1 /* Tags file maker to go with GNU Emacs -*- coding: utf-8 -*-
3 Copyright (C) 1984 The Regents of the University of California
5 Redistribution and use in source and binary forms, with or without
6 modification, are permitted provided that the following conditions are
8 1. Redistributions of source code must retain the above copyright
9 notice, this list of conditions and the following disclaimer.
10 2. Redistributions in binary form must reproduce the above copyright
11 notice, this list of conditions and the following disclaimer in the
12 documentation and/or other materials provided with the
14 3. Neither the name of the University nor the names of its
15 contributors may be used to endorse or promote products derived
16 from this software without specific prior written permission.
18 THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS''
19 AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO,
20 THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
21 PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS
22 BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
23 CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
24 SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
25 BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
26 WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE
27 OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN
28 IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
31 Copyright (C) 1984, 1987-1989, 1993-1995, 1998-2016 Free Software
34 This file is not considered part of GNU Emacs.
36 This program is free software: you can redistribute it and/or modify
37 it under the terms of the GNU General Public License as published by
38 the Free Software Foundation, either version 3 of the License, or
39 (at your option) any later version.
41 This program is distributed in the hope that it will be useful,
42 but WITHOUT ANY WARRANTY; without even the implied warranty of
43 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
44 GNU General Public License for more details.
46 You should have received a copy of the GNU General Public License
47 along with this program. If not, see <http://www.gnu.org/licenses/>. */
50 /* NB To comply with the above BSD license, copyright information is
51 reproduced in etc/ETAGS.README. That file should be updated when the
54 To the best of our knowledge, this code was originally based on the
55 ctags.c distributed with BSD4.2, which was copyrighted by the
56 University of California, as described above. */
61 * 1983 Ctags originally by Ken Arnold.
62 * 1984 Fortran added by Jim Kleckner.
63 * 1984 Ed Pelegri-Llopart added C typedefs.
64 * 1985 Emacs TAGS format by Richard Stallman.
65 * 1989 Sam Kendall added C++.
66 * 1992 Joseph B. Wells improved C and C++ parsing.
67 * 1993 Francesco Potortì reorganized C and C++.
68 * 1994 Line-by-line regexp tags by Tom Tromey.
69 * 2001 Nested classes by Francesco Potortì (concept by Mykola Dzyuba).
70 * 2002 #line directives by Francesco Potortì.
71 * Francesco Potortì maintained and improved it for many years
76 * If you want to add support for a new language, start by looking at the LUA
77 * language, which is the simplest. Alternatively, consider distributing etags
78 * together with a configuration file containing regexp definitions for etags.
81 char pot_etags_version
[] = "@(#) pot revision number is 17.38.1.4";
88 # define NDEBUG /* disable assert */
94 # define _GNU_SOURCE 1 /* enables some compiler checks on GNU */
97 /* WIN32_NATIVE is for XEmacs.
98 MSDOS, WINDOWSNT, DOS_NT are for Emacs. */
103 #endif /* WIN32_NATIVE */
108 # include <sys/param.h>
115 # define MAXPATHLEN _MAX_PATH
119 # define O_CLOEXEC O_NOINHERIT
120 #endif /* WINDOWSNT */
127 #include <sysstdio.h>
130 #include <binary-io.h>
132 #include <c-strcase.h>
136 # undef assert /* some systems have a buggy assert.h */
137 # define assert(x) ((void) 0)
143 /* Define CTAGS to make the program "ctags" compatible with the usual one.
144 Leave it undefined to make the program "etags", which makes emacs-style
145 tag tables and tags typedefs, #defines and struct/union/enum by default. */
154 streq (char const *s
, char const *t
)
156 return strcmp (s
, t
) == 0;
160 strcaseeq (char const *s
, char const *t
)
162 return c_strcasecmp (s
, t
) == 0;
166 strneq (char const *s
, char const *t
, size_t n
)
168 return strncmp (s
, t
, n
) == 0;
172 strncaseeq (char const *s
, char const *t
, size_t n
)
174 return c_strncasecmp (s
, t
, n
) == 0;
177 /* C is not in a name. */
179 notinname (unsigned char c
)
181 /* Look at make_tag before modifying! */
182 static bool const table
[UCHAR_MAX
+ 1] = {
183 ['\0']=1, ['\t']=1, ['\n']=1, ['\f']=1, ['\r']=1, [' ']=1,
184 ['(']=1, [')']=1, [',']=1, [';']=1, ['=']=1
189 /* C can start a token. */
191 begtoken (unsigned char c
)
193 static bool const table
[UCHAR_MAX
+ 1] = {
195 ['A']=1, ['B']=1, ['C']=1, ['D']=1, ['E']=1, ['F']=1, ['G']=1, ['H']=1,
196 ['I']=1, ['J']=1, ['K']=1, ['L']=1, ['M']=1, ['N']=1, ['O']=1, ['P']=1,
197 ['Q']=1, ['R']=1, ['S']=1, ['T']=1, ['U']=1, ['V']=1, ['W']=1, ['X']=1,
200 ['a']=1, ['b']=1, ['c']=1, ['d']=1, ['e']=1, ['f']=1, ['g']=1, ['h']=1,
201 ['i']=1, ['j']=1, ['k']=1, ['l']=1, ['m']=1, ['n']=1, ['o']=1, ['p']=1,
202 ['q']=1, ['r']=1, ['s']=1, ['t']=1, ['u']=1, ['v']=1, ['w']=1, ['x']=1,
209 /* C can be in the middle of a token. */
211 intoken (unsigned char c
)
213 static bool const table
[UCHAR_MAX
+ 1] = {
215 ['0']=1, ['1']=1, ['2']=1, ['3']=1, ['4']=1,
216 ['5']=1, ['6']=1, ['7']=1, ['8']=1, ['9']=1,
217 ['A']=1, ['B']=1, ['C']=1, ['D']=1, ['E']=1, ['F']=1, ['G']=1, ['H']=1,
218 ['I']=1, ['J']=1, ['K']=1, ['L']=1, ['M']=1, ['N']=1, ['O']=1, ['P']=1,
219 ['Q']=1, ['R']=1, ['S']=1, ['T']=1, ['U']=1, ['V']=1, ['W']=1, ['X']=1,
222 ['a']=1, ['b']=1, ['c']=1, ['d']=1, ['e']=1, ['f']=1, ['g']=1, ['h']=1,
223 ['i']=1, ['j']=1, ['k']=1, ['l']=1, ['m']=1, ['n']=1, ['o']=1, ['p']=1,
224 ['q']=1, ['r']=1, ['s']=1, ['t']=1, ['u']=1, ['v']=1, ['w']=1, ['x']=1,
230 /* C can end a token. */
232 endtoken (unsigned char c
)
234 static bool const table
[UCHAR_MAX
+ 1] = {
235 ['\0']=1, ['\t']=1, ['\n']=1, ['\r']=1, [' ']=1,
236 ['!']=1, ['"']=1, ['#']=1, ['%']=1, ['&']=1, ['\'']=1, ['(']=1, [')']=1,
237 ['*']=1, ['+']=1, [',']=1, ['-']=1, ['.']=1, ['/']=1, [':']=1, [';']=1,
238 ['<']=1, ['=']=1, ['>']=1, ['?']=1, ['[']=1, [']']=1, ['^']=1,
239 ['{']=1, ['|']=1, ['}']=1, ['~']=1
245 * xnew, xrnew -- allocate, reallocate storage
247 * SYNOPSIS: Type *xnew (int n, Type);
248 * void xrnew (OldPointer, int n, Type);
250 #define xnew(n, Type) ((Type *) xmalloc ((n) * sizeof (Type)))
251 #define xrnew(op, n, Type) ((op) = (Type *) xrealloc (op, (n) * sizeof (Type)))
253 typedef void Lang_function (FILE *);
257 const char *suffix
; /* file name suffix for this compressor */
258 const char *command
; /* takes one arg and decompresses to stdout */
263 const char *name
; /* language name */
264 const char *help
; /* detailed help for the language */
265 Lang_function
*function
; /* parse function */
266 const char **suffixes
; /* name suffixes of this language's files */
267 const char **filenames
; /* names of this language's files */
268 const char **interpreters
; /* interpreters for this language */
269 bool metasource
; /* source used to generate other sources */
274 struct fdesc
*next
; /* for the linked list */
275 char *infname
; /* uncompressed input file name */
276 char *infabsname
; /* absolute uncompressed input file name */
277 char *infabsdir
; /* absolute dir of input file */
278 char *taggedfname
; /* file name to write in tagfile */
279 language
*lang
; /* language of file */
280 char *prop
; /* file properties to write in tagfile */
281 bool usecharno
; /* etags tags shall contain char number */
282 bool written
; /* entry written in the tags file */
285 typedef struct node_st
286 { /* sorting structure */
287 struct node_st
*left
, *right
; /* left and right sons */
288 fdesc
*fdp
; /* description of file to whom tag belongs */
289 char *name
; /* tag name */
290 char *regex
; /* search regexp */
291 bool valid
; /* write this tag on the tag file */
292 bool is_func
; /* function tag: use regexp in CTAGS mode */
293 bool been_warned
; /* warning already given for duplicated tag */
294 int lno
; /* line number tag is on */
295 long cno
; /* character number line starts on */
299 * A `linebuffer' is a structure which holds a line of text.
300 * `readline_internal' reads a line from a stream into a linebuffer
301 * and works regardless of the length of the line.
302 * SIZE is the size of BUFFER, LEN is the length of the string in
303 * BUFFER after readline reads it.
312 /* Used to support mixing of --lang and file names. */
316 at_language
, /* a language specification */
317 at_regexp
, /* a regular expression */
318 at_filename
, /* a file name */
319 at_stdin
, /* read from stdin here */
320 at_end
/* stop parsing the list */
321 } arg_type
; /* argument type */
322 language
*lang
; /* language associated with the argument */
323 char *what
; /* the argument itself */
326 /* Structure defining a regular expression. */
327 typedef struct regexp
329 struct regexp
*p_next
; /* pointer to next in list */
330 language
*lang
; /* if set, use only for this language */
331 char *pattern
; /* the regexp pattern */
332 char *name
; /* tag name */
333 struct re_pattern_buffer
*pat
; /* the compiled pattern */
334 struct re_registers regs
; /* re registers */
335 bool error_signaled
; /* already signaled for this regexp */
336 bool force_explicit_name
; /* do not allow implicit tag name */
337 bool ignore_case
; /* ignore case when matching */
338 bool multi_line
; /* do a multi-line match on the whole file */
342 /* Many compilers barf on this:
343 Lang_function Ada_funcs;
344 so let's write it this way */
345 static void Ada_funcs (FILE *);
346 static void Asm_labels (FILE *);
347 static void C_entries (int c_ext
, FILE *);
348 static void default_C_entries (FILE *);
349 static void plain_C_entries (FILE *);
350 static void Cjava_entries (FILE *);
351 static void Cobol_paragraphs (FILE *);
352 static void Cplusplus_entries (FILE *);
353 static void Cstar_entries (FILE *);
354 static void Erlang_functions (FILE *);
355 static void Forth_words (FILE *);
356 static void Fortran_functions (FILE *);
357 static void Go_functions (FILE *);
358 static void HTML_labels (FILE *);
359 static void Lisp_functions (FILE *);
360 static void Lua_functions (FILE *);
361 static void Makefile_targets (FILE *);
362 static void Pascal_functions (FILE *);
363 static void Perl_functions (FILE *);
364 static void PHP_functions (FILE *);
365 static void PS_functions (FILE *);
366 static void Prolog_functions (FILE *);
367 static void Python_functions (FILE *);
368 static void Ruby_functions (FILE *);
369 static void Scheme_functions (FILE *);
370 static void TeX_commands (FILE *);
371 static void Texinfo_nodes (FILE *);
372 static void Yacc_entries (FILE *);
373 static void just_read_file (FILE *);
375 static language
*get_language_from_langname (const char *);
376 static void readline (linebuffer
*, FILE *);
377 static long readline_internal (linebuffer
*, FILE *, char const *);
378 static bool nocase_tail (const char *);
379 static void get_tag (char *, char **);
381 static void analyze_regex (char *);
382 static void free_regexps (void);
383 static void regex_tag_multiline (void);
384 static void error (const char *, ...) ATTRIBUTE_FORMAT_PRINTF (1, 2);
385 static void verror (char const *, va_list) ATTRIBUTE_FORMAT_PRINTF (1, 0);
386 static _Noreturn
void suggest_asking_for_help (void);
387 static _Noreturn
void fatal (char const *, ...) ATTRIBUTE_FORMAT_PRINTF (1, 2);
388 static _Noreturn
void pfatal (const char *);
389 static void add_node (node
*, node
**);
391 static void process_file_name (char *, language
*);
392 static void process_file (FILE *, char *, language
*);
393 static void find_entries (FILE *);
394 static void free_tree (node
*);
395 static void free_fdesc (fdesc
*);
396 static void pfnote (char *, bool, char *, int, int, long);
397 static void invalidate_nodes (fdesc
*, node
**);
398 static void put_entries (node
*);
400 static char *concat (const char *, const char *, const char *);
401 static char *skip_spaces (char *);
402 static char *skip_non_spaces (char *);
403 static char *skip_name (char *);
404 static char *savenstr (const char *, int);
405 static char *savestr (const char *);
406 static char *etags_getcwd (void);
407 static char *relative_filename (char *, char *);
408 static char *absolute_filename (char *, char *);
409 static char *absolute_dirname (char *, char *);
410 static bool filename_is_absolute (char *f
);
411 static void canonicalize_filename (char *);
412 static char *etags_mktmp (void);
413 static void linebuffer_init (linebuffer
*);
414 static void linebuffer_setlen (linebuffer
*, int);
415 static void *xmalloc (size_t);
416 static void *xrealloc (void *, size_t);
419 static char searchar
= '/'; /* use /.../ searches */
421 static char *tagfile
; /* output file */
422 static char *progname
; /* name this program was invoked with */
423 static char *cwd
; /* current working directory */
424 static char *tagfiledir
; /* directory of tagfile */
425 static FILE *tagf
; /* ioptr for tags file */
426 static ptrdiff_t whatlen_max
; /* maximum length of any 'what' member */
428 static fdesc
*fdhead
; /* head of file description list */
429 static fdesc
*curfdp
; /* current file description */
430 static char *infilename
; /* current input file name */
431 static int lineno
; /* line number of current line */
432 static long charno
; /* current character number */
433 static long linecharno
; /* charno of start of current line */
434 static char *dbp
; /* pointer to start of current tag */
436 static const int invalidcharno
= -1;
438 static node
*nodehead
; /* the head of the binary tree of tags */
439 static node
*last_node
; /* the last node created */
441 static linebuffer lb
; /* the current line */
442 static linebuffer filebuf
; /* a buffer containing the whole file */
443 static linebuffer token_name
; /* a buffer containing a tag name */
445 static bool append_to_tagfile
; /* -a: append to tags */
446 /* The next five default to true in C and derived languages. */
447 static bool typedefs
; /* -t: create tags for C and Ada typedefs */
448 static bool typedefs_or_cplusplus
; /* -T: create tags for C typedefs, level */
449 /* 0 struct/enum/union decls, and C++ */
450 /* member functions. */
451 static bool constantypedefs
; /* -d: create tags for C #define, enum */
452 /* constants and variables. */
453 /* -D: opposite of -d. Default under ctags. */
454 static int globals
; /* create tags for global variables */
455 static int members
; /* create tags for C member variables */
456 static int declarations
; /* --declarations: tag them and extern in C&Co*/
457 static int no_line_directive
; /* ignore #line directives (undocumented) */
458 static int no_duplicates
; /* no duplicate tags for ctags (undocumented) */
459 static bool update
; /* -u: update tags */
460 static bool vgrind_style
; /* -v: create vgrind style index output */
461 static bool no_warnings
; /* -w: suppress warnings (undocumented) */
462 static bool cxref_style
; /* -x: create cxref style output */
463 static bool cplusplus
; /* .[hc] means C++, not C (undocumented) */
464 static bool ignoreindent
; /* -I: ignore indentation in C */
465 static int packages_only
; /* --packages-only: in Ada, only tag packages*/
466 static int class_qualify
; /* -Q: produce class-qualified tags in C++/Java */
468 /* STDIN is defined in LynxOS system headers */
473 #define STDIN 0x1001 /* returned by getopt_long on --parse-stdin */
474 static bool parsing_stdin
; /* --parse-stdin used */
476 static regexp
*p_head
; /* list of all regexps */
477 static bool need_filebuf
; /* some regexes are multi-line */
479 static struct option longopts
[] =
481 { "append", no_argument
, NULL
, 'a' },
482 { "packages-only", no_argument
, &packages_only
, 1 },
483 { "c++", no_argument
, NULL
, 'C' },
484 { "declarations", no_argument
, &declarations
, 1 },
485 { "no-line-directive", no_argument
, &no_line_directive
, 1 },
486 { "no-duplicates", no_argument
, &no_duplicates
, 1 },
487 { "help", no_argument
, NULL
, 'h' },
488 { "help", no_argument
, NULL
, 'H' },
489 { "ignore-indentation", no_argument
, NULL
, 'I' },
490 { "language", required_argument
, NULL
, 'l' },
491 { "members", no_argument
, &members
, 1 },
492 { "no-members", no_argument
, &members
, 0 },
493 { "output", required_argument
, NULL
, 'o' },
494 { "class-qualify", no_argument
, &class_qualify
, 'Q' },
495 { "regex", required_argument
, NULL
, 'r' },
496 { "no-regex", no_argument
, NULL
, 'R' },
497 { "ignore-case-regex", required_argument
, NULL
, 'c' },
498 { "parse-stdin", required_argument
, NULL
, STDIN
},
499 { "version", no_argument
, NULL
, 'V' },
501 #if CTAGS /* Ctags options */
502 { "backward-search", no_argument
, NULL
, 'B' },
503 { "cxref", no_argument
, NULL
, 'x' },
504 { "defines", no_argument
, NULL
, 'd' },
505 { "globals", no_argument
, &globals
, 1 },
506 { "typedefs", no_argument
, NULL
, 't' },
507 { "typedefs-and-c++", no_argument
, NULL
, 'T' },
508 { "update", no_argument
, NULL
, 'u' },
509 { "vgrind", no_argument
, NULL
, 'v' },
510 { "no-warn", no_argument
, NULL
, 'w' },
512 #else /* Etags options */
513 { "no-defines", no_argument
, NULL
, 'D' },
514 { "no-globals", no_argument
, &globals
, 0 },
515 { "include", required_argument
, NULL
, 'i' },
520 static compressor compressors
[] =
522 { "z", "gzip -d -c"},
523 { "Z", "gzip -d -c"},
524 { "gz", "gzip -d -c"},
525 { "GZ", "gzip -d -c"},
526 { "bz2", "bzip2 -d -c" },
527 { "xz", "xz -d -c" },
536 static const char *Ada_suffixes
[] =
537 { "ads", "adb", "ada", NULL
};
538 static const char Ada_help
[] =
539 "In Ada code, functions, procedures, packages, tasks and types are\n\
540 tags. Use the '--packages-only' option to create tags for\n\
542 Ada tag names have suffixes indicating the type of entity:\n\
543 Entity type: Qualifier:\n\
544 ------------ ----------\n\
551 Thus, 'M-x find-tag <RET> bidule/b <RET>' will go directly to the\n\
552 body of the package 'bidule', while 'M-x find-tag <RET> bidule <RET>'\n\
553 will just search for any tag 'bidule'.";
556 static const char *Asm_suffixes
[] =
557 { "a", /* Unix assembler */
558 "asm", /* Microcontroller assembly */
559 "def", /* BSO/Tasking definition includes */
560 "inc", /* Microcontroller include files */
561 "ins", /* Microcontroller include files */
562 "s", "sa", /* Unix assembler */
563 "S", /* cpp-processed Unix assembler */
564 "src", /* BSO/Tasking C compiler output */
567 static const char Asm_help
[] =
568 "In assembler code, labels appearing at the beginning of a line,\n\
569 followed by a colon, are tags.";
572 /* Note that .c and .h can be considered C++, if the --c++ flag was
573 given, or if the `class' or `template' keywords are met inside the file.
