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
);
2010 * recurse on left children, iterate on right children.
2013 free_tree (register node
*np
)
2017 register node
*node_right
= np
->right
;
2018 free_tree (np
->left
);
2028 * delete a file description
2031 free_fdesc (register fdesc
*fdp
)
2033 free (fdp
->infname
);
2034 free (fdp
->infabsname
);
2035 free (fdp
->infabsdir
);
2036 free (fdp
->taggedfname
);
2043 * Adds a node to the tree of nodes. In etags mode, sort by file
2044 * name. In ctags mode, sort by tag name. Make no attempt at
2047 * add_node is the only function allowed to add nodes, so it can
2051 add_node (node
*np
, node
**cur_node_p
)
2054 register node
*cur_node
= *cur_node_p
;
2056 if (cur_node
== NULL
)
2066 /* For each file name, tags are in a linked sublist on the right
2067 pointer. The first tags of different files are a linked list
2068 on the left pointer. last_node points to the end of the last
2070 if (last_node
!= NULL
&& last_node
->fdp
== np
->fdp
)
2072 /* Let's use the same sublist as the last added node. */
2073 assert (last_node
->right
== NULL
);
2074 last_node
->right
= np
;
2077 else if (cur_node
->fdp
== np
->fdp
)
2079 /* Scanning the list we found the head of a sublist which is
2080 good for us. Let's scan this sublist. */
2081 add_node (np
, &cur_node
->right
);
2084 /* The head of this sublist is not good for us. Let's try the
2086 add_node (np
, &cur_node
->left
);
2087 } /* if ETAGS mode */
2092 dif
= strcmp (np
->name
, cur_node
->name
);
2095 * If this tag name matches an existing one, then
2096 * do not add the node, but maybe print a warning.
2098 if (no_duplicates
&& !dif
)
2100 if (np
->fdp
== cur_node
->fdp
)
2104 fprintf (stderr
, "Duplicate entry in file %s, line %d: %s\n",
2105 np
->fdp
->infname
, lineno
, np
->name
);
2106 fprintf (stderr
, "Second entry ignored\n");
2109 else if (!cur_node
->been_warned
&& !no_warnings
)
2113 "Duplicate entry in files %s and %s: %s (Warning only)\n",
2114 np
->fdp
->infname
, cur_node
->fdp
->infname
, np
->name
);
2115 cur_node
->been_warned
= true;
2120 /* Actually add the node */
2121 add_node (np
, dif
< 0 ? &cur_node
->left
: &cur_node
->right
);
2122 } /* if CTAGS mode */
2126 * invalidate_nodes ()
2127 * Scan the node tree and invalidate all nodes pointing to the
2128 * given file description (CTAGS case) or free them (ETAGS case).
2131 invalidate_nodes (fdesc
*badfdp
, node
**npp
)
2140 if (np
->left
!= NULL
)
2141 invalidate_nodes (badfdp
, &np
->left
);
2142 if (np
->fdp
== badfdp
)
2144 if (np
->right
!= NULL
)
2145 invalidate_nodes (badfdp
, &np
->right
);
2149 assert (np
->fdp
!= NULL
);
2150 if (np
->fdp
== badfdp
)
2152 *npp
= np
->left
; /* detach the sublist from the list */
2153 np
->left
= NULL
; /* isolate it */
2154 free_tree (np
); /* free it */
2155 invalidate_nodes (badfdp
, npp
);
2158 invalidate_nodes (badfdp
, &np
->left
);
2163 static int total_size_of_entries (node
*);
2164 static int number_len (long) ATTRIBUTE_CONST
;
2166 /* Length of a non-negative number's decimal representation. */
2168 number_len (long int num
)
2171 while ((num
/= 10) > 0)
2177 * Return total number of characters that put_entries will output for
2178 * the nodes in the linked list at the right of the specified node.
2179 * This count is irrelevant with etags.el since emacs 19.34 at least,
2180 * but is still supplied for backward compatibility.
2183 total_size_of_entries (register node
*np
)
2185 register int total
= 0;
2187 for (; np
!= NULL
; np
= np
->right
)
2190 total
+= strlen (np
->regex
) + 1; /* pat\177 */
2191 if (np
->name
!= NULL
)
2192 total
+= strlen (np
->name
) + 1; /* name\001 */
2193 total
+= number_len ((long) np
->lno
) + 1; /* lno, */
2194 if (np
->cno
!= invalidcharno
) /* cno */
2195 total
+= number_len (np
->cno
);
2196 total
+= 1; /* newline */
2203 put_entries (register node
*np
)
2206 static fdesc
*fdp
= NULL
;
2211 /* Output subentries that precede this one */
2213 put_entries (np
->left
);
2215 /* Output this entry */
2224 fprintf (tagf
, "\f\n%s,%d\n",
2225 fdp
->taggedfname
, total_size_of_entries (np
));
2226 fdp
->written
= true;
2228 fputs (np
->regex
, tagf
);
2229 fputc ('\177', tagf
);
2230 if (np
->name
!= NULL
)
2232 fputs (np
->name
, tagf
);
2233 fputc ('\001', tagf
);
2235 fprintf (tagf
, "%d,", np
->lno
);
2236 if (np
->cno
!= invalidcharno
)
2237 fprintf (tagf
, "%ld", np
->cno
);
2243 if (np
->name
== NULL
)
2244 error ("internal error: NULL name in ctags mode.");
2249 fprintf (stdout
, "%s %s %d\n",
2250 np
->name
, np
->fdp
->taggedfname
, (np
->lno
+ 63) / 64);
2252 fprintf (stdout
, "%-16s %3d %-16s %s\n",
2253 np
->name
, np
->lno
, np
->fdp
->taggedfname
, np
->regex
);
2257 fprintf (tagf
, "%s\t%s\t", np
->name
, np
->fdp
->taggedfname
);
2260 { /* function or #define macro with args */
2261 putc (searchar
, tagf
);
2264 for (sp
= np
->regex
; *sp
; sp
++)
2266 if (*sp
== '\\' || *sp
== searchar
)
2270 putc (searchar
, tagf
);
2273 { /* anything else; text pattern inadequate */
2274 fprintf (tagf
, "%d", np
->lno
);
2279 } /* if this node contains a valid tag */
2281 /* Output subentries that follow this one */
2282 put_entries (np
->right
);
2284 put_entries (np
->left
);
2289 #define C_EXT 0x00fff /* C extensions */
2290 #define C_PLAIN 0x00000 /* C */
2291 #define C_PLPL 0x00001 /* C++ */
2292 #define C_STAR 0x00003 /* C* */
2293 #define C_JAVA 0x00005 /* JAVA */
2294 #define C_AUTO 0x01000 /* C, but switch to C++ if `class' is met */
2295 #define YACC 0x10000 /* yacc file */
2298 * The C symbol tables.
2303 st_C_objprot
, st_C_objimpl
, st_C_objend
,
2305 st_C_ignore
, st_C_attribute
,
2308 st_C_class
, st_C_template
,
2309 st_C_struct
, st_C_extern
, st_C_enum
, st_C_define
, st_C_typedef
2312 /* Feed stuff between (but not including) %[ and %] lines to:
2318 struct C_stab_entry { char *name; int c_ext; enum sym_type type; }
2322 while, 0, st_C_ignore
2323 switch, 0, st_C_ignore
2324 return, 0, st_C_ignore
2325 __attribute__, 0, st_C_attribute
2326 GTY, 0, st_C_attribute
2327 @interface, 0, st_C_objprot
2328 @protocol, 0, st_C_objprot
2329 @implementation,0, st_C_objimpl
2330 @end, 0, st_C_objend
2331 import, (C_JAVA & ~C_PLPL), st_C_ignore
2332 package, (C_JAVA & ~C_PLPL), st_C_ignore
2333 friend, C_PLPL, st_C_ignore
2334 extends, (C_JAVA & ~C_PLPL), st_C_javastruct
2335 implements, (C_JAVA & ~C_PLPL), st_C_javastruct
2336 interface, (C_JAVA & ~C_PLPL), st_C_struct
2337 class, 0, st_C_class
2338 namespace, C_PLPL, st_C_struct
2339 domain, C_STAR, st_C_struct
2340 union, 0, st_C_struct
2341 struct, 0, st_C_struct
2342 extern, 0, st_C_extern
2344 typedef, 0, st_C_typedef
2345 define, 0, st_C_define
2346 undef, 0, st_C_define
2347 operator, C_PLPL, st_C_operator
2348 template, 0, st_C_template
2349 # DEFUN used in emacs, the next three used in glibc (SYSCALL only for mach).
2350 DEFUN, 0, st_C_gnumacro
2351 SYSCALL, 0, st_C_gnumacro
2352 ENTRY, 0, st_C_gnumacro
2353 PSEUDO, 0, st_C_gnumacro
2354 # These are defined inside C functions, so currently they are not met.
2355 # EXFUN used in glibc, DEFVAR_* in emacs.
2356 #EXFUN, 0, st_C_gnumacro
2357 #DEFVAR_, 0, st_C_gnumacro
2359 and replace lines between %< and %> with its output, then:
2360 - remove the #if characterset check
2361 - make in_word_set static and not inline. */
2363 /* C code produced by gperf version 3.0.1 */
2364 /* Command-line: gperf -m 5 */
2365 /* Computed positions: -k'2-3' */
2367 struct C_stab_entry
{ const char *name
; int c_ext
; enum sym_type type
; };
2368 /* maximum key range = 33, duplicates = 0 */
2371 hash (const char *str
, int len
)
2373 static char const asso_values
[] =
2375 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2376 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2377 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2378 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2379 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2380 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2381 35, 35, 35, 35, 35, 35, 35, 35, 35, 3,
2382 26, 35, 35, 35, 35, 35, 35, 35, 27, 35,
2383 35, 35, 35, 24, 0, 35, 35, 35, 35, 0,
2384 35, 35, 35, 35, 35, 1, 35, 16, 35, 6,
2385 23, 0, 0, 35, 22, 0, 35, 35, 5, 0,
2386 0, 15, 1, 35, 6, 35, 8, 19, 35, 16,
2387 4, 5, 35, 35, 35, 35, 35, 35, 35, 35,
2388 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2389 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2390 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2391 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2392 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2393 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2394 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2395 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2396 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2397 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2398 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2399 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2400 35, 35, 35, 35, 35, 35
2407 hval
+= asso_values
[(unsigned char) str
[2]];
2410 hval
+= asso_values
[(unsigned char) str
[1]];
2416 static struct C_stab_entry
*
2417 in_word_set (register const char *str
, register unsigned int len
)
2421 TOTAL_KEYWORDS
= 33,
2422 MIN_WORD_LENGTH
= 2,
2423 MAX_WORD_LENGTH
= 15,
2428 static struct C_stab_entry wordlist
[] =
2431 {"if", 0, st_C_ignore
},
2432 {"GTY", 0, st_C_attribute
},
2433 {"@end", 0, st_C_objend
},
2434 {"union", 0, st_C_struct
},
2435 {"define", 0, st_C_define
},
2436 {"import", (C_JAVA
& ~C_PLPL
), st_C_ignore
},
2437 {"template", 0, st_C_template
},
2438 {"operator", C_PLPL
, st_C_operator
},
2439 {"@interface", 0, st_C_objprot
},
2440 {"implements", (C_JAVA
& ~C_PLPL
), st_C_javastruct
},
2441 {"friend", C_PLPL
, st_C_ignore
},
2442 {"typedef", 0, st_C_typedef
},
2443 {"return", 0, st_C_ignore
},
2444 {"@implementation",0, st_C_objimpl
},
2445 {"@protocol", 0, st_C_objprot
},
2446 {"interface", (C_JAVA
& ~C_PLPL
), st_C_struct
},
2447 {"extern", 0, st_C_extern
},
2448 {"extends", (C_JAVA
& ~C_PLPL
), st_C_javastruct
},
2449 {"struct", 0, st_C_struct
},
2450 {"domain", C_STAR
, st_C_struct
},
2451 {"switch", 0, st_C_ignore
},
2452 {"enum", 0, st_C_enum
},
2453 {"for", 0, st_C_ignore
},
2454 {"namespace", C_PLPL
, st_C_struct
},
2455 {"class", 0, st_C_class
},
2456 {"while", 0, st_C_ignore
},
2457 {"undef", 0, st_C_define
},
2458 {"package", (C_JAVA
& ~C_PLPL
), st_C_ignore
},
2459 {"__attribute__", 0, st_C_attribute
},
2460 {"SYSCALL", 0, st_C_gnumacro
},
2461 {"ENTRY", 0, st_C_gnumacro
},
2462 {"PSEUDO", 0, st_C_gnumacro
},
2463 {"DEFUN", 0, st_C_gnumacro
}
2466 if (len
<= MAX_WORD_LENGTH
&& len
>= MIN_WORD_LENGTH
)
2468 int key
= hash (str
, len
);
2470 if (key
<= MAX_HASH_VALUE
&& key
>= 0)
2472 const char *s
= wordlist
[key
].name
;
2474 if (*str
== *s
&& !strncmp (str
+ 1, s
+ 1, len
- 1) && s
[len
] == '\0')
2475 return &wordlist
[key
];
2482 static enum sym_type
2483 C_symtype (char *str
, int len
, int c_ext
)
2485 register struct C_stab_entry
*se
= in_word_set (str
, len
);
2487 if (se
== NULL
|| (se
->c_ext
&& !(c_ext
& se
->c_ext
)))
2494 * Ignoring __attribute__ ((list))
2496 static bool inattribute
; /* looking at an __attribute__ construct */
2499 * C functions and variables are recognized using a simple
2500 * finite automaton. fvdef is its state variable.
