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
39 (at your option) any later version.
41 This program is distributed in the hope that it will be useful,
42 but WITHOUT ANY WARRANTY; without even the implied warranty of
43 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
44 GNU General Public License for more details.
46 You should have received a copy of the GNU General Public License
47 along with this program. If not, see <http://www.gnu.org/licenses/>. */
50 /* NB To comply with the above BSD license, copyright information is
51 reproduced in etc/ETAGS.README. That file should be updated when the
54 To the best of our knowledge, this code was originally based on the
55 ctags.c distributed with BSD4.2, which was copyrighted by the
56 University of California, as described above. */
61 * 1983 Ctags originally by Ken Arnold.
62 * 1984 Fortran added by Jim Kleckner.
63 * 1984 Ed Pelegri-Llopart added C typedefs.
64 * 1985 Emacs TAGS format by Richard Stallman.
65 * 1989 Sam Kendall added C++.
66 * 1992 Joseph B. Wells improved C and C++ parsing.
67 * 1993 Francesco Potortì reorganized C and C++.
68 * 1994 Line-by-line regexp tags by Tom Tromey.
69 * 2001 Nested classes by Francesco Potortì (concept by Mykola Dzyuba).
70 * 2002 #line directives by Francesco Potortì.
71 * Francesco Potortì maintained and improved it for many years
76 * If you want to add support for a new language, start by looking at the LUA
77 * language, which is the simplest. Alternatively, consider distributing etags
78 * together with a configuration file containing regexp definitions for etags.
81 char pot_etags_version
[] = "@(#) pot revision number is 17.38.1.4";
88 # define NDEBUG /* disable assert */
94 # define _GNU_SOURCE 1 /* enables some compiler checks on GNU */
97 /* WIN32_NATIVE is for XEmacs.
98 MSDOS, WINDOWSNT, DOS_NT are for Emacs. */
103 #endif /* WIN32_NATIVE */
108 # include <sys/param.h>
115 # define MAXPATHLEN _MAX_PATH
119 # define O_CLOEXEC O_NOINHERIT
120 #endif /* WINDOWSNT */
127 #include <sysstdio.h>
130 #include <binary-io.h>
132 #include <c-strcase.h>
136 # undef assert /* some systems have a buggy assert.h */
137 # define assert(x) ((void) 0)
143 /* Define CTAGS to make the program "ctags" compatible with the usual one.
144 Leave it undefined to make the program "etags", which makes emacs-style
145 tag tables and tags typedefs, #defines and struct/union/enum by default. */
154 streq (char const *s
, char const *t
)
156 return strcmp (s
, t
) == 0;
160 strcaseeq (char const *s
, char const *t
)
162 return c_strcasecmp (s
, t
) == 0;
166 strneq (char const *s
, char const *t
, size_t n
)
168 return strncmp (s
, t
, n
) == 0;
172 strncaseeq (char const *s
, char const *t
, size_t n
)
174 return c_strncasecmp (s
, t
, n
) == 0;
177 /* C is not in a name. */
179 notinname (unsigned char c
)
181 /* Look at make_tag before modifying! */
182 static bool const table
[UCHAR_MAX
+ 1] = {
183 ['\0']=1, ['\t']=1, ['\n']=1, ['\f']=1, ['\r']=1, [' ']=1,
184 ['(']=1, [')']=1, [',']=1, [';']=1, ['=']=1
189 /* C can start a token. */
191 begtoken (unsigned char c
)
193 static bool const table
[UCHAR_MAX
+ 1] = {
195 ['A']=1, ['B']=1, ['C']=1, ['D']=1, ['E']=1, ['F']=1, ['G']=1, ['H']=1,
196 ['I']=1, ['J']=1, ['K']=1, ['L']=1, ['M']=1, ['N']=1, ['O']=1, ['P']=1,
197 ['Q']=1, ['R']=1, ['S']=1, ['T']=1, ['U']=1, ['V']=1, ['W']=1, ['X']=1,
200 ['a']=1, ['b']=1, ['c']=1, ['d']=1, ['e']=1, ['f']=1, ['g']=1, ['h']=1,
201 ['i']=1, ['j']=1, ['k']=1, ['l']=1, ['m']=1, ['n']=1, ['o']=1, ['p']=1,
202 ['q']=1, ['r']=1, ['s']=1, ['t']=1, ['u']=1, ['v']=1, ['w']=1, ['x']=1,
209 /* C can be in the middle of a token. */
211 intoken (unsigned char c
)
213 static bool const table
[UCHAR_MAX
+ 1] = {
215 ['0']=1, ['1']=1, ['2']=1, ['3']=1, ['4']=1,
216 ['5']=1, ['6']=1, ['7']=1, ['8']=1, ['9']=1,
217 ['A']=1, ['B']=1, ['C']=1, ['D']=1, ['E']=1, ['F']=1, ['G']=1, ['H']=1,
218 ['I']=1, ['J']=1, ['K']=1, ['L']=1, ['M']=1, ['N']=1, ['O']=1, ['P']=1,
219 ['Q']=1, ['R']=1, ['S']=1, ['T']=1, ['U']=1, ['V']=1, ['W']=1, ['X']=1,
222 ['a']=1, ['b']=1, ['c']=1, ['d']=1, ['e']=1, ['f']=1, ['g']=1, ['h']=1,
223 ['i']=1, ['j']=1, ['k']=1, ['l']=1, ['m']=1, ['n']=1, ['o']=1, ['p']=1,
224 ['q']=1, ['r']=1, ['s']=1, ['t']=1, ['u']=1, ['v']=1, ['w']=1, ['x']=1,
230 /* C can end a token. */
232 endtoken (unsigned char c
)
234 static bool const table
[UCHAR_MAX
+ 1] = {
235 ['\0']=1, ['\t']=1, ['\n']=1, ['\r']=1, [' ']=1,
236 ['!']=1, ['"']=1, ['#']=1, ['%']=1, ['&']=1, ['\'']=1, ['(']=1, [')']=1,
237 ['*']=1, ['+']=1, [',']=1, ['-']=1, ['.']=1, ['/']=1, [':']=1, [';']=1,
238 ['<']=1, ['=']=1, ['>']=1, ['?']=1, ['[']=1, [']']=1, ['^']=1,
239 ['{']=1, ['|']=1, ['}']=1, ['~']=1
245 * xnew, xrnew -- allocate, reallocate storage
247 * SYNOPSIS: Type *xnew (int n, Type);
248 * void xrnew (OldPointer, int n, Type);
250 #define xnew(n, Type) ((Type *) xmalloc ((n) * sizeof (Type)))
251 #define xrnew(op, n, Type) ((op) = (Type *) xrealloc (op, (n) * sizeof (Type)))
253 typedef void Lang_function (FILE *);
257 const char *suffix
; /* file name suffix for this compressor */
258 const char *command
; /* takes one arg and decompresses to stdout */
263 const char *name
; /* language name */
264 const char *help
; /* detailed help for the language */
265 Lang_function
*function
; /* parse function */
266 const char **suffixes
; /* name suffixes of this language's files */
267 const char **filenames
; /* names of this language's files */
268 const char **interpreters
; /* interpreters for this language */
269 bool metasource
; /* source used to generate other sources */
274 struct fdesc
*next
; /* for the linked list */
275 char *infname
; /* uncompressed input file name */
276 char *infabsname
; /* absolute uncompressed input file name */
277 char *infabsdir
; /* absolute dir of input file */
278 char *taggedfname
; /* file name to write in tagfile */
279 language
*lang
; /* language of file */
280 char *prop
; /* file properties to write in tagfile */
281 bool usecharno
; /* etags tags shall contain char number */
282 bool written
; /* entry written in the tags file */
285 typedef struct node_st
286 { /* sorting structure */
287 struct node_st
*left
, *right
; /* left and right sons */
288 fdesc
*fdp
; /* description of file to whom tag belongs */
289 char *name
; /* tag name */
290 char *regex
; /* search regexp */
291 bool valid
; /* write this tag on the tag file */
292 bool is_func
; /* function tag: use regexp in CTAGS mode */
293 bool been_warned
; /* warning already given for duplicated tag */
294 int lno
; /* line number tag is on */
295 long cno
; /* character number line starts on */
299 * A `linebuffer' is a structure which holds a line of text.
300 * `readline_internal' reads a line from a stream into a linebuffer
301 * and works regardless of the length of the line.
302 * SIZE is the size of BUFFER, LEN is the length of the string in
303 * BUFFER after readline reads it.
312 /* Used to support mixing of --lang and file names. */
316 at_language
, /* a language specification */
317 at_regexp
, /* a regular expression */
318 at_filename
, /* a file name */
319 at_stdin
, /* read from stdin here */
320 at_end
/* stop parsing the list */
321 } arg_type
; /* argument type */
322 language
*lang
; /* language associated with the argument */
323 char *what
; /* the argument itself */
326 /* Structure defining a regular expression. */
327 typedef struct regexp
329 struct regexp
*p_next
; /* pointer to next in list */
330 language
*lang
; /* if set, use only for this language */
331 char *pattern
; /* the regexp pattern */
332 char *name
; /* tag name */
333 struct re_pattern_buffer
*pat
; /* the compiled pattern */
334 struct re_registers regs
; /* re registers */
335 bool error_signaled
; /* already signaled for this regexp */
336 bool force_explicit_name
; /* do not allow implicit tag name */
337 bool ignore_case
; /* ignore case when matching */
338 bool multi_line
; /* do a multi-line match on the whole file */
342 /* Many compilers barf on this:
343 Lang_function Ada_funcs;
344 so let's write it this way */
345 static void Ada_funcs (FILE *);
346 static void Asm_labels (FILE *);
347 static void C_entries (int c_ext
, FILE *);
348 static void default_C_entries (FILE *);
349 static void plain_C_entries (FILE *);
350 static void Cjava_entries (FILE *);
351 static void Cobol_paragraphs (FILE *);
352 static void Cplusplus_entries (FILE *);
353 static void Cstar_entries (FILE *);
354 static void Erlang_functions (FILE *);
355 static void Forth_words (FILE *);
356 static void Fortran_functions (FILE *);
357 static void Go_functions (FILE *);
358 static void HTML_labels (FILE *);
359 static void Lisp_functions (FILE *);
360 static void Lua_functions (FILE *);
361 static void Makefile_targets (FILE *);
362 static void Pascal_functions (FILE *);
363 static void Perl_functions (FILE *);
364 static void PHP_functions (FILE *);
365 static void PS_functions (FILE *);
366 static void Prolog_functions (FILE *);
367 static void Python_functions (FILE *);
368 static void Ruby_functions (FILE *);
369 static void Scheme_functions (FILE *);
370 static void TeX_commands (FILE *);
371 static void Texinfo_nodes (FILE *);
372 static void Yacc_entries (FILE *);
373 static void just_read_file (FILE *);
375 static language
*get_language_from_langname (const char *);
376 static void readline (linebuffer
*, FILE *);
377 static long readline_internal (linebuffer
*, FILE *, char const *);
378 static bool nocase_tail (const char *);
379 static void get_tag (char *, char **);
381 static void analyze_regex (char *);
382 static void free_regexps (void);
383 static void regex_tag_multiline (void);
384 static void error (const char *, ...) ATTRIBUTE_FORMAT_PRINTF (1, 2);
385 static void verror (char const *, va_list) ATTRIBUTE_FORMAT_PRINTF (1, 0);
386 static _Noreturn
void suggest_asking_for_help (void);
387 static _Noreturn
void fatal (char const *, ...) ATTRIBUTE_FORMAT_PRINTF (1, 2);
388 static _Noreturn
void pfatal (const char *);
389 static void add_node (node
*, node
**);
391 static void process_file_name (char *, language
*);
392 static void process_file (FILE *, char *, language
*);
393 static void find_entries (FILE *);
394 static void free_tree (node
*);
395 static void free_fdesc (fdesc
*);
396 static void pfnote (char *, bool, char *, int, int, long);
397 static void invalidate_nodes (fdesc
*, node
**);
398 static void put_entries (node
*);
400 static char *concat (const char *, const char *, const char *);
401 static char *skip_spaces (char *);
402 static char *skip_non_spaces (char *);
403 static char *skip_name (char *);
404 static char *savenstr (const char *, int);
405 static char *savestr (const char *);
406 static char *etags_getcwd (void);
407 static char *relative_filename (char *, char *);
408 static char *absolute_filename (char *, char *);
409 static char *absolute_dirname (char *, char *);
410 static bool filename_is_absolute (char *f
);
411 static void canonicalize_filename (char *);
412 static char *etags_mktmp (void);
413 static void linebuffer_init (linebuffer
*);
414 static void linebuffer_setlen (linebuffer
*, int);
415 static void *xmalloc (size_t);
416 static void *xrealloc (void *, size_t);
419 static char searchar
= '/'; /* use /.../ searches */
421 static char *tagfile
; /* output file */
422 static char *progname
; /* name this program was invoked with */
423 static char *cwd
; /* current working directory */
424 static char *tagfiledir
; /* directory of tagfile */
425 static FILE *tagf
; /* ioptr for tags file */
426 static ptrdiff_t whatlen_max
; /* maximum length of any 'what' member */
428 static fdesc
*fdhead
; /* head of file description list */
429 static fdesc
*curfdp
; /* current file description */
430 static char *infilename
; /* current input file name */
431 static int lineno
; /* line number of current line */
432 static long charno
; /* current character number */
433 static long linecharno
; /* charno of start of current line */
434 static char *dbp
; /* pointer to start of current tag */
436 static const int invalidcharno
= -1;
438 static node
*nodehead
; /* the head of the binary tree of tags */
439 static node
*last_node
; /* the last node created */
441 static linebuffer lb
; /* the current line */
442 static linebuffer filebuf
; /* a buffer containing the whole file */
443 static linebuffer token_name
; /* a buffer containing a tag name */
445 static bool append_to_tagfile
; /* -a: append to tags */
446 /* The next five default to true in C and derived languages. */
447 static bool typedefs
; /* -t: create tags for C and Ada typedefs */
448 static bool typedefs_or_cplusplus
; /* -T: create tags for C typedefs, level */
449 /* 0 struct/enum/union decls, and C++ */
450 /* member functions. */
451 static bool constantypedefs
; /* -d: create tags for C #define, enum */
452 /* constants and variables. */
453 /* -D: opposite of -d. Default under ctags. */
454 static int globals
; /* create tags for global variables */
455 static int members
; /* create tags for C member variables */
456 static int declarations
; /* --declarations: tag them and extern in C&Co*/
457 static int no_line_directive
; /* ignore #line directives (undocumented) */
458 static int no_duplicates
; /* no duplicate tags for ctags (undocumented) */
459 static bool update
; /* -u: update tags */
460 static bool vgrind_style
; /* -v: create vgrind style index output */
461 static bool no_warnings
; /* -w: suppress warnings (undocumented) */
462 static bool cxref_style
; /* -x: create cxref style output */
463 static bool cplusplus
; /* .[hc] means C++, not C (undocumented) */
464 static bool ignoreindent
; /* -I: ignore indentation in C */
465 static int packages_only
; /* --packages-only: in Ada, only tag packages*/
466 static int class_qualify
; /* -Q: produce class-qualified tags in C++/Java */
468 /* STDIN is defined in LynxOS system headers */
473 #define STDIN 0x1001 /* returned by getopt_long on --parse-stdin */
474 static bool parsing_stdin
; /* --parse-stdin used */
476 static regexp
*p_head
; /* list of all regexps */
477 static bool need_filebuf
; /* some regexes are multi-line */
479 static struct option longopts
[] =
481 { "append", no_argument
, NULL
, 'a' },
482 { "packages-only", no_argument
, &packages_only
, 1 },
483 { "c++", no_argument
, NULL
, 'C' },
484 { "declarations", no_argument
, &declarations
, 1 },
485 { "no-line-directive", no_argument
, &no_line_directive
, 1 },
486 { "no-duplicates", no_argument
, &no_duplicates
, 1 },
487 { "help", no_argument
, NULL
, 'h' },
488 { "help", no_argument
, NULL
, 'H' },
489 { "ignore-indentation", no_argument
, NULL
, 'I' },
490 { "language", required_argument
, NULL
, 'l' },
491 { "members", no_argument
, &members
, 1 },
492 { "no-members", no_argument
, &members
, 0 },
493 { "output", required_argument
, NULL
, 'o' },
494 { "class-qualify", no_argument
, &class_qualify
, 'Q' },
495 { "regex", required_argument
, NULL
, 'r' },
496 { "no-regex", no_argument
, NULL
, 'R' },
497 { "ignore-case-regex", required_argument
, NULL
, 'c' },
498 { "parse-stdin", required_argument
, NULL
, STDIN
},
499 { "version", no_argument
, NULL
, 'V' },
501 #if CTAGS /* Ctags options */
502 { "backward-search", no_argument
, NULL
, 'B' },
503 { "cxref", no_argument
, NULL
, 'x' },
504 { "defines", no_argument
, NULL
, 'd' },
505 { "globals", no_argument
, &globals
, 1 },
506 { "typedefs", no_argument
, NULL
, 't' },
507 { "typedefs-and-c++", no_argument
, NULL
, 'T' },
508 { "update", no_argument
, NULL
, 'u' },
509 { "vgrind", no_argument
, NULL
, 'v' },
510 { "no-warn", no_argument
, NULL
, 'w' },
512 #else /* Etags options */
513 { "no-defines", no_argument
, NULL
, 'D' },
514 { "no-globals", no_argument
, &globals
, 0 },
515 { "include", required_argument
, NULL
, 'i' },
520 static compressor compressors
[] =
522 { "z", "gzip -d -c"},
523 { "Z", "gzip -d -c"},
524 { "gz", "gzip -d -c"},
525 { "GZ", "gzip -d -c"},
526 { "bz2", "bzip2 -d -c" },
527 { "xz", "xz -d -c" },
536 static const char *Ada_suffixes
[] =
537 { "ads", "adb", "ada", NULL
};
538 static const char Ada_help
[] =
539 "In Ada code, functions, procedures, packages, tasks and types are\n\
540 tags. Use the '--packages-only' option to create tags for\n\
542 Ada tag names have suffixes indicating the type of entity:\n\
543 Entity type: Qualifier:\n\
544 ------------ ----------\n\
551 Thus, 'M-x find-tag <RET> bidule/b <RET>' will go directly to the\n\
552 body of the package 'bidule', while 'M-x find-tag <RET> bidule <RET>'\n\
553 will just search for any tag 'bidule'.";
556 static const char *Asm_suffixes
[] =
557 { "a", /* Unix assembler */
558 "asm", /* Microcontroller assembly */
559 "def", /* BSO/Tasking definition includes */
560 "inc", /* Microcontroller include files */
561 "ins", /* Microcontroller include files */
562 "s", "sa", /* Unix assembler */
563 "S", /* cpp-processed Unix assembler */
564 "src", /* BSO/Tasking C compiler output */
567 static const char Asm_help
[] =
568 "In assembler code, labels appearing at the beginning of a line,\n\
569 followed by a colon, are tags.";
572 /* Note that .c and .h can be considered C++, if the --c++ flag was
573 given, or if the `class' or `template' keywords are met inside the file.
