer
[1] == '!')
1789 /* Set lp to point at the first char after the last slash in the
1790 line or, if no slashes, at the first nonblank. Then set cp to
1791 the first successive blank and terminate the string. */
1792 lp
= strrchr (lb
.buffer
+2, '/');
1796 lp
= skip_spaces (lb
.buffer
+ 2);
1797 cp
= skip_non_spaces (lp
);
1800 if (strlen (lp
) > 0)
1802 lang
= get_language_from_interpreter (lp
);
1803 if (lang
!= NULL
&& lang
->function
!= NULL
)
1805 curfdp
->lang
= lang
;
1806 parser
= lang
->function
;
1813 /* Else try to guess the language given the case insensitive file name. */
1816 lang
= get_language_from_filename (curfdp
->infname
, false);
1817 if (lang
!= NULL
&& lang
->function
!= NULL
)
1819 curfdp
->lang
= lang
;
1820 parser
= lang
->function
;
1824 /* Else try Fortran or C. */
1827 node
*old_last_node
= last_node
;
1829 curfdp
->lang
= get_language_from_langname ("fortran");
1832 if (old_last_node
== last_node
)
1833 /* No Fortran entries found. Try C. */
1836 curfdp
->lang
= get_language_from_langname (cplusplus
? "c++" : "c");
1842 if (!no_line_directive
1843 && curfdp
->lang
!= NULL
&& curfdp
->lang
->metasource
)
1844 /* It may be that this is a bingo.y file, and we already parsed a bingo.c
1845 file, or anyway we parsed a file that is automatically generated from
1846 this one. If this is the case, the bingo.c file contained #line
1847 directives that generated tags pointing to this file. Let's delete
1848 them all before parsing this file, which is the real source. */
1850 fdesc
**fdpp
= &fdhead
;
1851 while (*fdpp
!= NULL
)
1853 && streq ((*fdpp
)->taggedfname
, curfdp
->taggedfname
))
1854 /* We found one of those! We must delete both the file description
1855 and all tags referring to it. */
1857 fdesc
*badfdp
= *fdpp
;
1859 /* Delete the tags referring to badfdp->taggedfname
1860 that were obtained from badfdp->infname. */
1861 invalidate_nodes (badfdp
, &nodehead
);
1863 *fdpp
= badfdp
->next
; /* remove the bad description from the list */
1864 free_fdesc (badfdp
);
1867 fdpp
= &(*fdpp
)->next
; /* advance the list pointer */
1870 assert (parser
!= NULL
);
1872 /* Generic initializations before reading from file. */
1873 linebuffer_setlen (&filebuf
, 0); /* reset the file buffer */
1875 /* Generic initializations before parsing file with readline. */
1876 lineno
= 0; /* reset global line number */
1877 charno
= 0; /* reset global char number */
1878 linecharno
= 0; /* reset global char number of line start */
1882 regex_tag_multiline ();
1887 * Check whether an implicitly named tag should be created,
1888 * then call `pfnote'.
1889 * NAME is a string that is internally copied by this function.
1891 * TAGS format specification
1892 * Idea by Sam Kendall <kendall@mv.mv.com> (1997)
1893 * The following is explained in some more detail in etc/ETAGS.EBNF.
1895 * make_tag creates tags with "implicit tag names" (unnamed tags)
1896 * if the following are all true, assuming NONAM=" \f\t\n\r()=,;":
1897 * 1. NAME does not contain any of the characters in NONAM;
1898 * 2. LINESTART contains name as either a rightmost, or rightmost but
1899 * one character, substring;
1900 * 3. the character, if any, immediately before NAME in LINESTART must
1901 * be a character in NONAM;
1902 * 4. the character, if any, immediately after NAME in LINESTART must
1903 * also be a character in NONAM.
1905 * The implementation uses the notinname() macro, which recognizes the
1906 * characters stored in the string `nonam'.
1907 * etags.el needs to use the same characters that are in NONAM.
1910 make_tag (const char *name
, /* tag name, or NULL if unnamed */
1911 int namelen
, /* tag length */
1912 bool is_func
, /* tag is a function */
1913 char *linestart
, /* start of the line where tag is */
1914 int linelen
, /* length of the line where tag is */
1915 int lno
, /* line number */
1916 long int cno
) /* character number */
1918 bool named
= (name
!= NULL
&& namelen
> 0);
1921 if (!CTAGS
&& named
) /* maybe set named to false */
1922 /* Let's try to make an implicit tag name, that is, create an unnamed tag
1923 such that etags.el can guess a name from it. */
1926 register const char *cp
= name
;
1928 for (i
= 0; i
< namelen
; i
++)
1929 if (notinname (*cp
++))
1931 if (i
== namelen
) /* rule #1 */
1933 cp
= linestart
+ linelen
- namelen
;
1934 if (notinname (linestart
[linelen
-1]))
1935 cp
-= 1; /* rule #4 */
1936 if (cp
>= linestart
/* rule #2 */
1938 || notinname (cp
[-1])) /* rule #3 */
1939 && strneq (name
, cp
, namelen
)) /* rule #2 */
1940 named
= false; /* use implicit tag name */
1945 nname
= savenstr (name
, namelen
);
1947 pfnote (nname
, is_func
, linestart
, linelen
, lno
, cno
);
1952 pfnote (char *name
, bool is_func
, char *linestart
, int linelen
, int lno
,
1954 /* tag name, or NULL if unnamed */
1955 /* tag is a function */
1956 /* start of the line where tag is */
1957 /* length of the line where tag is */
1959 /* character number */
1963 assert (name
== NULL
|| name
[0] != '\0');
1964 if (CTAGS
&& name
== NULL
)
1967 np
= xnew (1, node
);
1969 /* If ctags mode, change name "main" to M<thisfilename>. */
1970 if (CTAGS
&& !cxref_style
&& streq (name
, "main"))
1972 char *fp
= strrchr (curfdp
->taggedfname
, '/');
1973 np
->name
= concat ("M", fp
== NULL
? curfdp
->taggedfname
: fp
+ 1, "");
1974 fp
= strrchr (np
->name
, '.');
1975 if (fp
!= NULL
&& fp
[1] != '\0' && fp
[2] == '\0')
1981 np
->been_warned
= false;
1983 np
->is_func
= is_func
;
1985 if (np
->fdp
->usecharno
)
1986 /* Our char numbers are 0-base, because of C language tradition?
1987 ctags compatibility? old versions compatibility? I don't know.
1988 Anyway, since emacs's are 1-base we expect etags.el to take care
1989 of the difference. If we wanted to have 1-based numbers, we would
1990 uncomment the +1 below. */
1991 np
->cno
= cno
/* + 1 */ ;
1993 np
->cno
= invalidcharno
;
1994 np
->left
= np
->right
= NULL
;
1995 if (CTAGS
&& !cxref_style
)
1997 if (strlen (linestart
) < 50)
1998 np
->regex
= concat (linestart
, "$", "");
2000 np
->regex
= savenstr (linestart
, 50);
2003 np
->regex
= savenstr (linestart
, linelen
);
2005 add_node (np
, &nodehead
);
2009 * Utility functions and data to avoid recursion.
2012 typedef struct stack_entry
{
2014 struct stack_entry
*next
;
2018 push_node (node
*np
, stkentry
**stack_top
)
2022 stkentry
*new = xnew (1, stkentry
);
2025 new->next
= *stack_top
;
2031 pop_node (stkentry
**stack_top
)
2037 stkentry
*old_start
= *stack_top
;
2039 ret
= (*stack_top
)->np
;
2040 *stack_top
= (*stack_top
)->next
;
2048 * emulate recursion on left children, iterate on right children.
2051 free_tree (register node
*np
)
2053 stkentry
*stack
= NULL
;
2057 /* Descent on left children. */
2060 push_node (np
, &stack
);
2063 /* Free node without left children. */
2064 node
*node_right
= np
->right
;
2070 /* Backtrack to find a node with right children, while freeing nodes
2071 that don't have right children. */
2072 while (node_right
== NULL
&& (np
= pop_node (&stack
)) != NULL
)
2074 node_right
= np
->right
;
2080 /* Free right children. */
2087 * delete a file description
2090 free_fdesc (register fdesc
*fdp
)
2092 free (fdp
->infname
);
2093 free (fdp
->infabsname
);
2094 free (fdp
->infabsdir
);
2095 free (fdp
->taggedfname
);
2102 * Adds a node to the tree of nodes. In etags mode, sort by file
2103 * name. In ctags mode, sort by tag name. Make no attempt at
2106 * add_node is the only function allowed to add nodes, so it can
2110 add_node (node
*np
, node
**cur_node_p
)
2112 node
*cur_node
= *cur_node_p
;
2114 /* Make the first node. */
2115 if (cur_node
== NULL
)
2125 /* For each file name, tags are in a linked sublist on the right
2126 pointer. The first tags of different files are a linked list
2127 on the left pointer. last_node points to the end of the last
2129 if (last_node
!= NULL
&& last_node
->fdp
== np
->fdp
)
2131 /* Let's use the same sublist as the last added node. */
2132 assert (last_node
->right
== NULL
);
2133 last_node
->right
= np
;
2138 while (cur_node
->fdp
!= np
->fdp
)
2140 if (cur_node
->left
== NULL
)
2142 /* The head of this sublist is not good for us. Let's try the
2144 cur_node
= cur_node
->left
;
2148 /* Scanning the list we found the head of a sublist which is
2149 good for us. Let's scan this sublist. */
2150 if (cur_node
->right
)
2152 cur_node
= cur_node
->right
;
2153 while (cur_node
->right
)
2154 cur_node
= cur_node
->right
;
2156 /* Make a new node in this sublist. */
2157 cur_node
->right
= np
;
2161 /* Make a new sublist. */
2162 cur_node
->left
= np
;
2166 } /* if ETAGS mode */
2170 node
**next_node
= &cur_node
;
2172 while ((cur_node
= *next_node
) != NULL
)
2174 int dif
= strcmp (np
->name
, cur_node
->name
);
2176 * If this tag name matches an existing one, then
2177 * do not add the node, but maybe print a warning.
2179 if (!dif
&& no_duplicates
)
2181 if (np
->fdp
== cur_node
->fdp
)
2186 "Duplicate entry in file %s, line %d: %s\n",
2187 np
->fdp
->infname
, lineno
, np
->name
);
2188 fprintf (stderr
, "Second entry ignored\n");
2191 else if (!cur_node
->been_warned
&& !no_warnings
)
2195 "Duplicate entry in files %s and %s: %s (Warning only)\n",
2196 np
->fdp
->infname
, cur_node
->fdp
->infname
, np
->name
);
2197 cur_node
->been_warned
= true;
2202 next_node
= dif
< 0 ? &cur_node
->left
: &cur_node
->right
;
2206 } /* if CTAGS mode */
2210 * invalidate_nodes ()
2211 * Scan the node tree and invalidate all nodes pointing to the
2212 * given file description (CTAGS case) or free them (ETAGS case).
2215 invalidate_nodes (fdesc
*badfdp
, node
**npp
)
2218 stkentry
*stack
= NULL
;
2224 /* Push all the left children on the stack. */
2225 while (np
->left
!= NULL
)
2227 push_node (np
, &stack
);
2230 /* Invalidate this node. */
2231 if (np
->fdp
== badfdp
)
2235 /* Pop nodes from stack, invalidating them, until we find one
2236 with a right child. */
2237 while ((np
= pop_node (&stack
)) != NULL
)
2239 if (np
->fdp
== badfdp
)
2241 if (np
->right
!= NULL
)
2245 /* Process the right child, if any. */
2252 node super_root
, *np_parent
= NULL
;
2254 super_root
.left
= np
;
2255 super_root
.fdp
= (fdesc
*) -1;
2260 /* Descent on left children until node with BADFP. */
2261 while (np
&& np
->fdp
!= badfdp
)
2263 assert (np
->fdp
!= NULL
);
2269 np_parent
->left
= np
->left
; /* detach subtree from the tree */
2270 np
->left
= NULL
; /* isolate it */
2271 free_tree (np
); /* free it */
2273 /* Continue with rest of tree. */
2274 np
= np_parent
? np_parent
->left
: NULL
;
2277 *npp
= super_root
.left
;
2282 static int total_size_of_entries (node
*);
2283 static int number_len (long) ATTRIBUTE_CONST
;
2285 /* Length of a non-negative number's decimal representation. */
2287 number_len (long int num
)
2290 while ((num
/= 10) > 0)
2296 * Return total number of characters that put_entries will output for
2297 * the nodes in the linked list at the right of the specified node.
2298 * This count is irrelevant with etags.el since emacs 19.34 at least,
2299 * but is still supplied for backward compatibility.
2302 total_size_of_entries (register node
*np
)
2304 register int total
= 0;
2306 for (; np
!= NULL
; np
= np
->right
)
2309 total
+= strlen (np
->regex
) + 1; /* pat\177 */
2310 if (np
->name
!= NULL
)
2311 total
+= strlen (np
->name
) + 1; /* name\001 */
2312 total
+= number_len ((long) np
->lno
) + 1; /* lno, */
2313 if (np
->cno
!= invalidcharno
) /* cno */
2314 total
+= number_len (np
->cno
);
2315 total
+= 1; /* newline */
2322 put_entry (node
*np
)
2325 static fdesc
*fdp
= NULL
;
2327 /* Output this entry */
2336 fprintf (tagf
, "\f\n%s,%d\n",
2337 fdp
->taggedfname
, total_size_of_entries (np
));
2338 fdp
->written
= true;
2340 fputs (np
->regex
, tagf
);
2341 fputc ('\177', tagf
);
2342 if (np
->name
!= NULL
)
2344 fputs (np
->name
, tagf
);
2345 fputc ('\001', tagf
);
2347 fprintf (tagf
, "%d,", np
->lno
);
2348 if (np
->cno
!= invalidcharno
)
2349 fprintf (tagf
, "%ld", np
->cno
);
2355 if (np
->name
== NULL
)
2356 error ("internal error: NULL name in ctags mode.");
2361 fprintf (stdout
, "%s %s %d\n",
2362 np
->name
, np
->fdp
->taggedfname
, (np
->lno
+ 63) / 64);
2364 fprintf (stdout
, "%-16s %3d %-16s %s\n",
2365 np
->name
, np
->lno
, np
->fdp
->taggedfname
, np
->regex
);
2369 fprintf (tagf
, "%s\t%s\t", np
->name
, np
->fdp
->taggedfname
);
2372 { /* function or #define macro with args */
2373 putc (searchar
, tagf
);
2376 for (sp
= np
->regex
; *sp
; sp
++)
2378 if (*sp
== '\\' || *sp
== searchar
)
2382 putc (searchar
, tagf
);
2385 { /* anything else; text pattern inadequate */
2386 fprintf (tagf
, "%d", np
->lno
);
2391 } /* if this node contains a valid tag */
2395 put_entries (node
*np
)
2397 stkentry
*stack
= NULL
;
2406 /* Stack subentries that precede this one. */
2409 push_node (np
, &stack
);
2412 /* Output this subentry. */
2414 /* Stack subentries that follow this one. */
2417 /* Output subentries that precede the next one. */
2418 np
= pop_node (&stack
);
2429 push_node (np
, &stack
);
2430 while ((np
= pop_node (&stack
)) != NULL
)
2432 /* Output this subentry. */
2436 /* Output subentries that follow this one. */
2437 put_entry (np
->right
);
2438 /* Stack subentries from the following files. */
2439 push_node (np
->left
, &stack
);
2442 push_node (np
->left
, &stack
);
2449 #define C_EXT 0x00fff /* C extensions */
2450 #define C_PLAIN 0x00000 /* C */
2451 #define C_PLPL 0x00001 /* C++ */
2452 #define C_STAR 0x00003 /* C* */
2453 #define C_JAVA 0x00005 /* JAVA */
2454 #define C_AUTO 0x01000 /* C, but switch to C++ if `class' is met */
2455 #define YACC 0x10000 /* yacc file */
2458 * The C symbol tables.