2504 fvnone
, /* nothing seen */
2505 fdefunkey
, /* Emacs DEFUN keyword seen */
2506 fdefunname
, /* Emacs DEFUN name seen */
2507 foperator
, /* func: operator keyword seen (cplpl) */
2508 fvnameseen
, /* function or variable name seen */
2509 fstartlist
, /* func: just after open parenthesis */
2510 finlist
, /* func: in parameter list */
2511 flistseen
, /* func: after parameter list */
2512 fignore
, /* func: before open brace */
2513 vignore
/* var-like: ignore until ';' */
2516 static bool fvextern
; /* func or var: extern keyword seen; */
2519 * typedefs are recognized using a simple finite automaton.
2520 * typdef is its state variable.
2524 tnone
, /* nothing seen */
2525 tkeyseen
, /* typedef keyword seen */
2526 ttypeseen
, /* defined type seen */
2527 tinbody
, /* inside typedef body */
2528 tend
, /* just before typedef tag */
2529 tignore
/* junk after typedef tag */
2533 * struct-like structures (enum, struct and union) are recognized
2534 * using another simple finite automaton. `structdef' is its state
2539 snone
, /* nothing seen yet,
2540 or in struct body if bracelev > 0 */
2541 skeyseen
, /* struct-like keyword seen */
2542 stagseen
, /* struct-like tag seen */
2543 scolonseen
/* colon seen after struct-like tag */
2547 * When objdef is different from onone, objtag is the name of the class.
2549 static const char *objtag
= "<uninited>";
2552 * Yet another little state machine to deal with preprocessor lines.
2556 dnone
, /* nothing seen */
2557 dsharpseen
, /* '#' seen as first char on line */
2558 ddefineseen
, /* '#' and 'define' seen */
2559 dignorerest
/* ignore rest of line */
2563 * State machine for Objective C protocols and implementations.
2564 * Idea by Tom R.Hageman <tom@basil.icce.rug.nl> (1995)
2568 onone
, /* nothing seen */
2569 oprotocol
, /* @interface or @protocol seen */
2570 oimplementation
, /* @implementations seen */
2571 otagseen
, /* class name seen */
2572 oparenseen
, /* parenthesis before category seen */
2573 ocatseen
, /* category name seen */
2574 oinbody
, /* in @implementation body */
2575 omethodsign
, /* in @implementation body, after +/- */
2576 omethodtag
, /* after method name */
2577 omethodcolon
, /* after method colon */
2578 omethodparm
, /* after method parameter */
2579 oignore
/* wait for @end */
2584 * Use this structure to keep info about the token read, and how it
2585 * should be tagged. Used by the make_C_tag function to build a tag.
2589 char *line
; /* string containing the token */
2590 int offset
; /* where the token starts in LINE */
2591 int length
; /* token length */
2593 The previous members can be used to pass strings around for generic
2594 purposes. The following ones specifically refer to creating tags. In this
2595 case the token contained here is the pattern that will be used to create a
2598 bool valid
; /* do not create a tag; the token should be
2599 invalidated whenever a state machine is
2600 reset prematurely */
2601 bool named
; /* create a named tag */
2602 int lineno
; /* source line number of tag */
2603 long linepos
; /* source char number of tag */
2604 } token
; /* latest token read */
2607 * Variables and functions for dealing with nested structures.
2608 * Idea by Mykola Dzyuba <mdzyuba@yahoo.com> (2001)
2610 static void pushclass_above (int, char *, int);
2611 static void popclass_above (int);
2612 static void write_classname (linebuffer
*, const char *qualifier
);
2615 char **cname
; /* nested class names */
2616 int *bracelev
; /* nested class brace level */
2617 int nl
; /* class nesting level (elements used) */
2618 int size
; /* length of the array */
2619 } cstack
; /* stack for nested declaration tags */
2620 /* Current struct nesting depth (namespace, class, struct, union, enum). */
2621 #define nestlev (cstack.nl)
2622 /* After struct keyword or in struct body, not inside a nested function. */
2623 #define instruct (structdef == snone && nestlev > 0 \
2624 && bracelev == cstack.bracelev[nestlev-1] + 1)
2627 pushclass_above (int bracelev
, char *str
, int len
)
2631 popclass_above (bracelev
);
2633 if (nl
>= cstack
.size
)
2635 int size
= cstack
.size
*= 2;
2636 xrnew (cstack
.cname
, size
, char *);
2637 xrnew (cstack
.bracelev
, size
, int);
2639 assert (nl
== 0 || cstack
.bracelev
[nl
-1] < bracelev
);
2640 cstack
.cname
[nl
] = (str
== NULL
) ? NULL
: savenstr (str
, len
);
2641 cstack
.bracelev
[nl
] = bracelev
;
2646 popclass_above (int bracelev
)
2650 for (nl
= cstack
.nl
- 1;
2651 nl
>= 0 && cstack
.bracelev
[nl
] >= bracelev
;
2654 free (cstack
.cname
[nl
]);
2660 write_classname (linebuffer
*cn
, const char *qualifier
)
2663 int qlen
= strlen (qualifier
);
2665 if (cstack
.nl
== 0 || cstack
.cname
[0] == NULL
)
2669 cn
->buffer
[0] = '\0';
2673 len
= strlen (cstack
.cname
[0]);
2674 linebuffer_setlen (cn
, len
);
2675 strcpy (cn
->buffer
, cstack
.cname
[0]);
2677 for (i
= 1; i
< cstack
.nl
; i
++)
2679 char *s
= cstack
.cname
[i
];
2682 linebuffer_setlen (cn
, len
+ qlen
+ strlen (s
));
2683 len
+= sprintf (cn
->buffer
+ len
, "%s%s", qualifier
, s
);
2688 static bool consider_token (char *, int, int, int *, int, int, bool *);
2689 static void make_C_tag (bool);
2693 * checks to see if the current token is at the start of a
2694 * function or variable, or corresponds to a typedef, or
2695 * is a struct/union/enum tag, or #define, or an enum constant.
2697 * *IS_FUNC_OR_VAR gets true if the token is a function or #define macro
2698 * with args. C_EXTP points to which language we are looking at.
2709 consider_token (char *str
, int len
, int c
, int *c_extp
,
2710 int bracelev
, int parlev
, bool *is_func_or_var
)
2711 /* IN: token pointer */
2712 /* IN: token length */
2713 /* IN: first char after the token */
2714 /* IN, OUT: C extensions mask */
2715 /* IN: brace level */
2716 /* IN: parenthesis level */
2717 /* OUT: function or variable found */
2719 /* When structdef is stagseen, scolonseen, or snone with bracelev > 0,
2720 structtype is the type of the preceding struct-like keyword, and
2721 structbracelev is the brace level where it has been seen. */
2722 static enum sym_type structtype
;
2723 static int structbracelev
;
2724 static enum sym_type toktype
;
2727 toktype
= C_symtype (str
, len
, *c_extp
);
2730 * Skip __attribute__
2732 if (toktype
== st_C_attribute
)
2739 * Advance the definedef state machine.
2744 /* We're not on a preprocessor line. */
2745 if (toktype
== st_C_gnumacro
)
2752 if (toktype
== st_C_define
)
2754 definedef
= ddefineseen
;
2758 definedef
= dignorerest
;
2763 * Make a tag for any macro, unless it is a constant
2764 * and constantypedefs is false.
2766 definedef
= dignorerest
;
2767 *is_func_or_var
= (c
== '(');
2768 if (!*is_func_or_var
&& !constantypedefs
)
2775 error ("internal error: definedef value.");
2784 if (toktype
== st_C_typedef
)
2807 if (structdef
== snone
&& fvdef
== fvnone
)
2829 case st_C_javastruct
:
2830 if (structdef
== stagseen
)
2831 structdef
= scolonseen
;
2835 if ((*c_extp
& C_AUTO
) /* automatic detection of C++ language */
2837 && definedef
== dnone
&& structdef
== snone
2838 && typdef
== tnone
&& fvdef
== fvnone
)
2839 *c_extp
= (*c_extp
| C_PLPL
) & ~C_AUTO
;
2840 if (toktype
== st_C_template
)
2847 && (typdef
== tkeyseen
2848 || (typedefs_or_cplusplus
&& structdef
== snone
)))
2850 structdef
= skeyseen
;
2851 structtype
= toktype
;
2852 structbracelev
= bracelev
;
2853 if (fvdef
== fvnameseen
)
2861 if (structdef
== skeyseen
)
2863 structdef
= stagseen
;
2867 if (typdef
!= tnone
)
2870 /* Detect Objective C constructs. */
2880 objdef
= oimplementation
;
2886 case oimplementation
:
2887 /* Save the class tag for functions or variables defined inside. */
2888 objtag
= savenstr (str
, len
);
2892 /* Save the class tag for categories. */
2893 objtag
= savenstr (str
, len
);
2895 *is_func_or_var
= true;
2899 *is_func_or_var
= true;
2907 objdef
= omethodtag
;
2908 linebuffer_setlen (&token_name
, len
);
2909 memcpy (token_name
.buffer
, str
, len
);
2910 token_name
.buffer
[len
] = '\0';
2916 objdef
= omethodparm
;
2921 objdef
= omethodtag
;
2924 int oldlen
= token_name
.len
;
2926 linebuffer_setlen (&token_name
, oldlen
+ len
);
2927 memcpy (token_name
.buffer
+ oldlen
, str
, len
);
2928 token_name
.buffer
[oldlen
+ len
] = '\0';
2934 if (toktype
== st_C_objend
)
2936 /* Memory leakage here: the string pointed by objtag is
2937 never released, because many tests would be needed to
2938 avoid breaking on incorrect input code. The amount of
2939 memory leaked here is the sum of the lengths of the
2949 /* A function, variable or enum constant? */
2971 *is_func_or_var
= true;
2975 && structdef
== snone
2976 && structtype
== st_C_enum
&& bracelev
> structbracelev
2977 /* Don't tag tokens in expressions that assign values to enum
2979 && fvdef
!= vignore
)
2980 return true; /* enum constant */
2986 fvdef
= fdefunname
; /* GNU macro */
2987 *is_func_or_var
= true;
2995 if ((strneq (str
, "asm", 3) && endtoken (str
[3]))
2996 || (strneq (str
, "__asm__", 7) && endtoken (str
[7])))
3007 if (len
>= 10 && strneq (str
+len
-10, "::operator", 10))
3009 if (*c_extp
& C_AUTO
) /* automatic detection of C++ */
3010 *c_extp
= (*c_extp
| C_PLPL
) & ~C_AUTO
;
3012 *is_func_or_var
= true;
3015 if (bracelev
> 0 && !instruct
)
3017 fvdef
= fvnameseen
; /* function or variable */
3018 *is_func_or_var
= true;
3033 * C_entries often keeps pointers to tokens or lines which are older than
3034 * the line currently read. By keeping two line buffers, and switching
3035 * them at end of line, it is possible to use those pointers.
3043 #define current_lb_is_new (newndx == curndx)
3044 #define switch_line_buffers() (curndx = 1 - curndx)
3046 #define curlb (lbs[curndx].lb)
3047 #define newlb (lbs[newndx].lb)
3048 #define curlinepos (lbs[curndx].linepos)
3049 #define newlinepos (lbs[newndx].linepos)
3051 #define plainc ((c_ext & C_EXT) == C_PLAIN)
3052 #define cplpl (c_ext & C_PLPL)
3053 #define cjava ((c_ext & C_JAVA) == C_JAVA)
3055 #define CNL_SAVE_DEFINEDEF() \
3057 curlinepos = charno; \
3058 readline (&curlb, inf); \
3059 lp = curlb.buffer; \
3066 CNL_SAVE_DEFINEDEF (); \
3067 if (savetoken.valid) \
3069 token = savetoken; \
3070 savetoken.valid = false; \
3072 definedef = dnone; \
3077 make_C_tag (bool isfun
)
3079 /* This function is never called when token.valid is false, but
3080 we must protect against invalid input or internal errors. */
3082 make_tag (token_name
.buffer
, token_name
.len
, isfun
, token
.line
,
3083 token
.offset
+token
.length
+1, token
.lineno
, token
.linepos
);
3085 { /* this branch is optimized away if !DEBUG */
3086 make_tag (concat ("INVALID TOKEN:-->", token_name
.buffer
, ""),
3087 token_name
.len
+ 17, isfun
, token
.line
,
3088 token
.offset
+token
.length
+1, token
.lineno
, token
.linepos
);
3089 error ("INVALID TOKEN");
3092 token
.valid
= false;
3096 perhaps_more_input (FILE *inf
)
3098 return !feof (inf
) && !ferror (inf
);
3104 * This routine finds functions, variables, typedefs,
3105 * #define's, enum constants and struct/union/enum definitions in
3106 * C syntax and adds them to the list.