574 That is why default_C_entries is called for these. */
575 static const char *default_C_suffixes
[] =
577 #if CTAGS /* C help for Ctags */
578 static const char default_C_help
[] =
579 "In C code, any C function is a tag. Use -t to tag typedefs.\n\
580 Use -T to tag definitions of 'struct', 'union' and 'enum'.\n\
581 Use -d to tag '#define' macro definitions and 'enum' constants.\n\
582 Use --globals to tag global variables.\n\
583 You can tag function declarations and external variables by\n\
584 using '--declarations', and struct members by using '--members'.";
585 #else /* C help for Etags */
586 static const char default_C_help
[] =
587 "In C code, any C function or typedef is a tag, and so are\n\
588 definitions of 'struct', 'union' and 'enum'. '#define' macro\n\
589 definitions and 'enum' constants are tags unless you specify\n\
590 '--no-defines'. Global variables are tags unless you specify\n\
591 '--no-globals' and so are struct members unless you specify\n\
592 '--no-members'. Use of '--no-globals', '--no-defines' and\n\
593 '--no-members' can make the tags table file much smaller.\n\
594 You can tag function declarations and external variables by\n\
595 using '--declarations'.";
596 #endif /* C help for Ctags and Etags */
598 static const char *Cplusplus_suffixes
[] =
599 { "C", "c++", "cc", "cpp", "cxx", "H", "h++", "hh", "hpp", "hxx",
600 "M", /* Objective C++ */
601 "pdb", /* PostScript with C syntax */
603 static const char Cplusplus_help
[] =
604 "In C++ code, all the tag constructs of C code are tagged. (Use\n\
605 --help --lang=c --lang=c++ for full help.)\n\
606 In addition to C tags, member functions are also recognized. Member\n\
607 variables are recognized unless you use the '--no-members' option.\n\
608 Tags for variables and functions in classes are named 'CLASS::VARIABLE'\n\
609 and 'CLASS::FUNCTION'. 'operator' definitions have tag names like\n\
612 static const char *Cjava_suffixes
[] =
614 static char Cjava_help
[] =
615 "In Java code, all the tags constructs of C and C++ code are\n\
616 tagged. (Use --help --lang=c --lang=c++ --lang=java for full help.)";
619 static const char *Cobol_suffixes
[] =
620 { "COB", "cob", NULL
};
621 static char Cobol_help
[] =
622 "In Cobol code, tags are paragraph names; that is, any word\n\
623 starting in column 8 and followed by a period.";
625 static const char *Cstar_suffixes
[] =
626 { "cs", "hs", NULL
};
628 static const char *Erlang_suffixes
[] =
629 { "erl", "hrl", NULL
};
630 static const char Erlang_help
[] =
631 "In Erlang code, the tags are the functions, records and macros\n\
632 defined in the file.";
634 const char *Forth_suffixes
[] =
635 { "fth", "tok", NULL
};
636 static const char Forth_help
[] =
637 "In Forth code, tags are words defined by ':',\n\
638 constant, code, create, defer, value, variable, buffer:, field.";
640 static const char *Fortran_suffixes
[] =
641 { "F", "f", "f90", "for", NULL
};
642 static const char Fortran_help
[] =
643 "In Fortran code, functions, subroutines and block data are tags.";
645 static const char *Go_suffixes
[] = {"go", NULL
};
646 static const char Go_help
[] =
647 "In Go code, functions, interfaces and packages are tags.";
649 static const char *HTML_suffixes
[] =
650 { "htm", "html", "shtml", NULL
};
651 static const char HTML_help
[] =
652 "In HTML input files, the tags are the 'title' and the 'h1', 'h2',\n\
653 'h3' headers. Also, tags are 'name=' in anchors and all\n\
654 occurrences of 'id='.";
656 static const char *Lisp_suffixes
[] =
657 { "cl", "clisp", "el", "l", "lisp", "LSP", "lsp", "ml", NULL
};
658 static const char Lisp_help
[] =
659 "In Lisp code, any function defined with 'defun', any variable\n\
660 defined with 'defvar' or 'defconst', and in general the first\n\
661 argument of any expression that starts with '(def' in column zero\n\
663 The '--declarations' option tags \"(defvar foo)\" constructs too.";
665 static const char *Lua_suffixes
[] =
666 { "lua", "LUA", NULL
};
667 static const char Lua_help
[] =
668 "In Lua scripts, all functions are tags.";
670 static const char *Makefile_filenames
[] =
671 { "Makefile", "makefile", "GNUMakefile", "Makefile.in", "Makefile.am", NULL
};
672 static const char Makefile_help
[] =
673 "In makefiles, targets are tags; additionally, variables are tags\n\
674 unless you specify '--no-globals'.";
676 static const char *Objc_suffixes
[] =
677 { "lm", /* Objective lex file */
678 "m", /* Objective C file */
680 static const char Objc_help
[] =
681 "In Objective C code, tags include Objective C definitions for classes,\n\
682 class categories, methods and protocols. Tags for variables and\n\
683 functions in classes are named 'CLASS::VARIABLE' and 'CLASS::FUNCTION'.\n\
684 (Use --help --lang=c --lang=objc --lang=java for full help.)";
686 static const char *Pascal_suffixes
[] =
687 { "p", "pas", NULL
};
688 static const char Pascal_help
[] =
689 "In Pascal code, the tags are the functions and procedures defined\n\
691 /* " // this is for working around an Emacs highlighting bug... */
693 static const char *Perl_suffixes
[] =
694 { "pl", "pm", NULL
};
695 static const char *Perl_interpreters
[] =
696 { "perl", "@PERL@", NULL
};
697 static const char Perl_help
[] =
698 "In Perl code, the tags are the packages, subroutines and variables\n\
699 defined by the 'package', 'sub', 'my' and 'local' keywords. Use\n\
700 '--globals' if you want to tag global variables. Tags for\n\
701 subroutines are named 'PACKAGE::SUB'. The name for subroutines\n\
702 defined in the default package is 'main::SUB'.";
704 static const char *PHP_suffixes
[] =
705 { "php", "php3", "php4", NULL
};
706 static const char PHP_help
[] =
707 "In PHP code, tags are functions, classes and defines. Unless you use\n\
708 the '--no-members' option, vars are tags too.";
710 static const char *plain_C_suffixes
[] =
711 { "pc", /* Pro*C file */
714 static const char *PS_suffixes
[] =
715 { "ps", "psw", NULL
}; /* .psw is for PSWrap */
716 static const char PS_help
[] =
717 "In PostScript code, the tags are the functions.";
719 static const char *Prolog_suffixes
[] =
721 static const char Prolog_help
[] =
722 "In Prolog code, tags are predicates and rules at the beginning of\n\
725 static const char *Python_suffixes
[] =
727 static const char Python_help
[] =
728 "In Python code, 'def' or 'class' at the beginning of a line\n\
731 static const char *Ruby_suffixes
[] =
732 { "rb", "ruby", NULL
};
733 static const char Ruby_help
[] =
734 "In Ruby code, 'def' or 'class' or 'module' at the beginning of\n\
735 a line generate a tag. Constants also generate a tag.";
737 /* Can't do the `SCM' or `scm' prefix with a version number. */
738 static const char *Scheme_suffixes
[] =
739 { "oak", "sch", "scheme", "SCM", "scm", "SM", "sm", "ss", "t", NULL
};
740 static const char Scheme_help
[] =
741 "In Scheme code, tags include anything defined with 'def' or with a\n\
742 construct whose name starts with 'def'. They also include\n\
743 variables set with 'set!' at top level in the file.";
745 static const char *TeX_suffixes
[] =
746 { "bib", "clo", "cls", "ltx", "sty", "TeX", "tex", NULL
};
747 static const char TeX_help
[] =
748 "In LaTeX text, the argument of any of the commands '\\chapter',\n\
749 '\\section', '\\subsection', '\\subsubsection', '\\eqno', '\\label',\n\
750 '\\ref', '\\cite', '\\bibitem', '\\part', '\\appendix', '\\entry',\n\
751 '\\index', '\\def', '\\newcommand', '\\renewcommand',\n\
752 '\\newenvironment' or '\\renewenvironment' is a tag.\n\
754 Other commands can be specified by setting the environment variable\n\
755 'TEXTAGS' to a colon-separated list like, for example,\n\
756 TEXTAGS=\"mycommand:myothercommand\".";
759 static const char *Texinfo_suffixes
[] =
760 { "texi", "texinfo", "txi", NULL
};
761 static const char Texinfo_help
[] =
762 "for texinfo files, lines starting with @node are tagged.";
764 static const char *Yacc_suffixes
[] =
765 { "y", "y++", "ym", "yxx", "yy", NULL
}; /* .ym is Objective yacc file */
766 static const char Yacc_help
[] =
767 "In Bison or Yacc input files, each rule defines as a tag the\n\
768 nonterminal it constructs. The portions of the file that contain\n\
769 C code are parsed as C code (use --help --lang=c --lang=yacc\n\
772 static const char auto_help
[] =
773 "'auto' is not a real language, it indicates to use\n\
774 a default language for files base on file name suffix and file contents.";
776 static const char none_help
[] =
777 "'none' is not a real language, it indicates to only do\n\
778 regexp processing on files.";
780 static const char no_lang_help
[] =
781 "No detailed help available for this language.";
785 * Table of languages.
787 * It is ok for a given function to be listed under more than one
788 * name. I just didn't.
791 static language lang_names
[] =
793 { "ada", Ada_help
, Ada_funcs
, Ada_suffixes
},
794 { "asm", Asm_help
, Asm_labels
, Asm_suffixes
},
795 { "c", default_C_help
, default_C_entries
, default_C_suffixes
},
796 { "c++", Cplusplus_help
, Cplusplus_entries
, Cplusplus_suffixes
},
797 { "c*", no_lang_help
, Cstar_entries
, Cstar_suffixes
},
798 { "cobol", Cobol_help
, Cobol_paragraphs
, Cobol_suffixes
},
799 { "erlang", Erlang_help
, Erlang_functions
, Erlang_suffixes
},
800 { "forth", Forth_help
, Forth_words
, Forth_suffixes
},
801 { "fortran", Fortran_help
, Fortran_functions
, Fortran_suffixes
},
802 { "go", Go_help
, Go_functions
, Go_suffixes
},
803 { "html", HTML_help
, HTML_labels
, HTML_suffixes
},
804 { "java", Cjava_help
, Cjava_entries
, Cjava_suffixes
},
805 { "lisp", Lisp_help
, Lisp_functions
, Lisp_suffixes
},
806 { "lua", Lua_help
, Lua_functions
, Lua_suffixes
},
807 { "makefile", Makefile_help
,Makefile_targets
,NULL
,Makefile_filenames
},
808 { "objc", Objc_help
, plain_C_entries
, Objc_suffixes
},
809 { "pascal", Pascal_help
, Pascal_functions
, Pascal_suffixes
},
810 { "perl",Perl_help
,Perl_functions
,Perl_suffixes
,NULL
,Perl_interpreters
},
811 { "php", PHP_help
, PHP_functions
, PHP_suffixes
},
812 { "postscript",PS_help
, PS_functions
, PS_suffixes
},
813 { "proc", no_lang_help
, plain_C_entries
, plain_C_suffixes
},
814 { "prolog", Prolog_help
, Prolog_functions
, Prolog_suffixes
},
815 { "python", Python_help
, Python_functions
, Python_suffixes
},
816 { "ruby", Ruby_help
, Ruby_functions
, Ruby_suffixes
},
817 { "scheme", Scheme_help
, Scheme_functions
, Scheme_suffixes
},
818 { "tex", TeX_help
, TeX_commands
, TeX_suffixes
},
819 { "texinfo", Texinfo_help
, Texinfo_nodes
, Texinfo_suffixes
},
820 { "yacc", Yacc_help
,Yacc_entries
,Yacc_suffixes
,NULL
,NULL
,true},
821 { "auto", auto_help
}, /* default guessing scheme */
822 { "none", none_help
, just_read_file
}, /* regexp matching only */
823 { NULL
} /* end of list */
828 print_language_names (void)
831 const char **name
, **ext
;
833 puts ("\nThese are the currently supported languages, along with the\n\
834 default file names and dot suffixes:");
835 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
837 printf (" %-*s", 10, lang
->name
);
838 if (lang
->filenames
!= NULL
)
839 for (name
= lang
->filenames
; *name
!= NULL
; name
++)
840 printf (" %s", *name
);
841 if (lang
->suffixes
!= NULL
)
842 for (ext
= lang
->suffixes
; *ext
!= NULL
; ext
++)
843 printf (" .%s", *ext
);
846 puts ("where 'auto' means use default language for files based on file\n\
847 name suffix, and 'none' means only do regexp processing on files.\n\
848 If no language is specified and no matching suffix is found,\n\
849 the first line of the file is read for a sharp-bang (#!) sequence\n\
850 followed by the name of an interpreter. If no such sequence is found,\n\
851 Fortran is tried first; if no tags are found, C is tried next.\n\
852 When parsing any C file, a \"class\" or \"template\" keyword\n\
854 puts ("Compressed files are supported using gzip, bzip2, and xz.\n\
856 For detailed help on a given language use, for example,\n\
857 etags --help --lang=ada.");
861 # define EMACS_NAME "standalone"
864 # define VERSION "17.38.1.4"
866 static _Noreturn
void
869 char emacs_copyright
[] = COPYRIGHT
;
871 printf ("%s (%s %s)\n", (CTAGS
) ? "ctags" : "etags", EMACS_NAME
, VERSION
);
872 puts (emacs_copyright
);
873 puts ("This program is distributed under the terms in ETAGS.README");
878 #ifndef PRINT_UNDOCUMENTED_OPTIONS_HELP
879 # define PRINT_UNDOCUMENTED_OPTIONS_HELP false
882 static _Noreturn
void
883 print_help (argument
*argbuffer
)
885 bool help_for_lang
= false;
887 for (; argbuffer
->arg_type
!= at_end
; argbuffer
++)
888 if (argbuffer
->arg_type
== at_language
)
892 puts (argbuffer
->lang
->help
);
893 help_for_lang
= true;
899 printf ("Usage: %s [options] [[regex-option ...] file-name] ...\n\
901 These are the options accepted by %s.\n", progname
, progname
);
902 puts ("You may use unambiguous abbreviations for the long option names.");
903 puts (" A - as file name means read names from stdin (one per line).\n\
904 Absolute names are stored in the output file as they are.\n\
905 Relative ones are stored relative to the output file's directory.\n");
907 puts ("-a, --append\n\
908 Append tag entries to existing tags file.");
910 puts ("--packages-only\n\
911 For Ada files, only generate tags for packages.");
914 puts ("-B, --backward-search\n\
915 Write the search commands for the tag entries using '?', the\n\
916 backward-search command instead of '/', the forward-search command.");
918 /* This option is mostly obsolete, because etags can now automatically
919 detect C++. Retained for backward compatibility and for debugging and
920 experimentation. In principle, we could want to tag as C++ even
921 before any "class" or "template" keyword.
923 Treat files whose name suffix defaults to C language as C++ files.");
926 puts ("--declarations\n\
927 In C and derived languages, create tags for function declarations,");
929 puts ("\tand create tags for extern variables if --globals is used.");
932 ("\tand create tags for extern variables unless --no-globals is used.");
935 puts ("-d, --defines\n\
936 Create tag entries for C #define constants and enum constants, too.");
938 puts ("-D, --no-defines\n\
939 Don't create tag entries for C #define constants and enum constants.\n\
940 This makes the tags file smaller.");
943 puts ("-i FILE, --include=FILE\n\
944 Include a note in tag file indicating that, when searching for\n\
945 a tag, one should also consult the tags file FILE after\n\
946 checking the current file.");
948 puts ("-l LANG, --language=LANG\n\
949 Force the following files to be considered as written in the\n\
950 named language up to the next --language=LANG option.");
954 Create tag entries for global variables in some languages.");
956 puts ("--no-globals\n\
957 Do not create tag entries for global variables in some\n\
958 languages. This makes the tags file smaller.");
960 puts ("--no-line-directive\n\
961 Ignore #line preprocessor directives in C and derived languages.");
965 Create tag entries for members of structures in some languages.");
967 puts ("--no-members\n\
968 Do not create tag entries for members of structures\n\
969 in some languages.");
971 puts ("-Q, --class-qualify\n\
972 Qualify tag names with their class name in C++, ObjC, and Java.\n\
973 This produces tag names of the form \"class::member\" for C++,\n\
974 \"class(category)\" for Objective C, and \"class.member\" for Java.\n\
975 For Objective C, this also produces class methods qualified with\n\
976 their arguments, as in \"foo:bar:baz:more\".");
977 puts ("-r REGEXP, --regex=REGEXP or --regex=@regexfile\n\
978 Make a tag for each line matching a regular expression pattern\n\
979 in the following files. {LANGUAGE}REGEXP uses REGEXP for LANGUAGE\n\
980 files only. REGEXFILE is a file containing one REGEXP per line.\n\
981 REGEXP takes the form /TAGREGEXP/TAGNAME/MODS, where TAGNAME/ is\n\
982 optional. The TAGREGEXP pattern is anchored (as if preceded by ^).");
983 puts (" If TAGNAME/ is present, the tags created are named.\n\
984 For example Tcl named tags can be created with:\n\
985 --regex=\"/proc[ \\t]+\\([^ \\t]+\\)/\\1/.\".\n\
986 MODS are optional one-letter modifiers: 'i' means to ignore case,\n\
987 'm' means to allow multi-line matches, 's' implies 'm' and\n\
988 causes dot to match any character, including newline.");
990 puts ("-R, --no-regex\n\
991 Don't create tags from regexps for the following files.");
993 puts ("-I, --ignore-indentation\n\
994 In C and C++ do not assume that a closing brace in the first\n\
995 column is the final brace of a function or structure definition.");
997 puts ("-o FILE, --output=FILE\n\
998 Write the tags to FILE.");
1000 puts ("--parse-stdin=NAME\n\
1001 Read from standard input and record tags as belonging to file NAME.");
1005 puts ("-t, --typedefs\n\
1006 Generate tag entries for C and Ada typedefs.");
1007 puts ("-T, --typedefs-and-c++\n\
1008 Generate tag entries for C typedefs, C struct/enum/union tags,\n\
1009 and C++ member functions.");
1013 puts ("-u, --update\n\
1014 Update the tag entries for the given files, leaving tag\n\
1015 entries for other files in place. Currently, this is\n\
1016 implemented by deleting the existing entries for the given\n\
1017 files and then rewriting the new entries at the end of the\n\
1018 tags file. It is often faster to simply rebuild the entire\n\
1019 tag file than to use this.");
1023 puts ("-v, --vgrind\n\
1024 Print on the standard output an index of items intended for\n\
1025 human consumption, similar to the output of vgrind. The index\n\
1026 is sorted, and gives the page number of each item.");
1028 if (PRINT_UNDOCUMENTED_OPTIONS_HELP
)
1029 puts ("-w, --no-duplicates\n\
1030 Do not create duplicate tag entries, for compatibility with\n\
1031 traditional ctags.");
1033 if (PRINT_UNDOCUMENTED_OPTIONS_HELP
)
1034 puts ("-w, --no-warn\n\
1035 Suppress warning messages about duplicate tag entries.");
1037 puts ("-x, --cxref\n\
1038 Like --vgrind, but in the style of cxref, rather than vgrind.\n\
1039 The output uses line numbers instead of page numbers, but\n\
1040 beyond that the differences are cosmetic; try both to see\n\
1044 puts ("-V, --version\n\
1045 Print the version of the program.\n\
1047 Print this help message.\n\
1048 Followed by one or more '--language' options prints detailed\n\
1049 help about tag generation for the specified languages.");
1051 print_language_names ();
1054 puts ("Report bugs to bug-gnu-emacs@gnu.org");
1056 exit (EXIT_SUCCESS
);
1061 main (int argc
, char **argv
)
1064 unsigned int nincluded_files
;
1065 char **included_files
;
1066 argument
*argbuffer
;
1067 int current_arg
, file_count
;
1068 linebuffer filename_lb
;
1069 bool help_asked
= false;
1075 nincluded_files
= 0;
1076 included_files
= xnew (argc
, char *);
1080 /* Allocate enough no matter what happens. Overkill, but each one
1082 argbuffer
= xnew (argc
, argument
);
1085 * Always find typedefs and structure tags.
1086 * Also default to find macro constants, enum constants, struct
1087 * members and global variables. Do it for both etags and ctags.
1089 typedefs
= typedefs_or_cplusplus
= constantypedefs
= true;
1090 globals
= members
= true;
1092 /* When the optstring begins with a '-' getopt_long does not rearrange the
1093 non-options arguments to be at the end, but leaves them alone. */
1094 optstring
= concat ("-ac:Cf:Il:o:Qr:RSVhH",
1095 (CTAGS
) ? "BxdtTuvw" : "Di:",
1098 while ((opt
= getopt_long (argc
, argv
, optstring
, longopts
, NULL
)) != EOF
)
1102 /* If getopt returns 0, then it has already processed a
1103 long-named option. We should do nothing. */
1107 /* This means that a file name has been seen. Record it. */
1108 argbuffer
[current_arg
].arg_type
= at_filename
;
1109 argbuffer
[current_arg
].what
= optarg
;
1110 len
= strlen (optarg
);
1111 if (whatlen_max
< len
)
1118 /* Parse standard input. Idea by Vivek <vivek@etla.org>. */
1119 argbuffer
[current_arg
].arg_type
= at_stdin
;
1120 argbuffer
[current_arg
].what
= optarg
;
1121 len
= strlen (optarg
);
1122 if (whatlen_max
< len
)
1127 fatal ("cannot parse standard input more than once");
1128 parsing_stdin
= true;
1131 /* Common options. */
1132 case 'a': append_to_tagfile
= true; break;
1133 case 'C': cplusplus
= true; break;
1134 case 'f': /* for compatibility with old makefiles */
1138 error ("-o option may only be given once.");
1139 suggest_asking_for_help ();
1145 case 'S': /* for backward compatibility */
1146 ignoreindent
= true;
1150 language
*lang
= get_language_from_langname (optarg
);
1153 argbuffer
[current_arg
].lang
= lang
;
1154 argbuffer
[current_arg
].arg_type
= at_language
;
1160 /* Backward compatibility: support obsolete --ignore-case-regexp. */
1161 optarg
= concat (optarg
, "i", ""); /* memory leak here */
1164 argbuffer
[current_arg
].arg_type
= at_regexp
;
1165 argbuffer
[current_arg
].what
= optarg
;
1166 len
= strlen (optarg
);
1167 if (whatlen_max
< len
)
1172 argbuffer
[current_arg
].arg_type
= at_regexp
;
1173 argbuffer
[current_arg
].what
= NULL
;
1188 case 'D': constantypedefs
= false; break;
1189 case 'i': included_files
[nincluded_files
++] = optarg
; break;
1191 /* Ctags options. */
1192 case 'B': searchar
= '?'; break;
1193 case 'd': constantypedefs
= true; break;
1194 case 't': typedefs
= true; break;
1195 case 'T': typedefs
= typedefs_or_cplusplus
= true; break;
1196 case 'u': update
= true; break;
1197 case 'v': vgrind_style
= true; /*FALLTHRU*/
1198 case 'x': cxref_style
= true; break;
1199 case 'w': no_warnings
= true; break;
1201 suggest_asking_for_help ();
1205 /* No more options. Store the rest of arguments. */
1206 for (; optind
< argc
; optind
++)
1208 argbuffer
[current_arg
].arg_type
= at_filename
;
1209 argbuffer
[current_arg
].what
= argv
[optind
];
1210 len
= strlen (argv
[optind
]);
1211 if (whatlen_max
< len
)
1217 argbuffer
[current_arg
].arg_type
= at_end
;
1220 print_help (argbuffer
);
1223 if (nincluded_files
== 0 && file_count
== 0)
1225 error ("no input files specified.");
1226 suggest_asking_for_help ();
1230 if (tagfile
== NULL
)
1231 tagfile
= savestr (CTAGS
? "tags" : "TAGS");
1232 cwd
= etags_getcwd (); /* the current working directory */
1233 if (cwd
[strlen (cwd
) - 1] != '/')
1236 cwd
= concat (oldcwd
, "/", "");
1240 /* Compute base directory for relative file names. */
1241 if (streq (tagfile
, "-")
1242 || strneq (tagfile
, "/dev/", 5))
1243 tagfiledir
= cwd
; /* relative file names are relative to cwd */
1246 canonicalize_filename (tagfile
);
1247 tagfiledir
= absolute_dirname (tagfile
, cwd
);
1250 linebuffer_init (&lb
);
1251 linebuffer_init (&filename_lb
);
1252 linebuffer_init (&filebuf
);
1253 linebuffer_init (&token_name
);
1257 if (streq (tagfile
, "-"))
1260 SET_BINARY (fileno (stdout
));
1263 tagf
= fopen (tagfile
, append_to_tagfile
? "ab" : "wb");
1269 * Loop through files finding functions.
