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, 1998-2017 Free Software
34 This file is not considered part of GNU Emacs.
36 This program is free software: you can redistribute it and/or modify
37 it under the terms of the GNU General Public License as published by
38 the Free Software Foundation, either version 3 of the License, or (at
39 your option) any later version.
41 This program is distributed in the hope that it will be useful,
42 but WITHOUT ANY WARRANTY; without even the implied warranty of
43 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
44 GNU General Public License for more details.
46 You should have received a copy of the GNU General Public License
47 along with this program. If not, see <http://www.gnu.org/licenses/>. */
50 /* NB To comply with the above BSD license, copyright information is
51 reproduced in etc/ETAGS.README. That file should be updated when the
54 To the best of our knowledge, this code was originally based on the
55 ctags.c distributed with BSD4.2, which was copyrighted by the
56 University of California, as described above. */
61 * 1983 Ctags originally by Ken Arnold.
62 * 1984 Fortran added by Jim Kleckner.
63 * 1984 Ed Pelegri-Llopart added C typedefs.
64 * 1985 Emacs TAGS format by Richard Stallman.
65 * 1989 Sam Kendall added C++.
66 * 1992 Joseph B. Wells improved C and C++ parsing.
67 * 1993 Francesco Potortì reorganized C and C++.
68 * 1994 Line-by-line regexp tags by Tom Tromey.
69 * 2001 Nested classes by Francesco Potortì (concept by Mykola Dzyuba).
70 * 2002 #line directives by Francesco Potortì.
71 * Francesco Potortì maintained and improved it for many years
76 * If you want to add support for a new language, start by looking at the LUA
77 * language, which is the simplest. Alternatively, consider distributing etags
78 * together with a configuration file containing regexp definitions for etags.
81 char pot_etags_version
[] = "@(#) pot revision number is 17.38.1.4";
88 # define NDEBUG /* disable assert */
93 /* WIN32_NATIVE is for XEmacs.
94 MSDOS, WINDOWSNT, DOS_NT are for Emacs. */
99 #endif /* WIN32_NATIVE */
104 # include <sys/param.h>
114 # define O_CLOEXEC O_NOINHERIT
115 #endif /* WINDOWSNT */
122 #include <sysstdio.h>
125 #include <binary-io.h>
126 #include <unlocked-io.h>
128 #include <c-strcase.h>
132 # undef assert /* some systems have a buggy assert.h */
133 # define assert(x) ((void) 0)
139 /* Define CTAGS to make the program "ctags" compatible with the usual one.
140 Leave it undefined to make the program "etags", which makes emacs-style
141 tag tables and tags typedefs, #defines and struct/union/enum by default. */
150 streq (char const *s
, char const *t
)
152 return strcmp (s
, t
) == 0;
156 strcaseeq (char const *s
, char const *t
)
158 return c_strcasecmp (s
, t
) == 0;
162 strneq (char const *s
, char const *t
, size_t n
)
164 return strncmp (s
, t
, n
) == 0;
168 strncaseeq (char const *s
, char const *t
, size_t n
)
170 return c_strncasecmp (s
, t
, n
) == 0;
173 /* C is not in a name. */
175 notinname (unsigned char c
)
177 /* Look at make_tag before modifying! */
178 static bool const table
[UCHAR_MAX
+ 1] = {
179 ['\0']=1, ['\t']=1, ['\n']=1, ['\f']=1, ['\r']=1, [' ']=1,
180 ['(']=1, [')']=1, [',']=1, [';']=1, ['=']=1
185 /* C can start a token. */
187 begtoken (unsigned char c
)
189 static bool const table
[UCHAR_MAX
+ 1] = {
191 ['A']=1, ['B']=1, ['C']=1, ['D']=1, ['E']=1, ['F']=1, ['G']=1, ['H']=1,
192 ['I']=1, ['J']=1, ['K']=1, ['L']=1, ['M']=1, ['N']=1, ['O']=1, ['P']=1,
193 ['Q']=1, ['R']=1, ['S']=1, ['T']=1, ['U']=1, ['V']=1, ['W']=1, ['X']=1,
196 ['a']=1, ['b']=1, ['c']=1, ['d']=1, ['e']=1, ['f']=1, ['g']=1, ['h']=1,
197 ['i']=1, ['j']=1, ['k']=1, ['l']=1, ['m']=1, ['n']=1, ['o']=1, ['p']=1,
198 ['q']=1, ['r']=1, ['s']=1, ['t']=1, ['u']=1, ['v']=1, ['w']=1, ['x']=1,
205 /* C can be in the middle of a token. */
207 intoken (unsigned char c
)
209 static bool const table
[UCHAR_MAX
+ 1] = {
211 ['0']=1, ['1']=1, ['2']=1, ['3']=1, ['4']=1,
212 ['5']=1, ['6']=1, ['7']=1, ['8']=1, ['9']=1,
213 ['A']=1, ['B']=1, ['C']=1, ['D']=1, ['E']=1, ['F']=1, ['G']=1, ['H']=1,
214 ['I']=1, ['J']=1, ['K']=1, ['L']=1, ['M']=1, ['N']=1, ['O']=1, ['P']=1,
215 ['Q']=1, ['R']=1, ['S']=1, ['T']=1, ['U']=1, ['V']=1, ['W']=1, ['X']=1,
218 ['a']=1, ['b']=1, ['c']=1, ['d']=1, ['e']=1, ['f']=1, ['g']=1, ['h']=1,
219 ['i']=1, ['j']=1, ['k']=1, ['l']=1, ['m']=1, ['n']=1, ['o']=1, ['p']=1,
220 ['q']=1, ['r']=1, ['s']=1, ['t']=1, ['u']=1, ['v']=1, ['w']=1, ['x']=1,
226 /* C can end a token. */
228 endtoken (unsigned char c
)
230 static bool const table
[UCHAR_MAX
+ 1] = {
231 ['\0']=1, ['\t']=1, ['\n']=1, ['\r']=1, [' ']=1,
232 ['!']=1, ['"']=1, ['#']=1, ['%']=1, ['&']=1, ['\'']=1, ['(']=1, [')']=1,
233 ['*']=1, ['+']=1, [',']=1, ['-']=1, ['.']=1, ['/']=1, [':']=1, [';']=1,
234 ['<']=1, ['=']=1, ['>']=1, ['?']=1, ['[']=1, [']']=1, ['^']=1,
235 ['{']=1, ['|']=1, ['}']=1, ['~']=1
241 * xnew, xrnew -- allocate, reallocate storage
243 * SYNOPSIS: Type *xnew (int n, Type);
244 * void xrnew (OldPointer, int n, Type);
246 #define xnew(n, Type) ((Type *) xmalloc ((n) * sizeof (Type)))
247 #define xrnew(op, n, Type) ((op) = (Type *) xrealloc (op, (n) * sizeof (Type)))
249 typedef void Lang_function (FILE *);
253 const char *suffix
; /* file name suffix for this compressor */
254 const char *command
; /* takes one arg and decompresses to stdout */
259 const char *name
; /* language name */
260 const char *help
; /* detailed help for the language */
261 Lang_function
*function
; /* parse function */
262 const char **suffixes
; /* name suffixes of this language's files */
263 const char **filenames
; /* names of this language's files */
264 const char **interpreters
; /* interpreters for this language */
265 bool metasource
; /* source used to generate other sources */
270 struct fdesc
*next
; /* for the linked list */
271 char *infname
; /* uncompressed input file name */
272 char *infabsname
; /* absolute uncompressed input file name */
273 char *infabsdir
; /* absolute dir of input file */
274 char *taggedfname
; /* file name to write in tagfile */
275 language
*lang
; /* language of file */
276 char *prop
; /* file properties to write in tagfile */
277 bool usecharno
; /* etags tags shall contain char number */
278 bool written
; /* entry written in the tags file */
281 typedef struct node_st
282 { /* sorting structure */
283 struct node_st
*left
, *right
; /* left and right sons */
284 fdesc
*fdp
; /* description of file to whom tag belongs */
285 char *name
; /* tag name */
286 char *regex
; /* search regexp */
287 bool valid
; /* write this tag on the tag file */
288 bool is_func
; /* function tag: use regexp in CTAGS mode */
289 bool been_warned
; /* warning already given for duplicated tag */
290 int lno
; /* line number tag is on */
291 long cno
; /* character number line starts on */
295 * A `linebuffer' is a structure which holds a line of text.
296 * `readline_internal' reads a line from a stream into a linebuffer
297 * and works regardless of the length of the line.
298 * SIZE is the size of BUFFER, LEN is the length of the string in
299 * BUFFER after readline reads it.
