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", "ru", "rbw", NULL
};
733 static const char *Ruby_filenames
[] =
734 { "Rakefile", "Thorfile", NULL
};
735 static const char Ruby_help
[] =
736 "In Ruby code, 'def' or 'class' or 'module' at the beginning of\n\
737 a line generate a tag. Constants also generate a tag.";
739 /* Can't do the `SCM' or `scm' prefix with a version number. */
740 static const char *Scheme_suffixes
[] =
741 { "oak", "sch", "scheme", "SCM", "scm", "SM", "sm", "ss", "t", NULL
};
742 static const char Scheme_help
[] =
743 "In Scheme code, tags include anything defined with 'def' or with a\n\
744 construct whose name starts with 'def'. They also include\n\
745 variables set with 'set!' at top level in the file.";
747 static const char *TeX_suffixes
[] =
748 { "bib", "clo", "cls", "ltx", "sty", "TeX", "tex", NULL
};
749 static const char TeX_help
[] =
750 "In LaTeX text, the argument of any of the commands '\\chapter',\n\
751 '\\section', '\\subsection', '\\subsubsection', '\\eqno', '\\label',\n\
752 '\\ref', '\\cite', '\\bibitem', '\\part', '\\appendix', '\\entry',\n\
753 '\\index', '\\def', '\\newcommand', '\\renewcommand',\n\
754 '\\newenvironment' or '\\renewenvironment' is a tag.\n\
756 Other commands can be specified by setting the environment variable\n\
757 'TEXTAGS' to a colon-separated list like, for example,\n\
758 TEXTAGS=\"mycommand:myothercommand\".";
761 static const char *Texinfo_suffixes
[] =
762 { "texi", "texinfo", "txi", NULL
};
763 static const char Texinfo_help
[] =
764 "for texinfo files, lines starting with @node are tagged.";
766 static const char *Yacc_suffixes
[] =
767 { "y", "y++", "ym", "yxx", "yy", NULL
}; /* .ym is Objective yacc file */
768 static const char Yacc_help
[] =
769 "In Bison or Yacc input files, each rule defines as a tag the\n\
770 nonterminal it constructs. The portions of the file that contain\n\
771 C code are parsed as C code (use --help --lang=c --lang=yacc\n\
774 static const char auto_help
[] =
775 "'auto' is not a real language, it indicates to use\n\
776 a default language for files base on file name suffix and file contents.";
778 static const char none_help
[] =
779 "'none' is not a real language, it indicates to only do\n\
780 regexp processing on files.";
782 static const char no_lang_help
[] =
783 "No detailed help available for this language.";
787 * Table of languages.
789 * It is ok for a given function to be listed under more than one
790 * name. I just didn't.
793 static language lang_names
[] =
795 { "ada", Ada_help
, Ada_funcs
, Ada_suffixes
},
796 { "asm", Asm_help
, Asm_labels
, Asm_suffixes
},
797 { "c", default_C_help
, default_C_entries
, default_C_suffixes
},
798 { "c++", Cplusplus_help
, Cplusplus_entries
, Cplusplus_suffixes
},
799 { "c*", no_lang_help
, Cstar_entries
, Cstar_suffixes
},
800 { "cobol", Cobol_help
, Cobol_paragraphs
, Cobol_suffixes
},
801 { "erlang", Erlang_help
, Erlang_functions
, Erlang_suffixes
},
802 { "forth", Forth_help
, Forth_words
, Forth_suffixes
},
803 { "fortran", Fortran_help
, Fortran_functions
, Fortran_suffixes
},
804 { "go", Go_help
, Go_functions
, Go_suffixes
},
805 { "html", HTML_help
, HTML_labels
, HTML_suffixes
},
806 { "java", Cjava_help
, Cjava_entries
, Cjava_suffixes
},
807 { "lisp", Lisp_help
, Lisp_functions
, Lisp_suffixes
},
808 { "lua", Lua_help
, Lua_functions
, Lua_suffixes
},
809 { "makefile", Makefile_help
,Makefile_targets
,NULL
,Makefile_filenames
},
810 { "objc", Objc_help
, plain_C_entries
, Objc_suffixes
},
811 { "pascal", Pascal_help
, Pascal_functions
, Pascal_suffixes
},
812 { "perl",Perl_help
,Perl_functions
,Perl_suffixes
,NULL
,Perl_interpreters
},
813 { "php", PHP_help
, PHP_functions
, PHP_suffixes
},
814 { "postscript",PS_help
, PS_functions
, PS_suffixes
},
815 { "proc", no_lang_help
, plain_C_entries
, plain_C_suffixes
},
816 { "prolog", Prolog_help
, Prolog_functions
, Prolog_suffixes
},
817 { "python", Python_help
, Python_functions
, Python_suffixes
},
818 { "ruby", Ruby_help
,Ruby_functions
,Ruby_suffixes
,Ruby_filenames
},
819 { "scheme", Scheme_help
, Scheme_functions
, Scheme_suffixes
},
820 { "tex", TeX_help
, TeX_commands
, TeX_suffixes
},
821 { "texinfo", Texinfo_help
, Texinfo_nodes
, Texinfo_suffixes
},
822 { "yacc", Yacc_help
,Yacc_entries
,Yacc_suffixes
,NULL
,NULL
,true},
823 { "auto", auto_help
}, /* default guessing scheme */
824 { "none", none_help
, just_read_file
}, /* regexp matching only */
825 { NULL
} /* end of list */
830 print_language_names (void)
833 const char **name
, **ext
;
835 puts ("\nThese are the currently supported languages, along with the\n\
836 default file names and dot suffixes:");
837 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
839 printf (" %-*s", 10, lang
->name
);
840 if (lang
->filenames
!= NULL
)
841 for (name
= lang
->filenames
; *name
!= NULL
; name
++)
842 printf (" %s", *name
);
843 if (lang
->suffixes
!= NULL
)
844 for (ext
= lang
->suffixes
; *ext
!= NULL
; ext
++)
845 printf (" .%s", *ext
);
848 puts ("where 'auto' means use default language for files based on file\n\
849 name suffix, and 'none' means only do regexp processing on files.\n\
850 If no language is specified and no matching suffix is found,\n\
851 the first line of the file is read for a sharp-bang (#!) sequence\n\
852 followed by the name of an interpreter. If no such sequence is found,\n\
853 Fortran is tried first; if no tags are found, C is tried next.\n\
854 When parsing any C file, a \"class\" or \"template\" keyword\n\
856 puts ("Compressed files are supported using gzip, bzip2, and xz.\n\
858 For detailed help on a given language use, for example,\n\
859 etags --help --lang=ada.");
863 # define EMACS_NAME "standalone"
866 # define VERSION "17.38.1.4"
868 static _Noreturn
void
871 char emacs_copyright
[] = COPYRIGHT
;
873 printf ("%s (%s %s)\n", (CTAGS
) ? "ctags" : "etags", EMACS_NAME
, VERSION
);
874 puts (emacs_copyright
);
875 puts ("This program is distributed under the terms in ETAGS.README");
880 #ifndef PRINT_UNDOCUMENTED_OPTIONS_HELP
881 # define PRINT_UNDOCUMENTED_OPTIONS_HELP false
884 static _Noreturn
void
885 print_help (argument
*argbuffer
)
887 bool help_for_lang
= false;
889 for (; argbuffer
->arg_type
!= at_end
; argbuffer
++)
890 if (argbuffer
->arg_type
== at_language
)
894 puts (argbuffer
->lang
->help
);
895 help_for_lang
= true;
901 printf ("Usage: %s [options] [[regex-option ...] file-name] ...\n\
903 These are the options accepted by %s.\n", progname
, progname
);
904 puts ("You may use unambiguous abbreviations for the long option names.");
905 puts (" A - as file name means read names from stdin (one per line).\n\
906 Absolute names are stored in the output file as they are.\n\
907 Relative ones are stored relative to the output file's directory.\n");
909 puts ("-a, --append\n\
910 Append tag entries to existing tags file.");
912 puts ("--packages-only\n\
913 For Ada files, only generate tags for packages.");
916 puts ("-B, --backward-search\n\
917 Write the search commands for the tag entries using '?', the\n\
918 backward-search command instead of '/', the forward-search command.");
920 /* This option is mostly obsolete, because etags can now automatically
921 detect C++. Retained for backward compatibility and for debugging and
922 experimentation. In principle, we could want to tag as C++ even
923 before any "class" or "template" keyword.
925 Treat files whose name suffix defaults to C language as C++ files.");
928 puts ("--declarations\n\
929 In C and derived languages, create tags for function declarations,");
931 puts ("\tand create tags for extern variables if --globals is used.");
934 ("\tand create tags for extern variables unless --no-globals is used.");
937 puts ("-d, --defines\n\
938 Create tag entries for C #define constants and enum constants, too.");
940 puts ("-D, --no-defines\n\
941 Don't create tag entries for C #define constants and enum constants.\n\
942 This makes the tags file smaller.");
945 puts ("-i FILE, --include=FILE\n\
946 Include a note in tag file indicating that, when searching for\n\
947 a tag, one should also consult the tags file FILE after\n\
948 checking the current file.");
950 puts ("-l LANG, --language=LANG\n\
951 Force the following files to be considered as written in the\n\
952 named language up to the next --language=LANG option.");
956 Create tag entries for global variables in some languages.");
958 puts ("--no-globals\n\
959 Do not create tag entries for global variables in some\n\
960 languages. This makes the tags file smaller.");
962 puts ("--no-line-directive\n\
963 Ignore #line preprocessor directives in C and derived languages.");
967 Create tag entries for members of structures in some languages.");
969 puts ("--no-members\n\
970 Do not create tag entries for members of structures\n\
971 in some languages.");
973 puts ("-Q, --class-qualify\n\
974 Qualify tag names with their class name in C++, ObjC, and Java.\n\
975 This produces tag names of the form \"class::member\" for C++,\n\
976 \"class(category)\" for Objective C, and \"class.member\" for Java.\n\
977 For Objective C, this also produces class methods qualified with\n\
978 their arguments, as in \"foo:bar:baz:more\".");
979 puts ("-r REGEXP, --regex=REGEXP or --regex=@regexfile\n\
980 Make a tag for each line matching a regular expression pattern\n\
981 in the following files. {LANGUAGE}REGEXP uses REGEXP for LANGUAGE\n\
982 files only. REGEXFILE is a file containing one REGEXP per line.\n\
983 REGEXP takes the form /TAGREGEXP/TAGNAME/MODS, where TAGNAME/ is\n\
984 optional. The TAGREGEXP pattern is anchored (as if preceded by ^).");
985 puts (" If TAGNAME/ is present, the tags created are named.\n\
986 For example Tcl named tags can be created with:\n\
987 --regex=\"/proc[ \\t]+\\([^ \\t]+\\)/\\1/.\".\n\
988 MODS are optional one-letter modifiers: 'i' means to ignore case,\n\
989 'm' means to allow multi-line matches, 's' implies 'm' and\n\
990 causes dot to match any character, including newline.");
992 puts ("-R, --no-regex\n\
993 Don't create tags from regexps for the following files.");
995 puts ("-I, --ignore-indentation\n\
996 In C and C++ do not assume that a closing brace in the first\n\
997 column is the final brace of a function or structure definition.");
999 puts ("-o FILE, --output=FILE\n\
1000 Write the tags to FILE.");
1002 puts ("--parse-stdin=NAME\n\
1003 Read from standard input and record tags as belonging to file NAME.");
1007 puts ("-t, --typedefs\n\
1008 Generate tag entries for C and Ada typedefs.");
1009 puts ("-T, --typedefs-and-c++\n\
1010 Generate tag entries for C typedefs, C struct/enum/union tags,\n\
1011 and C++ member functions.");
1015 puts ("-u, --update\n\
1016 Update the tag entries for the given files, leaving tag\n\
1017 entries for other files in place. Currently, this is\n\
1018 implemented by deleting the existing entries for the given\n\
1019 files and then rewriting the new entries at the end of the\n\
1020 tags file. It is often faster to simply rebuild the entire\n\
1021 tag file than to use this.");
1025 puts ("-v, --vgrind\n\
1026 Print on the standard output an index of items intended for\n\
1027 human consumption, similar to the output of vgrind. The index\n\
1028 is sorted, and gives the page number of each item.");
1030 if (PRINT_UNDOCUMENTED_OPTIONS_HELP
)
1031 puts ("-w, --no-duplicates\n\
1032 Do not create duplicate tag entries, for compatibility with\n\
1033 traditional ctags.");
1035 if (PRINT_UNDOCUMENTED_OPTIONS_HELP
)
1036 puts ("-w, --no-warn\n\
1037 Suppress warning messages about duplicate tag entries.");
1039 puts ("-x, --cxref\n\
1040 Like --vgrind, but in the style of cxref, rather than vgrind.\n\
1041 The output uses line numbers instead of page numbers, but\n\
1042 beyond that the differences are cosmetic; try both to see\n\
1046 puts ("-V, --version\n\
1047 Print the version of the program.\n\
1049 Print this help message.\n\
1050 Followed by one or more '--language' options prints detailed\n\
1051 help about tag generation for the specified languages.");
1053 print_language_names ();
1056 puts ("Report bugs to bug-gnu-emacs@gnu.org");
1058 exit (EXIT_SUCCESS
);
1063 main (int argc
, char **argv
)
1066 unsigned int nincluded_files
;
1067 char **included_files
;
1068 argument
*argbuffer
;
1069 int current_arg
, file_count
;
1070 linebuffer filename_lb
;
1071 bool help_asked
= false;
1077 nincluded_files
= 0;
1078 included_files
= xnew (argc
, char *);
1082 /* Allocate enough no matter what happens. Overkill, but each one
1084 argbuffer
= xnew (argc
, argument
);
1087 * Always find typedefs and structure tags.
1088 * Also default to find macro constants, enum constants, struct
1089 * members and global variables. Do it for both etags and ctags.
1091 typedefs
= typedefs_or_cplusplus
= constantypedefs
= true;
1092 globals
= members
= true;
1094 /* When the optstring begins with a '-' getopt_long does not rearrange the
1095 non-options arguments to be at the end, but leaves them alone. */
1096 optstring
= concat ("-ac:Cf:Il:o:Qr:RSVhH",
1097 (CTAGS
) ? "BxdtTuvw" : "Di:",
1100 while ((opt
= getopt_long (argc
, argv
, optstring
, longopts
, NULL
)) != EOF
)
1104 /* If getopt returns 0, then it has already processed a
1105 long-named option. We should do nothing. */
1109 /* This means that a file name has been seen. Record it. */
1110 argbuffer
[current_arg
].arg_type
= at_filename
;
1111 argbuffer
[current_arg
].what
= optarg
;
1112 len
= strlen (optarg
);
1113 if (whatlen_max
< len
)
1120 /* Parse standard input. Idea by Vivek <vivek@etla.org>. */
1121 argbuffer
[current_arg
].arg_type
= at_stdin
;
1122 argbuffer
[current_arg
].what
= optarg
;
1123 len
= strlen (optarg
);
1124 if (whatlen_max
< len
)
1129 fatal ("cannot parse standard input more than once");
1130 parsing_stdin
= true;
1133 /* Common options. */
1134 case 'a': append_to_tagfile
= true; break;
1135 case 'C': cplusplus
= true; break;
1136 case 'f': /* for compatibility with old makefiles */
1140 error ("-o option may only be given once.");
1141 suggest_asking_for_help ();
1147 case 'S': /* for backward compatibility */
1148 ignoreindent
= true;
1152 language
*lang
= get_language_from_langname (optarg
);
1155 argbuffer
[current_arg
].lang
= lang
;
1156 argbuffer
[current_arg
].arg_type
= at_language
;
1162 /* Backward compatibility: support obsolete --ignore-case-regexp. */
1163 optarg
= concat (optarg
, "i", ""); /* memory leak here */
1166 argbuffer
[current_arg
].arg_type
= at_regexp
;
1167 argbuffer
[current_arg
].what
= optarg
;
1168 len
= strlen (optarg
);
1169 if (whatlen_max
< len
)
1174 argbuffer
[current_arg
].arg_type
= at_regexp
;
1175 argbuffer
[current_arg
].what
= NULL
;
1190 case 'D': constantypedefs
= false; break;
1191 case 'i': included_files
[nincluded_files
++] = optarg
; break;
1193 /* Ctags options. */
1194 case 'B': searchar
= '?'; break;
1195 case 'd': constantypedefs
= true; break;
1196 case 't': typedefs
= true; break;
1197 case 'T': typedefs
= typedefs_or_cplusplus
= true; break;
1198 case 'u': update
= true; break;
1199 case 'v': vgrind_style
= true; /*FALLTHRU*/
1200 case 'x': cxref_style
= true; break;
1201 case 'w': no_warnings
= true; break;
1203 suggest_asking_for_help ();
1207 /* No more options. Store the rest of arguments. */
1208 for (; optind
< argc
; optind
++)
1210 argbuffer
[current_arg
].arg_type
= at_filename
;
1211 argbuffer
[current_arg
].what
= argv
[optind
];
1212 len
= strlen (argv
[optind
]);
1213 if (whatlen_max
< len
)
1219 argbuffer
[current_arg
].arg_type
= at_end
;
1222 print_help (argbuffer
);
1225 if (nincluded_files
== 0 && file_count
== 0)
1227 error ("no input files specified.");
1228 suggest_asking_for_help ();
1232 if (tagfile
== NULL
)
1233 tagfile
= savestr (CTAGS
? "tags" : "TAGS");
1234 cwd
= etags_getcwd (); /* the current working directory */
1235 if (cwd
[strlen (cwd
) - 1] != '/')
1238 cwd
= concat (oldcwd
, "/", "");
1242 /* Compute base directory for relative file names. */
1243 if (streq (tagfile
, "-")
1244 || strneq (tagfile
, "/dev/", 5))
1245 tagfiledir
= cwd
; /* relative file names are relative to cwd */
1248 canonicalize_filename (tagfile
);
1249 tagfiledir
= absolute_dirname (tagfile
, cwd
);
1252 linebuffer_init (&lb
);
1253 linebuffer_init (&filename_lb
);
1254 linebuffer_init (&filebuf
);
1255 linebuffer_init (&token_name
);
1259 if (streq (tagfile
, "-"))
1262 SET_BINARY (fileno (stdout
));
1265 tagf
= fopen (tagfile
, append_to_tagfile
? "ab" : "wb");
1271 * Loop through files finding functions.