3109 C_entries (int c_ext
, FILE *inf
)
3110 /* extension of C */
3113 register char c
; /* latest char read; '\0' for end of line */
3114 register char *lp
; /* pointer one beyond the character `c' */
3115 int curndx
, newndx
; /* indices for current and new lb */
3116 register int tokoff
; /* offset in line of start of current token */
3117 register int toklen
; /* length of current token */
3118 const char *qualifier
; /* string used to qualify names */
3119 int qlen
; /* length of qualifier */
3120 int bracelev
; /* current brace level */
3121 int bracketlev
; /* current bracket level */
3122 int parlev
; /* current parenthesis level */
3123 int attrparlev
; /* __attribute__ parenthesis level */
3124 int templatelev
; /* current template level */
3125 int typdefbracelev
; /* bracelev where a typedef struct body begun */
3126 bool incomm
, inquote
, inchar
, quotednl
, midtoken
;
3127 bool yacc_rules
; /* in the rules part of a yacc file */
3128 struct tok savetoken
= {0}; /* token saved during preprocessor handling */
3131 linebuffer_init (&lbs
[0].lb
);
3132 linebuffer_init (&lbs
[1].lb
);
3133 if (cstack
.size
== 0)
3135 cstack
.size
= (DEBUG
) ? 1 : 4;
3137 cstack
.cname
= xnew (cstack
.size
, char *);
3138 cstack
.bracelev
= xnew (cstack
.size
, int);
3141 tokoff
= toklen
= typdefbracelev
= 0; /* keep compiler quiet */
3142 curndx
= newndx
= 0;
3146 fvdef
= fvnone
; fvextern
= false; typdef
= tnone
;
3147 structdef
= snone
; definedef
= dnone
; objdef
= onone
;
3149 midtoken
= inquote
= inchar
= incomm
= quotednl
= false;
3150 token
.valid
= savetoken
.valid
= false;
3151 bracelev
= bracketlev
= parlev
= attrparlev
= templatelev
= 0;
3153 { qualifier
= "."; qlen
= 1; }
3155 { qualifier
= "::"; qlen
= 2; }
3158 while (perhaps_more_input (inf
))
3163 /* If we are at the end of the line, the next character is a
3164 '\0'; do not skip it, because it is what tells us
3165 to read the next line. */
3186 /* Newlines inside comments do not end macro definitions in
3188 CNL_SAVE_DEFINEDEF ();
3201 /* Newlines inside strings do not end macro definitions
3202 in traditional cpp, even though compilers don't
3203 usually accept them. */
3204 CNL_SAVE_DEFINEDEF ();
3214 /* Hmmm, something went wrong. */
3250 if (fvdef
!= finlist
&& fvdef
!= fignore
&& fvdef
!= vignore
)
3265 else if (/* cplpl && */ *lp
== '/')
3271 if ((c_ext
& YACC
) && *lp
== '%')
3273 /* Entering or exiting rules section in yacc file. */
3275 definedef
= dnone
; fvdef
= fvnone
; fvextern
= false;
3276 typdef
= tnone
; structdef
= snone
;
3277 midtoken
= inquote
= inchar
= incomm
= quotednl
= false;
3279 yacc_rules
= !yacc_rules
;
3285 if (definedef
== dnone
)
3288 bool cpptoken
= true;
3290 /* Look back on this line. If all blanks, or nonblanks
3291 followed by an end of comment, this is a preprocessor
3293 for (cp
= newlb
.buffer
; cp
< lp
-1; cp
++)
3294 if (!c_isspace (*cp
))
3296 if (*cp
== '*' && cp
[1] == '/')
3306 definedef
= dsharpseen
;
3307 /* This is needed for tagging enum values: when there are
3308 preprocessor conditionals inside the enum, we need to
3309 reset the value of fvdef so that the next enum value is
3310 tagged even though the one before it did not end in a
3312 if (fvdef
== vignore
&& instruct
&& parlev
== 0)
3314 if (strneq (cp
, "#if", 3) || strneq (cp
, "#el", 3))
3318 } /* if (definedef == dnone) */
3329 CNL_SAVE_DEFINEDEF ();
3336 /* Consider token only if some involved conditions are satisfied. */
3337 if (typdef
!= tignore
3338 && definedef
!= dignorerest
3341 && (definedef
!= dnone
3342 || structdef
!= scolonseen
)
3349 if (c
== ':' && *lp
== ':' && begtoken (lp
[1]))
3350 /* This handles :: in the middle,
3351 but not at the beginning of an identifier.
3352 Also, space-separated :: is not recognized. */
3354 if (c_ext
& C_AUTO
) /* automatic detection of C++ */
3355 c_ext
= (c_ext
| C_PLPL
) & ~C_AUTO
;
3359 goto still_in_token
;
3363 bool funorvar
= false;
3366 || consider_token (newlb
.buffer
+ tokoff
, toklen
, c
,
3367 &c_ext
, bracelev
, parlev
,
3370 if (fvdef
== foperator
)
3373 lp
= skip_spaces (lp
-1);
3377 && !c_isspace (*lp
) && *lp
!= '(')
3380 toklen
+= lp
- oldlp
;
3382 token
.named
= false;
3384 && nestlev
> 0 && definedef
== dnone
)
3385 /* in struct body */
3390 write_classname (&token_name
, qualifier
);
3391 len
= token_name
.len
;
3392 linebuffer_setlen (&token_name
,
3393 len
+ qlen
+ toklen
);
3394 sprintf (token_name
.buffer
+ len
, "%s%.*s",
3396 newlb
.buffer
+ tokoff
);
3400 linebuffer_setlen (&token_name
, toklen
);
3401 sprintf (token_name
.buffer
, "%.*s",
3402 toklen
, newlb
.buffer
+ tokoff
);
3406 else if (objdef
== ocatseen
)
3407 /* Objective C category */
3411 int len
= strlen (objtag
) + 2 + toklen
;
3412 linebuffer_setlen (&token_name
, len
);
3413 sprintf (token_name
.buffer
, "%s(%.*s)",
3415 newlb
.buffer
+ tokoff
);
3419 linebuffer_setlen (&token_name
, toklen
);
3420 sprintf (token_name
.buffer
, "%.*s",
3421 toklen
, newlb
.buffer
+ tokoff
);
3425 else if (objdef
== omethodtag
3426 || objdef
== omethodparm
)
3427 /* Objective C method */
3431 else if (fvdef
== fdefunname
)
3432 /* GNU DEFUN and similar macros */
3434 bool defun
= (newlb
.buffer
[tokoff
] == 'F');
3438 /* Rewrite the tag so that emacs lisp DEFUNs
3439 can be found by their elisp name */
3445 linebuffer_setlen (&token_name
, len
);
3446 memcpy (token_name
.buffer
,
3447 newlb
.buffer
+ off
, len
);
3448 token_name
.buffer
[len
] = '\0';
3451 if (token_name
.buffer
[len
] == '_')
3452 token_name
.buffer
[len
] = '-';
3453 token
.named
= defun
;
3457 linebuffer_setlen (&token_name
, toklen
);
3458 memcpy (token_name
.buffer
,
3459 newlb
.buffer
+ tokoff
, toklen
);
3460 token_name
.buffer
[toklen
] = '\0';
3461 /* Name macros and members. */
3462 token
.named
= (structdef
== stagseen
3463 || typdef
== ttypeseen
3466 && definedef
== dignorerest
)
3468 && definedef
== dnone
3469 && structdef
== snone
3472 token
.lineno
= lineno
;
3473 token
.offset
= tokoff
;
3474 token
.length
= toklen
;
3475 token
.line
= newlb
.buffer
;
3476 token
.linepos
= newlinepos
;
3479 if (definedef
== dnone
3480 && (fvdef
== fvnameseen
3481 || fvdef
== foperator
3482 || structdef
== stagseen
3484 || typdef
== ttypeseen
3485 || objdef
!= onone
))
3487 if (current_lb_is_new
)
3488 switch_line_buffers ();
3490 else if (definedef
!= dnone
3491 || fvdef
== fdefunname
3493 make_C_tag (funorvar
);
3495 else /* not yacc and consider_token failed */
3497 if (inattribute
&& fvdef
== fignore
)
3499 /* We have just met __attribute__ after a
3500 function parameter list: do not tag the
3507 } /* if (endtoken (c)) */
3508 else if (intoken (c
))
3514 } /* if (midtoken) */
3515 else if (begtoken (c
))
3523 /* This prevents tagging fb in
3524 void (__attribute__((noreturn)) *fb) (void);
3525 Fixing this is not easy and not very important. */
3529 if (plainc
|| declarations
)
3531 make_C_tag (true); /* a function */
3538 if (structdef
== stagseen
&& !cjava
)
3540 popclass_above (bracelev
);
3550 if (!yacc_rules
|| lp
== newlb
.buffer
+ 1)
3552 tokoff
= lp
- 1 - newlb
.buffer
;
3557 } /* if (begtoken) */
3558 } /* if must look at token */
3561 /* Detect end of line, colon, comma, semicolon and various braces
3562 after having handled a token.*/
3568 if (yacc_rules
&& token
.offset
== 0 && token
.valid
)
3570 make_C_tag (false); /* a yacc function */
3573 if (definedef
!= dnone
)
3579 make_C_tag (true); /* an Objective C class */
3583 objdef
= omethodcolon
;
3586 int toklen
= token_name
.len
;
3587 linebuffer_setlen (&token_name
, toklen
+ 1);
3588 strcpy (token_name
.buffer
+ toklen
, ":");
3594 if (structdef
== stagseen
)
3596 structdef
= scolonseen
;
3599 /* Should be useless, but may be work as a safety net. */
3600 if (cplpl
&& fvdef
== flistseen
)
3602 make_C_tag (true); /* a function */
3608 if (definedef
!= dnone
|| inattribute
)
3614 make_C_tag (false); /* a typedef */
3624 if (typdef
== tignore
|| cplpl
)
3628 if ((globals
&& bracelev
== 0 && (!fvextern
|| declarations
))
3629 || (members
&& instruct
))
3630 make_C_tag (false); /* a variable */
3633 token
.valid
= false;
3637 && (cplpl
|| !instruct
)
3638 && (typdef
== tnone
|| (typdef
!= tignore
&& instruct
)))
3640 && plainc
&& instruct
))
3641 make_C_tag (true); /* a function */
3647 && cplpl
&& structdef
== stagseen
)
3648 make_C_tag (false); /* forward declaration */
3650 token
.valid
= false;
3651 } /* switch (fvdef) */
3657 if (structdef
== stagseen
)
3661 if (definedef
!= dnone
|| inattribute
)
3667 make_C_tag (true); /* an Objective C method */
3682 if (instruct
&& parlev
== 0)
3693 && (!fvextern
|| declarations
))
3694 || (members
&& instruct
)))
3695 make_C_tag (false); /* a variable */
3698 if ((declarations
&& typdef
== tnone
&& !instruct
)
3699 || (members
&& typdef
!= tignore
&& instruct
))
3701 make_C_tag (true); /* a function */
3704 else if (!declarations
)
3706 token
.valid
= false;
3711 if (structdef
== stagseen
)
3715 if (definedef
!= dnone
|| inattribute
)
3717 if (structdef
== stagseen
)
3724 make_C_tag (false); /* a typedef */
3736 if ((members
&& bracelev
== 1)
3737 || (globals
&& bracelev
== 0
3738 && (!fvextern
|| declarations
)))
3739 make_C_tag (false); /* a variable */
3755 if (definedef
!= dnone
)
3757 if (objdef
== otagseen
&& parlev
== 0)
3758 objdef
= oparenseen
;
3762 if (typdef
== ttypeseen
3766 /* This handles constructs like:
3767 typedef void OperatorFun (int fun); */
3788 if (--attrparlev
== 0)
3789 inattribute
= false;
3792 if (definedef
!= dnone
)
3794 if (objdef
== ocatseen
&& parlev
== 1)
3796 make_C_tag (true); /* an Objective C category */
3812 || typdef
== ttypeseen
))
3815 make_C_tag (false); /* a typedef */
3818 else if (parlev
< 0) /* can happen due to ill-conceived #if's. */
3822 if (definedef
!= dnone
)
3824 if (typdef
== ttypeseen
)
3826 /* Whenever typdef is set to tinbody (currently only
3827 here), typdefbracelev should be set to bracelev. */
3829 typdefbracelev
= bracelev
;
3834 if (cplpl
&& !class_qualify
)
3836 /* Remove class and namespace qualifiers from the token,
3837 leaving only the method/member name. */
3838 char *cc
, *uqname
= token_name
.buffer
;
3839 char *tok_end
= token_name
.buffer
+ token_name
.len
;
3841 for (cc
= token_name
.buffer
; cc
< tok_end
; cc
++)
3843 if (*cc
== ':' && cc
[1] == ':')
3849 if (uqname
> token_name
.buffer
)
3851 int uqlen
= strlen (uqname
);
3852 linebuffer_setlen (&token_name
, uqlen
);
3853 memmove (token_name
.buffer
, uqname
, uqlen
+ 1);
3856 make_C_tag (true); /* a function */
3865 make_C_tag (true); /* an Objective C class */
3870 make_C_tag (true); /* an Objective C method */
3874 /* Neutralize `extern "C" {' grot. */
3875 if (bracelev
== 0 && structdef
== snone
&& nestlev
== 0
3885 case skeyseen
: /* unnamed struct */
3886 pushclass_above (bracelev
, NULL
, 0);
3889 case stagseen
: /* named struct or enum */
3890 case scolonseen
: /* a class */
3891 pushclass_above (bracelev
,token
.line
+token
.offset
, token
.length
);
3893 make_C_tag (false); /* a struct or enum */
3901 if (definedef
!= dnone
)
3903 if (fvdef
== fstartlist
)
3905 fvdef
= fvnone
; /* avoid tagging `foo' in `foo (*bar()) ()' */
3906 token
.valid
= false;
3910 if (definedef
!= dnone
)
3913 if (!ignoreindent
&& lp
== newlb
.buffer
+ 1)
3916 token
.valid
= false; /* unexpected value, token unreliable */
3917 bracelev
= 0; /* reset brace level if first column */
3918 parlev
= 0; /* also reset paren level, just in case... */
3920 else if (bracelev
< 0)
3922 token
.valid
= false; /* something gone amiss, token unreliable */
3925 if (bracelev
== 0 && fvdef
== vignore
)
3926 fvdef
= fvnone
; /* end of function */
3927 popclass_above (bracelev
);
3929 /* Only if typdef == tinbody is typdefbracelev significant. */
3930 if (typdef
== tinbody
&& bracelev
<= typdefbracelev
)
3932 assert (bracelev
== typdefbracelev
);
3937 if (definedef
!= dnone
)
3947 if ((members
&& bracelev
== 1)
3948 || (globals
&& bracelev
== 0 && (!fvextern
|| declarations
)))
3949 make_C_tag (false); /* a variable */
3957 && (structdef
== stagseen
|| fvdef
== fvnameseen
))
3964 if (templatelev
> 0)
3972 if (objdef
== oinbody
&& bracelev
== 0)
3974 objdef
= omethodsign
;
3979 case '#': case '~': case '&': case '%': case '/':
3980 case '|': case '^': case '!': case '.': case '?':
3981 if (definedef
!= dnone
)
3983 /* These surely cannot follow a function tag in C. */
3996 if (objdef
== otagseen
)
3998 make_C_tag (true); /* an Objective C class */
4001 /* If a macro spans multiple lines don't reset its state. */
4003 CNL_SAVE_DEFINEDEF ();
4009 } /* while not eof */
4011 free (lbs
[0].lb
.buffer
);
4012 free (lbs
[1].lb
.buffer
);
4016 * Process either a C++ file or a C file depending on the setting
4020 default_C_entries (FILE *inf
)
4022 C_entries (cplusplus
? C_PLPL
: C_AUTO
, inf
);
4025 /* Always do plain C. */
4027 plain_C_entries (FILE *inf
)
4032 /* Always do C++. */
4034 Cplusplus_entries (FILE *inf
)
4036 C_entries (C_PLPL
, inf
);
4039 /* Always do Java. */
4041 Cjava_entries (FILE *inf
)
4043 C_entries (C_JAVA
, inf
);
4048 Cstar_entries (FILE *inf
)
4050 C_entries (C_STAR
, inf
);
4053 /* Always do Yacc. */
4055 Yacc_entries (FILE *inf
)
4057 C_entries (YACC
, inf
);
4061 /* Useful macros. */
4062 #define LOOP_ON_INPUT_LINES(file_pointer, line_buffer, char_pointer) \
4063 while (perhaps_more_input (file_pointer) \
4064 && (readline (&(line_buffer), file_pointer), \
4065 (char_pointer) = (line_buffer).buffer, \
4068 #define LOOKING_AT(cp, kw) /* kw is the keyword, a literal string */ \
4069 ((assert ("" kw), true) /* syntax error if not a literal string */ \
4070 && strneq ((cp), kw, sizeof (kw)-1) /* cp points at kw */ \
4071 && notinname ((cp)[sizeof (kw)-1]) /* end of kw */ \
4072 && ((cp) = skip_spaces ((cp) + sizeof (kw) - 1), true)) /* skip spaces */
4074 /* Similar to LOOKING_AT but does not use notinname, does not skip */
4075 #define LOOKING_AT_NOCASE(cp, kw) /* the keyword is a literal string */ \
4076 ((assert ("" kw), true) /* syntax error if not a literal string */ \
4077 && strncaseeq ((cp), kw, sizeof (kw)-1) /* cp points at kw */ \
4078 && ((cp) += sizeof (kw) - 1, true)) /* skip spaces */
4081 * Read a file, but do no processing. This is used to do regexp
4082 * matching on files that have no language defined.