308 /* Used to support mixing of --lang and file names. */
312 at_language
, /* a language specification */
313 at_regexp
, /* a regular expression */
314 at_filename
, /* a file name */
315 at_stdin
, /* read from stdin here */
316 at_end
/* stop parsing the list */
317 } arg_type
; /* argument type */
318 language
*lang
; /* language associated with the argument */
319 char *what
; /* the argument itself */
322 /* Structure defining a regular expression. */
323 typedef struct regexp
325 struct regexp
*p_next
; /* pointer to next in list */
326 language
*lang
; /* if set, use only for this language */
327 char *pattern
; /* the regexp pattern */
328 char *name
; /* tag name */
329 struct re_pattern_buffer
*pat
; /* the compiled pattern */
330 struct re_registers regs
; /* re registers */
331 bool error_signaled
; /* already signaled for this regexp */
332 bool force_explicit_name
; /* do not allow implicit tag name */
333 bool ignore_case
; /* ignore case when matching */
334 bool multi_line
; /* do a multi-line match on the whole file */
338 /* Many compilers barf on this:
339 Lang_function Ada_funcs;
340 so let's write it this way */
341 static void Ada_funcs (FILE *);
342 static void Asm_labels (FILE *);
343 static void C_entries (int c_ext
, FILE *);
344 static void default_C_entries (FILE *);
345 static void plain_C_entries (FILE *);
346 static void Cjava_entries (FILE *);
347 static void Cobol_paragraphs (FILE *);
348 static void Cplusplus_entries (FILE *);
349 static void Cstar_entries (FILE *);
350 static void Erlang_functions (FILE *);
351 static void Forth_words (FILE *);
352 static void Fortran_functions (FILE *);
353 static void Go_functions (FILE *);
354 static void HTML_labels (FILE *);
355 static void Lisp_functions (FILE *);
356 static void Lua_functions (FILE *);
357 static void Makefile_targets (FILE *);
358 static void Pascal_functions (FILE *);
359 static void Perl_functions (FILE *);
360 static void PHP_functions (FILE *);
361 static void PS_functions (FILE *);
362 static void Prolog_functions (FILE *);
363 static void Python_functions (FILE *);
364 static void Ruby_functions (FILE *);
365 static void Scheme_functions (FILE *);
366 static void TeX_commands (FILE *);
367 static void Texinfo_nodes (FILE *);
368 static void Yacc_entries (FILE *);
369 static void just_read_file (FILE *);
371 static language
*get_language_from_langname (const char *);
372 static void readline (linebuffer
*, FILE *);
373 static long readline_internal (linebuffer
*, FILE *, char const *);
374 static bool nocase_tail (const char *);
375 static void get_tag (char *, char **);
376 static void get_lispy_tag (char *);
378 static void analyze_regex (char *);
379 static void free_regexps (void);
380 static void regex_tag_multiline (void);
381 static void error (const char *, ...) ATTRIBUTE_FORMAT_PRINTF (1, 2);
382 static void verror (char const *, va_list) ATTRIBUTE_FORMAT_PRINTF (1, 0);
383 static _Noreturn
void suggest_asking_for_help (void);
384 static _Noreturn
void fatal (char const *, ...) ATTRIBUTE_FORMAT_PRINTF (1, 2);
385 static _Noreturn
void pfatal (const char *);
386 static void add_node (node
*, node
**);
388 static void process_file_name (char *, language
*);
389 static void process_file (FILE *, char *, language
*);
390 static void find_entries (FILE *);
391 static void free_tree (node
*);
392 static void free_fdesc (fdesc
*);
393 static void pfnote (char *, bool, char *, int, int, long);
394 static void invalidate_nodes (fdesc
*, node
**);
395 static void put_entries (node
*);
397 static char *concat (const char *, const char *, const char *);
398 static char *skip_spaces (char *);
399 static char *skip_non_spaces (char *);
400 static char *skip_name (char *);
401 static char *savenstr (const char *, int);
402 static char *savestr (const char *);
403 static char *etags_getcwd (void);
404 static char *relative_filename (char *, char *);
405 static char *absolute_filename (char *, char *);
406 static char *absolute_dirname (char *, char *);
407 static bool filename_is_absolute (char *f
);
408 static void canonicalize_filename (char *);
409 static char *etags_mktmp (void);
410 static void linebuffer_init (linebuffer
*);
411 static void linebuffer_setlen (linebuffer
*, int);
412 static void *xmalloc (size_t);
413 static void *xrealloc (void *, size_t);
416 static char searchar
= '/'; /* use /.../ searches */
418 static char *tagfile
; /* output file */
419 static char *progname
; /* name this program was invoked with */
420 static char *cwd
; /* current working directory */
421 static char *tagfiledir
; /* directory of tagfile */
422 static FILE *tagf
; /* ioptr for tags file */
423 static ptrdiff_t whatlen_max
; /* maximum length of any 'what' member */
425 static fdesc
*fdhead
; /* head of file description list */
426 static fdesc
*curfdp
; /* current file description */
427 static char *infilename
; /* current input file name */
428 static int lineno
; /* line number of current line */
429 static long charno
; /* current character number */
430 static long linecharno
; /* charno of start of current line */
431 static char *dbp
; /* pointer to start of current tag */
433 static const int invalidcharno
= -1;
435 static node
*nodehead
; /* the head of the binary tree of tags */
436 static node
*last_node
; /* the last node created */
438 static linebuffer lb
; /* the current line */
439 static linebuffer filebuf
; /* a buffer containing the whole file */
440 static linebuffer token_name
; /* a buffer containing a tag name */
442 static bool append_to_tagfile
; /* -a: append to tags */
443 /* The next five default to true in C and derived languages. */
444 static bool typedefs
; /* -t: create tags for C and Ada typedefs */
445 static bool typedefs_or_cplusplus
; /* -T: create tags for C typedefs, level */
446 /* 0 struct/enum/union decls, and C++ */
447 /* member functions. */
448 static bool constantypedefs
; /* -d: create tags for C #define, enum */
449 /* constants and variables. */
450 /* -D: opposite of -d. Default under ctags. */
451 static int globals
; /* create tags for global variables */
452 static int members
; /* create tags for C member variables */
453 static int declarations
; /* --declarations: tag them and extern in C&Co*/
454 static int no_line_directive
; /* ignore #line directives (undocumented) */
455 static int no_duplicates
; /* no duplicate tags for ctags (undocumented) */
456 static bool update
; /* -u: update tags */
457 static bool vgrind_style
; /* -v: create vgrind style index output */
458 static bool no_warnings
; /* -w: suppress warnings (undocumented) */
459 static bool cxref_style
; /* -x: create cxref style output */
460 static bool cplusplus
; /* .[hc] means C++, not C (undocumented) */
461 static bool ignoreindent
; /* -I: ignore indentation in C */
462 static int packages_only
; /* --packages-only: in Ada, only tag packages*/
463 static int class_qualify
; /* -Q: produce class-qualified tags in C++/Java */
465 /* STDIN is defined in LynxOS system headers */
470 #define STDIN 0x1001 /* returned by getopt_long on --parse-stdin */
471 static bool parsing_stdin
; /* --parse-stdin used */
473 static regexp
*p_head
; /* list of all regexps */
474 static bool need_filebuf
; /* some regexes are multi-line */
476 static struct option longopts
[] =
478 { "append", no_argument
, NULL
, 'a' },
479 { "packages-only", no_argument
, &packages_only
, 1 },
480 { "c++", no_argument
, NULL
, 'C' },
481 { "declarations", no_argument
, &declarations
, 1 },
482 { "no-line-directive", no_argument
, &no_line_directive
, 1 },
483 { "no-duplicates", no_argument
, &no_duplicates
, 1 },
484 { "help", no_argument
, NULL
, 'h' },
485 { "help", no_argument
, NULL
, 'H' },
486 { "ignore-indentation", no_argument
, NULL
, 'I' },
487 { "language", required_argument
, NULL
, 'l' },
488 { "members", no_argument
, &members
, 1 },
489 { "no-members", no_argument
, &members
, 0 },
490 { "output", required_argument
, NULL
, 'o' },
491 { "class-qualify", no_argument
, &class_qualify
, 'Q' },
492 { "regex", required_argument
, NULL
, 'r' },
493 { "no-regex", no_argument
, NULL
, 'R' },
494 { "ignore-case-regex", required_argument
, NULL
, 'c' },
495 { "parse-stdin", required_argument
, NULL
, STDIN
},
496 { "version", no_argument
, NULL
, 'V' },
498 #if CTAGS /* Ctags options */
499 { "backward-search", no_argument
, NULL
, 'B' },
500 { "cxref", no_argument
, NULL
, 'x' },
501 { "defines", no_argument
, NULL
, 'd' },
502 { "globals", no_argument
, &globals
, 1 },
503 { "typedefs", no_argument
, NULL
, 't' },
504 { "typedefs-and-c++", no_argument
, NULL
, 'T' },
505 { "update", no_argument
, NULL
, 'u' },
506 { "vgrind", no_argument
, NULL
, 'v' },
507 { "no-warn", no_argument
, NULL
, 'w' },
509 #else /* Etags options */
510 { "no-defines", no_argument
, NULL
, 'D' },
511 { "no-globals", no_argument
, &globals
, 0 },
512 { "include", required_argument
, NULL
, 'i' },
517 static compressor compressors
[] =
519 { "z", "gzip -d -c"},
520 { "Z", "gzip -d -c"},
521 { "gz", "gzip -d -c"},
522 { "GZ", "gzip -d -c"},
523 { "bz2", "bzip2 -d -c" },
524 { "xz", "xz -d -c" },
533 static const char *Ada_suffixes
[] =
534 { "ads", "adb", "ada", NULL
};
535 static const char Ada_help
[] =
536 "In Ada code, functions, procedures, packages, tasks and types are\n\
537 tags. Use the '--packages-only' option to create tags for\n\
539 Ada tag names have suffixes indicating the type of entity:\n\
540 Entity type: Qualifier:\n\
541 ------------ ----------\n\
548 Thus, 'M-x find-tag <RET> bidule/b <RET>' will go directly to the\n\
549 body of the package 'bidule', while 'M-x find-tag <RET> bidule <RET>'\n\
550 will just search for any tag 'bidule'.";
553 static const char *Asm_suffixes
[] =
554 { "a", /* Unix assembler */
555 "asm", /* Microcontroller assembly */
556 "def", /* BSO/Tasking definition includes */
557 "inc", /* Microcontroller include files */
558 "ins", /* Microcontroller include files */
559 "s", "sa", /* Unix assembler */
560 "S", /* cpp-processed Unix assembler */
561 "src", /* BSO/Tasking C compiler output */
564 static const char Asm_help
[] =
565 "In assembler code, labels appearing at the beginning of a line,\n\
566 followed by a colon, are tags.";
569 /* Note that .c and .h can be considered C++, if the --c++ flag was
570 given, or if the `class' or `template' keywords are met inside the file.
571 That is why default_C_entries is called for these. */
572 static const char *default_C_suffixes
[] =
574 #if CTAGS /* C help for Ctags */
575 static const char default_C_help
[] =
576 "In C code, any C function is a tag. Use -t to tag typedefs.\n\
577 Use -T to tag definitions of 'struct', 'union' and 'enum'.\n\
578 Use -d to tag '#define' macro definitions and 'enum' constants.\n\
579 Use --globals to tag global variables.\n\
580 You can tag function declarations and external variables by\n\
581 using '--declarations', and struct members by using '--members'.";
582 #else /* C help for Etags */
583 static const char default_C_help
[] =
584 "In C code, any C function or typedef is a tag, and so are\n\
585 definitions of 'struct', 'union' and 'enum'. '#define' macro\n\
586 definitions and 'enum' constants are tags unless you specify\n\
587 '--no-defines'. Global variables are tags unless you specify\n\
588 '--no-globals' and so are struct members unless you specify\n\
589 '--no-members'. Use of '--no-globals', '--no-defines' and\n\
590 '--no-members' can make the tags table file much smaller.\n\
591 You can tag function declarations and external variables by\n\
592 using '--declarations'.";
593 #endif /* C help for Ctags and Etags */
595 static const char *Cplusplus_suffixes
[] =
596 { "C", "c++", "cc", "cpp", "cxx", "H", "h++", "hh", "hpp", "hxx",
597 "M", /* Objective C++ */
598 "pdb", /* PostScript with C syntax */
600 static const char Cplusplus_help
[] =
601 "In C++ code, all the tag constructs of C code are tagged. (Use\n\
602 --help --lang=c --lang=c++ for full help.)\n\
603 In addition to C tags, member functions are also recognized. Member\n\