574 That is why default_C_entries is called for these. */
575 static const char *default_C_suffixes
[] =
577 #if CTAGS /* C help for Ctags */
578 static const char default_C_help
[] =
579 "In C code, any C function is a tag. Use -t to tag typedefs.\n\
580 Use -T to tag definitions of 'struct', 'union' and 'enum'.\n\
581 Use -d to tag '#define' macro definitions and 'enum' constants.\n\
582 Use --globals to tag global variables.\n\
583 You can tag function declarations and external variables by\n\
584 using '--declarations', and struct members by using '--members'.";
585 #else /* C help for Etags */
586 static const char default_C_help
[] =
587 "In C code, any C function or typedef is a tag, and so are\n\
588 definitions of 'struct', 'union' and 'enum'. '#define' macro\n\
589 definitions and 'enum' constants are tags unless you specify\n\
590 '--no-defines'. Global variables are tags unless you specify\n\
591 '--no-globals' and so are struct members unless you specify\n\
592 '--no-members'. Use of '--no-globals', '--no-defines' and\n\
593 '--no-members' can make the tags table file much smaller.\n\
594 You can tag function declarations and external variables by\n\
595 using '--declarations'.";
596 #endif /* C help for Ctags and Etags */
598 static const char *Cplusplus_suffixes
[] =
599 { "C", "c++", "cc", "cpp", "cxx", "H", "h++", "hh", "hpp", "hxx",
600 "M", /* Objective C++ */
601 "pdb", /* PostScript with C syntax */
603 static const char Cplusplus_help
[] =
604 "In C++ code, all the tag constructs of C code are tagged. (Use\n\
605 --help --lang=c --lang=c++ for full help.)\n\
606 In addition to C tags, member functions are also recognized. Member\n\
607 variables are recognized unless you use the '--no-members' option.\n\
608 Tags for variables and functions in classes are named 'CLASS::VARIABLE'\n\
609 and 'CLASS::FUNCTION'. 'operator' definitions have tag names like\n\
612 static const char *Cjava_suffixes
[] =
614 static char Cjava_help
[] =
615 "In Java code, all the tags constructs of C and C++ code are\n\
616 tagged. (Use --help --lang=c --lang=c++ --lang=java for full help.)";
619 static const char *Cobol_suffixes
[] =
620 { "COB", "cob", NULL
};
621 static char Cobol_help
[] =
622 "In Cobol code, tags are paragraph names; that is, any word\n\
623 starting in column 8 and followed by a period.";
625 static const char *Cstar_suffixes
[] =
626 { "cs", "hs", NULL
};
628 static const char *Erlang_suffixes
[] =
629 { "erl", "hrl", NULL
};
630 static const char Erlang_help
[] =
631 "In Erlang code, the tags are the functions, records and macros\n\
632 defined in the file.";
634 const char *Forth_suffixes
[] =
635 { "fth", "tok", NULL
};
636 static const char Forth_help
[] =
637 "In Forth code, tags are words defined by ':',\n\
638 constant, code, create, defer, value, variable, buffer:, field.";
640 static const char *Fortran_suffixes
[] =
641 { "F", "f", "f90", "for", NULL
};
642 static const char Fortran_help
[] =
643 "In Fortran code, functions, subroutines and block data are tags.";
645 static const char *Go_suffixes
[] = {"go", NULL
};
646 static const char Go_help
[] =
647 "In Go code, functions, interfaces and packages are tags.";
649 static const char *HTML_suffixes
[] =
650 { "htm", "html", "shtml", NULL
};
651 static const char HTML_help
[] =
652 "In HTML input files, the tags are the 'title' and the 'h1', 'h2',\n\
653 'h3' headers. Also, tags are 'name=' in anchors and all\n\
654 occurrences of 'id='.";
656 static const char *Lisp_suffixes
[] =
657 { "cl", "clisp", "el", "l", "lisp", "LSP", "lsp", "ml", NULL
};
658 static const char Lisp_help
[] =
659 "In Lisp code, any function defined with 'defun', any variable\n\
660 defined with 'defvar' or 'defconst', and in general the first\n\
661 argument of any expression that starts with '(def' in column zero\n\
663 The '--declarations' option tags \"(defvar foo)\" constructs too.";
665 static const char *Lua_suffixes
[] =
666 { "lua", "LUA", NULL
};
667 static const char Lua_help
[] =
668 "In Lua scripts, all functions are tags.";
670 static const char *Makefile_filenames
[] =
671 { "Makefile", "makefile", "GNUMakefile", "Makefile.in", "Makefile.am", NULL
};
672 static const char Makefile_help
[] =
673 "In makefiles, targets are tags; additionally, variables are tags\n\
674 unless you specify '--no-globals'.";
676 static const char *Objc_suffixes
[] =
677 { "lm", /* Objective lex file */
678 "m", /* Objective C file */
680 static const char Objc_help
[] =
681 "In Objective C code, tags include Objective C definitions for classes,\n\
682 class categories, methods and protocols. Tags for variables and\n\
683 functions in classes are named 'CLASS::VARIABLE' and 'CLASS::FUNCTION'.\n\
684 (Use --help --lang=c --lang=objc --lang=java for full help.)";
686 static const char *Pascal_suffixes
[] =
687 { "p", "pas", NULL
};
688 static const char Pascal_help
[] =
689 "In Pascal code, the tags are the functions and procedures defined\n\
691 /* " // this is for working around an Emacs highlighting bug... */
693 static const char *Perl_suffixes
[] =
694 { "pl", "pm", NULL
};
695 static const char *Perl_interpreters
[] =
696 { "perl", "@PERL@", NULL
};
697 static const char Perl_help
[] =
698 "In Perl code, the tags are the packages, subroutines and variables\n\
699 defined by the 'package', 'sub', 'my' and 'local' keywords. Use\n\
700 '--globals' if you want to tag global variables. Tags for\n\
701 subroutines are named 'PACKAGE::SUB'. The name for subroutines\n\
702 defined in the default package is 'main::SUB'.";
704 static const char *PHP_suffixes
[] =
705 { "php", "php3", "php4", NULL
};
706 static const char PHP_help
[] =
707 "In PHP code, tags are functions, classes and defines. Unless you use\n\
708 the '--no-members' option, vars are tags too.";
710 static const char *plain_C_suffixes
[] =
711 { "pc", /* Pro*C file */
714 static const char *PS_suffixes
[] =
715 { "ps", "psw", NULL
}; /* .psw is for PSWrap */
716 static const char PS_help
[] =
717 "In PostScript code, the tags are the functions.";
719 static const char *Prolog_suffixes
[] =
721 static const char Prolog_help
[] =
722 "In Prolog code, tags are predicates and rules at the beginning of\n\
725 static const char *Python_suffixes
[] =
727 static const char Python_help
[] =
728 "In Python code, 'def' or 'class' at the beginning of a line\n\
731 static const char *Ruby_suffixes
[] =
732 { "rb", "ruby", NULL
};
733 static const char Ruby_help
[] =
734 "In Ruby code, 'def' or 'class' or 'module' at the beginning of\n\
735 a line generate a tag. Constants also generate a tag.";
737 /* Can't do the `SCM' or `scm' prefix with a version number. */
738 static const char *Scheme_suffixes
[] =
739 { "oak", "sch", "scheme", "SCM", "scm", "SM", "sm", "ss", "t", NULL
};
740 static const char Scheme_help
[] =
741 "In Scheme code, tags include anything defined with 'def' or with a\n\
742 construct whose name starts with 'def'. They also include\n\
743 variables set with 'set!' at top level in the file.";
745 static const char *TeX_suffixes
[] =
746 { "bib", "clo", "cls", "ltx", "sty", "TeX", "tex", NULL
};
747 static const char TeX_help
[] =
748 "In LaTeX text, the argument of any of the commands '\\chapter',\n\
749 '\\section', '\\subsection', '\\subsubsection', '\\eqno', '\\label',\n\
750 '\\ref', '\\cite', '\\bibitem', '\\part', '\\appendix', '\\entry',\n\
751 '\\index', '\\def', '\\newcommand', '\\renewcommand',\n\
752 '\\newenvironment' or '\\renewenvironment' is a tag.\n\
754 Other commands can be specified by setting the environment variable\n\
755 'TEXTAGS' to a colon-separated list like, for example,\n\
756 TEXTAGS=\"mycommand:myothercommand\".";
759 static const char *Texinfo_suffixes
[] =
760 { "texi", "texinfo", "txi", NULL
};
761 static const char Texinfo_help
[] =
762 "for texinfo files, lines starting with @node are tagged.";
764 static const char *Yacc_suffixes
[] =
765 { "y", "y++", "ym", "yxx", "yy", NULL
}; /* .ym is Objective yacc file */
766 static const char Yacc_help
[] =
767 "In Bison or Yacc input files, each rule defines as a tag the\n\
768 nonterminal it constructs. The portions of the file that contain\n\
769 C code are parsed as C code (use --help --lang=c --lang=yacc\n\
772 static const char auto_help
[] =
773 "'auto' is not a real language, it indicates to use\n\
774 a default language for files base on file name suffix and file contents.";
776 static const char none_help
[] =
777 "'none' is not a real language, it indicates to only do\n\
778 regexp processing on files.";
780 static const char no_lang_help
[] =
781 "No detailed help available for this language.";
785 * Table of languages.
787 * It is ok for a given function to be listed under more than one
788 * name. I just didn't.
791 static language lang_names
[] =
793 { "ada", Ada_help
, Ada_funcs
, Ada_suffixes
},
794 { "asm", Asm_help
, Asm_labels
, Asm_suffixes
},
795 { "c", default_C_help
, default_C_entries
, default_C_suffixes
},
796 { "c++", Cplusplus_help
, Cplusplus_entries
, Cplusplus_suffixes
},
797 { "c*", no_lang_help
, Cstar_entries
, Cstar_suffixes
},
798 { "cobol", Cobol_help
, Cobol_paragraphs
, Cobol_suffixes
},
799 { "erlang", Erlang_help
, Erlang_functions
, Erlang_suffixes
},
800 { "forth", Forth_help
, Forth_words
, Forth_suffixes
},
801 { "fortran", Fortran_help
, Fortran_functions
, Fortran_suffixes
},
802 { "go", Go_help
, Go_functions
, Go_suffixes
},
803 { "html", HTML_help
, HTML_labels
, HTML_suffixes
},
804 { "java", Cjava_help
, Cjava_entries
, Cjava_suffixes
},
805 { "lisp", Lisp_help
, Lisp_functions
, Lisp_suffixes
},
806 { "lua", Lua_help
, Lua_functions
, Lua_suffixes
},
807 { "makefile", Makefile_help
,Makefile_targets
,NULL
,Makefile_filenames
},
808 { "objc", Objc_help
, plain_C_entries
, Objc_suffixes
},
809 { "pascal", Pascal_help
, Pascal_functions
, Pascal_suffixes
},
810 { "perl",Perl_help
,Perl_functions
,Perl_suffixes
,NULL
,Perl_interpreters
},
811 { "php", PHP_help
, PHP_functions
, PHP_suffixes
},
812 { "postscript",PS_help
, PS_functions
, PS_suffixes
},
813 { "proc", no_lang_help
, plain_C_entries
, plain_C_suffixes
},
814 { "prolog", Prolog_help
, Prolog_functions
, Prolog_suffixes
},
815 { "python", Python_help
, Python_functions
, Python_suffixes
},
816 { "ruby", Ruby_help
, Ruby_functions
, Ruby_suffixes
},
817 { "scheme", Scheme_help
, Scheme_functions
, Scheme_suffixes
},
818 { "tex", TeX_help
, TeX_commands
, TeX_suffixes
},
819 { "texinfo", Texinfo_help
, Texinfo_nodes
, Texinfo_suffixes
},
820 { "yacc", Yacc_help
,Yacc_entries
,Yacc_suffixes
,NULL
,NULL
,true},
821 { "auto", auto_help
}, /* default guessing scheme */
822 { "none", none_help
, just_read_file
}, /* regexp matching only */
823 { NULL
} /* end of list */
828 print_language_names (void)
831 const char **name
, **ext
;
833 puts ("\nThese are the currently supported languages, along with the\n\
834 default file names and dot suffixes:");
835 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
837 printf (" %-*s", 10, lang
->name
);
838 if (lang
->filenames
!= NULL
)
839 for (name
= lang
->filenames
; *name
!= NULL
; name
++)
840 printf (" %s", *name
);
841 if (lang
->suffixes
!= NULL
)
842 for (ext
= lang
->suffixes
; *ext
!= NULL
; ext
++)
843 printf (" .%s", *ext
);
846 puts ("where 'auto' means use default language for files based on file\n\
847 name suffix, and 'none' means only do regexp processing on files.\n\
848 If no language is specified and no matching suffix is found,\n\
849 the first line of the file is read for a sharp-bang (#!) sequence\n\
850 followed by the name of an interpreter. If no such sequence is found,\n\
851 Fortran is tried first; if no tags are found, C is tried next.\n\
852 When parsing any C file, a \"class\" or \"template\" keyword\n\
854 puts ("Compressed files are supported using gzip, bzip2, and xz.\n\
856 For detailed help on a given language use, for example,\n\
857 etags --help --lang=ada.");
861 # define EMACS_NAME "standalone"
864 # define VERSION "17.38.1.4"
866 static _Noreturn
void
869 char emacs_copyright
[] = COPYRIGHT
;
871 printf ("%s (%s %s)\n", (CTAGS
) ? "ctags" : "etags", EMACS_NAME
, VERSION
);
872 puts (emacs_copyright
);
873 puts ("This program is distributed under the terms in ETAGS.README");
878 #ifndef PRINT_UNDOCUMENTED_OPTIONS_HELP
879 # define PRINT_UNDOCUMENTED_OPTIONS_HELP false
882 static _Noreturn
void
883 print_help (argument
*argbuffer
)
885 bool help_for_lang
= false;
887 for (; argbuffer
->arg_type
!= at_end
; argbuffer
++)
888 if (argbuffer
->arg_type
== at_language
)
892 puts (argbuffer
->lang
->help
);
893 help_for_lang
= true;
899 printf ("Usage: %s [options] [[regex-option ...] file-name] ...\n\
901 These are the options accepted by %s.\n", progname
, progname
);
902 puts ("You may use unambiguous abbreviations for the long option names.");
903 puts (" A - as file name means read names from stdin (one per line).\n\
904 Absolute names are stored in the output file as they are.\n\
905 Relative ones are stored relative to the output file's directory.\n");
907 puts ("-a, --append\n\
908 Append tag entries to existing tags file.");
910 puts ("--packages-only\n\
911 For Ada files, only generate tags for packages.");
914 puts ("-B, --backward-search\n\
915 Write the search commands for the tag entries using '?', the\n\
916 backward-search command instead of '/', the forward-search command.");
918 /* This option is mostly obsolete, because etags can now automatically
919 detect C++. Retained for backward compatibility and for debugging and
920 experimentation. In principle, we could want to tag as C++ even
921 before any "class" or "template" keyword.
923 Treat files whose name suffix defaults to C language as C++ files.");
926 puts ("--declarations\n\
927 In C and derived languages, create tags for function declarations,");
929 puts ("\tand create tags for extern variables if --globals is used.");
932 ("\tand create tags for extern variables unless --no-globals is used.");
935 puts ("-d, --defines\n\
936 Create tag entries for C #define constants and enum constants, too.");
938 puts ("-D, --no-defines\n\
939 Don't create tag entries for C #define constants and enum constants.\n\
940 This makes the tags file smaller.");
943 puts ("-i FILE, --include=FILE\n\
944 Include a note in tag file indicating that, when searching for\n\
945 a tag, one should also consult the tags file FILE after\n\
946 checking the current file.");
948 puts ("-l LANG, --language=LANG\n\
949 Force the following files to be considered as written in the\n\
950 named language up to the next --language=LANG option.");
954 Create tag entries for global variables in some languages.");
956 puts ("--no-globals\n\
957 Do not create tag entries for global variables in some\n\
958 languages. This makes the tags file smaller.");
960 if (PRINT_UNDOCUMENTED_OPTIONS_HELP
)
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, and Java.\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\".");
978 puts ("-r REGEXP, --regex=REGEXP or --regex=@regexfile\n\
979 Make a tag for each line matching a regular expression pattern\n\
980 in the following files. {LANGUAGE}REGEXP uses REGEXP for LANGUAGE\n\
981 files only. REGEXFILE is a file containing one REGEXP per line.\n\
982 REGEXP takes the form /TAGREGEXP/TAGNAME/MODS, where TAGNAME/ is\n\
983 optional. The TAGREGEXP pattern is anchored (as if preceded by ^).");
984 puts (" If TAGNAME/ is present, the tags created are named.\n\
985 For example Tcl named tags can be created with:\n\
986 --regex=\"/proc[ \\t]+\\([^ \\t]+\\)/\\1/.\".\n\
987 MODS are optional one-letter modifiers: 'i' means to ignore case,\n\
988 'm' means to allow multi-line matches, 's' implies 'm' and\n\
989 causes dot to match any character, including newline.");
991 puts ("-R, --no-regex\n\
992 Don't create tags from regexps for the following files.");
994 puts ("-I, --ignore-indentation\n\
995 In C and C++ do not assume that a closing brace in the first\n\
996 column is the final brace of a function or structure definition.");
998 puts ("-o FILE, --output=FILE\n\
999 Write the tags to FILE.");
1001 puts ("--parse-stdin=NAME\n\
1002 Read from standard input and record tags as belonging to file NAME.");
1006 puts ("-t, --typedefs\n\
1007 Generate tag entries for C and Ada typedefs.");
1008 puts ("-T, --typedefs-and-c++\n\
1009 Generate tag entries for C typedefs, C struct/enum/union tags,\n\
1010 and C++ member functions.");
1014 puts ("-u, --update\n\
1015 Update the tag entries for the given files, leaving tag\n\
1016 entries for other files in place. Currently, this is\n\
1017 implemented by deleting the existing entries for the given\n\
1018 files and then rewriting the new entries at the end of the\n\
1019 tags file. It is often faster to simply rebuild the entire\n\
1020 tag file than to use this.");
1024 puts ("-v, --vgrind\n\
1025 Print on the standard output an index of items intended for\n\
1026 human consumption, similar to the output of vgrind. The index\n\
1027 is sorted, and gives the page number of each item.");
1029 if (PRINT_UNDOCUMENTED_OPTIONS_HELP
)
1030 puts ("-w, --no-duplicates\n\
1031 Do not create duplicate tag entries, for compatibility with\n\
1032 traditional ctags.");
1034 if (PRINT_UNDOCUMENTED_OPTIONS_HELP
)
1035 puts ("-w, --no-warn\n\
1036 Suppress warning messages about duplicate tag entries.");
1038 puts ("-x, --cxref\n\
1039 Like --vgrind, but in the style of cxref, rather than vgrind.\n\
1040 The output uses line numbers instead of page numbers, but\n\
1041 beyond that the differences are cosmetic; try both to see\n\
1045 puts ("-V, --version\n\
1046 Print the version of the program.\n\
1048 Print this help message.\n\
1049 Followed by one or more '--language' options prints detailed\n\
1050 help about tag generation for the specified languages.");
1052 print_language_names ();
1055 puts ("Report bugs to bug-gnu-emacs@gnu.org");
1057 exit (EXIT_SUCCESS
);
1062 main (int argc
, char **argv
)
1065 unsigned int nincluded_files
;
1066 char **included_files
;
1067 argument
*argbuffer
;
1068 int current_arg
, file_count
;
1069 linebuffer filename_lb
;
1070 bool help_asked
= false;
1076 nincluded_files
= 0;
1077 included_files
= xnew (argc
, char *);
1081 /* Allocate enough no matter what happens. Overkill, but each one
1083 argbuffer
= xnew (argc
, argument
);
1086 * Always find typedefs and structure tags.
1087 * Also default to find macro constants, enum constants, struct
1088 * members and global variables. Do it for both etags and ctags.
1090 typedefs
= typedefs_or_cplusplus
= constantypedefs
= true;
1091 globals
= members
= true;
1093 /* When the optstring begins with a '-' getopt_long does not rearrange the
1094 non-options arguments to be at the end, but leaves them alone. */
1095 optstring
= concat ("-ac:Cf:Il:o:Qr:RSVhH",
1096 (CTAGS
) ? "BxdtTuvw" : "Di:",
1099 while ((opt
= getopt_long (argc
, argv
, optstring
, longopts
, NULL
)) != EOF
)
1103 /* If getopt returns 0, then it has already processed a
1104 long-named option. We should do nothing. */
1108 /* This means that a file name has been seen. Record it. */
1109 argbuffer
[current_arg
].arg_type
= at_filename
;
1110 argbuffer
[current_arg
].what
= optarg
;
1111 len
= strlen (optarg
);
1112 if (whatlen_max
< len
)
1119 /* Parse standard input. Idea by Vivek <vivek@etla.org>. */
1120 argbuffer
[current_arg
].arg_type
= at_stdin
;
1121 argbuffer
[current_arg
].what
= optarg
;
1122 len
= strlen (optarg
);
1123 if (whatlen_max
< len
)
1128 fatal ("cannot parse standard input more than once");
1129 parsing_stdin
= true;
1132 /* Common options. */
1133 case 'a': append_to_tagfile
= true; break;
1134 case 'C': cplusplus
= true; break;
1135 case 'f': /* for compatibility with old makefiles */
1139 error ("-o option may only be given once.");
1140 suggest_asking_for_help ();
1146 case 'S': /* for backward compatibility */
1147 ignoreindent
= true;
1151 language
*lang
= get_language_from_langname (optarg
);
1154 argbuffer
[current_arg
].lang
= lang
;
1155 argbuffer
[current_arg
].arg_type
= at_language
;
1161 /* Backward compatibility: support obsolete --ignore-case-regexp. */
1162 optarg
= concat (optarg
, "i", ""); /* memory leak here */
1165 argbuffer
[current_arg
].arg_type
= at_regexp
;
1166 argbuffer
[current_arg
].what
= optarg
;
1167 len
= strlen (optarg
);
1168 if (whatlen_max
< len
)
1173 argbuffer
[current_arg
].arg_type
= at_regexp
;
1174 argbuffer
[current_arg
].what
= NULL
;
1189 case 'D': constantypedefs
= false; break;
1190 case 'i': included_files
[nincluded_files
++] = optarg
; break;
1192 /* Ctags options. */
1193 case 'B': searchar
= '?'; break;
1194 case 'd': constantypedefs
= true; break;
1195 case 't': typedefs
= true; break;
1196 case 'T': typedefs
= typedefs_or_cplusplus
= true; break;
1197 case 'u': update
= true; break;
1198 case 'v': vgrind_style
= true; /*FALLTHRU*/
1199 case 'x': cxref_style
= true; break;
1200 case 'w': no_warnings
= true; break;
1202 suggest_asking_for_help ();
1206 /* No more options. Store the rest of arguments. */
1207 for (; optind
< argc
; optind
++)
1209 argbuffer
[current_arg
].arg_type
= at_filename
;
1210 argbuffer
[current_arg
].what
= argv
[optind
];
1211 len
= strlen (argv
[optind
]);
1212 if (whatlen_max
< len
)
1218 argbuffer
[current_arg
].arg_type
= at_end
;
1221 print_help (argbuffer
);
1224 if (nincluded_files
== 0 && file_count
== 0)
1226 error ("no input files specified.");
1227 suggest_asking_for_help ();
1231 if (tagfile
== NULL
)
1232 tagfile
= savestr (CTAGS
? "tags" : "TAGS");
1233 cwd
= etags_getcwd (); /* the current working directory */
1234 if (cwd
[strlen (cwd
) - 1] != '/')
1237 cwd
= concat (oldcwd
, "/", "");
1241 /* Compute base directory for relative file names. */
1242 if (streq (tagfile
, "-")
1243 || strneq (tagfile
, "/dev/", 5))
1244 tagfiledir
= cwd
; /* relative file names are relative to cwd */
1247 canonicalize_filename (tagfile
);
1248 tagfiledir
= absolute_dirname (tagfile
, cwd
);
1251 linebuffer_init (&lb
);
1252 linebuffer_init (&filename_lb
);
1253 linebuffer_init (&filebuf
);
1254 linebuffer_init (&token_name
);
1258 if (streq (tagfile
, "-"))
1261 SET_BINARY (fileno (stdout
));
1264 tagf
= fopen (tagfile
, append_to_tagfile
? "ab" : "wb");
1270 * Loop through files finding functions.