4085 just_read_file (FILE *inf
)
4087 while (perhaps_more_input (inf
))
4088 readline (&lb
, inf
);
4092 /* Fortran parsing */
4094 static void F_takeprec (void);
4095 static void F_getit (FILE *);
4100 dbp
= skip_spaces (dbp
);
4104 dbp
= skip_spaces (dbp
);
4105 if (strneq (dbp
, "(*)", 3))
4110 if (!c_isdigit (*dbp
))
4112 --dbp
; /* force failure */
4117 while (c_isdigit (*dbp
));
4125 dbp
= skip_spaces (dbp
);
4128 readline (&lb
, inf
);
4133 dbp
= skip_spaces (dbp
);
4135 if (!c_isalpha (*dbp
) && *dbp
!= '_' && *dbp
!= '$')
4137 for (cp
= dbp
+ 1; *cp
!= '\0' && intoken (*cp
); cp
++)
4139 make_tag (dbp
, cp
-dbp
, true,
4140 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4145 Fortran_functions (FILE *inf
)
4147 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
4150 dbp
++; /* Ratfor escape to fortran */
4151 dbp
= skip_spaces (dbp
);
4155 if (LOOKING_AT_NOCASE (dbp
, "recursive"))
4156 dbp
= skip_spaces (dbp
);
4158 if (LOOKING_AT_NOCASE (dbp
, "pure"))
4159 dbp
= skip_spaces (dbp
);
4161 if (LOOKING_AT_NOCASE (dbp
, "elemental"))
4162 dbp
= skip_spaces (dbp
);
4164 switch (c_tolower (*dbp
))
4167 if (nocase_tail ("integer"))
4171 if (nocase_tail ("real"))
4175 if (nocase_tail ("logical"))
4179 if (nocase_tail ("complex") || nocase_tail ("character"))
4183 if (nocase_tail ("double"))
4185 dbp
= skip_spaces (dbp
);
4188 if (nocase_tail ("precision"))
4194 dbp
= skip_spaces (dbp
);
4197 switch (c_tolower (*dbp
))
4200 if (nocase_tail ("function"))
4204 if (nocase_tail ("subroutine"))
4208 if (nocase_tail ("entry"))
4212 if (nocase_tail ("blockdata") || nocase_tail ("block data"))
4214 dbp
= skip_spaces (dbp
);
4215 if (*dbp
== '\0') /* assume un-named */
4216 make_tag ("blockdata", 9, true,
4217 lb
.buffer
, dbp
- lb
.buffer
, lineno
, linecharno
);
4219 F_getit (inf
); /* look for name */
4228 * Go language support
4229 * Original code by Xi Lu <lx@shellcodes.org> (2016)
4232 Go_functions(FILE *inf
)
4236 LOOP_ON_INPUT_LINES(inf
, lb
, cp
)
4238 cp
= skip_spaces (cp
);
4240 if (LOOKING_AT (cp
, "package"))
4243 while (!notinname (*cp
) && *cp
!= '\0')
4245 make_tag (name
, cp
- name
, false, lb
.buffer
,
4246 cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4248 else if (LOOKING_AT (cp
, "func"))
4250 /* Go implementation of interface, such as:
4251 func (n *Integer) Add(m Integer) ...
4252 skip `(n *Integer)` part.
4258 cp
= skip_spaces (cp
+1);
4265 while (!notinname (*cp
))
4268 make_tag (name
, cp
- name
, true, lb
.buffer
,
4269 cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4272 else if (members
&& LOOKING_AT (cp
, "type"))
4276 /* Ignore the likes of the following:
4284 while (!notinname (*cp
) && *cp
!= '\0')
4287 make_tag (name
, cp
- name
, false, lb
.buffer
,
4288 cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4297 * Philippe Waroquiers (1998)
4300 /* Once we are positioned after an "interesting" keyword, let's get
4301 the real tag value necessary. */
4303 Ada_getit (FILE *inf
, const char *name_qualifier
)
4309 while (perhaps_more_input (inf
))
4311 dbp
= skip_spaces (dbp
);
4313 || (dbp
[0] == '-' && dbp
[1] == '-'))
4315 readline (&lb
, inf
);
4318 switch (c_tolower (*dbp
))
4321 if (nocase_tail ("body"))
4323 /* Skipping body of procedure body or package body or ....
4324 resetting qualifier to body instead of spec. */
4325 name_qualifier
= "/b";
4330 /* Skipping type of task type or protected type ... */
4331 if (nocase_tail ("type"))
4338 for (cp
= dbp
; *cp
!= '\0' && *cp
!= '"'; cp
++)
4343 dbp
= skip_spaces (dbp
);
4345 c_isalnum (*cp
) || *cp
== '_' || *cp
== '.';
4353 name
= concat (dbp
, name_qualifier
, "");
4355 make_tag (name
, strlen (name
), true,
4356 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4365 Ada_funcs (FILE *inf
)
4367 bool inquote
= false;
4368 bool skip_till_semicolumn
= false;
4370 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
4372 while (*dbp
!= '\0')
4374 /* Skip a string i.e. "abcd". */
4375 if (inquote
|| (*dbp
== '"'))
4377 dbp
= strchr (dbp
+ !inquote
, '"');
4382 continue; /* advance char */
4387 break; /* advance line */
4391 /* Skip comments. */
4392 if (dbp
[0] == '-' && dbp
[1] == '-')
4393 break; /* advance line */
4395 /* Skip character enclosed in single quote i.e. 'a'
4396 and skip single quote starting an attribute i.e. 'Image. */
4405 if (skip_till_semicolumn
)
4408 skip_till_semicolumn
= false;
4410 continue; /* advance char */
4413 /* Search for beginning of a token. */
4414 if (!begtoken (*dbp
))
4417 continue; /* advance char */
4420 /* We are at the beginning of a token. */
4421 switch (c_tolower (*dbp
))
4424 if (!packages_only
&& nocase_tail ("function"))
4425 Ada_getit (inf
, "/f");
4427 break; /* from switch */
4428 continue; /* advance char */
4430 if (!packages_only
&& nocase_tail ("procedure"))
4431 Ada_getit (inf
, "/p");
4432 else if (nocase_tail ("package"))
4433 Ada_getit (inf
, "/s");
4434 else if (nocase_tail ("protected")) /* protected type */
4435 Ada_getit (inf
, "/t");
4437 break; /* from switch */
4438 continue; /* advance char */
4441 if (typedefs
&& !packages_only
&& nocase_tail ("use"))
4443 /* when tagging types, avoid tagging use type Pack.Typename;
4444 for this, we will skip everything till a ; */
4445 skip_till_semicolumn
= true;
4446 continue; /* advance char */
4450 if (!packages_only
&& nocase_tail ("task"))
4451 Ada_getit (inf
, "/k");
4452 else if (typedefs
&& !packages_only
&& nocase_tail ("type"))
4454 Ada_getit (inf
, "/t");
4455 while (*dbp
!= '\0')
4459 break; /* from switch */
4460 continue; /* advance char */
4463 /* Look for the end of the token. */
4464 while (!endtoken (*dbp
))
4467 } /* advance char */
4468 } /* advance line */
4473 * Unix and microcontroller assembly tag handling
4474 * Labels: /^[a-zA-Z_.$][a-zA_Z0-9_.$]*[: ^I^J]/
4475 * Idea by Bob Weiner, Motorola Inc. (1994)
4478 Asm_labels (FILE *inf
)
4482 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4484 /* If first char is alphabetic or one of [_.$], test for colon
4485 following identifier. */
4486 if (c_isalpha (*cp
) || *cp
== '_' || *cp
== '.' || *cp
== '$')
4488 /* Read past label. */
4490 while (c_isalnum (*cp
) || *cp
== '_' || *cp
== '.' || *cp
== '$')
4492 if (*cp
== ':' || c_isspace (*cp
))
4493 /* Found end of label, so copy it and add it to the table. */
4494 make_tag (lb
.buffer
, cp
- lb
.buffer
, true,
4495 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4503 * Perl sub names: /^sub[ \t\n]+[^ \t\n{]+/
4504 * /^use constant[ \t\n]+[^ \t\n{=,;]+/
4505 * Perl variable names: /^(my|local).../
4506 * Original code by Bart Robinson <lomew@cs.utah.edu> (1995)
4507 * Additions by Michael Ernst <mernst@alum.mit.edu> (1997)
4508 * Ideas by Kai Großjohann <Kai.Grossjohann@CS.Uni-Dortmund.DE> (2001)
4511 Perl_functions (FILE *inf
)
4513 char *package
= savestr ("main"); /* current package name */
4516 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4518 cp
= skip_spaces (cp
);
4520 if (LOOKING_AT (cp
, "package"))
4523 get_tag (cp
, &package
);
4525 else if (LOOKING_AT (cp
, "sub"))
4531 while (!notinname (*cp
))
4534 continue; /* nothing found */
4535 pos
= strchr (sp
, ':');
4536 if (pos
&& pos
< cp
&& pos
[1] == ':')
4538 /* The name is already qualified. */
4541 char *q
= pos
+ 2, *qpos
;
4542 while ((qpos
= strchr (q
, ':')) != NULL
4548 make_tag (sp
, cp
- sp
, true,
4549 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4551 else if (class_qualify
)
4554 char savechar
, *name
;
4558 name
= concat (package
, "::", sp
);
4560 make_tag (name
, strlen (name
), true,
4561 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4565 make_tag (sp
, cp
- sp
, true,
4566 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4568 else if (LOOKING_AT (cp
, "use constant")
4569 || LOOKING_AT (cp
, "use constant::defer"))
4571 /* For hash style multi-constant like
4572 use constant { FOO => 123,
4574 only the first FOO is picked up. Parsing across the value
4575 expressions would be difficult in general, due to possible nested
4576 hashes, here-documents, etc. */
4578 cp
= skip_spaces (cp
+1);
4581 else if (globals
) /* only if we are tagging global vars */
4583 /* Skip a qualifier, if any. */
4584 bool qual
= LOOKING_AT (cp
, "my") || LOOKING_AT (cp
, "local");
4585 /* After "my" or "local", but before any following paren or space. */
4586 char *varstart
= cp
;
4588 if (qual
/* should this be removed? If yes, how? */
4589 && (*cp
== '$' || *cp
== '@' || *cp
== '%'))
4594 while (c_isalnum (*cp
) || *cp
== '_');
4598 /* Should be examining a variable list at this point;
4599 could insist on seeing an open parenthesis. */
4600 while (*cp
!= '\0' && *cp
!= ';' && *cp
!= '=' && *cp
!= ')')
4606 make_tag (varstart
, cp
- varstart
, false,
4607 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4616 * Look for /^[\t]*def[ \t\n]+[^ \t\n(:]+/ or /^class[ \t\n]+[^ \t\n(:]+/
4617 * Idea by Eric S. Raymond <esr@thyrsus.com> (1997)
4618 * More ideas by seb bacon <seb@jamkit.com> (2002)
4621 Python_functions (FILE *inf
)
4625 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4627 cp
= skip_spaces (cp
);
4628 if (LOOKING_AT (cp
, "def") || LOOKING_AT (cp
, "class"))
4631 while (!notinname (*cp
) && *cp
!= ':')
4633 make_tag (name
, cp
- name
, true,
4634 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4641 * Original code by Xi Lu <lx@shellcodes.org> (2015)
4644 Ruby_functions (FILE *inf
)
4647 bool reader
= false, writer
= false, alias
= false, continuation
= false;
4649 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4651 bool is_class
= false;
4652 bool is_method
= false;
4655 cp
= skip_spaces (cp
);
4658 && c_isalpha (*cp
) && c_isupper (*cp
))
4660 char *bp
, *colon
= NULL
;
4664 for (cp
++; c_isalnum (*cp
) || *cp
== '_' || *cp
== ':'; cp
++)
4671 bp
= skip_spaces (cp
);
4672 if (*bp
== '=' && !(bp
[1] == '=' || bp
[1] == '>'))
4674 if (colon
&& !c_isspace (colon
[1]))
4676 make_tag (name
, cp
- name
, false,
4677 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4681 else if (!continuation
4682 /* Modules, classes, methods. */
4683 && ((is_method
= LOOKING_AT (cp
, "def"))
4684 || (is_class
= LOOKING_AT (cp
, "class"))
4685 || LOOKING_AT (cp
, "module")))
4687 const char self_name
[] = "self.";
4688 const size_t self_size1
= sizeof (self_name
) - 1;
4692 /* Ruby method names can end in a '='. Also, operator overloading can
4693 define operators whose names include '='. */
4694 while (!notinname (*cp
) || *cp
== '=')
4697 /* Remove "self." from the method name. */
4698 if (cp
- name
> self_size1
4699 && strneq (name
, self_name
, self_size1
))
4702 /* Remove the class/module qualifiers from method names. */
4707 for (q
= name
; q
< cp
&& *q
!= '.'; q
++)
4709 if (q
< cp
- 1) /* punt if we see just "FOO." */
4713 /* Don't tag singleton classes. */
4714 if (is_class
&& strneq (name
, "<<", 2) && cp
== name
+ 2)
4717 make_tag (name
, cp
- name
, true,
4718 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4722 /* Tag accessors and aliases. */
4725 reader
= writer
= alias
= false;