1273 for (i
= 0; i
< current_arg
; i
++)
1275 static language
*lang
; /* non-NULL if language is forced */
1278 switch (argbuffer
[i
].arg_type
)
1281 lang
= argbuffer
[i
].lang
;
1284 analyze_regex (argbuffer
[i
].what
);
1287 this_file
= argbuffer
[i
].what
;
1288 /* Input file named "-" means read file names from stdin
1289 (one per line) and use them. */
1290 if (streq (this_file
, "-"))
1293 fatal ("cannot parse standard input "
1294 "AND read file names from it");
1295 while (readline_internal (&filename_lb
, stdin
, "-") > 0)
1296 process_file_name (filename_lb
.buffer
, lang
);
1299 process_file_name (this_file
, lang
);
1302 this_file
= argbuffer
[i
].what
;
1303 process_file (stdin
, this_file
, lang
);
1306 error ("internal error: arg_type");
1312 free (filebuf
.buffer
);
1313 free (token_name
.buffer
);
1315 if (!CTAGS
|| cxref_style
)
1317 /* Write the remaining tags to tagf (ETAGS) or stdout (CXREF). */
1318 put_entries (nodehead
);
1319 free_tree (nodehead
);
1325 /* Output file entries that have no tags. */
1326 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
1328 fprintf (tagf
, "\f\n%s,0\n", fdp
->taggedfname
);
1330 while (nincluded_files
-- > 0)
1331 fprintf (tagf
, "\f\n%s,include\n", *included_files
++);
1333 if (fclose (tagf
) == EOF
)
1337 exit (EXIT_SUCCESS
);
1340 /* From here on, we are in (CTAGS && !cxref_style) */
1344 xmalloc (strlen (tagfile
) + whatlen_max
+
1345 sizeof "mv..OTAGS;fgrep -v '\t\t' OTAGS >;rm OTAGS");
1346 for (i
= 0; i
< current_arg
; ++i
)
1348 switch (argbuffer
[i
].arg_type
)
1354 continue; /* the for loop */
1356 char *z
= stpcpy (cmd
, "mv ");
1357 z
= stpcpy (z
, tagfile
);
1358 z
= stpcpy (z
, " OTAGS;fgrep -v '\t");
1359 z
= stpcpy (z
, argbuffer
[i
].what
);
1360 z
= stpcpy (z
, "\t' OTAGS >");
1361 z
= stpcpy (z
, tagfile
);
1362 strcpy (z
, ";rm OTAGS");
1363 if (system (cmd
) != EXIT_SUCCESS
)
1364 fatal ("failed to execute shell command");
1367 append_to_tagfile
= true;
1370 tagf
= fopen (tagfile
, append_to_tagfile
? "ab" : "wb");
1373 put_entries (nodehead
); /* write all the tags (CTAGS) */
1374 free_tree (nodehead
);
1376 if (fclose (tagf
) == EOF
)
1380 if (append_to_tagfile
|| update
)
1382 char *cmd
= xmalloc (2 * strlen (tagfile
) + sizeof "sort -u -o..");
1383 /* Maybe these should be used:
1384 setenv ("LC_COLLATE", "C", 1);
1385 setenv ("LC_ALL", "C", 1); */
1386 char *z
= stpcpy (cmd
, "sort -u -o ");
1387 z
= stpcpy (z
, tagfile
);
1389 strcpy (z
, tagfile
);
1390 exit (system (cmd
));
1392 return EXIT_SUCCESS
;
1397 * Return a compressor given the file name. If EXTPTR is non-zero,
1398 * return a pointer into FILE where the compressor-specific
1399 * extension begins. If no compressor is found, NULL is returned
1400 * and EXTPTR is not significant.
1401 * Idea by Vladimir Alexiev <vladimir@cs.ualberta.ca> (1998)
1404 get_compressor_from_suffix (char *file
, char **extptr
)
1407 char *slash
, *suffix
;
1409 /* File has been processed by canonicalize_filename,
1410 so we don't need to consider backslashes on DOS_NT. */
1411 slash
= strrchr (file
, '/');
1412 suffix
= strrchr (file
, '.');
1413 if (suffix
== NULL
|| suffix
< slash
)
1418 /* Let those poor souls who live with DOS 8+3 file name limits get
1419 some solace by treating foo.cgz as if it were foo.c.gz, etc.
1420 Only the first do loop is run if not MSDOS */
1423 for (compr
= compressors
; compr
->suffix
!= NULL
; compr
++)
1424 if (streq (compr
->suffix
, suffix
))
1427 break; /* do it only once: not really a loop */
1430 } while (*suffix
!= '\0');
1437 * Return a language given the name.
1440 get_language_from_langname (const char *name
)
1445 error ("empty language name");
1448 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1449 if (streq (name
, lang
->name
))
1451 error ("unknown language \"%s\"", name
);
1459 * Return a language given the interpreter name.
1462 get_language_from_interpreter (char *interpreter
)
1467 if (interpreter
== NULL
)
1469 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1470 if (lang
->interpreters
!= NULL
)
1471 for (iname
= lang
->interpreters
; *iname
!= NULL
; iname
++)
1472 if (streq (*iname
, interpreter
))
1481 * Return a language given the file name.
1484 get_language_from_filename (char *file
, int case_sensitive
)
1487 const char **name
, **ext
, *suffix
;
1489 /* Try whole file name first. */
1490 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1491 if (lang
->filenames
!= NULL
)
1492 for (name
= lang
->filenames
; *name
!= NULL
; name
++)
1493 if ((case_sensitive
)
1494 ? streq (*name
, file
)
1495 : strcaseeq (*name
, file
))
1498 /* If not found, try suffix after last dot. */
1499 suffix
= strrchr (file
, '.');
1503 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1504 if (lang
->suffixes
!= NULL
)
1505 for (ext
= lang
->suffixes
; *ext
!= NULL
; ext
++)
1506 if ((case_sensitive
)
1507 ? streq (*ext
, suffix
)
1508 : strcaseeq (*ext
, suffix
))
1515 * This routine is called on each file argument.
1518 process_file_name (char *file
, language
*lang
)
1523 char *compressed_name
, *uncompressed_name
;
1524 char *ext
, *real_name
, *tmp_name
;
1527 canonicalize_filename (file
);
1528 if (streq (file
, tagfile
) && !streq (tagfile
, "-"))
1530 error ("skipping inclusion of %s in self.", file
);
1533 compr
= get_compressor_from_suffix (file
, &ext
);
1536 compressed_name
= file
;
1537 uncompressed_name
= savenstr (file
, ext
- file
);
1541 compressed_name
= NULL
;
1542 uncompressed_name
= file
;
1545 /* If the canonicalized uncompressed name
1546 has already been dealt with, skip it silently. */
1547 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
1549 assert (fdp
->infname
!= NULL
);
1550 if (streq (uncompressed_name
, fdp
->infname
))
1554 inf
= fopen (file
, "r" FOPEN_BINARY
);
1559 int file_errno
= errno
;
1560 if (compressed_name
)
1562 /* Try with the given suffix. */
1563 inf
= fopen (uncompressed_name
, "r" FOPEN_BINARY
);
1565 real_name
= uncompressed_name
;
1569 /* Try all possible suffixes. */
1570 for (compr
= compressors
; compr
->suffix
!= NULL
; compr
++)
1572 compressed_name
= concat (file
, ".", compr
->suffix
);
1573 inf
= fopen (compressed_name
, "r" FOPEN_BINARY
);
1576 real_name
= compressed_name
;
1581 char *suf
= compressed_name
+ strlen (file
);
1582 size_t suflen
= strlen (compr
->suffix
) + 1;
1583 for ( ; suf
[1]; suf
++, suflen
--)
1585 memmove (suf
, suf
+ 1, suflen
);
1586 inf
= fopen (compressed_name
, "r" FOPEN_BINARY
);
1589 real_name
= compressed_name
;
1596 free (compressed_name
);
1597 compressed_name
= NULL
;
1608 if (real_name
== compressed_name
)
1611 tmp_name
= etags_mktmp ();
1616 #if MSDOS || defined (DOS_NT)
1617 char *cmd1
= concat (compr
->command
, " \"", real_name
);
1618 char *cmd
= concat (cmd1
, "\" > ", tmp_name
);
1620 char *cmd1
= concat (compr
->command
, " '", real_name
);
1621 char *cmd
= concat (cmd1
, "' > ", tmp_name
);
1625 if (system (cmd
) == -1)
1632 inf
= fopen (tmp_name
, "r" FOPEN_BINARY
);
1646 process_file (inf
, uncompressed_name
, lang
);
1648 retval
= fclose (inf
);
1649 if (real_name
== compressed_name
)
1658 if (compressed_name
!= file
)
1659 free (compressed_name
);
1660 if (uncompressed_name
!= file
)
1661 free (uncompressed_name
);
1668 process_file (FILE *fh
, char *fn
, language
*lang
)
1670 static const fdesc emptyfdesc
;
1674 /* Create a new input file description entry. */
1675 fdp
= xnew (1, fdesc
);
1678 fdp
->infname
= savestr (fn
);
1680 fdp
->infabsname
= absolute_filename (fn
, cwd
);
1681 fdp
->infabsdir
= absolute_dirname (fn
, cwd
);
1682 if (filename_is_absolute (fn
))
1684 /* An absolute file name. Canonicalize it. */
1685 fdp
->taggedfname
= absolute_filename (fn
, NULL
);
1689 /* A file name relative to cwd. Make it relative
1690 to the directory of the tags file. */
1691 fdp
->taggedfname
= relative_filename (fn
, tagfiledir
);
1693 fdp
->usecharno
= true; /* use char position when making tags */
1695 fdp
->written
= false; /* not written on tags file yet */
1698 curfdp
= fdhead
; /* the current file description */
1702 /* If not Ctags, and if this is not metasource and if it contained no #line
1703 directives, we can write the tags and free all nodes pointing to
1706 && curfdp
->usecharno
/* no #line directives in this file */
1707 && !curfdp
->lang
->metasource
)
1711 /* Look for the head of the sublist relative to this file. See add_node
1712 for the structure of the node tree. */
1714 for (np
= nodehead
; np
!= NULL
; prev
= np
, np
= np
->left
)
1715 if (np
->fdp
== curfdp
)
1718 /* If we generated tags for this file, write and delete them. */
1721 /* This is the head of the last sublist, if any. The following
1722 instructions depend on this being true. */
1723 assert (np
->left
== NULL
);
1725 assert (fdhead
== curfdp
);
1726 assert (last_node
->fdp
== curfdp
);
1727 put_entries (np
); /* write tags for file curfdp->taggedfname */
1728 free_tree (np
); /* remove the written nodes */
1730 nodehead
= NULL
; /* no nodes left */
1732 prev
->left
= NULL
; /* delete the pointer to the sublist */
1738 reset_input (FILE *inf
)
1740 if (fseek (inf
, 0, SEEK_SET
) != 0)
1741 perror (infilename
);
1745 * This routine opens the specified file and calls the function
1746 * which finds the function and type definitions.
1749 find_entries (FILE *inf
)
1752 language
*lang
= curfdp
->lang
;
1753 Lang_function
*parser
= NULL
;
1755 /* If user specified a language, use it. */
1756 if (lang
!= NULL
&& lang
->function
!= NULL
)
1758 parser
= lang
->function
;
1761 /* Else try to guess the language given the file name. */
1764 lang
= get_language_from_filename (curfdp
->infname
, true);
1765 if (lang
!= NULL
&& lang
->function
!= NULL
)
1767 curfdp
->lang
= lang
;
1768 parser
= lang
->function
;
1772 /* Else look for sharp-bang as the first two characters. */
1774 && readline_internal (&lb
, inf
, infilename
) > 0
1776 && lb
.buffer
[0] == '#'
1777 && lb
.buffer
[1] == '!')
1781 /* Set lp to point at the first char after the last slash in the
1782 line or, if no slashes, at the first nonblank. Then set cp to
1783 the first successive blank and terminate the string. */
1784 lp
= strrchr (lb
.buffer
+2, '/');
1788 lp
= skip_spaces (lb
.buffer
+ 2);
1789 cp
= skip_non_spaces (lp
);
1792 if (strlen (lp
) > 0)
1794 lang
= get_language_from_interpreter (lp
);
1795 if (lang
!= NULL
&& lang
->function
!= NULL
)
1797 curfdp
->lang
= lang
;
1798 parser
= lang
->function
;
1805 /* Else try to guess the language given the case insensitive file name. */
1808 lang
= get_language_from_filename (curfdp
->infname
, false);
1809 if (lang
!= NULL
&& lang
->function
!= NULL
)
1811 curfdp
->lang
= lang
;
1812 parser
= lang
->function
;
1816 /* Else try Fortran or C. */
1819 node
*old_last_node
= last_node
;
1821 curfdp
->lang
= get_language_from_langname ("fortran");
1824 if (old_last_node
== last_node
)
1825 /* No Fortran entries found. Try C. */
1828 curfdp
->lang
= get_language_from_langname (cplusplus
? "c++" : "c");
1834 if (!no_line_directive
1835 && curfdp
->lang
!= NULL
&& curfdp
->lang
->metasource
)
1836 /* It may be that this is a bingo.y file, and we already parsed a bingo.c
1837 file, or anyway we parsed a file that is automatically generated from
1838 this one. If this is the case, the bingo.c file contained #line
1839 directives that generated tags pointing to this file. Let's delete
1840 them all before parsing this file, which is the real source. */
1842 fdesc
**fdpp
= &fdhead
;
1843 while (*fdpp
!= NULL
)
1845 && streq ((*fdpp
)->taggedfname
, curfdp
->taggedfname
))
1846 /* We found one of those! We must delete both the file description
1847 and all tags referring to it. */
1849 fdesc
*badfdp
= *fdpp
;
1851 /* Delete the tags referring to badfdp->taggedfname
1852 that were obtained from badfdp->infname. */
1853 invalidate_nodes (badfdp
, &nodehead
);
1855 *fdpp
= badfdp
->next
; /* remove the bad description from the list */
1856 free_fdesc (badfdp
);
1859 fdpp
= &(*fdpp
)->next
; /* advance the list pointer */
1862 assert (parser
!= NULL
);
1864 /* Generic initializations before reading from file. */
1865 linebuffer_setlen (&filebuf
, 0); /* reset the file buffer */
1867 /* Generic initializations before parsing file with readline. */
1868 lineno
= 0; /* reset global line number */
1869 charno
= 0; /* reset global char number */
1870 linecharno
= 0; /* reset global char number of line start */
1874 regex_tag_multiline ();
1879 * Check whether an implicitly named tag should be created,
1880 * then call `pfnote'.
1881 * NAME is a string that is internally copied by this function.
1883 * TAGS format specification
1884 * Idea by Sam Kendall <kendall@mv.mv.com> (1997)
1885 * The following is explained in some more detail in etc/ETAGS.EBNF.
1887 * make_tag creates tags with "implicit tag names" (unnamed tags)
1888 * if the following are all true, assuming NONAM=" \f\t\n\r()=,;":
1889 * 1. NAME does not contain any of the characters in NONAM;
1890 * 2. LINESTART contains name as either a rightmost, or rightmost but
1891 * one character, substring;
1892 * 3. the character, if any, immediately before NAME in LINESTART must
1893 * be a character in NONAM;
1894 * 4. the character, if any, immediately after NAME in LINESTART must
1895 * also be a character in NONAM.
1897 * The implementation uses the notinname() macro, which recognizes the
1898 * characters stored in the string `nonam'.
1899 * etags.el needs to use the same characters that are in NONAM.
1902 make_tag (const char *name
, /* tag name, or NULL if unnamed */
1903 int namelen
, /* tag length */
1904 bool is_func
, /* tag is a function */
1905 char *linestart
, /* start of the line where tag is */
1906 int linelen
, /* length of the line where tag is */
1907 int lno
, /* line number */
1908 long int cno
) /* character number */
1910 bool named
= (name
!= NULL
&& namelen
> 0);
1913 if (!CTAGS
&& named
) /* maybe set named to false */
1914 /* Let's try to make an implicit tag name, that is, create an unnamed tag
1915 such that etags.el can guess a name from it. */
1918 register const char *cp
= name
;
1920 for (i
= 0; i
< namelen
; i
++)
1921 if (notinname (*cp
++))
1923 if (i
== namelen
) /* rule #1 */
1925 cp
= linestart
+ linelen
- namelen
;
1926 if (notinname (linestart
[linelen
-1]))
1927 cp
-= 1; /* rule #4 */
1928 if (cp
>= linestart
/* rule #2 */
1930 || notinname (cp
[-1])) /* rule #3 */
1931 && strneq (name
, cp
, namelen
)) /* rule #2 */
1932 named
= false; /* use implicit tag name */
1937 nname
= savenstr (name
, namelen
);
1939 pfnote (nname
, is_func
, linestart
, linelen
, lno
, cno
);
1944 pfnote (char *name
, bool is_func
, char *linestart
, int linelen
, int lno
,
1946 /* tag name, or NULL if unnamed */
1947 /* tag is a function */
1948 /* start of the line where tag is */
1949 /* length of the line where tag is */
1951 /* character number */
1955 assert (name
== NULL
|| name
[0] != '\0');
1956 if (CTAGS
&& name
== NULL
)
1959 np
= xnew (1, node
);
1961 /* If ctags mode, change name "main" to M<thisfilename>. */
1962 if (CTAGS
&& !cxref_style
&& streq (name
, "main"))
1964 char *fp
= strrchr (curfdp
->taggedfname
, '/');
1965 np
->name
= concat ("M", fp
== NULL
? curfdp
->taggedfname
: fp
+ 1, "");
1966 fp
= strrchr (np
->name
, '.');
1967 if (fp
!= NULL
&& fp
[1] != '\0' && fp
[2] == '\0')
1973 np
->been_warned
= false;
1975 np
->is_func
= is_func
;
1977 if (np
->fdp
->usecharno
)
1978 /* Our char numbers are 0-base, because of C language tradition?
1979 ctags compatibility? old versions compatibility? I don't know.