1271 for (i
= 0; i
< current_arg
; i
++)
1273 static language
*lang
; /* non-NULL if language is forced */
1276 switch (argbuffer
[i
].arg_type
)
1279 lang
= argbuffer
[i
].lang
;
1282 analyze_regex (argbuffer
[i
].what
);
1285 this_file
= argbuffer
[i
].what
;
1286 /* Input file named "-" means read file names from stdin
1287 (one per line) and use them. */
1288 if (streq (this_file
, "-"))
1291 fatal ("cannot parse standard input "
1292 "AND read file names from it");
1293 while (readline_internal (&filename_lb
, stdin
, "-") > 0)
1294 process_file_name (filename_lb
.buffer
, lang
);
1297 process_file_name (this_file
, lang
);
1300 this_file
= argbuffer
[i
].what
;
1301 process_file (stdin
, this_file
, lang
);
1304 error ("internal error: arg_type");
1310 free (filebuf
.buffer
);
1311 free (token_name
.buffer
);
1313 if (!CTAGS
|| cxref_style
)
1315 /* Write the remaining tags to tagf (ETAGS) or stdout (CXREF). */
1316 put_entries (nodehead
);
1317 free_tree (nodehead
);
1323 /* Output file entries that have no tags. */
1324 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
1326 fprintf (tagf
, "\f\n%s,0\n", fdp
->taggedfname
);
1328 while (nincluded_files
-- > 0)
1329 fprintf (tagf
, "\f\n%s,include\n", *included_files
++);
1331 if (fclose (tagf
) == EOF
)
1335 exit (EXIT_SUCCESS
);
1338 /* From here on, we are in (CTAGS && !cxref_style) */
1342 xmalloc (strlen (tagfile
) + whatlen_max
+
1343 sizeof "mv..OTAGS;fgrep -v '\t\t' OTAGS >;rm OTAGS");
1344 for (i
= 0; i
< current_arg
; ++i
)
1346 switch (argbuffer
[i
].arg_type
)
1352 continue; /* the for loop */
1354 char *z
= stpcpy (cmd
, "mv ");
1355 z
= stpcpy (z
, tagfile
);
1356 z
= stpcpy (z
, " OTAGS;fgrep -v '\t");
1357 z
= stpcpy (z
, argbuffer
[i
].what
);
1358 z
= stpcpy (z
, "\t' OTAGS >");
1359 z
= stpcpy (z
, tagfile
);
1360 strcpy (z
, ";rm OTAGS");
1361 if (system (cmd
) != EXIT_SUCCESS
)
1362 fatal ("failed to execute shell command");
1365 append_to_tagfile
= true;
1368 tagf
= fopen (tagfile
, append_to_tagfile
? "ab" : "wb");
1371 put_entries (nodehead
); /* write all the tags (CTAGS) */
1372 free_tree (nodehead
);
1374 if (fclose (tagf
) == EOF
)
1378 if (append_to_tagfile
|| update
)
1380 char *cmd
= xmalloc (2 * strlen (tagfile
) + sizeof "sort -u -o..");
1381 /* Maybe these should be used:
1382 setenv ("LC_COLLATE", "C", 1);
1383 setenv ("LC_ALL", "C", 1); */
1384 char *z
= stpcpy (cmd
, "sort -u -o ");
1385 z
= stpcpy (z
, tagfile
);
1387 strcpy (z
, tagfile
);
1388 exit (system (cmd
));
1390 return EXIT_SUCCESS
;
1395 * Return a compressor given the file name. If EXTPTR is non-zero,
1396 * return a pointer into FILE where the compressor-specific
1397 * extension begins. If no compressor is found, NULL is returned
1398 * and EXTPTR is not significant.
1399 * Idea by Vladimir Alexiev <vladimir@cs.ualberta.ca> (1998)
1402 get_compressor_from_suffix (char *file
, char **extptr
)
1405 char *slash
, *suffix
;
1407 /* File has been processed by canonicalize_filename,
1408 so we don't need to consider backslashes on DOS_NT. */
1409 slash
= strrchr (file
, '/');
1410 suffix
= strrchr (file
, '.');
1411 if (suffix
== NULL
|| suffix
< slash
)
1416 /* Let those poor souls who live with DOS 8+3 file name limits get
1417 some solace by treating foo.cgz as if it were foo.c.gz, etc.
1418 Only the first do loop is run if not MSDOS */
1421 for (compr
= compressors
; compr
->suffix
!= NULL
; compr
++)
1422 if (streq (compr
->suffix
, suffix
))
1425 break; /* do it only once: not really a loop */
1428 } while (*suffix
!= '\0');
1435 * Return a language given the name.
1438 get_language_from_langname (const char *name
)
1443 error ("empty language name");
1446 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1447 if (streq (name
, lang
->name
))
1449 error ("unknown language \"%s\"", name
);
1457 * Return a language given the interpreter name.
1460 get_language_from_interpreter (char *interpreter
)
1465 if (interpreter
== NULL
)
1467 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1468 if (lang
->interpreters
!= NULL
)
1469 for (iname
= lang
->interpreters
; *iname
!= NULL
; iname
++)
1470 if (streq (*iname
, interpreter
))
1479 * Return a language given the file name.
1482 get_language_from_filename (char *file
, int case_sensitive
)
1485 const char **name
, **ext
, *suffix
;
1487 /* Try whole file name first. */
1488 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1489 if (lang
->filenames
!= NULL
)
1490 for (name
= lang
->filenames
; *name
!= NULL
; name
++)
1491 if ((case_sensitive
)
1492 ? streq (*name
, file
)
1493 : strcaseeq (*name
, file
))
1496 /* If not found, try suffix after last dot. */
1497 suffix
= strrchr (file
, '.');
1501 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1502 if (lang
->suffixes
!= NULL
)
1503 for (ext
= lang
->suffixes
; *ext
!= NULL
; ext
++)
1504 if ((case_sensitive
)
1505 ? streq (*ext
, suffix
)
1506 : strcaseeq (*ext
, suffix
))
1513 * This routine is called on each file argument.
1516 process_file_name (char *file
, language
*lang
)
1521 char *compressed_name
, *uncompressed_name
;
1522 char *ext
, *real_name
, *tmp_name
;
1525 canonicalize_filename (file
);
1526 if (streq (file
, tagfile
) && !streq (tagfile
, "-"))
1528 error ("skipping inclusion of %s in self.", file
);
1531 compr
= get_compressor_from_suffix (file
, &ext
);
1534 compressed_name
= file
;
1535 uncompressed_name
= savenstr (file
, ext
- file
);
1539 compressed_name
= NULL
;
1540 uncompressed_name
= file
;
1543 /* If the canonicalized uncompressed name
1544 has already been dealt with, skip it silently. */
1545 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
1547 assert (fdp
->infname
!= NULL
);
1548 if (streq (uncompressed_name
, fdp
->infname
))
1552 inf
= fopen (file
, "r" FOPEN_BINARY
);
1557 int file_errno
= errno
;
1558 if (compressed_name
)
1560 /* Try with the given suffix. */
1561 inf
= fopen (uncompressed_name
, "r" FOPEN_BINARY
);
1563 real_name
= uncompressed_name
;
1567 /* Try all possible suffixes. */
1568 for (compr
= compressors
; compr
->suffix
!= NULL
; compr
++)
1570 compressed_name
= concat (file
, ".", compr
->suffix
);
1571 inf
= fopen (compressed_name
, "r" FOPEN_BINARY
);
1574 real_name
= compressed_name
;
1579 char *suf
= compressed_name
+ strlen (file
);
1580 size_t suflen
= strlen (compr
->suffix
) + 1;
1581 for ( ; suf
[1]; suf
++, suflen
--)
1583 memmove (suf
, suf
+ 1, suflen
);
1584 inf
= fopen (compressed_name
, "r" FOPEN_BINARY
);
1587 real_name
= compressed_name
;
1594 free (compressed_name
);
1595 compressed_name
= NULL
;
1606 if (real_name
== compressed_name
)
1609 tmp_name
= etags_mktmp ();
1614 #if MSDOS || defined (DOS_NT)
1615 char *cmd1
= concat (compr
->command
, " \"", real_name
);
1616 char *cmd
= concat (cmd1
, "\" > ", tmp_name
);
1618 char *cmd1
= concat (compr
->command
, " '", real_name
);
1619 char *cmd
= concat (cmd1
, "' > ", tmp_name
);
1623 if (system (cmd
) == -1)
1630 inf
= fopen (tmp_name
, "r" FOPEN_BINARY
);
1644 process_file (inf
, uncompressed_name
, lang
);
1646 retval
= fclose (inf
);
1647 if (real_name
== compressed_name
)
1656 if (compressed_name
!= file
)
1657 free (compressed_name
);
1658 if (uncompressed_name
!= file
)
1659 free (uncompressed_name
);
1666 process_file (FILE *fh
, char *fn
, language
*lang
)
1668 static const fdesc emptyfdesc
;
1672 /* Create a new input file description entry. */
1673 fdp
= xnew (1, fdesc
);
1676 fdp
->infname
= savestr (fn
);
1678 fdp
->infabsname
= absolute_filename (fn
, cwd
);
1679 fdp
->infabsdir
= absolute_dirname (fn
, cwd
);
1680 if (filename_is_absolute (fn
))
1682 /* An absolute file name. Canonicalize it. */
1683 fdp
->taggedfname
= absolute_filename (fn
, NULL
);
1687 /* A file name relative to cwd. Make it relative
1688 to the directory of the tags file. */
1689 fdp
->taggedfname
= relative_filename (fn
, tagfiledir
);
1691 fdp
->usecharno
= true; /* use char position when making tags */
1693 fdp
->written
= false; /* not written on tags file yet */
1696 curfdp
= fdhead
; /* the current file description */
1700 /* If not Ctags, and if this is not metasource and if it contained no #line
1701 directives, we can write the tags and free all nodes pointing to
1704 && curfdp
->usecharno
/* no #line directives in this file */
1705 && !curfdp
->lang
->metasource
)
1709 /* Look for the head of the sublist relative to this file. See add_node
1710 for the structure of the node tree. */
1712 for (np
= nodehead
; np
!= NULL
; prev
= np
, np
= np
->left
)
1713 if (np
->fdp
== curfdp
)
1716 /* If we generated tags for this file, write and delete them. */
1719 /* This is the head of the last sublist, if any. The following
1720 instructions depend on this being true. */
1721 assert (np
->left
== NULL
);
1723 assert (fdhead
== curfdp
);
1724 assert (last_node
->fdp
== curfdp
);
1725 put_entries (np
); /* write tags for file curfdp->taggedfname */
1726 free_tree (np
); /* remove the written nodes */
1728 nodehead
= NULL
; /* no nodes left */
1730 prev
->left
= NULL
; /* delete the pointer to the sublist */
1736 reset_input (FILE *inf
)
1738 if (fseek (inf
, 0, SEEK_SET
) != 0)
1739 perror (infilename
);
1743 * This routine opens the specified file and calls the function
1744 * which finds the function and type definitions.
1747 find_entries (FILE *inf
)
1750 language
*lang
= curfdp
->lang
;
1751 Lang_function
*parser
= NULL
;
1753 /* If user specified a language, use it. */
1754 if (lang
!= NULL
&& lang
->function
!= NULL
)
1756 parser
= lang
->function
;
1759 /* Else try to guess the language given the file name. */
1762 lang
= get_language_from_filename (curfdp
->infname
, true);
1763 if (lang
!= NULL
&& lang
->function
!= NULL
)
1765 curfdp
->lang
= lang
;
1766 parser
= lang
->function
;
1770 /* Else look for sharp-bang as the first two characters. */
1772 && readline_internal (&lb
, inf
, infilename
) > 0
1774 && lb
.buffer
[0] == '#'
1775 && lb
.buffer
[1] == '!')
1779 /* Set lp to point at the first char after the last slash in the
1780 line or, if no slashes, at the first nonblank. Then set cp to
1781 the first successive blank and terminate the string. */
1782 lp
= strrchr (lb
.buffer
+2, '/');
1786 lp
= skip_spaces (lb
.buffer
+ 2);
1787 cp
= skip_non_spaces (lp
);
1790 if (strlen (lp
) > 0)
1792 lang
= get_language_from_interpreter (lp
);
1793 if (lang
!= NULL
&& lang
->function
!= NULL
)
1795 curfdp
->lang
= lang
;
1796 parser
= lang
->function
;
1803 /* Else try to guess the language given the case insensitive file name. */
1806 lang
= get_language_from_filename (curfdp
->infname
, false);
1807 if (lang
!= NULL
&& lang
->function
!= NULL
)
1809 curfdp
->lang
= lang
;
1810 parser
= lang
->function
;
1814 /* Else try Fortran or C. */
1817 node
*old_last_node
= last_node
;
1819 curfdp
->lang
= get_language_from_langname ("fortran");
1822 if (old_last_node
== last_node
)
1823 /* No Fortran entries found. Try C. */
1826 curfdp
->lang
= get_language_from_langname (cplusplus
? "c++" : "c");
1832 if (!no_line_directive
1833 && curfdp
->lang
!= NULL
&& curfdp
->lang
->metasource
)
1834 /* It may be that this is a bingo.y file, and we already parsed a bingo.c
1835 file, or anyway we parsed a file that is automatically generated from
1836 this one. If this is the case, the bingo.c file contained #line
1837 directives that generated tags pointing to this file. Let's delete
1838 them all before parsing this file, which is the real source. */
1840 fdesc
**fdpp
= &fdhead
;
1841 while (*fdpp
!= NULL
)
1843 && streq ((*fdpp
)->taggedfname
, curfdp
->taggedfname
))
1844 /* We found one of those! We must delete both the file description
1845 and all tags referring to it. */
1847 fdesc
*badfdp
= *fdpp
;
1849 /* Delete the tags referring to badfdp->taggedfname
1850 that were obtained from badfdp->infname. */
1851 invalidate_nodes (badfdp
, &nodehead
);
1853 *fdpp
= badfdp
->next
; /* remove the bad description from the list */
1854 free_fdesc (badfdp
);
1857 fdpp
= &(*fdpp
)->next
; /* advance the list pointer */
1860 assert (parser
!= NULL
);
1862 /* Generic initializations before reading from file. */
1863 linebuffer_setlen (&filebuf
, 0); /* reset the file buffer */
1865 /* Generic initializations before parsing file with readline. */
1866 lineno
= 0; /* reset global line number */
1867 charno
= 0; /* reset global char number */
1868 linecharno
= 0; /* reset global char number of line start */
1872 regex_tag_multiline ();
1877 * Check whether an implicitly named tag should be created,
1878 * then call `pfnote'.
1879 * NAME is a string that is internally copied by this function.
1881 * TAGS format specification
1882 * Idea by Sam Kendall <kendall@mv.mv.com> (1997)
1883 * The following is explained in some more detail in etc/ETAGS.EBNF.
1885 * make_tag creates tags with "implicit tag names" (unnamed tags)
1886 * if the following are all true, assuming NONAM=" \f\t\n\r()=,;":
1887 * 1. NAME does not contain any of the characters in NONAM;
1888 * 2. LINESTART contains name as either a rightmost, or rightmost but
1889 * one character, substring;
1890 * 3. the character, if any, immediately before NAME in LINESTART must
1891 * be a character in NONAM;
1892 * 4. the character, if any, immediately after NAME in LINESTART must
1893 * also be a character in NONAM.
1895 * The implementation uses the notinname() macro, which recognizes the
1896 * characters stored in the string `nonam'.
1897 * etags.el needs to use the same characters that are in NONAM.
1900 make_tag (const char *name
, /* tag name, or NULL if unnamed */
1901 int namelen
, /* tag length */
1902 bool is_func
, /* tag is a function */
1903 char *linestart
, /* start of the line where tag is */
1904 int linelen
, /* length of the line where tag is */
1905 int lno
, /* line number */
1906 long int cno
) /* character number */
1908 bool named
= (name
!= NULL
&& namelen
> 0);
1911 if (!CTAGS
&& named
) /* maybe set named to false */
1912 /* Let's try to make an implicit tag name, that is, create an unnamed tag
1913 such that etags.el can guess a name from it. */
1916 register const char *cp
= name
;
1918 for (i
= 0; i
< namelen
; i
++)
1919 if (notinname (*cp
++))
1921 if (i
== namelen
) /* rule #1 */
1923 cp
= linestart
+ linelen
- namelen
;
1924 if (notinname (linestart
[linelen
-1]))
1925 cp
-= 1; /* rule #4 */
1926 if (cp
>= linestart
/* rule #2 */
1928 || notinname (cp
[-1])) /* rule #3 */
1929 && strneq (name
, cp
, namelen
)) /* rule #2 */
1930 named
= false; /* use implicit tag name */
1935 nname
= savenstr (name
, namelen
);
1937 pfnote (nname
, is_func
, linestart
, linelen
, lno
, cno
);
1942 pfnote (char *name
, bool is_func
, char *linestart
, int linelen
, int lno
,
1944 /* tag name, or NULL if unnamed */
1945 /* tag is a function */
1946 /* start of the line where tag is */
1947 /* length of the line where tag is */
1949 /* character number */
1953 assert (name
== NULL
|| name
[0] != '\0');
1954 if (CTAGS
&& name
== NULL
)
1957 np
= xnew (1, node
);
1959 /* If ctags mode, change name "main" to M<thisfilename>. */
1960 if (CTAGS
&& !cxref_style
&& streq (name
, "main"))
1962 char *fp
= strrchr (curfdp
->taggedfname
, '/');
1963 np
->name
= concat ("M", fp
== NULL
? curfdp
->taggedfname
: fp
+ 1, "");
1964 fp
= strrchr (np
->name
, '.');
1965 if (fp
!= NULL
&& fp
[1] != '\0' && fp
[2] == '\0')
1971 np
->been_warned
= false;
1973 np
->is_func
= is_func
;
1975 if (np
->fdp
->usecharno
)
1976 /* Our char numbers are 0-base, because of C language tradition?
1977 ctags compatibility? old versions compatibility? I don't know.
1978 Anyway, since emacs's are 1-base we expect etags.el to take care
1979 of the difference. If we wanted to have 1-based numbers, we would
1980 uncomment the +1 below. */
1981 np
->cno
= cno
/* + 1 */ ;
1983 np
->cno
= invalidcharno
;
1984 np
->left
= np
->right
= NULL
;
1985 if (CTAGS
&& !cxref_style
)
1987 if (strlen (linestart
) < 50)
1988 np
->regex
= concat (linestart
, "$", "");
1990 np
->regex
= savenstr (linestart
, 50);
1993 np
->regex
= savenstr (linestart
, linelen
);
1995 add_node (np
, &nodehead
);
2000 * recurse on left children, iterate on right children.
2003 free_tree (register node
*np
)
2007 register node
*node_right
= np
->right
;
2008 free_tree (np
->left
);
2018 * delete a file description
2021 free_fdesc (register fdesc
*fdp
)
2023 free (fdp
->infname
);
2024 free (fdp
->infabsname
);
2025 free (fdp
->infabsdir
);
2026 free (fdp
->taggedfname
);
2033 * Adds a node to the tree of nodes. In etags mode, sort by file
2034 * name. In ctags mode, sort by tag name. Make no attempt at
2037 * add_node is the only function allowed to add nodes, so it can
2041 add_node (node
*np
, node
**cur_node_p
)
2044 register node
*cur_node
= *cur_node_p
;
2046 if (cur_node
== NULL
)
2056 /* For each file name, tags are in a linked sublist on the right
2057 pointer. The first tags of different files are a linked list
2058 on the left pointer. last_node points to the end of the last
2060 if (last_node
!= NULL
&& last_node
->fdp
== np
->fdp
)
2062 /* Let's use the same sublist as the last added node. */
2063 assert (last_node
->right
== NULL
);
2064 last_node
->right
= np
;
2067 else if (cur_node
->fdp
== np
->fdp
)
2069 /* Scanning the list we found the head of a sublist which is
2070 good for us. Let's scan this sublist. */
2071 add_node (np
, &cur_node
->right
);
2074 /* The head of this sublist is not good for us. Let's try the
2076 add_node (np
, &cur_node
->left
);
2077 } /* if ETAGS mode */
2082 dif
= strcmp (np
->name
, cur_node
->name
);
2085 * If this tag name matches an existing one, then
2086 * do not add the node, but maybe print a warning.