604 variables are recognized unless you use the '--no-members' option.\n\
605 Tags for variables and functions in classes are named 'CLASS::VARIABLE'\n\
606 and 'CLASS::FUNCTION'. 'operator' definitions have tag names like\n\
609 static const char *Cjava_suffixes
[] =
611 static char Cjava_help
[] =
612 "In Java code, all the tags constructs of C and C++ code are\n\
613 tagged. (Use --help --lang=c --lang=c++ --lang=java for full help.)";
616 static const char *Cobol_suffixes
[] =
617 { "COB", "cob", NULL
};
618 static char Cobol_help
[] =
619 "In Cobol code, tags are paragraph names; that is, any word\n\
620 starting in column 8 and followed by a period.";
622 static const char *Cstar_suffixes
[] =
623 { "cs", "hs", NULL
};
625 static const char *Erlang_suffixes
[] =
626 { "erl", "hrl", NULL
};
627 static const char Erlang_help
[] =
628 "In Erlang code, the tags are the functions, records and macros\n\
629 defined in the file.";
631 const char *Forth_suffixes
[] =
632 { "fth", "tok", NULL
};
633 static const char Forth_help
[] =
634 "In Forth code, tags are words defined by ':',\n\
635 constant, code, create, defer, value, variable, buffer:, field.";
637 static const char *Fortran_suffixes
[] =
638 { "F", "f", "f90", "for", NULL
};
639 static const char Fortran_help
[] =
640 "In Fortran code, functions, subroutines and block data are tags.";
642 static const char *Go_suffixes
[] = {"go", NULL
};
643 static const char Go_help
[] =
644 "In Go code, functions, interfaces and packages are tags.";
646 static const char *HTML_suffixes
[] =
647 { "htm", "html", "shtml", NULL
};
648 static const char HTML_help
[] =
649 "In HTML input files, the tags are the 'title' and the 'h1', 'h2',\n\
650 'h3' headers. Also, tags are 'name=' in anchors and all\n\
651 occurrences of 'id='.";
653 static const char *Lisp_suffixes
[] =
654 { "cl", "clisp", "el", "l", "lisp", "LSP", "lsp", "ml", NULL
};
655 static const char Lisp_help
[] =
656 "In Lisp code, any function defined with 'defun', any variable\n\
657 defined with 'defvar' or 'defconst', and in general the first\n\
658 argument of any expression that starts with '(def' in column zero\n\
660 The '--declarations' option tags \"(defvar foo)\" constructs too.";
662 static const char *Lua_suffixes
[] =
663 { "lua", "LUA", NULL
};
664 static const char Lua_help
[] =
665 "In Lua scripts, all functions are tags.";
667 static const char *Makefile_filenames
[] =
668 { "Makefile", "makefile", "GNUMakefile", "Makefile.in", "Makefile.am", NULL
};
669 static const char Makefile_help
[] =
670 "In makefiles, targets are tags; additionally, variables are tags\n\
671 unless you specify '--no-globals'.";
673 static const char *Objc_suffixes
[] =
674 { "lm", /* Objective lex file */
675 "m", /* Objective C file */
677 static const char Objc_help
[] =
678 "In Objective C code, tags include Objective C definitions for classes,\n\
679 class categories, methods and protocols. Tags for variables and\n\
680 functions in classes are named 'CLASS::VARIABLE' and 'CLASS::FUNCTION'.\
681 \n(Use --help --lang=c --lang=objc --lang=java for full help.)";
683 static const char *Pascal_suffixes
[] =
684 { "p", "pas", NULL
};
685 static const char Pascal_help
[] =
686 "In Pascal code, the tags are the functions and procedures defined\n\
688 /* " // this is for working around an Emacs highlighting bug... */
690 static const char *Perl_suffixes
[] =
691 { "pl", "pm", NULL
};
692 static const char *Perl_interpreters
[] =
693 { "perl", "@PERL@", NULL
};
694 static const char Perl_help
[] =
695 "In Perl code, the tags are the packages, subroutines and variables\n\
696 defined by the 'package', 'sub', 'my' and 'local' keywords. Use\n\
697 '--globals' if you want to tag global variables. Tags for\n\
698 subroutines are named 'PACKAGE::SUB'. The name for subroutines\n\
699 defined in the default package is 'main::SUB'.";
701 static const char *PHP_suffixes
[] =
702 { "php", "php3", "php4", NULL
};
703 static const char PHP_help
[] =
704 "In PHP code, tags are functions, classes and defines. Unless you use\n\
705 the '--no-members' option, vars are tags too.";
707 static const char *plain_C_suffixes
[] =
708 { "pc", /* Pro*C file */
711 static const char *PS_suffixes
[] =
712 { "ps", "psw", NULL
}; /* .psw is for PSWrap */
713 static const char PS_help
[] =
714 "In PostScript code, the tags are the functions.";
716 static const char *Prolog_suffixes
[] =
718 static const char Prolog_help
[] =
719 "In Prolog code, tags are predicates and rules at the beginning of\n\
722 static const char *Python_suffixes
[] =
724 static const char Python_help
[] =
725 "In Python code, 'def' or 'class' at the beginning of a line\n\
728 static const char *Ruby_suffixes
[] =
729 { "rb", "ru", "rbw", NULL
};
730 static const char *Ruby_filenames
[] =
731 { "Rakefile", "Thorfile", NULL
};
732 static const char Ruby_help
[] =
733 "In Ruby code, 'def' or 'class' or 'module' at the beginning of\n\
734 a line generate a tag. Constants also generate a tag.";
736 /* Can't do the `SCM' or `scm' prefix with a version number. */
737 static const char *Scheme_suffixes
[] =
738 { "oak", "sch", "scheme", "SCM", "scm", "SM", "sm", "ss", "t", NULL
};
739 static const char Scheme_help
[] =
740 "In Scheme code, tags include anything defined with 'def' or with a\n\
741 construct whose name starts with 'def'. They also include\n\
742 variables set with 'set!' at top level in the file.";
744 static const char *TeX_suffixes
[] =
745 { "bib", "clo", "cls", "ltx", "sty", "TeX", "tex", NULL
};
746 static const char TeX_help
[] =
747 "In LaTeX text, the argument of any of the commands '\\chapter',\n\
748 '\\section', '\\subsection', '\\subsubsection', '\\eqno', '\\label',\n\
749 '\\ref', '\\cite', '\\bibitem', '\\part', '\\appendix', '\\entry',\n\
750 '\\index', '\\def', '\\newcommand', '\\renewcommand',\n\
751 '\\newenvironment' or '\\renewenvironment' is a tag.\n\
753 Other commands can be specified by setting the environment variable\n\
754 'TEXTAGS' to a colon-separated list like, for example,\n\
755 TEXTAGS=\"mycommand:myothercommand\".";
758 static const char *Texinfo_suffixes
[] =
759 { "texi", "texinfo", "txi", NULL
};
760 static const char Texinfo_help
[] =
761 "for texinfo files, lines starting with @node are tagged.";
763 static const char *Yacc_suffixes
[] =
764 { "y", "y++", "ym", "yxx", "yy", NULL
}; /* .ym is Objective yacc file */
765 static const char Yacc_help
[] =
766 "In Bison or Yacc input files, each rule defines as a tag the\n\
767 nonterminal it constructs. The portions of the file that contain\n\
768 C code are parsed as C code (use --help --lang=c --lang=yacc\n\
771 static const char auto_help
[] =
772 "'auto' is not a real language, it indicates to use\n\
773 a default language for files base on file name suffix and file contents.";
775 static const char none_help
[] =
776 "'none' is not a real language, it indicates to only do\n\
777 regexp processing on files.";
779 static const char no_lang_help
[] =
780 "No detailed help available for this language.";
784 * Table of languages.
786 * It is ok for a given function to be listed under more than one
787 * name. I just didn't.
790 static language lang_names
[] =
792 { "ada", Ada_help
, Ada_funcs
, Ada_suffixes
},
793 { "asm", Asm_help
, Asm_labels
, Asm_suffixes
},
794 { "c", default_C_help
, default_C_entries
, default_C_suffixes
},
795 { "c++", Cplusplus_help
, Cplusplus_entries
, Cplusplus_suffixes
},
796 { "c*", no_lang_help
, Cstar_entries
, Cstar_suffixes
},
797 { "cobol", Cobol_help
, Cobol_paragraphs
, Cobol_suffixes
},
798 { "erlang", Erlang_help
, Erlang_functions
, Erlang_suffixes
},
799 { "forth", Forth_help
, Forth_words
, Forth_suffixes
},
800 { "fortran", Fortran_help
, Fortran_functions
, Fortran_suffixes
},
801 { "go", Go_help
, Go_functions
, Go_suffixes
},
802 { "html", HTML_help
, HTML_labels
, HTML_suffixes
},
803 { "java", Cjava_help
, Cjava_entries
, Cjava_suffixes
},
804 { "lisp", Lisp_help
, Lisp_functions
, Lisp_suffixes
},
805 { "lua", Lua_help
, Lua_functions
, Lua_suffixes
},
806 { "makefile", Makefile_help
,Makefile_targets
,NULL
,Makefile_filenames
},
807 { "objc", Objc_help
, plain_C_entries
, Objc_suffixes
},
808 { "pascal", Pascal_help
, Pascal_functions
, Pascal_suffixes
},
809 { "perl",Perl_help
,Perl_functions
,Perl_suffixes
,NULL
,Perl_interpreters
},
810 { "php", PHP_help
, PHP_functions
, PHP_suffixes
},
811 { "postscript",PS_help
, PS_functions
, PS_suffixes
},
812 { "proc", no_lang_help
, plain_C_entries
, plain_C_suffixes
},
813 { "prolog", Prolog_help
, Prolog_functions
, Prolog_suffixes
},
814 { "python", Python_help
, Python_functions
, Python_suffixes
},
815 { "ruby", Ruby_help
,Ruby_functions
,Ruby_suffixes
,Ruby_filenames
},
816 { "scheme", Scheme_help
, Scheme_functions
, Scheme_suffixes
},
817 { "tex", TeX_help
, TeX_commands
, TeX_suffixes
},
818 { "texinfo", Texinfo_help
, Texinfo_nodes
, Texinfo_suffixes
},
819 { "yacc", Yacc_help
,Yacc_entries
,Yacc_suffixes
,NULL
,NULL
,true},
820 { "auto", auto_help
}, /* default guessing scheme */
821 { "none", none_help
, just_read_file
}, /* regexp matching only */
822 { NULL
} /* end of list */
827 print_language_names (void)
830 const char **name
, **ext
;
832 puts ("\nThese are the currently supported languages, along with the\n\
833 default file names and dot suffixes:");
834 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
836 printf (" %-*s", 10, lang
->name
);
837 if (lang
->filenames
!= NULL
)
838 for (name
= lang
->filenames
; *name
!= NULL
; name
++)
839 printf (" %s", *name
);
840 if (lang
->suffixes
!= NULL
)
841 for (ext
= lang
->suffixes
; *ext
!= NULL
; ext
++)
842 printf (" .%s", *ext
);
845 puts ("where 'auto' means use default language for files based on file\n\
846 name suffix, and 'none' means only do regexp processing on files.\n\
847 If no language is specified and no matching suffix is found,\n\
848 the first line of the file is read for a sharp-bang (#!) sequence\n\
849 followed by the name of an interpreter. If no such sequence is found,\n\
850 Fortran is tried first; if no tags are found, C is tried next.\n\
851 When parsing any C file, a \"class\" or \"template\" keyword\n\
853 puts ("Compressed files are supported using gzip, bzip2, and xz.\n\
855 For detailed help on a given language use, for example,\n\
856 etags --help --lang=ada.");
860 # define EMACS_NAME "standalone"
863 # define VERSION "17.38.1.4"
865 static _Noreturn
void
868 char emacs_copyright
[] = COPYRIGHT
;
870 printf ("%s (%s %s)\n", (CTAGS
) ? "ctags" : "etags", EMACS_NAME
, VERSION
);
871 puts (emacs_copyright
);
872 puts ("This program is distributed under the terms in ETAGS.README");
877 #ifndef PRINT_UNDOCUMENTED_OPTIONS_HELP
878 # define PRINT_UNDOCUMENTED_OPTIONS_HELP false
881 static _Noreturn
void
882 print_help (argument
*argbuffer
)
884 bool help_for_lang
= false;
886 for (; argbuffer
->arg_type
!= at_end
; argbuffer
++)
887 if (argbuffer
->arg_type
== at_language
)
891 puts (argbuffer
->lang
->help
);
892 help_for_lang
= true;
898 printf ("Usage: %s [options] [[regex-option ...] file-name] ...\n\
900 These are the options accepted by %s.\n", progname
, progname
);
901 puts ("You may use unambiguous abbreviations for the long option names.");
902 puts (" A - as file name means read names from stdin (one per line).\n\
903 Absolute names are stored in the output file as they are.\n\
904 Relative ones are stored relative to the output file's directory.\n");
906 puts ("-a, --append\n\
907 Append tag entries to existing tags file.");
909 puts ("--packages-only\n\
910 For Ada files, only generate tags for packages.");
913 puts ("-B, --backward-search\n\
914 Write the search commands for the tag entries using '?', the\n\
915 backward-search command instead of '/', the forward-search command.");
917 /* This option is mostly obsolete, because etags can now automatically
918 detect C++. Retained for backward compatibility and for debugging and
919 experimentation. In principle, we could want to tag as C++ even
920 before any "class" or "template" keyword.