1980 Anyway, since emacs's are 1-base we expect etags.el to take care
1981 of the difference. If we wanted to have 1-based numbers, we would
1982 uncomment the +1 below. */
1983 np
->cno
= cno
/* + 1 */ ;
1985 np
->cno
= invalidcharno
;
1986 np
->left
= np
->right
= NULL
;
1987 if (CTAGS
&& !cxref_style
)
1989 if (strlen (linestart
) < 50)
1990 np
->regex
= concat (linestart
, "$", "");
1992 np
->regex
= savenstr (linestart
, 50);
1995 np
->regex
= savenstr (linestart
, linelen
);
1997 add_node (np
, &nodehead
);
2002 * recurse on left children, iterate on right children.
2005 free_tree (register node
*np
)
2009 register node
*node_right
= np
->right
;
2010 free_tree (np
->left
);
2020 * delete a file description
2023 free_fdesc (register fdesc
*fdp
)
2025 free (fdp
->infname
);
2026 free (fdp
->infabsname
);
2027 free (fdp
->infabsdir
);
2028 free (fdp
->taggedfname
);
2035 * Adds a node to the tree of nodes. In etags mode, sort by file
2036 * name. In ctags mode, sort by tag name. Make no attempt at
2039 * add_node is the only function allowed to add nodes, so it can
2043 add_node (node
*np
, node
**cur_node_p
)
2046 register node
*cur_node
= *cur_node_p
;
2048 if (cur_node
== NULL
)
2058 /* For each file name, tags are in a linked sublist on the right
2059 pointer. The first tags of different files are a linked list
2060 on the left pointer. last_node points to the end of the last
2062 if (last_node
!= NULL
&& last_node
->fdp
== np
->fdp
)
2064 /* Let's use the same sublist as the last added node. */
2065 assert (last_node
->right
== NULL
);
2066 last_node
->right
= np
;
2069 else if (cur_node
->fdp
== np
->fdp
)
2071 /* Scanning the list we found the head of a sublist which is
2072 good for us. Let's scan this sublist. */
2073 add_node (np
, &cur_node
->right
);
2076 /* The head of this sublist is not good for us. Let's try the
2078 add_node (np
, &cur_node
->left
);
2079 } /* if ETAGS mode */
2084 dif
= strcmp (np
->name
, cur_node
->name
);
2087 * If this tag name matches an existing one, then
2088 * do not add the node, but maybe print a warning.
2090 if (no_duplicates
&& !dif
)
2092 if (np
->fdp
== cur_node
->fdp
)
2096 fprintf (stderr
, "Duplicate entry in file %s, line %d: %s\n",
2097 np
->fdp
->infname
, lineno
, np
->name
);
2098 fprintf (stderr
, "Second entry ignored\n");
2101 else if (!cur_node
->been_warned
&& !no_warnings
)
2105 "Duplicate entry in files %s and %s: %s (Warning only)\n",
2106 np
->fdp
->infname
, cur_node
->fdp
->infname
, np
->name
);
2107 cur_node
->been_warned
= true;
2112 /* Actually add the node */
2113 add_node (np
, dif
< 0 ? &cur_node
->left
: &cur_node
->right
);
2114 } /* if CTAGS mode */
2118 * invalidate_nodes ()
2119 * Scan the node tree and invalidate all nodes pointing to the
2120 * given file description (CTAGS case) or free them (ETAGS case).
2123 invalidate_nodes (fdesc
*badfdp
, node
**npp
)
2132 if (np
->left
!= NULL
)
2133 invalidate_nodes (badfdp
, &np
->left
);
2134 if (np
->fdp
== badfdp
)
2136 if (np
->right
!= NULL
)
2137 invalidate_nodes (badfdp
, &np
->right
);
2141 assert (np
->fdp
!= NULL
);
2142 if (np
->fdp
== badfdp
)
2144 *npp
= np
->left
; /* detach the sublist from the list */
2145 np
->left
= NULL
; /* isolate it */
2146 free_tree (np
); /* free it */
2147 invalidate_nodes (badfdp
, npp
);
2150 invalidate_nodes (badfdp
, &np
->left
);
2155 static int total_size_of_entries (node
*);
2156 static int number_len (long) ATTRIBUTE_CONST
;
2158 /* Length of a non-negative number's decimal representation. */
2160 number_len (long int num
)
2163 while ((num
/= 10) > 0)
2169 * Return total number of characters that put_entries will output for
2170 * the nodes in the linked list at the right of the specified node.
2171 * This count is irrelevant with etags.el since emacs 19.34 at least,
2172 * but is still supplied for backward compatibility.
2175 total_size_of_entries (register node
*np
)
2177 register int total
= 0;
2179 for (; np
!= NULL
; np
= np
->right
)
2182 total
+= strlen (np
->regex
) + 1; /* pat\177 */
2183 if (np
->name
!= NULL
)
2184 total
+= strlen (np
->name
) + 1; /* name\001 */
2185 total
+= number_len ((long) np
->lno
) + 1; /* lno, */
2186 if (np
->cno
!= invalidcharno
) /* cno */
2187 total
+= number_len (np
->cno
);
2188 total
+= 1; /* newline */
2195 put_entries (register node
*np
)
2198 static fdesc
*fdp
= NULL
;
2203 /* Output subentries that precede this one */
2205 put_entries (np
->left
);
2207 /* Output this entry */
2216 fprintf (tagf
, "\f\n%s,%d\n",
2217 fdp
->taggedfname
, total_size_of_entries (np
));
2218 fdp
->written
= true;
2220 fputs (np
->regex
, tagf
);
2221 fputc ('\177', tagf
);
2222 if (np
->name
!= NULL
)
2224 fputs (np
->name
, tagf
);
2225 fputc ('\001', tagf
);
2227 fprintf (tagf
, "%d,", np
->lno
);
2228 if (np
->cno
!= invalidcharno
)
2229 fprintf (tagf
, "%ld", np
->cno
);
2235 if (np
->name
== NULL
)
2236 error ("internal error: NULL name in ctags mode.");
2241 fprintf (stdout
, "%s %s %d\n",
2242 np
->name
, np
->fdp
->taggedfname
, (np
->lno
+ 63) / 64);
2244 fprintf (stdout
, "%-16s %3d %-16s %s\n",
2245 np
->name
, np
->lno
, np
->fdp
->taggedfname
, np
->regex
);
2249 fprintf (tagf
, "%s\t%s\t", np
->name
, np
->fdp
->taggedfname
);
2252 { /* function or #define macro with args */
2253 putc (searchar
, tagf
);
2256 for (sp
= np
->regex
; *sp
; sp
++)
2258 if (*sp
== '\\' || *sp
== searchar
)
2262 putc (searchar
, tagf
);
2265 { /* anything else; text pattern inadequate */
2266 fprintf (tagf
, "%d", np
->lno
);
2271 } /* if this node contains a valid tag */
2273 /* Output subentries that follow this one */
2274 put_entries (np
->right
);
2276 put_entries (np
->left
);
2281 #define C_EXT 0x00fff /* C extensions */
2282 #define C_PLAIN 0x00000 /* C */
2283 #define C_PLPL 0x00001 /* C++ */
2284 #define C_STAR 0x00003 /* C* */
2285 #define C_JAVA 0x00005 /* JAVA */
2286 #define C_AUTO 0x01000 /* C, but switch to C++ if `class' is met */
2287 #define YACC 0x10000 /* yacc file */
2290 * The C symbol tables.
2295 st_C_objprot
, st_C_objimpl
, st_C_objend
,
2297 st_C_ignore
, st_C_attribute
,
2300 st_C_class
, st_C_template
,
2301 st_C_struct
, st_C_extern
, st_C_enum
, st_C_define
, st_C_typedef
2304 /* Feed stuff between (but not including) %[ and %] lines to:
2310 struct C_stab_entry { char *name; int c_ext; enum sym_type type; }
2314 while, 0, st_C_ignore
2315 switch, 0, st_C_ignore
2316 return, 0, st_C_ignore
2317 __attribute__, 0, st_C_attribute
2318 GTY, 0, st_C_attribute
2319 @interface, 0, st_C_objprot
2320 @protocol, 0, st_C_objprot
2321 @implementation,0, st_C_objimpl
2322 @end, 0, st_C_objend
2323 import, (C_JAVA & ~C_PLPL), st_C_ignore
2324 package, (C_JAVA & ~C_PLPL), st_C_ignore
2325 friend, C_PLPL, st_C_ignore
2326 extends, (C_JAVA & ~C_PLPL), st_C_javastruct
2327 implements, (C_JAVA & ~C_PLPL), st_C_javastruct
2328 interface, (C_JAVA & ~C_PLPL), st_C_struct
2329 class, 0, st_C_class
2330 namespace, C_PLPL, st_C_struct
2331 domain, C_STAR, st_C_struct
2332 union, 0, st_C_struct
2333 struct, 0, st_C_struct
2334 extern, 0, st_C_extern
2336 typedef, 0, st_C_typedef
2337 define, 0, st_C_define
2338 undef, 0, st_C_define
2339 operator, C_PLPL, st_C_operator
2340 template, 0, st_C_template
2341 # DEFUN used in emacs, the next three used in glibc (SYSCALL only for mach).
2342 DEFUN, 0, st_C_gnumacro
2343 SYSCALL, 0, st_C_gnumacro
2344 ENTRY, 0, st_C_gnumacro
2345 PSEUDO, 0, st_C_gnumacro
2346 # These are defined inside C functions, so currently they are not met.
2347 # EXFUN used in glibc, DEFVAR_* in emacs.
2348 #EXFUN, 0, st_C_gnumacro
2349 #DEFVAR_, 0, st_C_gnumacro
2351 and replace lines between %< and %> with its output, then:
2352 - remove the #if characterset check
2353 - make in_word_set static and not inline. */
2355 /* C code produced by gperf version 3.0.1 */
2356 /* Command-line: gperf -m 5 */
2357 /* Computed positions: -k'2-3' */
2359 struct C_stab_entry
{ const char *name
; int c_ext
; enum sym_type type
; };
2360 /* maximum key range = 33, duplicates = 0 */
2363 hash (const char *str
, int len
)
2365 static char const asso_values
[] =
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, 35,
2373 35, 35, 35, 35, 35, 35, 35, 35, 35, 3,
2374 26, 35, 35, 35, 35, 35, 35, 35, 27, 35,
2375 35, 35, 35, 24, 0, 35, 35, 35, 35, 0,
2376 35, 35, 35, 35, 35, 1, 35, 16, 35, 6,
2377 23, 0, 0, 35, 22, 0, 35, 35, 5, 0,
2378 0, 15, 1, 35, 6, 35, 8, 19, 35, 16,
2379 4, 5, 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, 35, 35, 35, 35,
2392 35, 35, 35, 35, 35, 35
2399 hval
+= asso_values
[(unsigned char) str
[2]];
2402 hval
+= asso_values
[(unsigned char) str
[1]];
2408 static struct C_stab_entry
*
2409 in_word_set (register const char *str
, register unsigned int len
)
2413 TOTAL_KEYWORDS
= 33,
2414 MIN_WORD_LENGTH
= 2,
2415 MAX_WORD_LENGTH
= 15,
2420 static struct C_stab_entry wordlist
[] =
2423 {"if", 0, st_C_ignore
},
2424 {"GTY", 0, st_C_attribute
},
2425 {"@end", 0, st_C_objend
},
2426 {"union", 0, st_C_struct
},
2427 {"define", 0, st_C_define
},
2428 {"import", (C_JAVA
& ~C_PLPL
), st_C_ignore
},
2429 {"template", 0, st_C_template
},
2430 {"operator", C_PLPL
, st_C_operator
},
2431 {"@interface", 0, st_C_objprot
},
2432 {"implements", (C_JAVA
& ~C_PLPL
), st_C_javastruct
},
2433 {"friend", C_PLPL
, st_C_ignore
},
2434 {"typedef", 0, st_C_typedef
},
2435 {"return", 0, st_C_ignore
},
2436 {"@implementation",0, st_C_objimpl
},
2437 {"@protocol", 0, st_C_objprot
},
2438 {"interface", (C_JAVA
& ~C_PLPL
), st_C_struct
},
2439 {"extern", 0, st_C_extern
},
2440 {"extends", (C_JAVA
& ~C_PLPL
), st_C_javastruct
},
2441 {"struct", 0, st_C_struct
},
2442 {"domain", C_STAR
, st_C_struct
},
2443 {"switch", 0, st_C_ignore
},
2444 {"enum", 0, st_C_enum
},
2445 {"for", 0, st_C_ignore
},
2446 {"namespace", C_PLPL
, st_C_struct
},
2447 {"class", 0, st_C_class
},
2448 {"while", 0, st_C_ignore
},
2449 {"undef", 0, st_C_define
},
2450 {"package", (C_JAVA
& ~C_PLPL
), st_C_ignore
},
2451 {"__attribute__", 0, st_C_attribute
},
2452 {"SYSCALL", 0, st_C_gnumacro
},
2453 {"ENTRY", 0, st_C_gnumacro
},
2454 {"PSEUDO", 0, st_C_gnumacro
},
2455 {"DEFUN", 0, st_C_gnumacro
}
2458 if (len
<= MAX_WORD_LENGTH
&& len
>= MIN_WORD_LENGTH
)
2460 int key
= hash (str
, len
);
2462 if (key
<= MAX_HASH_VALUE
&& key
>= 0)
2464 const char *s
= wordlist
[key
].name
;
2466 if (*str
== *s
&& !strncmp (str
+ 1, s
+ 1, len
- 1) && s
[len
] == '\0')
2467 return &wordlist
[key
];
2474 static enum sym_type
2475 C_symtype (char *str
, int len
, int c_ext
)
2477 register struct C_stab_entry
*se
= in_word_set (str
, len
);
2479 if (se
== NULL
|| (se
->c_ext
&& !(c_ext
& se
->c_ext
)))
2486 * Ignoring __attribute__ ((list))
2488 static bool inattribute
; /* looking at an __attribute__ construct */
2491 * C functions and variables are recognized using a simple
2492 * finite automaton. fvdef is its state variable.
2496 fvnone
, /* nothing seen */
2497 fdefunkey
, /* Emacs DEFUN keyword seen */
2498 fdefunname
, /* Emacs DEFUN name seen */
2499 foperator
, /* func: operator keyword seen (cplpl) */
2500 fvnameseen
, /* function or variable name seen */
2501 fstartlist
, /* func: just after open parenthesis */
2502 finlist
, /* func: in parameter list */
2503 flistseen
, /* func: after parameter list */
2504 fignore
, /* func: before open brace */
2505 vignore
/* var-like: ignore until ';' */
2508 static bool fvextern
; /* func or var: extern keyword seen; */
2511 * typedefs are recognized using a simple finite automaton.
2512 * typdef is its state variable.
2516 tnone
, /* nothing seen */
2517 tkeyseen
, /* typedef keyword seen */
2518 ttypeseen
, /* defined type seen */
2519 tinbody
, /* inside typedef body */
2520 tend
, /* just before typedef tag */
2521 tignore
/* junk after typedef tag */
2525 * struct-like structures (enum, struct and union) are recognized
2526 * using another simple finite automaton. `structdef' is its state
2531 snone
, /* nothing seen yet,
2532 or in struct body if bracelev > 0 */
2533 skeyseen
, /* struct-like keyword seen */
2534 stagseen
, /* struct-like tag seen */
2535 scolonseen
/* colon seen after struct-like tag */
2539 * When objdef is different from onone, objtag is the name of the class.
2541 static const char *objtag
= "<uninited>";
2544 * Yet another little state machine to deal with preprocessor lines.
2548 dnone
, /* nothing seen */
2549 dsharpseen
, /* '#' seen as first char on line */
2550 ddefineseen
, /* '#' and 'define' seen */
2551 dignorerest
/* ignore rest of line */
2555 * State machine for Objective C protocols and implementations.
2556 * Idea by Tom R.Hageman <tom@basil.icce.rug.nl> (1995)
2560 onone
, /* nothing seen */
2561 oprotocol
, /* @interface or @protocol seen */
2562 oimplementation
, /* @implementations seen */
2563 otagseen
, /* class name seen */
2564 oparenseen
, /* parenthesis before category seen */
2565 ocatseen
, /* category name seen */
2566 oinbody
, /* in @implementation body */
2567 omethodsign
, /* in @implementation body, after +/- */
2568 omethodtag
, /* after method name */
2569 omethodcolon
, /* after method colon */
2570 omethodparm
, /* after method parameter */
2571 oignore
/* wait for @end */
2576 * Use this structure to keep info about the token read, and how it
2577 * should be tagged. Used by the make_C_tag function to build a tag.
2581 char *line
; /* string containing the token */
2582 int offset
; /* where the token starts in LINE */
2583 int length
; /* token length */
2585 The previous members can be used to pass strings around for generic
2586 purposes. The following ones specifically refer to creating tags. In this
2587 case the token contained here is the pattern that will be used to create a
2590 bool valid
; /* do not create a tag; the token should be
2591 invalidated whenever a state machine is
2592 reset prematurely */
2593 bool named
; /* create a named tag */
2594 int lineno
; /* source line number of tag */
2595 long linepos
; /* source char number of tag */
2596 } token
; /* latest token read */
2599 * Variables and functions for dealing with nested structures.
2600 * Idea by Mykola Dzyuba <mdzyuba@yahoo.com> (2001)
2602 static void pushclass_above (int, char *, int);
2603 static void popclass_above (int);
2604 static void write_classname (linebuffer
*, const char *qualifier
);
2607 char **cname
; /* nested class names */
2608 int *bracelev
; /* nested class brace level */
2609 int nl
; /* class nesting level (elements used) */
2610 int size
; /* length of the array */
2611 } cstack
; /* stack for nested declaration tags */
2612 /* Current struct nesting depth (namespace, class, struct, union, enum). */
2613 #define nestlev (cstack.nl)
2614 /* After struct keyword or in struct body, not inside a nested function. */
2615 #define instruct (structdef == snone && nestlev > 0 \
2616 && bracelev == cstack.bracelev[nestlev-1] + 1)
2619 pushclass_above (int bracelev
, char *str
, int len
)
2623 popclass_above (bracelev
);
2625 if (nl
>= cstack
.size
)
2627 int size
= cstack
.size
*= 2;
2628 xrnew (cstack
.cname
, size
, char *);
2629 xrnew (cstack
.bracelev
, size
, int);
2631 assert (nl
== 0 || cstack
.bracelev
[nl
-1] < bracelev
);
2632 cstack
.cname
[nl
] = (str
== NULL
) ? NULL
: savenstr (str
, len
);
2633 cstack
.bracelev
[nl
] = bracelev
;
2638 popclass_above (int bracelev
)
2642 for (nl
= cstack
.nl
- 1;
2643 nl
>= 0 && cstack
.bracelev
[nl
] >= bracelev
;
2646 free (cstack
.cname
[nl
]);
2652 write_classname (linebuffer
*cn
, const char *qualifier
)
2655 int qlen
= strlen (qualifier
);
2657 if (cstack
.nl
== 0 || cstack
.cname
[0] == NULL
)
2661 cn
->buffer
[0] = '\0';
2665 len
= strlen (cstack
.cname
[0]);
2666 linebuffer_setlen (cn
, len
);
2667 strcpy (cn
->buffer
, cstack
.cname
[0]);
2669 for (i
= 1; i
< cstack
.nl
; i
++)
2671 char *s
= cstack
.cname
[i
];
2674 linebuffer_setlen (cn
, len
+ qlen
+ strlen (s
));
2675 len
+= sprintf (cn
->buffer
+ len
, "%s%s", qualifier
, s
);
2680 static bool consider_token (char *, int, int, int *, int, int, bool *);
2681 static void make_C_tag (bool);
2685 * checks to see if the current token is at the start of a
2686 * function or variable, or corresponds to a typedef, or
2687 * is a struct/union/enum tag, or #define, or an enum constant.