2463 st_C_objprot
, st_C_objimpl
, st_C_objend
,
2465 st_C_ignore
, st_C_attribute
,
2468 st_C_class
, st_C_template
,
2469 st_C_struct
, st_C_extern
, st_C_enum
, st_C_define
, st_C_typedef
2472 /* Feed stuff between (but not including) %[ and %] lines to:
2478 struct C_stab_entry { char *name; int c_ext; enum sym_type type; }
2482 while, 0, st_C_ignore
2483 switch, 0, st_C_ignore
2484 return, 0, st_C_ignore
2485 __attribute__, 0, st_C_attribute
2486 GTY, 0, st_C_attribute
2487 @interface, 0, st_C_objprot
2488 @protocol, 0, st_C_objprot
2489 @implementation,0, st_C_objimpl
2490 @end, 0, st_C_objend
2491 import, (C_JAVA & ~C_PLPL), st_C_ignore
2492 package, (C_JAVA & ~C_PLPL), st_C_ignore
2493 friend, C_PLPL, st_C_ignore
2494 extends, (C_JAVA & ~C_PLPL), st_C_javastruct
2495 implements, (C_JAVA & ~C_PLPL), st_C_javastruct
2496 interface, (C_JAVA & ~C_PLPL), st_C_struct
2497 class, 0, st_C_class
2498 namespace, C_PLPL, st_C_struct
2499 domain, C_STAR, st_C_struct
2500 union, 0, st_C_struct
2501 struct, 0, st_C_struct
2502 extern, 0, st_C_extern
2504 typedef, 0, st_C_typedef
2505 define, 0, st_C_define
2506 undef, 0, st_C_define
2507 operator, C_PLPL, st_C_operator
2508 template, 0, st_C_template
2509 # DEFUN used in emacs, the next three used in glibc (SYSCALL only for mach).
2510 DEFUN, 0, st_C_gnumacro
2511 SYSCALL, 0, st_C_gnumacro
2512 ENTRY, 0, st_C_gnumacro
2513 PSEUDO, 0, st_C_gnumacro
2514 # These are defined inside C functions, so currently they are not met.
2515 # EXFUN used in glibc, DEFVAR_* in emacs.
2516 #EXFUN, 0, st_C_gnumacro
2517 #DEFVAR_, 0, st_C_gnumacro
2519 and replace lines between %< and %> with its output, then:
2520 - remove the #if characterset check
2521 - make in_word_set static and not inline. */
2523 /* C code produced by gperf version 3.0.1 */
2524 /* Command-line: gperf -m 5 */
2525 /* Computed positions: -k'2-3' */
2527 struct C_stab_entry
{ const char *name
; int c_ext
; enum sym_type type
; };
2528 /* maximum key range = 33, duplicates = 0 */
2531 hash (const char *str
, int len
)
2533 static char const asso_values
[] =
2535 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2536 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2537 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2538 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2539 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2540 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2541 35, 35, 35, 35, 35, 35, 35, 35, 35, 3,
2542 26, 35, 35, 35, 35, 35, 35, 35, 27, 35,
2543 35, 35, 35, 24, 0, 35, 35, 35, 35, 0,
2544 35, 35, 35, 35, 35, 1, 35, 16, 35, 6,
2545 23, 0, 0, 35, 22, 0, 35, 35, 5, 0,
2546 0, 15, 1, 35, 6, 35, 8, 19, 35, 16,
2547 4, 5, 35, 35, 35, 35, 35, 35, 35, 35,
2548 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2549 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2550 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2551 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2552 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2553 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2554 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2555 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2556 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2557 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2558 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2559 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2560 35, 35, 35, 35, 35, 35
2567 hval
+= asso_values
[(unsigned char) str
[2]];
2570 hval
+= asso_values
[(unsigned char) str
[1]];
2576 static struct C_stab_entry
*
2577 in_word_set (register const char *str
, register unsigned int len
)
2581 TOTAL_KEYWORDS
= 33,
2582 MIN_WORD_LENGTH
= 2,
2583 MAX_WORD_LENGTH
= 15,
2588 static struct C_stab_entry wordlist
[] =
2591 {"if", 0, st_C_ignore
},
2592 {"GTY", 0, st_C_attribute
},
2593 {"@end", 0, st_C_objend
},
2594 {"union", 0, st_C_struct
},
2595 {"define", 0, st_C_define
},
2596 {"import", (C_JAVA
& ~C_PLPL
), st_C_ignore
},
2597 {"template", 0, st_C_template
},
2598 {"operator", C_PLPL
, st_C_operator
},
2599 {"@interface", 0, st_C_objprot
},
2600 {"implements", (C_JAVA
& ~C_PLPL
), st_C_javastruct
},
2601 {"friend", C_PLPL
, st_C_ignore
},
2602 {"typedef", 0, st_C_typedef
},
2603 {"return", 0, st_C_ignore
},
2604 {"@implementation",0, st_C_objimpl
},
2605 {"@protocol", 0, st_C_objprot
},
2606 {"interface", (C_JAVA
& ~C_PLPL
), st_C_struct
},
2607 {"extern", 0, st_C_extern
},
2608 {"extends", (C_JAVA
& ~C_PLPL
), st_C_javastruct
},
2609 {"struct", 0, st_C_struct
},
2610 {"domain", C_STAR
, st_C_struct
},
2611 {"switch", 0, st_C_ignore
},
2612 {"enum", 0, st_C_enum
},
2613 {"for", 0, st_C_ignore
},
2614 {"namespace", C_PLPL
, st_C_struct
},
2615 {"class", 0, st_C_class
},
2616 {"while", 0, st_C_ignore
},
2617 {"undef", 0, st_C_define
},
2618 {"package", (C_JAVA
& ~C_PLPL
), st_C_ignore
},
2619 {"__attribute__", 0, st_C_attribute
},
2620 {"SYSCALL", 0, st_C_gnumacro
},
2621 {"ENTRY", 0, st_C_gnumacro
},
2622 {"PSEUDO", 0, st_C_gnumacro
},
2623 {"DEFUN", 0, st_C_gnumacro
}
2626 if (len
<= MAX_WORD_LENGTH
&& len
>= MIN_WORD_LENGTH
)
2628 int key
= hash (str
, len
);
2630 if (key
<= MAX_HASH_VALUE
&& key
>= 0)
2632 const char *s
= wordlist
[key
].name
;
2634 if (*str
== *s
&& !strncmp (str
+ 1, s
+ 1, len
- 1) && s
[len
] == '\0')
2635 return &wordlist
[key
];
2642 static enum sym_type
2643 C_symtype (char *str
, int len
, int c_ext
)
2645 register struct C_stab_entry
*se
= in_word_set (str
, len
);
2647 if (se
== NULL
|| (se
->c_ext
&& !(c_ext
& se
->c_ext
)))
2654 * Ignoring __attribute__ ((list))
2656 static bool inattribute
; /* looking at an __attribute__ construct */
2659 * C functions and variables are recognized using a simple
2660 * finite automaton. fvdef is its state variable.
2664 fvnone
, /* nothing seen */
2665 fdefunkey
, /* Emacs DEFUN keyword seen */
2666 fdefunname
, /* Emacs DEFUN name seen */
2667 foperator
, /* func: operator keyword seen (cplpl) */
2668 fvnameseen
, /* function or variable name seen */
2669 fstartlist
, /* func: just after open parenthesis */
2670 finlist
, /* func: in parameter list */
2671 flistseen
, /* func: after parameter list */
2672 fignore
, /* func: before open brace */
2673 vignore
/* var-like: ignore until ';' */
2676 static bool fvextern
; /* func or var: extern keyword seen; */
2679 * typedefs are recognized using a simple finite automaton.
2680 * typdef is its state variable.
2684 tnone
, /* nothing seen */
2685 tkeyseen
, /* typedef keyword seen */
2686 ttypeseen
, /* defined type seen */
2687 tinbody
, /* inside typedef body */
2688 tend
, /* just before typedef tag */
2689 tignore
/* junk after typedef tag */
2693 * struct-like structures (enum, struct and union) are recognized
2694 * using another simple finite automaton. `structdef' is its state
2699 snone
, /* nothing seen yet,
2700 or in struct body if bracelev > 0 */
2701 skeyseen
, /* struct-like keyword seen */
2702 stagseen
, /* struct-like tag seen */
2703 scolonseen
/* colon seen after struct-like tag */
2707 * When objdef is different from onone, objtag is the name of the class.
2709 static const char *objtag
= "<uninited>";
2712 * Yet another little state machine to deal with preprocessor lines.
2716 dnone
, /* nothing seen */
2717 dsharpseen
, /* '#' seen as first char on line */
2718 ddefineseen
, /* '#' and 'define' seen */
2719 dignorerest
/* ignore rest of line */
2723 * State machine for Objective C protocols and implementations.
2724 * Idea by Tom R.Hageman <tom@basil.icce.rug.nl> (1995)
2728 onone
, /* nothing seen */
2729 oprotocol
, /* @interface or @protocol seen */
2730 oimplementation
, /* @implementations seen */
2731 otagseen
, /* class name seen */
2732 oparenseen
, /* parenthesis before category seen */
2733 ocatseen
, /* category name seen */
2734 oinbody
, /* in @implementation body */
2735 omethodsign
, /* in @implementation body, after +/- */
2736 omethodtag
, /* after method name */
2737 omethodcolon
, /* after method colon */
2738 omethodparm
, /* after method parameter */
2739 oignore
/* wait for @end */
2744 * Use this structure to keep info about the token read, and how it
2745 * should be tagged. Used by the make_C_tag function to build a tag.
2749 char *line
; /* string containing the token */
2750 int offset
; /* where the token starts in LINE */
2751 int length
; /* token length */
2753 The previous members can be used to pass strings around for generic
2754 purposes. The following ones specifically refer to creating tags. In this
2755 case the token contained here is the pattern that will be used to create a
2758 bool valid
; /* do not create a tag; the token should be
2759 invalidated whenever a state machine is
2760 reset prematurely */
2761 bool named
; /* create a named tag */
2762 int lineno
; /* source line number of tag */
2763 long linepos
; /* source char number of tag */
2764 } token
; /* latest token read */
2767 * Variables and functions for dealing with nested structures.
2768 * Idea by Mykola Dzyuba <mdzyuba@yahoo.com> (2001)
2770 static void pushclass_above (int, char *, int);
2771 static void popclass_above (int);
2772 static void write_classname (linebuffer
*, const char *qualifier
);
2775 char **cname
; /* nested class names */
2776 int *bracelev
; /* nested class brace level */
2777 int nl
; /* class nesting level (elements used) */
2778 int size
; /* length of the array */
2779 } cstack
; /* stack for nested declaration tags */
2780 /* Current struct nesting depth (namespace, class, struct, union, enum). */
2781 #define nestlev (cstack.nl)
2782 /* After struct keyword or in struct body, not inside a nested function. */
2783 #define instruct (structdef == snone && nestlev > 0 \
2784 && bracelev == cstack.bracelev[nestlev-1] + 1)
2787 pushclass_above (int bracelev
, char *str
, int len
)
2791 popclass_above (bracelev
);
2793 if (nl
>= cstack
.size
)
2795 int size
= cstack
.size
*= 2;
2796 xrnew (cstack
.cname
, size
, char *);
2797 xrnew (cstack
.bracelev
, size
, int);
2799 assert (nl
== 0 || cstack
.bracelev
[nl
-1] < bracelev
);
2800 cstack
.cname
[nl
] = (str
== NULL
) ? NULL
: savenstr (str
, len
);
2801 cstack
.bracelev
[nl
] = bracelev
;
2806 popclass_above (int bracelev
)
2810 for (nl
= cstack
.nl
- 1;
2811 nl
>= 0 && cstack
.bracelev
[nl
] >= bracelev
;
2814 free (cstack
.cname
[nl
]);
2820 write_classname (linebuffer
*cn
, const char *qualifier
)
2823 int qlen
= strlen (qualifier
);
2825 if (cstack
.nl
== 0 || cstack
.cname
[0] == NULL
)
2829 cn
->buffer
[0] = '\0';
2833 len
= strlen (cstack
.cname
[0]);
2834 linebuffer_setlen (cn
, len
);
2835 strcpy (cn
->buffer
, cstack
.cname
[0]);
2837 for (i
= 1; i
< cstack
.nl
; i
++)
2839 char *s
= cstack
.cname
[i
];
2842 linebuffer_setlen (cn
, len
+ qlen
+ strlen (s
));
2843 len
+= sprintf (cn
->buffer
+ len
, "%s%s", qualifier
, s
);
2848 static bool consider_token (char *, int, int, int *, int, int, bool *);
2849 static void make_C_tag (bool);
2853 * checks to see if the current token is at the start of a
2854 * function or variable, or corresponds to a typedef, or
2855 * is a struct/union/enum tag, or #define, or an enum constant.
2857 * *IS_FUNC_OR_VAR gets true if the token is a function or #define macro
2858 * with args. C_EXTP points to which language we are looking at.
2869 consider_token (char *str
, int len
, int c
, int *c_extp
,
2870 int bracelev
, int parlev
, bool *is_func_or_var
)
2871 /* IN: token pointer */
2872 /* IN: token length */
2873 /* IN: first char after the token */
2874 /* IN, OUT: C extensions mask */
2875 /* IN: brace level */
2876 /* IN: parenthesis level */
2877 /* OUT: function or variable found */
2879 /* When structdef is stagseen, scolonseen, or snone with bracelev > 0,
2880 structtype is the type of the preceding struct-like keyword, and
2881 structbracelev is the brace level where it has been seen. */
2882 static enum sym_type structtype
;
2883 static int structbracelev
;
2884 static enum sym_type toktype
;
2887 toktype
= C_symtype (str
, len
, *c_extp
);
2890 * Skip __attribute__
2892 if (toktype
== st_C_attribute
)
2899 * Advance the definedef state machine.
2904 /* We're not on a preprocessor line. */
2905 if (toktype
== st_C_gnumacro
)
2912 if (toktype
== st_C_define
)
2914 definedef
= ddefineseen
;
2918 definedef
= dignorerest
;
2923 * Make a tag for any macro, unless it is a constant
2924 * and constantypedefs is false.