2088 if (no_duplicates
&& !dif
)
2090 if (np
->fdp
== cur_node
->fdp
)
2094 fprintf (stderr
, "Duplicate entry in file %s, line %d: %s\n",
2095 np
->fdp
->infname
, lineno
, np
->name
);
2096 fprintf (stderr
, "Second entry ignored\n");
2099 else if (!cur_node
->been_warned
&& !no_warnings
)
2103 "Duplicate entry in files %s and %s: %s (Warning only)\n",
2104 np
->fdp
->infname
, cur_node
->fdp
->infname
, np
->name
);
2105 cur_node
->been_warned
= true;
2110 /* Actually add the node */
2111 add_node (np
, dif
< 0 ? &cur_node
->left
: &cur_node
->right
);
2112 } /* if CTAGS mode */
2116 * invalidate_nodes ()
2117 * Scan the node tree and invalidate all nodes pointing to the
2118 * given file description (CTAGS case) or free them (ETAGS case).
2121 invalidate_nodes (fdesc
*badfdp
, node
**npp
)
2130 if (np
->left
!= NULL
)
2131 invalidate_nodes (badfdp
, &np
->left
);
2132 if (np
->fdp
== badfdp
)
2134 if (np
->right
!= NULL
)
2135 invalidate_nodes (badfdp
, &np
->right
);
2139 assert (np
->fdp
!= NULL
);
2140 if (np
->fdp
== badfdp
)
2142 *npp
= np
->left
; /* detach the sublist from the list */
2143 np
->left
= NULL
; /* isolate it */
2144 free_tree (np
); /* free it */
2145 invalidate_nodes (badfdp
, npp
);
2148 invalidate_nodes (badfdp
, &np
->left
);
2153 static int total_size_of_entries (node
*);
2154 static int number_len (long) ATTRIBUTE_CONST
;
2156 /* Length of a non-negative number's decimal representation. */
2158 number_len (long int num
)
2161 while ((num
/= 10) > 0)
2167 * Return total number of characters that put_entries will output for
2168 * the nodes in the linked list at the right of the specified node.
2169 * This count is irrelevant with etags.el since emacs 19.34 at least,
2170 * but is still supplied for backward compatibility.
2173 total_size_of_entries (register node
*np
)
2175 register int total
= 0;
2177 for (; np
!= NULL
; np
= np
->right
)
2180 total
+= strlen (np
->regex
) + 1; /* pat\177 */
2181 if (np
->name
!= NULL
)
2182 total
+= strlen (np
->name
) + 1; /* name\001 */
2183 total
+= number_len ((long) np
->lno
) + 1; /* lno, */
2184 if (np
->cno
!= invalidcharno
) /* cno */
2185 total
+= number_len (np
->cno
);
2186 total
+= 1; /* newline */
2193 put_entries (register node
*np
)
2196 static fdesc
*fdp
= NULL
;
2201 /* Output subentries that precede this one */
2203 put_entries (np
->left
);
2205 /* Output this entry */
2214 fprintf (tagf
, "\f\n%s,%d\n",
2215 fdp
->taggedfname
, total_size_of_entries (np
));
2216 fdp
->written
= true;
2218 fputs (np
->regex
, tagf
);
2219 fputc ('\177', tagf
);
2220 if (np
->name
!= NULL
)
2222 fputs (np
->name
, tagf
);
2223 fputc ('\001', tagf
);
2225 fprintf (tagf
, "%d,", np
->lno
);
2226 if (np
->cno
!= invalidcharno
)
2227 fprintf (tagf
, "%ld", np
->cno
);
2233 if (np
->name
== NULL
)
2234 error ("internal error: NULL name in ctags mode.");
2239 fprintf (stdout
, "%s %s %d\n",
2240 np
->name
, np
->fdp
->taggedfname
, (np
->lno
+ 63) / 64);
2242 fprintf (stdout
, "%-16s %3d %-16s %s\n",
2243 np
->name
, np
->lno
, np
->fdp
->taggedfname
, np
->regex
);
2247 fprintf (tagf
, "%s\t%s\t", np
->name
, np
->fdp
->taggedfname
);
2250 { /* function or #define macro with args */
2251 putc (searchar
, tagf
);
2254 for (sp
= np
->regex
; *sp
; sp
++)
2256 if (*sp
== '\\' || *sp
== searchar
)
2260 putc (searchar
, tagf
);
2263 { /* anything else; text pattern inadequate */
2264 fprintf (tagf
, "%d", np
->lno
);
2269 } /* if this node contains a valid tag */
2271 /* Output subentries that follow this one */
2272 put_entries (np
->right
);
2274 put_entries (np
->left
);
2279 #define C_EXT 0x00fff /* C extensions */
2280 #define C_PLAIN 0x00000 /* C */
2281 #define C_PLPL 0x00001 /* C++ */
2282 #define C_STAR 0x00003 /* C* */
2283 #define C_JAVA 0x00005 /* JAVA */
2284 #define C_AUTO 0x01000 /* C, but switch to C++ if `class' is met */
2285 #define YACC 0x10000 /* yacc file */
2288 * The C symbol tables.
2293 st_C_objprot
, st_C_objimpl
, st_C_objend
,
2295 st_C_ignore
, st_C_attribute
,
2298 st_C_class
, st_C_template
,
2299 st_C_struct
, st_C_extern
, st_C_enum
, st_C_define
, st_C_typedef
2302 /* Feed stuff between (but not including) %[ and %] lines to:
2308 struct C_stab_entry { char *name; int c_ext; enum sym_type type; }
2312 while, 0, st_C_ignore
2313 switch, 0, st_C_ignore
2314 return, 0, st_C_ignore
2315 __attribute__, 0, st_C_attribute
2316 GTY, 0, st_C_attribute
2317 @interface, 0, st_C_objprot
2318 @protocol, 0, st_C_objprot
2319 @implementation,0, st_C_objimpl
2320 @end, 0, st_C_objend
2321 import, (C_JAVA & ~C_PLPL), st_C_ignore
2322 package, (C_JAVA & ~C_PLPL), st_C_ignore
2323 friend, C_PLPL, st_C_ignore
2324 extends, (C_JAVA & ~C_PLPL), st_C_javastruct
2325 implements, (C_JAVA & ~C_PLPL), st_C_javastruct
2326 interface, (C_JAVA & ~C_PLPL), st_C_struct
2327 class, 0, st_C_class
2328 namespace, C_PLPL, st_C_struct
2329 domain, C_STAR, st_C_struct
2330 union, 0, st_C_struct
2331 struct, 0, st_C_struct
2332 extern, 0, st_C_extern
2334 typedef, 0, st_C_typedef
2335 define, 0, st_C_define
2336 undef, 0, st_C_define
2337 operator, C_PLPL, st_C_operator
2338 template, 0, st_C_template
2339 # DEFUN used in emacs, the next three used in glibc (SYSCALL only for mach).
2340 DEFUN, 0, st_C_gnumacro
2341 SYSCALL, 0, st_C_gnumacro
2342 ENTRY, 0, st_C_gnumacro
2343 PSEUDO, 0, st_C_gnumacro
2344 # These are defined inside C functions, so currently they are not met.
2345 # EXFUN used in glibc, DEFVAR_* in emacs.
2346 #EXFUN, 0, st_C_gnumacro
2347 #DEFVAR_, 0, st_C_gnumacro
2349 and replace lines between %< and %> with its output, then:
2350 - remove the #if characterset check
2351 - make in_word_set static and not inline. */
2353 /* C code produced by gperf version 3.0.1 */
2354 /* Command-line: gperf -m 5 */
2355 /* Computed positions: -k'2-3' */
2357 struct C_stab_entry
{ const char *name
; int c_ext
; enum sym_type type
; };
2358 /* maximum key range = 33, duplicates = 0 */
2361 hash (const char *str
, int len
)
2363 static char const asso_values
[] =
2365 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2366 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2367 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2368 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2369 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2370 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2371 35, 35, 35, 35, 35, 35, 35, 35, 35, 3,
2372 26, 35, 35, 35, 35, 35, 35, 35, 27, 35,
2373 35, 35, 35, 24, 0, 35, 35, 35, 35, 0,
2374 35, 35, 35, 35, 35, 1, 35, 16, 35, 6,
2375 23, 0, 0, 35, 22, 0, 35, 35, 5, 0,
2376 0, 15, 1, 35, 6, 35, 8, 19, 35, 16,
2377 4, 5, 35, 35, 35, 35, 35, 35, 35, 35,
2378 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2379 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2380 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2381 35, 35, 35, 35, 35, 35, 35, 35, 35, 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
2397 hval
+= asso_values
[(unsigned char) str
[2]];
2400 hval
+= asso_values
[(unsigned char) str
[1]];
2406 static struct C_stab_entry
*
2407 in_word_set (register const char *str
, register unsigned int len
)
2411 TOTAL_KEYWORDS
= 33,
2412 MIN_WORD_LENGTH
= 2,
2413 MAX_WORD_LENGTH
= 15,
2418 static struct C_stab_entry wordlist
[] =
2421 {"if", 0, st_C_ignore
},
2422 {"GTY", 0, st_C_attribute
},
2423 {"@end", 0, st_C_objend
},
2424 {"union", 0, st_C_struct
},
2425 {"define", 0, st_C_define
},
2426 {"import", (C_JAVA
& ~C_PLPL
), st_C_ignore
},
2427 {"template", 0, st_C_template
},
2428 {"operator", C_PLPL
, st_C_operator
},
2429 {"@interface", 0, st_C_objprot
},
2430 {"implements", (C_JAVA
& ~C_PLPL
), st_C_javastruct
},
2431 {"friend", C_PLPL
, st_C_ignore
},
2432 {"typedef", 0, st_C_typedef
},
2433 {"return", 0, st_C_ignore
},
2434 {"@implementation",0, st_C_objimpl
},
2435 {"@protocol", 0, st_C_objprot
},
2436 {"interface", (C_JAVA
& ~C_PLPL
), st_C_struct
},
2437 {"extern", 0, st_C_extern
},
2438 {"extends", (C_JAVA
& ~C_PLPL
), st_C_javastruct
},
2439 {"struct", 0, st_C_struct
},
2440 {"domain", C_STAR
, st_C_struct
},
2441 {"switch", 0, st_C_ignore
},
2442 {"enum", 0, st_C_enum
},
2443 {"for", 0, st_C_ignore
},
2444 {"namespace", C_PLPL
, st_C_struct
},
2445 {"class", 0, st_C_class
},
2446 {"while", 0, st_C_ignore
},
2447 {"undef", 0, st_C_define
},
2448 {"package", (C_JAVA
& ~C_PLPL
), st_C_ignore
},
2449 {"__attribute__", 0, st_C_attribute
},
2450 {"SYSCALL", 0, st_C_gnumacro
},
2451 {"ENTRY", 0, st_C_gnumacro
},
2452 {"PSEUDO", 0, st_C_gnumacro
},
2453 {"DEFUN", 0, st_C_gnumacro
}
2456 if (len
<= MAX_WORD_LENGTH
&& len
>= MIN_WORD_LENGTH
)
2458 int key
= hash (str
, len
);
2460 if (key
<= MAX_HASH_VALUE
&& key
>= 0)
2462 const char *s
= wordlist
[key
].name
;
2464 if (*str
== *s
&& !strncmp (str
+ 1, s
+ 1, len
- 1) && s
[len
] == '\0')
2465 return &wordlist
[key
];
2472 static enum sym_type
2473 C_symtype (char *str
, int len
, int c_ext
)
2475 register struct C_stab_entry
*se
= in_word_set (str
, len
);
2477 if (se
== NULL
|| (se
->c_ext
&& !(c_ext
& se
->c_ext
)))
2484 * Ignoring __attribute__ ((list))
2486 static bool inattribute
; /* looking at an __attribute__ construct */
2489 * C functions and variables are recognized using a simple
2490 * finite automaton. fvdef is its state variable.
2494 fvnone
, /* nothing seen */
2495 fdefunkey
, /* Emacs DEFUN keyword seen */
2496 fdefunname
, /* Emacs DEFUN name seen */
2497 foperator
, /* func: operator keyword seen (cplpl) */
2498 fvnameseen
, /* function or variable name seen */
2499 fstartlist
, /* func: just after open parenthesis */
2500 finlist
, /* func: in parameter list */
2501 flistseen
, /* func: after parameter list */
2502 fignore
, /* func: before open brace */
2503 vignore
/* var-like: ignore until ';' */
2506 static bool fvextern
; /* func or var: extern keyword seen; */
2509 * typedefs are recognized using a simple finite automaton.
2510 * typdef is its state variable.
2514 tnone
, /* nothing seen */
2515 tkeyseen
, /* typedef keyword seen */
2516 ttypeseen
, /* defined type seen */
2517 tinbody
, /* inside typedef body */
2518 tend
, /* just before typedef tag */
2519 tignore
/* junk after typedef tag */
2523 * struct-like structures (enum, struct and union) are recognized
2524 * using another simple finite automaton. `structdef' is its state
2529 snone
, /* nothing seen yet,
2530 or in struct body if bracelev > 0 */
2531 skeyseen
, /* struct-like keyword seen */
2532 stagseen
, /* struct-like tag seen */
2533 scolonseen
/* colon seen after struct-like tag */
2537 * When objdef is different from onone, objtag is the name of the class.
2539 static const char *objtag
= "<uninited>";
2542 * Yet another little state machine to deal with preprocessor lines.
2546 dnone
, /* nothing seen */
2547 dsharpseen
, /* '#' seen as first char on line */
2548 ddefineseen
, /* '#' and 'define' seen */
2549 dignorerest
/* ignore rest of line */
2553 * State machine for Objective C protocols and implementations.
2554 * Idea by Tom R.Hageman <tom@basil.icce.rug.nl> (1995)
2558 onone
, /* nothing seen */
2559 oprotocol
, /* @interface or @protocol seen */
2560 oimplementation
, /* @implementations seen */
2561 otagseen
, /* class name seen */
2562 oparenseen
, /* parenthesis before category seen */
2563 ocatseen
, /* category name seen */
2564 oinbody
, /* in @implementation body */
2565 omethodsign
, /* in @implementation body, after +/- */
2566 omethodtag
, /* after method name */
2567 omethodcolon
, /* after method colon */
2568 omethodparm
, /* after method parameter */
2569 oignore
/* wait for @end */
2574 * Use this structure to keep info about the token read, and how it
2575 * should be tagged. Used by the make_C_tag function to build a tag.
2579 char *line
; /* string containing the token */
2580 int offset
; /* where the token starts in LINE */
2581 int length
; /* token length */
2583 The previous members can be used to pass strings around for generic
2584 purposes. The following ones specifically refer to creating tags. In this
2585 case the token contained here is the pattern that will be used to create a
2588 bool valid
; /* do not create a tag; the token should be
2589 invalidated whenever a state machine is
2590 reset prematurely */
2591 bool named
; /* create a named tag */
2592 int lineno
; /* source line number of tag */
2593 long linepos
; /* source char number of tag */
2594 } token
; /* latest token read */
2597 * Variables and functions for dealing with nested structures.
2598 * Idea by Mykola Dzyuba <mdzyuba@yahoo.com> (2001)
2600 static void pushclass_above (int, char *, int);
2601 static void popclass_above (int);
2602 static void write_classname (linebuffer
*, const char *qualifier
);
2605 char **cname
; /* nested class names */
2606 int *bracelev
; /* nested class brace level */
2607 int nl
; /* class nesting level (elements used) */
2608 int size
; /* length of the array */
2609 } cstack
; /* stack for nested declaration tags */
2610 /* Current struct nesting depth (namespace, class, struct, union, enum). */
2611 #define nestlev (cstack.nl)
2612 /* After struct keyword or in struct body, not inside a nested function. */
2613 #define instruct (structdef == snone && nestlev > 0 \
2614 && bracelev == cstack.bracelev[nestlev-1] + 1)
2617 pushclass_above (int bracelev
, char *str
, int len
)
2621 popclass_above (bracelev
);
2623 if (nl
>= cstack
.size
)
2625 int size
= cstack
.size
*= 2;
2626 xrnew (cstack
.cname
, size
, char *);
2627 xrnew (cstack
.bracelev
, size
, int);
2629 assert (nl
== 0 || cstack
.bracelev
[nl
-1] < bracelev
);
2630 cstack
.cname
[nl
] = (str
== NULL
) ? NULL
: savenstr (str
, len
);
2631 cstack
.bracelev
[nl
] = bracelev
;
2636 popclass_above (int bracelev
)
2640 for (nl
= cstack
.nl
- 1;
2641 nl
>= 0 && cstack
.bracelev
[nl
] >= bracelev
;
2644 free (cstack
.cname
[nl
]);
2650 write_classname (linebuffer
*cn
, const char *qualifier
)
2653 int qlen
= strlen (qualifier
);
2655 if (cstack
.nl
== 0 || cstack
.cname
[0] == NULL
)
2659 cn
->buffer
[0] = '\0';
2663 len
= strlen (cstack
.cname
[0]);
2664 linebuffer_setlen (cn
, len
);
2665 strcpy (cn
->buffer
, cstack
.cname
[0]);
2667 for (i
= 1; i
< cstack
.nl
; i
++)
2669 char *s
= cstack
.cname
[i
];
2672 linebuffer_setlen (cn
, len
+ qlen
+ strlen (s
));
2673 len
+= sprintf (cn
->buffer
+ len
, "%s%s", qualifier
, s
);
2678 static bool consider_token (char *, int, int, int *, int, int, bool *);
2679 static void make_C_tag (bool);
2683 * checks to see if the current token is at the start of a
2684 * function or variable, or corresponds to a typedef, or
2685 * is a struct/union/enum tag, or #define, or an enum constant.
2687 * *IS_FUNC_OR_VAR gets true if the token is a function or #define macro
2688 * with args. C_EXTP points to which language we are looking at.
2699 consider_token (char *str
, int len
, int c
, int *c_extp
,
2700 int bracelev
, int parlev
, bool *is_func_or_var
)
2701 /* IN: token pointer */
2702 /* IN: token length */
2703 /* IN: first char after the token */
2704 /* IN, OUT: C extensions mask */
2705 /* IN: brace level */
2706 /* IN: parenthesis level */
2707 /* OUT: function or variable found */
2709 /* When structdef is stagseen, scolonseen, or snone with bracelev > 0,
2710 structtype is the type of the preceding struct-like keyword, and
2711 structbracelev is the brace level where it has been seen. */
2712 static enum sym_type structtype
;
2713 static int structbracelev
;
2714 static enum sym_type toktype
;
2717 toktype
= C_symtype (str
, len
, *c_extp
);
2720 * Skip __attribute__
2722 if (toktype
== st_C_attribute
)
2729 * Advance the definedef state machine.
2734 /* We're not on a preprocessor line. */
2735 if (toktype
== st_C_gnumacro
)
2742 if (toktype
== st_C_define
)
2744 definedef
= ddefineseen
;
2748 definedef
= dignorerest
;
2753 * Make a tag for any macro, unless it is a constant
2754 * and constantypedefs is false.
2756 definedef
= dignorerest
;
2757 *is_func_or_var
= (c
== '(');
2758 if (!*is_func_or_var
&& !constantypedefs
)
2765 error ("internal error: definedef value.");
2774 if (toktype
== st_C_typedef
)
2797 if (structdef
== snone
&& fvdef
== fvnone
)
2819 case st_C_javastruct
:
2820 if (structdef
== stagseen
)
2821 structdef
= scolonseen
;
2825 if ((*c_extp
& C_AUTO
) /* automatic detection of C++ language */
2827 && definedef
== dnone
&& structdef
== snone
2828 && typdef
== tnone
&& fvdef
== fvnone
)
2829 *c_extp
= (*c_extp
| C_PLPL
) & ~C_AUTO
;
2830 if (toktype
== st_C_template
)
2837 && (typdef
== tkeyseen
2838 || (typedefs_or_cplusplus
&& structdef
== snone
)))
2840 structdef
= skeyseen
;
2841 structtype
= toktype
;
2842 structbracelev
= bracelev
;
2843 if (fvdef
== fvnameseen
)
2851 if (structdef
== skeyseen
)
2853 structdef
= stagseen
;
2857 if (typdef
!= tnone
)
2860 /* Detect Objective C constructs. */
2870 objdef
= oimplementation
;
2876 case oimplementation
:
2877 /* Save the class tag for functions or variables defined inside. */
2878 objtag
= savenstr (str
, len
);
2882 /* Save the class tag for categories. */
2883 objtag
= savenstr (str
, len
);
2885 *is_func_or_var
= true;
2889 *is_func_or_var
= true;
2897 objdef
= omethodtag
;
2898 linebuffer_setlen (&token_name
, len
);
2899 memcpy (token_name
.buffer
, str
, len
);
2900 token_name
.buffer
[len
] = '\0';
2906 objdef
= omethodparm
;
2911 objdef
= omethodtag
;
2914 int oldlen
= token_name
.len
;
2916 linebuffer_setlen (&token_name
, oldlen
+ len
);
2917 memcpy (token_name
.buffer
+ oldlen
, str
, len
);
2918 token_name
.buffer
[oldlen
+ len
] = '\0';
2924 if (toktype
== st_C_objend
)
2926 /* Memory leakage here: the string pointed by objtag is
2927 never released, because many tests would be needed to
2928 avoid breaking on incorrect input code. The amount of
2929 memory leaked here is the sum of the lengths of the
2939 /* A function, variable or enum constant? */
2961 *is_func_or_var
= true;
2965 && structdef
== snone
2966 && structtype
== st_C_enum
&& bracelev
> structbracelev
2967 /* Don't tag tokens in expressions that assign values to enum
2969 && fvdef
!= vignore
)
2970 return true; /* enum constant */
2976 fvdef
= fdefunname
; /* GNU macro */
2977 *is_func_or_var
= true;
2985 if ((strneq (str
, "asm", 3) && endtoken (str
[3]))
2986 || (strneq (str
, "__asm__", 7) && endtoken (str
[7])))
2997 if (len
>= 10 && strneq (str
+len
-10, "::operator", 10))
2999 if (*c_extp
& C_AUTO
) /* automatic detection of C++ */
3000 *c_extp
= (*c_extp
| C_PLPL
) & ~C_AUTO
;
3002 *is_func_or_var
= true;
3005 if (bracelev
> 0 && !instruct
)
3007 fvdef
= fvnameseen
; /* function or variable */
3008 *is_func_or_var
= true;
3023 * C_entries often keeps pointers to tokens or lines which are older than
3024 * the line currently read. By keeping two line buffers, and switching
3025 * them at end of line, it is possible to use those pointers.