922 Treat files whose name suffix defaults to C language as C++ files.");
925 puts ("--declarations\n\
926 In C and derived languages, create tags for function declarations,");
928 puts ("\tand create tags for extern variables if --globals is used.");
931 ("\tand create tags for extern variables unless --no-globals is used.");
934 puts ("-d, --defines\n\
935 Create tag entries for C #define constants and enum constants, too.");
937 puts ("-D, --no-defines\n\
938 Don't create tag entries for C #define constants and enum constants.\n\
939 This makes the tags file smaller.");
942 puts ("-i FILE, --include=FILE\n\
943 Include a note in tag file indicating that, when searching for\n\
944 a tag, one should also consult the tags file FILE after\n\
945 checking the current file.");
947 puts ("-l LANG, --language=LANG\n\
948 Force the following files to be considered as written in the\n\
949 named language up to the next --language=LANG option.");
953 Create tag entries for global variables in some languages.");
955 puts ("--no-globals\n\
956 Do not create tag entries for global variables in some\n\
957 languages. This makes the tags file smaller.");
959 puts ("--no-line-directive\n\
960 Ignore #line preprocessor directives in C and derived languages.");
964 Create tag entries for members of structures in some languages.");
966 puts ("--no-members\n\
967 Do not create tag entries for members of structures\n\
968 in some languages.");
970 puts ("-Q, --class-qualify\n\
971 Qualify tag names with their class name in C++, ObjC, Java, and Perl.\n\
972 This produces tag names of the form \"class::member\" for C++,\n\
973 \"class(category)\" for Objective C, and \"class.member\" for Java.\n\
974 For Objective C, this also produces class methods qualified with\n\
975 their arguments, as in \"foo:bar:baz:more\".\n\
976 For Perl, this produces \"package::member\".");
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; FALLTHROUGH
;
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_mode (STDOUT_FILENO
, O_BINARY
);
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 return EXIT_SUCCESS
;
1338 /* From here on, we are in (CTAGS && !cxref_style) */
1342 xmalloc (strlen (tagfile
) + whatlen_max
+
1343 sizeof "mv..OTAGS;grep -Fv '\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;grep -Fv '\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 return 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
;
1488 /* Try whole file name first. */
1489 slash
= strrchr (file
, '/');
1493 else if (file
[0] && file
[1] == ':')
1496 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1497 if (lang
->filenames
!= NULL
)
1498 for (name
= lang
->filenames
; *name
!= NULL
; name
++)
1499 if ((case_sensitive
)
1500 ? streq (*name
, file
)
1501 : strcaseeq (*name
, file
))
1504 /* If not found, try suffix after last dot. */
1505 suffix
= strrchr (file
, '.');
1509 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1510 if (lang
->suffixes
!= NULL
)
1511 for (ext
= lang
->suffixes
; *ext
!= NULL
; ext
++)
1512 if ((case_sensitive
)
1513 ? streq (*ext
, suffix
)
1514 : strcaseeq (*ext
, suffix
))
1521 * This routine is called on each file argument.
1524 process_file_name (char *file
, language
*lang
)
1529 char *compressed_name
, *uncompressed_name
;
1530 char *ext
, *real_name
, *tmp_name
;
1533 canonicalize_filename (file
);
1534 if (streq (file
, tagfile
) && !streq (tagfile
, "-"))
1536 error ("skipping inclusion of %s in self.", file
);
1539 compr
= get_compressor_from_suffix (file
, &ext
);
1542 compressed_name
= file
;
1543 uncompressed_name
= savenstr (file
, ext
- file
);
1547 compressed_name
= NULL
;
1548 uncompressed_name
= file
;
1551 /* If the canonicalized uncompressed name
1552 has already been dealt with, skip it silently. */
1553 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
1555 assert (fdp
->infname
!= NULL
);
1556 if (streq (uncompressed_name
, fdp
->infname
))
1560 inf
= fopen (file
, "r" FOPEN_BINARY
);
1565 int file_errno
= errno
;
1566 if (compressed_name
)
1568 /* Try with the given suffix. */
1569 inf
= fopen (uncompressed_name
, "r" FOPEN_BINARY
);
1571 real_name
= uncompressed_name
;
1575 /* Try all possible suffixes. */
1576 for (compr
= compressors
; compr
->suffix
!= NULL
; compr
++)
1578 compressed_name
= concat (file
, ".", compr
->suffix
);
1579 inf
= fopen (compressed_name
, "r" FOPEN_BINARY
);
1582 real_name
= compressed_name
;
1587 char *suf
= compressed_name
+ strlen (file
);
1588 size_t suflen
= strlen (compr
->suffix
) + 1;
1589 for ( ; suf
[1]; suf
++, suflen
--)
1591 memmove (suf
, suf
+ 1, suflen
);
1592 inf
= fopen (compressed_name
, "r" FOPEN_BINARY
);
1595 real_name
= compressed_name
;
1602 free (compressed_name
);
1603 compressed_name
= NULL
;
1614 if (real_name
== compressed_name
)
1617 tmp_name
= etags_mktmp ();
1622 #if MSDOS || defined (DOS_NT)
1623 char *cmd1
= concat (compr
->command
, " \"", real_name
);
1624 char *cmd
= concat (cmd1
, "\" > ", tmp_name
);
1626 char *cmd1
= concat (compr
->command
, " '", real_name
);
1627 char *cmd
= concat (cmd1
, "' > ", tmp_name
);
1631 if (system (cmd
) == -1)
1638 inf
= fopen (tmp_name
, "r" FOPEN_BINARY
);
1652 process_file (inf
, uncompressed_name
, lang
);
1654 retval
= fclose (inf
);
1655 if (real_name
== compressed_name
)
1664 if (compressed_name
!= file
)
1665 free (compressed_name
);
1666 if (uncompressed_name
!= file
)
1667 free (uncompressed_name
);
1674 process_file (FILE *fh
, char *fn
, language
*lang
)
1676 static const fdesc emptyfdesc
;
1680 /* Create a new input file description entry. */
1681 fdp
= xnew (1, fdesc
);
1684 fdp
->infname
= savestr (fn
);
1686 fdp
->infabsname
= absolute_filename (fn
, cwd
);
1687 fdp
->infabsdir
= absolute_dirname (fn
, cwd
);
1688 if (filename_is_absolute (fn
))
1690 /* An absolute file name. Canonicalize it. */
1691 fdp
->taggedfname
= absolute_filename (fn
, NULL
);
1695 /* A file name relative to cwd. Make it relative
1696 to the directory of the tags file. */
1697 fdp
->taggedfname
= relative_filename (fn
, tagfiledir
);
1699 fdp
->usecharno
= true; /* use char position when making tags */
1701 fdp
->written
= false; /* not written on tags file yet */
1704 curfdp
= fdhead
; /* the current file description */
1708 /* If not Ctags, and if this is not metasource and if it contained no #line
1709 directives, we can write the tags and free all nodes pointing to
1712 && curfdp
->usecharno
/* no #line directives in this file */
1713 && !curfdp
->lang
->metasource
)
1717 /* Look for the head of the sublist relative to this file. See add_node
1718 for the structure of the node tree. */
1720 for (np
= nodehead
; np
!= NULL
; prev
= np
, np
= np
->left
)
1721 if (np
->fdp
== curfdp
)
1724 /* If we generated tags for this file, write and delete them. */
1727 /* This is the head of the last sublist, if any. The following
1728 instructions depend on this being true. */
1729 assert (np
->left
== NULL
);
1731 assert (fdhead
== curfdp
);
1732 assert (last_node
->fdp
== curfdp
);
1733 put_entries (np
); /* write tags for file curfdp->taggedfname */
1734 free_tree (np
); /* remove the written nodes */
1736 nodehead
= NULL
; /* no nodes left */
1738 prev
->left
= NULL
; /* delete the pointer to the sublist */
1744 reset_input (FILE *inf
)
1746 if (fseek (inf
, 0, SEEK_SET
) != 0)
1747 perror (infilename
);
1751 * This routine opens the specified file and calls the function
1752 * which finds the function and type definitions.
1755 find_entries (FILE *inf
)
1758 language
*lang
= curfdp
->lang
;
1759 Lang_function
*parser
= NULL
;
1761 /* If user specified a language, use it. */
1762 if (lang
!= NULL
&& lang
->function
!= NULL
)
1764 parser
= lang
->function
;
1767 /* Else try to guess the language given the file name. */
1770 lang
= get_language_from_filename (curfdp
->infname
, true);
1771 if (lang
!= NULL
&& lang
->function
!= NULL
)
1773 curfdp
->lang
= lang
;
1774 parser
= lang
->function
;
1778 /* Else look for sharp-bang as the first two characters. */
1780 && readline_internal (&lb
, inf
, infilename
) > 0
1782 && lb
.buffer
[0] == '#'
1783 && lb
.buffer
[1] == '!')
1787 /* Set lp to point at the first char after the last slash in the
1788 line or, if no slashes, at the first nonblank. Then set cp to
1789 the first successive blank and terminate the string. */
1790 lp
= strrchr (lb
.buffer
+2, '/');
1794 lp
= skip_spaces (lb
.buffer
+ 2);
1795 cp
= skip_non_spaces (lp
);
1798 if (strlen (lp
) > 0)
1800 lang
= get_language_from_interpreter (lp
);
1801 if (lang
!= NULL
&& lang
->function
!= NULL
)
1803 curfdp
->lang
= lang
;
1804 parser
= lang
->function
;
1811 /* Else try to guess the language given the case insensitive file name. */
1814 lang
= get_language_from_filename (curfdp
->infname
, false);
1815 if (lang
!= NULL
&& lang
->function
!= NULL
)
1817 curfdp
->lang
= lang
;
1818 parser
= lang
->function
;
1822 /* Else try Fortran or C. */
1825 node
*old_last_node
= last_node
;
1827 curfdp
->lang
= get_language_from_langname ("fortran");
1830 if (old_last_node
== last_node
)
1831 /* No Fortran entries found. Try C. */
1834 curfdp
->lang
= get_language_from_langname (cplusplus
? "c++" : "c");
1840 if (!no_line_directive
1841 && curfdp
->lang
!= NULL
&& curfdp
->lang
->metasource
)
1842 /* It may be that this is a bingo.y file, and we already parsed a bingo.c
1843 file, or anyway we parsed a file that is automatically generated from
1844 this one. If this is the case, the bingo.c file contained #line
1845 directives that generated tags pointing to this file. Let's delete
1846 them all before parsing this file, which is the real source. */
1848 fdesc
**fdpp
= &fdhead
;
1849 while (*fdpp
!= NULL
)
1851 && streq ((*fdpp
)->taggedfname
, curfdp
->taggedfname
))
1852 /* We found one of those! We must delete both the file description
1853 and all tags referring to it. */
1855 fdesc
*badfdp
= *fdpp
;
1857 /* Delete the tags referring to badfdp->taggedfname
1858 that were obtained from badfdp->infname. */
1859 invalidate_nodes (badfdp
, &nodehead
);
1861 *fdpp
= badfdp
->next
; /* remove the bad description from the list */
1862 free_fdesc (badfdp
);
1865 fdpp
= &(*fdpp
)->next
; /* advance the list pointer */
1868 assert (parser
!= NULL
);
1870 /* Generic initializations before reading from file. */
1871 linebuffer_setlen (&filebuf
, 0); /* reset the file buffer */
1873 /* Generic initializations before parsing file with readline. */
1874 lineno
= 0; /* reset global line number */
1875 charno
= 0; /* reset global char number */
1876 linecharno
= 0; /* reset global char number of line start */
1880 regex_tag_multiline ();
1885 * Check whether an implicitly named tag should be created,
1886 * then call `pfnote'.