2689 * *IS_FUNC_OR_VAR gets true if the token is a function or #define macro
2690 * with args. C_EXTP points to which language we are looking at.
2701 consider_token (char *str
, int len
, int c
, int *c_extp
,
2702 int bracelev
, int parlev
, bool *is_func_or_var
)
2703 /* IN: token pointer */
2704 /* IN: token length */
2705 /* IN: first char after the token */
2706 /* IN, OUT: C extensions mask */
2707 /* IN: brace level */
2708 /* IN: parenthesis level */
2709 /* OUT: function or variable found */
2711 /* When structdef is stagseen, scolonseen, or snone with bracelev > 0,
2712 structtype is the type of the preceding struct-like keyword, and
2713 structbracelev is the brace level where it has been seen. */
2714 static enum sym_type structtype
;
2715 static int structbracelev
;
2716 static enum sym_type toktype
;
2719 toktype
= C_symtype (str
, len
, *c_extp
);
2722 * Skip __attribute__
2724 if (toktype
== st_C_attribute
)
2731 * Advance the definedef state machine.
2736 /* We're not on a preprocessor line. */
2737 if (toktype
== st_C_gnumacro
)
2744 if (toktype
== st_C_define
)
2746 definedef
= ddefineseen
;
2750 definedef
= dignorerest
;
2755 * Make a tag for any macro, unless it is a constant
2756 * and constantypedefs is false.
2758 definedef
= dignorerest
;
2759 *is_func_or_var
= (c
== '(');
2760 if (!*is_func_or_var
&& !constantypedefs
)
2767 error ("internal error: definedef value.");
2776 if (toktype
== st_C_typedef
)
2799 if (structdef
== snone
&& fvdef
== fvnone
)
2821 case st_C_javastruct
:
2822 if (structdef
== stagseen
)
2823 structdef
= scolonseen
;
2827 if ((*c_extp
& C_AUTO
) /* automatic detection of C++ language */
2829 && definedef
== dnone
&& structdef
== snone
2830 && typdef
== tnone
&& fvdef
== fvnone
)
2831 *c_extp
= (*c_extp
| C_PLPL
) & ~C_AUTO
;
2832 if (toktype
== st_C_template
)
2839 && (typdef
== tkeyseen
2840 || (typedefs_or_cplusplus
&& structdef
== snone
)))
2842 structdef
= skeyseen
;
2843 structtype
= toktype
;
2844 structbracelev
= bracelev
;
2845 if (fvdef
== fvnameseen
)
2853 if (structdef
== skeyseen
)
2855 structdef
= stagseen
;
2859 if (typdef
!= tnone
)
2862 /* Detect Objective C constructs. */
2872 objdef
= oimplementation
;
2878 case oimplementation
:
2879 /* Save the class tag for functions or variables defined inside. */
2880 objtag
= savenstr (str
, len
);
2884 /* Save the class tag for categories. */
2885 objtag
= savenstr (str
, len
);
2887 *is_func_or_var
= true;
2891 *is_func_or_var
= true;
2899 objdef
= omethodtag
;
2900 linebuffer_setlen (&token_name
, len
);
2901 memcpy (token_name
.buffer
, str
, len
);
2902 token_name
.buffer
[len
] = '\0';
2908 objdef
= omethodparm
;
2913 objdef
= omethodtag
;
2916 int oldlen
= token_name
.len
;
2918 linebuffer_setlen (&token_name
, oldlen
+ len
);
2919 memcpy (token_name
.buffer
+ oldlen
, str
, len
);
2920 token_name
.buffer
[oldlen
+ len
] = '\0';
2926 if (toktype
== st_C_objend
)
2928 /* Memory leakage here: the string pointed by objtag is
2929 never released, because many tests would be needed to
2930 avoid breaking on incorrect input code. The amount of
2931 memory leaked here is the sum of the lengths of the
2941 /* A function, variable or enum constant? */
2963 *is_func_or_var
= true;
2967 && structdef
== snone
2968 && structtype
== st_C_enum
&& bracelev
> structbracelev
2969 /* Don't tag tokens in expressions that assign values to enum
2971 && fvdef
!= vignore
)
2972 return true; /* enum constant */
2978 fvdef
= fdefunname
; /* GNU macro */
2979 *is_func_or_var
= true;
2987 if ((strneq (str
, "asm", 3) && endtoken (str
[3]))
2988 || (strneq (str
, "__asm__", 7) && endtoken (str
[7])))
2999 if (len
>= 10 && strneq (str
+len
-10, "::operator", 10))
3001 if (*c_extp
& C_AUTO
) /* automatic detection of C++ */
3002 *c_extp
= (*c_extp
| C_PLPL
) & ~C_AUTO
;
3004 *is_func_or_var
= true;
3007 if (bracelev
> 0 && !instruct
)
3009 fvdef
= fvnameseen
; /* function or variable */
3010 *is_func_or_var
= true;
3025 * C_entries often keeps pointers to tokens or lines which are older than
3026 * the line currently read. By keeping two line buffers, and switching
3027 * them at end of line, it is possible to use those pointers.
3035 #define current_lb_is_new (newndx == curndx)
3036 #define switch_line_buffers() (curndx = 1 - curndx)
3038 #define curlb (lbs[curndx].lb)
3039 #define newlb (lbs[newndx].lb)
3040 #define curlinepos (lbs[curndx].linepos)
3041 #define newlinepos (lbs[newndx].linepos)
3043 #define plainc ((c_ext & C_EXT) == C_PLAIN)
3044 #define cplpl (c_ext & C_PLPL)
3045 #define cjava ((c_ext & C_JAVA) == C_JAVA)
3047 #define CNL_SAVE_DEFINEDEF() \
3049 curlinepos = charno; \
3050 readline (&curlb, inf); \
3051 lp = curlb.buffer; \
3058 CNL_SAVE_DEFINEDEF (); \
3059 if (savetoken.valid) \
3061 token = savetoken; \
3062 savetoken.valid = false; \
3064 definedef = dnone; \
3069 make_C_tag (bool isfun
)
3071 /* This function is never called when token.valid is false, but
3072 we must protect against invalid input or internal errors. */
3074 make_tag (token_name
.buffer
, token_name
.len
, isfun
, token
.line
,
3075 token
.offset
+token
.length
+1, token
.lineno
, token
.linepos
);
3077 { /* this branch is optimized away if !DEBUG */
3078 make_tag (concat ("INVALID TOKEN:-->", token_name
.buffer
, ""),
3079 token_name
.len
+ 17, isfun
, token
.line
,
3080 token
.offset
+token
.length
+1, token
.lineno
, token
.linepos
);
3081 error ("INVALID TOKEN");
3084 token
.valid
= false;
3088 perhaps_more_input (FILE *inf
)
3090 return !feof (inf
) && !ferror (inf
);
3096 * This routine finds functions, variables, typedefs,
3097 * #define's, enum constants and struct/union/enum definitions in
3098 * C syntax and adds them to the list.
3101 C_entries (int c_ext
, FILE *inf
)
3102 /* extension of C */
3105 register char c
; /* latest char read; '\0' for end of line */
3106 register char *lp
; /* pointer one beyond the character `c' */
3107 int curndx
, newndx
; /* indices for current and new lb */
3108 register int tokoff
; /* offset in line of start of current token */
3109 register int toklen
; /* length of current token */
3110 const char *qualifier
; /* string used to qualify names */
3111 int qlen
; /* length of qualifier */
3112 int bracelev
; /* current brace level */
3113 int bracketlev
; /* current bracket level */
3114 int parlev
; /* current parenthesis level */
3115 int attrparlev
; /* __attribute__ parenthesis level */
3116 int templatelev
; /* current template level */
3117 int typdefbracelev
; /* bracelev where a typedef struct body begun */
3118 bool incomm
, inquote
, inchar
, quotednl
, midtoken
;
3119 bool yacc_rules
; /* in the rules part of a yacc file */
3120 struct tok savetoken
= {0}; /* token saved during preprocessor handling */
3123 linebuffer_init (&lbs
[0].lb
);
3124 linebuffer_init (&lbs
[1].lb
);
3125 if (cstack
.size
== 0)
3127 cstack
.size
= (DEBUG
) ? 1 : 4;
3129 cstack
.cname
= xnew (cstack
.size
, char *);
3130 cstack
.bracelev
= xnew (cstack
.size
, int);
3133 tokoff
= toklen
= typdefbracelev
= 0; /* keep compiler quiet */
3134 curndx
= newndx
= 0;
3138 fvdef
= fvnone
; fvextern
= false; typdef
= tnone
;
3139 structdef
= snone
; definedef
= dnone
; objdef
= onone
;
3141 midtoken
= inquote
= inchar
= incomm
= quotednl
= false;
3142 token
.valid
= savetoken
.valid
= false;
3143 bracelev
= bracketlev
= parlev
= attrparlev
= templatelev
= 0;
3145 { qualifier
= "."; qlen
= 1; }
3147 { qualifier
= "::"; qlen
= 2; }
3150 while (perhaps_more_input (inf
))
3155 /* If we are at the end of the line, the next character is a
3156 '\0'; do not skip it, because it is what tells us
3157 to read the next line. */
3178 /* Newlines inside comments do not end macro definitions in
3180 CNL_SAVE_DEFINEDEF ();
3193 /* Newlines inside strings do not end macro definitions
3194 in traditional cpp, even though compilers don't
3195 usually accept them. */
3196 CNL_SAVE_DEFINEDEF ();
3206 /* Hmmm, something went wrong. */
3242 if (fvdef
!= finlist
&& fvdef
!= fignore
&& fvdef
!= vignore
)
3257 else if (/* cplpl && */ *lp
== '/')
3263 if ((c_ext
& YACC
) && *lp
== '%')
3265 /* Entering or exiting rules section in yacc file. */
3267 definedef
= dnone
; fvdef
= fvnone
; fvextern
= false;
3268 typdef
= tnone
; structdef
= snone
;
3269 midtoken
= inquote
= inchar
= incomm
= quotednl
= false;
3271 yacc_rules
= !yacc_rules
;
3277 if (definedef
== dnone
)
3280 bool cpptoken
= true;
3282 /* Look back on this line. If all blanks, or nonblanks
3283 followed by an end of comment, this is a preprocessor
3285 for (cp
= newlb
.buffer
; cp
< lp
-1; cp
++)
3286 if (!c_isspace (*cp
))
3288 if (*cp
== '*' && cp
[1] == '/')
3298 definedef
= dsharpseen
;
3299 /* This is needed for tagging enum values: when there are
3300 preprocessor conditionals inside the enum, we need to
3301 reset the value of fvdef so that the next enum value is
3302 tagged even though the one before it did not end in a
3304 if (fvdef
== vignore
&& instruct
&& parlev
== 0)
3306 if (strneq (cp
, "#if", 3) || strneq (cp
, "#el", 3))
3310 } /* if (definedef == dnone) */
3321 CNL_SAVE_DEFINEDEF ();
3328 /* Consider token only if some involved conditions are satisfied. */
3329 if (typdef
!= tignore
3330 && definedef
!= dignorerest
3333 && (definedef
!= dnone
3334 || structdef
!= scolonseen
)
3341 if (c
== ':' && *lp
== ':' && begtoken (lp
[1]))
3342 /* This handles :: in the middle,
3343 but not at the beginning of an identifier.
3344 Also, space-separated :: is not recognized. */
3346 if (c_ext
& C_AUTO
) /* automatic detection of C++ */
3347 c_ext
= (c_ext
| C_PLPL
) & ~C_AUTO
;
3351 goto still_in_token
;
3355 bool funorvar
= false;
3358 || consider_token (newlb
.buffer
+ tokoff
, toklen
, c
,
3359 &c_ext
, bracelev
, parlev
,
3362 if (fvdef
== foperator
)
3365 lp
= skip_spaces (lp
-1);
3369 && !c_isspace (*lp
) && *lp
!= '(')
3372 toklen
+= lp
- oldlp
;
3374 token
.named
= false;
3376 && nestlev
> 0 && definedef
== dnone
)
3377 /* in struct body */
3382 write_classname (&token_name
, qualifier
);
3383 len
= token_name
.len
;
3384 linebuffer_setlen (&token_name
,
3385 len
+ qlen
+ toklen
);
3386 sprintf (token_name
.buffer
+ len
, "%s%.*s",
3388 newlb
.buffer
+ tokoff
);
3392 linebuffer_setlen (&token_name
, toklen
);
3393 sprintf (token_name
.buffer
, "%.*s",
3394 toklen
, newlb
.buffer
+ tokoff
);
3398 else if (objdef
== ocatseen
)
3399 /* Objective C category */
3403 int len
= strlen (objtag
) + 2 + toklen
;
3404 linebuffer_setlen (&token_name
, len
);
3405 sprintf (token_name
.buffer
, "%s(%.*s)",
3407 newlb
.buffer
+ tokoff
);
3411 linebuffer_setlen (&token_name
, toklen
);
3412 sprintf (token_name
.buffer
, "%.*s",
3413 toklen
, newlb
.buffer
+ tokoff
);
3417 else if (objdef
== omethodtag
3418 || objdef
== omethodparm
)
3419 /* Objective C method */
3423 else if (fvdef
== fdefunname
)
3424 /* GNU DEFUN and similar macros */
3426 bool defun
= (newlb
.buffer
[tokoff
] == 'F');
3430 /* Rewrite the tag so that emacs lisp DEFUNs
3431 can be found by their elisp name */
3437 linebuffer_setlen (&token_name
, len
);
3438 memcpy (token_name
.buffer
,
3439 newlb
.buffer
+ off
, len
);
3440 token_name
.buffer
[len
] = '\0';
3443 if (token_name
.buffer
[len
] == '_')
3444 token_name
.buffer
[len
] = '-';
3445 token
.named
= defun
;
3449 linebuffer_setlen (&token_name
, toklen
);
3450 memcpy (token_name
.buffer
,
3451 newlb
.buffer
+ tokoff
, toklen
);
3452 token_name
.buffer
[toklen
] = '\0';
3453 /* Name macros and members. */
3454 token
.named
= (structdef
== stagseen
3455 || typdef
== ttypeseen
3458 && definedef
== dignorerest
)
3460 && definedef
== dnone
3461 && structdef
== snone
3464 token
.lineno
= lineno
;
3465 token
.offset
= tokoff
;
3466 token
.length
= toklen
;
3467 token
.line
= newlb
.buffer
;
3468 token
.linepos
= newlinepos
;
3471 if (definedef
== dnone
3472 && (fvdef
== fvnameseen
3473 || fvdef
== foperator
3474 || structdef
== stagseen
3476 || typdef
== ttypeseen
3477 || objdef
!= onone
))
3479 if (current_lb_is_new
)
3480 switch_line_buffers ();
3482 else if (definedef
!= dnone
3483 || fvdef
== fdefunname
3485 make_C_tag (funorvar
);
3487 else /* not yacc and consider_token failed */
3489 if (inattribute
&& fvdef
== fignore
)
3491 /* We have just met __attribute__ after a
3492 function parameter list: do not tag the
3499 } /* if (endtoken (c)) */
3500 else if (intoken (c
))
3506 } /* if (midtoken) */
3507 else if (begtoken (c
))
3515 /* This prevents tagging fb in
3516 void (__attribute__((noreturn)) *fb) (void);
3517 Fixing this is not easy and not very important. */
3521 if (plainc
|| declarations
)
3523 make_C_tag (true); /* a function */
3530 if (structdef
== stagseen
&& !cjava
)
3532 popclass_above (bracelev
);
3542 if (!yacc_rules
|| lp
== newlb
.buffer
+ 1)
3544 tokoff
= lp
- 1 - newlb
.buffer
;
3549 } /* if (begtoken) */
3550 } /* if must look at token */
3553 /* Detect end of line, colon, comma, semicolon and various braces
3554 after having handled a token.*/
3560 if (yacc_rules
&& token
.offset
== 0 && token
.valid
)
3562 make_C_tag (false); /* a yacc function */
3565 if (definedef
!= dnone
)
3571 make_C_tag (true); /* an Objective C class */
3575 objdef
= omethodcolon
;
3578 int toklen
= token_name
.len
;
3579 linebuffer_setlen (&token_name
, toklen
+ 1);
3580 strcpy (token_name
.buffer
+ toklen
, ":");
3586 if (structdef
== stagseen
)
3588 structdef
= scolonseen
;
3591 /* Should be useless, but may be work as a safety net. */
3592 if (cplpl
&& fvdef
== flistseen
)
3594 make_C_tag (true); /* a function */
3600 if (definedef
!= dnone
|| inattribute
)
3606 make_C_tag (false); /* a typedef */
3616 if (typdef
== tignore
|| cplpl
)
3620 if ((globals
&& bracelev
== 0 && (!fvextern
|| declarations
))
3621 || (members
&& instruct
))
3622 make_C_tag (false); /* a variable */
3625 token
.valid
= false;
3629 && (cplpl
|| !instruct
)
3630 && (typdef
== tnone
|| (typdef
!= tignore
&& instruct
)))
3632 && plainc
&& instruct
))
3633 make_C_tag (true); /* a function */
3639 && cplpl
&& structdef
== stagseen
)
3640 make_C_tag (false); /* forward declaration */
3642 token
.valid
= false;
3643 } /* switch (fvdef) */
3649 if (structdef
== stagseen
)
3653 if (definedef
!= dnone
|| inattribute
)
3659 make_C_tag (true); /* an Objective C method */
3674 if (instruct
&& parlev
== 0)
3685 && (!fvextern
|| declarations
))
3686 || (members
&& instruct
)))
3687 make_C_tag (false); /* a variable */
3690 if ((declarations
&& typdef
== tnone
&& !instruct
)
3691 || (members
&& typdef
!= tignore
&& instruct
))
3693 make_C_tag (true); /* a function */
3696 else if (!declarations
)
3698 token
.valid
= false;
3703 if (structdef
== stagseen
)
3707 if (definedef
!= dnone
|| inattribute
)
3709 if (structdef
== stagseen
)
3716 make_C_tag (false); /* a typedef */
3728 if ((members
&& bracelev
== 1)
3729 || (globals
&& bracelev
== 0
3730 && (!fvextern
|| declarations
)))
3731 make_C_tag (false); /* a variable */
3747 if (definedef
!= dnone
)
3749 if (objdef
== otagseen
&& parlev
== 0)
3750 objdef
= oparenseen
;
3754 if (typdef
== ttypeseen
3758 /* This handles constructs like:
3759 typedef void OperatorFun (int fun); */
3780 if (--attrparlev
== 0)
3781 inattribute
= false;
3784 if (definedef
!= dnone
)
3786 if (objdef
== ocatseen
&& parlev
== 1)
3788 make_C_tag (true); /* an Objective C category */
3804 || typdef
== ttypeseen
))
3807 make_C_tag (false); /* a typedef */
3810 else if (parlev
< 0) /* can happen due to ill-conceived #if's. */
3814 if (definedef
!= dnone
)
3816 if (typdef
== ttypeseen
)
3818 /* Whenever typdef is set to tinbody (currently only
3819 here), typdefbracelev should be set to bracelev. */
3821 typdefbracelev
= bracelev
;
3826 if (cplpl
&& !class_qualify
)
3828 /* Remove class and namespace qualifiers from the token,
3829 leaving only the method/member name. */
3830 char *cc
, *uqname
= token_name
.buffer
;
3831 char *tok_end
= token_name
.buffer
+ token_name
.len
;
3833 for (cc
= token_name
.buffer
; cc
< tok_end
; cc
++)
3835 if (*cc
== ':' && cc
[1] == ':')
3841 if (uqname
> token_name
.buffer
)
3843 int uqlen
= strlen (uqname
);
3844 linebuffer_setlen (&token_name
, uqlen
);
3845 memmove (token_name
.buffer
, uqname
, uqlen
+ 1);
3848 make_C_tag (true); /* a function */
3857 make_C_tag (true); /* an Objective C class */
3862 make_C_tag (true); /* an Objective C method */
3866 /* Neutralize `extern "C" {' grot. */
3867 if (bracelev
== 0 && structdef
== snone
&& nestlev
== 0
3877 case skeyseen
: /* unnamed struct */
3878 pushclass_above (bracelev
, NULL
, 0);
3881 case stagseen
: /* named struct or enum */
3882 case scolonseen