3033 #define current_lb_is_new (newndx == curndx)
3034 #define switch_line_buffers() (curndx = 1 - curndx)
3036 #define curlb (lbs[curndx].lb)
3037 #define newlb (lbs[newndx].lb)
3038 #define curlinepos (lbs[curndx].linepos)
3039 #define newlinepos (lbs[newndx].linepos)
3041 #define plainc ((c_ext & C_EXT) == C_PLAIN)
3042 #define cplpl (c_ext & C_PLPL)
3043 #define cjava ((c_ext & C_JAVA) == C_JAVA)
3045 #define CNL_SAVE_DEFINEDEF() \
3047 curlinepos = charno; \
3048 readline (&curlb, inf); \
3049 lp = curlb.buffer; \
3056 CNL_SAVE_DEFINEDEF (); \
3057 if (savetoken.valid) \
3059 token = savetoken; \
3060 savetoken.valid = false; \
3062 definedef = dnone; \
3067 make_C_tag (bool isfun
)
3069 /* This function is never called when token.valid is false, but
3070 we must protect against invalid input or internal errors. */
3072 make_tag (token_name
.buffer
, token_name
.len
, isfun
, token
.line
,
3073 token
.offset
+token
.length
+1, token
.lineno
, token
.linepos
);
3075 { /* this branch is optimized away if !DEBUG */
3076 make_tag (concat ("INVALID TOKEN:-->", token_name
.buffer
, ""),
3077 token_name
.len
+ 17, isfun
, token
.line
,
3078 token
.offset
+token
.length
+1, token
.lineno
, token
.linepos
);
3079 error ("INVALID TOKEN");
3082 token
.valid
= false;
3086 perhaps_more_input (FILE *inf
)
3088 return !feof (inf
) && !ferror (inf
);
3094 * This routine finds functions, variables, typedefs,
3095 * #define's, enum constants and struct/union/enum definitions in
3096 * C syntax and adds them to the list.
3099 C_entries (int c_ext
, FILE *inf
)
3100 /* extension of C */
3103 register char c
; /* latest char read; '\0' for end of line */
3104 register char *lp
; /* pointer one beyond the character `c' */
3105 int curndx
, newndx
; /* indices for current and new lb */
3106 register int tokoff
; /* offset in line of start of current token */
3107 register int toklen
; /* length of current token */
3108 const char *qualifier
; /* string used to qualify names */
3109 int qlen
; /* length of qualifier */
3110 int bracelev
; /* current brace level */
3111 int bracketlev
; /* current bracket level */
3112 int parlev
; /* current parenthesis level */
3113 int attrparlev
; /* __attribute__ parenthesis level */
3114 int templatelev
; /* current template level */
3115 int typdefbracelev
; /* bracelev where a typedef struct body begun */
3116 bool incomm
, inquote
, inchar
, quotednl
, midtoken
;
3117 bool yacc_rules
; /* in the rules part of a yacc file */
3118 struct tok savetoken
= {0}; /* token saved during preprocessor handling */
3121 linebuffer_init (&lbs
[0].lb
);
3122 linebuffer_init (&lbs
[1].lb
);
3123 if (cstack
.size
== 0)
3125 cstack
.size
= (DEBUG
) ? 1 : 4;
3127 cstack
.cname
= xnew (cstack
.size
, char *);
3128 cstack
.bracelev
= xnew (cstack
.size
, int);
3131 tokoff
= toklen
= typdefbracelev
= 0; /* keep compiler quiet */
3132 curndx
= newndx
= 0;
3136 fvdef
= fvnone
; fvextern
= false; typdef
= tnone
;
3137 structdef
= snone
; definedef
= dnone
; objdef
= onone
;
3139 midtoken
= inquote
= inchar
= incomm
= quotednl
= false;
3140 token
.valid
= savetoken
.valid
= false;
3141 bracelev
= bracketlev
= parlev
= attrparlev
= templatelev
= 0;
3143 { qualifier
= "."; qlen
= 1; }
3145 { qualifier
= "::"; qlen
= 2; }
3148 while (perhaps_more_input (inf
))
3153 /* If we are at the end of the line, the next character is a
3154 '\0'; do not skip it, because it is what tells us
3155 to read the next line. */
3176 /* Newlines inside comments do not end macro definitions in
3178 CNL_SAVE_DEFINEDEF ();
3191 /* Newlines inside strings do not end macro definitions
3192 in traditional cpp, even though compilers don't
3193 usually accept them. */
3194 CNL_SAVE_DEFINEDEF ();
3204 /* Hmmm, something went wrong. */
3240 if (fvdef
!= finlist
&& fvdef
!= fignore
&& fvdef
!= vignore
)
3255 else if (/* cplpl && */ *lp
== '/')
3261 if ((c_ext
& YACC
) && *lp
== '%')
3263 /* Entering or exiting rules section in yacc file. */
3265 definedef
= dnone
; fvdef
= fvnone
; fvextern
= false;
3266 typdef
= tnone
; structdef
= snone
;
3267 midtoken
= inquote
= inchar
= incomm
= quotednl
= false;
3269 yacc_rules
= !yacc_rules
;
3275 if (definedef
== dnone
)
3278 bool cpptoken
= true;
3280 /* Look back on this line. If all blanks, or nonblanks
3281 followed by an end of comment, this is a preprocessor
3283 for (cp
= newlb
.buffer
; cp
< lp
-1; cp
++)
3284 if (!c_isspace (*cp
))
3286 if (*cp
== '*' && cp
[1] == '/')
3296 definedef
= dsharpseen
;
3297 /* This is needed for tagging enum values: when there are
3298 preprocessor conditionals inside the enum, we need to
3299 reset the value of fvdef so that the next enum value is
3300 tagged even though the one before it did not end in a
3302 if (fvdef
== vignore
&& instruct
&& parlev
== 0)
3304 if (strneq (cp
, "#if", 3) || strneq (cp
, "#el", 3))
3308 } /* if (definedef == dnone) */
3319 CNL_SAVE_DEFINEDEF ();
3326 /* Consider token only if some involved conditions are satisfied. */
3327 if (typdef
!= tignore
3328 && definedef
!= dignorerest
3331 && (definedef
!= dnone
3332 || structdef
!= scolonseen
)
3339 if (c
== ':' && *lp
== ':' && begtoken (lp
[1]))
3340 /* This handles :: in the middle,
3341 but not at the beginning of an identifier.
3342 Also, space-separated :: is not recognized. */
3344 if (c_ext
& C_AUTO
) /* automatic detection of C++ */
3345 c_ext
= (c_ext
| C_PLPL
) & ~C_AUTO
;
3349 goto still_in_token
;
3353 bool funorvar
= false;
3356 || consider_token (newlb
.buffer
+ tokoff
, toklen
, c
,
3357 &c_ext
, bracelev
, parlev
,
3360 if (fvdef
== foperator
)
3363 lp
= skip_spaces (lp
-1);
3367 && !c_isspace (*lp
) && *lp
!= '(')
3370 toklen
+= lp
- oldlp
;
3372 token
.named
= false;
3374 && nestlev
> 0 && definedef
== dnone
)
3375 /* in struct body */
3380 write_classname (&token_name
, qualifier
);
3381 len
= token_name
.len
;
3382 linebuffer_setlen (&token_name
,
3383 len
+ qlen
+ toklen
);
3384 sprintf (token_name
.buffer
+ len
, "%s%.*s",
3386 newlb
.buffer
+ tokoff
);
3390 linebuffer_setlen (&token_name
, toklen
);
3391 sprintf (token_name
.buffer
, "%.*s",
3392 toklen
, newlb
.buffer
+ tokoff
);
3396 else if (objdef
== ocatseen
)
3397 /* Objective C category */
3401 int len
= strlen (objtag
) + 2 + toklen
;
3402 linebuffer_setlen (&token_name
, len
);
3403 sprintf (token_name
.buffer
, "%s(%.*s)",
3405 newlb
.buffer
+ tokoff
);
3409 linebuffer_setlen (&token_name
, toklen
);
3410 sprintf (token_name
.buffer
, "%.*s",
3411 toklen
, newlb
.buffer
+ tokoff
);
3415 else if (objdef
== omethodtag
3416 || objdef
== omethodparm
)
3417 /* Objective C method */
3421 else if (fvdef
== fdefunname
)
3422 /* GNU DEFUN and similar macros */
3424 bool defun
= (newlb
.buffer
[tokoff
] == 'F');
3428 /* Rewrite the tag so that emacs lisp DEFUNs
3429 can be found by their elisp name */
3435 linebuffer_setlen (&token_name
, len
);
3436 memcpy (token_name
.buffer
,
3437 newlb
.buffer
+ off
, len
);
3438 token_name
.buffer
[len
] = '\0';
3441 if (token_name
.buffer
[len
] == '_')
3442 token_name
.buffer
[len
] = '-';
3443 token
.named
= defun
;
3447 linebuffer_setlen (&token_name
, toklen
);
3448 memcpy (token_name
.buffer
,
3449 newlb
.buffer
+ tokoff
, toklen
);
3450 token_name
.buffer
[toklen
] = '\0';
3451 /* Name macros and members. */
3452 token
.named
= (structdef
== stagseen
3453 || typdef
== ttypeseen
3456 && definedef
== dignorerest
)
3458 && definedef
== dnone
3459 && structdef
== snone
3462 token
.lineno
= lineno
;
3463 token
.offset
= tokoff
;
3464 token
.length
= toklen
;
3465 token
.line
= newlb
.buffer
;
3466 token
.linepos
= newlinepos
;
3469 if (definedef
== dnone
3470 && (fvdef
== fvnameseen
3471 || fvdef
== foperator
3472 || structdef
== stagseen
3474 || typdef
== ttypeseen
3475 || objdef
!= onone
))
3477 if (current_lb_is_new
)
3478 switch_line_buffers ();
3480 else if (definedef
!= dnone
3481 || fvdef
== fdefunname
3483 make_C_tag (funorvar
);
3485 else /* not yacc and consider_token failed */
3487 if (inattribute
&& fvdef
== fignore
)
3489 /* We have just met __attribute__ after a
3490 function parameter list: do not tag the
3497 } /* if (endtoken (c)) */
3498 else if (intoken (c
))
3504 } /* if (midtoken) */
3505 else if (begtoken (c
))
3513 /* This prevents tagging fb in
3514 void (__attribute__((noreturn)) *fb) (void);
3515 Fixing this is not easy and not very important. */
3519 if (plainc
|| declarations
)
3521 make_C_tag (true); /* a function */
3528 if (structdef
== stagseen
&& !cjava
)
3530 popclass_above (bracelev
);
3540 if (!yacc_rules
|| lp
== newlb
.buffer
+ 1)
3542 tokoff
= lp
- 1 - newlb
.buffer
;
3547 } /* if (begtoken) */
3548 } /* if must look at token */
3551 /* Detect end of line, colon, comma, semicolon and various braces
3552 after having handled a token.*/
3558 if (yacc_rules
&& token
.offset
== 0 && token
.valid
)
3560 make_C_tag (false); /* a yacc function */
3563 if (definedef
!= dnone
)
3569 make_C_tag (true); /* an Objective C class */
3573 objdef
= omethodcolon
;
3576 int toklen
= token_name
.len
;
3577 linebuffer_setlen (&token_name
, toklen
+ 1);
3578 strcpy (token_name
.buffer
+ toklen
, ":");
3584 if (structdef
== stagseen
)
3586 structdef
= scolonseen
;
3589 /* Should be useless, but may be work as a safety net. */
3590 if (cplpl
&& fvdef
== flistseen
)
3592 make_C_tag (true); /* a function */
3598 if (definedef
!= dnone
|| inattribute
)
3604 make_C_tag (false); /* a typedef */
3614 if (typdef
== tignore
|| cplpl
)
3618 if ((globals
&& bracelev
== 0 && (!fvextern
|| declarations
))
3619 || (members
&& instruct
))
3620 make_C_tag (false); /* a variable */
3623 token
.valid
= false;
3627 && (cplpl
|| !instruct
)
3628 && (typdef
== tnone
|| (typdef
!= tignore
&& instruct
)))
3630 && plainc
&& instruct
))
3631 make_C_tag (true); /* a function */
3637 && cplpl
&& structdef
== stagseen
)
3638 make_C_tag (false); /* forward declaration */
3640 token
.valid
= false;
3641 } /* switch (fvdef) */
3647 if (structdef
== stagseen
)
3651 if (definedef
!= dnone
|| inattribute
)
3657 make_C_tag (true); /* an Objective C method */
3672 if (instruct
&& parlev
== 0)
3683 && (!fvextern
|| declarations
))
3684 || (members
&& instruct
)))
3685 make_C_tag (false); /* a variable */
3688 if ((declarations
&& typdef
== tnone
&& !instruct
)
3689 || (members
&& typdef
!= tignore
&& instruct
))
3691 make_C_tag (true); /* a function */
3694 else if (!declarations
)
3696 token
.valid
= false;
3701 if (structdef
== stagseen
)
3705 if (definedef
!= dnone
|| inattribute
)
3707 if (structdef
== stagseen
)
3714 make_C_tag (false); /* a typedef */
3726 if ((members
&& bracelev
== 1)
3727 || (globals
&& bracelev
== 0
3728 && (!fvextern
|| declarations
)))
3729 make_C_tag (false); /* a variable */
3745 if (definedef
!= dnone
)
3747 if (objdef
== otagseen
&& parlev
== 0)
3748 objdef
= oparenseen
;
3752 if (typdef
== ttypeseen
3756 /* This handles constructs like:
3757 typedef void OperatorFun (int fun); */
3778 if (--attrparlev
== 0)
3779 inattribute
= false;
3782 if (definedef
!= dnone
)
3784 if (objdef
== ocatseen
&& parlev
== 1)
3786 make_C_tag (true); /* an Objective C category */
3802 || typdef
== ttypeseen
))
3805 make_C_tag (false); /* a typedef */
3808 else if (parlev
< 0) /* can happen due to ill-conceived #if's. */
3812 if (definedef
!= dnone
)
3814 if (typdef
== ttypeseen
)
3816 /* Whenever typdef is set to tinbody (currently only
3817 here), typdefbracelev should be set to bracelev. */
3819 typdefbracelev
= bracelev
;
3824 if (cplpl
&& !class_qualify
)
3826 /* Remove class and namespace qualifiers from the token,
3827 leaving only the method/member name. */
3828 char *cc
, *uqname
= token_name
.buffer
;
3829 char *tok_end
= token_name
.buffer
+ token_name
.len
;
3831 for (cc
= token_name
.buffer
; cc
< tok_end
; cc
++)
3833 if (*cc
== ':' && cc
[1] == ':')
3839 if (uqname
> token_name
.buffer
)
3841 int uqlen
= strlen (uqname
);
3842 linebuffer_setlen (&token_name
, uqlen
);
3843 memmove (token_name
.buffer
, uqname
, uqlen
+ 1);
3846 make_C_tag (true); /* a function */
3855 make_C_tag (true); /* an Objective C class */
3860 make_C_tag (true); /* an Objective C method */
3864 /* Neutralize `extern "C" {' grot. */
3865 if (bracelev
== 0 && structdef
== snone
&& nestlev
== 0
3875 case skeyseen
: /* unnamed struct */
3876 pushclass_above (bracelev
, NULL
, 0);
3879 case stagseen
: /* named struct or enum */
3880 case scolonseen
: /* a class */
3881 pushclass_above (bracelev
,token
.line
+token
.offset
, token
.length
);
3883 make_C_tag (false); /* a struct or enum */
3891 if (definedef
!= dnone
)
3893 if (fvdef
== fstartlist
)
3895 fvdef
= fvnone
; /* avoid tagging `foo' in `foo (*bar()) ()' */
3896 token
.valid
= false;
3900 if (definedef
!= dnone
)
3903 if (!ignoreindent
&& lp
== newlb
.buffer
+ 1)
3906 token
.valid
= false; /* unexpected value, token unreliable */
3907 bracelev
= 0; /* reset brace level if first column */
3908 parlev
= 0; /* also reset paren level, just in case... */
3910 else if (bracelev
< 0)
3912 token
.valid
= false; /* something gone amiss, token unreliable */
3915 if (bracelev
== 0 && fvdef
== vignore
)
3916 fvdef
= fvnone
; /* end of function */
3917 popclass_above (bracelev
);
3919 /* Only if typdef == tinbody is typdefbracelev significant. */
3920 if (typdef
== tinbody
&& bracelev
<= typdefbracelev
)
3922 assert (bracelev
== typdefbracelev
);
3927 if (definedef
!= dnone
)
3937 if ((members
&& bracelev
== 1)
3938 || (globals
&& bracelev
== 0 && (!fvextern
|| declarations
)))
3939 make_C_tag (false); /* a variable */
3947 && (structdef
== stagseen
|| fvdef
== fvnameseen
))
3954 if (templatelev
> 0)
3962 if (objdef
== oinbody
&& bracelev
== 0)
3964 objdef
= omethodsign
;
3969 case '#': case '~': case '&': case '%': case '/':
3970 case '|': case '^': case '!': case '.': case '?':
3971 if (definedef
!= dnone
)
3973 /* These surely cannot follow a function tag in C. */
3986 if (objdef
== otagseen
)
3988 make_C_tag (true); /* an Objective C class */
3991 /* If a macro spans multiple lines don't reset its state. */
3993 CNL_SAVE_DEFINEDEF ();
3999 } /* while not eof */
4001 free (lbs
[0].lb
.buffer
);
4002 free (lbs
[1].lb
.buffer
);
4006 * Process either a C++ file or a C file depending on the setting
4010 default_C_entries (FILE *inf
)
4012 C_entries (cplusplus
? C_PLPL
: C_AUTO
, inf
);
4015 /* Always do plain C. */
4017 plain_C_entries (FILE *inf
)
4022 /* Always do C++. */
4024 Cplusplus_entries (FILE *inf
)
4026 C_entries (C_PLPL
, inf
);
4029 /* Always do Java. */
4031 Cjava_entries (FILE *inf
)
4033 C_entries (C_JAVA
, inf
);
4038 Cstar_entries (FILE *inf
)
4040 C_entries (C_STAR
, inf
);
4043 /* Always do Yacc. */
4045 Yacc_entries (FILE *inf
)
4047 C_entries (YACC
, inf
);
4051 /* Useful macros. */
4052 #define LOOP_ON_INPUT_LINES(file_pointer, line_buffer, char_pointer) \
4053 while (perhaps_more_input (file_pointer) \
4054 && (readline (&(line_buffer), file_pointer), \
4055 (char_pointer) = (line_buffer).buffer, \
4058 #define LOOKING_AT(cp, kw) /* kw is the keyword, a literal string */ \
4059 ((assert ("" kw), true) /* syntax error if not a literal string */ \
4060 && strneq ((cp), kw, sizeof (kw)-1) /* cp points at kw */ \
4061 && notinname ((cp)[sizeof (kw)-1]) /* end of kw */ \
4062 && ((cp) = skip_spaces ((cp)+sizeof (kw)-1))) /* skip spaces */
4064 /* Similar to LOOKING_AT but does not use notinname, does not skip */
4065 #define LOOKING_AT_NOCASE(cp, kw) /* the keyword is a literal string */ \
4066 ((assert ("" kw), true) /* syntax error if not a literal string */ \
4067 && strncaseeq ((cp), kw, sizeof (kw)-1) /* cp points at kw */ \
4068 && ((cp) += sizeof (kw)-1)) /* skip spaces */
4071 * Read a file, but do no processing. This is used to do regexp
4072 * matching on files that have no language defined.