1272 for (i
= 0; i
< current_arg
; i
++)
1274 static language
*lang
; /* non-NULL if language is forced */
1277 switch (argbuffer
[i
].arg_type
)
1280 lang
= argbuffer
[i
].lang
;
1283 analyze_regex (argbuffer
[i
].what
);
1286 this_file
= argbuffer
[i
].what
;
1287 /* Input file named "-" means read file names from stdin
1288 (one per line) and use them. */
1289 if (streq (this_file
, "-"))
1292 fatal ("cannot parse standard input "
1293 "AND read file names from it");
1294 while (readline_internal (&filename_lb
, stdin
, "-") > 0)
1295 process_file_name (filename_lb
.buffer
, lang
);
1298 process_file_name (this_file
, lang
);
1301 this_file
= argbuffer
[i
].what
;
1302 process_file (stdin
, this_file
, lang
);
1305 error ("internal error: arg_type");
1311 free (filebuf
.buffer
);
1312 free (token_name
.buffer
);
1314 if (!CTAGS
|| cxref_style
)
1316 /* Write the remaining tags to tagf (ETAGS) or stdout (CXREF). */
1317 put_entries (nodehead
);
1318 free_tree (nodehead
);
1324 /* Output file entries that have no tags. */
1325 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
1327 fprintf (tagf
, "\f\n%s,0\n", fdp
->taggedfname
);
1329 while (nincluded_files
-- > 0)
1330 fprintf (tagf
, "\f\n%s,include\n", *included_files
++);
1332 if (fclose (tagf
) == EOF
)
1336 exit (EXIT_SUCCESS
);
1339 /* From here on, we are in (CTAGS && !cxref_style) */
1343 xmalloc (strlen (tagfile
) + whatlen_max
+
1344 sizeof "mv..OTAGS;fgrep -v '\t\t' OTAGS >;rm OTAGS");
1345 for (i
= 0; i
< current_arg
; ++i
)
1347 switch (argbuffer
[i
].arg_type
)
1353 continue; /* the for loop */
1355 char *z
= stpcpy (cmd
, "mv ");
1356 z
= stpcpy (z
, tagfile
);
1357 z
= stpcpy (z
, " OTAGS;fgrep -v '\t");
1358 z
= stpcpy (z
, argbuffer
[i
].what
);
1359 z
= stpcpy (z
, "\t' OTAGS >");
1360 z
= stpcpy (z
, tagfile
);
1361 strcpy (z
, ";rm OTAGS");
1362 if (system (cmd
) != EXIT_SUCCESS
)
1363 fatal ("failed to execute shell command");
1366 append_to_tagfile
= true;
1369 tagf
= fopen (tagfile
, append_to_tagfile
? "ab" : "wb");
1372 put_entries (nodehead
); /* write all the tags (CTAGS) */
1373 free_tree (nodehead
);
1375 if (fclose (tagf
) == EOF
)
1379 if (append_to_tagfile
|| update
)
1381 char *cmd
= xmalloc (2 * strlen (tagfile
) + sizeof "sort -u -o..");
1382 /* Maybe these should be used:
1383 setenv ("LC_COLLATE", "C", 1);
1384 setenv ("LC_ALL", "C", 1); */
1385 char *z
= stpcpy (cmd
, "sort -u -o ");
1386 z
= stpcpy (z
, tagfile
);
1388 strcpy (z
, tagfile
);
1389 exit (system (cmd
));
1391 return EXIT_SUCCESS
;
1396 * Return a compressor given the file name. If EXTPTR is non-zero,
1397 * return a pointer into FILE where the compressor-specific
1398 * extension begins. If no compressor is found, NULL is returned
1399 * and EXTPTR is not significant.
1400 * Idea by Vladimir Alexiev <vladimir@cs.ualberta.ca> (1998)
1403 get_compressor_from_suffix (char *file
, char **extptr
)
1406 char *slash
, *suffix
;
1408 /* File has been processed by canonicalize_filename,
1409 so we don't need to consider backslashes on DOS_NT. */
1410 slash
= strrchr (file
, '/');
1411 suffix
= strrchr (file
, '.');
1412 if (suffix
== NULL
|| suffix
< slash
)
1417 /* Let those poor souls who live with DOS 8+3 file name limits get
1418 some solace by treating foo.cgz as if it were foo.c.gz, etc.
1419 Only the first do loop is run if not MSDOS */
1422 for (compr
= compressors
; compr
->suffix
!= NULL
; compr
++)
1423 if (streq (compr
->suffix
, suffix
))
1426 break; /* do it only once: not really a loop */
1429 } while (*suffix
!= '\0');
1436 * Return a language given the name.
1439 get_language_from_langname (const char *name
)
1444 error ("empty language name");
1447 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1448 if (streq (name
, lang
->name
))
1450 error ("unknown language \"%s\"", name
);
1458 * Return a language given the interpreter name.
1461 get_language_from_interpreter (char *interpreter
)
1466 if (interpreter
== NULL
)
1468 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1469 if (lang
->interpreters
!= NULL
)
1470 for (iname
= lang
->interpreters
; *iname
!= NULL
; iname
++)
1471 if (streq (*iname
, interpreter
))
1480 * Return a language given the file name.
1483 get_language_from_filename (char *file
, int case_sensitive
)
1486 const char **name
, **ext
, *suffix
;
1488 /* Try whole file name first. */
1489 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1490 if (lang
->filenames
!= NULL
)
1491 for (name
= lang
->filenames
; *name
!= NULL
; name
++)
1492 if ((case_sensitive
)
1493 ? streq (*name
, file
)
1494 : strcaseeq (*name
, file
))
1497 /* If not found, try suffix after last dot. */
1498 suffix
= strrchr (file
, '.');
1502 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1503 if (lang
->suffixes
!= NULL
)
1504 for (ext
= lang
->suffixes
; *ext
!= NULL
; ext
++)
1505 if ((case_sensitive
)
1506 ? streq (*ext
, suffix
)
1507 : strcaseeq (*ext
, suffix
))
1514 * This routine is called on each file argument.
1517 process_file_name (char *file
, language
*lang
)
1522 char *compressed_name
, *uncompressed_name
;
1523 char *ext
, *real_name
, *tmp_name
;
1526 canonicalize_filename (file
);
1527 if (streq (file
, tagfile
) && !streq (tagfile
, "-"))
1529 error ("skipping inclusion of %s in self.", file
);
1532 compr
= get_compressor_from_suffix (file
, &ext
);
1535 compressed_name
= file
;
1536 uncompressed_name
= savenstr (file
, ext
- file
);
1540 compressed_name
= NULL
;
1541 uncompressed_name
= file
;
1544 /* If the canonicalized uncompressed name
1545 has already been dealt with, skip it silently. */
1546 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
1548 assert (fdp
->infname
!= NULL
);
1549 if (streq (uncompressed_name
, fdp
->infname
))
1553 inf
= fopen (file
, "r" FOPEN_BINARY
);
1558 int file_errno
= errno
;
1559 if (compressed_name
)
1561 /* Try with the given suffix. */
1562 inf
= fopen (uncompressed_name
, "r" FOPEN_BINARY
);
1564 real_name
= uncompressed_name
;
1568 /* Try all possible suffixes. */
1569 for (compr
= compressors
; compr
->suffix
!= NULL
; compr
++)
1571 compressed_name
= concat (file
, ".", compr
->suffix
);
1572 inf
= fopen (compressed_name
, "r" FOPEN_BINARY
);
1575 real_name
= compressed_name
;
1580 char *suf
= compressed_name
+ strlen (file
);
1581 size_t suflen
= strlen (compr
->suffix
) + 1;
1582 for ( ; suf
[1]; suf
++, suflen
--)
1584 memmove (suf
, suf
+ 1, suflen
);
1585 inf
= fopen (compressed_name
, "r" FOPEN_BINARY
);
1588 real_name
= compressed_name
;
1595 free (compressed_name
);
1596 compressed_name
= NULL
;
1607 if (real_name
== compressed_name
)
1610 tmp_name
= etags_mktmp ();
1615 #if MSDOS || defined (DOS_NT)
1616 char *cmd1
= concat (compr
->command
, " \"", real_name
);
1617 char *cmd
= concat (cmd1
, "\" > ", tmp_name
);
1619 char *cmd1
= concat (compr
->command
, " '", real_name
);
1620 char *cmd
= concat (cmd1
, "' > ", tmp_name
);
1624 if (system (cmd
) == -1)
1631 inf
= fopen (tmp_name
, "r" FOPEN_BINARY
);
1645 process_file (inf
, uncompressed_name
, lang
);
1647 retval
= fclose (inf
);
1648 if (real_name
== compressed_name
)
1657 if (compressed_name
!= file
)
1658 free (compressed_name
);
1659 if (uncompressed_name
!= file
)
1660 free (uncompressed_name
);
1667 process_file (FILE *fh
, char *fn
, language
*lang
)
1669 static const fdesc emptyfdesc
;
1673 /* Create a new input file description entry. */
1674 fdp
= xnew (1, fdesc
);
1677 fdp
->infname
= savestr (fn
);
1679 fdp
->infabsname
= absolute_filename (fn
, cwd
);
1680 fdp
->infabsdir
= absolute_dirname (fn
, cwd
);
1681 if (filename_is_absolute (fn
))
1683 /* An absolute file name. Canonicalize it. */
1684 fdp
->taggedfname
= absolute_filename (fn
, NULL
);
1688 /* A file name relative to cwd. Make it relative
1689 to the directory of the tags file. */
1690 fdp
->taggedfname
= relative_filename (fn
, tagfiledir
);
1692 fdp
->usecharno
= true; /* use char position when making tags */
1694 fdp
->written
= false; /* not written on tags file yet */
1697 curfdp
= fdhead
; /* the current file description */
1701 /* If not Ctags, and if this is not metasource and if it contained no #line
1702 directives, we can write the tags and free all nodes pointing to
1705 && curfdp
->usecharno
/* no #line directives in this file */
1706 && !curfdp
->lang
->metasource
)
1710 /* Look for the head of the sublist relative to this file. See add_node
1711 for the structure of the node tree. */
1713 for (np
= nodehead
; np
!= NULL
; prev
= np
, np
= np
->left
)
1714 if (np
->fdp
== curfdp
)
1717 /* If we generated tags for this file, write and delete them. */
1720 /* This is the head of the last sublist, if any. The following
1721 instructions depend on this being true. */
1722 assert (np
->left
== NULL
);
1724 assert (fdhead
== curfdp
);
1725 assert (last_node
->fdp
== curfdp
);
1726 put_entries (np
); /* write tags for file curfdp->taggedfname */
1727 free_tree (np
); /* remove the written nodes */
1729 nodehead
= NULL
; /* no nodes left */
1731 prev
->left
= NULL
; /* delete the pointer to the sublist */
1737 reset_input (FILE *inf
)
1739 if (fseek (inf
, 0, SEEK_SET
) != 0)
1740 perror (infilename
);
1744 * This routine opens the specified file and calls the function
1745 * which finds the function and type definitions.
1748 find_entries (FILE *inf
)
1751 language
*lang
= curfdp
->lang
;
1752 Lang_function
*parser
= NULL
;
1754 /* If user specified a language, use it. */
1755 if (lang
!= NULL
&& lang
->function
!= NULL
)
1757 parser
= lang
->function
;
1760 /* Else try to guess the language given the file name. */
1763 lang
= get_language_from_filename (curfdp
->infname
, true);
1764 if (lang
!= NULL
&& lang
->function
!= NULL
)
1766 curfdp
->lang
= lang
;
1767 parser
= lang
->function
;
1771 /* Else look for sharp-bang as the first two characters. */
1773 && readline_internal (&lb
, inf
, infilename
) > 0
1775 && lb
.buffer
[0] == '#'
1776 && lb
.buffer
[1] == '!')
1780 /* Set lp to point at the first char after the last slash in the
1781 line or, if no slashes, at the first nonblank. Then set cp to
1782 the first successive blank and terminate the string. */
1783 lp
= strrchr (lb
.buffer
+2, '/');
1787 lp
= skip_spaces (lb
.buffer
+ 2);
1788 cp
= skip_non_spaces (lp
);
1791 if (strlen (lp
) > 0)
1793 lang
= get_language_from_interpreter (lp
);
1794 if (lang
!= NULL
&& lang
->function
!= NULL
)
1796 curfdp
->lang
= lang
;
1797 parser
= lang
->function
;
1804 /* Else try to guess the language given the case insensitive file name. */
1807 lang
= get_language_from_filename (curfdp
->infname
, false);
1808 if (lang
!= NULL
&& lang
->function
!= NULL
)
1810 curfdp
->lang
= lang
;
1811 parser
= lang
->function
;
1815 /* Else try Fortran or C. */
1818 node
*old_last_node
= last_node
;
1820 curfdp
->lang
= get_language_from_langname ("fortran");
1823 if (old_last_node
== last_node
)
1824 /* No Fortran entries found. Try C. */
1827 curfdp
->lang
= get_language_from_langname (cplusplus
? "c++" : "c");
1833 if (!no_line_directive
1834 && curfdp
->lang
!= NULL
&& curfdp
->lang
->metasource
)
1835 /* It may be that this is a bingo.y file, and we already parsed a bingo.c
1836 file, or anyway we parsed a file that is automatically generated from
1837 this one. If this is the case, the bingo.c file contained #line
1838 directives that generated tags pointing to this file. Let's delete
1839 them all before parsing this file, which is the real source. */
1841 fdesc
**fdpp
= &fdhead
;
1842 while (*fdpp
!= NULL
)
1844 && streq ((*fdpp
)->taggedfname
, curfdp
->taggedfname
))
1845 /* We found one of those! We must delete both the file description
1846 and all tags referring to it. */
1848 fdesc
*badfdp
= *fdpp
;
1850 /* Delete the tags referring to badfdp->taggedfname
1851 that were obtained from badfdp->infname. */
1852 invalidate_nodes (badfdp
, &nodehead
);
1854 *fdpp
= badfdp
->next
; /* remove the bad description from the list */
1855 free_fdesc (badfdp
);
1858 fdpp
= &(*fdpp
)->next
; /* advance the list pointer */
1861 assert (parser
!= NULL
);
1863 /* Generic initializations before reading from file. */
1864 linebuffer_setlen (&filebuf
, 0); /* reset the file buffer */
1866 /* Generic initializations before parsing file with readline. */
1867 lineno
= 0; /* reset global line number */
1868 charno
= 0; /* reset global char number */
1869 linecharno
= 0; /* reset global char number of line start */
1873 regex_tag_multiline ();
1878 * Check whether an implicitly named tag should be created,
1879 * then call `pfnote'.
1880 * NAME is a string that is internally copied by this function.
1882 * TAGS format specification
1883 * Idea by Sam Kendall <kendall@mv.mv.com> (1997)
1884 * The following is explained in some more detail in etc/ETAGS.EBNF.
1886 * make_tag creates tags with "implicit tag names" (unnamed tags)
1887 * if the following are all true, assuming NONAM=" \f\t\n\r()=,;":
1888 * 1. NAME does not contain any of the characters in NONAM;
1889 * 2. LINESTART contains name as either a rightmost, or rightmost but
1890 * one character, substring;
1891 * 3. the character, if any, immediately before NAME in LINESTART must
1892 * be a character in NONAM;
1893 * 4. the character, if any, immediately after NAME in LINESTART must
1894 * also be a character in NONAM.
1896 * The implementation uses the notinname() macro, which recognizes the
1897 * characters stored in the string `nonam'.
1898 * etags.el needs to use the same characters that are in NONAM.
1901 make_tag (const char *name
, /* tag name, or NULL if unnamed */
1902 int namelen
, /* tag length */
1903 bool is_func
, /* tag is a function */
1904 char *linestart
, /* start of the line where tag is */
1905 int linelen
, /* length of the line where tag is */
1906 int lno
, /* line number */
1907 long int cno
) /* character number */
1909 bool named
= (name
!= NULL
&& namelen
> 0);
1912 if (!CTAGS
&& named
) /* maybe set named to false */
1913 /* Let's try to make an implicit tag name, that is, create an unnamed tag
1914 such that etags.el can guess a name from it. */
1917 register const char *cp
= name
;
1919 for (i
= 0; i
< namelen
; i
++)
1920 if (notinname (*cp
++))
1922 if (i
== namelen
) /* rule #1 */
1924 cp
= linestart
+ linelen
- namelen
;
1925 if (notinname (linestart
[linelen
-1]))
1926 cp
-= 1; /* rule #4 */
1927 if (cp
>= linestart
/* rule #2 */
1929 || notinname (cp
[-1])) /* rule #3 */
1930 && strneq (name
, cp
, namelen
)) /* rule #2 */
1931 named
= false; /* use implicit tag name */
1936 nname
= savenstr (name
, namelen
);
1938 pfnote (nname
, is_func
, linestart
, linelen
, lno
, cno
);
1943 pfnote (char *name
, bool is_func
, char *linestart
, int linelen
, int lno
,
1945 /* tag name, or NULL if unnamed */
1946 /* tag is a function */
1947 /* start of the line where tag is */
1948 /* length of the line where tag is */
1950 /* character number */
1954 assert (name
== NULL
|| name
[0] != '\0');
1955 if (CTAGS
&& name
== NULL
)
1958 np
= xnew (1, node
);
1960 /* If ctags mode, change name "main" to M<thisfilename>. */
1961 if (CTAGS
&& !cxref_style
&& streq (name
, "main"))
1963 char *fp
= strrchr (curfdp
->taggedfname
, '/');
1964 np
->name
= concat ("M", fp
== NULL
? curfdp
->taggedfname
: fp
+ 1, "");
1965 fp
= strrchr (np
->name
, '.');
1966 if (fp
!= NULL
&& fp
[1] != '\0' && fp
[2] == '\0')
1972 np
->been_warned
= false;
1974 np
->is_func
= is_func
;
1976 if (np
->fdp
->usecharno
)
1977 /* Our char numbers are 0-base, because of C language tradition?
1978 ctags compatibility? old versions compatibility? I don't know.
1979 Anyway, since emacs's are 1-base we expect etags.el to take care
1980 of the difference. If we wanted to have 1-based numbers, we would
1981 uncomment the +1 below. */
1982 np
->cno
= cno
/* + 1 */ ;
1984 np
->cno
= invalidcharno
;
1985 np
->left
= np
->right
= NULL
;
1986 if (CTAGS
&& !cxref_style
)
1988 if (strlen (linestart
) < 50)
1989 np
->regex
= concat (linestart
, "$", "");
1991 np
->regex
= savenstr (linestart
, 50);
1994 np
->regex
= savenstr (linestart
, linelen
);
1996 add_node (np
, &nodehead
);
2001 * recurse on left children, iterate on right children.
2004 free_tree (register node
*np
)
2008 register node
*node_right
= np
->right
;
2009 free_tree (np
->left
);
2019 * delete a file description
2022 free_fdesc (register fdesc
*fdp
)
2024 free (fdp
->infname
);
2025 free (fdp
->infabsname
);
2026 free (fdp
->infabsdir
);
2027 free (fdp
->taggedfname
);
2034 * Adds a node to the tree of nodes. In etags mode, sort by file
2035 * name. In ctags mode, sort by tag name. Make no attempt at
2038 * add_node is the only function allowed to add nodes, so it can
2042 add_node (node
*np
, node
**cur_node_p
)
2045 register node
*cur_node
= *cur_node_p
;
2047 if (cur_node
== NULL
)
2057 /* For each file name, tags are in a linked sublist on the right
2058 pointer. The first tags of different files are a linked list
2059 on the left pointer. last_node points to the end of the last
2061 if (last_node
!= NULL
&& last_node
->fdp
== np
->fdp
)
2063 /* Let's use the same sublist as the last added node. */
2064 assert (last_node
->right
== NULL
);
2065 last_node
->right
= np
;
2068 else if (cur_node
->fdp
== np
->fdp
)
2070 /* Scanning the list we found the head of a sublist which is
2071 good for us. Let's scan this sublist. */
2072 add_node (np
, &cur_node
->right
);
2075 /* The head of this sublist is not good for us. Let's try the
2077 add_node (np
, &cur_node
->left
);
2078 } /* if ETAGS mode */
2083 dif
= strcmp (np
->name
, cur_node
->name
);
2086 * If this tag name matches an existing one, then
2087 * do not add the node, but maybe print a warning.