1887 * NAME is a string that is internally copied by this function.
1889 * TAGS format specification
1890 * Idea by Sam Kendall <kendall@mv.mv.com> (1997)
1891 * The following is explained in some more detail in etc/ETAGS.EBNF.
1893 * make_tag creates tags with "implicit tag names" (unnamed tags)
1894 * if the following are all true, assuming NONAM=" \f\t\n\r()=,;":
1895 * 1. NAME does not contain any of the characters in NONAM;
1896 * 2. LINESTART contains name as either a rightmost, or rightmost but
1897 * one character, substring;
1898 * 3. the character, if any, immediately before NAME in LINESTART must
1899 * be a character in NONAM;
1900 * 4. the character, if any, immediately after NAME in LINESTART must
1901 * also be a character in NONAM.
1903 * The implementation uses the notinname() macro, which recognizes the
1904 * characters stored in the string `nonam'.
1905 * etags.el needs to use the same characters that are in NONAM.
1908 make_tag (const char *name
, /* tag name, or NULL if unnamed */
1909 int namelen
, /* tag length */
1910 bool is_func
, /* tag is a function */
1911 char *linestart
, /* start of the line where tag is */
1912 int linelen
, /* length of the line where tag is */
1913 int lno
, /* line number */
1914 long int cno
) /* character number */
1916 bool named
= (name
!= NULL
&& namelen
> 0);
1919 if (!CTAGS
&& named
) /* maybe set named to false */
1920 /* Let's try to make an implicit tag name, that is, create an unnamed tag
1921 such that etags.el can guess a name from it. */
1924 register const char *cp
= name
;
1926 for (i
= 0; i
< namelen
; i
++)
1927 if (notinname (*cp
++))
1929 if (i
== namelen
) /* rule #1 */
1931 cp
= linestart
+ linelen
- namelen
;
1932 if (notinname (linestart
[linelen
-1]))
1933 cp
-= 1; /* rule #4 */
1934 if (cp
>= linestart
/* rule #2 */
1936 || notinname (cp
[-1])) /* rule #3 */
1937 && strneq (name
, cp
, namelen
)) /* rule #2 */
1938 named
= false; /* use implicit tag name */
1943 nname
= savenstr (name
, namelen
);
1945 pfnote (nname
, is_func
, linestart
, linelen
, lno
, cno
);
1950 pfnote (char *name
, bool is_func
, char *linestart
, int linelen
, int lno
,
1952 /* tag name, or NULL if unnamed */
1953 /* tag is a function */
1954 /* start of the line where tag is */
1955 /* length of the line where tag is */
1957 /* character number */
1961 assert (name
== NULL
|| name
[0] != '\0');
1962 if (CTAGS
&& name
== NULL
)
1965 np
= xnew (1, node
);
1967 /* If ctags mode, change name "main" to M<thisfilename>. */
1968 if (CTAGS
&& !cxref_style
&& streq (name
, "main"))
1970 char *fp
= strrchr (curfdp
->taggedfname
, '/');
1971 np
->name
= concat ("M", fp
== NULL
? curfdp
->taggedfname
: fp
+ 1, "");
1972 fp
= strrchr (np
->name
, '.');
1973 if (fp
!= NULL
&& fp
[1] != '\0' && fp
[2] == '\0')
1979 np
->been_warned
= false;
1981 np
->is_func
= is_func
;
1983 if (np
->fdp
->usecharno
)
1984 /* Our char numbers are 0-base, because of C language tradition?
1985 ctags compatibility? old versions compatibility? I don't know.
1986 Anyway, since emacs's are 1-base we expect etags.el to take care
1987 of the difference. If we wanted to have 1-based numbers, we would
1988 uncomment the +1 below. */
1989 np
->cno
= cno
/* + 1 */ ;
1991 np
->cno
= invalidcharno
;
1992 np
->left
= np
->right
= NULL
;
1993 if (CTAGS
&& !cxref_style
)
1995 if (strlen (linestart
) < 50)
1996 np
->regex
= concat (linestart
, "$", "");
1998 np
->regex
= savenstr (linestart
, 50);
2001 np
->regex
= savenstr (linestart
, linelen
);
2003 add_node (np
, &nodehead
);
2007 * Utility functions and data to avoid recursion.
2010 typedef struct stack_entry
{
2012 struct stack_entry
*next
;
2016 push_node (node
*np
, stkentry
**stack_top
)
2020 stkentry
*new = xnew (1, stkentry
);
2023 new->next
= *stack_top
;
2029 pop_node (stkentry
**stack_top
)
2035 stkentry
*old_start
= *stack_top
;
2037 ret
= (*stack_top
)->np
;
2038 *stack_top
= (*stack_top
)->next
;
2046 * emulate recursion on left children, iterate on right children.
2049 free_tree (register node
*np
)
2051 stkentry
*stack
= NULL
;
2055 /* Descent on left children. */
2058 push_node (np
, &stack
);
2061 /* Free node without left children. */
2062 node
*node_right
= np
->right
;
2068 /* Backtrack to find a node with right children, while freeing nodes
2069 that don't have right children. */
2070 while (node_right
== NULL
&& (np
= pop_node (&stack
)) != NULL
)
2072 node_right
= np
->right
;
2078 /* Free right children. */
2085 * delete a file description
2088 free_fdesc (register fdesc
*fdp
)
2090 free (fdp
->infname
);
2091 free (fdp
->infabsname
);
2092 free (fdp
->infabsdir
);
2093 free (fdp
->taggedfname
);
2100 * Adds a node to the tree of nodes. In etags mode, sort by file
2101 * name. In ctags mode, sort by tag name. Make no attempt at
2104 * add_node is the only function allowed to add nodes, so it can
2108 add_node (node
*np
, node
**cur_node_p
)
2110 node
*cur_node
= *cur_node_p
;
2112 /* Make the first node. */
2113 if (cur_node
== NULL
)
2123 /* For each file name, tags are in a linked sublist on the right
2124 pointer. The first tags of different files are a linked list
2125 on the left pointer. last_node points to the end of the last
2127 if (last_node
!= NULL
&& last_node
->fdp
== np
->fdp
)
2129 /* Let's use the same sublist as the last added node. */
2130 assert (last_node
->right
== NULL
);
2131 last_node
->right
= np
;
2136 while (cur_node
->fdp
!= np
->fdp
)
2138 if (cur_node
->left
== NULL
)
2140 /* The head of this sublist is not good for us. Let's try the
2142 cur_node
= cur_node
->left
;
2146 /* Scanning the list we found the head of a sublist which is
2147 good for us. Let's scan this sublist. */
2148 if (cur_node
->right
)
2150 cur_node
= cur_node
->right
;
2151 while (cur_node
->right
)
2152 cur_node
= cur_node
->right
;
2154 /* Make a new node in this sublist. */
2155 cur_node
->right
= np
;
2159 /* Make a new sublist. */
2160 cur_node
->left
= np
;
2164 } /* if ETAGS mode */
2168 node
**next_node
= &cur_node
;
2170 while ((cur_node
= *next_node
) != NULL
)
2172 int dif
= strcmp (np
->name
, cur_node
->name
);
2174 * If this tag name matches an existing one, then
2175 * do not add the node, but maybe print a warning.
2177 if (!dif
&& no_duplicates
)
2179 if (np
->fdp
== cur_node
->fdp
)
2184 "Duplicate entry in file %s, line %d: %s\n",
2185 np
->fdp
->infname
, lineno
, np
->name
);
2186 fprintf (stderr
, "Second entry ignored\n");
2189 else if (!cur_node
->been_warned
&& !no_warnings
)
2193 "Duplicate entry in files %s and %s: %s (Warning only)\n",
2194 np
->fdp
->infname
, cur_node
->fdp
->infname
, np
->name
);
2195 cur_node
->been_warned
= true;
2200 next_node
= dif
< 0 ? &cur_node
->left
: &cur_node
->right
;
2204 } /* if CTAGS mode */
2208 * invalidate_nodes ()
2209 * Scan the node tree and invalidate all nodes pointing to the
2210 * given file description (CTAGS case) or free them (ETAGS case).
2213 invalidate_nodes (fdesc
*badfdp
, node
**npp
)
2216 stkentry
*stack
= NULL
;
2222 /* Push all the left children on the stack. */
2223 while (np
->left
!= NULL
)
2225 push_node (np
, &stack
);
2228 /* Invalidate this node. */
2229 if (np
->fdp
== badfdp
)
2233 /* Pop nodes from stack, invalidating them, until we find one
2234 with a right child. */
2235 while ((np
= pop_node (&stack
)) != NULL
)
2237 if (np
->fdp
== badfdp
)
2239 if (np
->right
!= NULL
)
2243 /* Process the right child, if any. */
2250 node super_root
, *np_parent
= NULL
;
2252 super_root
.left
= np
;
2253 super_root
.fdp
= (fdesc
*) -1;
2258 /* Descent on left children until node with BADFP. */
2259 while (np
&& np
->fdp
!= badfdp
)
2261 assert (np
->fdp
!= NULL
);
2267 np_parent
->left
= np
->left
; /* detach subtree from the tree */
2268 np
->left
= NULL
; /* isolate it */
2269 free_tree (np
); /* free it */
2271 /* Continue with rest of tree. */
2272 np
= np_parent
->left
;
2275 *npp
= super_root
.left
;
2280 static int total_size_of_entries (node
*);
2281 static int number_len (long) ATTRIBUTE_CONST
;
2283 /* Length of a non-negative number's decimal representation. */
2285 number_len (long int num
)
2288 while ((num
/= 10) > 0)
2294 * Return total number of characters that put_entries will output for
2295 * the nodes in the linked list at the right of the specified node.
2296 * This count is irrelevant with etags.el since emacs 19.34 at least,
2297 * but is still supplied for backward compatibility.
2300 total_size_of_entries (register node
*np
)
2302 register int total
= 0;
2304 for (; np
!= NULL
; np
= np
->right
)
2307 total
+= strlen (np
->regex
) + 1; /* pat\177 */
2308 if (np
->name
!= NULL
)
2309 total
+= strlen (np
->name
) + 1; /* name\001 */
2310 total
+= number_len ((long) np
->lno
) + 1; /* lno, */
2311 if (np
->cno
!= invalidcharno
) /* cno */
2312 total
+= number_len (np
->cno
);
2313 total
+= 1; /* newline */
2320 put_entry (node
*np
)
2323 static fdesc
*fdp
= NULL
;
2325 /* Output this entry */
2334 fprintf (tagf
, "\f\n%s,%d\n",
2335 fdp
->taggedfname
, total_size_of_entries (np
));
2336 fdp
->written
= true;
2338 fputs (np
->regex
, tagf
);
2339 fputc ('\177', tagf
);
2340 if (np
->name
!= NULL
)
2342 fputs (np
->name
, tagf
);
2343 fputc ('\001', tagf
);
2345 fprintf (tagf
, "%d,", np
->lno
);
2346 if (np
->cno
!= invalidcharno
)
2347 fprintf (tagf
, "%ld", np
->cno
);
2353 if (np
->name
== NULL
)
2354 error ("internal error: NULL name in ctags mode.");
2359 fprintf (stdout
, "%s %s %d\n",
2360 np
->name
, np
->fdp
->taggedfname
, (np
->lno
+ 63) / 64);
2362 fprintf (stdout
, "%-16s %3d %-16s %s\n",
2363 np
->name
, np
->lno
, np
->fdp
->taggedfname
, np
->regex
);
2367 fprintf (tagf
, "%s\t%s\t", np
->name
, np
->fdp
->taggedfname
);
2370 { /* function or #define macro with args */
2371 putc (searchar
, tagf
);
2374 for (sp
= np
->regex
; *sp
; sp
++)
2376 if (*sp
== '\\' || *sp
== searchar
)
2380 putc (searchar
, tagf
);
2383 { /* anything else; text pattern inadequate */
2384 fprintf (tagf
, "%d", np
->lno
);
2389 } /* if this node contains a valid tag */
2393 put_entries (node
*np
)
2395 stkentry
*stack
= NULL
;
2404 /* Stack subentries that precede this one. */
2407 push_node (np
, &stack
);
2410 /* Output this subentry. */
2412 /* Stack subentries that follow this one. */
2415 /* Output subentries that precede the next one. */
2416 np
= pop_node (&stack
);
2427 push_node (np
, &stack
);
2428 while ((np
= pop_node (&stack
)) != NULL
)
2430 /* Output this subentry. */
2434 /* Output subentries that follow this one. */
2435 put_entry (np
->right
);
2436 /* Stack subentries from the following files. */
2437 push_node (np
->left
, &stack
);
2440 push_node (np
->left
, &stack
);
2447 #define C_EXT 0x00fff /* C extensions */
2448 #define C_PLAIN 0x00000 /* C */
2449 #define C_PLPL 0x00001 /* C++ */
2450 #define C_STAR 0x00003 /* C* */
2451 #define C_JAVA 0x00005 /* JAVA */
2452 #define C_AUTO 0x01000 /* C, but switch to C++ if `class' is met */
2453 #define YACC 0x10000 /* yacc file */
2456 * The C symbol tables.