4727 while (*cp
&& *cp
!= '#')
4731 reader
= writer
= alias
= false;
4732 if (LOOKING_AT (cp
, "attr_reader"))
4734 else if (LOOKING_AT (cp
, "attr_writer"))
4736 else if (LOOKING_AT (cp
, "attr_accessor"))
4741 else if (LOOKING_AT (cp
, "alias_method"))
4744 if (reader
|| writer
|| alias
)
4749 cp
= skip_spaces (cp
);
4751 cp
= skip_spaces (cp
+ 1);
4753 cp
= skip_name (cp
);
4759 make_tag (np
, cp
- np
, true,
4760 lb
.buffer
, cp
- lb
.buffer
+ 1,
4761 lineno
, linecharno
);
4762 continuation
= false;
4766 size_t name_len
= cp
- np
+ 1;
4767 char *wr_name
= xnew (name_len
+ 1, char);
4769 memcpy (wr_name
, np
, name_len
- 1);
4770 memcpy (wr_name
+ name_len
- 1, "=", 2);
4771 pfnote (wr_name
, true, lb
.buffer
, cp
- lb
.buffer
+ 1,
4772 lineno
, linecharno
);
4773 continuation
= false;
4778 make_tag (np
, cp
- np
, true,
4779 lb
.buffer
, cp
- lb
.buffer
+ 1,
4780 lineno
, linecharno
);
4781 continuation
= false;
4782 while (*cp
&& *cp
!= '#' && *cp
!= ';')
4785 continuation
= true;
4786 else if (!c_isspace (*cp
))
4787 continuation
= false;
4791 continuation
= false;
4793 cp
= skip_spaces (cp
);
4796 : (continuation
= (*cp
== ',')))
4797 && (cp
= skip_spaces (cp
+ 1), *cp
&& *cp
!= '#'));
4800 cp
= skip_name (cp
);
4801 while (*cp
&& *cp
!= '#' && notinname (*cp
))
4812 * - /^[ \t]*function[ \t\n]+[^ \t\n(]+/
4813 * - /^[ \t]*class[ \t\n]+[^ \t\n]+/
4814 * - /^[ \t]*define\(\"[^\"]+/
4815 * Only with --members:
4816 * - /^[ \t]*var[ \t\n]+\$[^ \t\n=;]/
4817 * Idea by Diez B. Roggisch (2001)
4820 PHP_functions (FILE *inf
)
4823 bool search_identifier
= false;
4825 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4827 cp
= skip_spaces (cp
);
4829 if (search_identifier
4832 while (!notinname (*cp
))
4834 make_tag (name
, cp
- name
, true,
4835 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4836 search_identifier
= false;
4838 else if (LOOKING_AT (cp
, "function"))
4841 cp
= skip_spaces (cp
+1);
4845 while (!notinname (*cp
))
4847 make_tag (name
, cp
- name
, true,
4848 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4851 search_identifier
= true;
4853 else if (LOOKING_AT (cp
, "class"))
4858 while (*cp
!= '\0' && !c_isspace (*cp
))
4860 make_tag (name
, cp
- name
, false,
4861 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4864 search_identifier
= true;
4866 else if (strneq (cp
, "define", 6)
4867 && (cp
= skip_spaces (cp
+6))
4869 && (*cp
== '"' || *cp
== '\''))
4873 while (*cp
!= quote
&& *cp
!= '\0')
4875 make_tag (name
, cp
- name
, false,
4876 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4879 && LOOKING_AT (cp
, "var")
4883 while (!notinname (*cp
))
4885 make_tag (name
, cp
- name
, false,
4886 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4893 * Cobol tag functions
4894 * We could look for anything that could be a paragraph name.
4895 * i.e. anything that starts in column 8 is one word and ends in a full stop.
4896 * Idea by Corny de Souza (1993)
4899 Cobol_paragraphs (FILE *inf
)
4901 register char *bp
, *ep
;
4903 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4909 /* If eoln, compiler option or comment ignore whole line. */
4910 if (bp
[-1] != ' ' || !c_isalnum (bp
[0]))
4913 for (ep
= bp
; c_isalnum (*ep
) || *ep
== '-'; ep
++)
4916 make_tag (bp
, ep
- bp
, true,
4917 lb
.buffer
, ep
- lb
.buffer
+ 1, lineno
, linecharno
);
4924 * Ideas by Assar Westerlund <assar@sics.se> (2001)
4927 Makefile_targets (FILE *inf
)
4931 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4933 if (*bp
== '\t' || *bp
== '#')
4935 while (*bp
!= '\0' && *bp
!= '=' && *bp
!= ':')
4937 if (*bp
== ':' || (globals
&& *bp
== '='))
4939 /* We should detect if there is more than one tag, but we do not.
4940 We just skip initial and final spaces. */
4941 char * namestart
= skip_spaces (lb
.buffer
);
4942 while (--bp
> namestart
)
4943 if (!notinname (*bp
))
4945 make_tag (namestart
, bp
- namestart
+ 1, true,
4946 lb
.buffer
, bp
- lb
.buffer
+ 2, lineno
, linecharno
);
4954 * Original code by Mosur K. Mohan (1989)
4956 * Locates tags for procedures & functions. Doesn't do any type- or
4957 * var-definitions. It does look for the keyword "extern" or
4958 * "forward" immediately following the procedure statement; if found,
4959 * the tag is skipped.
4962 Pascal_functions (FILE *inf
)
4964 linebuffer tline
; /* mostly copied from C_entries */
4966 int save_lineno
, namelen
, taglen
;
4969 bool /* each of these flags is true if: */
4970 incomment
, /* point is inside a comment */
4971 inquote
, /* point is inside '..' string */
4972 get_tagname
, /* point is after PROCEDURE/FUNCTION
4973 keyword, so next item = potential tag */
4974 found_tag
, /* point is after a potential tag */
4975 inparms
, /* point is within parameter-list */
4976 verify_tag
; /* point has passed the parm-list, so the
4977 next token will determine whether this
4978 is a FORWARD/EXTERN to be ignored, or
4979 whether it is a real tag */
4981 save_lcno
= save_lineno
= namelen
= taglen
= 0; /* keep compiler quiet */
4982 name
= NULL
; /* keep compiler quiet */
4985 linebuffer_init (&tline
);
4987 incomment
= inquote
= false;
4988 found_tag
= false; /* have a proc name; check if extern */
4989 get_tagname
= false; /* found "procedure" keyword */
4990 inparms
= false; /* found '(' after "proc" */
4991 verify_tag
= false; /* check if "extern" is ahead */
4994 while (perhaps_more_input (inf
)) /* long main loop to get next char */
4997 if (c
== '\0') /* if end of line */
4999 readline (&lb
, inf
);
5003 if (!((found_tag
&& verify_tag
)
5005 c
= *dbp
++; /* only if don't need *dbp pointing
5006 to the beginning of the name of
5007 the procedure or function */
5011 if (c
== '}') /* within { } comments */
5013 else if (c
== '*' && *dbp
== ')') /* within (* *) comments */
5030 inquote
= true; /* found first quote */
5032 case '{': /* found open { comment */
5036 if (*dbp
== '*') /* found open (* comment */
5041 else if (found_tag
) /* found '(' after tag, i.e., parm-list */
5044 case ')': /* end of parms list */
5049 if (found_tag
&& !inparms
) /* end of proc or fn stmt */
5056 if (found_tag
&& verify_tag
&& (*dbp
!= ' '))
5058 /* Check if this is an "extern" declaration. */
5061 if (c_tolower (*dbp
) == 'e')
5063 if (nocase_tail ("extern")) /* superfluous, really! */
5069 else if (c_tolower (*dbp
) == 'f')
5071 if (nocase_tail ("forward")) /* check for forward reference */
5077 if (found_tag
&& verify_tag
) /* not external proc, so make tag */
5081 make_tag (name
, namelen
, true,
5082 tline
.buffer
, taglen
, save_lineno
, save_lcno
);
5086 if (get_tagname
) /* grab name of proc or fn */
5093 /* Find block name. */
5094 for (cp
= dbp
+ 1; *cp
!= '\0' && !endtoken (*cp
); cp
++)
5097 /* Save all values for later tagging. */
5098 linebuffer_setlen (&tline
, lb
.len
);
5099 strcpy (tline
.buffer
, lb
.buffer
);
5100 save_lineno
= lineno
;
5101 save_lcno
= linecharno
;
5102 name
= tline
.buffer
+ (dbp
- lb
.buffer
);
5104 taglen
= cp
- lb
.buffer
+ 1;
5106 dbp
= cp
; /* set dbp to e-o-token */
5107 get_tagname
= false;
5111 /* And proceed to check for "extern". */
5113 else if (!incomment
&& !inquote
&& !found_tag
)
5115 /* Check for proc/fn keywords. */
5116 switch (c_tolower (c
))
5119 if (nocase_tail ("rocedure")) /* c = 'p', dbp has advanced */
5123 if (nocase_tail ("unction"))
5128 } /* while not eof */
5130 free (tline
.buffer
);
5135 * Lisp tag functions
5136 * look for (def or (DEF, quote or QUOTE
5139 static void L_getit (void);
5144 if (*dbp
== '\'') /* Skip prefix quote */
5146 else if (*dbp
== '(')
5149 /* Try to skip "(quote " */
5150 if (!LOOKING_AT (dbp
, "quote") && !LOOKING_AT (dbp
, "QUOTE"))
5151 /* Ok, then skip "(" before name in (defstruct (foo)) */
5152 dbp
= skip_spaces (dbp
);
5154 get_tag (dbp
, NULL
);
5158 Lisp_functions (FILE *inf
)
5160 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
5165 /* "(defvar foo)" is a declaration rather than a definition. */
5169 if (LOOKING_AT (p
, "defvar"))
5171 p
= skip_name (p
); /* past var name */
5172 p
= skip_spaces (p
);
5178 if (strneq (dbp
+ 1, "cl-", 3) || strneq (dbp
+ 1, "CL-", 3))
5181 if (strneq (dbp
+1, "def", 3) || strneq (dbp
+1, "DEF", 3))
5183 dbp
= skip_non_spaces (dbp
);
5184 dbp
= skip_spaces (dbp
);
5189 /* Check for (foo::defmumble name-defined ... */
5192 while (!notinname (*dbp
) && *dbp
!= ':');
5197 while (*dbp
== ':');
5199 if (strneq (dbp
, "def", 3) || strneq (dbp
, "DEF", 3))
5201 dbp
= skip_non_spaces (dbp
);
5202 dbp
= skip_spaces (dbp
);
5212 * Lua script language parsing
5213 * Original code by David A. Capello <dacap@users.sourceforge.net> (2004)
5215 * "function" and "local function" are tags if they start at column 1.
5218 Lua_functions (FILE *inf
)
5222 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
5224 bp
= skip_spaces (bp
);
5225 if (bp
[0] != 'f' && bp
[0] != 'l')
5228 (void)LOOKING_AT (bp
, "local"); /* skip possible "local" */
5230 if (LOOKING_AT (bp
, "function"))
5232 char *tag_name
, *tp_dot
, *tp_colon
;
5234 get_tag (bp
, &tag_name
);
5235 /* If the tag ends with ".foo" or ":foo", make an additional tag for
5237 tp_dot
= strrchr (tag_name
, '.');
5238 tp_colon
= strrchr (tag_name
, ':');
5239 if (tp_dot
|| tp_colon
)
5241 char *p
= tp_dot
> tp_colon
? tp_dot
: tp_colon
;
5242 int len_add
= p
- tag_name
+ 1;
5244 get_tag (bp
+ len_add
, NULL
);
5253 * Just look for lines where the first character is '/'
5254 * Also look at "defineps" for PSWrap
5256 * Richard Mlynarik <mly@adoc.xerox.com> (1997)
5257 * Masatake Yamato <masata-y@is.aist-nara.ac.jp> (1999)
5260 PS_functions (FILE *inf
)
5262 register char *bp
, *ep
;
5264 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
5269 *ep
!= '\0' && *ep
!= ' ' && *ep
!= '{';
5272 make_tag (bp
, ep
- bp
, true,
5273 lb
.buffer
, ep
- lb
.buffer
+ 1, lineno
, linecharno
);
5275 else if (LOOKING_AT (bp
, "defineps"))
5283 * Ignore anything after \ followed by space or in ( )
5284 * Look for words defined by :
5285 * Look for constant, code, create, defer, value, and variable
5286 * OBP extensions: Look for buffer:, field,
5287 * Ideas by Eduardo Horvath <eeh@netbsd.org> (2004)
5290 Forth_words (FILE *inf
)
5294 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
5295 while ((bp
= skip_spaces (bp
))[0] != '\0')
5296 if (bp
[0] == '\\' && c_isspace (bp
[1]))
5297 break; /* read next line */
5298 else if (bp
[0] == '(' && c_isspace (bp
[1]))
5299 do /* skip to ) or eol */
5301 while (*bp
!= ')' && *bp
!= '\0');
5302 else if ((bp
[0] == ':' && c_isspace (bp
[1]) && bp
++)
5303 || LOOKING_AT_NOCASE (bp
, "constant")
5304 || LOOKING_AT_NOCASE (bp
, "code")
5305 || LOOKING_AT_NOCASE (bp
, "create")
5306 || LOOKING_AT_NOCASE (bp
, "defer")
5307 || LOOKING_AT_NOCASE (bp
, "value")
5308 || LOOKING_AT_NOCASE (bp
, "variable")
5309 || LOOKING_AT_NOCASE (bp
, "buffer:")
5310 || LOOKING_AT_NOCASE (bp
, "field"))
5311 get_tag (skip_spaces (bp
), NULL
); /* Yay! A definition! */
5313 bp
= skip_non_spaces (bp
);
5318 * Scheme tag functions
5319 * look for (def... xyzzy
5321 * (def ... ((...(xyzzy ....