4075 just_read_file (FILE *inf
)
4077 while (perhaps_more_input (inf
))
4078 readline (&lb
, inf
);
4082 /* Fortran parsing */
4084 static void F_takeprec (void);
4085 static void F_getit (FILE *);
4090 dbp
= skip_spaces (dbp
);
4094 dbp
= skip_spaces (dbp
);
4095 if (strneq (dbp
, "(*)", 3))
4100 if (!c_isdigit (*dbp
))
4102 --dbp
; /* force failure */
4107 while (c_isdigit (*dbp
));
4115 dbp
= skip_spaces (dbp
);
4118 readline (&lb
, inf
);
4123 dbp
= skip_spaces (dbp
);
4125 if (!c_isalpha (*dbp
) && *dbp
!= '_' && *dbp
!= '$')
4127 for (cp
= dbp
+ 1; *cp
!= '\0' && intoken (*cp
); cp
++)
4129 make_tag (dbp
, cp
-dbp
, true,
4130 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4135 Fortran_functions (FILE *inf
)
4137 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
4140 dbp
++; /* Ratfor escape to fortran */
4141 dbp
= skip_spaces (dbp
);
4145 if (LOOKING_AT_NOCASE (dbp
, "recursive"))
4146 dbp
= skip_spaces (dbp
);
4148 if (LOOKING_AT_NOCASE (dbp
, "pure"))
4149 dbp
= skip_spaces (dbp
);
4151 if (LOOKING_AT_NOCASE (dbp
, "elemental"))
4152 dbp
= skip_spaces (dbp
);
4154 switch (c_tolower (*dbp
))
4157 if (nocase_tail ("integer"))
4161 if (nocase_tail ("real"))
4165 if (nocase_tail ("logical"))
4169 if (nocase_tail ("complex") || nocase_tail ("character"))
4173 if (nocase_tail ("double"))
4175 dbp
= skip_spaces (dbp
);
4178 if (nocase_tail ("precision"))
4184 dbp
= skip_spaces (dbp
);
4187 switch (c_tolower (*dbp
))
4190 if (nocase_tail ("function"))
4194 if (nocase_tail ("subroutine"))
4198 if (nocase_tail ("entry"))
4202 if (nocase_tail ("blockdata") || nocase_tail ("block data"))
4204 dbp
= skip_spaces (dbp
);
4205 if (*dbp
== '\0') /* assume un-named */
4206 make_tag ("blockdata", 9, true,
4207 lb
.buffer
, dbp
- lb
.buffer
, lineno
, linecharno
);
4209 F_getit (inf
); /* look for name */
4218 * Go language support
4219 * Original code by Xi Lu <lx@shellcodes.org> (2016)
4222 Go_functions(FILE *inf
)
4226 LOOP_ON_INPUT_LINES(inf
, lb
, cp
)
4228 cp
= skip_spaces (cp
);
4230 if (LOOKING_AT (cp
, "package"))
4233 while (!notinname (*cp
) && *cp
!= '\0')
4235 make_tag (name
, cp
- name
, false, lb
.buffer
,
4236 cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4238 else if (LOOKING_AT (cp
, "func"))
4240 /* Go implementation of interface, such as:
4241 func (n *Integer) Add(m Integer) ...
4242 skip `(n *Integer)` part.
4248 cp
= skip_spaces (cp
+1);
4255 while (!notinname (*cp
))
4258 make_tag (name
, cp
- name
, true, lb
.buffer
,
4259 cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4262 else if (members
&& LOOKING_AT (cp
, "type"))
4266 /* Ignore the likes of the following:
4274 while (!notinname (*cp
) && *cp
!= '\0')
4277 make_tag (name
, cp
- name
, false, lb
.buffer
,
4278 cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4287 * Philippe Waroquiers (1998)
4290 /* Once we are positioned after an "interesting" keyword, let's get
4291 the real tag value necessary. */
4293 Ada_getit (FILE *inf
, const char *name_qualifier
)
4299 while (perhaps_more_input (inf
))
4301 dbp
= skip_spaces (dbp
);
4303 || (dbp
[0] == '-' && dbp
[1] == '-'))
4305 readline (&lb
, inf
);
4308 switch (c_tolower (*dbp
))
4311 if (nocase_tail ("body"))
4313 /* Skipping body of procedure body or package body or ....
4314 resetting qualifier to body instead of spec. */
4315 name_qualifier
= "/b";
4320 /* Skipping type of task type or protected type ... */
4321 if (nocase_tail ("type"))
4328 for (cp
= dbp
; *cp
!= '\0' && *cp
!= '"'; cp
++)
4333 dbp
= skip_spaces (dbp
);
4335 c_isalnum (*cp
) || *cp
== '_' || *cp
== '.';
4343 name
= concat (dbp
, name_qualifier
, "");
4345 make_tag (name
, strlen (name
), true,
4346 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4355 Ada_funcs (FILE *inf
)
4357 bool inquote
= false;
4358 bool skip_till_semicolumn
= false;
4360 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
4362 while (*dbp
!= '\0')
4364 /* Skip a string i.e. "abcd". */
4365 if (inquote
|| (*dbp
== '"'))
4367 dbp
= strchr (dbp
+ !inquote
, '"');
4372 continue; /* advance char */
4377 break; /* advance line */
4381 /* Skip comments. */
4382 if (dbp
[0] == '-' && dbp
[1] == '-')
4383 break; /* advance line */
4385 /* Skip character enclosed in single quote i.e. 'a'
4386 and skip single quote starting an attribute i.e. 'Image. */
4395 if (skip_till_semicolumn
)
4398 skip_till_semicolumn
= false;
4400 continue; /* advance char */
4403 /* Search for beginning of a token. */
4404 if (!begtoken (*dbp
))
4407 continue; /* advance char */
4410 /* We are at the beginning of a token. */
4411 switch (c_tolower (*dbp
))
4414 if (!packages_only
&& nocase_tail ("function"))
4415 Ada_getit (inf
, "/f");
4417 break; /* from switch */
4418 continue; /* advance char */
4420 if (!packages_only
&& nocase_tail ("procedure"))
4421 Ada_getit (inf
, "/p");
4422 else if (nocase_tail ("package"))
4423 Ada_getit (inf
, "/s");
4424 else if (nocase_tail ("protected")) /* protected type */
4425 Ada_getit (inf
, "/t");
4427 break; /* from switch */
4428 continue; /* advance char */
4431 if (typedefs
&& !packages_only
&& nocase_tail ("use"))
4433 /* when tagging types, avoid tagging use type Pack.Typename;
4434 for this, we will skip everything till a ; */
4435 skip_till_semicolumn
= true;
4436 continue; /* advance char */
4440 if (!packages_only
&& nocase_tail ("task"))
4441 Ada_getit (inf
, "/k");
4442 else if (typedefs
&& !packages_only
&& nocase_tail ("type"))
4444 Ada_getit (inf
, "/t");
4445 while (*dbp
!= '\0')
4449 break; /* from switch */
4450 continue; /* advance char */
4453 /* Look for the end of the token. */
4454 while (!endtoken (*dbp
))
4457 } /* advance char */
4458 } /* advance line */
4463 * Unix and microcontroller assembly tag handling
4464 * Labels: /^[a-zA-Z_.$][a-zA_Z0-9_.$]*[: ^I^J]/
4465 * Idea by Bob Weiner, Motorola Inc. (1994)
4468 Asm_labels (FILE *inf
)
4472 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4474 /* If first char is alphabetic or one of [_.$], test for colon
4475 following identifier. */
4476 if (c_isalpha (*cp
) || *cp
== '_' || *cp
== '.' || *cp
== '$')
4478 /* Read past label. */
4480 while (c_isalnum (*cp
) || *cp
== '_' || *cp
== '.' || *cp
== '$')
4482 if (*cp
== ':' || c_isspace (*cp
))
4483 /* Found end of label, so copy it and add it to the table. */
4484 make_tag (lb
.buffer
, cp
- lb
.buffer
, true,
4485 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4493 * Perl sub names: /^sub[ \t\n]+[^ \t\n{]+/
4494 * /^use constant[ \t\n]+[^ \t\n{=,;]+/
4495 * Perl variable names: /^(my|local).../
4496 * Original code by Bart Robinson <lomew@cs.utah.edu> (1995)
4497 * Additions by Michael Ernst <mernst@alum.mit.edu> (1997)
4498 * Ideas by Kai Großjohann <Kai.Grossjohann@CS.Uni-Dortmund.DE> (2001)
4501 Perl_functions (FILE *inf
)
4503 char *package
= savestr ("main"); /* current package name */
4506 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4508 cp
= skip_spaces (cp
);
4510 if (LOOKING_AT (cp
, "package"))
4513 get_tag (cp
, &package
);
4515 else if (LOOKING_AT (cp
, "sub"))
4521 while (!notinname (*cp
))
4524 continue; /* nothing found */
4525 pos
= strchr (sp
, ':');
4526 if (pos
&& pos
< cp
&& pos
[1] == ':')
4527 /* The name is already qualified. */
4528 make_tag (sp
, cp
- sp
, true,
4529 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4533 char savechar
, *name
;
4537 name
= concat (package
, "::", sp
);
4539 make_tag (name
, strlen (name
), true,
4540 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4544 else if (LOOKING_AT (cp
, "use constant")
4545 || LOOKING_AT (cp
, "use constant::defer"))
4547 /* For hash style multi-constant like
4548 use constant { FOO => 123,
4550 only the first FOO is picked up. Parsing across the value
4551 expressions would be difficult in general, due to possible nested
4552 hashes, here-documents, etc. */
4554 cp
= skip_spaces (cp
+1);
4557 else if (globals
) /* only if we are tagging global vars */
4559 /* Skip a qualifier, if any. */
4560 bool qual
= LOOKING_AT (cp
, "my") || LOOKING_AT (cp
, "local");
4561 /* After "my" or "local", but before any following paren or space. */
4562 char *varstart
= cp
;
4564 if (qual
/* should this be removed? If yes, how? */
4565 && (*cp
== '$' || *cp
== '@' || *cp
== '%'))
4570 while (c_isalnum (*cp
) || *cp
== '_');
4574 /* Should be examining a variable list at this point;
4575 could insist on seeing an open parenthesis. */
4576 while (*cp
!= '\0' && *cp
!= ';' && *cp
!= '=' && *cp
!= ')')
4582 make_tag (varstart
, cp
- varstart
, false,
4583 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4592 * Look for /^[\t]*def[ \t\n]+[^ \t\n(:]+/ or /^class[ \t\n]+[^ \t\n(:]+/
4593 * Idea by Eric S. Raymond <esr@thyrsus.com> (1997)
4594 * More ideas by seb bacon <seb@jamkit.com> (2002)
4597 Python_functions (FILE *inf
)
4601 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4603 cp
= skip_spaces (cp
);
4604 if (LOOKING_AT (cp
, "def") || LOOKING_AT (cp
, "class"))
4607 while (!notinname (*cp
) && *cp
!= ':')
4609 make_tag (name
, cp
- name
, true,
4610 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4617 * Original code by Xi Lu <lx@shellcodes.org> (2015)
4620 Ruby_functions (FILE *inf
)
4624 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4626 bool is_class
= false;
4627 bool is_method
= false;
4630 cp
= skip_spaces (cp
);
4631 if (c_isalpha (*cp
) && c_isupper (*cp
)) /* constants */
4633 char *bp
, *colon
= NULL
;
4637 for (cp
++; c_isalnum (*cp
) || *cp
== '_' || *cp
== ':'; cp
++)
4644 bp
= skip_spaces (cp
);
4645 if (*bp
== '=' && !(bp
[1] == '=' || bp
[1] == '>'))
4647 if (colon
&& !c_isspace (colon
[1]))
4649 make_tag (name
, cp
- name
, false,
4650 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4654 else if ((is_method
= LOOKING_AT (cp
, "def")) /* module/class/method */
4655 || (is_class
= LOOKING_AT (cp
, "class"))
4656 || LOOKING_AT (cp
, "module"))
4658 const char self_name
[] = "self.";
4659 const size_t self_size1
= sizeof (self_name
) - 1;
4663 /* Ruby method names can end in a '='. Also, operator overloading can
4664 define operators whose names include '='. */
4665 while (!notinname (*cp
) || *cp
== '=')
4668 /* Remove "self." from the method name. */
4669 if (cp
- name
> self_size1
4670 && strneq (name
, self_name
, self_size1
))
4673 /* Remove the class/module qualifiers from method names. */
4678 for (q
= name
; q
< cp
&& *q
!= '.'; q
++)
4680 if (q
< cp
- 1) /* punt if we see just "FOO." */
4684 /* Don't tag singleton classes. */
4685 if (is_class
&& strneq (name
, "<<", 2) && cp
== name
+ 2)
4688 make_tag (name
, cp
- name
, true,
4689 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4693 /* Tag accessors and aliases. */
4694 while (*cp
&& *cp
!= '#')
4696 bool reader
= false, writer
= false, alias
= false;
4698 if (LOOKING_AT (cp
, "attr_reader"))
4700 else if (LOOKING_AT (cp
, "attr_writer"))
4702 else if (LOOKING_AT (cp
, "attr_accessor"))
4707 else if (LOOKING_AT (cp
, "alias_method"))
4709 if (reader
|| writer
|| alias
)
4714 cp
= skip_name (cp
);
4716 make_tag (np
, cp
- np
, true,
4717 lb
.buffer
, cp
- lb
.buffer
+ 1,
4718 lineno
, linecharno
);
4721 size_t name_len
= cp
- np
+ 1;
4722 char *wr_name
= xnew (name_len
+ 1, char);
4724 memcpy (wr_name
, np
, name_len
- 1);
4725 memcpy (wr_name
+ name_len
- 1, "=", 2);
4726 pfnote (wr_name
, true, lb
.buffer
, cp
- lb
.buffer
+ 1,
4727 lineno
, linecharno
);
4731 make_tag (np
, cp
- np
, true,
4732 lb
.buffer
, cp
- lb
.buffer
+ 1,
4733 lineno
, linecharno
);
4734 while (*cp
&& *cp
!= '#' && *cp
!= ';')
4738 && (cp
= skip_spaces (cp
+ 1), *cp
&& *cp
!= '#'));
4740 cp
= skip_name (cp
);
4741 while (*cp
&& *cp
!= '#' && notinname (*cp
))
4752 * - /^[ \t]*function[ \t\n]+[^ \t\n(]+/
4753 * - /^[ \t]*class[ \t\n]+[^ \t\n]+/
4754 * - /^[ \t]*define\(\"[^\"]+/
4755 * Only with --members:
4756 * - /^[ \t]*var[ \t\n]+\$[^ \t\n=;]/
4757 * Idea by Diez B. Roggisch (2001)
4760 PHP_functions (FILE *inf
)
4763 bool search_identifier
= false;
4765 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4767 cp
= skip_spaces (cp
);
4769 if (search_identifier
4772 while (!notinname (*cp
))
4774 make_tag (name
, cp
- name
, true,
4775 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4776 search_identifier
= false;
4778 else if (LOOKING_AT (cp
, "function"))
4781 cp
= skip_spaces (cp
+1);
4785 while (!notinname (*cp
))
4787 make_tag (name
, cp
- name
, true,
4788 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4791 search_identifier
= true;
4793 else if (LOOKING_AT (cp
, "class"))
4798 while (*cp
!= '\0' && !c_isspace (*cp
))
4800 make_tag (name
, cp
- name
, false,
4801 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4804 search_identifier
= true;
4806 else if (strneq (cp
, "define", 6)
4807 && (cp
= skip_spaces (cp
+6))
4809 && (*cp
== '"' || *cp
== '\''))
4813 while (*cp
!= quote
&& *cp
!= '\0')
4815 make_tag (name
, cp
- name
, false,
4816 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4819 && LOOKING_AT (cp
, "var")
4823 while (!notinname (*cp
))
4825 make_tag (name
, cp
- name
, false,
4826 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4833 * Cobol tag functions
4834 * We could look for anything that could be a paragraph name.
4835 * i.e. anything that starts in column 8 is one word and ends in a full stop.
4836 * Idea by Corny de Souza (1993)
4839 Cobol_paragraphs (FILE *inf
)
4841 register char *bp
, *ep
;
4843 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4849 /* If eoln, compiler option or comment ignore whole line. */
4850 if (bp
[-1] != ' ' || !c_isalnum (bp
[0]))
4853 for (ep
= bp
; c_isalnum (*ep
) || *ep
== '-'; ep
++)
4856 make_tag (bp
, ep
- bp
, true,
4857 lb
.buffer
, ep
- lb
.buffer
+ 1, lineno
, linecharno
);
4864 * Ideas by Assar Westerlund <assar@sics.se> (2001)
4867 Makefile_targets (FILE *inf
)
4871 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4873 if (*bp
== '\t' || *bp
== '#')
4875 while (*bp
!= '\0' && *bp
!= '=' && *bp
!= ':')
4877 if (*bp
== ':' || (globals
&& *bp
== '='))
4879 /* We should detect if there is more than one tag, but we do not.
4880 We just skip initial and final spaces. */
4881 char * namestart
= skip_spaces (lb
.buffer
);
4882 while (--bp
> namestart
)
4883 if (!notinname (*bp
))
4885 make_tag (namestart
, bp
- namestart
+ 1, true,
4886 lb
.buffer
, bp
- lb
.buffer
+ 2, lineno
, linecharno
);
4894 * Original code by Mosur K. Mohan (1989)
4896 * Locates tags for procedures & functions. Doesn't do any type- or
4897 * var-definitions. It does look for the keyword "extern" or
4898 * "forward" immediately following the procedure statement; if found,
4899 * the tag is skipped.
4902 Pascal_functions (FILE *inf
)
4904 linebuffer tline
; /* mostly copied from C_entries */
4906 int save_lineno
, namelen
, taglen
;
4909 bool /* each of these flags is true if: */
4910 incomment
, /* point is inside a comment */
4911 inquote
, /* point is inside '..' string */
4912 get_tagname
, /* point is after PROCEDURE/FUNCTION
4913 keyword, so next item = potential tag */
4914 found_tag
, /* point is after a potential tag */
4915 inparms
, /* point is within parameter-list */
4916 verify_tag
; /* point has passed the parm-list, so the
4917 next token will determine whether this
4918 is a FORWARD/EXTERN to be ignored, or
4919 whether it is a real tag */
4921 save_lcno
= save_lineno
= namelen
= taglen
= 0; /* keep compiler quiet */
4922 name
= NULL
; /* keep compiler quiet */
4925 linebuffer_init (&tline
);
4927 incomment
= inquote
= false;
4928 found_tag
= false; /* have a proc name; check if extern */
4929 get_tagname
= false; /* found "procedure" keyword */
4930 inparms
= false; /* found '(' after "proc" */
4931 verify_tag
= false; /* check if "extern" is ahead */
4934 while (perhaps_more_input (inf
)) /* long main loop to get next char */
4937 if (c
== '\0') /* if end of line */
4939 readline (&lb
, inf
);
4943 if (!((found_tag
&& verify_tag
)
4945 c
= *dbp
++; /* only if don't need *dbp pointing
4946 to the beginning of the name of
4947 the procedure or function */
4951 if (c
== '}') /* within { } comments */
4953 else if (c
== '*' && *dbp
== ')') /* within (* *) comments */
4970 inquote
= true; /* found first quote */
4972 case '{': /* found open { comment */
4976 if (*dbp
== '*') /* found open (* comment */
4981 else if (found_tag
) /* found '(' after tag, i.e., parm-list */
4984 case ')': /* end of parms list */
4989 if (found_tag
&& !inparms
) /* end of proc or fn stmt */
4996 if (found_tag
&& verify_tag
&& (*dbp
!= ' '))
4998 /* Check if this is an "extern" declaration. */
5001 if (c_tolower (*dbp
) == 'e')
5003 if (nocase_tail ("extern")) /* superfluous, really! */
5009 else if (c_tolower (*dbp
) == 'f')
5011 if (nocase_tail ("forward")) /* check for forward reference */
5017 if (found_tag
&& verify_tag
) /* not external proc, so make tag */
5021 make_tag (name
, namelen
, true,
5022 tline
.buffer
, taglen
, save_lineno
, save_lcno
);
5026 if (get_tagname
) /* grab name of proc or fn */
5033 /* Find block name. */
5034 for (cp
= dbp
+ 1; *cp
!= '\0' && !endtoken (*cp
); cp
++)
5037 /* Save all values for later tagging. */
5038 linebuffer_setlen (&tline
, lb
.len
);
5039 strcpy (tline
.buffer
, lb
.buffer
);
5040 save_lineno
= lineno
;
5041 save_lcno
= linecharno
;
5042 name
= tline
.buffer
+ (dbp
- lb
.buffer
);
5044 taglen
= cp
- lb
.buffer
+ 1;
5046 dbp
= cp
; /* set dbp to e-o-token */
5047 get_tagname
= false;
5051 /* And proceed to check for "extern". */
5053 else if (!incomment
&& !inquote
&& !found_tag
)
5055 /* Check for proc/fn keywords. */
5056 switch (c_tolower (c
))
5059 if (nocase_tail ("rocedure")) /* c = 'p', dbp has advanced */
5063 if (nocase_tail ("unction"))
5068 } /* while not eof */
5070 free (tline
.buffer
);
5075 * Lisp tag functions
5076 * look for (def or (DEF, quote or QUOTE
5079 static void L_getit (void);
5084 if (*dbp
== '\'') /* Skip prefix quote */
5086 else if (*dbp
== '(')
5089 /* Try to skip "(quote " */
5090 if (!LOOKING_AT (dbp
, "quote") && !LOOKING_AT (dbp
, "QUOTE"))
5091 /* Ok, then skip "(" before name in (defstruct (foo)) */
5092 dbp
= skip_spaces (dbp
);
5094 get_tag (dbp
, NULL
);
5098 Lisp_functions (FILE *inf
)
5100 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
5105 /* "(defvar foo)" is a declaration rather than a definition. */
5109 if (LOOKING_AT (p
, "defvar"))
5111 p
= skip_name (p
); /* past var name */
5112 p
= skip_spaces (p
);
5118 if (strneq (dbp
+ 1, "cl-", 3) || strneq (dbp
+ 1, "CL-", 3))
5121 if (strneq (dbp
+1, "def", 3) || strneq (dbp
+1, "DEF", 3))
5123 dbp
= skip_non_spaces (dbp
);
5124 dbp
= skip_spaces (dbp
);
5129 /* Check for (foo::defmumble name-defined ... */
5132 while (!notinname (*dbp
) && *dbp
!= ':');
5137 while (*dbp
== ':');
5139 if (strneq (dbp
, "def", 3) || strneq (dbp
, "DEF", 3))
5141 dbp
= skip_non_spaces (dbp
);
5142 dbp
= skip_spaces (dbp
);
5152 * Lua script language parsing
5153 * Original code by David A. Capello <dacap@users.sourceforge.net> (2004)
5155 * "function" and "local function" are tags if they start at column 1.