2089 if (no_duplicates
&& !dif
)
2091 if (np
->fdp
== cur_node
->fdp
)
2095 fprintf (stderr
, "Duplicate entry in file %s, line %d: %s\n",
2096 np
->fdp
->infname
, lineno
, np
->name
);
2097 fprintf (stderr
, "Second entry ignored\n");
2100 else if (!cur_node
->been_warned
&& !no_warnings
)
2104 "Duplicate entry in files %s and %s: %s (Warning only)\n",
2105 np
->fdp
->infname
, cur_node
->fdp
->infname
, np
->name
);
2106 cur_node
->been_warned
= true;
2111 /* Actually add the node */
2112 add_node (np
, dif
< 0 ? &cur_node
->left
: &cur_node
->right
);
2113 } /* if CTAGS mode */
2117 * invalidate_nodes ()
2118 * Scan the node tree and invalidate all nodes pointing to the
2119 * given file description (CTAGS case) or free them (ETAGS case).
2122 invalidate_nodes (fdesc
*badfdp
, node
**npp
)
2131 if (np
->left
!= NULL
)
2132 invalidate_nodes (badfdp
, &np
->left
);
2133 if (np
->fdp
== badfdp
)
2135 if (np
->right
!= NULL
)
2136 invalidate_nodes (badfdp
, &np
->right
);
2140 assert (np
->fdp
!= NULL
);
2141 if (np
->fdp
== badfdp
)
2143 *npp
= np
->left
; /* detach the sublist from the list */
2144 np
->left
= NULL
; /* isolate it */
2145 free_tree (np
); /* free it */
2146 invalidate_nodes (badfdp
, npp
);
2149 invalidate_nodes (badfdp
, &np
->left
);
2154 static int total_size_of_entries (node
*);
2155 static int number_len (long) ATTRIBUTE_CONST
;
2157 /* Length of a non-negative number's decimal representation. */
2159 number_len (long int num
)
2162 while ((num
/= 10) > 0)
2168 * Return total number of characters that put_entries will output for
2169 * the nodes in the linked list at the right of the specified node.
2170 * This count is irrelevant with etags.el since emacs 19.34 at least,
2171 * but is still supplied for backward compatibility.
2174 total_size_of_entries (register node
*np
)
2176 register int total
= 0;
2178 for (; np
!= NULL
; np
= np
->right
)
2181 total
+= strlen (np
->regex
) + 1; /* pat\177 */
2182 if (np
->name
!= NULL
)
2183 total
+= strlen (np
->name
) + 1; /* name\001 */
2184 total
+= number_len ((long) np
->lno
) + 1; /* lno, */
2185 if (np
->cno
!= invalidcharno
) /* cno */
2186 total
+= number_len (np
->cno
);
2187 total
+= 1; /* newline */
2194 put_entries (register node
*np
)
2197 static fdesc
*fdp
= NULL
;
2202 /* Output subentries that precede this one */
2204 put_entries (np
->left
);
2206 /* Output this entry */
2215 fprintf (tagf
, "\f\n%s,%d\n",
2216 fdp
->taggedfname
, total_size_of_entries (np
));
2217 fdp
->written
= true;
2219 fputs (np
->regex
, tagf
);
2220 fputc ('\177', tagf
);
2221 if (np
->name
!= NULL
)
2223 fputs (np
->name
, tagf
);
2224 fputc ('\001', tagf
);
2226 fprintf (tagf
, "%d,", np
->lno
);
2227 if (np
->cno
!= invalidcharno
)
2228 fprintf (tagf
, "%ld", np
->cno
);
2234 if (np
->name
== NULL
)
2235 error ("internal error: NULL name in ctags mode.");
2240 fprintf (stdout
, "%s %s %d\n",
2241 np
->name
, np
->fdp
->taggedfname
, (np
->lno
+ 63) / 64);
2243 fprintf (stdout
, "%-16s %3d %-16s %s\n",
2244 np
->name
, np
->lno
, np
->fdp
->taggedfname
, np
->regex
);
2248 fprintf (tagf
, "%s\t%s\t", np
->name
, np
->fdp
->taggedfname
);
2251 { /* function or #define macro with args */
2252 putc (searchar
, tagf
);
2255 for (sp
= np
->regex
; *sp
; sp
++)
2257 if (*sp
== '\\' || *sp
== searchar
)
2261 putc (searchar
, tagf
);
2264 { /* anything else; text pattern inadequate */
2265 fprintf (tagf
, "%d", np
->lno
);
2270 } /* if this node contains a valid tag */
2272 /* Output subentries that follow this one */
2273 put_entries (np
->right
);
2275 put_entries (np
->left
);
2280 #define C_EXT 0x00fff /* C extensions */
2281 #define C_PLAIN 0x00000 /* C */
2282 #define C_PLPL 0x00001 /* C++ */
2283 #define C_STAR 0x00003 /* C* */
2284 #define C_JAVA 0x00005 /* JAVA */
2285 #define C_AUTO 0x01000 /* C, but switch to C++ if `class' is met */
2286 #define YACC 0x10000 /* yacc file */
2289 * The C symbol tables.
2294 st_C_objprot
, st_C_objimpl
, st_C_objend
,
2296 st_C_ignore
, st_C_attribute
,
2299 st_C_class
, st_C_template
,
2300 st_C_struct
, st_C_extern
, st_C_enum
, st_C_define
, st_C_typedef
2303 /* Feed stuff between (but not including) %[ and %] lines to:
2309 struct C_stab_entry { char *name; int c_ext; enum sym_type type; }
2313 while, 0, st_C_ignore
2314 switch, 0, st_C_ignore
2315 return, 0, st_C_ignore
2316 __attribute__, 0, st_C_attribute
2317 GTY, 0, st_C_attribute
2318 @interface, 0, st_C_objprot
2319 @protocol, 0, st_C_objprot
2320 @implementation,0, st_C_objimpl
2321 @end, 0, st_C_objend
2322 import, (C_JAVA & ~C_PLPL), st_C_ignore
2323 package, (C_JAVA & ~C_PLPL), st_C_ignore
2324 friend, C_PLPL, st_C_ignore
2325 extends, (C_JAVA & ~C_PLPL), st_C_javastruct
2326 implements, (C_JAVA & ~C_PLPL), st_C_javastruct
2327 interface, (C_JAVA & ~C_PLPL), st_C_struct
2328 class, 0, st_C_class
2329 namespace, C_PLPL, st_C_struct
2330 domain, C_STAR, st_C_struct
2331 union, 0, st_C_struct
2332 struct, 0, st_C_struct
2333 extern, 0, st_C_extern
2335 typedef, 0, st_C_typedef
2336 define, 0, st_C_define
2337 undef, 0, st_C_define
2338 operator, C_PLPL, st_C_operator
2339 template, 0, st_C_template
2340 # DEFUN used in emacs, the next three used in glibc (SYSCALL only for mach).
2341 DEFUN, 0, st_C_gnumacro
2342 SYSCALL, 0, st_C_gnumacro
2343 ENTRY, 0, st_C_gnumacro
2344 PSEUDO, 0, st_C_gnumacro
2345 # These are defined inside C functions, so currently they are not met.
2346 # EXFUN used in glibc, DEFVAR_* in emacs.
2347 #EXFUN, 0, st_C_gnumacro
2348 #DEFVAR_, 0, st_C_gnumacro
2350 and replace lines between %< and %> with its output, then:
2351 - remove the #if characterset check
2352 - make in_word_set static and not inline. */
2354 /* C code produced by gperf version 3.0.1 */
2355 /* Command-line: gperf -m 5 */
2356 /* Computed positions: -k'2-3' */
2358 struct C_stab_entry
{ const char *name
; int c_ext
; enum sym_type type
; };
2359 /* maximum key range = 33, duplicates = 0 */
2362 hash (const char *str
, int len
)
2364 static char const asso_values
[] =
2366 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2367 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2368 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2369 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2370 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2371 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2372 35, 35, 35, 35, 35, 35, 35, 35, 35, 3,
2373 26, 35, 35, 35, 35, 35, 35, 35, 27, 35,
2374 35, 35, 35, 24, 0, 35, 35, 35, 35, 0,
2375 35, 35, 35, 35, 35, 1, 35, 16, 35, 6,
2376 23, 0, 0, 35, 22, 0, 35, 35, 5, 0,
2377 0, 15, 1, 35, 6, 35, 8, 19, 35, 16,
2378 4, 5, 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, 35,
2382 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2383 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2384 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2385 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2386 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2387 35, 35, 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
2398 hval
+= asso_values
[(unsigned char) str
[2]];
2401 hval
+= asso_values
[(unsigned char) str
[1]];
2407 static struct C_stab_entry
*
2408 in_word_set (register const char *str
, register unsigned int len
)
2412 TOTAL_KEYWORDS
= 33,
2413 MIN_WORD_LENGTH
= 2,
2414 MAX_WORD_LENGTH
= 15,
2419 static struct C_stab_entry wordlist
[] =
2422 {"if", 0, st_C_ignore
},
2423 {"GTY", 0, st_C_attribute
},
2424 {"@end", 0, st_C_objend
},
2425 {"union", 0, st_C_struct
},
2426 {"define", 0, st_C_define
},
2427 {"import", (C_JAVA
& ~C_PLPL
), st_C_ignore
},
2428 {"template", 0, st_C_template
},
2429 {"operator", C_PLPL
, st_C_operator
},
2430 {"@interface", 0, st_C_objprot
},
2431 {"implements", (C_JAVA
& ~C_PLPL
), st_C_javastruct
},
2432 {"friend", C_PLPL
, st_C_ignore
},
2433 {"typedef", 0, st_C_typedef
},
2434 {"return", 0, st_C_ignore
},
2435 {"@implementation",0, st_C_objimpl
},
2436 {"@protocol", 0, st_C_objprot
},
2437 {"interface", (C_JAVA
& ~C_PLPL
), st_C_struct
},
2438 {"extern", 0, st_C_extern
},
2439 {"extends", (C_JAVA
& ~C_PLPL
), st_C_javastruct
},
2440 {"struct", 0, st_C_struct
},
2441 {"domain", C_STAR
, st_C_struct
},
2442 {"switch", 0, st_C_ignore
},
2443 {"enum", 0, st_C_enum
},
2444 {"for", 0, st_C_ignore
},
2445 {"namespace", C_PLPL
, st_C_struct
},
2446 {"class", 0, st_C_class
},
2447 {"while", 0, st_C_ignore
},
2448 {"undef", 0, st_C_define
},
2449 {"package", (C_JAVA
& ~C_PLPL
), st_C_ignore
},
2450 {"__attribute__", 0, st_C_attribute
},
2451 {"SYSCALL", 0, st_C_gnumacro
},
2452 {"ENTRY", 0, st_C_gnumacro
},
2453 {"PSEUDO", 0, st_C_gnumacro
},
2454 {"DEFUN", 0, st_C_gnumacro
}
2457 if (len
<= MAX_WORD_LENGTH
&& len
>= MIN_WORD_LENGTH
)
2459 int key
= hash (str
, len
);
2461 if (key
<= MAX_HASH_VALUE
&& key
>= 0)
2463 const char *s
= wordlist
[key
].name
;
2465 if (*str
== *s
&& !strncmp (str
+ 1, s
+ 1, len
- 1) && s
[len
] == '\0')
2466 return &wordlist
[key
];
2473 static enum sym_type
2474 C_symtype (char *str
, int len
, int c_ext
)
2476 register struct C_stab_entry
*se
= in_word_set (str
, len
);
2478 if (se
== NULL
|| (se
->c_ext
&& !(c_ext
& se
->c_ext
)))
2485 * Ignoring __attribute__ ((list))
2487 static bool inattribute
; /* looking at an __attribute__ construct */
2490 * C functions and variables are recognized using a simple
2491 * finite automaton. fvdef is its state variable.
2495 fvnone
, /* nothing seen */
2496 fdefunkey
, /* Emacs DEFUN keyword seen */
2497 fdefunname
, /* Emacs DEFUN name seen */
2498 foperator
, /* func: operator keyword seen (cplpl) */
2499 fvnameseen
, /* function or variable name seen */
2500 fstartlist
, /* func: just after open parenthesis */
2501 finlist
, /* func: in parameter list */
2502 flistseen
, /* func: after parameter list */
2503 fignore
, /* func: before open brace */
2504 vignore
/* var-like: ignore until ';' */
2507 static bool fvextern
; /* func or var: extern keyword seen; */
2510 * typedefs are recognized using a simple finite automaton.
2511 * typdef is its state variable.
2515 tnone
, /* nothing seen */
2516 tkeyseen
, /* typedef keyword seen */
2517 ttypeseen
, /* defined type seen */
2518 tinbody
, /* inside typedef body */
2519 tend
, /* just before typedef tag */
2520 tignore
/* junk after typedef tag */
2524 * struct-like structures (enum, struct and union) are recognized
2525 * using another simple finite automaton. `structdef' is its state
2530 snone
, /* nothing seen yet,
2531 or in struct body if bracelev > 0 */
2532 skeyseen
, /* struct-like keyword seen */
2533 stagseen
, /* struct-like tag seen */
2534 scolonseen
/* colon seen after struct-like tag */
2538 * When objdef is different from onone, objtag is the name of the class.
2540 static const char *objtag
= "<uninited>";
2543 * Yet another little state machine to deal with preprocessor lines.
2547 dnone
, /* nothing seen */
2548 dsharpseen
, /* '#' seen as first char on line */
2549 ddefineseen
, /* '#' and 'define' seen */
2550 dignorerest
/* ignore rest of line */
2554 * State machine for Objective C protocols and implementations.
2555 * Idea by Tom R.Hageman <tom@basil.icce.rug.nl> (1995)
2559 onone
, /* nothing seen */
2560 oprotocol
, /* @interface or @protocol seen */
2561 oimplementation
, /* @implementations seen */
2562 otagseen
, /* class name seen */
2563 oparenseen
, /* parenthesis before category seen */
2564 ocatseen
, /* category name seen */
2565 oinbody
, /* in @implementation body */
2566 omethodsign
, /* in @implementation body, after +/- */
2567 omethodtag
, /* after method name */
2568 omethodcolon
, /* after method colon */
2569 omethodparm
, /* after method parameter */
2570 oignore
/* wait for @end */
2575 * Use this structure to keep info about the token read, and how it
2576 * should be tagged. Used by the make_C_tag function to build a tag.
2580 char *line
; /* string containing the token */
2581 int offset
; /* where the token starts in LINE */
2582 int length
; /* token length */
2584 The previous members can be used to pass strings around for generic
2585 purposes. The following ones specifically refer to creating tags. In this
2586 case the token contained here is the pattern that will be used to create a
2589 bool valid
; /* do not create a tag; the token should be
2590 invalidated whenever a state machine is
2591 reset prematurely */
2592 bool named
; /* create a named tag */
2593 int lineno
; /* source line number of tag */
2594 long linepos
; /* source char number of tag */
2595 } token
; /* latest token read */
2598 * Variables and functions for dealing with nested structures.
2599 * Idea by Mykola Dzyuba <mdzyuba@yahoo.com> (2001)
2601 static void pushclass_above (int, char *, int);
2602 static void popclass_above (int);
2603 static void write_classname (linebuffer
*, const char *qualifier
);
2606 char **cname
; /* nested class names */
2607 int *bracelev
; /* nested class brace level */
2608 int nl
; /* class nesting level (elements used) */
2609 int size
; /* length of the array */
2610 } cstack
; /* stack for nested declaration tags */
2611 /* Current struct nesting depth (namespace, class, struct, union, enum). */
2612 #define nestlev (cstack.nl)
2613 /* After struct keyword or in struct body, not inside a nested function. */
2614 #define instruct (structdef == snone && nestlev > 0 \
2615 && bracelev == cstack.bracelev[nestlev-1] + 1)
2618 pushclass_above (int bracelev
, char *str
, int len
)
2622 popclass_above (bracelev
);
2624 if (nl
>= cstack
.size
)
2626 int size
= cstack
.size
*= 2;
2627 xrnew (cstack
.cname
, size
, char *);
2628 xrnew (cstack
.bracelev
, size
, int);
2630 assert (nl
== 0 || cstack
.bracelev
[nl
-1] < bracelev
);
2631 cstack
.cname
[nl
] = (str
== NULL
) ? NULL
: savenstr (str
, len
);
2632 cstack
.bracelev
[nl
] = bracelev
;
2637 popclass_above (int bracelev
)
2641 for (nl
= cstack
.nl
- 1;
2642 nl
>= 0 && cstack
.bracelev
[nl
] >= bracelev
;
2645 free (cstack
.cname
[nl
]);
2651 write_classname (linebuffer
*cn
, const char *qualifier
)
2654 int qlen
= strlen (qualifier
);
2656 if (cstack
.nl
== 0 || cstack
.cname
[0] == NULL
)
2660 cn
->buffer
[0] = '\0';
2664 len
= strlen (cstack
.cname
[0]);
2665 linebuffer_setlen (cn
, len
);
2666 strcpy (cn
->buffer
, cstack
.cname
[0]);
2668 for (i
= 1; i
< cstack
.nl
; i
++)
2670 char *s
= cstack
.cname
[i
];
2673 linebuffer_setlen (cn
, len
+ qlen
+ strlen (s
));
2674 len
+= sprintf (cn
->buffer
+ len
, "%s%s", qualifier
, s
);
2679 static bool consider_token (char *, int, int, int *, int, int, bool *);
2680 static void make_C_tag (bool);
2684 * checks to see if the current token is at the start of a
2685 * function or variable, or corresponds to a typedef, or
2686 * is a struct/union/enum tag, or #define, or an enum constant.
2688 * *IS_FUNC_OR_VAR gets true if the token is a function or #define macro
2689 * with args. C_EXTP points to which language we are looking at.
2700 consider_token (char *str
, int len
, int c
, int *c_extp
,
2701 int bracelev
, int parlev
, bool *is_func_or_var
)
2702 /* IN: token pointer */
2703 /* IN: token length */
2704 /* IN: first char after the token */
2705 /* IN, OUT: C extensions mask */
2706 /* IN: brace level */
2707 /* IN: parenthesis level */
2708 /* OUT: function or variable found */
2710 /* When structdef is stagseen, scolonseen, or snone with bracelev > 0,
2711 structtype is the type of the preceding struct-like keyword, and
2712 structbracelev is the brace level where it has been seen. */
2713 static enum sym_type structtype
;
2714 static int structbracelev
;
2715 static enum sym_type toktype
;
2718 toktype
= C_symtype (str
, len
, *c_extp
);
2721 * Skip __attribute__
2723 if (toktype
== st_C_attribute
)
2730 * Advance the definedef state machine.
2735 /* We're not on a preprocessor line. */
2736 if (toktype
== st_C_gnumacro
)
2743 if (toktype
== st_C_define
)
2745 definedef
= ddefineseen
;
2749 definedef
= dignorerest
;
2754 * Make a tag for any macro, unless it is a constant
2755 * and constantypedefs is false.
2757 definedef
= dignorerest
;
2758 *is_func_or_var
= (c
== '(');
2759 if (!*is_func_or_var
&& !constantypedefs
)
2766 error ("internal error: definedef value.");
2775 if (toktype
== st_C_typedef
)
2798 if (structdef
== snone
&& fvdef
== fvnone
)
2820 case st_C_javastruct
:
2821 if (structdef
== stagseen
)
2822 structdef
= scolonseen
;
2826 if ((*c_extp
& C_AUTO
) /* automatic detection of C++ language */
2828 && definedef
== dnone
&& structdef
== snone
2829 && typdef
== tnone
&& fvdef
== fvnone
)
2830 *c_extp
= (*c_extp
| C_PLPL
) & ~C_AUTO
;
2831 if (toktype
== st_C_template
)
2838 && (typdef
== tkeyseen
2839 || (typedefs_or_cplusplus
&& structdef
== snone
)))
2841 structdef
= skeyseen
;
2842 structtype
= toktype
;
2843 structbracelev
= bracelev
;
2844 if (fvdef
== fvnameseen
)
2852 if (structdef
== skeyseen
)
2854 structdef
= stagseen
;
2858 if (typdef
!= tnone
)
2861 /* Detect Objective C constructs. */
2871 objdef
= oimplementation
;
2877 case oimplementation
:
2878 /* Save the class tag for functions or variables defined inside. */
2879 objtag
= savenstr (str
, len
);
2883 /* Save the class tag for categories. */
2884 objtag
= savenstr (str
, len
);
2886 *is_func_or_var
= true;
2890 *is_func_or_var
= true;
2898 objdef
= omethodtag
;
2899 linebuffer_setlen (&token_name
, len
);
2900 memcpy (token_name
.buffer
, str
, len
);
2901 token_name
.buffer
[len
] = '\0';
2907 objdef
= omethodparm
;
2912 objdef
= omethodtag
;
2915 int oldlen
= token_name
.len
;
2917 linebuffer_setlen (&token_name
, oldlen
+ len
);
2918 memcpy (token_name
.buffer
+ oldlen
, str
, len
);
2919 token_name
.buffer
[oldlen
+ len
] = '\0';
2925 if (toktype
== st_C_objend
)
2927 /* Memory leakage here: the string pointed by objtag is
2928 never released, because many tests would be needed to
2929 avoid breaking on incorrect input code. The amount of
2930 memory leaked here is the sum of the lengths of the
2940 /* A function, variable or enum constant? */
2962 *is_func_or_var
= true;
2966 && structdef
== snone
2967 && structtype
== st_C_enum
&& bracelev
> structbracelev
2968 /* Don't tag tokens in expressions that assign values to enum
2970 && fvdef
!= vignore
)
2971 return true; /* enum constant */
2977 fvdef
= fdefunname
; /* GNU macro */
2978 *is_func_or_var
= true;
2986 if ((strneq (str
, "asm", 3) && endtoken (str
[3]))
2987 || (strneq (str
, "__asm__", 7) && endtoken (str
[7])))
2998 if (len
>= 10 && strneq (str
+len
-10, "::operator", 10))
3000 if (*c_extp
& C_AUTO
) /* automatic detection of C++ */
3001 *c_extp
= (*c_extp
| C_PLPL
) & ~C_AUTO
;
3003 *is_func_or_var
= true;
3006 if (bracelev
> 0 && !instruct
)
3008 fvdef
= fvnameseen
; /* function or variable */
3009 *is_func_or_var
= true;
3024 * C_entries often keeps pointers to tokens or lines which are older than
3025 * the line currently read. By keeping two line buffers, and switching
3026 * them at end of line, it is possible to use those pointers.