: /* a class */
3883 pushclass_above (bracelev
,token
.line
+token
.offset
, token
.length
);
3885 make_C_tag (false); /* a struct or enum */
3893 if (definedef
!= dnone
)
3895 if (fvdef
== fstartlist
)
3897 fvdef
= fvnone
; /* avoid tagging `foo' in `foo (*bar()) ()' */
3898 token
.valid
= false;
3902 if (definedef
!= dnone
)
3905 if (!ignoreindent
&& lp
== newlb
.buffer
+ 1)
3908 token
.valid
= false; /* unexpected value, token unreliable */
3909 bracelev
= 0; /* reset brace level if first column */
3910 parlev
= 0; /* also reset paren level, just in case... */
3912 else if (bracelev
< 0)
3914 token
.valid
= false; /* something gone amiss, token unreliable */
3917 if (bracelev
== 0 && fvdef
== vignore
)
3918 fvdef
= fvnone
; /* end of function */
3919 popclass_above (bracelev
);
3921 /* Only if typdef == tinbody is typdefbracelev significant. */
3922 if (typdef
== tinbody
&& bracelev
<= typdefbracelev
)
3924 assert (bracelev
== typdefbracelev
);
3929 if (definedef
!= dnone
)
3939 if ((members
&& bracelev
== 1)
3940 || (globals
&& bracelev
== 0 && (!fvextern
|| declarations
)))
3941 make_C_tag (false); /* a variable */
3949 && (structdef
== stagseen
|| fvdef
== fvnameseen
))
3956 if (templatelev
> 0)
3964 if (objdef
== oinbody
&& bracelev
== 0)
3966 objdef
= omethodsign
;
3971 case '#': case '~': case '&': case '%': case '/':
3972 case '|': case '^': case '!': case '.': case '?':
3973 if (definedef
!= dnone
)
3975 /* These surely cannot follow a function tag in C. */
3988 if (objdef
== otagseen
)
3990 make_C_tag (true); /* an Objective C class */
3993 /* If a macro spans multiple lines don't reset its state. */
3995 CNL_SAVE_DEFINEDEF ();
4001 } /* while not eof */
4003 free (lbs
[0].lb
.buffer
);
4004 free (lbs
[1].lb
.buffer
);
4008 * Process either a C++ file or a C file depending on the setting
4012 default_C_entries (FILE *inf
)
4014 C_entries (cplusplus
? C_PLPL
: C_AUTO
, inf
);
4017 /* Always do plain C. */
4019 plain_C_entries (FILE *inf
)
4024 /* Always do C++. */
4026 Cplusplus_entries (FILE *inf
)
4028 C_entries (C_PLPL
, inf
);
4031 /* Always do Java. */
4033 Cjava_entries (FILE *inf
)
4035 C_entries (C_JAVA
, inf
);
4040 Cstar_entries (FILE *inf
)
4042 C_entries (C_STAR
, inf
);
4045 /* Always do Yacc. */
4047 Yacc_entries (FILE *inf
)
4049 C_entries (YACC
, inf
);
4053 /* Useful macros. */
4054 #define LOOP_ON_INPUT_LINES(file_pointer, line_buffer, char_pointer) \
4055 while (perhaps_more_input (file_pointer) \
4056 && (readline (&(line_buffer), file_pointer), \
4057 (char_pointer) = (line_buffer).buffer, \
4060 #define LOOKING_AT(cp, kw) /* kw is the keyword, a literal string */ \
4061 ((assert ("" kw), true) /* syntax error if not a literal string */ \
4062 && strneq ((cp), kw, sizeof (kw)-1) /* cp points at kw */ \
4063 && notinname ((cp)[sizeof (kw)-1]) /* end of kw */ \
4064 && ((cp) = skip_spaces ((cp)+sizeof (kw)-1))) /* skip spaces */
4066 /* Similar to LOOKING_AT but does not use notinname, does not skip */
4067 #define LOOKING_AT_NOCASE(cp, kw) /* the keyword is a literal string */ \
4068 ((assert ("" kw), true) /* syntax error if not a literal string */ \
4069 && strncaseeq ((cp), kw, sizeof (kw)-1) /* cp points at kw */ \
4070 && ((cp) += sizeof (kw)-1)) /* skip spaces */
4073 * Read a file, but do no processing. This is used to do regexp
4074 * matching on files that have no language defined.
4077 just_read_file (FILE *inf
)
4079 while (perhaps_more_input (inf
))
4080 readline (&lb
, inf
);
4084 /* Fortran parsing */
4086 static void F_takeprec (void);
4087 static void F_getit (FILE *);
4092 dbp
= skip_spaces (dbp
);
4096 dbp
= skip_spaces (dbp
);
4097 if (strneq (dbp
, "(*)", 3))
4102 if (!c_isdigit (*dbp
))
4104 --dbp
; /* force failure */
4109 while (c_isdigit (*dbp
));
4117 dbp
= skip_spaces (dbp
);
4120 readline (&lb
, inf
);
4125 dbp
= skip_spaces (dbp
);
4127 if (!c_isalpha (*dbp
) && *dbp
!= '_' && *dbp
!= '$')
4129 for (cp
= dbp
+ 1; *cp
!= '\0' && intoken (*cp
); cp
++)
4131 make_tag (dbp
, cp
-dbp
, true,
4132 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4137 Fortran_functions (FILE *inf
)
4139 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
4142 dbp
++; /* Ratfor escape to fortran */
4143 dbp
= skip_spaces (dbp
);
4147 if (LOOKING_AT_NOCASE (dbp
, "recursive"))
4148 dbp
= skip_spaces (dbp
);
4150 if (LOOKING_AT_NOCASE (dbp
, "pure"))
4151 dbp
= skip_spaces (dbp
);
4153 if (LOOKING_AT_NOCASE (dbp
, "elemental"))
4154 dbp
= skip_spaces (dbp
);
4156 switch (c_tolower (*dbp
))
4159 if (nocase_tail ("integer"))
4163 if (nocase_tail ("real"))
4167 if (nocase_tail ("logical"))
4171 if (nocase_tail ("complex") || nocase_tail ("character"))
4175 if (nocase_tail ("double"))
4177 dbp
= skip_spaces (dbp
);
4180 if (nocase_tail ("precision"))
4186 dbp
= skip_spaces (dbp
);
4189 switch (c_tolower (*dbp
))
4192 if (nocase_tail ("function"))
4196 if (nocase_tail ("subroutine"))
4200 if (nocase_tail ("entry"))
4204 if (nocase_tail ("blockdata") || nocase_tail ("block data"))
4206 dbp
= skip_spaces (dbp
);
4207 if (*dbp
== '\0') /* assume un-named */
4208 make_tag ("blockdata", 9, true,
4209 lb
.buffer
, dbp
- lb
.buffer
, lineno
, linecharno
);
4211 F_getit (inf
); /* look for name */
4220 * Go language support
4221 * Original code by Xi Lu <lx@shellcodes.org> (2016)
4224 Go_functions(FILE *inf
)
4228 LOOP_ON_INPUT_LINES(inf
, lb
, cp
)
4230 cp
= skip_spaces (cp
);
4232 if (LOOKING_AT (cp
, "package"))
4235 while (!notinname (*cp
) && *cp
!= '\0')
4237 make_tag (name
, cp
- name
, false, lb
.buffer
,
4238 cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4240 else if (LOOKING_AT (cp
, "func"))
4242 /* Go implementation of interface, such as:
4243 func (n *Integer) Add(m Integer) ...
4244 skip `(n *Integer)` part.
4250 cp
= skip_spaces (cp
+1);
4257 while (!notinname (*cp
))
4260 make_tag (name
, cp
- name
, true, lb
.buffer
,
4261 cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4264 else if (members
&& LOOKING_AT (cp
, "type"))
4268 /* Ignore the likes of the following:
4276 while (!notinname (*cp
) && *cp
!= '\0')
4279 make_tag (name
, cp
- name
, false, lb
.buffer
,
4280 cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4289 * Philippe Waroquiers (1998)
4292 /* Once we are positioned after an "interesting" keyword, let's get
4293 the real tag value necessary. */
4295 Ada_getit (FILE *inf
, const char *name_qualifier
)
4301 while (perhaps_more_input (inf
))
4303 dbp
= skip_spaces (dbp
);
4305 || (dbp
[0] == '-' && dbp
[1] == '-'))
4307 readline (&lb
, inf
);
4310 switch (c_tolower (*dbp
))
4313 if (nocase_tail ("body"))
4315 /* Skipping body of procedure body or package body or ....
4316 resetting qualifier to body instead of spec. */
4317 name_qualifier
= "/b";
4322 /* Skipping type of task type or protected type ... */
4323 if (nocase_tail ("type"))
4330 for (cp
= dbp
; *cp
!= '\0' && *cp
!= '"'; cp
++)
4335 dbp
= skip_spaces (dbp
);
4337 c_isalnum (*cp
) || *cp
== '_' || *cp
== '.';
4345 name
= concat (dbp
, name_qualifier
, "");
4347 make_tag (name
, strlen (name
), true,
4348 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4357 Ada_funcs (FILE *inf
)
4359 bool inquote
= false;
4360 bool skip_till_semicolumn
= false;
4362 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
4364 while (*dbp
!= '\0')
4366 /* Skip a string i.e. "abcd". */
4367 if (inquote
|| (*dbp
== '"'))
4369 dbp
= strchr (dbp
+ !inquote
, '"');
4374 continue; /* advance char */
4379 break; /* advance line */
4383 /* Skip comments. */
4384 if (dbp
[0] == '-' && dbp
[1] == '-')
4385 break; /* advance line */
4387 /* Skip character enclosed in single quote i.e. 'a'
4388 and skip single quote starting an attribute i.e. 'Image. */
4397 if (skip_till_semicolumn
)
4400 skip_till_semicolumn
= false;
4402 continue; /* advance char */
4405 /* Search for beginning of a token. */
4406 if (!begtoken (*dbp
))
4409 continue; /* advance char */
4412 /* We are at the beginning of a token. */
4413 switch (c_tolower (*dbp
))
4416 if (!packages_only
&& nocase_tail ("function"))
4417 Ada_getit (inf
, "/f");
4419 break; /* from switch */
4420 continue; /* advance char */
4422 if (!packages_only
&& nocase_tail ("procedure"))
4423 Ada_getit (inf
, "/p");
4424 else if (nocase_tail ("package"))
4425 Ada_getit (inf
, "/s");
4426 else if (nocase_tail ("protected")) /* protected type */
4427 Ada_getit (inf
, "/t");
4429 break; /* from switch */
4430 continue; /* advance char */
4433 if (typedefs
&& !packages_only
&& nocase_tail ("use"))
4435 /* when tagging types, avoid tagging use type Pack.Typename;
4436 for this, we will skip everything till a ; */
4437 skip_till_semicolumn
= true;
4438 continue; /* advance char */
4442 if (!packages_only
&& nocase_tail ("task"))
4443 Ada_getit (inf
, "/k");
4444 else if (typedefs
&& !packages_only
&& nocase_tail ("type"))
4446 Ada_getit (inf
, "/t");
4447 while (*dbp
!= '\0')
4451 break; /* from switch */
4452 continue; /* advance char */
4455 /* Look for the end of the token. */
4456 while (!endtoken (*dbp
))
4459 } /* advance char */
4460 } /* advance line */
4465 * Unix and microcontroller assembly tag handling
4466 * Labels: /^[a-zA-Z_.$][a-zA_Z0-9_.$]*[: ^I^J]/
4467 * Idea by Bob Weiner, Motorola Inc. (1994)
4470 Asm_labels (FILE *inf
)
4474 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4476 /* If first char is alphabetic or one of [_.$], test for colon
4477 following identifier. */
4478 if (c_isalpha (*cp
) || *cp
== '_' || *cp
== '.' || *cp
== '$')
4480 /* Read past label. */
4482 while (c_isalnum (*cp
) || *cp
== '_' || *cp
== '.' || *cp
== '$')
4484 if (*cp
== ':' || c_isspace (*cp
))
4485 /* Found end of label, so copy it and add it to the table. */
4486 make_tag (lb
.buffer
, cp
- lb
.buffer
, true,
4487 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4495 * Perl sub names: /^sub[ \t\n]+[^ \t\n{]+/
4496 * /^use constant[ \t\n]+[^ \t\n{=,;]+/
4497 * Perl variable names: /^(my|local).../
4498 * Original code by Bart Robinson <lomew@cs.utah.edu> (1995)
4499 * Additions by Michael Ernst <mernst@alum.mit.edu> (1997)
4500 * Ideas by Kai Großjohann <Kai.Grossjohann@CS.Uni-Dortmund.DE> (2001)
4503 Perl_functions (FILE *inf
)
4505 char *package
= savestr ("main"); /* current package name */
4508 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4510 cp
= skip_spaces (cp
);
4512 if (LOOKING_AT (cp
, "package"))
4515 get_tag (cp
, &package
);
4517 else if (LOOKING_AT (cp
, "sub"))
4523 while (!notinname (*cp
))
4526 continue; /* nothing found */
4527 pos
= strchr (sp
, ':');
4528 if (pos
&& pos
< cp
&& pos
[1] == ':')
4529 /* The name is already qualified. */
4530 make_tag (sp
, cp
- sp
, true,
4531 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4535 char savechar
, *name
;
4539 name
= concat (package
, "::", sp
);
4541 make_tag (name
, strlen (name
), true,
4542 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4546 else if (LOOKING_AT (cp
, "use constant")
4547 || LOOKING_AT (cp
, "use constant::defer"))
4549 /* For hash style multi-constant like
4550 use constant { FOO => 123,
4552 only the first FOO is picked up. Parsing across the value
4553 expressions would be difficult in general, due to possible nested
4554 hashes, here-documents, etc. */
4556 cp
= skip_spaces (cp
+1);
4559 else if (globals
) /* only if we are tagging global vars */
4561 /* Skip a qualifier, if any. */
4562 bool qual
= LOOKING_AT (cp
, "my") || LOOKING_AT (cp
, "local");
4563 /* After "my" or "local", but before any following paren or space. */
4564 char *varstart
= cp
;
4566 if (qual
/* should this be removed? If yes, how? */
4567 && (*cp
== '$' || *cp
== '@' || *cp
== '%'))
4572 while (c_isalnum (*cp
) || *cp
== '_');
4576 /* Should be examining a variable list at this point;
4577 could insist on seeing an open parenthesis. */
4578 while (*cp
!= '\0' && *cp
!= ';' && *cp
!= '=' && *cp
!= ')')
4584 make_tag (varstart
, cp
- varstart
, false,
4585 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4594 * Look for /^[\t]*def[ \t\n]+[^ \t\n(:]+/ or /^class[ \t\n]+[^ \t\n(:]+/
4595 * Idea by Eric S. Raymond <esr@thyrsus.com> (1997)
4596 * More ideas by seb bacon <seb@jamkit.com> (2002)
4599 Python_functions (FILE *inf
)
4603 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4605 cp
= skip_spaces (cp
);
4606 if (LOOKING_AT (cp
, "def") || LOOKING_AT (cp
, "class"))
4609 while (!notinname (*cp
) && *cp
!= ':')
4611 make_tag (name
, cp
- name
, true,
4612 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4619 * Original code by Xi Lu <lx@shellcodes.org> (2015)
4622 Ruby_functions (FILE *inf
)
4626 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4628 bool is_class
= false;
4629 bool is_method
= false;
4632 cp
= skip_spaces (cp
);
4633 if (c_isalpha (*cp
) && c_isupper (*cp
)) /* constants */
4635 char *bp
, *colon
= NULL
;
4639 for (cp
++; c_isalnum (*cp
) || *cp
== '_' || *cp
== ':'; cp
++)
4646 bp
= skip_spaces (cp
);
4647 if (*bp
== '=' && !(bp
[1] == '=' || bp
[1] == '>'))
4649 if (colon
&& !c_isspace (colon
[1]))
4651 make_tag (name
, cp
- name
, false,
4652 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4656 else if ((is_method
= LOOKING_AT (cp
, "def")) /* module/class/method */
4657 || (is_class
= LOOKING_AT (cp
, "class"))
4658 || LOOKING_AT (cp
, "module"))
4660 const char self_name
[] = "self.";
4661 const size_t self_size1
= sizeof (self_name
) - 1;
4665 /* Ruby method names can end in a '='. Also, operator overloading can
4666 define operators whose names include '='. */
4667 while (!notinname (*cp
) || *cp
== '=')
4670 /* Remove "self." from the method name. */
4671 if (cp
- name
> self_size1
4672 && strneq (name
, self_name
, self_size1
))
4675 /* Remove the class/module qualifiers from method names. */
4680 for (q
= name
; q
< cp
&& *q
!= '.'; q
++)
4682 if (q
< cp
- 1) /* punt if we see just "FOO." */
4686 /* Don't tag singleton classes. */
4687 if (is_class
&& strneq (name
, "<<", 2) && cp
== name
+ 2)
4690 make_tag (name
, cp
- name
, true,
4691 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4695 /* Tag accessors and aliases. */
4696 while (*cp
&& *cp
!= '#')
4698 bool reader
= false, writer
= false, alias
= false;
4700 if (LOOKING_AT (cp
, "attr_reader"))
4702 else if (LOOKING_AT (cp
, "attr_writer"))
4704 else if (LOOKING_AT (cp
, "attr_accessor"))
4709 else if (LOOKING_AT (cp
, "alias_method"))
4711 if (reader
|| writer
|| alias
)
4716 cp
= skip_name (cp
);
4718 make_tag (np
, cp
- np
, true,
4719 lb
.buffer
, cp
- lb
.buffer
+ 1,
4720 lineno
, linecharno
);
4723 size_t name_len
= cp
- np
+ 1;
4724 char *wr_name
= xnew (name_len
+ 1, char);
4726 memcpy (wr_name
, np
, name_len
- 1);
4727 memcpy (wr_name
+ name_len
- 1, "=", 2);
4728 pfnote (wr_name
, true, lb
.buffer
, cp
- lb
.buffer
+ 1,
4729 lineno
, linecharno
);
4733 make_tag (np
, cp
- np
, true,
4734 lb
.buffer
, cp
- lb
.buffer
+ 1,
4735 lineno
, linecharno
);
4736 while (*cp
&& *cp
!= '#' && *cp
!= ';')
4740 && (cp
= skip_spaces (cp
+ 1), *cp
&& *cp
!= '#'));
4742 cp
= skip_name (cp
);
4743 while (*cp
&& *cp
!= '#' && notinname (*cp
))
4754 * - /^[ \t]*function[ \t\n]+[^ \t\n(]+/
4755 * - /^[ \t]*class[ \t\n]+[^ \t\n]+/
4756 * - /^[ \t]*define\(\"[^\"]+/
4757 * Only with --members:
4758 * - /^[ \t]*var[ \t\n]+\$[^ \t\n=;]/
4759 * Idea by Diez B. Roggisch (2001)
4762 PHP_functions (FILE *inf
)
4765 bool search_identifier
= false;
4767 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4769 cp
= skip_spaces (cp
);
4771 if (search_identifier
4774 while (!notinname (*cp
))
4776 make_tag (name
, cp
- name
, true,
4777 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4778 search_identifier
= false;
4780 else if (LOOKING_AT (cp
, "function"))
4783 cp
= skip_spaces (cp
+1);
4787 while (!notinname (*cp
))
4789 make_tag (name
, cp
- name
, true,
4790 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4793 search_identifier
= true;
4795 else if (LOOKING_AT (cp
, "class"))
4800 while (*cp
!= '\0' && !c_isspace (*cp
))
4802 make_tag (name
, cp
- name
, false,
4803 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4806 search_identifier
= true;
4808 else if (strneq (cp
, "define", 6)
4809 && (cp
= skip_spaces (cp
+6))
4811 && (*cp
== '"' || *cp
== '\''))
4815 while (*cp
!= quote
&& *cp
!= '\0')
4817 make_tag (name
, cp
- name
, false,
4818 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4821 && LOOKING_AT (cp
, "var")
4825 while (!notinname (*cp
))
4827 make_tag (name
, cp
- name
, false,
4828 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4835 * Cobol tag functions
4836 * We could look for anything that could be a paragraph name.