5323 * Original code by Ken Haase (1985?)
5326 Scheme_functions (FILE *inf
)
5330 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
5332 if (strneq (bp
, "(def", 4) || strneq (bp
, "(DEF", 4))
5334 bp
= skip_non_spaces (bp
+4);
5335 /* Skip over open parens and white space. Don't continue past
5337 while (*bp
&& notinname (*bp
))
5341 if (LOOKING_AT (bp
, "(SET!") || LOOKING_AT (bp
, "(set!"))
5347 /* Find tags in TeX and LaTeX input files. */
5349 /* TEX_toktab is a table of TeX control sequences that define tags.
5350 * Each entry records one such control sequence.
5352 * Original code from who knows whom.
5354 * Stefan Monnier (2002)
5357 static linebuffer
*TEX_toktab
= NULL
; /* Table with tag tokens */
5359 /* Default set of control sequences to put into TEX_toktab.
5360 The value of environment var TEXTAGS is prepended to this. */
5361 static const char *TEX_defenv
= "\
5362 :chapter:section:subsection:subsubsection:eqno:label:ref:cite:bibitem\
5363 :part:appendix:entry:index:def\
5364 :newcommand:renewcommand:newenvironment:renewenvironment";
5366 static void TEX_decode_env (const char *, const char *);
5369 * TeX/LaTeX scanning loop.
5372 TeX_commands (FILE *inf
)
5377 char TEX_esc
= '\0';
5378 char TEX_opgrp
, TEX_clgrp
;
5380 /* Initialize token table once from environment. */
5381 if (TEX_toktab
== NULL
)
5382 TEX_decode_env ("TEXTAGS", TEX_defenv
);
5384 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5386 /* Look at each TEX keyword in line. */
5389 /* Look for a TEX escape. */
5393 if (c
== '\0' || c
== '%')
5396 /* Select either \ or ! as escape character, whichever comes
5397 first outside a comment. */
5418 for (key
= TEX_toktab
; key
->buffer
!= NULL
; key
++)
5419 if (strneq (cp
, key
->buffer
, key
->len
))
5422 int namelen
, linelen
;
5425 cp
= skip_spaces (cp
+ key
->len
);
5426 if (*cp
== TEX_opgrp
)
5432 (!c_isspace (*p
) && *p
!= '#' &&
5433 *p
!= TEX_opgrp
&& *p
!= TEX_clgrp
);
5438 if (!opgrp
|| *p
== TEX_clgrp
)
5440 while (*p
!= '\0' && *p
!= TEX_opgrp
&& *p
!= TEX_clgrp
)
5442 linelen
= p
- lb
.buffer
+ 1;
5444 make_tag (cp
, namelen
, true,
5445 lb
.buffer
, linelen
, lineno
, linecharno
);
5446 goto tex_next_line
; /* We only tag a line once */
5454 /* Read environment and prepend it to the default string.
5455 Build token table. */
5457 TEX_decode_env (const char *evarname
, const char *defenv
)
5459 register const char *env
, *p
;
5462 /* Append default string to environment. */
5463 env
= getenv (evarname
);
5467 env
= concat (env
, defenv
, "");
5469 /* Allocate a token table */
5470 for (len
= 1, p
= env
; (p
= strchr (p
, ':')); )
5473 TEX_toktab
= xnew (len
, linebuffer
);
5475 /* Unpack environment string into token table. Be careful about */
5476 /* zero-length strings (leading ':', "::" and trailing ':') */
5477 for (i
= 0; *env
!= '\0';)
5479 p
= strchr (env
, ':');
5480 if (!p
) /* End of environment string. */
5481 p
= env
+ strlen (env
);
5483 { /* Only non-zero strings. */
5484 TEX_toktab
[i
].buffer
= savenstr (env
, p
- env
);
5485 TEX_toktab
[i
].len
= p
- env
;
5492 TEX_toktab
[i
].buffer
= NULL
; /* Mark end of table. */
5493 TEX_toktab
[i
].len
= 0;
5500 /* Texinfo support. Dave Love, Mar. 2000. */
5502 Texinfo_nodes (FILE *inf
)
5505 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5506 if (LOOKING_AT (cp
, "@node"))
5509 while (*cp
!= '\0' && *cp
!= ',')
5511 make_tag (start
, cp
- start
, true,
5512 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
5519 * Contents of <title>, <h1>, <h2>, <h3> are tags.
5520 * Contents of <a name=xxx> are tags with name xxx.
5522 * Francesco Potortì, 2002.
5525 HTML_labels (FILE *inf
)
5527 bool getnext
= false; /* next text outside of HTML tags is a tag */
5528 bool skiptag
= false; /* skip to the end of the current HTML tag */
5529 bool intag
= false; /* inside an html tag, looking for ID= */
5530 bool inanchor
= false; /* when INTAG, is an anchor, look for NAME= */
5534 linebuffer_setlen (&token_name
, 0); /* no name in buffer */
5536 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
5537 for (;;) /* loop on the same line */
5539 if (skiptag
) /* skip HTML tag */
5541 while (*dbp
!= '\0' && *dbp
!= '>')
5547 continue; /* look on the same line */
5549 break; /* go to next line */
5552 else if (intag
) /* look for "name=" or "id=" */
5554 while (*dbp
!= '\0' && *dbp
!= '>'
5555 && c_tolower (*dbp
) != 'n' && c_tolower (*dbp
) != 'i')
5558 break; /* go to next line */
5563 continue; /* look on the same line */
5565 if ((inanchor
&& LOOKING_AT_NOCASE (dbp
, "name="))
5566 || LOOKING_AT_NOCASE (dbp
, "id="))
5568 bool quoted
= (dbp
[0] == '"');
5571 for (end
= ++dbp
; *end
!= '\0' && *end
!= '"'; end
++)
5574 for (end
= dbp
; *end
!= '\0' && intoken (*end
); end
++)
5576 linebuffer_setlen (&token_name
, end
- dbp
);
5577 memcpy (token_name
.buffer
, dbp
, end
- dbp
);
5578 token_name
.buffer
[end
- dbp
] = '\0';
5581 intag
= false; /* we found what we looked for */
5582 skiptag
= true; /* skip to the end of the tag */
5583 getnext
= true; /* then grab the text */
5584 continue; /* look on the same line */
5589 else if (getnext
) /* grab next tokens and tag them */
5591 dbp
= skip_spaces (dbp
);
5593 break; /* go to next line */
5597 inanchor
= (c_tolower (dbp
[1]) == 'a' && !intoken (dbp
[2]));
5598 continue; /* look on the same line */
5601 for (end
= dbp
+ 1; *end
!= '\0' && *end
!= '<'; end
++)
5603 make_tag (token_name
.buffer
, token_name
.len
, true,
5604 dbp
, end
- dbp
, lineno
, linecharno
);
5605 linebuffer_setlen (&token_name
, 0); /* no name in buffer */
5607 break; /* go to next line */
5610 else /* look for an interesting HTML tag */
5612 while (*dbp
!= '\0' && *dbp
!= '<')
5615 break; /* go to next line */
5617 if (c_tolower (dbp
[1]) == 'a' && !intoken (dbp
[2]))
5620 continue; /* look on the same line */
5622 else if (LOOKING_AT_NOCASE (dbp
, "<title>")
5623 || LOOKING_AT_NOCASE (dbp
, "<h1>")
5624 || LOOKING_AT_NOCASE (dbp
, "<h2>")
5625 || LOOKING_AT_NOCASE (dbp
, "<h3>"))
5629 continue; /* look on the same line */
5640 * Assumes that the predicate or rule starts at column 0.
5641 * Only the first clause of a predicate or rule is added.
5642 * Original code by Sunichirou Sugou (1989)
5643 * Rewritten by Anders Lindgren (1996)
5645 static size_t prolog_pr (char *, char *);
5646 static void prolog_skip_comment (linebuffer
*, FILE *);
5647 static size_t prolog_atom (char *, size_t);
5650 Prolog_functions (FILE *inf
)
5660 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5662 if (cp
[0] == '\0') /* Empty line */
5664 else if (c_isspace (cp
[0])) /* Not a predicate */
5666 else if (cp
[0] == '/' && cp
[1] == '*') /* comment. */
5667 prolog_skip_comment (&lb
, inf
);
5668 else if ((len
= prolog_pr (cp
, last
)) > 0)
5670 /* Predicate or rule. Store the function name so that we
5671 only generate a tag for the first clause. */
5673 last
= xnew (len
+ 1, char);
5674 else if (len
+ 1 > allocated
)
5675 xrnew (last
, len
+ 1, char);
5676 allocated
= len
+ 1;
5677 memcpy (last
, cp
, len
);
5686 prolog_skip_comment (linebuffer
*plb
, FILE *inf
)
5692 for (cp
= plb
->buffer
; *cp
!= '\0'; cp
++)
5693 if (cp
[0] == '*' && cp
[1] == '/')
5695 readline (plb
, inf
);
5697 while (perhaps_more_input (inf
));
5701 * A predicate or rule definition is added if it matches:
5702 * <beginning of line><Prolog Atom><whitespace>(
5703 * or <beginning of line><Prolog Atom><whitespace>:-
5705 * It is added to the tags database if it doesn't match the
5706 * name of the previous clause header.
5708 * Return the size of the name of the predicate or rule, or 0 if no
5712 prolog_pr (char *s
, char *last
)
5714 /* Name of last clause. */
5719 pos
= prolog_atom (s
, 0);
5724 pos
= skip_spaces (s
+ pos
) - s
;
5727 || (s
[pos
] == '(' && (pos
+= 1))
5728 || (s
[pos
] == ':' && s
[pos
+ 1] == '-' && (pos
+= 2)))
5729 && (last
== NULL
/* save only the first clause */
5730 || len
!= strlen (last
)
5731 || !strneq (s
, last
, len
)))
5733 make_tag (s
, len
, true, s
, pos
, lineno
, linecharno
);
5741 * Consume a Prolog atom.
5742 * Return the number of bytes consumed, or 0 if there was an error.
5744 * A prolog atom, in this context, could be one of:
5745 * - An alphanumeric sequence, starting with a lower case letter.
5746 * - A quoted arbitrary string. Single quotes can escape themselves.
5747 * Backslash quotes everything.
5750 prolog_atom (char *s
, size_t pos
)
5756 if (c_islower (s
[pos
]) || s
[pos
] == '_')
5758 /* The atom is unquoted. */
5760 while (c_isalnum (s
[pos
]) || s
[pos
] == '_')
5764 return pos
- origpos
;
5766 else if (s
[pos
] == '\'')
5777 pos
++; /* A double quote */
5779 else if (s
[pos
] == '\0')
5780 /* Multiline quoted atoms are ignored. */
5782 else if (s
[pos
] == '\\')
5784 if (s
[pos
+1] == '\0')
5791 return pos
- origpos
;
5799 * Support for Erlang
5801 * Generates tags for functions, defines, and records.
5802 * Assumes that Erlang functions start at column 0.
5803 * Original code by Anders Lindgren (1996)
5805 static int erlang_func (char *, char *);
5806 static void erlang_attribute (char *);
5807 static int erlang_atom (char *);
5810 Erlang_functions (FILE *inf
)
5820 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5822 if (cp
[0] == '\0') /* Empty line */
5824 else if (c_isspace (cp
[0])) /* Not function nor attribute */
5826 else if (cp
[0] == '%') /* comment */
5828 else if (cp
[0] == '"') /* Sometimes, strings start in column one */
5830 else if (cp
[0] == '-') /* attribute, e.g. "-define" */
5832 erlang_attribute (cp
);
5839 else if ((len
= erlang_func (cp
, last
)) > 0)
5842 * Function. Store the function name so that we only
5843 * generates a tag for the first clause.
5846 last
= xnew (len
+ 1, char);
5847 else if (len
+ 1 > allocated
)
5848 xrnew (last
, len
+ 1, char);
5849 allocated
= len
+ 1;
5850 memcpy (last
, cp
, len
);
5859 * A function definition is added if it matches:
5860 * <beginning of line><Erlang Atom><whitespace>(
5862 * It is added to the tags database if it doesn't match the
5863 * name of the previous clause header.