3034 #define current_lb_is_new (newndx == curndx)
3035 #define switch_line_buffers() (curndx = 1 - curndx)
3037 #define curlb (lbs[curndx].lb)
3038 #define newlb (lbs[newndx].lb)
3039 #define curlinepos (lbs[curndx].linepos)
3040 #define newlinepos (lbs[newndx].linepos)
3042 #define plainc ((c_ext & C_EXT) == C_PLAIN)
3043 #define cplpl (c_ext & C_PLPL)
3044 #define cjava ((c_ext & C_JAVA) == C_JAVA)
3046 #define CNL_SAVE_DEFINEDEF() \
3048 curlinepos = charno; \
3049 readline (&curlb, inf); \
3050 lp = curlb.buffer; \
3057 CNL_SAVE_DEFINEDEF (); \
3058 if (savetoken.valid) \
3060 token = savetoken; \
3061 savetoken.valid = false; \
3063 definedef = dnone; \
3068 make_C_tag (bool isfun
)
3070 /* This function is never called when token.valid is false, but
3071 we must protect against invalid input or internal errors. */
3073 make_tag (token_name
.buffer
, token_name
.len
, isfun
, token
.line
,
3074 token
.offset
+token
.length
+1, token
.lineno
, token
.linepos
);
3076 { /* this branch is optimized away if !DEBUG */
3077 make_tag (concat ("INVALID TOKEN:-->", token_name
.buffer
, ""),
3078 token_name
.len
+ 17, isfun
, token
.line
,
3079 token
.offset
+token
.length
+1, token
.lineno
, token
.linepos
);
3080 error ("INVALID TOKEN");
3083 token
.valid
= false;
3087 perhaps_more_input (FILE *inf
)
3089 return !feof (inf
) && !ferror (inf
);
3095 * This routine finds functions, variables, typedefs,
3096 * #define's, enum constants and struct/union/enum definitions in
3097 * C syntax and adds them to the list.
3100 C_entries (int c_ext
, FILE *inf
)
3101 /* extension of C */
3104 register char c
; /* latest char read; '\0' for end of line */
3105 register char *lp
; /* pointer one beyond the character `c' */
3106 int curndx
, newndx
; /* indices for current and new lb */
3107 register int tokoff
; /* offset in line of start of current token */
3108 register int toklen
; /* length of current token */
3109 const char *qualifier
; /* string used to qualify names */
3110 int qlen
; /* length of qualifier */
3111 int bracelev
; /* current brace level */
3112 int bracketlev
; /* current bracket level */
3113 int parlev
; /* current parenthesis level */
3114 int attrparlev
; /* __attribute__ parenthesis level */
3115 int templatelev
; /* current template level */
3116 int typdefbracelev
; /* bracelev where a typedef struct body begun */
3117 bool incomm
, inquote
, inchar
, quotednl
, midtoken
;
3118 bool yacc_rules
; /* in the rules part of a yacc file */
3119 struct tok savetoken
= {0}; /* token saved during preprocessor handling */
3122 linebuffer_init (&lbs
[0].lb
);
3123 linebuffer_init (&lbs
[1].lb
);
3124 if (cstack
.size
== 0)
3126 cstack
.size
= (DEBUG
) ? 1 : 4;
3128 cstack
.cname
= xnew (cstack
.size
, char *);
3129 cstack
.bracelev
= xnew (cstack
.size
, int);
3132 tokoff
= toklen
= typdefbracelev
= 0; /* keep compiler quiet */
3133 curndx
= newndx
= 0;
3137 fvdef
= fvnone
; fvextern
= false; typdef
= tnone
;
3138 structdef
= snone
; definedef
= dnone
; objdef
= onone
;
3140 midtoken
= inquote
= inchar
= incomm
= quotednl
= false;
3141 token
.valid
= savetoken
.valid
= false;
3142 bracelev
= bracketlev
= parlev
= attrparlev
= templatelev
= 0;
3144 { qualifier
= "."; qlen
= 1; }
3146 { qualifier
= "::"; qlen
= 2; }
3149 while (perhaps_more_input (inf
))
3154 /* If we are at the end of the line, the next character is a
3155 '\0'; do not skip it, because it is what tells us
3156 to read the next line. */
3177 /* Newlines inside comments do not end macro definitions in
3179 CNL_SAVE_DEFINEDEF ();
3192 /* Newlines inside strings do not end macro definitions
3193 in traditional cpp, even though compilers don't
3194 usually accept them. */
3195 CNL_SAVE_DEFINEDEF ();
3205 /* Hmmm, something went wrong. */
3241 if (fvdef
!= finlist
&& fvdef
!= fignore
&& fvdef
!= vignore
)
3256 else if (/* cplpl && */ *lp
== '/')
3262 if ((c_ext
& YACC
) && *lp
== '%')
3264 /* Entering or exiting rules section in yacc file. */
3266 definedef
= dnone
; fvdef
= fvnone
; fvextern
= false;
3267 typdef
= tnone
; structdef
= snone
;
3268 midtoken
= inquote
= inchar
= incomm
= quotednl
= false;
3270 yacc_rules
= !yacc_rules
;
3276 if (definedef
== dnone
)
3279 bool cpptoken
= true;
3281 /* Look back on this line. If all blanks, or nonblanks
3282 followed by an end of comment, this is a preprocessor
3284 for (cp
= newlb
.buffer
; cp
< lp
-1; cp
++)
3285 if (!c_isspace (*cp
))
3287 if (*cp
== '*' && cp
[1] == '/')
3297 definedef
= dsharpseen
;
3298 /* This is needed for tagging enum values: when there are
3299 preprocessor conditionals inside the enum, we need to
3300 reset the value of fvdef so that the next enum value is
3301 tagged even though the one before it did not end in a
3303 if (fvdef
== vignore
&& instruct
&& parlev
== 0)
3305 if (strneq (cp
, "#if", 3) || strneq (cp
, "#el", 3))
3309 } /* if (definedef == dnone) */
3320 CNL_SAVE_DEFINEDEF ();
3327 /* Consider token only if some involved conditions are satisfied. */
3328 if (typdef
!= tignore
3329 && definedef
!= dignorerest
3332 && (definedef
!= dnone
3333 || structdef
!= scolonseen
)
3340 if (c
== ':' && *lp
== ':' && begtoken (lp
[1]))
3341 /* This handles :: in the middle,
3342 but not at the beginning of an identifier.
3343 Also, space-separated :: is not recognized. */
3345 if (c_ext
& C_AUTO
) /* automatic detection of C++ */
3346 c_ext
= (c_ext
| C_PLPL
) & ~C_AUTO
;
3350 goto still_in_token
;
3354 bool funorvar
= false;
3357 || consider_token (newlb
.buffer
+ tokoff
, toklen
, c
,
3358 &c_ext
, bracelev
, parlev
,
3361 if (fvdef
== foperator
)
3364 lp
= skip_spaces (lp
-1);
3368 && !c_isspace (*lp
) && *lp
!= '(')
3371 toklen
+= lp
- oldlp
;
3373 token
.named
= false;
3375 && nestlev
> 0 && definedef
== dnone
)
3376 /* in struct body */
3381 write_classname (&token_name
, qualifier
);
3382 len
= token_name
.len
;
3383 linebuffer_setlen (&token_name
,
3384 len
+ qlen
+ toklen
);
3385 sprintf (token_name
.buffer
+ len
, "%s%.*s",
3387 newlb
.buffer
+ tokoff
);
3391 linebuffer_setlen (&token_name
, toklen
);
3392 sprintf (token_name
.buffer
, "%.*s",
3393 toklen
, newlb
.buffer
+ tokoff
);
3397 else if (objdef
== ocatseen
)
3398 /* Objective C category */
3402 int len
= strlen (objtag
) + 2 + toklen
;
3403 linebuffer_setlen (&token_name
, len
);
3404 sprintf (token_name
.buffer
, "%s(%.*s)",
3406 newlb
.buffer
+ tokoff
);
3410 linebuffer_setlen (&token_name
, toklen
);
3411 sprintf (token_name
.buffer
, "%.*s",
3412 toklen
, newlb
.buffer
+ tokoff
);
3416 else if (objdef
== omethodtag
3417 || objdef
== omethodparm
)
3418 /* Objective C method */
3422 else if (fvdef
== fdefunname
)
3423 /* GNU DEFUN and similar macros */
3425 bool defun
= (newlb
.buffer
[tokoff
] == 'F');
3429 /* Rewrite the tag so that emacs lisp DEFUNs
3430 can be found by their elisp name */
3436 linebuffer_setlen (&token_name
, len
);
3437 memcpy (token_name
.buffer
,
3438 newlb
.buffer
+ off
, len
);
3439 token_name
.buffer
[len
] = '\0';
3442 if (token_name
.buffer
[len
] == '_')
3443 token_name
.buffer
[len
] = '-';
3444 token
.named
= defun
;
3448 linebuffer_setlen (&token_name
, toklen
);
3449 memcpy (token_name
.buffer
,
3450 newlb
.buffer
+ tokoff
, toklen
);
3451 token_name
.buffer
[toklen
] = '\0';
3452 /* Name macros and members. */
3453 token
.named
= (structdef
== stagseen
3454 || typdef
== ttypeseen
3457 && definedef
== dignorerest
)
3459 && definedef
== dnone
3460 && structdef
== snone
3463 token
.lineno
= lineno
;
3464 token
.offset
= tokoff
;
3465 token
.length
= toklen
;
3466 token
.line
= newlb
.buffer
;
3467 token
.linepos
= newlinepos
;
3470 if (definedef
== dnone
3471 && (fvdef
== fvnameseen
3472 || fvdef
== foperator
3473 || structdef
== stagseen
3475 || typdef
== ttypeseen
3476 || objdef
!= onone
))
3478 if (current_lb_is_new
)
3479 switch_line_buffers ();
3481 else if (definedef
!= dnone
3482 || fvdef
== fdefunname
3484 make_C_tag (funorvar
);
3486 else /* not yacc and consider_token failed */
3488 if (inattribute
&& fvdef
== fignore
)
3490 /* We have just met __attribute__ after a
3491 function parameter list: do not tag the
3498 } /* if (endtoken (c)) */
3499 else if (intoken (c
))
3505 } /* if (midtoken) */
3506 else if (begtoken (c
))
3514 /* This prevents tagging fb in
3515 void (__attribute__((noreturn)) *fb) (void);
3516 Fixing this is not easy and not very important. */
3520 if (plainc
|| declarations
)
3522 make_C_tag (true); /* a function */
3529 if (structdef
== stagseen
&& !cjava
)
3531 popclass_above (bracelev
);
3541 if (!yacc_rules
|| lp
== newlb
.buffer
+ 1)
3543 tokoff
= lp
- 1 - newlb
.buffer
;
3548 } /* if (begtoken) */
3549 } /* if must look at token */
3552 /* Detect end of line, colon, comma, semicolon and various braces
3553 after having handled a token.*/
3559 if (yacc_rules
&& token
.offset
== 0 && token
.valid
)
3561 make_C_tag (false); /* a yacc function */
3564 if (definedef
!= dnone
)
3570 make_C_tag (true); /* an Objective C class */
3574 objdef
= omethodcolon
;
3577 int toklen
= token_name
.len
;
3578 linebuffer_setlen (&token_name
, toklen
+ 1);
3579 strcpy (token_name
.buffer
+ toklen
, ":");
3585 if (structdef
== stagseen
)
3587 structdef
= scolonseen
;
3590 /* Should be useless, but may be work as a safety net. */
3591 if (cplpl
&& fvdef
== flistseen
)
3593 make_C_tag (true); /* a function */
3599 if (definedef
!= dnone
|| inattribute
)
3605 make_C_tag (false); /* a typedef */
3615 if (typdef
== tignore
|| cplpl
)
3619 if ((globals
&& bracelev
== 0 && (!fvextern
|| declarations
))
3620 || (members
&& instruct
))
3621 make_C_tag (false); /* a variable */
3624 token
.valid
= false;
3628 && (cplpl
|| !instruct
)
3629 && (typdef
== tnone
|| (typdef
!= tignore
&& instruct
)))
3631 && plainc
&& instruct
))
3632 make_C_tag (true); /* a function */
3638 && cplpl
&& structdef
== stagseen
)
3639 make_C_tag (false); /* forward declaration */
3641 token
.valid
= false;
3642 } /* switch (fvdef) */
3648 if (structdef
== stagseen
)
3652 if (definedef
!= dnone
|| inattribute
)
3658 make_C_tag (true); /* an Objective C method */
3673 if (instruct
&& parlev
== 0)
3684 && (!fvextern
|| declarations
))
3685 || (members
&& instruct
)))
3686 make_C_tag (false); /* a variable */
3689 if ((declarations
&& typdef
== tnone
&& !instruct
)
3690 || (members
&& typdef
!= tignore
&& instruct
))
3692 make_C_tag (true); /* a function */
3695 else if (!declarations
)
3697 token
.valid
= false;
3702 if (structdef
== stagseen
)
3706 if (definedef
!= dnone
|| inattribute
)
3708 if (structdef
== stagseen
)
3715 make_C_tag (false); /* a typedef */
3727 if ((members
&& bracelev
== 1)
3728 || (globals
&& bracelev
== 0
3729 && (!fvextern
|| declarations
)))
3730 make_C_tag (false); /* a variable */
3746 if (definedef
!= dnone
)
3748 if (objdef
== otagseen
&& parlev
== 0)
3749 objdef
= oparenseen
;
3753 if (typdef
== ttypeseen
3757 /* This handles constructs like:
3758 typedef void OperatorFun (int fun); */
3779 if (--attrparlev
== 0)
3780 inattribute
= false;
3783 if (definedef
!= dnone
)
3785 if (objdef
== ocatseen
&& parlev
== 1)
3787 make_C_tag (true); /* an Objective C category */
3803 || typdef
== ttypeseen
))
3806 make_C_tag (false); /* a typedef */
3809 else if (parlev
< 0) /* can happen due to ill-conceived #if's. */
3813 if (definedef
!= dnone
)
3815 if (typdef
== ttypeseen
)
3817 /* Whenever typdef is set to tinbody (currently only
3818 here), typdefbracelev should be set to bracelev. */
3820 typdefbracelev
= bracelev
;
3825 if (cplpl
&& !class_qualify
)
3827 /* Remove class and namespace qualifiers from the token,
3828 leaving only the method/member name. */
3829 char *cc
, *uqname
= token_name
.buffer
;
3830 char *tok_end
= token_name
.buffer
+ token_name
.len
;
3832 for (cc
= token_name
.buffer
; cc
< tok_end
; cc
++)
3834 if (*cc
== ':' && cc
[1] == ':')
3840 if (uqname
> token_name
.buffer
)
3842 int uqlen
= strlen (uqname
);
3843 linebuffer_setlen (&token_name
, uqlen
);
3844 memmove (token_name
.buffer
, uqname
, uqlen
+ 1);
3847 make_C_tag (true); /* a function */
3856 make_C_tag (true); /* an Objective C class */
3861 make_C_tag (true); /* an Objective C method */
3865 /* Neutralize `extern "C" {' grot. */
3866 if (bracelev
== 0 && structdef
== snone
&& nestlev
== 0
3876 case skeyseen
: /* unnamed struct */
3877 pushclass_above (bracelev
, NULL
, 0);
3880 case stagseen
: /* named struct or enum */
3881 case scolonseen
: /* a class */
3882 pushclass_above (bracelev
,token
.line
+token
.offset
, token
.length
);
3884 make_C_tag (false); /* a struct or enum */
3892 if (definedef
!= dnone
)
3894 if (fvdef
== fstartlist
)
3896 fvdef
= fvnone
; /* avoid tagging `foo' in `foo (*bar()) ()' */
3897 token
.valid
= false;
3901 if (definedef
!= dnone
)
3904 if (!ignoreindent
&& lp
== newlb
.buffer
+ 1)
3907 token
.valid
= false; /* unexpected value, token unreliable */
3908 bracelev
= 0; /* reset brace level if first column */
3909 parlev
= 0; /* also reset paren level, just in case... */
3911 else if (bracelev
< 0)
3913 token
.valid
= false; /* something gone amiss, token unreliable */
3916 if (bracelev
== 0 && fvdef
== vignore
)
3917 fvdef
= fvnone
; /* end of function */
3918 popclass_above (bracelev
);
3920 /* Only if typdef == tinbody is typdefbracelev significant. */
3921 if (typdef
== tinbody
&& bracelev
<= typdefbracelev
)
3923 assert (bracelev
== typdefbracelev
);
3928 if (definedef
!= dnone
)
3938 if ((members
&& bracelev
== 1)
3939 || (globals
&& bracelev
== 0 && (!fvextern
|| declarations
)))
3940 make_C_tag (false); /* a variable */
3948 && (structdef
== stagseen
|| fvdef
== fvnameseen
))
3955 if (templatelev
> 0)
3963 if (objdef
== oinbody
&& bracelev
== 0)
3965 objdef
= omethodsign
;
3970 case '#': case '~': case '&': case '%': case '/':
3971 case '|': case '^': case '!': case '.': case '?':
3972 if (definedef
!= dnone
)
3974 /* These surely cannot follow a function tag in C. */
3987 if (objdef
== otagseen
)
3989 make_C_tag (true); /* an Objective C class */
3992 /* If a macro spans multiple lines don't reset its state. */
3994 CNL_SAVE_DEFINEDEF ();
4000 } /* while not eof */
4002 free (lbs
[0].lb
.buffer
);
4003 free (lbs
[1].lb
.buffer
);
4007 * Process either a C++ file or a C file depending on the setting
4011 default_C_entries (FILE *inf
)
4013 C_entries (cplusplus
? C_PLPL
: C_AUTO
, inf
);
4016 /* Always do plain C. */
4018 plain_C_entries (FILE *inf
)
4023 /* Always do C++. */
4025 Cplusplus_entries (FILE *inf
)
4027 C_entries (C_PLPL
, inf
);
4030 /* Always do Java. */
4032 Cjava_entries (FILE *inf
)
4034 C_entries (C_JAVA
, inf
);
4039 Cstar_entries (FILE *inf
)
4041 C_entries (C_STAR
, inf
);
4044 /* Always do Yacc. */
4046 Yacc_entries (FILE *inf
)
4048 C_entries (YACC
, inf
);
4052 /* Useful macros. */
4053 #define LOOP_ON_INPUT_LINES(file_pointer, line_buffer, char_pointer) \
4054 while (perhaps_more_input (file_pointer) \
4055 && (readline (&(line_buffer), file_pointer), \
4056 (char_pointer) = (line_buffer).buffer, \
4059 #define LOOKING_AT(cp, kw) /* kw is the keyword, a literal string */ \
4060 ((assert ("" kw), true) /* syntax error if not a literal string */ \
4061 && strneq ((cp), kw, sizeof (kw)-1) /* cp points at kw */ \
4062 && notinname ((cp)[sizeof (kw)-1]) /* end of kw */ \
4063 && ((cp) = skip_spaces ((cp)+sizeof (kw)-1))) /* skip spaces */
4065 /* Similar to LOOKING_AT but does not use notinname, does not skip */
4066 #define LOOKING_AT_NOCASE(cp, kw) /* the keyword is a literal string */ \
4067 ((assert ("" kw), true) /* syntax error if not a literal string */ \
4068 && strncaseeq ((cp), kw, sizeof (kw)-1) /* cp points at kw */ \
4069 && ((cp) += sizeof (kw)-1)) /* skip spaces */
4072 * Read a file, but do no processing. This is used to do regexp
4073 * matching on files that have no language defined.