2926 definedef
= dignorerest
;
2927 *is_func_or_var
= (c
== '(');
2928 if (!*is_func_or_var
&& !constantypedefs
)
2935 error ("internal error: definedef value.");
2944 if (toktype
== st_C_typedef
)
2967 if (structdef
== snone
&& fvdef
== fvnone
)
2989 case st_C_javastruct
:
2990 if (structdef
== stagseen
)
2991 structdef
= scolonseen
;
2995 if ((*c_extp
& C_AUTO
) /* automatic detection of C++ language */
2997 && definedef
== dnone
&& structdef
== snone
2998 && typdef
== tnone
&& fvdef
== fvnone
)
2999 *c_extp
= (*c_extp
| C_PLPL
) & ~C_AUTO
;
3000 if (toktype
== st_C_template
)
3007 && (typdef
== tkeyseen
3008 || (typedefs_or_cplusplus
&& structdef
== snone
)))
3010 structdef
= skeyseen
;
3011 structtype
= toktype
;
3012 structbracelev
= bracelev
;
3013 if (fvdef
== fvnameseen
)
3021 if (structdef
== skeyseen
)
3023 structdef
= stagseen
;
3027 if (typdef
!= tnone
)
3030 /* Detect Objective C constructs. */
3040 objdef
= oimplementation
;
3046 case oimplementation
:
3047 /* Save the class tag for functions or variables defined inside. */
3048 objtag
= savenstr (str
, len
);
3052 /* Save the class tag for categories. */
3053 objtag
= savenstr (str
, len
);
3055 *is_func_or_var
= true;
3059 *is_func_or_var
= true;
3067 objdef
= omethodtag
;
3068 linebuffer_setlen (&token_name
, len
);
3069 memcpy (token_name
.buffer
, str
, len
);
3070 token_name
.buffer
[len
] = '\0';
3076 objdef
= omethodparm
;
3081 objdef
= omethodtag
;
3084 int oldlen
= token_name
.len
;
3086 linebuffer_setlen (&token_name
, oldlen
+ len
);
3087 memcpy (token_name
.buffer
+ oldlen
, str
, len
);
3088 token_name
.buffer
[oldlen
+ len
] = '\0';
3094 if (toktype
== st_C_objend
)
3096 /* Memory leakage here: the string pointed by objtag is
3097 never released, because many tests would be needed to
3098 avoid breaking on incorrect input code. The amount of
3099 memory leaked here is the sum of the lengths of the
3109 /* A function, variable or enum constant? */
3131 *is_func_or_var
= true;
3135 && structdef
== snone
3136 && structtype
== st_C_enum
&& bracelev
> structbracelev
3137 /* Don't tag tokens in expressions that assign values to enum
3139 && fvdef
!= vignore
)
3140 return true; /* enum constant */
3146 fvdef
= fdefunname
; /* GNU macro */
3147 *is_func_or_var
= true;
3155 if ((strneq (str
, "asm", 3) && endtoken (str
[3]))
3156 || (strneq (str
, "__asm__", 7) && endtoken (str
[7])))
3167 if (len
>= 10 && strneq (str
+len
-10, "::operator", 10))
3169 if (*c_extp
& C_AUTO
) /* automatic detection of C++ */
3170 *c_extp
= (*c_extp
| C_PLPL
) & ~C_AUTO
;
3172 *is_func_or_var
= true;
3175 if (bracelev
> 0 && !instruct
)
3177 fvdef
= fvnameseen
; /* function or variable */
3178 *is_func_or_var
= true;
3193 * C_entries often keeps pointers to tokens or lines which are older than
3194 * the line currently read. By keeping two line buffers, and switching
3195 * them at end of line, it is possible to use those pointers.
3203 #define current_lb_is_new (newndx == curndx)
3204 #define switch_line_buffers() (curndx = 1 - curndx)
3206 #define curlb (lbs[curndx].lb)
3207 #define newlb (lbs[newndx].lb)
3208 #define curlinepos (lbs[curndx].linepos)
3209 #define newlinepos (lbs[newndx].linepos)
3211 #define plainc ((c_ext & C_EXT) == C_PLAIN)
3212 #define cplpl (c_ext & C_PLPL)
3213 #define cjava ((c_ext & C_JAVA) == C_JAVA)
3215 #define CNL_SAVE_DEFINEDEF() \
3217 curlinepos = charno; \
3218 readline (&curlb, inf); \
3219 lp = curlb.buffer; \
3226 CNL_SAVE_DEFINEDEF (); \
3227 if (savetoken.valid) \
3229 token = savetoken; \
3230 savetoken.valid = false; \
3232 definedef = dnone; \
3237 make_C_tag (bool isfun
)
3239 /* This function is never called when token.valid is false, but
3240 we must protect against invalid input or internal errors. */
3242 make_tag (token_name
.buffer
, token_name
.len
, isfun
, token
.line
,
3243 token
.offset
+token
.length
+1, token
.lineno
, token
.linepos
);
3245 { /* this branch is optimized away if !DEBUG */
3246 make_tag (concat ("INVALID TOKEN:-->", token_name
.buffer
, ""),
3247 token_name
.len
+ 17, isfun
, token
.line
,
3248 token
.offset
+token
.length
+1, token
.lineno
, token
.linepos
);
3249 error ("INVALID TOKEN");
3252 token
.valid
= false;
3256 perhaps_more_input (FILE *inf
)
3258 return !feof (inf
) && !ferror (inf
);
3264 * This routine finds functions, variables, typedefs,
3265 * #define's, enum constants and struct/union/enum definitions in
3266 * C syntax and adds them to the list.
3269 C_entries (int c_ext
, FILE *inf
)
3270 /* extension of C */
3273 register char c
; /* latest char read; '\0' for end of line */
3274 register char *lp
; /* pointer one beyond the character `c' */
3275 int curndx
, newndx
; /* indices for current and new lb */
3276 register int tokoff
; /* offset in line of start of current token */
3277 register int toklen
; /* length of current token */
3278 const char *qualifier
; /* string used to qualify names */
3279 int qlen
; /* length of qualifier */
3280 int bracelev
; /* current brace level */
3281 int bracketlev
; /* current bracket level */
3282 int parlev
; /* current parenthesis level */
3283 int attrparlev
; /* __attribute__ parenthesis level */
3284 int templatelev
; /* current template level */
3285 int typdefbracelev
; /* bracelev where a typedef struct body begun */
3286 bool incomm
, inquote
, inchar
, quotednl
, midtoken
;
3287 bool yacc_rules
; /* in the rules part of a yacc file */
3288 struct tok savetoken
= {0}; /* token saved during preprocessor handling */
3291 linebuffer_init (&lbs
[0].lb
);
3292 linebuffer_init (&lbs
[1].lb
);
3293 if (cstack
.size
== 0)
3295 cstack
.size
= (DEBUG
) ? 1 : 4;
3297 cstack
.cname
= xnew (cstack
.size
, char *);
3298 cstack
.bracelev
= xnew (cstack
.size
, int);
3301 tokoff
= toklen
= typdefbracelev
= 0; /* keep compiler quiet */
3302 curndx
= newndx
= 0;
3306 fvdef
= fvnone
; fvextern
= false; typdef
= tnone
;
3307 structdef
= snone
; definedef
= dnone
; objdef
= onone
;
3309 midtoken
= inquote
= inchar
= incomm
= quotednl
= false;
3310 token
.valid
= savetoken
.valid
= false;
3311 bracelev
= bracketlev
= parlev
= attrparlev
= templatelev
= 0;
3313 { qualifier
= "."; qlen
= 1; }
3315 { qualifier
= "::"; qlen
= 2; }
3318 while (perhaps_more_input (inf
))
3323 /* If we are at the end of the line, the next character is a
3324 '\0'; do not skip it, because it is what tells us
3325 to read the next line. */
3346 /* Newlines inside comments do not end macro definitions in
3348 CNL_SAVE_DEFINEDEF ();
3361 /* Newlines inside strings do not end macro definitions
3362 in traditional cpp, even though compilers don't
3363 usually accept them. */
3364 CNL_SAVE_DEFINEDEF ();
3374 /* Hmmm, something went wrong. */
3410 if (fvdef
!= finlist
&& fvdef
!= fignore
&& fvdef
!= vignore
)
3425 else if (/* cplpl && */ *lp
== '/')
3431 if ((c_ext
& YACC
) && *lp
== '%')
3433 /* Entering or exiting rules section in yacc file. */
3435 definedef
= dnone
; fvdef
= fvnone
; fvextern
= false;
3436 typdef
= tnone
; structdef
= snone
;
3437 midtoken
= inquote
= inchar
= incomm
= quotednl
= false;
3439 yacc_rules
= !yacc_rules
;
3445 if (definedef
== dnone
)
3448 bool cpptoken
= true;
3450 /* Look back on this line. If all blanks, or nonblanks
3451 followed by an end of comment, this is a preprocessor
3453 for (cp
= newlb
.buffer
; cp
< lp
-1; cp
++)
3454 if (!c_isspace (*cp
))
3456 if (*cp
== '*' && cp
[1] == '/')
3466 definedef
= dsharpseen
;
3467 /* This is needed for tagging enum values: when there are
3468 preprocessor conditionals inside the enum, we need to
3469 reset the value of fvdef so that the next enum value is
3470 tagged even though the one before it did not end in a
3472 if (fvdef
== vignore
&& instruct
&& parlev
== 0)
3474 if (strneq (cp
, "#if", 3) || strneq (cp
, "#el", 3))
3478 } /* if (definedef == dnone) */
3489 CNL_SAVE_DEFINEDEF ();
3496 /* Consider token only if some involved conditions are satisfied. */
3497 if (typdef
!= tignore
3498 && definedef
!= dignorerest
3501 && (definedef
!= dnone
3502 || structdef
!= scolonseen
)
3509 if (c
== ':' && *lp
== ':' && begtoken (lp
[1]))
3510 /* This handles :: in the middle,
3511 but not at the beginning of an identifier.
3512 Also, space-separated :: is not recognized. */
3514 if (c_ext
& C_AUTO
) /* automatic detection of C++ */
3515 c_ext
= (c_ext
| C_PLPL
) & ~C_AUTO
;
3519 goto still_in_token
;
3523 bool funorvar
= false;
3526 || consider_token (newlb
.buffer
+ tokoff
, toklen
, c
,
3527 &c_ext
, bracelev
, parlev
,
3530 if (fvdef
== foperator
)
3533 lp
= skip_spaces (lp
-1);
3537 && !c_isspace (*lp
) && *lp
!= '(')
3540 toklen
+= lp
- oldlp
;
3542 token
.named
= false;
3544 && nestlev
> 0 && definedef
== dnone
)
3545 /* in struct body */
3550 write_classname (&token_name
, qualifier
);
3551 len
= token_name
.len
;
3552 linebuffer_setlen (&token_name
,
3553 len
+ qlen
+ toklen
);
3554 sprintf (token_name
.buffer
+ len
, "%s%.*s",
3556 newlb
.buffer
+ tokoff
);
3560 linebuffer_setlen (&token_name
, toklen
);
3561 sprintf (token_name
.buffer
, "%.*s",
3562 toklen
, newlb
.buffer
+ tokoff
);
3566 else if (objdef
== ocatseen
)
3567 /* Objective C category */
3571 int len
= strlen (objtag
) + 2 + toklen
;
3572 linebuffer_setlen (&token_name
, len
);
3573 sprintf (token_name
.buffer
, "%s(%.*s)",
3575 newlb
.buffer
+ tokoff
);
3579 linebuffer_setlen (&token_name
, toklen
);
3580 sprintf (token_name
.buffer
, "%.*s",
3581 toklen
, newlb
.buffer
+ tokoff
);
3585 else if (objdef
== omethodtag
3586 || objdef
== omethodparm
)
3587 /* Objective C method */
3591 else if (fvdef
== fdefunname
)
3592 /* GNU DEFUN and similar macros */
3594 bool defun
= (newlb
.buffer
[tokoff
] == 'F');
3603 /* First, tag it as its C name */
3604 linebuffer_setlen (&token_name
, toklen
);
3605 memcpy (token_name
.buffer
,
3606 newlb
.buffer
+ tokoff
, toklen
);
3607 token_name
.buffer
[toklen
] = '\0';
3609 token
.lineno
= lineno
;
3610 token
.offset
= tokoff
;
3611 token
.length
= toklen
;
3612 token
.line
= newlb
.buffer
;
3613 token
.linepos
= newlinepos
;
3615 make_C_tag (funorvar
);
3617 /* Rewrite the tag so that emacs lisp DEFUNs
3618 can be found also by their elisp name */
3619 linebuffer_setlen (&token_name
, len
);
3620 memcpy (token_name
.buffer
,
3621 newlb
.buffer
+ off
, len
);
3622 token_name
.buffer
[len
] = '\0';
3625 if (token_name
.buffer
[len
] == '_')
3626 token_name
.buffer
[len
] = '-';
3627 token
.named
= defun
;
3631 linebuffer_setlen (&token_name
, toklen
);
3632 memcpy (token_name
.buffer
,
3633 newlb
.buffer
+ tokoff
, toklen
);
3634 token_name
.buffer
[toklen
] = '\0';
3635 /* Name macros and members. */
3636 token
.named
= (structdef
== stagseen
3637 || typdef
== ttypeseen
3640 && definedef
== dignorerest
)
3642 && definedef
== dnone
3643 && structdef
== snone
3646 token
.lineno
= lineno
;
3647 token
.offset
= tokoff
;
3648 token
.length
= toklen
;
3649 token
.line
= newlb
.buffer
;
3650 token
.linepos
= newlinepos
;
3653 if (definedef
== dnone
3654 && (fvdef
== fvnameseen
3655 || fvdef
== foperator
3656 || structdef
== stagseen
3658 || typdef
== ttypeseen
3659 || objdef
!= onone
))
3661 if (current_lb_is_new
)
3662 switch_line_buffers ();
3664 else if (definedef
!= dnone