2461 st_C_objprot
, st_C_objimpl
, st_C_objend
,
2463 st_C_ignore
, st_C_attribute
, st_C_enum_bf
,
2466 st_C_class
, st_C_template
,
2467 st_C_struct
, st_C_extern
, st_C_enum
, st_C_define
, st_C_typedef
2470 /* Feed stuff between (but not including) %[ and %] lines to:
2476 struct C_stab_entry { char *name; int c_ext; enum sym_type type; }
2480 while, 0, st_C_ignore
2481 switch, 0, st_C_ignore
2482 return, 0, st_C_ignore
2483 __attribute__, 0, st_C_attribute
2484 GTY, 0, st_C_attribute
2485 @interface, 0, st_C_objprot
2486 @protocol, 0, st_C_objprot
2487 @implementation,0, st_C_objimpl
2488 @end, 0, st_C_objend
2489 import, (C_JAVA & ~C_PLPL), st_C_ignore
2490 package, (C_JAVA & ~C_PLPL), st_C_ignore
2491 friend, C_PLPL, st_C_ignore
2492 extends, (C_JAVA & ~C_PLPL), st_C_javastruct
2493 implements, (C_JAVA & ~C_PLPL), st_C_javastruct
2494 interface, (C_JAVA & ~C_PLPL), st_C_struct
2495 class, 0, st_C_class
2496 namespace, C_PLPL, st_C_struct
2497 domain, C_STAR, st_C_struct
2498 union, 0, st_C_struct
2499 struct, 0, st_C_struct
2500 extern, 0, st_C_extern
2502 typedef, 0, st_C_typedef
2503 define, 0, st_C_define
2504 undef, 0, st_C_define
2505 operator, C_PLPL, st_C_operator
2506 template, 0, st_C_template
2507 # DEFUN used in emacs, the next three used in glibc (SYSCALL only for mach).
2508 DEFUN, 0, st_C_gnumacro
2509 SYSCALL, 0, st_C_gnumacro
2510 ENTRY, 0, st_C_gnumacro
2511 PSEUDO, 0, st_C_gnumacro
2512 ENUM_BF, 0, st_C_enum_bf
2513 # These are defined inside C functions, so currently they are not met.
2514 # EXFUN used in glibc, DEFVAR_* in emacs.
2515 #EXFUN, 0, st_C_gnumacro
2516 #DEFVAR_, 0, st_C_gnumacro
2518 and replace lines between %< and %> with its output, then:
2519 - remove the #if characterset check
2520 - remove any #line directives
2521 - make in_word_set static and not inline
2522 - remove any 'register' qualifications from variable decls. */
2524 /* C code produced by gperf version 3.0.1 */
2525 /* Command-line: gperf -m 5 */
2526 /* Computed positions: -k'2-3' */
2528 struct C_stab_entry
{ const char *name
; int c_ext
; enum sym_type type
; };
2529 /* maximum key range = 34, duplicates = 0 */
2532 hash (const char *str
, int len
)
2534 static char const asso_values
[] =
2536 36, 36, 36, 36, 36, 36, 36, 36, 36, 36,
2537 36, 36, 36, 36, 36, 36, 36, 36, 36, 36,
2538 36, 36, 36, 36, 36, 36, 36, 36, 36, 36,
2539 36, 36, 36, 36, 36, 36, 36, 36, 36, 36,
2540 36, 36, 36, 36, 36, 36, 36, 36, 36, 36,
2541 36, 36, 36, 36, 36, 36, 36, 36, 36, 36,
2542 36, 36, 36, 36, 36, 36, 36, 36, 36, 3,
2543 27, 36, 36, 36, 36, 36, 36, 36, 26, 36,
2544 36, 36, 36, 25, 0, 0, 36, 36, 36, 0,
2545 36, 36, 36, 36, 36, 1, 36, 16, 36, 6,
2546 23, 0, 0, 36, 22, 0, 36, 36, 5, 0,
2547 0, 15, 1, 36, 6, 36, 8, 19, 36, 16,
2548 4, 5, 36, 36, 36, 36, 36, 36, 36, 36,
2549 36, 36, 36, 36, 36, 36, 36, 36, 36, 36,
2550 36, 36, 36, 36, 36, 36, 36, 36, 36, 36,
2551 36, 36, 36, 36, 36, 36, 36, 36, 36, 36,
2552 36, 36, 36, 36, 36, 36, 36, 36, 36, 36,
2553 36, 36, 36, 36, 36, 36, 36, 36, 36, 36,
2554 36, 36, 36, 36, 36, 36, 36, 36, 36, 36,
2555 36, 36, 36, 36, 36, 36, 36, 36, 36, 36,
2556 36, 36, 36, 36, 36, 36, 36, 36, 36, 36,
2557 36, 36, 36, 36, 36, 36, 36, 36, 36, 36,
2558 36, 36, 36, 36, 36, 36, 36, 36, 36, 36,
2559 36, 36, 36, 36, 36, 36, 36, 36, 36, 36,
2560 36, 36, 36, 36, 36, 36, 36, 36, 36, 36,
2561 36, 36, 36, 36, 36, 36
2568 hval
+= asso_values
[(unsigned char) str
[2]];
2571 hval
+= asso_values
[(unsigned char) str
[1]];
2577 static struct C_stab_entry
*
2578 in_word_set (register const char *str
, register unsigned int len
)
2582 TOTAL_KEYWORDS
= 34,
2583 MIN_WORD_LENGTH
= 2,
2584 MAX_WORD_LENGTH
= 15,
2589 static struct C_stab_entry wordlist
[] =
2592 {"if", 0, st_C_ignore
},
2593 {"GTY", 0, st_C_attribute
},
2594 {"@end", 0, st_C_objend
},
2595 {"union", 0, st_C_struct
},
2596 {"define", 0, st_C_define
},
2597 {"import", (C_JAVA
& ~C_PLPL
), st_C_ignore
},
2598 {"template", 0, st_C_template
},
2599 {"operator", C_PLPL
, st_C_operator
},
2600 {"@interface", 0, st_C_objprot
},
2601 {"implements", (C_JAVA
& ~C_PLPL
), st_C_javastruct
},
2602 {"friend", C_PLPL
, st_C_ignore
},
2603 {"typedef", 0, st_C_typedef
},
2604 {"return", 0, st_C_ignore
},
2605 {"@implementation",0, st_C_objimpl
},
2606 {"@protocol", 0, st_C_objprot
},
2607 {"interface", (C_JAVA
& ~C_PLPL
), st_C_struct
},
2608 {"extern", 0, st_C_extern
},
2609 {"extends", (C_JAVA
& ~C_PLPL
), st_C_javastruct
},
2610 {"struct", 0, st_C_struct
},
2611 {"domain", C_STAR
, st_C_struct
},
2612 {"switch", 0, st_C_ignore
},
2613 {"enum", 0, st_C_enum
},
2614 {"for", 0, st_C_ignore
},
2615 {"namespace", C_PLPL
, st_C_struct
},
2616 {"class", 0, st_C_class
},
2617 {"while", 0, st_C_ignore
},
2618 {"undef", 0, st_C_define
},
2619 {"package", (C_JAVA
& ~C_PLPL
), st_C_ignore
},
2620 {"__attribute__", 0, st_C_attribute
},
2621 {"ENTRY", 0, st_C_gnumacro
},
2622 {"SYSCALL", 0, st_C_gnumacro
},
2623 {"ENUM_BF", 0, st_C_enum_bf
},
2624 {"PSEUDO", 0, st_C_gnumacro
},
2625 {"DEFUN", 0, st_C_gnumacro
}
2628 if (len
<= MAX_WORD_LENGTH
&& len
>= MIN_WORD_LENGTH
)
2630 int key
= hash (str
, len
);
2632 if (key
<= MAX_HASH_VALUE
&& key
>= 0)
2634 const char *s
= wordlist
[key
].name
;
2636 if (*str
== *s
&& !strncmp (str
+ 1, s
+ 1, len
- 1) && s
[len
] == '\0')
2637 return &wordlist
[key
];
2644 static enum sym_type
2645 C_symtype (char *str
, int len
, int c_ext
)
2647 register struct C_stab_entry
*se
= in_word_set (str
, len
);
2649 if (se
== NULL
|| (se
->c_ext
&& !(c_ext
& se
->c_ext
)))
2656 * Ignoring __attribute__ ((list))
2658 static bool inattribute
; /* looking at an __attribute__ construct */
2660 /* Ignoring ENUM_BF (type)
2663 static bool in_enum_bf
; /* inside parentheses following ENUM_BF */
2666 * C functions and variables are recognized using a simple
2667 * finite automaton. fvdef is its state variable.
2671 fvnone
, /* nothing seen */
2672 fdefunkey
, /* Emacs DEFUN keyword seen */
2673 fdefunname
, /* Emacs DEFUN name seen */
2674 foperator
, /* func: operator keyword seen (cplpl) */
2675 fvnameseen
, /* function or variable name seen */
2676 fstartlist
, /* func: just after open parenthesis */
2677 finlist
, /* func: in parameter list */
2678 flistseen
, /* func: after parameter list */
2679 fignore
, /* func: before open brace */
2680 vignore
/* var-like: ignore until ';' */
2683 static bool fvextern
; /* func or var: extern keyword seen; */
2686 * typedefs are recognized using a simple finite automaton.
2687 * typdef is its state variable.
2691 tnone
, /* nothing seen */
2692 tkeyseen
, /* typedef keyword seen */
2693 ttypeseen
, /* defined type seen */
2694 tinbody
, /* inside typedef body */
2695 tend
, /* just before typedef tag */
2696 tignore
/* junk after typedef tag */
2700 * struct-like structures (enum, struct and union) are recognized
2701 * using another simple finite automaton. `structdef' is its state
2706 snone
, /* nothing seen yet,
2707 or in struct body if bracelev > 0 */
2708 skeyseen
, /* struct-like keyword seen */
2709 stagseen
, /* struct-like tag seen */
2710 scolonseen
/* colon seen after struct-like tag */
2714 * When objdef is different from onone, objtag is the name of the class.