4076 just_read_file (FILE *inf
)
4078 while (perhaps_more_input (inf
))
4079 readline (&lb
, inf
);
4083 /* Fortran parsing */
4085 static void F_takeprec (void);
4086 static void F_getit (FILE *);
4091 dbp
= skip_spaces (dbp
);
4095 dbp
= skip_spaces (dbp
);
4096 if (strneq (dbp
, "(*)", 3))
4101 if (!c_isdigit (*dbp
))
4103 --dbp
; /* force failure */
4108 while (c_isdigit (*dbp
));
4116 dbp
= skip_spaces (dbp
);
4119 readline (&lb
, inf
);
4124 dbp
= skip_spaces (dbp
);
4126 if (!c_isalpha (*dbp
) && *dbp
!= '_' && *dbp
!= '$')
4128 for (cp
= dbp
+ 1; *cp
!= '\0' && intoken (*cp
); cp
++)
4130 make_tag (dbp
, cp
-dbp
, true,
4131 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4136 Fortran_functions (FILE *inf
)
4138 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
4141 dbp
++; /* Ratfor escape to fortran */
4142 dbp
= skip_spaces (dbp
);
4146 if (LOOKING_AT_NOCASE (dbp
, "recursive"))
4147 dbp
= skip_spaces (dbp
);
4149 if (LOOKING_AT_NOCASE (dbp
, "pure"))
4150 dbp
= skip_spaces (dbp
);
4152 if (LOOKING_AT_NOCASE (dbp
, "elemental"))
4153 dbp
= skip_spaces (dbp
);
4155 switch (c_tolower (*dbp
))
4158 if (nocase_tail ("integer"))
4162 if (nocase_tail ("real"))
4166 if (nocase_tail ("logical"))
4170 if (nocase_tail ("complex") || nocase_tail ("character"))
4174 if (nocase_tail ("double"))
4176 dbp
= skip_spaces (dbp
);
4179 if (nocase_tail ("precision"))
4185 dbp
= skip_spaces (dbp
);
4188 switch (c_tolower (*dbp
))
4191 if (nocase_tail ("function"))
4195 if (nocase_tail ("subroutine"))
4199 if (nocase_tail ("entry"))
4203 if (nocase_tail ("blockdata") || nocase_tail ("block data"))
4205 dbp
= skip_spaces (dbp
);
4206 if (*dbp
== '\0') /* assume un-named */
4207 make_tag ("blockdata", 9, true,
4208 lb
.buffer
, dbp
- lb
.buffer
, lineno
, linecharno
);
4210 F_getit (inf
); /* look for name */
4219 * Go language support
4220 * Original code by Xi Lu <lx@shellcodes.org> (2016)
4223 Go_functions(FILE *inf
)
4227 LOOP_ON_INPUT_LINES(inf
, lb
, cp
)
4229 cp
= skip_spaces (cp
);
4231 if (LOOKING_AT (cp
, "package"))
4234 while (!notinname (*cp
) && *cp
!= '\0')
4236 make_tag (name
, cp
- name
, false, lb
.buffer
,
4237 cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4239 else if (LOOKING_AT (cp
, "func"))
4241 /* Go implementation of interface, such as:
4242 func (n *Integer) Add(m Integer) ...
4243 skip `(n *Integer)` part.
4249 cp
= skip_spaces (cp
+1);
4256 while (!notinname (*cp
))
4259 make_tag (name
, cp
- name
, true, lb
.buffer
,
4260 cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4263 else if (members
&& LOOKING_AT (cp
, "type"))
4267 /* Ignore the likes of the following:
4275 while (!notinname (*cp
) && *cp
!= '\0')
4278 make_tag (name
, cp
- name
, false, lb
.buffer
,
4279 cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4288 * Philippe Waroquiers (1998)
4291 /* Once we are positioned after an "interesting" keyword, let's get
4292 the real tag value necessary. */
4294 Ada_getit (FILE *inf
, const char *name_qualifier
)
4300 while (perhaps_more_input (inf
))
4302 dbp
= skip_spaces (dbp
);
4304 || (dbp
[0] == '-' && dbp
[1] == '-'))
4306 readline (&lb
, inf
);
4309 switch (c_tolower (*dbp
))
4312 if (nocase_tail ("body"))
4314 /* Skipping body of procedure body or package body or ....
4315 resetting qualifier to body instead of spec. */
4316 name_qualifier
= "/b";
4321 /* Skipping type of task type or protected type ... */
4322 if (nocase_tail ("type"))
4329 for (cp
= dbp
; *cp
!= '\0' && *cp
!= '"'; cp
++)
4334 dbp
= skip_spaces (dbp
);
4336 c_isalnum (*cp
) || *cp
== '_' || *cp
== '.';
4344 name
= concat (dbp
, name_qualifier
, "");
4346 make_tag (name
, strlen (name
), true,
4347 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4356 Ada_funcs (FILE *inf
)
4358 bool inquote
= false;
4359 bool skip_till_semicolumn
= false;
4361 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
4363 while (*dbp
!= '\0')
4365 /* Skip a string i.e. "abcd". */
4366 if (inquote
|| (*dbp
== '"'))
4368 dbp
= strchr (dbp
+ !inquote
, '"');
4373 continue; /* advance char */
4378 break; /* advance line */
4382 /* Skip comments. */
4383 if (dbp
[0] == '-' && dbp
[1] == '-')
4384 break; /* advance line */
4386 /* Skip character enclosed in single quote i.e. 'a'
4387 and skip single quote starting an attribute i.e. 'Image. */
4396 if (skip_till_semicolumn
)
4399 skip_till_semicolumn
= false;
4401 continue; /* advance char */
4404 /* Search for beginning of a token. */
4405 if (!begtoken (*dbp
))
4408 continue; /* advance char */
4411 /* We are at the beginning of a token. */
4412 switch (c_tolower (*dbp
))
4415 if (!packages_only
&& nocase_tail ("function"))
4416 Ada_getit (inf
, "/f");
4418 break; /* from switch */
4419 continue; /* advance char */
4421 if (!packages_only
&& nocase_tail ("procedure"))
4422 Ada_getit (inf
, "/p");
4423 else if (nocase_tail ("package"))
4424 Ada_getit (inf
, "/s");
4425 else if (nocase_tail ("protected")) /* protected type */
4426 Ada_getit (inf
, "/t");
4428 break; /* from switch */
4429 continue; /* advance char */
4432 if (typedefs
&& !packages_only
&& nocase_tail ("use"))
4434 /* when tagging types, avoid tagging use type Pack.Typename;
4435 for this, we will skip everything till a ; */
4436 skip_till_semicolumn
= true;
4437 continue; /* advance char */
4441 if (!packages_only
&& nocase_tail ("task"))
4442 Ada_getit (inf
, "/k");
4443 else if (typedefs
&& !packages_only
&& nocase_tail ("type"))
4445 Ada_getit (inf
, "/t");
4446 while (*dbp
!= '\0')
4450 break; /* from switch */
4451 continue; /* advance char */
4454 /* Look for the end of the token. */
4455 while (!endtoken (*dbp
))
4458 } /* advance char */
4459 } /* advance line */
4464 * Unix and microcontroller assembly tag handling
4465 * Labels: /^[a-zA-Z_.$][a-zA_Z0-9_.$]*[: ^I^J]/
4466 * Idea by Bob Weiner, Motorola Inc. (1994)
4469 Asm_labels (FILE *inf
)
4473 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4475 /* If first char is alphabetic or one of [_.$], test for colon
4476 following identifier. */
4477 if (c_isalpha (*cp
) || *cp
== '_' || *cp
== '.' || *cp
== '$')
4479 /* Read past label. */
4481 while (c_isalnum (*cp
) || *cp
== '_' || *cp
== '.' || *cp
== '$')
4483 if (*cp
== ':' || c_isspace (*cp
))
4484 /* Found end of label, so copy it and add it to the table. */
4485 make_tag (lb
.buffer
, cp
- lb
.buffer
, true,
4486 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4494 * Perl sub names: /^sub[ \t\n]+[^ \t\n{]+/
4495 * /^use constant[ \t\n]+[^ \t\n{=,;]+/
4496 * Perl variable names: /^(my|local).../
4497 * Original code by Bart Robinson <lomew@cs.utah.edu> (1995)
4498 * Additions by Michael Ernst <mernst@alum.mit.edu> (1997)
4499 * Ideas by Kai Großjohann <Kai.Grossjohann@CS.Uni-Dortmund.DE> (2001)
4502 Perl_functions (FILE *inf
)
4504 char *package
= savestr ("main"); /* current package name */
4507 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4509 cp
= skip_spaces (cp
);
4511 if (LOOKING_AT (cp
, "package"))
4514 get_tag (cp
, &package
);
4516 else if (LOOKING_AT (cp
, "sub"))
4522 while (!notinname (*cp
))
4525 continue; /* nothing found */
4526 pos
= strchr (sp
, ':');
4527 if (pos
&& pos
< cp
&& pos
[1] == ':')
4528 /* The name is already qualified. */
4529 make_tag (sp
, cp
- sp
, true,
4530 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4534 char savechar
, *name
;
4538 name
= concat (package
, "::", sp
);
4540 make_tag (name
, strlen (name
), true,
4541 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4545 else if (LOOKING_AT (cp
, "use constant")
4546 || LOOKING_AT (cp
, "use constant::defer"))
4548 /* For hash style multi-constant like
4549 use constant { FOO => 123,
4551 only the first FOO is picked up. Parsing across the value
4552 expressions would be difficult in general, due to possible nested
4553 hashes, here-documents, etc. */
4555 cp
= skip_spaces (cp
+1);
4558 else if (globals
) /* only if we are tagging global vars */
4560 /* Skip a qualifier, if any. */
4561 bool qual
= LOOKING_AT (cp
, "my") || LOOKING_AT (cp
, "local");
4562 /* After "my" or "local", but before any following paren or space. */
4563 char *varstart
= cp
;
4565 if (qual
/* should this be removed? If yes, how? */
4566 && (*cp
== '$' || *cp
== '@' || *cp
== '%'))
4571 while (c_isalnum (*cp
) || *cp
== '_');
4575 /* Should be examining a variable list at this point;
4576 could insist on seeing an open parenthesis. */
4577 while (*cp
!= '\0' && *cp
!= ';' && *cp
!= '=' && *cp
!= ')')
4583 make_tag (varstart
, cp
- varstart
, false,
4584 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4593 * Look for /^[\t]*def[ \t\n]+[^ \t\n(:]+/ or /^class[ \t\n]+[^ \t\n(:]+/
4594 * Idea by Eric S. Raymond <esr@thyrsus.com> (1997)
4595 * More ideas by seb bacon <seb@jamkit.com> (2002)
4598 Python_functions (FILE *inf
)
4602 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4604 cp
= skip_spaces (cp
);
4605 if (LOOKING_AT (cp
, "def") || LOOKING_AT (cp
, "class"))
4608 while (!notinname (*cp
) && *cp
!= ':')
4610 make_tag (name
, cp
- name
, true,
4611 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4618 * Original code by Xi Lu <lx@shellcodes.org> (2015)
4621 Ruby_functions (FILE *inf
)
4625 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4627 bool is_class
= false;
4628 bool is_method
= false;
4631 cp
= skip_spaces (cp
);
4632 if (c_isalpha (*cp
) && c_isupper (*cp
)) /* constants */
4634 char *bp
, *colon
= NULL
;
4638 for (cp
++; c_isalnum (*cp
) || *cp
== '_' || *cp
== ':'; cp
++)
4645 bp
= skip_spaces (cp
);
4646 if (*bp
== '=' && c_isspace (bp
[1]))
4648 if (colon
&& !c_isspace (colon
[1]))
4650 make_tag (name
, cp
- name
, false,
4651 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4655 else if ((is_method
= LOOKING_AT (cp
, "def")) /* module/class/method */
4656 || (is_class
= LOOKING_AT (cp
, "class"))
4657 || LOOKING_AT (cp
, "module"))
4659 const char self_name
[] = "self.";
4660 const size_t self_size1
= sizeof ("self.") - 1;
4664 /* Ruby method names can end in a '='. Also, operator overloading can
4665 define operators whose names include '='. */
4666 while (!notinname (*cp
) || *cp
== '=')
4669 /* Remove "self." from the method name. */
4670 if (cp
- name
> self_size1
4671 && strneq (name
, self_name
, self_size1
))
4674 /* Remove the class/module qualifiers from method names. */
4679 for (q
= name
; q
< cp
&& *q
!= '.'; q
++)
4681 if (q
< cp
- 1) /* punt if we see just "FOO." */
4685 /* Don't tag singleton classes. */
4686 if (is_class
&& strneq (name
, "<<", 2) && cp
== name
+ 2)
4689 make_tag (name
, cp
- name
, true,
4690 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4699 * - /^[ \t]*function[ \t\n]+[^ \t\n(]+/
4700 * - /^[ \t]*class[ \t\n]+[^ \t\n]+/
4701 * - /^[ \t]*define\(\"[^\"]+/
4702 * Only with --members:
4703 * - /^[ \t]*var[ \t\n]+\$[^ \t\n=;]/
4704 * Idea by Diez B. Roggisch (2001)
4707 PHP_functions (FILE *inf
)
4710 bool search_identifier
= false;
4712 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4714 cp
= skip_spaces (cp
);
4716 if (search_identifier
4719 while (!notinname (*cp
))
4721 make_tag (name
, cp
- name
, true,
4722 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4723 search_identifier
= false;
4725 else if (LOOKING_AT (cp
, "function"))
4728 cp
= skip_spaces (cp
+1);
4732 while (!notinname (*cp
))
4734 make_tag (name
, cp
- name
, true,
4735 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4738 search_identifier
= true;
4740 else if (LOOKING_AT (cp
, "class"))
4745 while (*cp
!= '\0' && !c_isspace (*cp
))
4747 make_tag (name
, cp
- name
, false,
4748 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4751 search_identifier
= true;
4753 else if (strneq (cp
, "define", 6)
4754 && (cp
= skip_spaces (cp
+6))
4756 && (*cp
== '"' || *cp
== '\''))
4760 while (*cp
!= quote
&& *cp
!= '\0')
4762 make_tag (name
, cp
- name
, false,
4763 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4766 && LOOKING_AT (cp
, "var")
4770 while (!notinname (*cp
))
4772 make_tag (name
, cp
- name
, false,
4773 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4780 * Cobol tag functions
4781 * We could look for anything that could be a paragraph name.
4782 * i.e. anything that starts in column 8 is one word and ends in a full stop.
4783 * Idea by Corny de Souza (1993)
4786 Cobol_paragraphs (FILE *inf
)
4788 register char *bp
, *ep
;
4790 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4796 /* If eoln, compiler option or comment ignore whole line. */
4797 if (bp
[-1] != ' ' || !c_isalnum (bp
[0]))
4800 for (ep
= bp
; c_isalnum (*ep
) || *ep
== '-'; ep
++)
4803 make_tag (bp
, ep
- bp
, true,
4804 lb
.buffer
, ep
- lb
.buffer
+ 1, lineno
, linecharno
);
4811 * Ideas by Assar Westerlund <assar@sics.se> (2001)
4814 Makefile_targets (FILE *inf
)
4818 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4820 if (*bp
== '\t' || *bp
== '#')
4822 while (*bp
!= '\0' && *bp
!= '=' && *bp
!= ':')
4824 if (*bp
== ':' || (globals
&& *bp
== '='))
4826 /* We should detect if there is more than one tag, but we do not.
4827 We just skip initial and final spaces. */
4828 char * namestart
= skip_spaces (lb
.buffer
);
4829 while (--bp
> namestart
)
4830 if (!notinname (*bp
))
4832 make_tag (namestart
, bp
- namestart
+ 1, true,
4833 lb
.buffer
, bp
- lb
.buffer
+ 2, lineno
, linecharno
);
4841 * Original code by Mosur K. Mohan (1989)
4843 * Locates tags for procedures & functions. Doesn't do any type- or
4844 * var-definitions. It does look for the keyword "extern" or
4845 * "forward" immediately following the procedure statement; if found,
4846 * the tag is skipped.
4849 Pascal_functions (FILE *inf
)
4851 linebuffer tline
; /* mostly copied from C_entries */
4853 int save_lineno
, namelen
, taglen
;
4856 bool /* each of these flags is true if: */
4857 incomment
, /* point is inside a comment */
4858 inquote
, /* point is inside '..' string */
4859 get_tagname
, /* point is after PROCEDURE/FUNCTION
4860 keyword, so next item = potential tag */
4861 found_tag
, /* point is after a potential tag */
4862 inparms
, /* point is within parameter-list */
4863 verify_tag
; /* point has passed the parm-list, so the
4864 next token will determine whether this
4865 is a FORWARD/EXTERN to be ignored, or
4866 whether it is a real tag */
4868 save_lcno
= save_lineno
= namelen
= taglen
= 0; /* keep compiler quiet */
4869 name
= NULL
; /* keep compiler quiet */
4872 linebuffer_init (&tline
);
4874 incomment
= inquote
= false;
4875 found_tag
= false; /* have a proc name; check if extern */
4876 get_tagname
= false; /* found "procedure" keyword */
4877 inparms
= false; /* found '(' after "proc" */
4878 verify_tag
= false; /* check if "extern" is ahead */
4881 while (perhaps_more_input (inf
)) /* long main loop to get next char */
4884 if (c
== '\0') /* if end of line */
4886 readline (&lb
, inf
);
4890 if (!((found_tag
&& verify_tag
)
4892 c
= *dbp
++; /* only if don't need *dbp pointing
4893 to the beginning of the name of
4894 the procedure or function */
4898 if (c
== '}') /* within { } comments */
4900 else if (c
== '*' && *dbp
== ')') /* within (* *) comments */
4917 inquote
= true; /* found first quote */
4919 case '{': /* found open { comment */
4923 if (*dbp
== '*') /* found open (* comment */
4928 else if (found_tag
) /* found '(' after tag, i.e., parm-list */
4931 case ')': /* end of parms list */
4936 if (found_tag
&& !inparms
) /* end of proc or fn stmt */
4943 if (found_tag
&& verify_tag
&& (*dbp
!= ' '))
4945 /* Check if this is an "extern" declaration. */
4948 if (c_tolower (*dbp
) == 'e')
4950 if (nocase_tail ("extern")) /* superfluous, really! */
4956 else if (c_tolower (*dbp
) == 'f')
4958 if (nocase_tail ("forward")) /* check for forward reference */
4964 if (found_tag
&& verify_tag
) /* not external proc, so make tag */
4968 make_tag (name
, namelen
, true,
4969 tline
.buffer
, taglen
, save_lineno
, save_lcno
);
4973 if (get_tagname
) /* grab name of proc or fn */
4980 /* Find block name. */
4981 for (cp
= dbp
+ 1; *cp
!= '\0' && !endtoken (*cp
); cp
++)
4984 /* Save all values for later tagging. */
4985 linebuffer_setlen (&tline
, lb
.len
);
4986 strcpy (tline
.buffer
, lb
.buffer
);
4987 save_lineno
= lineno
;
4988 save_lcno
= linecharno
;
4989 name
= tline
.buffer
+ (dbp
- lb
.buffer
);
4991 taglen
= cp
- lb
.buffer
+ 1;
4993 dbp
= cp
; /* set dbp to e-o-token */
4994 get_tagname
= false;
4998 /* And proceed to check for "extern". */
5000 else if (!incomment
&& !inquote
&& !found_tag
)
5002 /* Check for proc/fn keywords. */
5003 switch (c_tolower (c
))
5006 if (nocase_tail ("rocedure")) /* c = 'p', dbp has advanced */
5010 if (nocase_tail ("unction"))
5015 } /* while not eof */
5017 free (tline
.buffer
);
5022 * Lisp tag functions
5023 * look for (def or (DEF, quote or QUOTE
5026 static void L_getit (void);
5031 if (*dbp
== '\'') /* Skip prefix quote */
5033 else if (*dbp
== '(')
5036 /* Try to skip "(quote " */
5037 if (!LOOKING_AT (dbp
, "quote") && !LOOKING_AT (dbp
, "QUOTE"))
5038 /* Ok, then skip "(" before name in (defstruct (foo)) */
5039 dbp
= skip_spaces (dbp
);
5041 get_tag (dbp
, NULL
);
5045 Lisp_functions (FILE *inf
)
5047 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
5052 /* "(defvar foo)" is a declaration rather than a definition. */
5056 if (LOOKING_AT (p
, "defvar"))
5058 p
= skip_name (p
); /* past var name */
5059 p
= skip_spaces (p
);
5065 if (strneq (dbp
+ 1, "cl-", 3) || strneq (dbp
+ 1, "CL-", 3))
5068 if (strneq (dbp
+1, "def", 3) || strneq (dbp
+1, "DEF", 3))
5070 dbp
= skip_non_spaces (dbp
);
5071 dbp
= skip_spaces (dbp
);
5076 /* Check for (foo::defmumble name-defined ... */
5079 while (!notinname (*dbp
) && *dbp
!= ':');
5084 while (*dbp
== ':');
5086 if (strneq (dbp
, "def", 3) || strneq (dbp
, "DEF", 3))
5088 dbp
= skip_non_spaces (dbp
);
5089 dbp
= skip_spaces (dbp
);
5099 * Lua script language parsing
5100 * Original code by David A. Capello <dacap@users.sourceforge.net> (2004)
5102 * "function" and "local function" are tags if they start at column 1.