3665 || fvdef
== fdefunname
3667 make_C_tag (funorvar
);
3669 else /* not yacc and consider_token failed */
3671 if (inattribute
&& fvdef
== fignore
)
3673 /* We have just met __attribute__ after a
3674 function parameter list: do not tag the
3681 } /* if (endtoken (c)) */
3682 else if (intoken (c
))
3688 } /* if (midtoken) */
3689 else if (begtoken (c
))
3697 /* This prevents tagging fb in
3698 void (__attribute__((noreturn)) *fb) (void);
3699 Fixing this is not easy and not very important. */
3703 if (plainc
|| declarations
)
3705 make_C_tag (true); /* a function */
3712 if (structdef
== stagseen
&& !cjava
)
3714 popclass_above (bracelev
);
3724 if (!yacc_rules
|| lp
== newlb
.buffer
+ 1)
3726 tokoff
= lp
- 1 - newlb
.buffer
;
3731 } /* if (begtoken) */
3732 } /* if must look at token */
3735 /* Detect end of line, colon, comma, semicolon and various braces
3736 after having handled a token.*/
3742 if (yacc_rules
&& token
.offset
== 0 && token
.valid
)
3744 make_C_tag (false); /* a yacc function */
3747 if (definedef
!= dnone
)
3753 make_C_tag (true); /* an Objective C class */
3757 objdef
= omethodcolon
;
3760 int toklen
= token_name
.len
;
3761 linebuffer_setlen (&token_name
, toklen
+ 1);
3762 strcpy (token_name
.buffer
+ toklen
, ":");
3768 if (structdef
== stagseen
)
3770 structdef
= scolonseen
;
3773 /* Should be useless, but may be work as a safety net. */
3774 if (cplpl
&& fvdef
== flistseen
)
3776 make_C_tag (true); /* a function */
3782 if (definedef
!= dnone
|| inattribute
)
3788 make_C_tag (false); /* a typedef */
3798 if (typdef
== tignore
|| cplpl
)
3802 if ((globals
&& bracelev
== 0 && (!fvextern
|| declarations
))
3803 || (members
&& instruct
))
3804 make_C_tag (false); /* a variable */
3807 token
.valid
= false;
3811 && (cplpl
|| !instruct
)
3812 && (typdef
== tnone
|| (typdef
!= tignore
&& instruct
)))
3814 && plainc
&& instruct
))
3815 make_C_tag (true); /* a function */
3821 && cplpl
&& structdef
== stagseen
)
3822 make_C_tag (false); /* forward declaration */
3824 token
.valid
= false;
3825 } /* switch (fvdef) */
3831 if (structdef
== stagseen
)
3835 if (definedef
!= dnone
|| inattribute
)
3841 make_C_tag (true); /* an Objective C method */
3856 if (instruct
&& parlev
== 0)
3867 && (!fvextern
|| declarations
))
3868 || (members
&& instruct
)))
3869 make_C_tag (false); /* a variable */
3872 if ((declarations
&& typdef
== tnone
&& !instruct
)
3873 || (members
&& typdef
!= tignore
&& instruct
))
3875 make_C_tag (true); /* a function */
3878 else if (!declarations
)
3880 token
.valid
= false;
3885 if (structdef
== stagseen
)
3889 if (definedef
!= dnone
|| inattribute
)
3891 if (structdef
== stagseen
)
3898 make_C_tag (false); /* a typedef */
3910 if ((members
&& bracelev
== 1)
3911 || (globals
&& bracelev
== 0
3912 && (!fvextern
|| declarations
)))
3913 make_C_tag (false); /* a variable */
3929 if (definedef
!= dnone
)
3931 if (objdef
== otagseen
&& parlev
== 0)
3932 objdef
= oparenseen
;
3936 if (typdef
== ttypeseen
3940 /* This handles constructs like:
3941 typedef void OperatorFun (int fun); */
3962 if (--attrparlev
== 0)
3963 inattribute
= false;
3966 if (definedef
!= dnone
)
3968 if (objdef
== ocatseen
&& parlev
== 1)
3970 make_C_tag (true); /* an Objective C category */
3986 || typdef
== ttypeseen
))
3989 make_C_tag (false); /* a typedef */
3992 else if (parlev
< 0) /* can happen due to ill-conceived #if's. */
3996 if (definedef
!= dnone
)
3998 if (typdef
== ttypeseen
)
4000 /* Whenever typdef is set to tinbody (currently only
4001 here), typdefbracelev should be set to bracelev. */
4003 typdefbracelev
= bracelev
;
4008 if (cplpl
&& !class_qualify
)
4010 /* Remove class and namespace qualifiers from the token,
4011 leaving only the method/member name. */
4012 char *cc
, *uqname
= token_name
.buffer
;
4013 char *tok_end
= token_name
.buffer
+ token_name
.len
;
4015 for (cc
= token_name
.buffer
; cc
< tok_end
; cc
++)
4017 if (*cc
== ':' && cc
[1] == ':')
4023 if (uqname
> token_name
.buffer
)
4025 int uqlen
= strlen (uqname
);
4026 linebuffer_setlen (&token_name
, uqlen
);
4027 memmove (token_name
.buffer
, uqname
, uqlen
+ 1);
4030 make_C_tag (true); /* a function */
4039 make_C_tag (true); /* an Objective C class */
4044 make_C_tag (true); /* an Objective C method */
4048 /* Neutralize `extern "C" {' grot. */
4049 if (bracelev
== 0 && structdef
== snone
&& nestlev
== 0
4059 case skeyseen
: /* unnamed struct */
4060 pushclass_above (bracelev
, NULL
, 0);
4063 case stagseen
: /* named struct or enum */
4064 case scolonseen
: /* a class */
4065 pushclass_above (bracelev
,token
.line
+token
.offset
, token
.length
);
4067 make_C_tag (false); /* a struct or enum */
4075 if (definedef
!= dnone
)
4077 if (fvdef
== fstartlist
)
4079 fvdef
= fvnone
; /* avoid tagging `foo' in `foo (*bar()) ()' */
4080 token
.valid
= false;
4084 if (definedef
!= dnone
)
4087 if (!ignoreindent
&& lp
== newlb
.buffer
+ 1)
4090 token
.valid
= false; /* unexpected value, token unreliable */
4091 bracelev
= 0; /* reset brace level if first column */
4092 parlev
= 0; /* also reset paren level, just in case... */
4094 else if (bracelev
< 0)
4096 token
.valid
= false; /* something gone amiss, token unreliable */
4099 if (bracelev
== 0 && fvdef
== vignore
)
4100 fvdef
= fvnone
; /* end of function */
4101 popclass_above (bracelev
);
4103 /* Only if typdef == tinbody is typdefbracelev significant. */
4104 if (typdef
== tinbody
&& bracelev
<= typdefbracelev
)
4106 assert (bracelev
== typdefbracelev
);
4111 if (definedef
!= dnone
)
4121 if ((members
&& bracelev
== 1)
4122 || (globals
&& bracelev
== 0 && (!fvextern
|| declarations
)))
4123 make_C_tag (false); /* a variable */
4131 && (structdef
== stagseen
|| fvdef
== fvnameseen
))
4138 if (templatelev
> 0)
4146 if (objdef
== oinbody
&& bracelev
== 0)
4148 objdef
= omethodsign
;
4153 case '#': case '~': case '&': case '%': case '/':
4154 case '|': case '^': case '!': case '.': case '?':
4155 if (definedef
!= dnone
)
4157 /* These surely cannot follow a function tag in C. */
4170 if (objdef
== otagseen
)
4172 make_C_tag (true); /* an Objective C class */
4175 /* If a macro spans multiple lines don't reset its state. */
4177 CNL_SAVE_DEFINEDEF ();
4183 } /* while not eof */
4185 free (lbs
[0].lb
.buffer
);
4186 free (lbs
[1].lb
.buffer
);
4190 * Process either a C++ file or a C file depending on the setting
4194 default_C_entries (FILE *inf
)
4196 C_entries (cplusplus
? C_PLPL
: C_AUTO
, inf
);
4199 /* Always do plain C. */
4201 plain_C_entries (FILE *inf
)
4206 /* Always do C++. */
4208 Cplusplus_entries (FILE *inf
)
4210 C_entries (C_PLPL
, inf
);
4213 /* Always do Java. */
4215 Cjava_entries (FILE *inf
)
4217 C_entries (C_JAVA
, inf
);
4222 Cstar_entries (FILE *inf
)
4224 C_entries (C_STAR
, inf
);
4227 /* Always do Yacc. */
4229 Yacc_entries (FILE *inf
)
4231 C_entries (YACC
, inf
);
4235 /* Useful macros. */
4236 #define LOOP_ON_INPUT_LINES(file_pointer, line_buffer, char_pointer) \
4237 while (perhaps_more_input (file_pointer) \
4238 && (readline (&(line_buffer), file_pointer), \
4239 (char_pointer) = (line_buffer).buffer, \
4242 #define LOOKING_AT(cp, kw) /* kw is the keyword, a literal string */ \
4243 ((assert ("" kw), true) /* syntax error if not a literal string */ \
4244 && strneq ((cp), kw, sizeof (kw)-1) /* cp points at kw */ \
4245 && notinname ((cp)[sizeof (kw)-1]) /* end of kw */ \
4246 && ((cp) = skip_spaces ((cp) + sizeof (kw) - 1), true)) /* skip spaces */
4248 /* Similar to LOOKING_AT but does not use notinname, does not skip */
4249 #define LOOKING_AT_NOCASE(cp, kw) /* the keyword is a literal string */ \
4250 ((assert ("" kw), true) /* syntax error if not a literal string */ \
4251 && strncaseeq ((cp), kw, sizeof (kw)-1) /* cp points at kw */ \
4252 && ((cp) += sizeof (kw) - 1, true)) /* skip spaces */
4255 * Read a file, but do no processing. This is used to do regexp
4256 * matching on files that have no language defined.
4259 just_read_file (FILE *inf
)
4261 while (perhaps_more_input (inf
))
4262 readline (&lb
, inf
);
4266 /* Fortran parsing */
4268 static void F_takeprec (void);
4269 static void F_getit (FILE *);
4274 dbp
= skip_spaces (dbp
);
4278 dbp
= skip_spaces (dbp
);
4279 if (strneq (dbp
, "(*)", 3))
4284 if (!c_isdigit (*dbp
))
4286 --dbp
; /* force failure */
4291 while (c_isdigit (*dbp
));
4299 dbp
= skip_spaces (dbp
);
4302 readline (&lb
, inf
);
4307 dbp
= skip_spaces (dbp
);
4309 if (!c_isalpha (*dbp
) && *dbp
!= '_' && *dbp
!= '$')
4311 for (cp
= dbp
+ 1; *cp
!= '\0' && intoken (*cp
); cp
++)
4313 make_tag (dbp
, cp
-dbp
, true,
4314 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4319 Fortran_functions (FILE *inf
)
4321 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
4324 dbp
++; /* Ratfor escape to fortran */
4325 dbp
= skip_spaces (dbp
);
4329 if (LOOKING_AT_NOCASE (dbp
, "recursive"))
4330 dbp
= skip_spaces (dbp
);
4332 if (LOOKING_AT_NOCASE (dbp
, "pure"))
4333 dbp
= skip_spaces (dbp
);
4335 if (LOOKING_AT_NOCASE (dbp
, "elemental"))
4336 dbp
= skip_spaces (dbp
);
4338 switch (c_tolower (*dbp
))
4341 if (nocase_tail ("integer"))
4345 if (nocase_tail ("real"))
4349 if (nocase_tail ("logical"))
4353 if (nocase_tail ("complex") || nocase_tail ("character"))
4357 if (nocase_tail ("double"))
4359 dbp
= skip_spaces (dbp
);
4362 if (nocase_tail ("precision"))
4368 dbp
= skip_spaces (dbp
);
4371 switch (c_tolower (*dbp
))
4374 if (nocase_tail ("function"))
4378 if (nocase_tail ("subroutine"))
4382 if (nocase_tail ("entry"))
4386 if (nocase_tail ("blockdata") || nocase_tail ("block data"))
4388 dbp
= skip_spaces (dbp
);
4389 if (*dbp
== '\0') /* assume un-named */
4390 make_tag ("blockdata", 9, true,
4391 lb
.buffer
, dbp
- lb
.buffer
, lineno
, linecharno
);
4393 F_getit (inf
); /* look for name */
4402 * Go language support
4403 * Original code by Xi Lu <lx@shellcodes.org> (2016)
4406 Go_functions(FILE *inf
)
4410 LOOP_ON_INPUT_LINES(inf
, lb
, cp
)
4412 cp
= skip_spaces (cp
);
4414 if (LOOKING_AT (cp
, "package"))
4417 while (!notinname (*cp
) && *cp
!= '\0')
4419 make_tag (name
, cp
- name
, false, lb
.buffer
,
4420 cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4422 else if (LOOKING_AT (cp
, "func"))
4424 /* Go implementation of interface, such as:
4425 func (n *Integer) Add(m Integer) ...
4426 skip `(n *Integer)` part.
4432 cp
= skip_spaces (cp
+1);
4439 while (!notinname (*cp
))
4442 make_tag (name
, cp
- name
, true, lb
.buffer
,
4443 cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4446 else if (members
&& LOOKING_AT (cp
, "type"))
4450 /* Ignore the likes of the following:
4458 while (!notinname (*cp
) && *cp
!= '\0')
4461 make_tag (name
, cp
- name
, false, lb
.buffer
,
4462 cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4471 * Philippe Waroquiers (1998)
4474 /* Once we are positioned after an "interesting" keyword, let's get
4475 the real tag value necessary. */
4477 Ada_getit (FILE *inf
, const char *name_qualifier
)
4483 while (perhaps_more_input (inf
))
4485 dbp
= skip_spaces (dbp
);
4487 || (dbp
[0] == '-' && dbp
[1] == '-'))
4489 readline (&lb
, inf
);
4492 switch (c_tolower (*dbp
))
4495 if (nocase_tail ("body"))
4497 /* Skipping body of procedure body or package body or ....