2716 static const char *objtag
= "<uninited>";
2719 * Yet another little state machine to deal with preprocessor lines.
2723 dnone
, /* nothing seen */
2724 dsharpseen
, /* '#' seen as first char on line */
2725 ddefineseen
, /* '#' and 'define' seen */
2726 dignorerest
/* ignore rest of line */
2730 * State machine for Objective C protocols and implementations.
2731 * Idea by Tom R.Hageman <tom@basil.icce.rug.nl> (1995)
2735 onone
, /* nothing seen */
2736 oprotocol
, /* @interface or @protocol seen */
2737 oimplementation
, /* @implementations seen */
2738 otagseen
, /* class name seen */
2739 oparenseen
, /* parenthesis before category seen */
2740 ocatseen
, /* category name seen */
2741 oinbody
, /* in @implementation body */
2742 omethodsign
, /* in @implementation body, after +/- */
2743 omethodtag
, /* after method name */
2744 omethodcolon
, /* after method colon */
2745 omethodparm
, /* after method parameter */
2746 oignore
/* wait for @end */
2751 * Use this structure to keep info about the token read, and how it
2752 * should be tagged. Used by the make_C_tag function to build a tag.
2756 char *line
; /* string containing the token */
2757 int offset
; /* where the token starts in LINE */
2758 int length
; /* token length */
2760 The previous members can be used to pass strings around for generic
2761 purposes. The following ones specifically refer to creating tags. In this
2762 case the token contained here is the pattern that will be used to create a
2765 bool valid
; /* do not create a tag; the token should be
2766 invalidated whenever a state machine is
2767 reset prematurely */
2768 bool named
; /* create a named tag */
2769 int lineno
; /* source line number of tag */
2770 long linepos
; /* source char number of tag */
2771 } token
; /* latest token read */
2774 * Variables and functions for dealing with nested structures.
2775 * Idea by Mykola Dzyuba <mdzyuba@yahoo.com> (2001)
2777 static void pushclass_above (int, char *, int);
2778 static void popclass_above (int);
2779 static void write_classname (linebuffer
*, const char *qualifier
);
2782 char **cname
; /* nested class names */
2783 int *bracelev
; /* nested class brace level */
2784 int nl
; /* class nesting level (elements used) */
2785 int size
; /* length of the array */
2786 } cstack
; /* stack for nested declaration tags */
2787 /* Current struct nesting depth (namespace, class, struct, union, enum). */
2788 #define nestlev (cstack.nl)
2789 /* After struct keyword or in struct body, not inside a nested function. */
2790 #define instruct (structdef == snone && nestlev > 0 \
2791 && bracelev == cstack.bracelev[nestlev-1] + 1)
2794 pushclass_above (int bracelev
, char *str
, int len
)
2798 popclass_above (bracelev
);
2800 if (nl
>= cstack
.size
)
2802 int size
= cstack
.size
*= 2;
2803 xrnew (cstack
.cname
, size
, char *);
2804 xrnew (cstack
.bracelev
, size
, int);
2806 assert (nl
== 0 || cstack
.bracelev
[nl
-1] < bracelev
);
2807 cstack
.cname
[nl
] = (str
== NULL
) ? NULL
: savenstr (str
, len
);
2808 cstack
.bracelev
[nl
] = bracelev
;
2813 popclass_above (int bracelev
)
2817 for (nl
= cstack
.nl
- 1;
2818 nl
>= 0 && cstack
.bracelev
[nl
] >= bracelev
;
2821 free (cstack
.cname
[nl
]);
2827 write_classname (linebuffer
*cn
, const char *qualifier
)
2830 int qlen
= strlen (qualifier
);
2832 if (cstack
.nl
== 0 || cstack
.cname
[0] == NULL
)
2836 cn
->buffer
[0] = '\0';
2840 len
= strlen (cstack
.cname
[0]);
2841 linebuffer_setlen (cn
, len
);
2842 strcpy (cn
->buffer
, cstack
.cname
[0]);
2844 for (i
= 1; i
< cstack
.nl
; i
++)
2846 char *s
= cstack
.cname
[i
];
2849 linebuffer_setlen (cn
, len
+ qlen
+ strlen (s
));
2850 len
+= sprintf (cn
->buffer
+ len
, "%s%s", qualifier
, s
);
2855 static bool consider_token (char *, int, int, int *, int, int, bool *);
2856 static void make_C_tag (bool);
2860 * checks to see if the current token is at the start of a
2861 * function or variable, or corresponds to a typedef, or
2862 * is a struct/union/enum tag, or #define, or an enum constant.
2864 * *IS_FUNC_OR_VAR gets true if the token is a function or #define macro
2865 * with args. C_EXTP points to which language we are looking at.
2876 consider_token (char *str
, int len
, int c
, int *c_extp
,
2877 int bracelev
, int parlev
, bool *is_func_or_var
)
2878 /* IN: token pointer */
2879 /* IN: token length */
2880 /* IN: first char after the token */
2881 /* IN, OUT: C extensions mask */
2882 /* IN: brace level */
2883 /* IN: parenthesis level */
2884 /* OUT: function or variable found */
2886 /* When structdef is stagseen, scolonseen, or snone with bracelev > 0,
2887 structtype is the type of the preceding struct-like keyword, and
2888 structbracelev is the brace level where it has been seen. */
2889 static enum sym_type structtype
;
2890 static int structbracelev
;
2891 static enum sym_type toktype
;
2894 toktype
= C_symtype (str
, len
, *c_extp
);
2897 * Skip __attribute__
2899 if (toktype
== st_C_attribute
)
2908 if (toktype
== st_C_enum_bf
&& definedef
== dnone
)
2915 * Advance the definedef state machine.
2920 /* We're not on a preprocessor line. */
2921 if (toktype
== st_C_gnumacro
)
2928 if (toktype
== st_C_define
)
2930 definedef
= ddefineseen
;
2934 definedef
= dignorerest
;
2939 * Make a tag for any macro, unless it is a constant
2940 * and constantypedefs is false.
2942 definedef
= dignorerest
;
2943 *is_func_or_var
= (c
== '(');
2944 if (!*is_func_or_var
&& !constantypedefs
)
2951 error ("internal error: definedef value.");
2960 if (toktype
== st_C_typedef
)
2983 if (structdef
== snone
&& fvdef
== fvnone
)
3005 case st_C_javastruct
:
3006 if (structdef
== stagseen
)
3007 structdef
= scolonseen
;
3011 if ((*c_extp
& C_AUTO
) /* automatic detection of C++ language */
3013 && definedef
== dnone
&& structdef
== snone
3014 && typdef
== tnone
&& fvdef
== fvnone
)
3015 *c_extp
= (*c_extp
| C_PLPL
) & ~C_AUTO
;
3016 if (toktype
== st_C_template
)
3023 && (typdef
== tkeyseen
3024 || (typedefs_or_cplusplus
&& structdef
== snone
)))
3026 structdef
= skeyseen
;
3027 structtype
= toktype
;
3028 structbracelev
= bracelev
;
3029 if (fvdef
== fvnameseen
)
3037 if (structdef
== skeyseen
)
3039 structdef
= stagseen
;
3043 if (typdef
!= tnone
)
3046 /* Detect Objective C constructs. */
3056 objdef
= oimplementation
;
3062 case oimplementation
:
3063 /* Save the class tag for functions or variables defined inside. */
3064 objtag
= savenstr (str
, len
);
3068 /* Save the class tag for categories. */
3069 objtag
= savenstr (str
, len
);
3071 *is_func_or_var
= true;
3075 *is_func_or_var
= true;
3083 objdef
= omethodtag
;
3084 linebuffer_setlen (&token_name
, len
);
3085 memcpy (token_name
.buffer
, str
, len
);
3086 token_name
.buffer
[len
] = '\0';
3092 objdef
= omethodparm
;
3097 objdef
= omethodtag
;
3100 int oldlen
= token_name
.len
;
3102 linebuffer_setlen (&token_name
, oldlen
+ len
);
3103 memcpy (token_name
.buffer
+ oldlen
, str
, len
);
3104 token_name
.buffer
[oldlen
+ len
] = '\0';
3110 if (toktype
== st_C_objend
)
3112 /* Memory leakage here: the string pointed by objtag is
3113 never released, because many tests would be needed to
3114 avoid breaking on incorrect input code. The amount of
3115 memory leaked here is the sum of the lengths of the
3125 /* A function, variable or enum constant? */
3147 *is_func_or_var
= true;
3151 && structdef
== snone
3152 && structtype
== st_C_enum
&& bracelev
> structbracelev
3153 /* Don't tag tokens in expressions that assign values to enum
3155 && fvdef
!= vignore
)
3156 return true; /* enum constant */
3162 fvdef
= fdefunname
; /* GNU macro */
3163 *is_func_or_var
= true;
3171 if ((strneq (str
, "asm", 3) && endtoken (str
[3]))
3172 || (strneq (str
, "__asm__", 7) && endtoken (str
[7])))
3183 if (len
>= 10 && strneq (str
+len
-10, "::operator", 10))
3185 if (*c_extp
& C_AUTO
) /* automatic detection of C++ */
3186 *c_extp
= (*c_extp
| C_PLPL
) & ~C_AUTO
;
3188 *is_func_or_var
= true;
3191 if (bracelev
> 0 && !instruct
)
3193 fvdef
= fvnameseen
; /* function or variable */
3194 *is_func_or_var
= true;
3209 * C_entries often keeps pointers to tokens or lines which are older than
3210 * the line currently read. By keeping two line buffers, and switching
3211 * them at end of line, it is possible to use those pointers.
3219 #define current_lb_is_new (newndx == curndx)
3220 #define switch_line_buffers() (curndx = 1 - curndx)
3222 #define curlb (lbs[curndx].lb)
3223 #define newlb (lbs[newndx].lb)
3224 #define curlinepos (lbs[curndx].linepos)
3225 #define newlinepos (lbs[newndx].linepos)
3227 #define plainc ((c_ext & C_EXT) == C_PLAIN)
3228 #define cplpl (c_ext & C_PLPL)
3229 #define cjava ((c_ext & C_JAVA) == C_JAVA)
3231 #define CNL_SAVE_DEFINEDEF() \
3233 curlinepos = charno; \
3234 readline (&curlb, inf); \
3235 lp = curlb.buffer; \
3242 CNL_SAVE_DEFINEDEF (); \
3243 if (savetoken.valid) \
3245 token = savetoken; \
3246 savetoken.valid = false; \
3248 definedef = dnone; \
3253 make_C_tag (bool isfun
)
3255 /* This function is never called when token.valid is false, but
3256 we must protect against invalid input or internal errors. */
3258 make_tag (token_name
.buffer
, token_name
.len
, isfun
, token
.line
,
3259 token
.offset
+token
.length
+1, token
.lineno
, token
.linepos
);
3261 { /* this branch is optimized away if !DEBUG */
3262 make_tag (concat ("INVALID TOKEN:-->", token_name
.buffer
, ""),
3263 token_name
.len
+ 17, isfun
, token
.line
,
3264 token
.offset
+token
.length
+1, token
.lineno
, token
.linepos
);
3265 error ("INVALID TOKEN");
3268 token
.valid
= false;
3272 perhaps_more_input (FILE *inf
)
3274 return !feof (inf
) && !ferror (inf
);
3280 * This routine finds functions, variables, typedefs,
3281 * #define's, enum constants and struct/union/enum definitions in
3282 * C syntax and adds them to the list.