5105 Lua_functions (FILE *inf
)
5109 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
5111 bp
= skip_spaces (bp
);
5112 if (bp
[0] != 'f' && bp
[0] != 'l')
5115 (void)LOOKING_AT (bp
, "local"); /* skip possible "local" */
5117 if (LOOKING_AT (bp
, "function"))
5119 char *tag_name
, *tp_dot
, *tp_colon
;
5121 get_tag (bp
, &tag_name
);
5122 /* If the tag ends with ".foo" or ":foo", make an additional tag for
5124 tp_dot
= strrchr (tag_name
, '.');
5125 tp_colon
= strrchr (tag_name
, ':');
5126 if (tp_dot
|| tp_colon
)
5128 char *p
= tp_dot
> tp_colon
? tp_dot
: tp_colon
;
5129 int len_add
= p
- tag_name
+ 1;
5131 get_tag (bp
+ len_add
, NULL
);
5140 * Just look for lines where the first character is '/'
5141 * Also look at "defineps" for PSWrap
5143 * Richard Mlynarik <mly@adoc.xerox.com> (1997)
5144 * Masatake Yamato <masata-y@is.aist-nara.ac.jp> (1999)
5147 PS_functions (FILE *inf
)
5149 register char *bp
, *ep
;
5151 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
5156 *ep
!= '\0' && *ep
!= ' ' && *ep
!= '{';
5159 make_tag (bp
, ep
- bp
, true,
5160 lb
.buffer
, ep
- lb
.buffer
+ 1, lineno
, linecharno
);
5162 else if (LOOKING_AT (bp
, "defineps"))
5170 * Ignore anything after \ followed by space or in ( )
5171 * Look for words defined by :
5172 * Look for constant, code, create, defer, value, and variable
5173 * OBP extensions: Look for buffer:, field,
5174 * Ideas by Eduardo Horvath <eeh@netbsd.org> (2004)
5177 Forth_words (FILE *inf
)
5181 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
5182 while ((bp
= skip_spaces (bp
))[0] != '\0')
5183 if (bp
[0] == '\\' && c_isspace (bp
[1]))
5184 break; /* read next line */
5185 else if (bp
[0] == '(' && c_isspace (bp
[1]))
5186 do /* skip to ) or eol */
5188 while (*bp
!= ')' && *bp
!= '\0');
5189 else if ((bp
[0] == ':' && c_isspace (bp
[1]) && bp
++)
5190 || LOOKING_AT_NOCASE (bp
, "constant")
5191 || LOOKING_AT_NOCASE (bp
, "code")
5192 || LOOKING_AT_NOCASE (bp
, "create")
5193 || LOOKING_AT_NOCASE (bp
, "defer")
5194 || LOOKING_AT_NOCASE (bp
, "value")
5195 || LOOKING_AT_NOCASE (bp
, "variable")
5196 || LOOKING_AT_NOCASE (bp
, "buffer:")
5197 || LOOKING_AT_NOCASE (bp
, "field"))
5198 get_tag (skip_spaces (bp
), NULL
); /* Yay! A definition! */
5200 bp
= skip_non_spaces (bp
);
5205 * Scheme tag functions
5206 * look for (def... xyzzy
5208 * (def ... ((...(xyzzy ....
5210 * Original code by Ken Haase (1985?)
5213 Scheme_functions (FILE *inf
)
5217 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
5219 if (strneq (bp
, "(def", 4) || strneq (bp
, "(DEF", 4))
5221 bp
= skip_non_spaces (bp
+4);
5222 /* Skip over open parens and white space. Don't continue past
5224 while (*bp
&& notinname (*bp
))
5228 if (LOOKING_AT (bp
, "(SET!") || LOOKING_AT (bp
, "(set!"))
5234 /* Find tags in TeX and LaTeX input files. */
5236 /* TEX_toktab is a table of TeX control sequences that define tags.
5237 * Each entry records one such control sequence.
5239 * Original code from who knows whom.
5241 * Stefan Monnier (2002)
5244 static linebuffer
*TEX_toktab
= NULL
; /* Table with tag tokens */
5246 /* Default set of control sequences to put into TEX_toktab.
5247 The value of environment var TEXTAGS is prepended to this. */
5248 static const char *TEX_defenv
= "\
5249 :chapter:section:subsection:subsubsection:eqno:label:ref:cite:bibitem\
5250 :part:appendix:entry:index:def\
5251 :newcommand:renewcommand:newenvironment:renewenvironment";
5253 static void TEX_decode_env (const char *, const char *);
5256 * TeX/LaTeX scanning loop.
5259 TeX_commands (FILE *inf
)
5264 char TEX_esc
= '\0';
5265 char TEX_opgrp
, TEX_clgrp
;
5267 /* Initialize token table once from environment. */
5268 if (TEX_toktab
== NULL
)
5269 TEX_decode_env ("TEXTAGS", TEX_defenv
);
5271 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5273 /* Look at each TEX keyword in line. */
5276 /* Look for a TEX escape. */
5280 if (c
== '\0' || c
== '%')
5283 /* Select either \ or ! as escape character, whichever comes
5284 first outside a comment. */
5305 for (key
= TEX_toktab
; key
->buffer
!= NULL
; key
++)
5306 if (strneq (cp
, key
->buffer
, key
->len
))
5309 int namelen
, linelen
;
5312 cp
= skip_spaces (cp
+ key
->len
);
5313 if (*cp
== TEX_opgrp
)
5319 (!c_isspace (*p
) && *p
!= '#' &&
5320 *p
!= TEX_opgrp
&& *p
!= TEX_clgrp
);
5325 if (!opgrp
|| *p
== TEX_clgrp
)
5327 while (*p
!= '\0' && *p
!= TEX_opgrp
&& *p
!= TEX_clgrp
)
5329 linelen
= p
- lb
.buffer
+ 1;
5331 make_tag (cp
, namelen
, true,
5332 lb
.buffer
, linelen
, lineno
, linecharno
);
5333 goto tex_next_line
; /* We only tag a line once */
5341 /* Read environment and prepend it to the default string.
5342 Build token table. */
5344 TEX_decode_env (const char *evarname
, const char *defenv
)
5346 register const char *env
, *p
;
5349 /* Append default string to environment. */
5350 env
= getenv (evarname
);
5354 env
= concat (env
, defenv
, "");
5356 /* Allocate a token table */
5357 for (len
= 1, p
= env
; (p
= strchr (p
, ':')); )
5360 TEX_toktab
= xnew (len
, linebuffer
);
5362 /* Unpack environment string into token table. Be careful about */
5363 /* zero-length strings (leading ':', "::" and trailing ':') */
5364 for (i
= 0; *env
!= '\0';)
5366 p
= strchr (env
, ':');
5367 if (!p
) /* End of environment string. */
5368 p
= env
+ strlen (env
);
5370 { /* Only non-zero strings. */
5371 TEX_toktab
[i
].buffer
= savenstr (env
, p
- env
);
5372 TEX_toktab
[i
].len
= p
- env
;
5379 TEX_toktab
[i
].buffer
= NULL
; /* Mark end of table. */
5380 TEX_toktab
[i
].len
= 0;
5387 /* Texinfo support. Dave Love, Mar. 2000. */
5389 Texinfo_nodes (FILE *inf
)
5392 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5393 if (LOOKING_AT (cp
, "@node"))
5396 while (*cp
!= '\0' && *cp
!= ',')
5398 make_tag (start
, cp
- start
, true,
5399 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
5406 * Contents of <title>, <h1>, <h2>, <h3> are tags.
5407 * Contents of <a name=xxx> are tags with name xxx.
5409 * Francesco Potortì, 2002.
5412 HTML_labels (FILE *inf
)
5414 bool getnext
= false; /* next text outside of HTML tags is a tag */
5415 bool skiptag
= false; /* skip to the end of the current HTML tag */
5416 bool intag
= false; /* inside an html tag, looking for ID= */
5417 bool inanchor
= false; /* when INTAG, is an anchor, look for NAME= */
5421 linebuffer_setlen (&token_name
, 0); /* no name in buffer */
5423 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
5424 for (;;) /* loop on the same line */
5426 if (skiptag
) /* skip HTML tag */
5428 while (*dbp
!= '\0' && *dbp
!= '>')
5434 continue; /* look on the same line */
5436 break; /* go to next line */
5439 else if (intag
) /* look for "name=" or "id=" */
5441 while (*dbp
!= '\0' && *dbp
!= '>'
5442 && c_tolower (*dbp
) != 'n' && c_tolower (*dbp
) != 'i')
5445 break; /* go to next line */
5450 continue; /* look on the same line */
5452 if ((inanchor
&& LOOKING_AT_NOCASE (dbp
, "name="))
5453 || LOOKING_AT_NOCASE (dbp
, "id="))
5455 bool quoted
= (dbp
[0] == '"');
5458 for (end
= ++dbp
; *end
!= '\0' && *end
!= '"'; end
++)
5461 for (end
= dbp
; *end
!= '\0' && intoken (*end
); end
++)
5463 linebuffer_setlen (&token_name
, end
- dbp
);
5464 memcpy (token_name
.buffer
, dbp
, end
- dbp
);
5465 token_name
.buffer
[end
- dbp
] = '\0';
5468 intag
= false; /* we found what we looked for */
5469 skiptag
= true; /* skip to the end of the tag */
5470 getnext
= true; /* then grab the text */
5471 continue; /* look on the same line */
5476 else if (getnext
) /* grab next tokens and tag them */
5478 dbp
= skip_spaces (dbp
);
5480 break; /* go to next line */
5484 inanchor
= (c_tolower (dbp
[1]) == 'a' && !intoken (dbp
[2]));
5485 continue; /* look on the same line */
5488 for (end
= dbp
+ 1; *end
!= '\0' && *end
!= '<'; end
++)
5490 make_tag (token_name
.buffer
, token_name
.len
, true,
5491 dbp
, end
- dbp
, lineno
, linecharno
);
5492 linebuffer_setlen (&token_name
, 0); /* no name in buffer */
5494 break; /* go to next line */
5497 else /* look for an interesting HTML tag */
5499 while (*dbp
!= '\0' && *dbp
!= '<')
5502 break; /* go to next line */
5504 if (c_tolower (dbp
[1]) == 'a' && !intoken (dbp
[2]))
5507 continue; /* look on the same line */
5509 else if (LOOKING_AT_NOCASE (dbp
, "<title>")
5510 || LOOKING_AT_NOCASE (dbp
, "<h1>")
5511 || LOOKING_AT_NOCASE (dbp
, "<h2>")
5512 || LOOKING_AT_NOCASE (dbp
, "<h3>"))
5516 continue; /* look on the same line */
5527 * Assumes that the predicate or rule starts at column 0.
5528 * Only the first clause of a predicate or rule is added.
5529 * Original code by Sunichirou Sugou (1989)
5530 * Rewritten by Anders Lindgren (1996)
5532 static size_t prolog_pr (char *, char *);
5533 static void prolog_skip_comment (linebuffer
*, FILE *);
5534 static size_t prolog_atom (char *, size_t);
5537 Prolog_functions (FILE *inf
)
5547 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5549 if (cp
[0] == '\0') /* Empty line */
5551 else if (c_isspace (cp
[0])) /* Not a predicate */
5553 else if (cp
[0] == '/' && cp
[1] == '*') /* comment. */
5554 prolog_skip_comment (&lb
, inf
);
5555 else if ((len
= prolog_pr (cp
, last
)) > 0)
5557 /* Predicate or rule. Store the function name so that we
5558 only generate a tag for the first clause. */
5560 last
= xnew (len
+ 1, char);
5561 else if (len
+ 1 > allocated
)
5562 xrnew (last
, len
+ 1, char);
5563 allocated
= len
+ 1;
5564 memcpy (last
, cp
, len
);
5573 prolog_skip_comment (linebuffer
*plb
, FILE *inf
)
5579 for (cp
= plb
->buffer
; *cp
!= '\0'; cp
++)
5580 if (cp
[0] == '*' && cp
[1] == '/')
5582 readline (plb
, inf
);
5584 while (perhaps_more_input (inf
));
5588 * A predicate or rule definition is added if it matches:
5589 * <beginning of line><Prolog Atom><whitespace>(
5590 * or <beginning of line><Prolog Atom><whitespace>:-
5592 * It is added to the tags database if it doesn't match the
5593 * name of the previous clause header.
5595 * Return the size of the name of the predicate or rule, or 0 if no
5599 prolog_pr (char *s
, char *last
)
5601 /* Name of last clause. */
5606 pos
= prolog_atom (s
, 0);
5611 pos
= skip_spaces (s
+ pos
) - s
;
5614 || (s
[pos
] == '(' && (pos
+= 1))
5615 || (s
[pos
] == ':' && s
[pos
+ 1] == '-' && (pos
+= 2)))
5616 && (last
== NULL
/* save only the first clause */
5617 || len
!= strlen (last
)
5618 || !strneq (s
, last
, len
)))
5620 make_tag (s
, len
, true, s
, pos
, lineno
, linecharno
);
5628 * Consume a Prolog atom.
5629 * Return the number of bytes consumed, or 0 if there was an error.
5631 * A prolog atom, in this context, could be one of:
5632 * - An alphanumeric sequence, starting with a lower case letter.
5633 * - A quoted arbitrary string. Single quotes can escape themselves.
5634 * Backslash quotes everything.
5637 prolog_atom (char *s
, size_t pos
)
5643 if (c_islower (s
[pos
]) || s
[pos
] == '_')
5645 /* The atom is unquoted. */
5647 while (c_isalnum (s
[pos
]) || s
[pos
] == '_')
5651 return pos
- origpos
;
5653 else if (s
[pos
] == '\'')
5664 pos
++; /* A double quote */
5666 else if (s
[pos
] == '\0')
5667 /* Multiline quoted atoms are ignored. */
5669 else if (s
[pos
] == '\\')
5671 if (s
[pos
+1] == '\0')
5678 return pos
- origpos
;
5686 * Support for Erlang
5688 * Generates tags for functions, defines, and records.
5689 * Assumes that Erlang functions start at column 0.
5690 * Original code by Anders Lindgren (1996)
5692 static int erlang_func (char *, char *);
5693 static void erlang_attribute (char *);
5694 static int erlang_atom (char *);
5697 Erlang_functions (FILE *inf
)
5707 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5709 if (cp
[0] == '\0') /* Empty line */
5711 else if (c_isspace (cp
[0])) /* Not function nor attribute */
5713 else if (cp
[0] == '%') /* comment */
5715 else if (cp
[0] == '"') /* Sometimes, strings start in column one */
5717 else if (cp
[0] == '-') /* attribute, e.g. "-define" */
5719 erlang_attribute (cp
);
5726 else if ((len
= erlang_func (cp
, last
)) > 0)
5729 * Function. Store the function name so that we only
5730 * generates a tag for the first clause.
5733 last
= xnew (len
+ 1, char);
5734 else if (len
+ 1 > allocated
)
5735 xrnew (last
, len
+ 1, char);
5736 allocated
= len
+ 1;
5737 memcpy (last
, cp
, len
);
5746 * A function definition is added if it matches:
5747 * <beginning of line><Erlang Atom><whitespace>(
5749 * It is added to the tags database if it doesn't match the
5750 * name of the previous clause header.
5752 * Return the size of the name of the function, or 0 if no function
5756 erlang_func (char *s
, char *last
)
5758 /* Name of last clause. */
5763 pos
= erlang_atom (s
);
5768 pos
= skip_spaces (s
+ pos
) - s
;
5770 /* Save only the first clause. */
5773 || len
!= (int)strlen (last
)
5774 || !strneq (s
, last
, len
)))
5776 make_tag (s
, len
, true, s
, pos
, lineno
, linecharno
);
5785 * Handle attributes. Currently, tags are generated for defines
5788 * They are on the form:
5789 * -define(foo, bar).
5790 * -define(Foo(M, N), M+N).
5791 * -record(graph, {vtab = notable, cyclic = true}).
5794 erlang_attribute (char *s
)
5798 if ((LOOKING_AT (cp
, "-define") || LOOKING_AT (cp
, "-record"))
5801 int len
= erlang_atom (skip_spaces (cp
));
5803 make_tag (cp
, len
, true, s
, cp
+ len
- s
, lineno
, linecharno
);
5810 * Consume an Erlang atom (or variable).
5811 * Return the number of bytes consumed, or -1 if there was an error.
5814 erlang_atom (char *s
)
5818 if (c_isalpha (s
[pos
]) || s
[pos
] == '_')
5820 /* The atom is unquoted. */
5823 while (c_isalnum (s
[pos
]) || s
[pos
] == '_');
5825 else if (s
[pos
] == '\'')
5827 for (pos
++; s
[pos
] != '\''; pos
++)
5828 if (s
[pos
] == '\0' /* multiline quoted atoms are ignored */
5829 || (s
[pos
] == '\\' && s
[++pos
] == '\0'))
5838 static char *scan_separators (char *);
5839 static void add_regex (char *, language
*);
5840 static char *substitute (char *, char *, struct re_registers
*);
5843 * Take a string like "/blah/" and turn it into "blah", verifying
5844 * that the first and last characters are the same, and handling
5845 * quoted separator characters. Actually, stops on the occurrence of
5846 * an unquoted separator. Also process \t, \n, etc. and turn into
5847 * appropriate characters. Works in place. Null terminates name string.
5848 * Returns pointer to terminating separator, or NULL for
5849 * unterminated regexps.
5852 scan_separators (char *name
)
5855 char *copyto
= name
;
5856 bool quoted
= false;
5858 for (++name
; *name
!= '\0'; ++name
)
5864 case 'a': *copyto
++ = '\007'; break; /* BEL (bell) */
5865 case 'b': *copyto
++ = '\b'; break; /* BS (back space) */
5866 case 'd': *copyto
++ = 0177; break; /* DEL (delete) */
5867 case 'e': *copyto
++ = 033; break; /* ESC (delete) */
5868 case 'f': *copyto
++ = '\f'; break; /* FF (form feed) */
5869 case 'n': *copyto
++ = '\n'; break; /* NL (new line) */
5870 case 'r': *copyto
++ = '\r'; break; /* CR (carriage return) */
5871 case 't': *copyto
++ = '\t'; break; /* TAB (horizontal tab) */
5872 case 'v': *copyto
++ = '\v'; break; /* VT (vertical tab) */
5878 /* Something else is quoted, so preserve the quote. */
5886 else if (*name
== '\\')
5888 else if (*name
== sep
)
5894 name
= NULL
; /* signal unterminated regexp */
5896 /* Terminate copied string. */
5901 /* Look at the argument of --regex or --no-regex and do the right
5902 thing. Same for each line of a regexp file. */
5904 analyze_regex (char *regex_arg
)
5906 if (regex_arg
== NULL
)
5908 free_regexps (); /* --no-regex: remove existing regexps */
5912 /* A real --regexp option or a line in a regexp file. */
5913 switch (regex_arg
[0])
5915 /* Comments in regexp file or null arg to --regex. */
5921 /* Read a regex file. This is recursive and may result in a
5922 loop, which will stop when the file descriptors are exhausted. */
5926 linebuffer regexbuf
;
5927 char *regexfile
= regex_arg
+ 1;
5929 /* regexfile is a file containing regexps, one per line. */
5930 regexfp
= fopen (regexfile
, "r" FOPEN_BINARY
);
5931 if (regexfp
== NULL
)
5933 linebuffer_init (®exbuf
);
5934 while (readline_internal (®exbuf
, regexfp
, regexfile
) > 0)
5935 analyze_regex (regexbuf
.buffer
);
5936 free (regexbuf
.buffer
);
5937 if (fclose (regexfp
) != 0)
5942 /* Regexp to be used for a specific language only. */
5946 char *lang_name
= regex_arg
+ 1;
5949 for (cp
= lang_name
; *cp
!= '}'; cp
++)
5952 error ("unterminated language name in regex: %s", regex_arg
);
5956 lang
= get_language_from_langname (lang_name
);
5959 add_regex (cp
, lang
);
5963 /* Regexp to be used for any language. */
5965 add_regex (regex_arg
, NULL
);
5970 /* Separate the regexp pattern, compile it,
5971 and care for optional name and modifiers. */
5973 add_regex (char *regexp_pattern
, language
*lang
)
5975 static struct re_pattern_buffer zeropattern
;
5976 char sep
, *pat
, *name
, *modifiers
;
5979 struct re_pattern_buffer
*patbuf
;
5982 force_explicit_name
= true, /* do not use implicit tag names */
5983 ignore_case
= false, /* case is significant */
5984 multi_line
= false, /* matches are done one line at a time */
5985 single_line
= false; /* dot does not match newline */
5988 if (strlen (regexp_pattern
) < 3)
5990 error ("null regexp");
5993 sep
= regexp_pattern
[0];
5994 name
= scan_separators (regexp_pattern
);
5997 error ("%s: unterminated regexp", regexp_pattern
);
6002 error ("null name for regexp \"%s\"", regexp_pattern
);
6005 modifiers
= scan_separators (name
);
6006 if (modifiers
== NULL
) /* no terminating separator --> no name */
6012 modifiers
+= 1; /* skip separator */
6014 /* Parse regex modifiers. */
6015 for (; modifiers
[0] != '\0'; modifiers
++)
6016 switch (modifiers
[0])
6019 if (modifiers
== name
)
6020 error ("forcing explicit tag name but no name, ignoring");
6021 force_explicit_name
= true;
6031 need_filebuf
= true;
6034 error ("invalid regexp modifier '%c', ignoring", modifiers
[0]);
6038 patbuf
= xnew (1, struct re_pattern_buffer
);
6039 *patbuf
= zeropattern
;
6042 static char lc_trans
[UCHAR_MAX
+ 1];
6044 for (i
= 0; i
< UCHAR_MAX
+ 1; i
++)
6045 lc_trans
[i
] = c_tolower (i
);
6046 patbuf
->translate
= lc_trans
; /* translation table to fold case */
6050 pat
= concat ("^", regexp_pattern
, ""); /* anchor to beginning of line */
6052 pat
= regexp_pattern
;
6055 re_set_syntax (RE_SYNTAX_EMACS
| RE_DOT_NEWLINE
);
6057 re_set_syntax (RE_SYNTAX_EMACS
);
6059 err
= re_compile_pattern (pat
, strlen (pat
), patbuf
);
6064 error ("%s while compiling pattern", err
);
6069 p_head
= xnew (1, regexp
);
6070 p_head
->pattern
= savestr (regexp_pattern
);
6071 p_head
->p_next
= rp
;
6072 p_head
->lang
= lang
;
6073 p_head
->pat
= patbuf
;
6074 p_head
->name
= savestr (name
);
6075 p_head
->error_signaled
= false;
6076 p_head
->force_explicit_name
= force_explicit_name
;
6077 p_head
->ignore_case
= ignore_case
;
6078 p_head
->multi_line
= multi_line
;
6082 * Do the substitutions indicated by the regular expression and
6086 substitute (char *in
, char *out
, struct re_registers
*regs
)
6089 int size
, dig
, diglen
;
6092 size
= strlen (out
);
6094 /* Pass 1: figure out how much to allocate by finding all \N strings. */
6095 if (out
[size
- 1] == '\\')
6096 fatal ("pattern error in \"%s\"", out
);
6097 for (t
= strchr (out
, '\\');
6099 t
= strchr (t
+ 2, '\\'))
6100 if (c_isdigit (t
[1]))
6103 diglen
= regs
->end
[dig
] - regs
->start
[dig
];
6109 /* Allocate space and do the substitutions. */
6111 result
= xnew (size
+ 1, char);
6113 for (t
= result
; *out
!= '\0'; out
++)
6114 if (*out
== '\\' && c_isdigit (*++out
))
6117 diglen
= regs
->end
[dig
] - regs
->start
[dig
];
6118 memcpy (t
, in
+ regs
->start
[dig
], diglen
);
6125 assert (t
<= result
+ size
);
6126 assert (t
- result
== (int)strlen (result
));
6131 /* Deallocate all regexps. */
6136 while (p_head
!= NULL
)
6138 rp
= p_head
->p_next
;
6139 free (p_head
->pattern
);
6140 free (p_head
->name
);
6148 * Reads the whole file as a single string from `filebuf' and looks for
6149 * multi-line regular expressions, creating tags on matches.