4498 resetting qualifier to body instead of spec. */
4499 name_qualifier
= "/b";
4504 /* Skipping type of task type or protected type ... */
4505 if (nocase_tail ("type"))
4512 for (cp
= dbp
; *cp
!= '\0' && *cp
!= '"'; cp
++)
4517 dbp
= skip_spaces (dbp
);
4519 c_isalnum (*cp
) || *cp
== '_' || *cp
== '.';
4527 name
= concat (dbp
, name_qualifier
, "");
4529 make_tag (name
, strlen (name
), true,
4530 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4539 Ada_funcs (FILE *inf
)
4541 bool inquote
= false;
4542 bool skip_till_semicolumn
= false;
4544 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
4546 while (*dbp
!= '\0')
4548 /* Skip a string i.e. "abcd". */
4549 if (inquote
|| (*dbp
== '"'))
4551 dbp
= strchr (dbp
+ !inquote
, '"');
4556 continue; /* advance char */
4561 break; /* advance line */
4565 /* Skip comments. */
4566 if (dbp
[0] == '-' && dbp
[1] == '-')
4567 break; /* advance line */
4569 /* Skip character enclosed in single quote i.e. 'a'
4570 and skip single quote starting an attribute i.e. 'Image. */
4579 if (skip_till_semicolumn
)
4582 skip_till_semicolumn
= false;
4584 continue; /* advance char */
4587 /* Search for beginning of a token. */
4588 if (!begtoken (*dbp
))
4591 continue; /* advance char */
4594 /* We are at the beginning of a token. */
4595 switch (c_tolower (*dbp
))
4598 if (!packages_only
&& nocase_tail ("function"))
4599 Ada_getit (inf
, "/f");
4601 break; /* from switch */
4602 continue; /* advance char */
4604 if (!packages_only
&& nocase_tail ("procedure"))
4605 Ada_getit (inf
, "/p");
4606 else if (nocase_tail ("package"))
4607 Ada_getit (inf
, "/s");
4608 else if (nocase_tail ("protected")) /* protected type */
4609 Ada_getit (inf
, "/t");
4611 break; /* from switch */
4612 continue; /* advance char */
4615 if (typedefs
&& !packages_only
&& nocase_tail ("use"))
4617 /* when tagging types, avoid tagging use type Pack.Typename;
4618 for this, we will skip everything till a ; */
4619 skip_till_semicolumn
= true;
4620 continue; /* advance char */
4624 if (!packages_only
&& nocase_tail ("task"))
4625 Ada_getit (inf
, "/k");
4626 else if (typedefs
&& !packages_only
&& nocase_tail ("type"))
4628 Ada_getit (inf
, "/t");
4629 while (*dbp
!= '\0')
4633 break; /* from switch */
4634 continue; /* advance char */
4637 /* Look for the end of the token. */
4638 while (!endtoken (*dbp
))
4641 } /* advance char */
4642 } /* advance line */
4647 * Unix and microcontroller assembly tag handling
4648 * Labels: /^[a-zA-Z_.$][a-zA_Z0-9_.$]*[: ^I^J]/
4649 * Idea by Bob Weiner, Motorola Inc. (1994)
4652 Asm_labels (FILE *inf
)
4656 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4658 /* If first char is alphabetic or one of [_.$], test for colon
4659 following identifier. */
4660 if (c_isalpha (*cp
) || *cp
== '_' || *cp
== '.' || *cp
== '$')
4662 /* Read past label. */
4664 while (c_isalnum (*cp
) || *cp
== '_' || *cp
== '.' || *cp
== '$')
4666 if (*cp
== ':' || c_isspace (*cp
))
4667 /* Found end of label, so copy it and add it to the table. */
4668 make_tag (lb
.buffer
, cp
- lb
.buffer
, true,
4669 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4677 * Perl sub names: /^sub[ \t\n]+[^ \t\n{]+/
4678 * /^use constant[ \t\n]+[^ \t\n{=,;]+/
4679 * Perl variable names: /^(my|local).../
4680 * Original code by Bart Robinson <lomew@cs.utah.edu> (1995)
4681 * Additions by Michael Ernst <mernst@alum.mit.edu> (1997)
4682 * Ideas by Kai Großjohann <Kai.Grossjohann@CS.Uni-Dortmund.DE> (2001)
4685 Perl_functions (FILE *inf
)
4687 char *package
= savestr ("main"); /* current package name */
4690 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4692 cp
= skip_spaces (cp
);
4694 if (LOOKING_AT (cp
, "package"))
4697 get_tag (cp
, &package
);
4699 else if (LOOKING_AT (cp
, "sub"))
4705 while (!notinname (*cp
))
4708 continue; /* nothing found */
4709 pos
= strchr (sp
, ':');
4710 if (pos
&& pos
< cp
&& pos
[1] == ':')
4712 /* The name is already qualified. */
4715 char *q
= pos
+ 2, *qpos
;
4716 while ((qpos
= strchr (q
, ':')) != NULL
4722 make_tag (sp
, cp
- sp
, true,
4723 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4725 else if (class_qualify
)
4728 char savechar
, *name
;
4732 name
= concat (package
, "::", sp
);
4734 make_tag (name
, strlen (name
), true,
4735 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4739 make_tag (sp
, cp
- sp
, true,
4740 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4742 else if (LOOKING_AT (cp
, "use constant")
4743 || LOOKING_AT (cp
, "use constant::defer"))
4745 /* For hash style multi-constant like
4746 use constant { FOO => 123,
4748 only the first FOO is picked up. Parsing across the value
4749 expressions would be difficult in general, due to possible nested
4750 hashes, here-documents, etc. */
4752 cp
= skip_spaces (cp
+1);
4755 else if (globals
) /* only if we are tagging global vars */
4757 /* Skip a qualifier, if any. */
4758 bool qual
= LOOKING_AT (cp
, "my") || LOOKING_AT (cp
, "local");
4759 /* After "my" or "local", but before any following paren or space. */
4760 char *varstart
= cp
;
4762 if (qual
/* should this be removed? If yes, how? */
4763 && (*cp
== '$' || *cp
== '@' || *cp
== '%'))
4768 while (c_isalnum (*cp
) || *cp
== '_');
4772 /* Should be examining a variable list at this point;
4773 could insist on seeing an open parenthesis. */
4774 while (*cp
!= '\0' && *cp
!= ';' && *cp
!= '=' && *cp
!= ')')
4780 make_tag (varstart
, cp
- varstart
, false,
4781 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4790 * Look for /^[\t]*def[ \t\n]+[^ \t\n(:]+/ or /^class[ \t\n]+[^ \t\n(:]+/
4791 * Idea by Eric S. Raymond <esr@thyrsus.com> (1997)
4792 * More ideas by seb bacon <seb@jamkit.com> (2002)
4795 Python_functions (FILE *inf
)
4799 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4801 cp
= skip_spaces (cp
);
4802 if (LOOKING_AT (cp
, "def") || LOOKING_AT (cp
, "class"))
4805 while (!notinname (*cp
) && *cp
!= ':')
4807 make_tag (name
, cp
- name
, true,
4808 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4815 * Original code by Xi Lu <lx@shellcodes.org> (2015)
4818 Ruby_functions (FILE *inf
)
4821 bool reader
= false, writer
= false, alias
= false, continuation
= false;
4823 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4825 bool is_class
= false;
4826 bool is_method
= false;
4829 cp
= skip_spaces (cp
);
4832 && c_isalpha (*cp
) && c_isupper (*cp
))
4834 char *bp
, *colon
= NULL
;
4838 for (cp
++; c_isalnum (*cp
) || *cp
== '_' || *cp
== ':'; cp
++)
4845 bp
= skip_spaces (cp
);
4846 if (*bp
== '=' && !(bp
[1] == '=' || bp
[1] == '>'))
4848 if (colon
&& !c_isspace (colon
[1]))
4850 make_tag (name
, cp
- name
, false,
4851 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4855 else if (!continuation
4856 /* Modules, classes, methods. */
4857 && ((is_method
= LOOKING_AT (cp
, "def"))
4858 || (is_class
= LOOKING_AT (cp
, "class"))
4859 || LOOKING_AT (cp
, "module")))
4861 const char self_name
[] = "self.";
4862 const size_t self_size1
= sizeof (self_name
) - 1;
4866 /* Ruby method names can end in a '='. Also, operator overloading can
4867 define operators whose names include '='. */
4868 while (!notinname (*cp
) || *cp
== '=')
4871 /* Remove "self." from the method name. */
4872 if (cp
- name
> self_size1
4873 && strneq (name
, self_name
, self_size1
))
4876 /* Remove the class/module qualifiers from method names. */
4881 for (q
= name
; q
< cp
&& *q
!= '.'; q
++)
4883 if (q
< cp
- 1) /* punt if we see just "FOO." */
4887 /* Don't tag singleton classes. */
4888 if (is_class
&& strneq (name
, "<<", 2) && cp
== name
+ 2)
4891 make_tag (name
, cp
- name
, true,
4892 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4896 /* Tag accessors and aliases. */
4899 reader
= writer
= alias
= false;
4901 while (*cp
&& *cp
!= '#')
4905 reader
= writer
= alias
= false;
4906 if (LOOKING_AT (cp
, "attr_reader"))
4908 else if (LOOKING_AT (cp
, "attr_writer"))
4910 else if (LOOKING_AT (cp
, "attr_accessor"))
4915 else if (LOOKING_AT (cp
, "alias_method"))
4918 if (reader
|| writer
|| alias
)
4923 cp
= skip_spaces (cp
);
4925 cp
= skip_spaces (cp
+ 1);
4927 cp
= skip_name (cp
);
4933 make_tag (np
, cp
- np
, true,
4934 lb
.buffer
, cp
- lb
.buffer
+ 1,
4935 lineno
, linecharno
);
4936 continuation
= false;
4940 size_t name_len
= cp
- np
+ 1;
4941 char *wr_name
= xnew (name_len
+ 1, char);
4943 memcpy (wr_name
, np
, name_len
- 1);
4944 memcpy (wr_name
+ name_len
- 1, "=", 2);
4945 pfnote (wr_name
, true, lb
.buffer
, cp
- lb
.buffer
+ 1,
4946 lineno
, linecharno
);
4947 continuation
= false;
4952 make_tag (np
, cp
- np
, true,
4953 lb
.buffer
, cp
- lb
.buffer
+ 1,
4954 lineno
, linecharno
);
4955 continuation
= false;
4956 while (*cp
&& *cp
!= '#' && *cp
!= ';')
4959 continuation
= true;
4960 else if (!c_isspace (*cp
))
4961 continuation
= false;
4965 continuation
= false;
4967 cp
= skip_spaces (cp
);
4970 : (continuation
= (*cp
== ',')))
4971 && (cp
= skip_spaces (cp
+ 1), *cp
&& *cp
!= '#'));
4974 cp
= skip_name (cp
);
4975 while (*cp
&& *cp
!= '#' && notinname (*cp
))
4986 * - /^[ \t]*function[ \t\n]+[^ \t\n(]+/
4987 * - /^[ \t]*class[ \t\n]+[^ \t\n]+/
4988 * - /^[ \t]*define\(\"[^\"]+/
4989 * Only with --members:
4990 * - /^[ \t]*var[ \t\n]+\$[^ \t\n=;]/
4991 * Idea by Diez B. Roggisch (2001)
4994 PHP_functions (FILE *inf
)
4997 bool search_identifier
= false;
4999 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5001 cp
= skip_spaces (cp
);
5003 if (search_identifier
5006 while (!notinname (*cp
))
5008 make_tag (name
, cp
- name
, true,
5009 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
5010 search_identifier
= false;
5012 else if (LOOKING_AT (cp
, "function"))
5015 cp
= skip_spaces (cp
+1);
5019 while (!notinname (*cp
))
5021 make_tag (name
, cp
- name
, true,
5022 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
5025 search_identifier
= true;
5027 else if (LOOKING_AT (cp
, "class"))
5032 while (*cp
!= '\0' && !c_isspace (*cp
))
5034 make_tag (name
, cp
- name
, false,
5035 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
5038 search_identifier
= true;
5040 else if (strneq (cp
, "define", 6)
5041 && (cp
= skip_spaces (cp
+6))
5043 && (*cp
== '"' || *cp
== '\''))
5047 while (*cp
!= quote
&& *cp
!= '\0')
5049 make_tag (name
, cp
- name
, false,
5050 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
5053 && LOOKING_AT (cp
, "var")
5057 while (!notinname (*cp
))
5059 make_tag (name
, cp
- name
, false,
5060 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
5067 * Cobol tag functions
5068 * We could look for anything that could be a paragraph name.
5069 * i.e. anything that starts in column 8 is one word and ends in a full stop.
5070 * Idea by Corny de Souza (1993)
5073 Cobol_paragraphs (FILE *inf
)
5075 register char *bp
, *ep
;
5077 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
5083 /* If eoln, compiler option or comment ignore whole line. */
5084 if (bp
[-1] != ' ' || !c_isalnum (bp
[0]))
5087 for (ep
= bp
; c_isalnum (*ep
) || *ep
== '-'; ep
++)
5090 make_tag (bp
, ep
- bp
, true,
5091 lb
.buffer
, ep
- lb
.buffer
+ 1, lineno
, linecharno
);
5098 * Ideas by Assar Westerlund <assar@sics.se> (2001)
5101 Makefile_targets (FILE *inf
)
5105 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
5107 if (*bp
== '\t' || *bp
== '#')
5109 while (*bp
!= '\0' && *bp
!= '=' && *bp
!= ':')
5111 if (*bp
== ':' || (globals
&& *bp
== '='))
5113 /* We should detect if there is more than one tag, but we do not.
5114 We just skip initial and final spaces. */
5115 char * namestart
= skip_spaces (lb
.buffer
);
5116 while (--bp
> namestart
)
5117 if (!notinname (*bp
))
5119 make_tag (namestart
, bp
- namestart
+ 1, true,
5120 lb
.buffer
, bp
- lb
.buffer
+ 2, lineno
, linecharno
);
5128 * Original code by Mosur K. Mohan (1989)
5130 * Locates tags for procedures & functions. Doesn't do any type- or
5131 * var-definitions. It does look for the keyword "extern" or
5132 * "forward" immediately following the procedure statement; if found,
5133 * the tag is skipped.
5136 Pascal_functions (FILE *inf
)
5138 linebuffer tline
; /* mostly copied from C_entries */
5140 int save_lineno
, namelen
, taglen
;
5143 bool /* each of these flags is true if: */
5144 incomment
, /* point is inside a comment */
5145 inquote
, /* point is inside '..' string */
5146 get_tagname
, /* point is after PROCEDURE/FUNCTION
5147 keyword, so next item = potential tag */
5148 found_tag
, /* point is after a potential tag */
5149 inparms
, /* point is within parameter-list */
5150 verify_tag
; /* point has passed the parm-list, so the
5151 next token will determine whether this
5152 is a FORWARD/EXTERN to be ignored, or
5153 whether it is a real tag */
5155 save_lcno
= save_lineno
= namelen
= taglen
= 0; /* keep compiler quiet */
5156 name
= NULL
; /* keep compiler quiet */
5159 linebuffer_init (&tline
);
5161 incomment
= inquote
= false;
5162 found_tag
= false; /* have a proc name; check if extern */
5163 get_tagname
= false; /* found "procedure" keyword */
5164 inparms
= false; /* found '(' after "proc" */
5165 verify_tag
= false; /* check if "extern" is ahead */
5168 while (perhaps_more_input (inf
)) /* long main loop to get next char */
5171 if (c
== '\0') /* if end of line */
5173 readline (&lb
, inf
);
5177 if (!((found_tag
&& verify_tag
)
5179 c
= *dbp
++; /* only if don't need *dbp pointing
5180 to the beginning of the name of
5181 the procedure or function */
5185 if (c
== '}') /* within { } comments */
5187 else if (c
== '*' && *dbp
== ')') /* within (* *) comments */
5204 inquote
= true; /* found first quote */
5206 case '{': /* found open { comment */
5210 if (*dbp
== '*') /* found open (* comment */
5215 else if (found_tag
) /* found '(' after tag, i.e., parm-list */
5218 case ')': /* end of parms list */
5223 if (found_tag
&& !inparms
) /* end of proc or fn stmt */
5230 if (found_tag
&& verify_tag
&& (*dbp
!= ' '))
5232 /* Check if this is an "extern" declaration. */
5235 if (c_tolower (*dbp
) == 'e')
5237 if (nocase_tail ("extern")) /* superfluous, really! */
5243 else if (c_tolower (*dbp
) == 'f')
5245 if (nocase_tail ("forward")) /* check for forward reference */
5251 if (found_tag
&& verify_tag
) /* not external proc, so make tag */
5255 make_tag (name
, namelen
, true,
5256 tline
.buffer
, taglen
, save_lineno
, save_lcno
);
5260 if (get_tagname
) /* grab name of proc or fn */
5267 /* Find block name. */
5268 for (cp
= dbp
+ 1; *cp
!= '\0' && !endtoken (*cp
); cp
++)
5271 /* Save all values for later tagging. */
5272 linebuffer_setlen (&tline
, lb
.len
);
5273 strcpy (tline
.buffer
, lb
.buffer
);
5274 save_lineno
= lineno
;
5275 save_lcno
= linecharno
;
5276 name
= tline
.buffer
+ (dbp
- lb
.buffer
);
5278 taglen
= cp
- lb
.buffer
+ 1;
5280 dbp
= cp
; /* set dbp to e-o-token */
5281 get_tagname
= false;
5285 /* And proceed to check for "extern". */
5287 else if (!incomment
&& !inquote
&& !found_tag
)
5289 /* Check for proc/fn keywords. */
5290 switch (c_tolower (c
))
5293 if (nocase_tail ("rocedure")) /* c = 'p', dbp has advanced */
5297 if (nocase_tail ("unction"))
5302 } /* while not eof */
5304 free (tline
.buffer
);
5309 * Lisp tag functions
5310 * look for (def or (DEF, quote or QUOTE
5313 static void L_getit (void);
5318 if (*dbp
== '\'') /* Skip prefix quote */
5320 else if (*dbp
== '(')
5323 /* Try to skip "(quote " */
5324 if (!LOOKING_AT (dbp
, "quote") && !LOOKING_AT (dbp
, "QUOTE"))
5325 /* Ok, then skip "(" before name in (defstruct (foo)) */
5326 dbp
= skip_spaces (dbp
);
5328 get_tag (dbp
, NULL
);
5332 Lisp_functions (FILE *inf
)
5334 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
5339 /* "(defvar foo)" is a declaration rather than a definition. */
5343 if (LOOKING_AT (p
, "defvar"))
5345 p
= skip_name (p
); /* past var name */
5346 p
= skip_spaces (p
);
5352 if (strneq (dbp
+ 1, "cl-", 3) || strneq (dbp
+ 1, "CL-", 3))
5355 if (strneq (dbp
+1, "def", 3) || strneq (dbp
+1, "DEF", 3))
5357 dbp
= skip_non_spaces (dbp
);
5358 dbp
= skip_spaces (dbp
);
5363 /* Check for (foo::defmumble name-defined ... */
5366 while (!notinname (*dbp
) && *dbp
!= ':');
5371 while (*dbp
== ':');
5373 if (strneq (dbp
, "def", 3) || strneq (dbp
, "DEF", 3))
5375 dbp
= skip_non_spaces (dbp
);
5376 dbp
= skip_spaces (dbp
);
5386 * Lua script language parsing
5387 * Original code by David A. Capello <dacap@users.sourceforge.net> (2004)
5389 * "function" and "local function" are tags if they start at column 1.