4837 * i.e. anything that starts in column 8 is one word and ends in a full stop.
4838 * Idea by Corny de Souza (1993)
4841 Cobol_paragraphs (FILE *inf
)
4843 register char *bp
, *ep
;
4845 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4851 /* If eoln, compiler option or comment ignore whole line. */
4852 if (bp
[-1] != ' ' || !c_isalnum (bp
[0]))
4855 for (ep
= bp
; c_isalnum (*ep
) || *ep
== '-'; ep
++)
4858 make_tag (bp
, ep
- bp
, true,
4859 lb
.buffer
, ep
- lb
.buffer
+ 1, lineno
, linecharno
);
4866 * Ideas by Assar Westerlund <assar@sics.se> (2001)
4869 Makefile_targets (FILE *inf
)
4873 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4875 if (*bp
== '\t' || *bp
== '#')
4877 while (*bp
!= '\0' && *bp
!= '=' && *bp
!= ':')
4879 if (*bp
== ':' || (globals
&& *bp
== '='))
4881 /* We should detect if there is more than one tag, but we do not.
4882 We just skip initial and final spaces. */
4883 char * namestart
= skip_spaces (lb
.buffer
);
4884 while (--bp
> namestart
)
4885 if (!notinname (*bp
))
4887 make_tag (namestart
, bp
- namestart
+ 1, true,
4888 lb
.buffer
, bp
- lb
.buffer
+ 2, lineno
, linecharno
);
4896 * Original code by Mosur K. Mohan (1989)
4898 * Locates tags for procedures & functions. Doesn't do any type- or
4899 * var-definitions. It does look for the keyword "extern" or
4900 * "forward" immediately following the procedure statement; if found,
4901 * the tag is skipped.
4904 Pascal_functions (FILE *inf
)
4906 linebuffer tline
; /* mostly copied from C_entries */
4908 int save_lineno
, namelen
, taglen
;
4911 bool /* each of these flags is true if: */
4912 incomment
, /* point is inside a comment */
4913 inquote
, /* point is inside '..' string */
4914 get_tagname
, /* point is after PROCEDURE/FUNCTION
4915 keyword, so next item = potential tag */
4916 found_tag
, /* point is after a potential tag */
4917 inparms
, /* point is within parameter-list */
4918 verify_tag
; /* point has passed the parm-list, so the
4919 next token will determine whether this
4920 is a FORWARD/EXTERN to be ignored, or
4921 whether it is a real tag */
4923 save_lcno
= save_lineno
= namelen
= taglen
= 0; /* keep compiler quiet */
4924 name
= NULL
; /* keep compiler quiet */
4927 linebuffer_init (&tline
);
4929 incomment
= inquote
= false;
4930 found_tag
= false; /* have a proc name; check if extern */
4931 get_tagname
= false; /* found "procedure" keyword */
4932 inparms
= false; /* found '(' after "proc" */
4933 verify_tag
= false; /* check if "extern" is ahead */
4936 while (perhaps_more_input (inf
)) /* long main loop to get next char */
4939 if (c
== '\0') /* if end of line */
4941 readline (&lb
, inf
);
4945 if (!((found_tag
&& verify_tag
)
4947 c
= *dbp
++; /* only if don't need *dbp pointing
4948 to the beginning of the name of
4949 the procedure or function */
4953 if (c
== '}') /* within { } comments */
4955 else if (c
== '*' && *dbp
== ')') /* within (* *) comments */
4972 inquote
= true; /* found first quote */
4974 case '{': /* found open { comment */
4978 if (*dbp
== '*') /* found open (* comment */
4983 else if (found_tag
) /* found '(' after tag, i.e., parm-list */
4986 case ')': /* end of parms list */
4991 if (found_tag
&& !inparms
) /* end of proc or fn stmt */
4998 if (found_tag
&& verify_tag
&& (*dbp
!= ' '))
5000 /* Check if this is an "extern" declaration. */
5003 if (c_tolower (*dbp
) == 'e')
5005 if (nocase_tail ("extern")) /* superfluous, really! */
5011 else if (c_tolower (*dbp
) == 'f')
5013 if (nocase_tail ("forward")) /* check for forward reference */
5019 if (found_tag
&& verify_tag
) /* not external proc, so make tag */
5023 make_tag (name
, namelen
, true,
5024 tline
.buffer
, taglen
, save_lineno
, save_lcno
);
5028 if (get_tagname
) /* grab name of proc or fn */
5035 /* Find block name. */
5036 for (cp
= dbp
+ 1; *cp
!= '\0' && !endtoken (*cp
); cp
++)
5039 /* Save all values for later tagging. */
5040 linebuffer_setlen (&tline
, lb
.len
);
5041 strcpy (tline
.buffer
, lb
.buffer
);
5042 save_lineno
= lineno
;
5043 save_lcno
= linecharno
;
5044 name
= tline
.buffer
+ (dbp
- lb
.buffer
);
5046 taglen
= cp
- lb
.buffer
+ 1;
5048 dbp
= cp
; /* set dbp to e-o-token */
5049 get_tagname
= false;
5053 /* And proceed to check for "extern". */
5055 else if (!incomment
&& !inquote
&& !found_tag
)
5057 /* Check for proc/fn keywords. */
5058 switch (c_tolower (c
))
5061 if (nocase_tail ("rocedure")) /* c = 'p', dbp has advanced */
5065 if (nocase_tail ("unction"))
5070 } /* while not eof */
5072 free (tline
.buffer
);
5077 * Lisp tag functions
5078 * look for (def or (DEF, quote or QUOTE
5081 static void L_getit (void);
5086 if (*dbp
== '\'') /* Skip prefix quote */
5088 else if (*dbp
== '(')
5091 /* Try to skip "(quote " */
5092 if (!LOOKING_AT (dbp
, "quote") && !LOOKING_AT (dbp
, "QUOTE"))
5093 /* Ok, then skip "(" before name in (defstruct (foo)) */
5094 dbp
= skip_spaces (dbp
);
5096 get_tag (dbp
, NULL
);
5100 Lisp_functions (FILE *inf
)
5102 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
5107 /* "(defvar foo)" is a declaration rather than a definition. */
5111 if (LOOKING_AT (p
, "defvar"))
5113 p
= skip_name (p
); /* past var name */
5114 p
= skip_spaces (p
);
5120 if (strneq (dbp
+ 1, "cl-", 3) || strneq (dbp
+ 1, "CL-", 3))
5123 if (strneq (dbp
+1, "def", 3) || strneq (dbp
+1, "DEF", 3))
5125 dbp
= skip_non_spaces (dbp
);
5126 dbp
= skip_spaces (dbp
);
5131 /* Check for (foo::defmumble name-defined ... */
5134 while (!notinname (*dbp
) && *dbp
!= ':');
5139 while (*dbp
== ':');
5141 if (strneq (dbp
, "def", 3) || strneq (dbp
, "DEF", 3))
5143 dbp
= skip_non_spaces (dbp
);
5144 dbp
= skip_spaces (dbp
);
5154 * Lua script language parsing
5155 * Original code by David A. Capello <dacap@users.sourceforge.net> (2004)
5157 * "function" and "local function" are tags if they start at column 1.
5160 Lua_functions (FILE *inf
)
5164 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
5166 bp
= skip_spaces (bp
);
5167 if (bp
[0] != 'f' && bp
[0] != 'l')
5170 (void)LOOKING_AT (bp
, "local"); /* skip possible "local" */
5172 if (LOOKING_AT (bp
, "function"))
5174 char *tag_name
, *tp_dot
, *tp_colon
;
5176 get_tag (bp
, &tag_name
);
5177 /* If the tag ends with ".foo" or ":foo", make an additional tag for
5179 tp_dot
= strrchr (tag_name
, '.');
5180 tp_colon
= strrchr (tag_name
, ':');
5181 if (tp_dot
|| tp_colon
)
5183 char *p
= tp_dot
> tp_colon
? tp_dot
: tp_colon
;
5184 int len_add
= p
- tag_name
+ 1;
5186 get_tag (bp
+ len_add
, NULL
);
5195 * Just look for lines where the first character is '/'
5196 * Also look at "defineps" for PSWrap
5198 * Richard Mlynarik <mly@adoc.xerox.com> (1997)
5199 * Masatake Yamato <masata-y@is.aist-nara.ac.jp> (1999)
5202 PS_functions (FILE *inf
)
5204 register char *bp
, *ep
;
5206 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
5211 *ep
!= '\0' && *ep
!= ' ' && *ep
!= '{';
5214 make_tag (bp
, ep
- bp
, true,
5215 lb
.buffer
, ep
- lb
.buffer
+ 1, lineno
, linecharno
);
5217 else if (LOOKING_AT (bp
, "defineps"))
5225 * Ignore anything after \ followed by space or in ( )
5226 * Look for words defined by :
5227 * Look for constant, code, create, defer, value, and variable
5228 * OBP extensions: Look for buffer:, field,
5229 * Ideas by Eduardo Horvath <eeh@netbsd.org> (2004)
5232 Forth_words (FILE *inf
)
5236 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
5237 while ((bp
= skip_spaces (bp
))[0] != '\0')
5238 if (bp
[0] == '\\' && c_isspace (bp
[1]))
5239 break; /* read next line */
5240 else if (bp
[0] == '(' && c_isspace (bp
[1]))
5241 do /* skip to ) or eol */
5243 while (*bp
!= ')' && *bp
!= '\0');
5244 else if ((bp
[0] == ':' && c_isspace (bp
[1]) && bp
++)
5245 || LOOKING_AT_NOCASE (bp
, "constant")
5246 || LOOKING_AT_NOCASE (bp
, "code")
5247 || LOOKING_AT_NOCASE (bp
, "create")
5248 || LOOKING_AT_NOCASE (bp
, "defer")
5249 || LOOKING_AT_NOCASE (bp
, "value")
5250 || LOOKING_AT_NOCASE (bp
, "variable")
5251 || LOOKING_AT_NOCASE (bp
, "buffer:")
5252 || LOOKING_AT_NOCASE (bp
, "field"))
5253 get_tag (skip_spaces (bp
), NULL
); /* Yay! A definition! */
5255 bp
= skip_non_spaces (bp
);
5260 * Scheme tag functions
5261 * look for (def... xyzzy
5263 * (def ... ((...(xyzzy ....
5265 * Original code by Ken Haase (1985?)
5268 Scheme_functions (FILE *inf
)
5272 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
5274 if (strneq (bp
, "(def", 4) || strneq (bp
, "(DEF", 4))
5276 bp
= skip_non_spaces (bp
+4);
5277 /* Skip over open parens and white space. Don't continue past
5279 while (*bp
&& notinname (*bp
))
5283 if (LOOKING_AT (bp
, "(SET!") || LOOKING_AT (bp
, "(set!"))
5289 /* Find tags in TeX and LaTeX input files. */
5291 /* TEX_toktab is a table of TeX control sequences that define tags.
5292 * Each entry records one such control sequence.
5294 * Original code from who knows whom.
5296 * Stefan Monnier (2002)
5299 static linebuffer
*TEX_toktab
= NULL
; /* Table with tag tokens */
5301 /* Default set of control sequences to put into TEX_toktab.
5302 The value of environment var TEXTAGS is prepended to this. */
5303 static const char *TEX_defenv
= "\
5304 :chapter:section:subsection:subsubsection:eqno:label:ref:cite:bibitem\
5305 :part:appendix:entry:index:def\
5306 :newcommand:renewcommand:newenvironment:renewenvironment";
5308 static void TEX_decode_env (const char *, const char *);
5311 * TeX/LaTeX scanning loop.
5314 TeX_commands (FILE *inf
)
5319 char TEX_esc
= '\0';
5320 char TEX_opgrp
, TEX_clgrp
;
5322 /* Initialize token table once from environment. */
5323 if (TEX_toktab
== NULL
)
5324 TEX_decode_env ("TEXTAGS", TEX_defenv
);
5326 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5328 /* Look at each TEX keyword in line. */
5331 /* Look for a TEX escape. */
5335 if (c
== '\0' || c
== '%')
5338 /* Select either \ or ! as escape character, whichever comes
5339 first outside a comment. */
5360 for (key
= TEX_toktab
; key
->buffer
!= NULL
; key
++)
5361 if (strneq (cp
, key
->buffer
, key
->len
))
5364 int namelen
, linelen
;
5367 cp
= skip_spaces (cp
+ key
->len
);
5368 if (*cp
== TEX_opgrp
)
5374 (!c_isspace (*p
) && *p
!= '#' &&
5375 *p
!= TEX_opgrp
&& *p
!= TEX_clgrp
);
5380 if (!opgrp
|| *p
== TEX_clgrp
)
5382 while (*p
!= '\0' && *p
!= TEX_opgrp
&& *p
!= TEX_clgrp
)
5384 linelen
= p
- lb
.buffer
+ 1;
5386 make_tag (cp
, namelen
, true,
5387 lb
.buffer
, linelen
, lineno
, linecharno
);
5388 goto tex_next_line
; /* We only tag a line once */
5396 /* Read environment and prepend it to the default string.