6150 * readline already dealt with normal regexps.
6152 * Idea by Ben Wing <ben@666.com> (2002).
6155 regex_tag_multiline (void)
6157 char *buffer
= filebuf
.buffer
;
6161 for (rp
= p_head
; rp
!= NULL
; rp
= rp
->p_next
)
6165 if (!rp
->multi_line
)
6166 continue; /* skip normal regexps */
6168 /* Generic initializations before parsing file from memory. */
6169 lineno
= 1; /* reset global line number */
6170 charno
= 0; /* reset global char number */
6171 linecharno
= 0; /* reset global char number of line start */
6173 /* Only use generic regexps or those for the current language. */
6174 if (rp
->lang
!= NULL
&& rp
->lang
!= curfdp
->lang
)
6177 while (match
>= 0 && match
< filebuf
.len
)
6179 match
= re_search (rp
->pat
, buffer
, filebuf
.len
, charno
,
6180 filebuf
.len
- match
, &rp
->regs
);
6185 if (!rp
->error_signaled
)
6187 error ("regexp stack overflow while matching \"%s\"",
6189 rp
->error_signaled
= true;
6196 if (match
== rp
->regs
.end
[0])
6198 if (!rp
->error_signaled
)
6200 error ("regexp matches the empty string: \"%s\"",
6202 rp
->error_signaled
= true;
6204 match
= -3; /* exit from while loop */
6208 /* Match occurred. Construct a tag. */
6209 while (charno
< rp
->regs
.end
[0])
6210 if (buffer
[charno
++] == '\n')
6211 lineno
++, linecharno
= charno
;
6213 if (name
[0] == '\0')
6215 else /* make a named tag */
6216 name
= substitute (buffer
, rp
->name
, &rp
->regs
);
6217 if (rp
->force_explicit_name
)
6218 /* Force explicit tag name, if a name is there. */
6219 pfnote (name
, true, buffer
+ linecharno
,
6220 charno
- linecharno
+ 1, lineno
, linecharno
);
6222 make_tag (name
, strlen (name
), true, buffer
+ linecharno
,
6223 charno
- linecharno
+ 1, lineno
, linecharno
);
6232 nocase_tail (const char *cp
)
6236 while (*cp
!= '\0' && c_tolower (*cp
) == c_tolower (dbp
[len
]))
6238 if (*cp
== '\0' && !intoken (dbp
[len
]))
6247 get_tag (register char *bp
, char **namepp
)
6249 register char *cp
= bp
;
6253 /* Go till you get to white space or a syntactic break */
6254 for (cp
= bp
+ 1; !notinname (*cp
); cp
++)
6256 make_tag (bp
, cp
- bp
, true,
6257 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
6261 *namepp
= savenstr (bp
, cp
- bp
);
6265 * Read a line of text from `stream' into `lbp', excluding the
6266 * newline or CR-NL, if any. Return the number of characters read from
6267 * `stream', which is the length of the line including the newline.
6269 * On DOS or Windows we do not count the CR character, if any before the
6270 * NL, in the returned length; this mirrors the behavior of Emacs on those
6271 * platforms (for text files, it translates CR-NL to NL as it reads in the
6274 * If multi-line regular expressions are requested, each line read is
6275 * appended to `filebuf'.
6278 readline_internal (linebuffer
*lbp
, FILE *stream
, char const *filename
)
6280 char *buffer
= lbp
->buffer
;
6281 char *p
= lbp
->buffer
;
6285 pend
= p
+ lbp
->size
; /* Separate to avoid 386/IX compiler bug. */
6289 register int c
= getc (stream
);
6292 /* We're at the end of linebuffer: expand it. */
6294 xrnew (buffer
, lbp
->size
, char);
6295 p
+= buffer
- lbp
->buffer
;
6296 pend
= buffer
+ lbp
->size
;
6297 lbp
->buffer
= buffer
;
6301 if (ferror (stream
))
6309 if (p
> buffer
&& p
[-1] == '\r')
6323 lbp
->len
= p
- buffer
;
6325 if (need_filebuf
/* we need filebuf for multi-line regexps */
6326 && chars_deleted
> 0) /* not at EOF */
6328 while (filebuf
.size
<= filebuf
.len
+ lbp
->len
+ 1) /* +1 for \n */
6330 /* Expand filebuf. */
6332 xrnew (filebuf
.buffer
, filebuf
.size
, char);
6334 memcpy (filebuf
.buffer
+ filebuf
.len
, lbp
->buffer
, lbp
->len
);
6335 filebuf
.len
+= lbp
->len
;
6336 filebuf
.buffer
[filebuf
.len
++] = '\n';
6337 filebuf
.buffer
[filebuf
.len
] = '\0';
6340 return lbp
->len
+ chars_deleted
;
6344 * Like readline_internal, above, but in addition try to match the
6345 * input line against relevant regular expressions and manage #line
6349 readline (linebuffer
*lbp
, FILE *stream
)
6353 linecharno
= charno
; /* update global char number of line start */
6354 result
= readline_internal (lbp
, stream
, infilename
); /* read line */
6355 lineno
+= 1; /* increment global line number */
6356 charno
+= result
; /* increment global char number */
6358 /* Honor #line directives. */
6359 if (!no_line_directive
)
6361 static bool discard_until_line_directive
;
6363 /* Check whether this is a #line directive. */
6364 if (result
> 12 && strneq (lbp
->buffer
, "#line ", 6))
6369 if (sscanf (lbp
->buffer
, "#line %u \"%n", &lno
, &start
) >= 1
6370 && start
> 0) /* double quote character found */
6372 char *endp
= lbp
->buffer
+ start
;
6374 while ((endp
= strchr (endp
, '"')) != NULL
6375 && endp
[-1] == '\\')
6378 /* Ok, this is a real #line directive. Let's deal with it. */
6380 char *taggedabsname
; /* absolute name of original file */
6381 char *taggedfname
; /* name of original file as given */
6382 char *name
; /* temp var */
6384 discard_until_line_directive
= false; /* found it */
6385 name
= lbp
->buffer
+ start
;
6387 canonicalize_filename (name
);
6388 taggedabsname
= absolute_filename (name
, tagfiledir
);
6389 if (filename_is_absolute (name
)
6390 || filename_is_absolute (curfdp
->infname
))
6391 taggedfname
= savestr (taggedabsname
);
6393 taggedfname
= relative_filename (taggedabsname
,tagfiledir
);
6395 if (streq (curfdp
->taggedfname
, taggedfname
))
6396 /* The #line directive is only a line number change. We
6397 deal with this afterwards. */
6400 /* The tags following this #line directive should be
6401 attributed to taggedfname. In order to do this, set
6402 curfdp accordingly. */
6404 fdesc
*fdp
; /* file description pointer */
6406 /* Go look for a file description already set up for the
6407 file indicated in the #line directive. If there is
6408 one, use it from now until the next #line
6410 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
6411 if (streq (fdp
->infname
, curfdp
->infname
)
6412 && streq (fdp
->taggedfname
, taggedfname
))
6413 /* If we remove the second test above (after the &&)
6414 then all entries pertaining to the same file are
6415 coalesced in the tags file. If we use it, then
6416 entries pertaining to the same file but generated
6417 from different files (via #line directives) will
6418 go into separate sections in the tags file. These
6419 alternatives look equivalent. The first one
6420 destroys some apparently useless information. */
6426 /* Else, if we already tagged the real file, skip all
6427 input lines until the next #line directive. */
6428 if (fdp
== NULL
) /* not found */
6429 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
6430 if (streq (fdp
->infabsname
, taggedabsname
))
6432 discard_until_line_directive
= true;
6436 /* Else create a new file description and use that from
6437 now on, until the next #line directive. */
6438 if (fdp
== NULL
) /* not found */
6441 fdhead
= xnew (1, fdesc
);
6442 *fdhead
= *curfdp
; /* copy curr. file description */
6444 fdhead
->infname
= savestr (curfdp
->infname
);
6445 fdhead
->infabsname
= savestr (curfdp
->infabsname
);
6446 fdhead
->infabsdir
= savestr (curfdp
->infabsdir
);
6447 fdhead
->taggedfname
= taggedfname
;
6448 fdhead
->usecharno
= false;
6449 fdhead
->prop
= NULL
;
6450 fdhead
->written
= false;
6454 free (taggedabsname
);
6456 readline (lbp
, stream
);
6458 } /* if a real #line directive */
6459 } /* if #line is followed by a number */
6460 } /* if line begins with "#line " */
6462 /* If we are here, no #line directive was found. */
6463 if (discard_until_line_directive
)
6467 /* Do a tail recursion on ourselves, thus discarding the contents
6468 of the line buffer. */
6469 readline (lbp
, stream
);
6473 discard_until_line_directive
= false;
6476 } /* if #line directives should be considered */
6483 /* Match against relevant regexps. */
6485 for (rp
= p_head
; rp
!= NULL
; rp
= rp
->p_next
)
6487 /* Only use generic regexps or those for the current language.
6488 Also do not use multiline regexps, which is the job of
6489 regex_tag_multiline. */
6490 if ((rp
->lang
!= NULL
&& rp
->lang
!= fdhead
->lang
)
6494 match
= re_match (rp
->pat
, lbp
->buffer
, lbp
->len
, 0, &rp
->regs
);
6499 if (!rp
->error_signaled
)
6501 error ("regexp stack overflow while matching \"%s\"",
6503 rp
->error_signaled
= true;
6510 /* Empty string matched. */
6511 if (!rp
->error_signaled
)
6513 error ("regexp matches the empty string: \"%s\"", rp
->pattern
);
6514 rp
->error_signaled
= true;
6518 /* Match occurred. Construct a tag. */
6520 if (name
[0] == '\0')
6522 else /* make a named tag */
6523 name
= substitute (lbp
->buffer
, rp
->name
, &rp
->regs
);
6524 if (rp
->force_explicit_name
)
6525 /* Force explicit tag name, if a name is there. */
6526 pfnote (name
, true, lbp
->buffer
, match
, lineno
, linecharno
);
6528 make_tag (name
, strlen (name
), true,
6529 lbp
->buffer
, match
, lineno
, linecharno
);
6538 * Return a pointer to a space of size strlen(cp)+1 allocated
6539 * with xnew where the string CP has been copied.
6542 savestr (const char *cp
)
6544 return savenstr (cp
, strlen (cp
));
6548 * Return a pointer to a space of size LEN+1 allocated with xnew where
6549 * the string CP has been copied for at most the first LEN characters.
6552 savenstr (const char *cp
, int len
)
6554 char *dp
= xnew (len
+ 1, char);
6556 return memcpy (dp
, cp
, len
);
6559 /* Skip spaces (end of string is not space), return new pointer. */
6561 skip_spaces (char *cp
)
6563 while (c_isspace (*cp
))
6568 /* Skip non spaces, except end of string, return new pointer. */
6570 skip_non_spaces (char *cp
)
6572 while (*cp
!= '\0' && !c_isspace (*cp
))
6577 /* Skip any chars in the "name" class.*/
6579 skip_name (char *cp
)
6581 /* '\0' is a notinname() so loop stops there too */
6582 while (! notinname (*cp
))
6587 /* Print error message and exit. */
6589 fatal (char const *format
, ...)
6592 va_start (ap
, format
);
6593 verror (format
, ap
);
6595 exit (EXIT_FAILURE
);
6599 pfatal (const char *s1
)
6602 exit (EXIT_FAILURE
);
6606 suggest_asking_for_help (void)
6608 fprintf (stderr
, "\tTry '%s --help' for a complete list of options.\n",
6610 exit (EXIT_FAILURE
);
6613 /* Output a diagnostic with printf-style FORMAT and args. */
6615 error (const char *format
, ...)
6618 va_start (ap
, format
);
6619 verror (format
, ap
);
6624 verror (char const *format
, va_list ap
)
6626 fprintf (stderr
, "%s: ", progname
);
6627 vfprintf (stderr
, format
, ap
);
6628 fprintf (stderr
, "\n");
6631 /* Return a newly-allocated string whose contents
6632 concatenate those of s1, s2, s3. */
6634 concat (const char *s1
, const char *s2
, const char *s3
)
6636 int len1
= strlen (s1
), len2
= strlen (s2
), len3
= strlen (s3
);
6637 char *result
= xnew (len1
+ len2
+ len3
+ 1, char);
6639 strcpy (result
, s1
);
6640 strcpy (result
+ len1
, s2
);
6641 strcpy (result
+ len1
+ len2
, s3
);
6647 /* Does the same work as the system V getcwd, but does not need to
6648 guess the buffer size in advance. */
6653 char *path
= xnew (bufsize
, char);
6655 while (getcwd (path
, bufsize
) == NULL
)
6657 if (errno
!= ERANGE
)
6661 path
= xnew (bufsize
, char);
6664 canonicalize_filename (path
);
6668 /* Return a newly allocated string containing a name of a temporary file. */
6672 const char *tmpdir
= getenv ("TMPDIR");
6673 const char *slash
= "/";
6675 #if MSDOS || defined (DOS_NT)
6677 tmpdir
= getenv ("TEMP");
6679 tmpdir
= getenv ("TMP");
6682 if (tmpdir
[strlen (tmpdir
) - 1] == '/'
6683 || tmpdir
[strlen (tmpdir
) - 1] == '\\')
6688 if (tmpdir
[strlen (tmpdir
) - 1] == '/')
6692 char *templt
= concat (tmpdir
, slash
, "etXXXXXX");
6693 int fd
= mkostemp (templt
, O_CLOEXEC
);
6694 if (fd
< 0 || close (fd
) != 0)
6696 int temp_errno
= errno
;
6702 #if defined (DOS_NT)
6703 /* The file name will be used in shell redirection, so it needs to have
6704 DOS-style backslashes, or else the Windows shell will barf. */
6706 for (p
= templt
; *p
; p
++)
6714 /* Return a newly allocated string containing the file name of FILE
6715 relative to the absolute directory DIR (which should end with a slash). */
6717 relative_filename (char *file
, char *dir
)
6719 char *fp
, *dp
, *afn
, *res
;
6722 /* Find the common root of file and dir (with a trailing slash). */
6723 afn
= absolute_filename (file
, cwd
);
6726 while (*fp
++ == *dp
++)
6728 fp
--, dp
--; /* back to the first differing char */
6730 if (fp
== afn
&& afn
[0] != '/') /* cannot build a relative name */
6733 do /* look at the equal chars until '/' */
6737 /* Build a sequence of "../" strings for the resulting relative file name. */
6739 while ((dp
= strchr (dp
+ 1, '/')) != NULL
)
6741 res
= xnew (3*i
+ strlen (fp
+ 1) + 1, char);
6744 z
= stpcpy (z
, "../");
6746 /* Add the file name relative to the common root of file and dir. */
6753 /* Return a newly allocated string containing the absolute file name
6754 of FILE given DIR (which should end with a slash). */
6756 absolute_filename (char *file
, char *dir
)
6758 char *slashp
, *cp
, *res
;
6760 if (filename_is_absolute (file
))
6761 res
= savestr (file
);
6763 /* We don't support non-absolute file names with a drive
6764 letter, like `d:NAME' (it's too much hassle). */
6765 else if (file
[1] == ':')
6766 fatal ("%s: relative file names with drive letters not supported", file
);
6769 res
= concat (dir
, file
, "");
6771 /* Delete the "/dirname/.." and "/." substrings. */
6772 slashp
= strchr (res
, '/');
6773 while (slashp
!= NULL
&& slashp
[0] != '\0')
6775 if (slashp
[1] == '.')
6777 if (slashp
[2] == '.'
6778 && (slashp
[3] == '/' || slashp
[3] == '\0'))
6783 while (cp
>= res
&& !filename_is_absolute (cp
));
6785 cp
= slashp
; /* the absolute name begins with "/.." */
6787 /* Under MSDOS and NT we get `d:/NAME' as absolute
6788 file name, so the luser could say `d:/../NAME'.
6789 We silently treat this as `d:/NAME'. */
6790 else if (cp
[0] != '/')
6793 memmove (cp
, slashp
+ 3, strlen (slashp
+ 2));
6797 else if (slashp
[2] == '/' || slashp
[2] == '\0')
6799 memmove (slashp
, slashp
+ 2, strlen (slashp
+ 1));
6804 slashp
= strchr (slashp
+ 1, '/');
6807 if (res
[0] == '\0') /* just a safety net: should never happen */
6810 return savestr ("/");
6816 /* Return a newly allocated string containing the absolute
6817 file name of dir where FILE resides given DIR (which should
6818 end with a slash). */
6820 absolute_dirname (char *file
, char *dir
)
6825 slashp
= strrchr (file
, '/');
6827 return savestr (dir
);
6830 res
= absolute_filename (file
, dir
);
6836 /* Whether the argument string is an absolute file name. The argument
6837 string must have been canonicalized with canonicalize_filename. */
6839 filename_is_absolute (char *fn
)
6841 return (fn
[0] == '/'
6843 || (c_isalpha (fn
[0]) && fn
[1] == ':' && fn
[2] == '/')
6848 /* Downcase DOS drive letter and collapse separators into single slashes.
6851 canonicalize_filename (register char *fn
)
6856 /* Canonicalize drive letter case. */
6857 if (c_isupper (fn
[0]) && fn
[1] == ':')
6858 fn
[0] = c_tolower (fn
[0]);
6860 /* Collapse multiple forward- and back-slashes into a single forward
6862 for (cp
= fn
; *cp
!= '\0'; cp
++, fn
++)
6863 if (*cp
== '/' || *cp
== '\\')
6866 while (cp
[1] == '/' || cp
[1] == '\\')
6874 /* Collapse multiple slashes into a single slash. */
6875 for (cp
= fn
; *cp
!= '\0'; cp
++, fn
++)
6879 while (cp
[1] == '/')
6885 #endif /* !DOS_NT */
6891 /* Initialize a linebuffer for use. */
6893 linebuffer_init (linebuffer
*lbp
)
6895 lbp
->size
= (DEBUG
) ? 3 : 200;
6896 lbp
->buffer
= xnew (lbp
->size
, char);
6897 lbp
->buffer
[0] = '\0';
6901 /* Set the minimum size of a string contained in a linebuffer. */
6903 linebuffer_setlen (linebuffer
*lbp
, int toksize
)
6905 while (lbp
->size
<= toksize
)
6908 xrnew (lbp
->buffer
, lbp
->size
, char);
6913 /* Like malloc but get fatal error if memory is exhausted. */
6915 xmalloc (size_t size
)
6917 void *result
= malloc (size
);
6919 fatal ("virtual memory exhausted");
6924 xrealloc (void *ptr
, size_t size
)
6926 void *result
= realloc (ptr
, size
);
6928 fatal ("virtual memory exhausted");
6934 * indent-tabs-mode: t
6937 * c-font-lock-extra-types: ("FILE" "bool" "language" "linebuffer" "fdesc" "node" "regexp")
6938 * c-file-style: "gnu"
6942 /* etags.c ends here */