5392 Lua_functions (FILE *inf
)
5396 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
5398 bp
= skip_spaces (bp
);
5399 if (bp
[0] != 'f' && bp
[0] != 'l')
5402 (void)LOOKING_AT (bp
, "local"); /* skip possible "local" */
5404 if (LOOKING_AT (bp
, "function"))
5406 char *tag_name
, *tp_dot
, *tp_colon
;
5408 get_tag (bp
, &tag_name
);
5409 /* If the tag ends with ".foo" or ":foo", make an additional tag for
5411 tp_dot
= strrchr (tag_name
, '.');
5412 tp_colon
= strrchr (tag_name
, ':');
5413 if (tp_dot
|| tp_colon
)
5415 char *p
= tp_dot
> tp_colon
? tp_dot
: tp_colon
;
5416 int len_add
= p
- tag_name
+ 1;
5418 get_tag (bp
+ len_add
, NULL
);
5427 * Just look for lines where the first character is '/'
5428 * Also look at "defineps" for PSWrap
5430 * Richard Mlynarik <mly@adoc.xerox.com> (1997)
5431 * Masatake Yamato <masata-y@is.aist-nara.ac.jp> (1999)
5434 PS_functions (FILE *inf
)
5436 register char *bp
, *ep
;
5438 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
5443 *ep
!= '\0' && *ep
!= ' ' && *ep
!= '{';
5446 make_tag (bp
, ep
- bp
, true,
5447 lb
.buffer
, ep
- lb
.buffer
+ 1, lineno
, linecharno
);
5449 else if (LOOKING_AT (bp
, "defineps"))
5457 * Ignore anything after \ followed by space or in ( )
5458 * Look for words defined by :
5459 * Look for constant, code, create, defer, value, and variable
5460 * OBP extensions: Look for buffer:, field,
5461 * Ideas by Eduardo Horvath <eeh@netbsd.org> (2004)
5464 Forth_words (FILE *inf
)
5468 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
5469 while ((bp
= skip_spaces (bp
))[0] != '\0')
5470 if (bp
[0] == '\\' && c_isspace (bp
[1]))
5471 break; /* read next line */
5472 else if (bp
[0] == '(' && c_isspace (bp
[1]))
5473 do /* skip to ) or eol */
5475 while (*bp
!= ')' && *bp
!= '\0');
5476 else if ((bp
[0] == ':' && c_isspace (bp
[1]) && bp
++)
5477 || LOOKING_AT_NOCASE (bp
, "constant")
5478 || LOOKING_AT_NOCASE (bp
, "code")
5479 || LOOKING_AT_NOCASE (bp
, "create")
5480 || LOOKING_AT_NOCASE (bp
, "defer")
5481 || LOOKING_AT_NOCASE (bp
, "value")
5482 || LOOKING_AT_NOCASE (bp
, "variable")
5483 || LOOKING_AT_NOCASE (bp
, "buffer:")
5484 || LOOKING_AT_NOCASE (bp
, "field"))
5485 get_tag (skip_spaces (bp
), NULL
); /* Yay! A definition! */
5487 bp
= skip_non_spaces (bp
);
5492 * Scheme tag functions
5493 * look for (def... xyzzy
5495 * (def ... ((...(xyzzy ....
5497 * Original code by Ken Haase (1985?)
5500 Scheme_functions (FILE *inf
)
5504 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
5506 if (strneq (bp
, "(def", 4) || strneq (bp
, "(DEF", 4))
5508 bp
= skip_non_spaces (bp
+4);
5509 /* Skip over open parens and white space. Don't continue past
5511 while (*bp
&& notinname (*bp
))
5515 if (LOOKING_AT (bp
, "(SET!") || LOOKING_AT (bp
, "(set!"))
5521 /* Find tags in TeX and LaTeX input files. */
5523 /* TEX_toktab is a table of TeX control sequences that define tags.
5524 * Each entry records one such control sequence.
5526 * Original code from who knows whom.
5528 * Stefan Monnier (2002)
5531 static linebuffer
*TEX_toktab
= NULL
; /* Table with tag tokens */
5533 /* Default set of control sequences to put into TEX_toktab.
5534 The value of environment var TEXTAGS is prepended to this. */
5535 static const char *TEX_defenv
= "\
5536 :chapter:section:subsection:subsubsection:eqno:label:ref:cite:bibitem\
5537 :part:appendix:entry:index:def\
5538 :newcommand:renewcommand:newenvironment:renewenvironment";
5540 static void TEX_decode_env (const char *, const char *);
5543 * TeX/LaTeX scanning loop.
5546 TeX_commands (FILE *inf
)
5551 char TEX_esc
= '\0';
5552 char TEX_opgrp
, TEX_clgrp
;
5554 /* Initialize token table once from environment. */
5555 if (TEX_toktab
== NULL
)
5556 TEX_decode_env ("TEXTAGS", TEX_defenv
);
5558 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5560 /* Look at each TEX keyword in line. */
5563 /* Look for a TEX escape. */
5567 if (c
== '\0' || c
== '%')
5570 /* Select either \ or ! as escape character, whichever comes
5571 first outside a comment. */
5592 for (key
= TEX_toktab
; key
->buffer
!= NULL
; key
++)
5593 if (strneq (cp
, key
->buffer
, key
->len
))
5596 int namelen
, linelen
;
5599 cp
= skip_spaces (cp
+ key
->len
);
5600 if (*cp
== TEX_opgrp
)
5606 (!c_isspace (*p
) && *p
!= '#' &&
5607 *p
!= TEX_opgrp
&& *p
!= TEX_clgrp
);
5612 if (!opgrp
|| *p
== TEX_clgrp
)
5614 while (*p
!= '\0' && *p
!= TEX_opgrp
&& *p
!= TEX_clgrp
)
5616 linelen
= p
- lb
.buffer
+ 1;
5618 make_tag (cp
, namelen
, true,
5619 lb
.buffer
, linelen
, lineno
, linecharno
);
5620 goto tex_next_line
; /* We only tag a line once */
5628 /* Read environment and prepend it to the default string.
5629 Build token table. */
5631 TEX_decode_env (const char *evarname
, const char *defenv
)
5633 register const char *env
, *p
;
5636 /* Append default string to environment. */
5637 env
= getenv (evarname
);
5641 env
= concat (env
, defenv
, "");
5643 /* Allocate a token table */
5644 for (len
= 1, p
= env
; (p
= strchr (p
, ':')); )
5647 TEX_toktab
= xnew (len
, linebuffer
);
5649 /* Unpack environment string into token table. Be careful about */
5650 /* zero-length strings (leading ':', "::" and trailing ':') */
5651 for (i
= 0; *env
!= '\0';)
5653 p
= strchr (env
, ':');
5654 if (!p
) /* End of environment string. */
5655 p
= env
+ strlen (env
);
5657 { /* Only non-zero strings. */
5658 TEX_toktab
[i
].buffer
= savenstr (env
, p
- env
);
5659 TEX_toktab
[i
].len
= p
- env
;
5666 TEX_toktab
[i
].buffer
= NULL
; /* Mark end of table. */
5667 TEX_toktab
[i
].len
= 0;
5674 /* Texinfo support. Dave Love, Mar. 2000. */
5676 Texinfo_nodes (FILE *inf
)
5679 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5680 if (LOOKING_AT (cp
, "@node"))
5683 while (*cp
!= '\0' && *cp
!= ',')
5685 make_tag (start
, cp
- start
, true,
5686 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
5693 * Contents of <title>, <h1>, <h2>, <h3> are tags.
5694 * Contents of <a name=xxx> are tags with name xxx.
5696 * Francesco Potortì, 2002.
5699 HTML_labels (FILE *inf
)
5701 bool getnext
= false; /* next text outside of HTML tags is a tag */
5702 bool skiptag
= false; /* skip to the end of the current HTML tag */
5703 bool intag
= false; /* inside an html tag, looking for ID= */
5704 bool inanchor
= false; /* when INTAG, is an anchor, look for NAME= */
5708 linebuffer_setlen (&token_name
, 0); /* no name in buffer */
5710 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
5711 for (;;) /* loop on the same line */
5713 if (skiptag
) /* skip HTML tag */
5715 while (*dbp
!= '\0' && *dbp
!= '>')
5721 continue; /* look on the same line */
5723 break; /* go to next line */
5726 else if (intag
) /* look for "name=" or "id=" */
5728 while (*dbp
!= '\0' && *dbp
!= '>'
5729 && c_tolower (*dbp
) != 'n' && c_tolower (*dbp
) != 'i')
5732 break; /* go to next line */
5737 continue; /* look on the same line */
5739 if ((inanchor
&& LOOKING_AT_NOCASE (dbp
, "name="))
5740 || LOOKING_AT_NOCASE (dbp
, "id="))
5742 bool quoted
= (dbp
[0] == '"');
5745 for (end
= ++dbp
; *end
!= '\0' && *end
!= '"'; end
++)
5748 for (end
= dbp
; *end
!= '\0' && intoken (*end
); end
++)
5750 linebuffer_setlen (&token_name
, end
- dbp
);
5751 memcpy (token_name
.buffer
, dbp
, end
- dbp
);
5752 token_name
.buffer
[end
- dbp
] = '\0';
5755 intag
= false; /* we found what we looked for */
5756 skiptag
= true; /* skip to the end of the tag */
5757 getnext
= true; /* then grab the text */
5758 continue; /* look on the same line */
5763 else if (getnext
) /* grab next tokens and tag them */
5765 dbp
= skip_spaces (dbp
);
5767 break; /* go to next line */
5771 inanchor
= (c_tolower (dbp
[1]) == 'a' && !intoken (dbp
[2]));
5772 continue; /* look on the same line */
5775 for (end
= dbp
+ 1; *end
!= '\0' && *end
!= '<'; end
++)
5777 make_tag (token_name
.buffer
, token_name
.len
, true,
5778 dbp
, end
- dbp
, lineno
, linecharno
);
5779 linebuffer_setlen (&token_name
, 0); /* no name in buffer */
5781 break; /* go to next line */
5784 else /* look for an interesting HTML tag */
5786 while (*dbp
!= '\0' && *dbp
!= '<')
5789 break; /* go to next line */
5791 if (c_tolower (dbp
[1]) == 'a' && !intoken (dbp
[2]))
5794 continue; /* look on the same line */
5796 else if (LOOKING_AT_NOCASE (dbp
, "<title>")
5797 || LOOKING_AT_NOCASE (dbp
, "<h1>")
5798 || LOOKING_AT_NOCASE (dbp
, "<h2>")
5799 || LOOKING_AT_NOCASE (dbp
, "<h3>"))
5803 continue; /* look on the same line */
5814 * Assumes that the predicate or rule starts at column 0.
5815 * Only the first clause of a predicate or rule is added.
5816 * Original code by Sunichirou Sugou (1989)
5817 * Rewritten by Anders Lindgren (1996)
5819 static size_t prolog_pr (char *, char *);
5820 static void prolog_skip_comment (linebuffer
*, FILE *);
5821 static size_t prolog_atom (char *, size_t);
5824 Prolog_functions (FILE *inf
)
5834 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5836 if (cp
[0] == '\0') /* Empty line */
5838 else if (c_isspace (cp
[0])) /* Not a predicate */
5840 else if (cp
[0] == '/' && cp
[1] == '*') /* comment. */
5841 prolog_skip_comment (&lb
, inf
);
5842 else if ((len
= prolog_pr (cp
, last
)) > 0)
5844 /* Predicate or rule. Store the function name so that we
5845 only generate a tag for the first clause. */
5847 last
= xnew (len
+ 1, char);
5848 else if (len
+ 1 > allocated
)
5849 xrnew (last
, len
+ 1, char);
5850 allocated
= len
+ 1;
5851 memcpy (last
, cp
, len
);
5860 prolog_skip_comment (linebuffer
*plb
, FILE *inf
)
5866 for (cp
= plb
->buffer
; *cp
!= '\0'; cp
++)
5867 if (cp
[0] == '*' && cp
[1] == '/')
5869 readline (plb
, inf
);
5871 while (perhaps_more_input (inf
));
5875 * A predicate or rule definition is added if it matches:
5876 * <beginning of line><Prolog Atom><whitespace>(
5877 * or <beginning of line><Prolog Atom><whitespace>:-
5879 * It is added to the tags database if it doesn't match the
5880 * name of the previous clause header.
5882 * Return the size of the name of the predicate or rule, or 0 if no
5886 prolog_pr (char *s
, char *last
)
5888 /* Name of last clause. */
5893 pos
= prolog_atom (s
, 0);
5898 pos
= skip_spaces (s
+ pos
) - s
;
5901 || (s
[pos
] == '(' && (pos
+= 1))
5902 || (s
[pos
] == ':' && s
[pos
+ 1] == '-' && (pos
+= 2)))
5903 && (last
== NULL
/* save only the first clause */
5904 || len
!= strlen (last
)
5905 || !strneq (s
, last
, len
)))
5907 make_tag (s
, len
, true, s
, pos
, lineno
, linecharno
);
5915 * Consume a Prolog atom.
5916 * Return the number of bytes consumed, or 0 if there was an error.