5158 Lua_functions (FILE *inf
)
5162 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
5164 bp
= skip_spaces (bp
);
5165 if (bp
[0] != 'f' && bp
[0] != 'l')
5168 (void)LOOKING_AT (bp
, "local"); /* skip possible "local" */
5170 if (LOOKING_AT (bp
, "function"))
5172 char *tag_name
, *tp_dot
, *tp_colon
;
5174 get_tag (bp
, &tag_name
);
5175 /* If the tag ends with ".foo" or ":foo", make an additional tag for
5177 tp_dot
= strrchr (tag_name
, '.');
5178 tp_colon
= strrchr (tag_name
, ':');
5179 if (tp_dot
|| tp_colon
)
5181 char *p
= tp_dot
> tp_colon
? tp_dot
: tp_colon
;
5182 int len_add
= p
- tag_name
+ 1;
5184 get_tag (bp
+ len_add
, NULL
);
5193 * Just look for lines where the first character is '/'
5194 * Also look at "defineps" for PSWrap
5196 * Richard Mlynarik <mly@adoc.xerox.com> (1997)
5197 * Masatake Yamato <masata-y@is.aist-nara.ac.jp> (1999)
5200 PS_functions (FILE *inf
)
5202 register char *bp
, *ep
;
5204 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
5209 *ep
!= '\0' && *ep
!= ' ' && *ep
!= '{';
5212 make_tag (bp
, ep
- bp
, true,
5213 lb
.buffer
, ep
- lb
.buffer
+ 1, lineno
, linecharno
);
5215 else if (LOOKING_AT (bp
, "defineps"))
5223 * Ignore anything after \ followed by space or in ( )
5224 * Look for words defined by :
5225 * Look for constant, code, create, defer, value, and variable
5226 * OBP extensions: Look for buffer:, field,
5227 * Ideas by Eduardo Horvath <eeh@netbsd.org> (2004)
5230 Forth_words (FILE *inf
)
5234 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
5235 while ((bp
= skip_spaces (bp
))[0] != '\0')
5236 if (bp
[0] == '\\' && c_isspace (bp
[1]))
5237 break; /* read next line */
5238 else if (bp
[0] == '(' && c_isspace (bp
[1]))
5239 do /* skip to ) or eol */
5241 while (*bp
!= ')' && *bp
!= '\0');
5242 else if ((bp
[0] == ':' && c_isspace (bp
[1]) && bp
++)
5243 || LOOKING_AT_NOCASE (bp
, "constant")
5244 || LOOKING_AT_NOCASE (bp
, "code")
5245 || LOOKING_AT_NOCASE (bp
, "create")
5246 || LOOKING_AT_NOCASE (bp
, "defer")
5247 || LOOKING_AT_NOCASE (bp
, "value")
5248 || LOOKING_AT_NOCASE (bp
, "variable")
5249 || LOOKING_AT_NOCASE (bp
, "buffer:")
5250 || LOOKING_AT_NOCASE (bp
, "field"))
5251 get_tag (skip_spaces (bp
), NULL
); /* Yay! A definition! */
5253 bp
= skip_non_spaces (bp
);
5258 * Scheme tag functions
5259 * look for (def... xyzzy
5261 * (def ... ((...(xyzzy ....
5263 * Original code by Ken Haase (1985?)
5266 Scheme_functions (FILE *inf
)
5270 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
5272 if (strneq (bp
, "(def", 4) || strneq (bp
, "(DEF", 4))
5274 bp
= skip_non_spaces (bp
+4);
5275 /* Skip over open parens and white space. Don't continue past
5277 while (*bp
&& notinname (*bp
))
5281 if (LOOKING_AT (bp
, "(SET!") || LOOKING_AT (bp
, "(set!"))
5287 /* Find tags in TeX and LaTeX input files. */
5289 /* TEX_toktab is a table of TeX control sequences that define tags.
5290 * Each entry records one such control sequence.
5292 * Original code from who knows whom.
5294 * Stefan Monnier (2002)
5297 static linebuffer
*TEX_toktab
= NULL
; /* Table with tag tokens */
5299 /* Default set of control sequences to put into TEX_toktab.
5300 The value of environment var TEXTAGS is prepended to this. */
5301 static const char *TEX_defenv
= "\
5302 :chapter:section:subsection:subsubsection:eqno:label:ref:cite:bibitem\
5303 :part:appendix:entry:index:def\
5304 :newcommand:renewcommand:newenvironment:renewenvironment";
5306 static void TEX_decode_env (const char *, const char *);
5309 * TeX/LaTeX scanning loop.
5312 TeX_commands (FILE *inf
)
5317 char TEX_esc
= '\0';
5318 char TEX_opgrp
, TEX_clgrp
;
5320 /* Initialize token table once from environment. */
5321 if (TEX_toktab
== NULL
)
5322 TEX_decode_env ("TEXTAGS", TEX_defenv
);
5324 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5326 /* Look at each TEX keyword in line. */
5329 /* Look for a TEX escape. */
5333 if (c
== '\0' || c
== '%')
5336 /* Select either \ or ! as escape character, whichever comes
5337 first outside a comment. */
5358 for (key
= TEX_toktab
; key
->buffer
!= NULL
; key
++)
5359 if (strneq (cp
, key
->buffer
, key
->len
))
5362 int namelen
, linelen
;
5365 cp
= skip_spaces (cp
+ key
->len
);
5366 if (*cp
== TEX_opgrp
)
5372 (!c_isspace (*p
) && *p
!= '#' &&
5373 *p
!= TEX_opgrp
&& *p
!= TEX_clgrp
);
5378 if (!opgrp
|| *p
== TEX_clgrp
)
5380 while (*p
!= '\0' && *p
!= TEX_opgrp
&& *p
!= TEX_clgrp
)
5382 linelen
= p
- lb
.buffer
+ 1;
5384 make_tag (cp
, namelen
, true,
5385 lb
.buffer
, linelen
, lineno
, linecharno
);
5386 goto tex_next_line
; /* We only tag a line once */
5394 /* Read environment and prepend it to the default string.
5395 Build token table. */
5397 TEX_decode_env (const char *evarname
, const char *defenv
)
5399 register const char *env
, *p
;
5402 /* Append default string to environment. */
5403 env
= getenv (evarname
);
5407 env
= concat (env
, defenv
, "");
5409 /* Allocate a token table */
5410 for (len
= 1, p
= env
; (p
= strchr (p
, ':')); )
5413 TEX_toktab
= xnew (len
, linebuffer
);
5415 /* Unpack environment string into token table. Be careful about */
5416 /* zero-length strings (leading ':', "::" and trailing ':') */
5417 for (i
= 0; *env
!= '\0';)
5419 p
= strchr (env
, ':');
5420 if (!p
) /* End of environment string. */
5421 p
= env
+ strlen (env
);
5423 { /* Only non-zero strings. */
5424 TEX_toktab
[i
].buffer
= savenstr (env
, p
- env
);
5425 TEX_toktab
[i
].len
= p
- env
;
5432 TEX_toktab
[i
].buffer
= NULL
; /* Mark end of table. */
5433 TEX_toktab
[i
].len
= 0;
5440 /* Texinfo support. Dave Love, Mar. 2000. */
5442 Texinfo_nodes (FILE *inf
)
5445 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5446 if (LOOKING_AT (cp
, "@node"))
5449 while (*cp
!= '\0' && *cp
!= ',')
5451 make_tag (start
, cp
- start
, true,
5452 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
5459 * Contents of <title>, <h1>, <h2>, <h3> are tags.
5460 * Contents of <a name=xxx> are tags with name xxx.
5462 * Francesco Potortì, 2002.
5465 HTML_labels (FILE *inf
)
5467 bool getnext
= false; /* next text outside of HTML tags is a tag */
5468 bool skiptag
= false; /* skip to the end of the current HTML tag */
5469 bool intag
= false; /* inside an html tag, looking for ID= */
5470 bool inanchor
= false; /* when INTAG, is an anchor, look for NAME= */
5474 linebuffer_setlen (&token_name
, 0); /* no name in buffer */
5476 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
5477 for (;;) /* loop on the same line */
5479 if (skiptag
) /* skip HTML tag */
5481 while (*dbp
!= '\0' && *dbp
!= '>')
5487 continue; /* look on the same line */
5489 break; /* go to next line */
5492 else if (intag
) /* look for "name=" or "id=" */
5494 while (*dbp
!= '\0' && *dbp
!= '>'
5495 && c_tolower (*dbp
) != 'n' && c_tolower (*dbp
) != 'i')
5498 break; /* go to next line */
5503 continue; /* look on the same line */
5505 if ((inanchor
&& LOOKING_AT_NOCASE (dbp
, "name="))
5506 || LOOKING_AT_NOCASE (dbp
, "id="))
5508 bool quoted
= (dbp
[0] == '"');
5511 for (end
= ++dbp
; *end
!= '\0' && *end
!= '"'; end
++)
5514 for (end
= dbp
; *end
!= '\0' && intoken (*end
); end
++)
5516 linebuffer_setlen (&token_name
, end
- dbp
);
5517 memcpy (token_name
.buffer
, dbp
, end
- dbp
);
5518 token_name
.buffer
[end
- dbp
] = '\0';
5521 intag
= false; /* we found what we looked for */
5522 skiptag
= true; /* skip to the end of the tag */
5523 getnext
= true; /* then grab the text */
5524 continue; /* look on the same line */
5529 else if (getnext
) /* grab next tokens and tag them */
5531 dbp
= skip_spaces (dbp
);
5533 break; /* go to next line */
5537 inanchor
= (c_tolower (dbp
[1]) == 'a' && !intoken (dbp
[2]));
5538 continue; /* look on the same line */
5541 for (end
= dbp
+ 1; *end
!= '\0' && *end
!= '<'; end
++)
5543 make_tag (token_name
.buffer
, token_name
.len
, true,
5544 dbp
, end
- dbp
, lineno
, linecharno
);
5545 linebuffer_setlen (&token_name
, 0); /* no name in buffer */
5547 break; /* go to next line */
5550 else /* look for an interesting HTML tag */
5552 while (*dbp
!= '\0' && *dbp
!= '<')
5555 break; /* go to next line */
5557 if (c_tolower (dbp
[1]) == 'a' && !intoken (dbp
[2]))
5560 continue; /* look on the same line */
5562 else if (LOOKING_AT_NOCASE (dbp
, "<title>")
5563 || LOOKING_AT_NOCASE (dbp
, "<h1>")
5564 || LOOKING_AT_NOCASE (dbp
, "<h2>")
5565 || LOOKING_AT_NOCASE (dbp
, "<h3>"))
5569 continue; /* look on the same line */
5580 * Assumes that the predicate or rule starts at column 0.
5581 * Only the first clause of a predicate or rule is added.
5582 * Original code by Sunichirou Sugou (1989)
5583 * Rewritten by Anders Lindgren (1996)
5585 static size_t prolog_pr (char *, char *);
5586 static void prolog_skip_comment (linebuffer
*, FILE *);
5587 static size_t prolog_atom (char *, size_t);
5590 Prolog_functions (FILE *inf
)
5600 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5602 if (cp
[0] == '\0') /* Empty line */
5604 else if (c_isspace (cp
[0])) /* Not a predicate */
5606 else if (cp
[0] == '/' && cp
[1] == '*') /* comment. */
5607 prolog_skip_comment (&lb
, inf
);
5608 else if ((len
= prolog_pr (cp
, last
)) > 0)
5610 /* Predicate or rule. Store the function name so that we
5611 only generate a tag for the first clause. */
5613 last
= xnew (len
+ 1, char);
5614 else if (len
+ 1 > allocated
)
5615 xrnew (last
, len
+ 1, char);
5616 allocated
= len
+ 1;
5617 memcpy (last
, cp
, len
);
5626 prolog_skip_comment (linebuffer
*plb
, FILE *inf
)
5632 for (cp
= plb
->buffer
; *cp
!= '\0'; cp
++)
5633 if (cp
[0] == '*' && cp
[1] == '/')
5635 readline (plb
, inf
);
5637 while (perhaps_more_input (inf
));
5641 * A predicate or rule definition is added if it matches:
5642 * <beginning of line><Prolog Atom><whitespace>(
5643 * or <beginning of line><Prolog Atom><whitespace>:-
5645 * It is added to the tags database if it doesn't match the
5646 * name of the previous clause header.
5648 * Return the size of the name of the predicate or rule, or 0 if no
5652 prolog_pr (char *s
, char *last
)
5654 /* Name of last clause. */
5659 pos
= prolog_atom (s
, 0);
5664 pos
= skip_spaces (s
+ pos
) - s
;
5667 || (s
[pos
] == '(' && (pos
+= 1))
5668 || (s
[pos
] == ':' && s
[pos
+ 1] == '-' && (pos
+= 2)))
5669 && (last
== NULL
/* save only the first clause */
5670 || len
!= strlen (last
)
5671 || !strneq (s
, last
, len
)))
5673 make_tag (s
, len
, true, s
, pos
, lineno
, linecharno
);
5681 * Consume a Prolog atom.
5682 * Return the number of bytes consumed, or 0 if there was an error.
5684 * A prolog atom, in this context, could be one of:
5685 * - An alphanumeric sequence, starting with a lower case letter.
5686 * - A quoted arbitrary string. Single quotes can escape themselves.
5687 * Backslash quotes everything.
5690 prolog_atom (char *s
, size_t pos
)
5696 if (c_islower (s
[pos
]) || s
[pos
] == '_')
5698 /* The atom is unquoted. */
5700 while (c_isalnum (s
[pos
]) || s
[pos
] == '_')
5704 return pos
- origpos
;
5706 else if (s
[pos
] == '\'')
5717 pos
++; /* A double quote */
5719 else if (s
[pos
] == '\0')
5720 /* Multiline quoted atoms are ignored. */
5722 else if (s
[pos
] == '\\')
5724 if (s
[pos
+1] == '\0')
5731 return pos
- origpos
;
5739 * Support for Erlang
5741 * Generates tags for functions, defines, and records.
5742 * Assumes that Erlang functions start at column 0.
5743 * Original code by Anders Lindgren (1996)
5745 static int erlang_func (char *, char *);
5746 static void erlang_attribute (char *);
5747 static int erlang_atom (char *);
5750 Erlang_functions (FILE *inf
)
5760 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5762 if (cp
[0] == '\0') /* Empty line */
5764 else if (c_isspace (cp
[0])) /* Not function nor attribute */
5766 else if (cp
[0] == '%') /* comment */
5768 else if (cp
[0] == '"') /* Sometimes, strings start in column one */
5770 else if (cp
[0] == '-') /* attribute, e.g. "-define" */
5772 erlang_attribute (cp
);
5779 else if ((len
= erlang_func (cp
, last
)) > 0)
5782 * Function. Store the function name so that we only
5783 * generates a tag for the first clause.
5786 last
= xnew (len
+ 1, char);
5787 else if (len
+ 1 > allocated
)
5788 xrnew (last
, len
+ 1, char);
5789 allocated
= len
+ 1;
5790 memcpy (last
, cp
, len
);
5799 * A function definition is added if it matches:
5800 * <beginning of line><Erlang Atom><whitespace>(
5802 * It is added to the tags database if it doesn't match the
5803 * name of the previous clause header.
5805 * Return the size of the name of the function, or 0 if no function
5809 erlang_func (char *s
, char *last
)
5811 /* Name of last clause. */
5816 pos
= erlang_atom (s
);
5821 pos
= skip_spaces (s
+ pos
) - s
;
5823 /* Save only the first clause. */
5826 || len
!= (int)strlen (last
)
5827 || !strneq (s
, last
, len
)))
5829 make_tag (s
, len
, true, s
, pos
, lineno
, linecharno
);
5838 * Handle attributes. Currently, tags are generated for defines
5841 * They are on the form:
5842 * -define(foo, bar).
5843 * -define(Foo(M, N), M+N).
5844 * -record(graph, {vtab = notable, cyclic = true}).
5847 erlang_attribute (char *s
)
5851 if ((LOOKING_AT (cp
, "-define") || LOOKING_AT (cp
, "-record"))
5854 int len
= erlang_atom (skip_spaces (cp
));
5856 make_tag (cp
, len
, true, s
, cp
+ len
- s
, lineno
, linecharno
);
5863 * Consume an Erlang atom (or variable).
5864 * Return the number of bytes consumed, or -1 if there was an error.
5867 erlang_atom (char *s
)
5871 if (c_isalpha (s
[pos
]) || s
[pos
] == '_')
5873 /* The atom is unquoted. */
5876 while (c_isalnum (s
[pos
]) || s
[pos
] == '_');
5878 else if (s
[pos
] == '\'')
5880 for (pos
++; s
[pos
] != '\''; pos
++)
5881 if (s
[pos
] == '\0' /* multiline quoted atoms are ignored */
5882 || (s
[pos
] == '\\' && s
[++pos
] == '\0'))
5891 static char *scan_separators (char *);
5892 static void add_regex (char *, language
*);
5893 static char *substitute (char *, char *, struct re_registers
*);
5896 * Take a string like "/blah/" and turn it into "blah", verifying
5897 * that the first and last characters are the same, and handling
5898 * quoted separator characters. Actually, stops on the occurrence of
5899 * an unquoted separator. Also process \t, \n, etc. and turn into
5900 * appropriate characters. Works in place. Null terminates name string.
5901 * Returns pointer to terminating separator, or NULL for
5902 * unterminated regexps.
5905 scan_separators (char *name
)
5908 char *copyto
= name
;
5909 bool quoted
= false;
5911 for (++name
; *name
!= '\0'; ++name
)
5917 case 'a': *copyto
++ = '\007'; break; /* BEL (bell) */
5918 case 'b': *copyto
++ = '\b'; break; /* BS (back space) */
5919 case 'd': *copyto
++ = 0177; break; /* DEL (delete) */
5920 case 'e': *copyto
++ = 033; break; /* ESC (delete) */
5921 case 'f': *copyto
++ = '\f'; break; /* FF (form feed) */
5922 case 'n': *copyto
++ = '\n'; break; /* NL (new line) */
5923 case 'r': *copyto
++ = '\r'; break; /* CR (carriage return) */
5924 case 't': *copyto
++ = '\t'; break; /* TAB (horizontal tab) */
5925 case 'v': *copyto
++ = '\v'; break; /* VT (vertical tab) */
5931 /* Something else is quoted, so preserve the quote. */
5939 else if (*name
== '\\')
5941 else if (*name
== sep
)
5947 name
= NULL
; /* signal unterminated regexp */
5949 /* Terminate copied string. */
5954 /* Look at the argument of --regex or --no-regex and do the right
5955 thing. Same for each line of a regexp file. */
5957 analyze_regex (char *regex_arg
)
5959 if (regex_arg
== NULL
)
5961 free_regexps (); /* --no-regex: remove existing regexps */
5965 /* A real --regexp option or a line in a regexp file. */
5966 switch (regex_arg
[0])
5968 /* Comments in regexp file or null arg to --regex. */
5974 /* Read a regex file. This is recursive and may result in a
5975 loop, which will stop when the file descriptors are exhausted. */
5979 linebuffer regexbuf
;
5980 char *regexfile
= regex_arg
+ 1;
5982 /* regexfile is a file containing regexps, one per line. */
5983 regexfp
= fopen (regexfile
, "r" FOPEN_BINARY
);
5984 if (regexfp
== NULL
)
5986 linebuffer_init (®exbuf
);
5987 while (readline_internal (®exbuf
, regexfp
, regexfile
) > 0)
5988 analyze_regex (regexbuf
.buffer
);
5989 free (regexbuf
.buffer
);
5990 if (fclose (regexfp
) != 0)
5995 /* Regexp to be used for a specific language only. */
5999 char *lang_name
= regex_arg
+ 1;
6002 for (cp
= lang_name
; *cp
!= '}'; cp
++)
6005 error ("unterminated language name in regex: %s", regex_arg
);
6009 lang
= get_language_from_langname (lang_name
);
6012 add_regex (cp
, lang
);
6016 /* Regexp to be used for any language. */
6018 add_regex (regex_arg
, NULL
);
6023 /* Separate the regexp pattern, compile it,
6024 and care for optional name and modifiers. */
6026 add_regex (char *regexp_pattern
, language
*lang
)
6028 static struct re_pattern_buffer zeropattern
;
6029 char sep
, *pat
, *name
, *modifiers
;
6032 struct re_pattern_buffer
*patbuf
;
6035 force_explicit_name
= true, /* do not use implicit tag names */
6036 ignore_case
= false, /* case is significant */
6037 multi_line
= false, /* matches are done one line at a time */
6038 single_line
= false; /* dot does not match newline */
6041 if (strlen (regexp_pattern
) < 3)
6043 error ("null regexp");
6046 sep
= regexp_pattern
[0];
6047 name
= scan_separators (regexp_pattern
);
6050 error ("%s: unterminated regexp", regexp_pattern
);
6055 error ("null name for regexp \"%s\"", regexp_pattern
);
6058 modifiers
= scan_separators (name
);
6059 if (modifiers
== NULL
) /* no terminating separator --> no name */
6065 modifiers
+= 1; /* skip separator */
6067 /* Parse regex modifiers. */
6068 for (; modifiers
[0] != '\0'; modifiers
++)
6069 switch (modifiers
[0])
6072 if (modifiers
== name
)
6073 error ("forcing explicit tag name but no name, ignoring");
6074 force_explicit_name
= true;
6084 need_filebuf
= true;
6087 error ("invalid regexp modifier '%c', ignoring", modifiers
[0]);
6091 patbuf
= xnew (1, struct re_pattern_buffer
);
6092 *patbuf
= zeropattern
;
6095 static char lc_trans
[UCHAR_MAX
+ 1];
6097 for (i
= 0; i
< UCHAR_MAX
+ 1; i
++)
6098 lc_trans
[i
] = c_tolower (i
);
6099 patbuf
->translate
= lc_trans
; /* translation table to fold case */
6103 pat
= concat ("^", regexp_pattern
, ""); /* anchor to beginning of line */
6105 pat
= regexp_pattern
;
6108 re_set_syntax (RE_SYNTAX_EMACS
| RE_DOT_NEWLINE
);
6110 re_set_syntax (RE_SYNTAX_EMACS
);
6112 err
= re_compile_pattern (pat
, strlen (pat
), patbuf
);
6117 error ("%s while compiling pattern", err
);
6122 p_head
= xnew (1, regexp
);
6123 p_head
->pattern
= savestr (regexp_pattern
);
6124 p_head
->p_next
= rp
;
6125 p_head
->lang
= lang
;
6126 p_head
->pat
= patbuf
;
6127 p_head
->name
= savestr (name
);
6128 p_head
->error_signaled
= false;
6129 p_head
->force_explicit_name
= force_explicit_name
;
6130 p_head
->ignore_case
= ignore_case
;
6131 p_head
->multi_line
= multi_line
;
6135 * Do the substitutions indicated by the regular expression and
6139 substitute (char *in
, char *out
, struct re_registers
*regs
)
6142 int size
, dig
, diglen
;
6145 size
= strlen (out
);
6147 /* Pass 1: figure out how much to allocate by finding all \N strings. */
6148 if (out
[size
- 1] == '\\')
6149 fatal ("pattern error in \"%s\"", out
);
6150 for (t
= strchr (out
, '\\');
6152 t
= strchr (t
+ 2, '\\'))
6153 if (c_isdigit (t
[1]))
6156 diglen
= regs
->end
[dig
] - regs
->start
[dig
];
6162 /* Allocate space and do the substitutions. */
6164 result
= xnew (size
+ 1, char);
6166 for (t
= result
; *out
!= '\0'; out
++)
6167 if (*out
== '\\' && c_isdigit (*++out
))
6170 diglen
= regs
->end
[dig
] - regs
->start
[dig
];
6171 memcpy (t
, in
+ regs
->start
[dig
], diglen
);
6178 assert (t
<= result
+ size
);
6179 assert (t
- result
== (int)strlen (result
));
6184 /* Deallocate all regexps. */
6189 while (p_head
!= NULL
)
6191 rp
= p_head
->p_next
;
6192 free (p_head
->pattern
);
6193 free (p_head
->name
);
6201 * Reads the whole file as a single string from `filebuf' and looks for
6202 * multi-line regular expressions, creating tags on matches.