5397 Build token table. */
5399 TEX_decode_env (const char *evarname
, const char *defenv
)
5401 register const char *env
, *p
;
5404 /* Append default string to environment. */
5405 env
= getenv (evarname
);
5409 env
= concat (env
, defenv
, "");
5411 /* Allocate a token table */
5412 for (len
= 1, p
= env
; (p
= strchr (p
, ':')); )
5415 TEX_toktab
= xnew (len
, linebuffer
);
5417 /* Unpack environment string into token table. Be careful about */
5418 /* zero-length strings (leading ':', "::" and trailing ':') */
5419 for (i
= 0; *env
!= '\0';)
5421 p
= strchr (env
, ':');
5422 if (!p
) /* End of environment string. */
5423 p
= env
+ strlen (env
);
5425 { /* Only non-zero strings. */
5426 TEX_toktab
[i
].buffer
= savenstr (env
, p
- env
);
5427 TEX_toktab
[i
].len
= p
- env
;
5434 TEX_toktab
[i
].buffer
= NULL
; /* Mark end of table. */
5435 TEX_toktab
[i
].len
= 0;
5442 /* Texinfo support. Dave Love, Mar. 2000. */
5444 Texinfo_nodes (FILE *inf
)
5447 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5448 if (LOOKING_AT (cp
, "@node"))
5451 while (*cp
!= '\0' && *cp
!= ',')
5453 make_tag (start
, cp
- start
, true,
5454 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
5461 * Contents of <title>, <h1>, <h2>, <h3> are tags.
5462 * Contents of <a name=xxx> are tags with name xxx.
5464 * Francesco Potortì, 2002.
5467 HTML_labels (FILE *inf
)
5469 bool getnext
= false; /* next text outside of HTML tags is a tag */
5470 bool skiptag
= false; /* skip to the end of the current HTML tag */
5471 bool intag
= false; /* inside an html tag, looking for ID= */
5472 bool inanchor
= false; /* when INTAG, is an anchor, look for NAME= */
5476 linebuffer_setlen (&token_name
, 0); /* no name in buffer */
5478 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
5479 for (;;) /* loop on the same line */
5481 if (skiptag
) /* skip HTML tag */
5483 while (*dbp
!= '\0' && *dbp
!= '>')
5489 continue; /* look on the same line */
5491 break; /* go to next line */
5494 else if (intag
) /* look for "name=" or "id=" */
5496 while (*dbp
!= '\0' && *dbp
!= '>'
5497 && c_tolower (*dbp
) != 'n' && c_tolower (*dbp
) != 'i')
5500 break; /* go to next line */
5505 continue; /* look on the same line */
5507 if ((inanchor
&& LOOKING_AT_NOCASE (dbp
, "name="))
5508 || LOOKING_AT_NOCASE (dbp
, "id="))
5510 bool quoted
= (dbp
[0] == '"');
5513 for (end
= ++dbp
; *end
!= '\0' && *end
!= '"'; end
++)
5516 for (end
= dbp
; *end
!= '\0' && intoken (*end
); end
++)
5518 linebuffer_setlen (&token_name
, end
- dbp
);
5519 memcpy (token_name
.buffer
, dbp
, end
- dbp
);
5520 token_name
.buffer
[end
- dbp
] = '\0';
5523 intag
= false; /* we found what we looked for */
5524 skiptag
= true; /* skip to the end of the tag */
5525 getnext
= true; /* then grab the text */
5526 continue; /* look on the same line */
5531 else if (getnext
) /* grab next tokens and tag them */
5533 dbp
= skip_spaces (dbp
);
5535 break; /* go to next line */
5539 inanchor
= (c_tolower (dbp
[1]) == 'a' && !intoken (dbp
[2]));
5540 continue; /* look on the same line */
5543 for (end
= dbp
+ 1; *end
!= '\0' && *end
!= '<'; end
++)
5545 make_tag (token_name
.buffer
, token_name
.len
, true,
5546 dbp
, end
- dbp
, lineno
, linecharno
);
5547 linebuffer_setlen (&token_name
, 0); /* no name in buffer */
5549 break; /* go to next line */
5552 else /* look for an interesting HTML tag */
5554 while (*dbp
!= '\0' && *dbp
!= '<')
5557 break; /* go to next line */
5559 if (c_tolower (dbp
[1]) == 'a' && !intoken (dbp
[2]))
5562 continue; /* look on the same line */
5564 else if (LOOKING_AT_NOCASE (dbp
, "<title>")
5565 || LOOKING_AT_NOCASE (dbp
, "<h1>")
5566 || LOOKING_AT_NOCASE (dbp
, "<h2>")
5567 || LOOKING_AT_NOCASE (dbp
, "<h3>"))
5571 continue; /* look on the same line */
5582 * Assumes that the predicate or rule starts at column 0.
5583 * Only the first clause of a predicate or rule is added.
5584 * Original code by Sunichirou Sugou (1989)
5585 * Rewritten by Anders Lindgren (1996)
5587 static size_t prolog_pr (char *, char *);
5588 static void prolog_skip_comment (linebuffer
*, FILE *);
5589 static size_t prolog_atom (char *, size_t);
5592 Prolog_functions (FILE *inf
)
5602 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5604 if (cp
[0] == '\0') /* Empty line */
5606 else if (c_isspace (cp
[0])) /* Not a predicate */
5608 else if (cp
[0] == '/' && cp
[1] == '*') /* comment. */
5609 prolog_skip_comment (&lb
, inf
);
5610 else if ((len
= prolog_pr (cp
, last
)) > 0)
5612 /* Predicate or rule. Store the function name so that we
5613 only generate a tag for the first clause. */
5615 last
= xnew (len
+ 1, char);
5616 else if (len
+ 1 > allocated
)
5617 xrnew (last
, len
+ 1, char);
5618 allocated
= len
+ 1;
5619 memcpy (last
, cp
, len
);
5628 prolog_skip_comment (linebuffer
*plb
, FILE *inf
)
5634 for (cp
= plb
->buffer
; *cp
!= '\0'; cp
++)
5635 if (cp
[0] == '*' && cp
[1] == '/')
5637 readline (plb
, inf
);
5639 while (perhaps_more_input (inf
));
5643 * A predicate or rule definition is added if it matches:
5644 * <beginning of line><Prolog Atom><whitespace>(
5645 * or <beginning of line><Prolog Atom><whitespace>:-
5647 * It is added to the tags database if it doesn't match the
5648 * name of the previous clause header.
5650 * Return the size of the name of the predicate or rule, or 0 if no
5654 prolog_pr (char *s
, char *last
)
5656 /* Name of last clause. */
5661 pos
= prolog_atom (s
, 0);
5666 pos
= skip_spaces (s
+ pos
) - s
;
5669 || (s
[pos
] == '(' && (pos
+= 1))
5670 || (s
[pos
] == ':' && s
[pos
+ 1] == '-' && (pos
+= 2)))
5671 && (last
== NULL
/* save only the first clause */
5672 || len
!= strlen (last
)
5673 || !strneq (s
, last
, len
)))
5675 make_tag (s
, len
, true, s
, pos
, lineno
, linecharno
);
5683 * Consume a Prolog atom.
5684 * Return the number of bytes consumed, or 0 if there was an error.
5686 * A prolog atom, in this context, could be one of:
5687 * - An alphanumeric sequence, starting with a lower case letter.
5688 * - A quoted arbitrary string. Single quotes can escape themselves.
5689 * Backslash quotes everything.
5692 prolog_atom (char *s
, size_t pos
)
5698 if (c_islower (s
[pos
]) || s
[pos
] == '_')
5700 /* The atom is unquoted. */
5702 while (c_isalnum (s
[pos
]) || s
[pos
] == '_')
5706 return pos
- origpos
;
5708 else if (s
[pos
] == '\'')
5719 pos
++; /* A double quote */
5721 else if (s
[pos
] == '\0')
5722 /* Multiline quoted atoms are ignored. */
5724 else if (s
[pos
] == '\\')
5726 if (s
[pos
+1] == '\0')
5733 return pos
- origpos
;
5741 * Support for Erlang
5743 * Generates tags for functions, defines, and records.
5744 * Assumes that Erlang functions start at column 0.
5745 * Original code by Anders Lindgren (1996)
5747 static int erlang_func (char *, char *);
5748 static void erlang_attribute (char *);
5749 static int erlang_atom (char *);
5752 Erlang_functions (FILE *inf
)
5762 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5764 if (cp
[0] == '\0') /* Empty line */
5766 else if (c_isspace (cp
[0])) /* Not function nor attribute */
5768 else if (cp
[0] == '%') /* comment */
5770 else if (cp
[0] == '"') /* Sometimes, strings start in column one */
5772 else if (cp
[0] == '-') /* attribute, e.g. "-define" */
5774 erlang_attribute (cp
);
5781 else if ((len
= erlang_func (cp
, last
)) > 0)
5784 * Function. Store the function name so that we only
5785 * generates a tag for the first clause.
5788 last
= xnew (len
+ 1, char);
5789 else if (len
+ 1 > allocated
)
5790 xrnew (last
, len
+ 1, char);
5791 allocated
= len
+ 1;
5792 memcpy (last
, cp
, len
);
5801 * A function definition is added if it matches:
5802 * <beginning of line><Erlang Atom><whitespace>(
5804 * It is added to the tags database if it doesn't match the
5805 * name of the previous clause header.
5807 * Return the size of the name of the function, or 0 if no function
5811 erlang_func (char *s
, char *last
)
5813 /* Name of last clause. */
5818 pos
= erlang_atom (s
);
5823 pos
= skip_spaces (s
+ pos
) - s
;
5825 /* Save only the first clause. */
5828 || len
!= (int)strlen (last
)
5829 || !strneq (s
, last
, len
)))
5831 make_tag (s
, len
, true, s
, pos
, lineno
, linecharno
);
5840 * Handle attributes. Currently, tags are generated for defines
5843 * They are on the form:
5844 * -define(foo, bar).
5845 * -define(Foo(M, N), M+N).
5846 * -record(graph, {vtab = notable, cyclic = true}).
5849 erlang_attribute (char *s
)
5853 if ((LOOKING_AT (cp
, "-define") || LOOKING_AT (cp
, "-record"))
5856 int len
= erlang_atom (skip_spaces (cp
));
5858 make_tag (cp
, len
, true, s
, cp
+ len
- s
, lineno
, linecharno
);
5865 * Consume an Erlang atom (or variable).
5866 * Return the number of bytes consumed, or -1 if there was an error.
5869 erlang_atom (char *s
)
5873 if (c_isalpha (s
[pos
]) || s
[pos
] == '_')
5875 /* The atom is unquoted. */
5878 while (c_isalnum (s
[pos
]) || s
[pos
] == '_');
5880 else if (s
[pos
] == '\'')
5882 for (pos
++; s
[pos
] != '\''; pos
++)
5883 if (s
[pos
] == '\0' /* multiline quoted atoms are ignored */
5884 || (s
[pos
] == '\\' && s
[++pos
] == '\0'))
5893 static char *scan_separators (char *);
5894 static void add_regex (char *, language
*);
5895 static char *substitute (char *, char *, struct re_registers
*);
5898 * Take a string like "/blah/" and turn it into "blah", verifying
5899 * that the first and last characters are the same, and handling
5900 * quoted separator characters. Actually, stops on the occurrence of
5901 * an unquoted separator. Also process \t, \n, etc. and turn into
5902 * appropriate characters. Works in place. Null terminates name string.
5903 * Returns pointer to terminating separator, or NULL for
5904 * unterminated regexps.
5907 scan_separators (char *name
)
5910 char *copyto
= name
;
5911 bool quoted
= false;
5913 for (++name
; *name
!= '\0'; ++name
)
5919 case 'a': *copyto
++ = '\007'; break; /* BEL (bell) */
5920 case 'b': *copyto
++ = '\b'; break; /* BS (back space) */
5921 case 'd': *copyto
++ = 0177; break; /* DEL (delete) */
5922 case 'e': *copyto
++ = 033; break; /* ESC (delete) */
5923 case 'f': *copyto
++ = '\f'; break; /* FF (form feed) */
5924 case 'n': *copyto
++ = '\n'; break; /* NL (new line) */
5925 case 'r': *copyto
++ = '\r'; break; /* CR (carriage return) */
5926 case 't': *copyto
++ = '\t'; break; /* TAB (horizontal tab) */
5927 case 'v': *copyto
++ = '\v'; break; /* VT (vertical tab) */
5933 /* Something else is quoted, so preserve the quote. */
5941 else if (*name
== '\\')
5943 else if (*name
== sep
)
5949 name
= NULL
; /* signal unterminated regexp */
5951 /* Terminate copied string. */
5956 /* Look at the argument of --regex or --no-regex and do the right
5957 thing. Same for each line of a regexp file. */
5959 analyze_regex (char *regex_arg
)
5961 if (regex_arg
== NULL
)
5963 free_regexps (); /* --no-regex: remove existing regexps */
5967 /* A real --regexp option or a line in a regexp file. */
5968 switch (regex_arg
[0])
5970 /* Comments in regexp file or null arg to --regex. */
5976 /* Read a regex file. This is recursive and may result in a
5977 loop, which will stop when the file descriptors are exhausted. */
5981 linebuffer regexbuf
;
5982 char *regexfile
= regex_arg
+ 1;
5984 /* regexfile is a file containing regexps, one per line. */
5985 regexfp
= fopen (regexfile
, "r" FOPEN_BINARY
);
5986 if (regexfp
== NULL
)
5988 linebuffer_init (®exbuf
);
5989 while (readline_internal (®exbuf
, regexfp
, regexfile
) > 0)
5990 analyze_regex (regexbuf
.buffer
);
5991 free (regexbuf
.buffer
);
5992 if (fclose (regexfp
) != 0)
5997 /* Regexp to be used for a specific language only. */
6001 char *lang_name
= regex_arg
+ 1;
6004 for (cp
= lang_name
; *cp
!= '}'; cp
++)
6007 error ("unterminated language name in regex: %s", regex_arg
);
6011 lang
= get_language_from_langname (lang_name
);
6014 add_regex (cp
, lang
);
6018 /* Regexp to be used for any language. */
6020 add_regex (regex_arg
, NULL
);
6025 /* Separate the regexp pattern, compile it,
6026 and care for optional name and modifiers. */
6028 add_regex (char *regexp_pattern
, language
*lang
)
6030 static struct re_pattern_buffer zeropattern
;
6031 char sep
, *pat
, *name
, *modifiers
;
6034 struct re_pattern_buffer
*patbuf
;
6037 force_explicit_name
= true, /* do not use implicit tag names */
6038 ignore_case
= false, /* case is significant */
6039 multi_line
= false, /* matches are done one line at a time */
6040 single_line
= false; /* dot does not match newline */
6043 if (strlen (regexp_pattern
) < 3)
6045 error ("null regexp");
6048 sep
= regexp_pattern
[0];
6049 name
= scan_separators (regexp_pattern
);
6052 error ("%s: unterminated regexp", regexp_pattern
);
6057 error ("null name for regexp \"%s\"", regexp_pattern
);
6060 modifiers
= scan_separators (name
);
6061 if (modifiers
== NULL
) /* no terminating separator --> no name */
6067 modifiers
+= 1; /* skip separator */
6069 /* Parse regex modifiers. */
6070 for (; modifiers
[0] != '\0'; modifiers
++)
6071 switch (modifiers
[0])
6074 if (modifiers
== name
)
6075 error ("forcing explicit tag name but no name, ignoring");
6076 force_explicit_name
= true;
6086 need_filebuf
= true;
6089 error ("invalid regexp modifier '%c', ignoring", modifiers
[0]);
6093 patbuf
= xnew (1, struct re_pattern_buffer
);
6094 *patbuf
= zeropattern
;
6097 static char lc_trans
[UCHAR_MAX
+ 1];
6099 for (i
= 0; i
< UCHAR_MAX
+ 1; i
++)
6100 lc_trans
[i
] = c_tolower (i
);
6101 patbuf
->translate
= lc_trans
; /* translation table to fold case */
6105 pat
= concat ("^", regexp_pattern
, ""); /* anchor to beginning of line */
6107 pat
= regexp_pattern
;
6110 re_set_syntax (RE_SYNTAX_EMACS
| RE_DOT_NEWLINE
);
6112 re_set_syntax (RE_SYNTAX_EMACS
);
6114 err
= re_compile_pattern (pat
, strlen (pat
), patbuf
);
6119 error ("%s while compiling pattern", err
);
6124 p_head
= xnew (1, regexp
);
6125 p_head
->pattern
= savestr (regexp_pattern
);
6126 p_head
->p_next
= rp
;
6127 p_head
->lang
= lang
;
6128 p_head
->pat
= patbuf
;
6129 p_head
->name
= savestr (name
);
6130 p_head
->error_signaled
= false;
6131 p_head
->force_explicit_name
= force_explicit_name
;
6132 p_head
->ignore_case
= ignore_case
;
6133 p_head
->multi_line
= multi_line
;
6137 * Do the substitutions indicated by the regular expression and
6141 substitute (char *in
, char *out
, struct re_registers
*regs
)
6144 int size
, dig
, diglen
;
6147 size
= strlen (out
);
6149 /* Pass 1: figure out how much to allocate by finding all \N strings. */
6150 if (out
[size
- 1] == '\\')
6151 fatal ("pattern error in \"%s\"", out
);
6152 for (t
= strchr (out
, '\\');
6154 t
= strchr (t
+ 2, '\\'))
6155 if (c_isdigit (t
[1]))
6158 diglen
= regs
->end
[dig
] - regs
->start
[dig
];
6164 /* Allocate space and do the substitutions. */
6166 result
= xnew (size
+ 1, char);
6168 for (t
= result
; *out
!= '\0'; out
++)
6169 if (*out
== '\\' && c_isdigit (*++out
))
6172 diglen
= regs
->end
[dig
] - regs
->start
[dig
];
6173 memcpy (t
, in
+ regs
->start
[dig
], diglen
);
6180 assert (t
<= result
+ size
);
6181 assert (t
- result
== (int)strlen (result
));
6186 /* Deallocate all regexps. */
6191 while (p_head
!= NULL
)
6193 rp
= p_head
->p_next
;
6194 free (p_head
->pattern
);
6195 free (p_head
->name
);
6203 * Reads the whole file as a single string from `filebuf' and looks for
6204 * multi-line regular expressions, creating tags on matches.
6205 * readline already dealt with normal regexps.
6207 * Idea by Ben Wing <ben@666.com> (2002).