5918 * A prolog atom, in this context, could be one of:
5919 * - An alphanumeric sequence, starting with a lower case letter.
5920 * - A quoted arbitrary string. Single quotes can escape themselves.
5921 * Backslash quotes everything.
5924 prolog_atom (char *s
, size_t pos
)
5930 if (c_islower (s
[pos
]) || s
[pos
] == '_')
5932 /* The atom is unquoted. */
5934 while (c_isalnum (s
[pos
]) || s
[pos
] == '_')
5938 return pos
- origpos
;
5940 else if (s
[pos
] == '\'')
5951 pos
++; /* A double quote */
5953 else if (s
[pos
] == '\0')
5954 /* Multiline quoted atoms are ignored. */
5956 else if (s
[pos
] == '\\')
5958 if (s
[pos
+1] == '\0')
5965 return pos
- origpos
;
5973 * Support for Erlang
5975 * Generates tags for functions, defines, and records.
5976 * Assumes that Erlang functions start at column 0.
5977 * Original code by Anders Lindgren (1996)
5979 static int erlang_func (char *, char *);
5980 static void erlang_attribute (char *);
5981 static int erlang_atom (char *);
5984 Erlang_functions (FILE *inf
)
5994 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5996 if (cp
[0] == '\0') /* Empty line */
5998 else if (c_isspace (cp
[0])) /* Not function nor attribute */
6000 else if (cp
[0] == '%') /* comment */
6002 else if (cp
[0] == '"') /* Sometimes, strings start in column one */
6004 else if (cp
[0] == '-') /* attribute, e.g. "-define" */
6006 erlang_attribute (cp
);
6013 else if ((len
= erlang_func (cp
, last
)) > 0)
6016 * Function. Store the function name so that we only
6017 * generates a tag for the first clause.
6020 last
= xnew (len
+ 1, char);
6021 else if (len
+ 1 > allocated
)
6022 xrnew (last
, len
+ 1, char);
6023 allocated
= len
+ 1;
6024 memcpy (last
, cp
, len
);
6033 * A function definition is added if it matches:
6034 * <beginning of line><Erlang Atom><whitespace>(
6036 * It is added to the tags database if it doesn't match the
6037 * name of the previous clause header.
6039 * Return the size of the name of the function, or 0 if no function
6043 erlang_func (char *s
, char *last
)
6045 /* Name of last clause. */
6050 pos
= erlang_atom (s
);
6055 pos
= skip_spaces (s
+ pos
) - s
;
6057 /* Save only the first clause. */
6060 || len
!= (int)strlen (last
)
6061 || !strneq (s
, last
, len
)))
6063 make_tag (s
, len
, true, s
, pos
, lineno
, linecharno
);
6072 * Handle attributes. Currently, tags are generated for defines
6075 * They are on the form:
6076 * -define(foo, bar).
6077 * -define(Foo(M, N), M+N).
6078 * -record(graph, {vtab = notable, cyclic = true}).
6081 erlang_attribute (char *s
)
6085 if ((LOOKING_AT (cp
, "-define") || LOOKING_AT (cp
, "-record"))
6088 int len
= erlang_atom (skip_spaces (cp
));
6090 make_tag (cp
, len
, true, s
, cp
+ len
- s
, lineno
, linecharno
);
6097 * Consume an Erlang atom (or variable).
6098 * Return the number of bytes consumed, or -1 if there was an error.
6101 erlang_atom (char *s
)
6105 if (c_isalpha (s
[pos
]) || s
[pos
] == '_')
6107 /* The atom is unquoted. */
6110 while (c_isalnum (s
[pos
]) || s
[pos
] == '_');
6112 else if (s
[pos
] == '\'')
6114 for (pos
++; s
[pos
] != '\''; pos
++)
6115 if (s
[pos
] == '\0' /* multiline quoted atoms are ignored */
6116 || (s
[pos
] == '\\' && s
[++pos
] == '\0'))
6125 static char *scan_separators (char *);
6126 static void add_regex (char *, language
*);
6127 static char *substitute (char *, char *, struct re_registers
*);
6130 * Take a string like "/blah/" and turn it into "blah", verifying
6131 * that the first and last characters are the same, and handling
6132 * quoted separator characters. Actually, stops on the occurrence of
6133 * an unquoted separator. Also process \t, \n, etc. and turn into
6134 * appropriate characters. Works in place. Null terminates name string.
6135 * Returns pointer to terminating separator, or NULL for
6136 * unterminated regexps.
6139 scan_separators (char *name
)
6142 char *copyto
= name
;
6143 bool quoted
= false;
6145 for (++name
; *name
!= '\0'; ++name
)
6151 case 'a': *copyto
++ = '\007'; break; /* BEL (bell) */
6152 case 'b': *copyto
++ = '\b'; break; /* BS (back space) */
6153 case 'd': *copyto
++ = 0177; break; /* DEL (delete) */
6154 case 'e': *copyto
++ = 033; break; /* ESC (delete) */
6155 case 'f': *copyto
++ = '\f'; break; /* FF (form feed) */
6156 case 'n': *copyto
++ = '\n'; break; /* NL (new line) */
6157 case 'r': *copyto
++ = '\r'; break; /* CR (carriage return) */
6158 case 't': *copyto
++ = '\t'; break; /* TAB (horizontal tab) */
6159 case 'v': *copyto
++ = '\v'; break; /* VT (vertical tab) */
6165 /* Something else is quoted, so preserve the quote. */
6173 else if (*name
== '\\')
6175 else if (*name
== sep
)
6181 name
= NULL
; /* signal unterminated regexp */
6183 /* Terminate copied string. */
6188 /* Look at the argument of --regex or --no-regex and do the right
6189 thing. Same for each line of a regexp file. */
6191 analyze_regex (char *regex_arg
)
6193 if (regex_arg
== NULL
)
6195 free_regexps (); /* --no-regex: remove existing regexps */
6199 /* A real --regexp option or a line in a regexp file. */
6200 switch (regex_arg
[0])
6202 /* Comments in regexp file or null arg to --regex. */
6208 /* Read a regex file. This is recursive and may result in a
6209 loop, which will stop when the file descriptors are exhausted. */
6213 linebuffer regexbuf
;
6214 char *regexfile
= regex_arg
+ 1;
6216 /* regexfile is a file containing regexps, one per line. */
6217 regexfp
= fopen (regexfile
, "r" FOPEN_BINARY
);
6218 if (regexfp
== NULL
)
6220 linebuffer_init (®exbuf
);
6221 while (readline_internal (®exbuf
, regexfp
, regexfile
) > 0)
6222 analyze_regex (regexbuf
.buffer
);
6223 free (regexbuf
.buffer
);
6224 if (fclose (regexfp
) != 0)
6229 /* Regexp to be used for a specific language only. */
6233 char *lang_name
= regex_arg
+ 1;
6236 for (cp
= lang_name
; *cp
!= '}'; cp
++)
6239 error ("unterminated language name in regex: %s", regex_arg
);
6243 lang
= get_language_from_langname (lang_name
);
6246 add_regex (cp
, lang
);
6250 /* Regexp to be used for any language. */
6252 add_regex (regex_arg
, NULL
);
6257 /* Separate the regexp pattern, compile it,
6258 and care for optional name and modifiers. */
6260 add_regex (char *regexp_pattern
, language
*lang
)
6262 static struct re_pattern_buffer zeropattern
;
6263 char sep
, *pat
, *name
, *modifiers
;
6266 struct re_pattern_buffer
*patbuf
;
6269 force_explicit_name
= true, /* do not use implicit tag names */
6270 ignore_case
= false, /* case is significant */
6271 multi_line
= false, /* matches are done one line at a time */
6272 single_line
= false; /* dot does not match newline */
6275 if (strlen (regexp_pattern
) < 3)
6277 error ("null regexp");
6280 sep
= regexp_pattern
[0];
6281 name
= scan_separators (regexp_pattern
);
6284 error ("%s: unterminated regexp", regexp_pattern
);
6289 error ("null name for regexp \"%s\"", regexp_pattern
);
6292 modifiers
= scan_separators (name
);
6293 if (modifiers
== NULL
) /* no terminating separator --> no name */
6299 modifiers
+= 1; /* skip separator */
6301 /* Parse regex modifiers. */
6302 for (; modifiers
[0] != '\0'; modifiers
++)
6303 switch (modifiers
[0])
6306 if (modifiers
== name
)
6307 error ("forcing explicit tag name but no name, ignoring");
6308 force_explicit_name
= true;
6318 need_filebuf
= true;
6321 error ("invalid regexp modifier '%c', ignoring", modifiers
[0]);
6325 patbuf
= xnew (1, struct re_pattern_buffer
);
6326 *patbuf
= zeropattern
;
6329 static char lc_trans
[UCHAR_MAX
+ 1];
6331 for (i
= 0; i
< UCHAR_MAX
+ 1; i
++)
6332 lc_trans
[i
] = c_tolower (i
);
6333 patbuf
->translate
= lc_trans
; /* translation table to fold case */
6337 pat
= concat ("^", regexp_pattern
, ""); /* anchor to beginning of line */
6339 pat
= regexp_pattern
;
6342 re_set_syntax (RE_SYNTAX_EMACS
| RE_DOT_NEWLINE
);
6344 re_set_syntax (RE_SYNTAX_EMACS
);
6346 err
= re_compile_pattern (pat
, strlen (pat
), patbuf
);
6351 error ("%s while compiling pattern", err
);
6356 p_head
= xnew (1, regexp
);
6357 p_head
->pattern
= savestr (regexp_pattern
);
6358 p_head
->p_next
= rp
;
6359 p_head
->lang
= lang
;
6360 p_head
->pat
= patbuf
;
6361 p_head
->name
= savestr (name
);
6362 p_head
->error_signaled
= false;
6363 p_head
->force_explicit_name
= force_explicit_name
;
6364 p_head
->ignore_case
= ignore_case
;
6365 p_head
->multi_line
= multi_line
;
6369 * Do the substitutions indicated by the regular expression and
6373 substitute (char *in
, char *out
, struct re_registers
*regs
)
6376 int size
, dig
, diglen
;
6379 size
= strlen (out
);
6381 /* Pass 1: figure out how much to allocate by finding all \N strings. */
6382 if (out
[size
- 1] == '\\')
6383 fatal ("pattern error in \"%s\"", out
);
6384 for (t
= strchr (out
, '\\');
6386 t
= strchr (t
+ 2, '\\'))
6387 if (c_isdigit (t
[1]))
6390 diglen
= regs
->end
[dig
] - regs
->start
[dig
];
6396 /* Allocate space and do the substitutions. */
6398 result
= xnew (size
+ 1, char);
6400 for (t
= result
; *out
!= '\0'; out
++)
6401 if (*out
== '\\' && c_isdigit (*++out
))
6404 diglen
= regs
->end
[dig
] - regs
->start
[dig
];
6405 memcpy (t
, in
+ regs
->start
[dig
], diglen
);
6412 assert (t
<= result
+ size
);
6413 assert (t
- result
== (int)strlen (result
));
6418 /* Deallocate all regexps. */
6423 while (p_head
!= NULL
)
6425 rp
= p_head
->p_next
;
6426 free (p_head
->pattern
);
6427 free (p_head
->name
);
6435 * Reads the whole file as a single string from `filebuf' and looks for
6436 * multi-line regular expressions, creating tags on matches.
6437 * readline already dealt with normal regexps.
6439 * Idea by Ben Wing <ben@666.com> (2002).
6442 regex_tag_multiline (void)
6444 char *buffer
= filebuf
.buffer
;
6448 for (rp
= p_head
; rp
!= NULL
; rp
= rp
->p_next
)
6452 if (!rp
->multi_line
)
6453 continue; /* skip normal regexps */
6455 /* Generic initializations before parsing file from memory. */
6456 lineno
= 1; /* reset global line number */
6457 charno
= 0; /* reset global char number */
6458 linecharno
= 0; /* reset global char number of line start */
6460 /* Only use generic regexps or those for the current language. */
6461 if (rp
->lang
!= NULL
&& rp
->lang
!= curfdp
->lang
)
6464 while (match
>= 0 && match
< filebuf
.len
)
6466 match
= re_search (rp
->pat
, buffer
, filebuf
.len
, charno
,
6467 filebuf
.len
- match
, &rp
->regs
);
6472 if (!rp
->error_signaled
)
6474 error ("regexp stack overflow while matching \"%s\"",
6476 rp
->error_signaled
= true;
6483 if (match
== rp
->regs
.end
[0])
6485 if (!rp
->error_signaled
)
6487 error ("regexp matches the empty string: \"%s\"",
6489 rp
->error_signaled
= true;
6491 match
= -3; /* exit from while loop */
6495 /* Match occurred. Construct a tag. */
6496 while (charno
< rp
->regs
.end
[0])
6497 if (buffer
[charno
++] == '\n')
6498 lineno
++, linecharno
= charno
;
6500 if (name
[0] == '\0')
6502 else /* make a named tag */
6503 name
= substitute (buffer
, rp
->name
, &rp
->regs
);
6504 if (rp
->force_explicit_name
)
6505 /* Force explicit tag name, if a name is there. */
6506 pfnote (name
, true, buffer
+ linecharno
,
6507 charno
- linecharno
+ 1, lineno
, linecharno
);
6509 make_tag (name
, strlen (name
), true, buffer
+ linecharno
,
6510 charno
- linecharno
+ 1, lineno
, linecharno
);
6519 nocase_tail (const char *cp
)
6523 while (*cp
!= '\0' && c_tolower (*cp
) == c_tolower (dbp
[len
]))
6525 if (*cp
== '\0' && !intoken (dbp
[len
]))
6534 get_tag (register char *bp
, char **namepp
)
6536 register char *cp
= bp
;
6540 /* Go till you get to white space or a syntactic break */
6541 for (cp
= bp
+ 1; !notinname (*cp
); cp
++)
6543 make_tag (bp
, cp
- bp
, true,
6544 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
6548 *namepp
= savenstr (bp
, cp
- bp
);
6552 * Read a line of text from `stream' into `lbp', excluding the
6553 * newline or CR-NL, if any. Return the number of characters read from
6554 * `stream', which is the length of the line including the newline.
6556 * On DOS or Windows we do not count the CR character, if any before the
6557 * NL, in the returned length; this mirrors the behavior of Emacs on those
6558 * platforms (for text files, it translates CR-NL to NL as it reads in the
6561 * If multi-line regular expressions are requested, each line read is
6562 * appended to `filebuf'.