5865 * Return the size of the name of the function, or 0 if no function
5869 erlang_func (char *s
, char *last
)
5871 /* Name of last clause. */
5876 pos
= erlang_atom (s
);
5881 pos
= skip_spaces (s
+ pos
) - s
;
5883 /* Save only the first clause. */
5886 || len
!= (int)strlen (last
)
5887 || !strneq (s
, last
, len
)))
5889 make_tag (s
, len
, true, s
, pos
, lineno
, linecharno
);
5898 * Handle attributes. Currently, tags are generated for defines
5901 * They are on the form:
5902 * -define(foo, bar).
5903 * -define(Foo(M, N), M+N).
5904 * -record(graph, {vtab = notable, cyclic = true}).
5907 erlang_attribute (char *s
)
5911 if ((LOOKING_AT (cp
, "-define") || LOOKING_AT (cp
, "-record"))
5914 int len
= erlang_atom (skip_spaces (cp
));
5916 make_tag (cp
, len
, true, s
, cp
+ len
- s
, lineno
, linecharno
);
5923 * Consume an Erlang atom (or variable).
5924 * Return the number of bytes consumed, or -1 if there was an error.
5927 erlang_atom (char *s
)
5931 if (c_isalpha (s
[pos
]) || s
[pos
] == '_')
5933 /* The atom is unquoted. */
5936 while (c_isalnum (s
[pos
]) || s
[pos
] == '_');
5938 else if (s
[pos
] == '\'')
5940 for (pos
++; s
[pos
] != '\''; pos
++)
5941 if (s
[pos
] == '\0' /* multiline quoted atoms are ignored */
5942 || (s
[pos
] == '\\' && s
[++pos
] == '\0'))
5951 static char *scan_separators (char *);
5952 static void add_regex (char *, language
*);
5953 static char *substitute (char *, char *, struct re_registers
*);
5956 * Take a string like "/blah/" and turn it into "blah", verifying
5957 * that the first and last characters are the same, and handling
5958 * quoted separator characters. Actually, stops on the occurrence of
5959 * an unquoted separator. Also process \t, \n, etc. and turn into
5960 * appropriate characters. Works in place. Null terminates name string.
5961 * Returns pointer to terminating separator, or NULL for
5962 * unterminated regexps.
5965 scan_separators (char *name
)
5968 char *copyto
= name
;
5969 bool quoted
= false;
5971 for (++name
; *name
!= '\0'; ++name
)
5977 case 'a': *copyto
++ = '\007'; break; /* BEL (bell) */
5978 case 'b': *copyto
++ = '\b'; break; /* BS (back space) */
5979 case 'd': *copyto
++ = 0177; break; /* DEL (delete) */
5980 case 'e': *copyto
++ = 033; break; /* ESC (delete) */
5981 case 'f': *copyto
++ = '\f'; break; /* FF (form feed) */
5982 case 'n': *copyto
++ = '\n'; break; /* NL (new line) */
5983 case 'r': *copyto
++ = '\r'; break; /* CR (carriage return) */
5984 case 't': *copyto
++ = '\t'; break; /* TAB (horizontal tab) */
5985 case 'v': *copyto
++ = '\v'; break; /* VT (vertical tab) */
5991 /* Something else is quoted, so preserve the quote. */
5999 else if (*name
== '\\')
6001 else if (*name
== sep
)
6007 name
= NULL
; /* signal unterminated regexp */
6009 /* Terminate copied string. */
6014 /* Look at the argument of --regex or --no-regex and do the right
6015 thing. Same for each line of a regexp file. */
6017 analyze_regex (char *regex_arg
)
6019 if (regex_arg
== NULL
)
6021 free_regexps (); /* --no-regex: remove existing regexps */
6025 /* A real --regexp option or a line in a regexp file. */
6026 switch (regex_arg
[0])
6028 /* Comments in regexp file or null arg to --regex. */
6034 /* Read a regex file. This is recursive and may result in a
6035 loop, which will stop when the file descriptors are exhausted. */
6039 linebuffer regexbuf
;
6040 char *regexfile
= regex_arg
+ 1;
6042 /* regexfile is a file containing regexps, one per line. */
6043 regexfp
= fopen (regexfile
, "r" FOPEN_BINARY
);
6044 if (regexfp
== NULL
)
6046 linebuffer_init (®exbuf
);
6047 while (readline_internal (®exbuf
, regexfp
, regexfile
) > 0)
6048 analyze_regex (regexbuf
.buffer
);
6049 free (regexbuf
.buffer
);
6050 if (fclose (regexfp
) != 0)
6055 /* Regexp to be used for a specific language only. */
6059 char *lang_name
= regex_arg
+ 1;
6062 for (cp
= lang_name
; *cp
!= '}'; cp
++)
6065 error ("unterminated language name in regex: %s", regex_arg
);
6069 lang
= get_language_from_langname (lang_name
);
6072 add_regex (cp
, lang
);
6076 /* Regexp to be used for any language. */
6078 add_regex (regex_arg
, NULL
);
6083 /* Separate the regexp pattern, compile it,
6084 and care for optional name and modifiers. */
6086 add_regex (char *regexp_pattern
, language
*lang
)
6088 static struct re_pattern_buffer zeropattern
;
6089 char sep
, *pat
, *name
, *modifiers
;
6092 struct re_pattern_buffer
*patbuf
;
6095 force_explicit_name
= true, /* do not use implicit tag names */
6096 ignore_case
= false, /* case is significant */
6097 multi_line
= false, /* matches are done one line at a time */
6098 single_line
= false; /* dot does not match newline */
6101 if (strlen (regexp_pattern
) < 3)
6103 error ("null regexp");
6106 sep
= regexp_pattern
[0];
6107 name
= scan_separators (regexp_pattern
);
6110 error ("%s: unterminated regexp", regexp_pattern
);
6115 error ("null name for regexp \"%s\"", regexp_pattern
);
6118 modifiers
= scan_separators (name
);
6119 if (modifiers
== NULL
) /* no terminating separator --> no name */
6125 modifiers
+= 1; /* skip separator */
6127 /* Parse regex modifiers. */
6128 for (; modifiers
[0] != '\0'; modifiers
++)
6129 switch (modifiers
[0])
6132 if (modifiers
== name
)
6133 error ("forcing explicit tag name but no name, ignoring");
6134 force_explicit_name
= true;
6144 need_filebuf
= true;
6147 error ("invalid regexp modifier '%c', ignoring", modifiers
[0]);
6151 patbuf
= xnew (1, struct re_pattern_buffer
);
6152 *patbuf
= zeropattern
;
6155 static char lc_trans
[UCHAR_MAX
+ 1];
6157 for (i
= 0; i
< UCHAR_MAX
+ 1; i
++)
6158 lc_trans
[i
] = c_tolower (i
);
6159 patbuf
->translate
= lc_trans
; /* translation table to fold case */
6163 pat
= concat ("^", regexp_pattern
, ""); /* anchor to beginning of line */
6165 pat
= regexp_pattern
;
6168 re_set_syntax (RE_SYNTAX_EMACS
| RE_DOT_NEWLINE
);
6170 re_set_syntax (RE_SYNTAX_EMACS
);
6172 err
= re_compile_pattern (pat
, strlen (pat
), patbuf
);
6177 error ("%s while compiling pattern", err
);
6182 p_head
= xnew (1, regexp
);
6183 p_head
->pattern
= savestr (regexp_pattern
);
6184 p_head
->p_next
= rp
;
6185 p_head
->lang
= lang
;
6186 p_head
->pat
= patbuf
;
6187 p_head
->name
= savestr (name
);
6188 p_head
->error_signaled
= false;
6189 p_head
->force_explicit_name
= force_explicit_name
;
6190 p_head
->ignore_case
= ignore_case
;
6191 p_head
->multi_line
= multi_line
;
6195 * Do the substitutions indicated by the regular expression and
6199 substitute (char *in
, char *out
, struct re_registers
*regs
)
6202 int size
, dig
, diglen
;
6205 size
= strlen (out
);
6207 /* Pass 1: figure out how much to allocate by finding all \N strings. */
6208 if (out
[size
- 1] == '\\')
6209 fatal ("pattern error in \"%s\"", out
);
6210 for (t
= strchr (out
, '\\');
6212 t
= strchr (t
+ 2, '\\'))
6213 if (c_isdigit (t
[1]))
6216 diglen
= regs
->end
[dig
] - regs
->start
[dig
];
6222 /* Allocate space and do the substitutions. */
6224 result
= xnew (size
+ 1, char);
6226 for (t
= result
; *out
!= '\0'; out
++)
6227 if (*out
== '\\' && c_isdigit (*++out
))
6230 diglen
= regs
->end
[dig
] - regs
->start
[dig
];
6231 memcpy (t
, in
+ regs
->start
[dig
], diglen
);
6238 assert (t
<= result
+ size
);
6239 assert (t
- result
== (int)strlen (result
));
6244 /* Deallocate all regexps. */
6249 while (p_head
!= NULL
)
6251 rp
= p_head
->p_next
;
6252 free (p_head
->pattern
);
6253 free (p_head
->name
);
6261 * Reads the whole file as a single string from `filebuf' and looks for
6262 * multi-line regular expressions, creating tags on matches.
6263 * readline already dealt with normal regexps.
6265 * Idea by Ben Wing <ben@666.com> (2002).
6268 regex_tag_multiline (void)
6270 char *buffer
= filebuf
.buffer
;
6274 for (rp
= p_head
; rp
!= NULL
; rp
= rp
->p_next
)
6278 if (!rp
->multi_line
)
6279 continue; /* skip normal regexps */
6281 /* Generic initializations before parsing file from memory. */
6282 lineno
= 1; /* reset global line number */
6283 charno
= 0; /* reset global char number */
6284 linecharno
= 0; /* reset global char number of line start */
6286 /* Only use generic regexps or those for the current language. */
6287 if (rp
->lang
!= NULL
&& rp
->lang
!= curfdp
->lang
)
6290 while (match
>= 0 && match
< filebuf
.len
)
6292 match
= re_search (rp
->pat
, buffer
, filebuf
.len
, charno
,
6293 filebuf
.len
- match
, &rp
->regs
);
6298 if (!rp
->error_signaled
)
6300 error ("regexp stack overflow while matching \"%s\"",
6302 rp
->error_signaled
= true;
6309 if (match
== rp
->regs
.end
[0])
6311 if (!rp
->error_signaled
)
6313 error ("regexp matches the empty string: \"%s\"",
6315 rp
->error_signaled
= true;
6317 match
= -3; /* exit from while loop */
6321 /* Match occurred. Construct a tag. */
6322 while (charno
< rp
->regs
.end
[0])
6323 if (buffer
[charno
++] == '\n')
6324 lineno
++, linecharno
= charno
;
6326 if (name
[0] == '\0')
6328 else /* make a named tag */
6329 name
= substitute (buffer
, rp
->name
, &rp
->regs
);
6330 if (rp
->force_explicit_name
)
6331 /* Force explicit tag name, if a name is there. */
6332 pfnote (name
, true, buffer
+ linecharno
,
6333 charno
- linecharno
+ 1, lineno
, linecharno
);
6335 make_tag (name
, strlen (name
), true, buffer
+ linecharno
,
6336 charno
- linecharno
+ 1, lineno
, linecharno
);
6345 nocase_tail (const char *cp
)
6349 while (*cp
!= '\0' && c_tolower (*cp
) == c_tolower (dbp
[len
]))
6351 if (*cp
== '\0' && !intoken (dbp
[len
]))
6360 get_tag (register char *bp
, char **namepp
)
6362 register char *cp
= bp
;
6366 /* Go till you get to white space or a syntactic break */
6367 for (cp
= bp
+ 1; !notinname (*cp
); cp
++)
6369 make_tag (bp
, cp
- bp
, true,
6370 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
6374 *namepp
= savenstr (bp
, cp
- bp
);
6378 * Read a line of text from `stream' into `lbp', excluding the
6379 * newline or CR-NL, if any. Return the number of characters read from
6380 * `stream', which is the length of the line including the newline.
6382 * On DOS or Windows we do not count the CR character, if any before the
6383 * NL, in the returned length; this mirrors the behavior of Emacs on those
6384 * platforms (for text files, it translates CR-NL to NL as it reads in the
6387 * If multi-line regular expressions are requested, each line read is
6388 * appended to `filebuf'.