6210 regex_tag_multiline (void)
6212 char *buffer
= filebuf
.buffer
;
6216 for (rp
= p_head
; rp
!= NULL
; rp
= rp
->p_next
)
6220 if (!rp
->multi_line
)
6221 continue; /* skip normal regexps */
6223 /* Generic initializations before parsing file from memory. */
6224 lineno
= 1; /* reset global line number */
6225 charno
= 0; /* reset global char number */
6226 linecharno
= 0; /* reset global char number of line start */
6228 /* Only use generic regexps or those for the current language. */
6229 if (rp
->lang
!= NULL
&& rp
->lang
!= curfdp
->lang
)
6232 while (match
>= 0 && match
< filebuf
.len
)
6234 match
= re_search (rp
->pat
, buffer
, filebuf
.len
, charno
,
6235 filebuf
.len
- match
, &rp
->regs
);
6240 if (!rp
->error_signaled
)
6242 error ("regexp stack overflow while matching \"%s\"",
6244 rp
->error_signaled
= true;
6251 if (match
== rp
->regs
.end
[0])
6253 if (!rp
->error_signaled
)
6255 error ("regexp matches the empty string: \"%s\"",
6257 rp
->error_signaled
= true;
6259 match
= -3; /* exit from while loop */
6263 /* Match occurred. Construct a tag. */
6264 while (charno
< rp
->regs
.end
[0])
6265 if (buffer
[charno
++] == '\n')
6266 lineno
++, linecharno
= charno
;
6268 if (name
[0] == '\0')
6270 else /* make a named tag */
6271 name
= substitute (buffer
, rp
->name
, &rp
->regs
);
6272 if (rp
->force_explicit_name
)
6273 /* Force explicit tag name, if a name is there. */
6274 pfnote (name
, true, buffer
+ linecharno
,
6275 charno
- linecharno
+ 1, lineno
, linecharno
);
6277 make_tag (name
, strlen (name
), true, buffer
+ linecharno
,
6278 charno
- linecharno
+ 1, lineno
, linecharno
);
6287 nocase_tail (const char *cp
)
6291 while (*cp
!= '\0' && c_tolower (*cp
) == c_tolower (dbp
[len
]))
6293 if (*cp
== '\0' && !intoken (dbp
[len
]))
6302 get_tag (register char *bp
, char **namepp
)
6304 register char *cp
= bp
;
6308 /* Go till you get to white space or a syntactic break */
6309 for (cp
= bp
+ 1; !notinname (*cp
); cp
++)
6311 make_tag (bp
, cp
- bp
, true,
6312 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
6316 *namepp
= savenstr (bp
, cp
- bp
);
6320 * Read a line of text from `stream' into `lbp', excluding the
6321 * newline or CR-NL, if any. Return the number of characters read from
6322 * `stream', which is the length of the line including the newline.
6324 * On DOS or Windows we do not count the CR character, if any before the
6325 * NL, in the returned length; this mirrors the behavior of Emacs on those
6326 * platforms (for text files, it translates CR-NL to NL as it reads in the
6329 * If multi-line regular expressions are requested, each line read is
6330 * appended to `filebuf'.
6333 readline_internal (linebuffer
*lbp
, FILE *stream
, char const *filename
)
6335 char *buffer
= lbp
->buffer
;
6336 char *p
= lbp
->buffer
;
6340 pend
= p
+ lbp
->size
; /* Separate to avoid 386/IX compiler bug. */
6344 register int c
= getc (stream
);
6347 /* We're at the end of linebuffer: expand it. */
6349 xrnew (buffer
, lbp
->size
, char);
6350 p
+= buffer
- lbp
->buffer
;
6351 pend
= buffer
+ lbp
->size
;
6352 lbp
->buffer
= buffer
;
6356 if (ferror (stream
))
6364 if (p
> buffer
&& p
[-1] == '\r')
6378 lbp
->len
= p
- buffer
;
6380 if (need_filebuf
/* we need filebuf for multi-line regexps */
6381 && chars_deleted
> 0) /* not at EOF */
6383 while (filebuf
.size
<= filebuf
.len
+ lbp
->len
+ 1) /* +1 for \n */
6385 /* Expand filebuf. */
6387 xrnew (filebuf
.buffer
, filebuf
.size
, char);
6389 memcpy (filebuf
.buffer
+ filebuf
.len
, lbp
->buffer
, lbp
->len
);
6390 filebuf
.len
+= lbp
->len
;
6391 filebuf
.buffer
[filebuf
.len
++] = '\n';
6392 filebuf
.buffer
[filebuf
.len
] = '\0';
6395 return lbp
->len
+ chars_deleted
;
6399 * Like readline_internal, above, but in addition try to match the
6400 * input line against relevant regular expressions and manage #line
6404 readline (linebuffer
*lbp
, FILE *stream
)
6408 linecharno
= charno
; /* update global char number of line start */
6409 result
= readline_internal (lbp
, stream
, infilename
); /* read line */
6410 lineno
+= 1; /* increment global line number */
6411 charno
+= result
; /* increment global char number */
6413 /* Honor #line directives. */
6414 if (!no_line_directive
)
6416 static bool discard_until_line_directive
;
6418 /* Check whether this is a #line directive. */
6419 if (result
> 12 && strneq (lbp
->buffer
, "#line ", 6))
6424 if (sscanf (lbp
->buffer
, "#line %u \"%n", &lno
, &start
) >= 1
6425 && start
> 0) /* double quote character found */
6427 char *endp
= lbp
->buffer
+ start
;
6429 while ((endp
= strchr (endp
, '"')) != NULL
6430 && endp
[-1] == '\\')
6433 /* Ok, this is a real #line directive. Let's deal with it. */
6435 char *taggedabsname
; /* absolute name of original file */
6436 char *taggedfname
; /* name of original file as given */
6437 char *name
; /* temp var */
6439 discard_until_line_directive
= false; /* found it */
6440 name
= lbp
->buffer
+ start
;
6442 canonicalize_filename (name
);
6443 taggedabsname
= absolute_filename (name
, tagfiledir
);
6444 if (filename_is_absolute (name
)
6445 || filename_is_absolute (curfdp
->infname
))
6446 taggedfname
= savestr (taggedabsname
);
6448 taggedfname
= relative_filename (taggedabsname
,tagfiledir
);
6450 if (streq (curfdp
->taggedfname
, taggedfname
))
6451 /* The #line directive is only a line number change. We
6452 deal with this afterwards. */
6455 /* The tags following this #line directive should be
6456 attributed to taggedfname. In order to do this, set
6457 curfdp accordingly. */
6459 fdesc
*fdp
; /* file description pointer */
6461 /* Go look for a file description already set up for the
6462 file indicated in the #line directive. If there is
6463 one, use it from now until the next #line
6465 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
6466 if (streq (fdp
->infname
, curfdp
->infname
)
6467 && streq (fdp
->taggedfname
, taggedfname
))
6468 /* If we remove the second test above (after the &&)
6469 then all entries pertaining to the same file are
6470 coalesced in the tags file. If we use it, then
6471 entries pertaining to the same file but generated
6472 from different files (via #line directives) will
6473 go into separate sections in the tags file. These
6474 alternatives look equivalent. The first one
6475 destroys some apparently useless information. */
6481 /* Else, if we already tagged the real file, skip all
6482 input lines until the next #line directive. */
6483 if (fdp
== NULL
) /* not found */
6484 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
6485 if (streq (fdp
->infabsname
, taggedabsname
))
6487 discard_until_line_directive
= true;
6491 /* Else create a new file description and use that from
6492 now on, until the next #line directive. */
6493 if (fdp
== NULL
) /* not found */
6496 fdhead
= xnew (1, fdesc
);
6497 *fdhead
= *curfdp
; /* copy curr. file description */
6499 fdhead
->infname
= savestr (curfdp
->infname
);
6500 fdhead
->infabsname
= savestr (curfdp
->infabsname
);
6501 fdhead
->infabsdir
= savestr (curfdp
->infabsdir
);
6502 fdhead
->taggedfname
= taggedfname
;
6503 fdhead
->usecharno
= false;
6504 fdhead
->prop
= NULL
;
6505 fdhead
->written
= false;
6509 free (taggedabsname
);
6511 readline (lbp
, stream
);
6513 } /* if a real #line directive */
6514 } /* if #line is followed by a number */
6515 } /* if line begins with "#line " */
6517 /* If we are here, no #line directive was found. */
6518 if (discard_until_line_directive
)
6522 /* Do a tail recursion on ourselves, thus discarding the contents
6523 of the line buffer. */
6524 readline (lbp
, stream
);
6528 discard_until_line_directive
= false;
6531 } /* if #line directives should be considered */
6538 /* Match against relevant regexps. */
6540 for (rp
= p_head
; rp
!= NULL
; rp
= rp
->p_next
)
6542 /* Only use generic regexps or those for the current language.
6543 Also do not use multiline regexps, which is the job of
6544 regex_tag_multiline. */
6545 if ((rp
->lang
!= NULL
&& rp
->lang
!= fdhead
->lang
)
6549 match
= re_match (rp
->pat
, lbp
->buffer
, lbp
->len
, 0, &rp
->regs
);
6554 if (!rp
->error_signaled
)
6556 error ("regexp stack overflow while matching \"%s\"",
6558 rp
->error_signaled
= true;
6565 /* Empty string matched. */
6566 if (!rp
->error_signaled
)
6568 error ("regexp matches the empty string: \"%s\"", rp
->pattern
);
6569 rp
->error_signaled
= true;
6573 /* Match occurred. Construct a tag. */
6575 if (name
[0] == '\0')
6577 else /* make a named tag */
6578 name
= substitute (lbp
->buffer
, rp
->name
, &rp
->regs
);
6579 if (rp
->force_explicit_name
)
6580 /* Force explicit tag name, if a name is there. */
6581 pfnote (name
, true, lbp
->buffer
, match
, lineno
, linecharno
);
6583 make_tag (name
, strlen (name
), true,
6584 lbp
->buffer
, match
, lineno
, linecharno
);
6593 * Return a pointer to a space of size strlen(cp)+1 allocated
6594 * with xnew where the string CP has been copied.
6597 savestr (const char *cp
)
6599 return savenstr (cp
, strlen (cp
));
6603 * Return a pointer to a space of size LEN+1 allocated with xnew where
6604 * the string CP has been copied for at most the first LEN characters.
6607 savenstr (const char *cp
, int len
)
6609 char *dp
= xnew (len
+ 1, char);
6611 return memcpy (dp
, cp
, len
);
6614 /* Skip spaces (end of string is not space), return new pointer. */
6616 skip_spaces (char *cp
)
6618 while (c_isspace (*cp
))
6623 /* Skip non spaces, except end of string, return new pointer. */
6625 skip_non_spaces (char *cp
)
6627 while (*cp
!= '\0' && !c_isspace (*cp
))
6632 /* Skip any chars in the "name" class.*/
6634 skip_name (char *cp
)
6636 /* '\0' is a notinname() so loop stops there too */
6637 while (! notinname (*cp
))
6642 /* Print error message and exit. */
6644 fatal (char const *format
, ...)
6647 va_start (ap
, format
);
6648 verror (format
, ap
);
6650 exit (EXIT_FAILURE
);
6654 pfatal (const char *s1
)
6657 exit (EXIT_FAILURE
);
6661 suggest_asking_for_help (void)
6663 fprintf (stderr
, "\tTry '%s --help' for a complete list of options.\n",
6665 exit (EXIT_FAILURE
);
6668 /* Output a diagnostic with printf-style FORMAT and args. */
6670 error (const char *format
, ...)
6673 va_start (ap
, format
);
6674 verror (format
, ap
);
6679 verror (char const *format
, va_list ap
)
6681 fprintf (stderr
, "%s: ", progname
);
6682 vfprintf (stderr
, format
, ap
);
6683 fprintf (stderr
, "\n");
6686 /* Return a newly-allocated string whose contents
6687 concatenate those of s1, s2, s3. */
6689 concat (const char *s1
, const char *s2
, const char *s3
)
6691 int len1
= strlen (s1
), len2
= strlen (s2
), len3
= strlen (s3
);
6692 char *result
= xnew (len1
+ len2
+ len3
+ 1, char);
6694 strcpy (result
, s1
);
6695 strcpy (result
+ len1
, s2
);
6696 strcpy (result
+ len1
+ len2
, s3
);
6702 /* Does the same work as the system V getcwd, but does not need to
6703 guess the buffer size in advance. */
6708 char *path
= xnew (bufsize
, char);
6710 while (getcwd (path
, bufsize
) == NULL
)
6712 if (errno
!= ERANGE
)
6716 path
= xnew (bufsize
, char);
6719 canonicalize_filename (path
);
6723 /* Return a newly allocated string containing a name of a temporary file. */
6727 const char *tmpdir
= getenv ("TMPDIR");
6728 const char *slash
= "/";
6730 #if MSDOS || defined (DOS_NT)
6732 tmpdir
= getenv ("TEMP");
6734 tmpdir
= getenv ("TMP");
6737 if (tmpdir
[strlen (tmpdir
) - 1] == '/'
6738 || tmpdir
[strlen (tmpdir
) - 1] == '\\')
6743 if (tmpdir
[strlen (tmpdir
) - 1] == '/')
6747 char *templt
= concat (tmpdir
, slash
, "etXXXXXX");
6748 int fd
= mkostemp (templt
, O_CLOEXEC
);
6749 if (fd
< 0 || close (fd
) != 0)
6751 int temp_errno
= errno
;
6757 #if defined (DOS_NT)
6758 /* The file name will be used in shell redirection, so it needs to have
6759 DOS-style backslashes, or else the Windows shell will barf. */
6761 for (p
= templt
; *p
; p
++)
6769 /* Return a newly allocated string containing the file name of FILE
6770 relative to the absolute directory DIR (which should end with a slash). */
6772 relative_filename (char *file
, char *dir
)
6774 char *fp
, *dp
, *afn
, *res
;
6777 /* Find the common root of file and dir (with a trailing slash). */
6778 afn
= absolute_filename (file
, cwd
);
6781 while (*fp
++ == *dp
++)
6783 fp
--, dp
--; /* back to the first differing char */
6785 if (fp
== afn
&& afn
[0] != '/') /* cannot build a relative name */
6788 do /* look at the equal chars until '/' */
6792 /* Build a sequence of "../" strings for the resulting relative file name. */
6794 while ((dp
= strchr (dp
+ 1, '/')) != NULL
)
6796 res
= xnew (3*i
+ strlen (fp
+ 1) + 1, char);
6799 z
= stpcpy (z
, "../");
6801 /* Add the file name relative to the common root of file and dir. */
6808 /* Return a newly allocated string containing the absolute file name
6809 of FILE given DIR (which should end with a slash). */
6811 absolute_filename (char *file
, char *dir
)
6813 char *slashp
, *cp
, *res
;
6815 if (filename_is_absolute (file
))
6816 res
= savestr (file
);
6818 /* We don't support non-absolute file names with a drive
6819 letter, like `d:NAME' (it's too much hassle). */
6820 else if (file
[1] == ':')
6821 fatal ("%s: relative file names with drive letters not supported", file
);
6824 res
= concat (dir
, file
, "");
6826 /* Delete the "/dirname/.." and "/." substrings. */
6827 slashp
= strchr (res
, '/');
6828 while (slashp
!= NULL
&& slashp
[0] != '\0')
6830 if (slashp
[1] == '.')
6832 if (slashp
[2] == '.'
6833 && (slashp
[3] == '/' || slashp
[3] == '\0'))
6838 while (cp
>= res
&& !filename_is_absolute (cp
));
6840 cp
= slashp
; /* the absolute name begins with "/.." */
6842 /* Under MSDOS and NT we get `d:/NAME' as absolute
6843 file name, so the luser could say `d:/../NAME'.
6844 We silently treat this as `d:/NAME'. */
6845 else if (cp
[0] != '/')
6848 memmove (cp
, slashp
+ 3, strlen (slashp
+ 2));
6852 else if (slashp
[2] == '/' || slashp
[2] == '\0')
6854 memmove (slashp
, slashp
+ 2, strlen (slashp
+ 1));
6859 slashp
= strchr (slashp
+ 1, '/');
6862 if (res
[0] == '\0') /* just a safety net: should never happen */
6865 return savestr ("/");
6871 /* Return a newly allocated string containing the absolute
6872 file name of dir where FILE resides given DIR (which should
6873 end with a slash). */
6875 absolute_dirname (char *file
, char *dir
)
6880 slashp
= strrchr (file
, '/');
6882 return savestr (dir
);
6885 res
= absolute_filename (file
, dir
);
6891 /* Whether the argument string is an absolute file name. The argument
6892 string must have been canonicalized with canonicalize_filename. */
6894 filename_is_absolute (char *fn
)
6896 return (fn
[0] == '/'
6898 || (c_isalpha (fn
[0]) && fn
[1] == ':' && fn
[2] == '/')
6903 /* Downcase DOS drive letter and collapse separators into single slashes.
6906 canonicalize_filename (register char *fn
)
6911 /* Canonicalize drive letter case. */
6912 if (c_isupper (fn
[0]) && fn
[1] == ':')
6913 fn
[0] = c_tolower (fn
[0]);
6915 /* Collapse multiple forward- and back-slashes into a single forward
6917 for (cp
= fn
; *cp
!= '\0'; cp
++, fn
++)
6918 if (*cp
== '/' || *cp
== '\\')
6921 while (cp
[1] == '/' || cp
[1] == '\\')
6929 /* Collapse multiple slashes into a single slash. */
6930 for (cp
= fn
; *cp
!= '\0'; cp
++, fn
++)
6934 while (cp
[1] == '/')
6940 #endif /* !DOS_NT */
6946 /* Initialize a linebuffer for use. */
6948 linebuffer_init (linebuffer
*lbp
)
6950 lbp
->size
= (DEBUG
) ? 3 : 200;
6951 lbp
->buffer
= xnew (lbp
->size
, char);
6952 lbp
->buffer
[0] = '\0';
6956 /* Set the minimum size of a string contained in a linebuffer. */
6958 linebuffer_setlen (linebuffer
*lbp
, int toksize
)
6960 while (lbp
->size
<= toksize
)
6963 xrnew (lbp
->buffer
, lbp
->size
, char);
6968 /* Like malloc but get fatal error if memory is exhausted. */
6970 xmalloc (size_t size
)
6972 void *result
= malloc (size
);
6974 fatal ("virtual memory exhausted");
6979 xrealloc (void *ptr
, size_t size
)
6981 void *result
= realloc (ptr
, size
);
6983 fatal ("virtual memory exhausted");
6989 * indent-tabs-mode: t
6992 * c-font-lock-extra-types: ("FILE" "bool" "language" "linebuffer" "fdesc" "node" "regexp")
6993 * c-file-style: "gnu"
6997 /* etags.c ends here */