6565 readline_internal (linebuffer
*lbp
, FILE *stream
, char const *filename
)
6567 char *buffer
= lbp
->buffer
;
6568 char *p
= lbp
->buffer
;
6572 pend
= p
+ lbp
->size
; /* Separate to avoid 386/IX compiler bug. */
6576 register int c
= getc (stream
);
6579 /* We're at the end of linebuffer: expand it. */
6581 xrnew (buffer
, lbp
->size
, char);
6582 p
+= buffer
- lbp
->buffer
;
6583 pend
= buffer
+ lbp
->size
;
6584 lbp
->buffer
= buffer
;
6588 if (ferror (stream
))
6596 if (p
> buffer
&& p
[-1] == '\r')
6610 lbp
->len
= p
- buffer
;
6612 if (need_filebuf
/* we need filebuf for multi-line regexps */
6613 && chars_deleted
> 0) /* not at EOF */
6615 while (filebuf
.size
<= filebuf
.len
+ lbp
->len
+ 1) /* +1 for \n */
6617 /* Expand filebuf. */
6619 xrnew (filebuf
.buffer
, filebuf
.size
, char);
6621 memcpy (filebuf
.buffer
+ filebuf
.len
, lbp
->buffer
, lbp
->len
);
6622 filebuf
.len
+= lbp
->len
;
6623 filebuf
.buffer
[filebuf
.len
++] = '\n';
6624 filebuf
.buffer
[filebuf
.len
] = '\0';
6627 return lbp
->len
+ chars_deleted
;
6631 * Like readline_internal, above, but in addition try to match the
6632 * input line against relevant regular expressions and manage #line
6636 readline (linebuffer
*lbp
, FILE *stream
)
6640 linecharno
= charno
; /* update global char number of line start */
6641 result
= readline_internal (lbp
, stream
, infilename
); /* read line */
6642 lineno
+= 1; /* increment global line number */
6643 charno
+= result
; /* increment global char number */
6645 /* Honor #line directives. */
6646 if (!no_line_directive
)
6648 static bool discard_until_line_directive
;
6650 /* Check whether this is a #line directive. */
6651 if (result
> 12 && strneq (lbp
->buffer
, "#line ", 6))
6656 if (sscanf (lbp
->buffer
, "#line %u \"%n", &lno
, &start
) >= 1
6657 && start
> 0) /* double quote character found */
6659 char *endp
= lbp
->buffer
+ start
;
6661 while ((endp
= strchr (endp
, '"')) != NULL
6662 && endp
[-1] == '\\')
6665 /* Ok, this is a real #line directive. Let's deal with it. */
6667 char *taggedabsname
; /* absolute name of original file */
6668 char *taggedfname
; /* name of original file as given */
6669 char *name
; /* temp var */
6671 discard_until_line_directive
= false; /* found it */
6672 name
= lbp
->buffer
+ start
;
6674 canonicalize_filename (name
);
6675 taggedabsname
= absolute_filename (name
, tagfiledir
);
6676 if (filename_is_absolute (name
)
6677 || filename_is_absolute (curfdp
->infname
))
6678 taggedfname
= savestr (taggedabsname
);
6680 taggedfname
= relative_filename (taggedabsname
,tagfiledir
);
6682 if (streq (curfdp
->taggedfname
, taggedfname
))
6683 /* The #line directive is only a line number change. We
6684 deal with this afterwards. */
6687 /* The tags following this #line directive should be
6688 attributed to taggedfname. In order to do this, set
6689 curfdp accordingly. */
6691 fdesc
*fdp
; /* file description pointer */
6693 /* Go look for a file description already set up for the
6694 file indicated in the #line directive. If there is
6695 one, use it from now until the next #line
6697 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
6698 if (streq (fdp
->infname
, curfdp
->infname
)
6699 && streq (fdp
->taggedfname
, taggedfname
))
6700 /* If we remove the second test above (after the &&)
6701 then all entries pertaining to the same file are
6702 coalesced in the tags file. If we use it, then
6703 entries pertaining to the same file but generated
6704 from different files (via #line directives) will
6705 go into separate sections in the tags file. These
6706 alternatives look equivalent. The first one
6707 destroys some apparently useless information. */
6713 /* Else, if we already tagged the real file, skip all
6714 input lines until the next #line directive. */
6715 if (fdp
== NULL
) /* not found */
6716 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
6717 if (streq (fdp
->infabsname
, taggedabsname
))
6719 discard_until_line_directive
= true;
6723 /* Else create a new file description and use that from
6724 now on, until the next #line directive. */
6725 if (fdp
== NULL
) /* not found */
6728 fdhead
= xnew (1, fdesc
);
6729 *fdhead
= *curfdp
; /* copy curr. file description */
6731 fdhead
->infname
= savestr (curfdp
->infname
);
6732 fdhead
->infabsname
= savestr (curfdp
->infabsname
);
6733 fdhead
->infabsdir
= savestr (curfdp
->infabsdir
);
6734 fdhead
->taggedfname
= taggedfname
;
6735 fdhead
->usecharno
= false;
6736 fdhead
->prop
= NULL
;
6737 fdhead
->written
= false;
6741 free (taggedabsname
);
6743 readline (lbp
, stream
);
6745 } /* if a real #line directive */
6746 } /* if #line is followed by a number */
6747 } /* if line begins with "#line " */
6749 /* If we are here, no #line directive was found. */
6750 if (discard_until_line_directive
)
6754 /* Do a tail recursion on ourselves, thus discarding the contents
6755 of the line buffer. */
6756 readline (lbp
, stream
);
6760 discard_until_line_directive
= false;
6763 } /* if #line directives should be considered */
6770 /* Match against relevant regexps. */
6772 for (rp
= p_head
; rp
!= NULL
; rp
= rp
->p_next
)
6774 /* Only use generic regexps or those for the current language.
6775 Also do not use multiline regexps, which is the job of
6776 regex_tag_multiline. */
6777 if ((rp
->lang
!= NULL
&& rp
->lang
!= fdhead
->lang
)
6781 match
= re_match (rp
->pat
, lbp
->buffer
, lbp
->len
, 0, &rp
->regs
);
6786 if (!rp
->error_signaled
)
6788 error ("regexp stack overflow while matching \"%s\"",
6790 rp
->error_signaled
= true;
6797 /* Empty string matched. */
6798 if (!rp
->error_signaled
)
6800 error ("regexp matches the empty string: \"%s\"", rp
->pattern
);
6801 rp
->error_signaled
= true;
6805 /* Match occurred. Construct a tag. */
6807 if (name
[0] == '\0')
6809 else /* make a named tag */
6810 name
= substitute (lbp
->buffer
, rp
->name
, &rp
->regs
);
6811 if (rp
->force_explicit_name
)
6812 /* Force explicit tag name, if a name is there. */
6813 pfnote (name
, true, lbp
->buffer
, match
, lineno
, linecharno
);
6815 make_tag (name
, strlen (name
), true,
6816 lbp
->buffer
, match
, lineno
, linecharno
);
6825 * Return a pointer to a space of size strlen(cp)+1 allocated
6826 * with xnew where the string CP has been copied.
6829 savestr (const char *cp
)
6831 return savenstr (cp
, strlen (cp
));
6835 * Return a pointer to a space of size LEN+1 allocated with xnew where
6836 * the string CP has been copied for at most the first LEN characters.
6839 savenstr (const char *cp
, int len
)
6841 char *dp
= xnew (len
+ 1, char);
6843 return memcpy (dp
, cp
, len
);
6846 /* Skip spaces (end of string is not space), return new pointer. */
6848 skip_spaces (char *cp
)
6850 while (c_isspace (*cp
))
6855 /* Skip non spaces, except end of string, return new pointer. */
6857 skip_non_spaces (char *cp
)
6859 while (*cp
!= '\0' && !c_isspace (*cp
))
6864 /* Skip any chars in the "name" class.*/
6866 skip_name (char *cp
)
6868 /* '\0' is a notinname() so loop stops there too */
6869 while (! notinname (*cp
))
6874 /* Print error message and exit. */
6876 fatal (char const *format
, ...)
6879 va_start (ap
, format
);
6880 verror (format
, ap
);
6882 exit (EXIT_FAILURE
);
6886 pfatal (const char *s1
)
6889 exit (EXIT_FAILURE
);
6893 suggest_asking_for_help (void)
6895 fprintf (stderr
, "\tTry '%s --help' for a complete list of options.\n",
6897 exit (EXIT_FAILURE
);
6900 /* Output a diagnostic with printf-style FORMAT and args. */
6902 error (const char *format
, ...)
6905 va_start (ap
, format
);
6906 verror (format
, ap
);
6911 verror (char const *format
, va_list ap
)
6913 fprintf (stderr
, "%s: ", progname
);
6914 vfprintf (stderr
, format
, ap
);
6915 fprintf (stderr
, "\n");
6918 /* Return a newly-allocated string whose contents
6919 concatenate those of s1, s2, s3. */
6921 concat (const char *s1
, const char *s2
, const char *s3
)
6923 int len1
= strlen (s1
), len2
= strlen (s2
), len3
= strlen (s3
);
6924 char *result
= xnew (len1
+ len2
+ len3
+ 1, char);
6926 strcpy (result
, s1
);
6927 strcpy (result
+ len1
, s2
);
6928 strcpy (result
+ len1
+ len2
, s3
);
6934 /* Does the same work as the system V getcwd, but does not need to
6935 guess the buffer size in advance. */
6940 char *path
= xnew (bufsize
, char);
6942 while (getcwd (path
, bufsize
) == NULL
)
6944 if (errno
!= ERANGE
)
6948 path
= xnew (bufsize
, char);
6951 canonicalize_filename (path
);
6955 /* Return a newly allocated string containing a name of a temporary file. */
6959 const char *tmpdir
= getenv ("TMPDIR");
6960 const char *slash
= "/";
6962 #if MSDOS || defined (DOS_NT)
6964 tmpdir
= getenv ("TEMP");
6966 tmpdir
= getenv ("TMP");
6969 if (tmpdir
[strlen (tmpdir
) - 1] == '/'
6970 || tmpdir
[strlen (tmpdir
) - 1] == '\\')
6975 if (tmpdir
[strlen (tmpdir
) - 1] == '/')
6979 char *templt
= concat (tmpdir
, slash
, "etXXXXXX");
6980 int fd
= mkostemp (templt
, O_CLOEXEC
);
6981 if (fd
< 0 || close (fd
) != 0)
6983 int temp_errno
= errno
;
6989 #if defined (DOS_NT)
6990 /* The file name will be used in shell redirection, so it needs to have
6991 DOS-style backslashes, or else the Windows shell will barf. */
6993 for (p
= templt
; *p
; p
++)
7001 /* Return a newly allocated string containing the file name of FILE
7002 relative to the absolute directory DIR (which should end with a slash). */
7004 relative_filename (char *file
, char *dir
)
7006 char *fp
, *dp
, *afn
, *res
;
7009 /* Find the common root of file and dir (with a trailing slash). */
7010 afn
= absolute_filename (file
, cwd
);
7013 while (*fp
++ == *dp
++)
7015 fp
--, dp
--; /* back to the first differing char */
7017 if (fp
== afn
&& afn
[0] != '/') /* cannot build a relative name */
7020 do /* look at the equal chars until '/' */
7024 /* Build a sequence of "../" strings for the resulting relative file name. */
7026 while ((dp
= strchr (dp
+ 1, '/')) != NULL
)
7028 res
= xnew (3*i
+ strlen (fp
+ 1) + 1, char);
7031 z
= stpcpy (z
, "../");
7033 /* Add the file name relative to the common root of file and dir. */
7040 /* Return a newly allocated string containing the absolute file name
7041 of FILE given DIR (which should end with a slash). */
7043 absolute_filename (char *file
, char *dir
)
7045 char *slashp
, *cp
, *res
;
7047 if (filename_is_absolute (file
))
7048 res
= savestr (file
);
7050 /* We don't support non-absolute file names with a drive
7051 letter, like `d:NAME' (it's too much hassle). */
7052 else if (file
[1] == ':')
7053 fatal ("%s: relative file names with drive letters not supported", file
);
7056 res
= concat (dir
, file
, "");
7058 /* Delete the "/dirname/.." and "/." substrings. */
7059 slashp
= strchr (res
, '/');
7060 while (slashp
!= NULL
&& slashp
[0] != '\0')
7062 if (slashp
[1] == '.')
7064 if (slashp
[2] == '.'
7065 && (slashp
[3] == '/' || slashp
[3] == '\0'))
7070 while (cp
>= res
&& !filename_is_absolute (cp
));
7072 cp
= slashp
; /* the absolute name begins with "/.." */
7074 /* Under MSDOS and NT we get `d:/NAME' as absolute
7075 file name, so the luser could say `d:/../NAME'.
7076 We silently treat this as `d:/NAME'. */
7077 else if (cp
[0] != '/')
7080 memmove (cp
, slashp
+ 3, strlen (slashp
+ 2));
7084 else if (slashp
[2] == '/' || slashp
[2] == '\0')
7086 memmove (slashp
, slashp
+ 2, strlen (slashp
+ 1));
7091 slashp
= strchr (slashp
+ 1, '/');
7094 if (res
[0] == '\0') /* just a safety net: should never happen */
7097 return savestr ("/");
7103 /* Return a newly allocated string containing the absolute
7104 file name of dir where FILE resides given DIR (which should
7105 end with a slash). */
7107 absolute_dirname (char *file
, char *dir
)
7112 slashp
= strrchr (file
, '/');
7114 return savestr (dir
);
7117 res
= absolute_filename (file
, dir
);
7123 /* Whether the argument string is an absolute file name. The argument
7124 string must have been canonicalized with canonicalize_filename. */
7126 filename_is_absolute (char *fn
)
7128 return (fn
[0] == '/'
7130 || (c_isalpha (fn
[0]) && fn
[1] == ':' && fn
[2] == '/')
7135 /* Downcase DOS drive letter and collapse separators into single slashes.
7138 canonicalize_filename (register char *fn
)
7143 /* Canonicalize drive letter case. */
7144 if (c_isupper (fn
[0]) && fn
[1] == ':')
7145 fn
[0] = c_tolower (fn
[0]);
7147 /* Collapse multiple forward- and back-slashes into a single forward
7149 for (cp
= fn
; *cp
!= '\0'; cp
++, fn
++)
7150 if (*cp
== '/' || *cp
== '\\')
7153 while (cp
[1] == '/' || cp
[1] == '\\')
7161 /* Collapse multiple slashes into a single slash. */
7162 for (cp
= fn
; *cp
!= '\0'; cp
++, fn
++)
7166 while (cp
[1] == '/')
7172 #endif /* !DOS_NT */
7178 /* Initialize a linebuffer for use. */
7180 linebuffer_init (linebuffer
*lbp
)
7182 lbp
->size
= (DEBUG
) ? 3 : 200;
7183 lbp
->buffer
= xnew (lbp
->size
, char);
7184 lbp
->buffer
[0] = '\0';
7188 /* Set the minimum size of a string contained in a linebuffer. */
7190 linebuffer_setlen (linebuffer
*lbp
, int toksize
)
7192 while (lbp
->size
<= toksize
)
7195 xrnew (lbp
->buffer
, lbp
->size
, char);
7200 /* Like malloc but get fatal error if memory is exhausted. */
7202 xmalloc (size_t size
)
7204 void *result
= malloc (size
);
7206 fatal ("virtual memory exhausted");
7211 xrealloc (void *ptr
, size_t size
)
7213 void *result
= realloc (ptr
, size
);
7215 fatal ("virtual memory exhausted");
7221 * indent-tabs-mode: t
7224 * c-font-lock-extra-types: ("FILE" "bool" "language" "linebuffer" "fdesc" "node" "regexp")
7225 * c-file-style: "gnu"
7229 /* etags.c ends here */