6203 * readline already dealt with normal regexps.
6205 * Idea by Ben Wing <ben@666.com> (2002).
6208 regex_tag_multiline (void)
6210 char *buffer
= filebuf
.buffer
;
6214 for (rp
= p_head
; rp
!= NULL
; rp
= rp
->p_next
)
6218 if (!rp
->multi_line
)
6219 continue; /* skip normal regexps */
6221 /* Generic initializations before parsing file from memory. */
6222 lineno
= 1; /* reset global line number */
6223 charno
= 0; /* reset global char number */
6224 linecharno
= 0; /* reset global char number of line start */
6226 /* Only use generic regexps or those for the current language. */
6227 if (rp
->lang
!= NULL
&& rp
->lang
!= curfdp
->lang
)
6230 while (match
>= 0 && match
< filebuf
.len
)
6232 match
= re_search (rp
->pat
, buffer
, filebuf
.len
, charno
,
6233 filebuf
.len
- match
, &rp
->regs
);
6238 if (!rp
->error_signaled
)
6240 error ("regexp stack overflow while matching \"%s\"",
6242 rp
->error_signaled
= true;
6249 if (match
== rp
->regs
.end
[0])
6251 if (!rp
->error_signaled
)
6253 error ("regexp matches the empty string: \"%s\"",
6255 rp
->error_signaled
= true;
6257 match
= -3; /* exit from while loop */
6261 /* Match occurred. Construct a tag. */
6262 while (charno
< rp
->regs
.end
[0])
6263 if (buffer
[charno
++] == '\n')
6264 lineno
++, linecharno
= charno
;
6266 if (name
[0] == '\0')
6268 else /* make a named tag */
6269 name
= substitute (buffer
, rp
->name
, &rp
->regs
);
6270 if (rp
->force_explicit_name
)
6271 /* Force explicit tag name, if a name is there. */
6272 pfnote (name
, true, buffer
+ linecharno
,
6273 charno
- linecharno
+ 1, lineno
, linecharno
);
6275 make_tag (name
, strlen (name
), true, buffer
+ linecharno
,
6276 charno
- linecharno
+ 1, lineno
, linecharno
);
6285 nocase_tail (const char *cp
)
6289 while (*cp
!= '\0' && c_tolower (*cp
) == c_tolower (dbp
[len
]))
6291 if (*cp
== '\0' && !intoken (dbp
[len
]))
6300 get_tag (register char *bp
, char **namepp
)
6302 register char *cp
= bp
;
6306 /* Go till you get to white space or a syntactic break */
6307 for (cp
= bp
+ 1; !notinname (*cp
); cp
++)
6309 make_tag (bp
, cp
- bp
, true,
6310 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
6314 *namepp
= savenstr (bp
, cp
- bp
);
6318 * Read a line of text from `stream' into `lbp', excluding the
6319 * newline or CR-NL, if any. Return the number of characters read from
6320 * `stream', which is the length of the line including the newline.
6322 * On DOS or Windows we do not count the CR character, if any before the
6323 * NL, in the returned length; this mirrors the behavior of Emacs on those
6324 * platforms (for text files, it translates CR-NL to NL as it reads in the
6327 * If multi-line regular expressions are requested, each line read is
6328 * appended to `filebuf'.
6331 readline_internal (linebuffer
*lbp
, FILE *stream
, char const *filename
)
6333 char *buffer
= lbp
->buffer
;
6334 char *p
= lbp
->buffer
;
6338 pend
= p
+ lbp
->size
; /* Separate to avoid 386/IX compiler bug. */
6342 register int c
= getc (stream
);
6345 /* We're at the end of linebuffer: expand it. */
6347 xrnew (buffer
, lbp
->size
, char);
6348 p
+= buffer
- lbp
->buffer
;
6349 pend
= buffer
+ lbp
->size
;
6350 lbp
->buffer
= buffer
;
6354 if (ferror (stream
))
6362 if (p
> buffer
&& p
[-1] == '\r')
6376 lbp
->len
= p
- buffer
;
6378 if (need_filebuf
/* we need filebuf for multi-line regexps */
6379 && chars_deleted
> 0) /* not at EOF */
6381 while (filebuf
.size
<= filebuf
.len
+ lbp
->len
+ 1) /* +1 for \n */
6383 /* Expand filebuf. */
6385 xrnew (filebuf
.buffer
, filebuf
.size
, char);
6387 memcpy (filebuf
.buffer
+ filebuf
.len
, lbp
->buffer
, lbp
->len
);
6388 filebuf
.len
+= lbp
->len
;
6389 filebuf
.buffer
[filebuf
.len
++] = '\n';
6390 filebuf
.buffer
[filebuf
.len
] = '\0';
6393 return lbp
->len
+ chars_deleted
;
6397 * Like readline_internal, above, but in addition try to match the
6398 * input line against relevant regular expressions and manage #line
6402 readline (linebuffer
*lbp
, FILE *stream
)
6406 linecharno
= charno
; /* update global char number of line start */
6407 result
= readline_internal (lbp
, stream
, infilename
); /* read line */
6408 lineno
+= 1; /* increment global line number */
6409 charno
+= result
; /* increment global char number */
6411 /* Honor #line directives. */
6412 if (!no_line_directive
)
6414 static bool discard_until_line_directive
;
6416 /* Check whether this is a #line directive. */
6417 if (result
> 12 && strneq (lbp
->buffer
, "#line ", 6))
6422 if (sscanf (lbp
->buffer
, "#line %u \"%n", &lno
, &start
) >= 1
6423 && start
> 0) /* double quote character found */
6425 char *endp
= lbp
->buffer
+ start
;
6427 while ((endp
= strchr (endp
, '"')) != NULL
6428 && endp
[-1] == '\\')
6431 /* Ok, this is a real #line directive. Let's deal with it. */
6433 char *taggedabsname
; /* absolute name of original file */
6434 char *taggedfname
; /* name of original file as given */
6435 char *name
; /* temp var */
6437 discard_until_line_directive
= false; /* found it */
6438 name
= lbp
->buffer
+ start
;
6440 canonicalize_filename (name
);
6441 taggedabsname
= absolute_filename (name
, tagfiledir
);
6442 if (filename_is_absolute (name
)
6443 || filename_is_absolute (curfdp
->infname
))
6444 taggedfname
= savestr (taggedabsname
);
6446 taggedfname
= relative_filename (taggedabsname
,tagfiledir
);
6448 if (streq (curfdp
->taggedfname
, taggedfname
))
6449 /* The #line directive is only a line number change. We
6450 deal with this afterwards. */
6453 /* The tags following this #line directive should be
6454 attributed to taggedfname. In order to do this, set
6455 curfdp accordingly. */
6457 fdesc
*fdp
; /* file description pointer */
6459 /* Go look for a file description already set up for the
6460 file indicated in the #line directive. If there is
6461 one, use it from now until the next #line
6463 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
6464 if (streq (fdp
->infname
, curfdp
->infname
)
6465 && streq (fdp
->taggedfname
, taggedfname
))
6466 /* If we remove the second test above (after the &&)
6467 then all entries pertaining to the same file are
6468 coalesced in the tags file. If we use it, then
6469 entries pertaining to the same file but generated
6470 from different files (via #line directives) will
6471 go into separate sections in the tags file. These
6472 alternatives look equivalent. The first one
6473 destroys some apparently useless information. */
6479 /* Else, if we already tagged the real file, skip all
6480 input lines until the next #line directive. */
6481 if (fdp
== NULL
) /* not found */
6482 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
6483 if (streq (fdp
->infabsname
, taggedabsname
))
6485 discard_until_line_directive
= true;
6489 /* Else create a new file description and use that from
6490 now on, until the next #line directive. */
6491 if (fdp
== NULL
) /* not found */
6494 fdhead
= xnew (1, fdesc
);
6495 *fdhead
= *curfdp
; /* copy curr. file description */
6497 fdhead
->infname
= savestr (curfdp
->infname
);
6498 fdhead
->infabsname
= savestr (curfdp
->infabsname
);
6499 fdhead
->infabsdir
= savestr (curfdp
->infabsdir
);
6500 fdhead
->taggedfname
= taggedfname
;
6501 fdhead
->usecharno
= false;
6502 fdhead
->prop
= NULL
;
6503 fdhead
->written
= false;
6507 free (taggedabsname
);
6509 readline (lbp
, stream
);
6511 } /* if a real #line directive */
6512 } /* if #line is followed by a number */
6513 } /* if line begins with "#line " */
6515 /* If we are here, no #line directive was found. */
6516 if (discard_until_line_directive
)
6520 /* Do a tail recursion on ourselves, thus discarding the contents
6521 of the line buffer. */
6522 readline (lbp
, stream
);
6526 discard_until_line_directive
= false;
6529 } /* if #line directives should be considered */
6536 /* Match against relevant regexps. */
6538 for (rp
= p_head
; rp
!= NULL
; rp
= rp
->p_next
)
6540 /* Only use generic regexps or those for the current language.
6541 Also do not use multiline regexps, which is the job of
6542 regex_tag_multiline. */
6543 if ((rp
->lang
!= NULL
&& rp
->lang
!= fdhead
->lang
)
6547 match
= re_match (rp
->pat
, lbp
->buffer
, lbp
->len
, 0, &rp
->regs
);
6552 if (!rp
->error_signaled
)
6554 error ("regexp stack overflow while matching \"%s\"",
6556 rp
->error_signaled
= true;
6563 /* Empty string matched. */
6564 if (!rp
->error_signaled
)
6566 error ("regexp matches the empty string: \"%s\"", rp
->pattern
);
6567 rp
->error_signaled
= true;
6571 /* Match occurred. Construct a tag. */
6573 if (name
[0] == '\0')
6575 else /* make a named tag */
6576 name
= substitute (lbp
->buffer
, rp
->name
, &rp
->regs
);
6577 if (rp
->force_explicit_name
)
6578 /* Force explicit tag name, if a name is there. */
6579 pfnote (name
, true, lbp
->buffer
, match
, lineno
, linecharno
);
6581 make_tag (name
, strlen (name
), true,
6582 lbp
->buffer
, match
, lineno
, linecharno
);
6591 * Return a pointer to a space of size strlen(cp)+1 allocated
6592 * with xnew where the string CP has been copied.
6595 savestr (const char *cp
)
6597 return savenstr (cp
, strlen (cp
));
6601 * Return a pointer to a space of size LEN+1 allocated with xnew where
6602 * the string CP has been copied for at most the first LEN characters.
6605 savenstr (const char *cp
, int len
)
6607 char *dp
= xnew (len
+ 1, char);
6609 return memcpy (dp
, cp
, len
);
6612 /* Skip spaces (end of string is not space), return new pointer. */
6614 skip_spaces (char *cp
)
6616 while (c_isspace (*cp
))
6621 /* Skip non spaces, except end of string, return new pointer. */
6623 skip_non_spaces (char *cp
)
6625 while (*cp
!= '\0' && !c_isspace (*cp
))
6630 /* Skip any chars in the "name" class.*/
6632 skip_name (char *cp
)
6634 /* '\0' is a notinname() so loop stops there too */
6635 while (! notinname (*cp
))
6640 /* Print error message and exit. */
6642 fatal (char const *format
, ...)
6645 va_start (ap
, format
);
6646 verror (format
, ap
);
6648 exit (EXIT_FAILURE
);
6652 pfatal (const char *s1
)
6655 exit (EXIT_FAILURE
);
6659 suggest_asking_for_help (void)
6661 fprintf (stderr
, "\tTry '%s --help' for a complete list of options.\n",
6663 exit (EXIT_FAILURE
);
6666 /* Output a diagnostic with printf-style FORMAT and args. */
6668 error (const char *format
, ...)
6671 va_start (ap
, format
);
6672 verror (format
, ap
);
6677 verror (char const *format
, va_list ap
)
6679 fprintf (stderr
, "%s: ", progname
);
6680 vfprintf (stderr
, format
, ap
);
6681 fprintf (stderr
, "\n");
6684 /* Return a newly-allocated string whose contents
6685 concatenate those of s1, s2, s3. */
6687 concat (const char *s1
, const char *s2
, const char *s3
)
6689 int len1
= strlen (s1
), len2
= strlen (s2
), len3
= strlen (s3
);
6690 char *result
= xnew (len1
+ len2
+ len3
+ 1, char);
6692 strcpy (result
, s1
);
6693 strcpy (result
+ len1
, s2
);
6694 strcpy (result
+ len1
+ len2
, s3
);
6700 /* Does the same work as the system V getcwd, but does not need to
6701 guess the buffer size in advance. */
6706 char *path
= xnew (bufsize
, char);
6708 while (getcwd (path
, bufsize
) == NULL
)
6710 if (errno
!= ERANGE
)
6714 path
= xnew (bufsize
, char);
6717 canonicalize_filename (path
);
6721 /* Return a newly allocated string containing a name of a temporary file. */
6725 const char *tmpdir
= getenv ("TMPDIR");
6726 const char *slash
= "/";
6728 #if MSDOS || defined (DOS_NT)
6730 tmpdir
= getenv ("TEMP");
6732 tmpdir
= getenv ("TMP");
6735 if (tmpdir
[strlen (tmpdir
) - 1] == '/'
6736 || tmpdir
[strlen (tmpdir
) - 1] == '\\')
6741 if (tmpdir
[strlen (tmpdir
) - 1] == '/')
6745 char *templt
= concat (tmpdir
, slash
, "etXXXXXX");
6746 int fd
= mkostemp (templt
, O_CLOEXEC
);
6747 if (fd
< 0 || close (fd
) != 0)
6749 int temp_errno
= errno
;
6755 #if defined (DOS_NT)
6756 /* The file name will be used in shell redirection, so it needs to have
6757 DOS-style backslashes, or else the Windows shell will barf. */
6759 for (p
= templt
; *p
; p
++)
6767 /* Return a newly allocated string containing the file name of FILE
6768 relative to the absolute directory DIR (which should end with a slash). */
6770 relative_filename (char *file
, char *dir
)
6772 char *fp
, *dp
, *afn
, *res
;
6775 /* Find the common root of file and dir (with a trailing slash). */
6776 afn
= absolute_filename (file
, cwd
);
6779 while (*fp
++ == *dp
++)
6781 fp
--, dp
--; /* back to the first differing char */
6783 if (fp
== afn
&& afn
[0] != '/') /* cannot build a relative name */
6786 do /* look at the equal chars until '/' */
6790 /* Build a sequence of "../" strings for the resulting relative file name. */
6792 while ((dp
= strchr (dp
+ 1, '/')) != NULL
)
6794 res
= xnew (3*i
+ strlen (fp
+ 1) + 1, char);
6797 z
= stpcpy (z
, "../");
6799 /* Add the file name relative to the common root of file and dir. */
6806 /* Return a newly allocated string containing the absolute file name
6807 of FILE given DIR (which should end with a slash). */
6809 absolute_filename (char *file
, char *dir
)
6811 char *slashp
, *cp
, *res
;
6813 if (filename_is_absolute (file
))
6814 res
= savestr (file
);
6816 /* We don't support non-absolute file names with a drive
6817 letter, like `d:NAME' (it's too much hassle). */
6818 else if (file
[1] == ':')
6819 fatal ("%s: relative file names with drive letters not supported", file
);
6822 res
= concat (dir
, file
, "");
6824 /* Delete the "/dirname/.." and "/." substrings. */
6825 slashp
= strchr (res
, '/');
6826 while (slashp
!= NULL
&& slashp
[0] != '\0')
6828 if (slashp
[1] == '.')
6830 if (slashp
[2] == '.'
6831 && (slashp
[3] == '/' || slashp
[3] == '\0'))
6836 while (cp
>= res
&& !filename_is_absolute (cp
));
6838 cp
= slashp
; /* the absolute name begins with "/.." */
6840 /* Under MSDOS and NT we get `d:/NAME' as absolute
6841 file name, so the luser could say `d:/../NAME'.
6842 We silently treat this as `d:/NAME'. */
6843 else if (cp
[0] != '/')
6846 memmove (cp
, slashp
+ 3, strlen (slashp
+ 2));
6850 else if (slashp
[2] == '/' || slashp
[2] == '\0')
6852 memmove (slashp
, slashp
+ 2, strlen (slashp
+ 1));
6857 slashp
= strchr (slashp
+ 1, '/');
6860 if (res
[0] == '\0') /* just a safety net: should never happen */
6863 return savestr ("/");
6869 /* Return a newly allocated string containing the absolute
6870 file name of dir where FILE resides given DIR (which should
6871 end with a slash). */
6873 absolute_dirname (char *file
, char *dir
)
6878 slashp
= strrchr (file
, '/');
6880 return savestr (dir
);
6883 res
= absolute_filename (file
, dir
);
6889 /* Whether the argument string is an absolute file name. The argument
6890 string must have been canonicalized with canonicalize_filename. */
6892 filename_is_absolute (char *fn
)
6894 return (fn
[0] == '/'
6896 || (c_isalpha (fn
[0]) && fn
[1] == ':' && fn
[2] == '/')
6901 /* Downcase DOS drive letter and collapse separators into single slashes.
6904 canonicalize_filename (register char *fn
)
6909 /* Canonicalize drive letter case. */
6910 if (c_isupper (fn
[0]) && fn
[1] == ':')
6911 fn
[0] = c_tolower (fn
[0]);
6913 /* Collapse multiple forward- and back-slashes into a single forward
6915 for (cp
= fn
; *cp
!= '\0'; cp
++, fn
++)
6916 if (*cp
== '/' || *cp
== '\\')
6919 while (cp
[1] == '/' || cp
[1] == '\\')
6927 /* Collapse multiple slashes into a single slash. */
6928 for (cp
= fn
; *cp
!= '\0'; cp
++, fn
++)
6932 while (cp
[1] == '/')
6938 #endif /* !DOS_NT */
6944 /* Initialize a linebuffer for use. */
6946 linebuffer_init (linebuffer
*lbp
)
6948 lbp
->size
= (DEBUG
) ? 3 : 200;
6949 lbp
->buffer
= xnew (lbp
->size
, char);
6950 lbp
->buffer
[0] = '\0';
6954 /* Set the minimum size of a string contained in a linebuffer. */
6956 linebuffer_setlen (linebuffer
*lbp
, int toksize
)
6958 while (lbp
->size
<= toksize
)
6961 xrnew (lbp
->buffer
, lbp
->size
, char);
6966 /* Like malloc but get fatal error if memory is exhausted. */
6968 xmalloc (size_t size
)
6970 void *result
= malloc (size
);
6972 fatal ("virtual memory exhausted");
6977 xrealloc (void *ptr
, size_t size
)
6979 void *result
= realloc (ptr
, size
);
6981 fatal ("virtual memory exhausted");
6987 * indent-tabs-mode: t
6990 * c-font-lock-extra-types: ("FILE" "bool" "language" "linebuffer" "fdesc" "node" "regexp")
6991 * c-file-style: "gnu"
6995 /* etags.c ends here */