6391 readline_internal (linebuffer
*lbp
, FILE *stream
, char const *filename
)
6393 char *buffer
= lbp
->buffer
;
6394 char *p
= lbp
->buffer
;
6398 pend
= p
+ lbp
->size
; /* Separate to avoid 386/IX compiler bug. */
6402 register int c
= getc (stream
);
6405 /* We're at the end of linebuffer: expand it. */
6407 xrnew (buffer
, lbp
->size
, char);
6408 p
+= buffer
- lbp
->buffer
;
6409 pend
= buffer
+ lbp
->size
;
6410 lbp
->buffer
= buffer
;
6414 if (ferror (stream
))
6422 if (p
> buffer
&& p
[-1] == '\r')
6436 lbp
->len
= p
- buffer
;
6438 if (need_filebuf
/* we need filebuf for multi-line regexps */
6439 && chars_deleted
> 0) /* not at EOF */
6441 while (filebuf
.size
<= filebuf
.len
+ lbp
->len
+ 1) /* +1 for \n */
6443 /* Expand filebuf. */
6445 xrnew (filebuf
.buffer
, filebuf
.size
, char);
6447 memcpy (filebuf
.buffer
+ filebuf
.len
, lbp
->buffer
, lbp
->len
);
6448 filebuf
.len
+= lbp
->len
;
6449 filebuf
.buffer
[filebuf
.len
++] = '\n';
6450 filebuf
.buffer
[filebuf
.len
] = '\0';
6453 return lbp
->len
+ chars_deleted
;
6457 * Like readline_internal, above, but in addition try to match the
6458 * input line against relevant regular expressions and manage #line
6462 readline (linebuffer
*lbp
, FILE *stream
)
6466 linecharno
= charno
; /* update global char number of line start */
6467 result
= readline_internal (lbp
, stream
, infilename
); /* read line */
6468 lineno
+= 1; /* increment global line number */
6469 charno
+= result
; /* increment global char number */
6471 /* Honor #line directives. */
6472 if (!no_line_directive
)
6474 static bool discard_until_line_directive
;
6476 /* Check whether this is a #line directive. */
6477 if (result
> 12 && strneq (lbp
->buffer
, "#line ", 6))
6482 if (sscanf (lbp
->buffer
, "#line %u \"%n", &lno
, &start
) >= 1
6483 && start
> 0) /* double quote character found */
6485 char *endp
= lbp
->buffer
+ start
;
6487 while ((endp
= strchr (endp
, '"')) != NULL
6488 && endp
[-1] == '\\')
6491 /* Ok, this is a real #line directive. Let's deal with it. */
6493 char *taggedabsname
; /* absolute name of original file */
6494 char *taggedfname
; /* name of original file as given */
6495 char *name
; /* temp var */
6497 discard_until_line_directive
= false; /* found it */
6498 name
= lbp
->buffer
+ start
;
6500 canonicalize_filename (name
);
6501 taggedabsname
= absolute_filename (name
, tagfiledir
);
6502 if (filename_is_absolute (name
)
6503 || filename_is_absolute (curfdp
->infname
))
6504 taggedfname
= savestr (taggedabsname
);
6506 taggedfname
= relative_filename (taggedabsname
,tagfiledir
);
6508 if (streq (curfdp
->taggedfname
, taggedfname
))
6509 /* The #line directive is only a line number change. We
6510 deal with this afterwards. */
6513 /* The tags following this #line directive should be
6514 attributed to taggedfname. In order to do this, set
6515 curfdp accordingly. */
6517 fdesc
*fdp
; /* file description pointer */
6519 /* Go look for a file description already set up for the
6520 file indicated in the #line directive. If there is
6521 one, use it from now until the next #line
6523 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
6524 if (streq (fdp
->infname
, curfdp
->infname
)
6525 && streq (fdp
->taggedfname
, taggedfname
))
6526 /* If we remove the second test above (after the &&)
6527 then all entries pertaining to the same file are
6528 coalesced in the tags file. If we use it, then
6529 entries pertaining to the same file but generated
6530 from different files (via #line directives) will
6531 go into separate sections in the tags file. These
6532 alternatives look equivalent. The first one
6533 destroys some apparently useless information. */
6539 /* Else, if we already tagged the real file, skip all
6540 input lines until the next #line directive. */
6541 if (fdp
== NULL
) /* not found */
6542 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
6543 if (streq (fdp
->infabsname
, taggedabsname
))
6545 discard_until_line_directive
= true;
6549 /* Else create a new file description and use that from
6550 now on, until the next #line directive. */
6551 if (fdp
== NULL
) /* not found */
6554 fdhead
= xnew (1, fdesc
);
6555 *fdhead
= *curfdp
; /* copy curr. file description */
6557 fdhead
->infname
= savestr (curfdp
->infname
);
6558 fdhead
->infabsname
= savestr (curfdp
->infabsname
);
6559 fdhead
->infabsdir
= savestr (curfdp
->infabsdir
);
6560 fdhead
->taggedfname
= taggedfname
;
6561 fdhead
->usecharno
= false;
6562 fdhead
->prop
= NULL
;
6563 fdhead
->written
= false;
6567 free (taggedabsname
);
6569 readline (lbp
, stream
);
6571 } /* if a real #line directive */
6572 } /* if #line is followed by a number */
6573 } /* if line begins with "#line " */
6575 /* If we are here, no #line directive was found. */
6576 if (discard_until_line_directive
)
6580 /* Do a tail recursion on ourselves, thus discarding the contents
6581 of the line buffer. */
6582 readline (lbp
, stream
);
6586 discard_until_line_directive
= false;
6589 } /* if #line directives should be considered */
6596 /* Match against relevant regexps. */
6598 for (rp
= p_head
; rp
!= NULL
; rp
= rp
->p_next
)
6600 /* Only use generic regexps or those for the current language.
6601 Also do not use multiline regexps, which is the job of
6602 regex_tag_multiline. */
6603 if ((rp
->lang
!= NULL
&& rp
->lang
!= fdhead
->lang
)
6607 match
= re_match (rp
->pat
, lbp
->buffer
, lbp
->len
, 0, &rp
->regs
);
6612 if (!rp
->error_signaled
)
6614 error ("regexp stack overflow while matching \"%s\"",
6616 rp
->error_signaled
= true;
6623 /* Empty string matched. */
6624 if (!rp
->error_signaled
)
6626 error ("regexp matches the empty string: \"%s\"", rp
->pattern
);
6627 rp
->error_signaled
= true;
6631 /* Match occurred. Construct a tag. */
6633 if (name
[0] == '\0')
6635 else /* make a named tag */
6636 name
= substitute (lbp
->buffer
, rp
->name
, &rp
->regs
);
6637 if (rp
->force_explicit_name
)
6638 /* Force explicit tag name, if a name is there. */
6639 pfnote (name
, true, lbp
->buffer
, match
, lineno
, linecharno
);
6641 make_tag (name
, strlen (name
), true,
6642 lbp
->buffer
, match
, lineno
, linecharno
);
6651 * Return a pointer to a space of size strlen(cp)+1 allocated
6652 * with xnew where the string CP has been copied.
6655 savestr (const char *cp
)
6657 return savenstr (cp
, strlen (cp
));
6661 * Return a pointer to a space of size LEN+1 allocated with xnew where
6662 * the string CP has been copied for at most the first LEN characters.
6665 savenstr (const char *cp
, int len
)
6667 char *dp
= xnew (len
+ 1, char);
6669 return memcpy (dp
, cp
, len
);
6672 /* Skip spaces (end of string is not space), return new pointer. */
6674 skip_spaces (char *cp
)
6676 while (c_isspace (*cp
))
6681 /* Skip non spaces, except end of string, return new pointer. */
6683 skip_non_spaces (char *cp
)
6685 while (*cp
!= '\0' && !c_isspace (*cp
))
6690 /* Skip any chars in the "name" class.*/
6692 skip_name (char *cp
)
6694 /* '\0' is a notinname() so loop stops there too */
6695 while (! notinname (*cp
))
6700 /* Print error message and exit. */
6702 fatal (char const *format
, ...)
6705 va_start (ap
, format
);
6706 verror (format
, ap
);
6708 exit (EXIT_FAILURE
);
6712 pfatal (const char *s1
)
6715 exit (EXIT_FAILURE
);
6719 suggest_asking_for_help (void)
6721 fprintf (stderr
, "\tTry '%s --help' for a complete list of options.\n",
6723 exit (EXIT_FAILURE
);
6726 /* Output a diagnostic with printf-style FORMAT and args. */
6728 error (const char *format
, ...)
6731 va_start (ap
, format
);
6732 verror (format
, ap
);
6737 verror (char const *format
, va_list ap
)
6739 fprintf (stderr
, "%s: ", progname
);
6740 vfprintf (stderr
, format
, ap
);
6741 fprintf (stderr
, "\n");
6744 /* Return a newly-allocated string whose contents
6745 concatenate those of s1, s2, s3. */
6747 concat (const char *s1
, const char *s2
, const char *s3
)
6749 int len1
= strlen (s1
), len2
= strlen (s2
), len3
= strlen (s3
);
6750 char *result
= xnew (len1
+ len2
+ len3
+ 1, char);
6752 strcpy (result
, s1
);
6753 strcpy (result
+ len1
, s2
);
6754 strcpy (result
+ len1
+ len2
, s3
);
6760 /* Does the same work as the system V getcwd, but does not need to
6761 guess the buffer size in advance. */
6766 char *path
= xnew (bufsize
, char);
6768 while (getcwd (path
, bufsize
) == NULL
)
6770 if (errno
!= ERANGE
)
6774 path
= xnew (bufsize
, char);
6777 canonicalize_filename (path
);
6781 /* Return a newly allocated string containing a name of a temporary file. */
6785 const char *tmpdir
= getenv ("TMPDIR");
6786 const char *slash
= "/";
6788 #if MSDOS || defined (DOS_NT)
6790 tmpdir
= getenv ("TEMP");
6792 tmpdir
= getenv ("TMP");
6795 if (tmpdir
[strlen (tmpdir
) - 1] == '/'
6796 || tmpdir
[strlen (tmpdir
) - 1] == '\\')
6801 if (tmpdir
[strlen (tmpdir
) - 1] == '/')
6805 char *templt
= concat (tmpdir
, slash
, "etXXXXXX");
6806 int fd
= mkostemp (templt
, O_CLOEXEC
);
6807 if (fd
< 0 || close (fd
) != 0)
6809 int temp_errno
= errno
;
6815 #if defined (DOS_NT)
6816 /* The file name will be used in shell redirection, so it needs to have
6817 DOS-style backslashes, or else the Windows shell will barf. */
6819 for (p
= templt
; *p
; p
++)
6827 /* Return a newly allocated string containing the file name of FILE
6828 relative to the absolute directory DIR (which should end with a slash). */
6830 relative_filename (char *file
, char *dir
)
6832 char *fp
, *dp
, *afn
, *res
;
6835 /* Find the common root of file and dir (with a trailing slash). */
6836 afn
= absolute_filename (file
, cwd
);
6839 while (*fp
++ == *dp
++)
6841 fp
--, dp
--; /* back to the first differing char */
6843 if (fp
== afn
&& afn
[0] != '/') /* cannot build a relative name */
6846 do /* look at the equal chars until '/' */
6850 /* Build a sequence of "../" strings for the resulting relative file name. */
6852 while ((dp
= strchr (dp
+ 1, '/')) != NULL
)
6854 res
= xnew (3*i
+ strlen (fp
+ 1) + 1, char);
6857 z
= stpcpy (z
, "../");
6859 /* Add the file name relative to the common root of file and dir. */
6866 /* Return a newly allocated string containing the absolute file name
6867 of FILE given DIR (which should end with a slash). */
6869 absolute_filename (char *file
, char *dir
)
6871 char *slashp
, *cp
, *res
;
6873 if (filename_is_absolute (file
))
6874 res
= savestr (file
);
6876 /* We don't support non-absolute file names with a drive
6877 letter, like `d:NAME' (it's too much hassle). */
6878 else if (file
[1] == ':')
6879 fatal ("%s: relative file names with drive letters not supported", file
);
6882 res
= concat (dir
, file
, "");
6884 /* Delete the "/dirname/.." and "/." substrings. */
6885 slashp
= strchr (res
, '/');
6886 while (slashp
!= NULL
&& slashp
[0] != '\0')
6888 if (slashp
[1] == '.')
6890 if (slashp
[2] == '.'
6891 && (slashp
[3] == '/' || slashp
[3] == '\0'))
6896 while (cp
>= res
&& !filename_is_absolute (cp
));
6898 cp
= slashp
; /* the absolute name begins with "/.." */
6900 /* Under MSDOS and NT we get `d:/NAME' as absolute
6901 file name, so the luser could say `d:/../NAME'.
6902 We silently treat this as `d:/NAME'. */
6903 else if (cp
[0] != '/')
6906 memmove (cp
, slashp
+ 3, strlen (slashp
+ 2));
6910 else if (slashp
[2] == '/' || slashp
[2] == '\0')
6912 memmove (slashp
, slashp
+ 2, strlen (slashp
+ 1));
6917 slashp
= strchr (slashp
+ 1, '/');
6920 if (res
[0] == '\0') /* just a safety net: should never happen */
6923 return savestr ("/");
6929 /* Return a newly allocated string containing the absolute
6930 file name of dir where FILE resides given DIR (which should
6931 end with a slash). */
6933 absolute_dirname (char *file
, char *dir
)
6938 slashp
= strrchr (file
, '/');
6940 return savestr (dir
);
6943 res
= absolute_filename (file
, dir
);
6949 /* Whether the argument string is an absolute file name. The argument
6950 string must have been canonicalized with canonicalize_filename. */
6952 filename_is_absolute (char *fn
)
6954 return (fn
[0] == '/'
6956 || (c_isalpha (fn
[0]) && fn
[1] == ':' && fn
[2] == '/')
6961 /* Downcase DOS drive letter and collapse separators into single slashes.
6964 canonicalize_filename (register char *fn
)
6969 /* Canonicalize drive letter case. */
6970 if (c_isupper (fn
[0]) && fn
[1] == ':')
6971 fn
[0] = c_tolower (fn
[0]);
6973 /* Collapse multiple forward- and back-slashes into a single forward
6975 for (cp
= fn
; *cp
!= '\0'; cp
++, fn
++)
6976 if (*cp
== '/' || *cp
== '\\')
6979 while (cp
[1] == '/' || cp
[1] == '\\')
6987 /* Collapse multiple slashes into a single slash. */
6988 for (cp
= fn
; *cp
!= '\0'; cp
++, fn
++)
6992 while (cp
[1] == '/')
6998 #endif /* !DOS_NT */
7004 /* Initialize a linebuffer for use. */
7006 linebuffer_init (linebuffer
*lbp
)
7008 lbp
->size
= (DEBUG
) ? 3 : 200;
7009 lbp
->buffer
= xnew (lbp
->size
, char);
7010 lbp
->buffer
[0] = '\0';
7014 /* Set the minimum size of a string contained in a linebuffer. */
7016 linebuffer_setlen (linebuffer
*lbp
, int toksize
)
7018 while (lbp
->size
<= toksize
)
7021 xrnew (lbp
->buffer
, lbp
->size
, char);
7026 /* Like malloc but get fatal error if memory is exhausted. */
7028 xmalloc (size_t size
)
7030 void *result
= malloc (size
);
7032 fatal ("virtual memory exhausted");
7037 xrealloc (void *ptr
, size_t size
)
7039 void *result
= realloc (ptr
, size
);
7041 fatal ("virtual memory exhausted");
7047 * indent-tabs-mode: t
7050 * c-font-lock-extra-types: ("FILE" "bool" "language" "linebuffer" "fdesc" "node" "regexp")
7051 * c-file-style: "gnu"
7055 /* etags.c ends here */