3285 C_entries (int c_ext
, FILE *inf
)
3286 /* extension of C */
3289 register char c
; /* latest char read; '\0' for end of line */
3290 register char *lp
; /* pointer one beyond the character `c' */
3291 int curndx
, newndx
; /* indices for current and new lb */
3292 register int tokoff
; /* offset in line of start of current token */
3293 register int toklen
; /* length of current token */
3294 const char *qualifier
; /* string used to qualify names */
3295 int qlen
; /* length of qualifier */
3296 int bracelev
; /* current brace level */
3297 int bracketlev
; /* current bracket level */
3298 int parlev
; /* current parenthesis level */
3299 int attrparlev
; /* __attribute__ parenthesis level */
3300 int templatelev
; /* current template level */
3301 int typdefbracelev
; /* bracelev where a typedef struct body begun */
3302 bool incomm
, inquote
, inchar
, quotednl
, midtoken
;
3303 bool yacc_rules
; /* in the rules part of a yacc file */
3304 struct tok savetoken
= {0}; /* token saved during preprocessor handling */
3307 linebuffer_init (&lbs
[0].lb
);
3308 linebuffer_init (&lbs
[1].lb
);
3309 if (cstack
.size
== 0)
3311 cstack
.size
= (DEBUG
) ? 1 : 4;
3313 cstack
.cname
= xnew (cstack
.size
, char *);
3314 cstack
.bracelev
= xnew (cstack
.size
, int);
3317 tokoff
= toklen
= typdefbracelev
= 0; /* keep compiler quiet */
3318 curndx
= newndx
= 0;
3322 fvdef
= fvnone
; fvextern
= false; typdef
= tnone
;
3323 structdef
= snone
; definedef
= dnone
; objdef
= onone
;
3325 midtoken
= inquote
= inchar
= incomm
= quotednl
= false;
3326 token
.valid
= savetoken
.valid
= false;
3327 bracelev
= bracketlev
= parlev
= attrparlev
= templatelev
= 0;
3329 { qualifier
= "."; qlen
= 1; }
3331 { qualifier
= "::"; qlen
= 2; }
3334 while (perhaps_more_input (inf
))
3339 /* If we are at the end of the line, the next character is a
3340 '\0'; do not skip it, because it is what tells us
3341 to read the next line. */
3362 /* Newlines inside comments do not end macro definitions in
3364 CNL_SAVE_DEFINEDEF ();
3377 /* Newlines inside strings do not end macro definitions
3378 in traditional cpp, even though compilers don't
3379 usually accept them. */
3380 CNL_SAVE_DEFINEDEF ();
3390 /* Hmmm, something went wrong. */
3426 if (fvdef
!= finlist
&& fvdef
!= fignore
&& fvdef
!= vignore
)
3441 else if (/* cplpl && */ *lp
== '/')
3447 if ((c_ext
& YACC
) && *lp
== '%')
3449 /* Entering or exiting rules section in yacc file. */
3451 definedef
= dnone
; fvdef
= fvnone
; fvextern
= false;
3452 typdef
= tnone
; structdef
= snone
;
3453 midtoken
= inquote
= inchar
= incomm
= quotednl
= false;
3455 yacc_rules
= !yacc_rules
;
3461 if (definedef
== dnone
)
3464 bool cpptoken
= true;
3466 /* Look back on this line. If all blanks, or nonblanks
3467 followed by an end of comment, this is a preprocessor
3469 for (cp
= newlb
.buffer
; cp
< lp
-1; cp
++)
3470 if (!c_isspace (*cp
))
3472 if (*cp
== '*' && cp
[1] == '/')
3482 definedef
= dsharpseen
;
3483 /* This is needed for tagging enum values: when there are
3484 preprocessor conditionals inside the enum, we need to
3485 reset the value of fvdef so that the next enum value is
3486 tagged even though the one before it did not end in a
3488 if (fvdef
== vignore
&& instruct
&& parlev
== 0)
3490 if (strneq (cp
, "#if", 3) || strneq (cp
, "#el", 3))
3494 } /* if (definedef == dnone) */
3505 CNL_SAVE_DEFINEDEF ();
3512 /* Consider token only if some involved conditions are satisfied. */
3513 if (typdef
!= tignore
3514 && definedef
!= dignorerest
3517 && (definedef
!= dnone
3518 || structdef
!= scolonseen
)
3526 if (c
== ':' && *lp
== ':' && begtoken (lp
[1]))
3527 /* This handles :: in the middle,
3528 but not at the beginning of an identifier.
3529 Also, space-separated :: is not recognized. */
3531 if (c_ext
& C_AUTO
) /* automatic detection of C++ */
3532 c_ext
= (c_ext
| C_PLPL
) & ~C_AUTO
;
3536 goto still_in_token
;
3540 bool funorvar
= false;
3543 || consider_token (newlb
.buffer
+ tokoff
, toklen
, c
,
3544 &c_ext
, bracelev
, parlev
,
3547 if (fvdef
== foperator
)
3550 lp
= skip_spaces (lp
-1);
3554 && !c_isspace (*lp
) && *lp
!= '(')
3557 toklen
+= lp
- oldlp
;
3559 token
.named
= false;
3561 && nestlev
> 0 && definedef
== dnone
)
3562 /* in struct body */
3567 write_classname (&token_name
, qualifier
);
3568 len
= token_name
.len
;
3569 linebuffer_setlen (&token_name
,
3570 len
+ qlen
+ toklen
);
3571 sprintf (token_name
.buffer
+ len
, "%s%.*s",
3573 newlb
.buffer
+ tokoff
);
3577 linebuffer_setlen (&token_name
, toklen
);
3578 sprintf (token_name
.buffer
, "%.*s",
3579 toklen
, newlb
.buffer
+ tokoff
);
3583 else if (objdef
== ocatseen
)
3584 /* Objective C category */
3588 int len
= strlen (objtag
) + 2 + toklen
;
3589 linebuffer_setlen (&token_name
, len
);
3590 sprintf (token_name
.buffer
, "%s(%.*s)",
3592 newlb
.buffer
+ tokoff
);
3596 linebuffer_setlen (&token_name
, toklen
);
3597 sprintf (token_name
.buffer
, "%.*s",
3598 toklen
, newlb
.buffer
+ tokoff
);
3602 else if (objdef
== omethodtag
3603 || objdef
== omethodparm
)
3604 /* Objective C method */
3608 else if (fvdef
== fdefunname
)
3609 /* GNU DEFUN and similar macros */
3611 bool defun
= (newlb
.buffer
[tokoff
] == 'F');
3620 /* First, tag it as its C name */
3621 linebuffer_setlen (&token_name
, toklen
);
3622 memcpy (token_name
.buffer
,
3623 newlb
.buffer
+ tokoff
, toklen
);
3624 token_name
.buffer
[toklen
] = '\0';
3626 token
.lineno
= lineno
;
3627 token
.offset
= tokoff
;
3628 token
.length
= toklen
;
3629 token
.line
= newlb
.buffer
;
3630 token
.linepos
= newlinepos
;
3632 make_C_tag (funorvar
);
3634 /* Rewrite the tag so that emacs lisp DEFUNs
3635 can be found also by their elisp name */
3636 linebuffer_setlen (&token_name
, len
);
3637 memcpy (token_name
.buffer
,
3638 newlb
.buffer
+ off
, len
);
3639 token_name
.buffer
[len
] = '\0';
3642 if (token_name
.buffer
[len
] == '_')
3643 token_name
.buffer
[len
] = '-';
3644 token
.named
= defun
;
3648 linebuffer_setlen (&token_name
, toklen
);
3649 memcpy (token_name
.buffer
,
3650 newlb
.buffer
+ tokoff
, toklen
);
3651 token_name
.buffer
[toklen
] = '\0';
3652 /* Name macros and members. */
3653 token
.named
= (structdef
== stagseen
3654 || typdef
== ttypeseen
3657 && definedef
== dignorerest
)
3659 && definedef
== dnone
3660 && structdef
== snone
3663 token
.lineno
= lineno
;
3664 token
.offset
= tokoff
;
3665 token
.length
= toklen
;
3666 token
.line
= newlb
.buffer
;
3667 token
.linepos
= newlinepos
;
3670 if (definedef
== dnone
3671 && (fvdef
== fvnameseen
3672 || fvdef
== foperator
3673 || structdef
== stagseen
3675 || typdef
== ttypeseen
3676 || objdef
!= onone
))
3678 if (current_lb_is_new
)
3679 switch_line_buffers ();
3681 else if (definedef
!= dnone
3682 || fvdef
== fdefunname
3684 make_C_tag (funorvar
);
3686 else /* not yacc and consider_token failed */
3688 if (inattribute
&& fvdef
== fignore
)
3690 /* We have just met __attribute__ after a
3691 function parameter list: do not tag the
3698 } /* if (endtoken (c)) */
3699 else if (intoken (c
))
3705 } /* if (midtoken) */
3706 else if (begtoken (c
))
3714 /* This prevents tagging fb in
3715 void (__attribute__((noreturn)) *fb) (void);
3716 Fixing this is not easy and not very important. */
3720 if (plainc
|| declarations
)
3722 make_C_tag (true); /* a function */
3729 if (structdef
== stagseen
&& !cjava
)
3731 popclass_above (bracelev
);
3741 if (!yacc_rules
|| lp
== newlb
.buffer
+ 1)
3743 tokoff
= lp
- 1 - newlb
.buffer
;
3748 } /* if (begtoken) */
3749 } /* if must look at token */
3752 /* Detect end of line, colon, comma, semicolon and various braces
3753 after having handled a token.*/
3759 if (yacc_rules
&& token
.offset
== 0 && token
.valid
)
3761 make_C_tag (false); /* a yacc function */
3764 if (definedef
!= dnone
)
3770 make_C_tag (true); /* an Objective C class */
3774 objdef
= omethodcolon
;
3777 int toklen
= token_name
.len
;
3778 linebuffer_setlen (&token_name
, toklen
+ 1);
3779 strcpy (token_name
.buffer
+ toklen
, ":");
3785 if (structdef
== stagseen
)
3787 structdef
= scolonseen
;
3790 /* Should be useless, but may be work as a safety net. */
3791 if (cplpl
&& fvdef
== flistseen
)
3793 make_C_tag (true); /* a function */
3799 if (definedef
!= dnone
|| inattribute
)
3805 make_C_tag (false); /* a typedef */
3815 if (typdef
== tignore
|| cplpl
)
3819 if ((globals
&& bracelev
== 0 && (!fvextern
|| declarations
))
3820 || (members
&& instruct
))
3821 make_C_tag (false); /* a variable */
3824 token
.valid
= false;
3828 && (cplpl
|| !instruct
)
3829 && (typdef
== tnone
|| (typdef
!= tignore
&& instruct
)))
3831 && plainc
&& instruct
))
3832 make_C_tag (true); /* a function */
3838 && cplpl
&& structdef
== stagseen
)
3839 make_C_tag (false); /* forward declaration */
3841 token
.valid
= false;
3842 } /* switch (fvdef) */
3848 if (structdef
== stagseen
)
3852 if (definedef
!= dnone
|| inattribute
)
3858 make_C_tag (true); /* an Objective C method */
3873 if (instruct
&& parlev
== 0)
3884 && (!fvextern
|| declarations
))
3885 || (members
&& instruct
)))
3886 make_C_tag (false); /* a variable */
3889 if ((declarations
&& typdef
== tnone
&& !instruct
)
3890 || (members
&& typdef
!= tignore
&& instruct
))
3892 make_C_tag (true); /* a function */
3895 else if (!declarations
)
3897 token
.valid
= false;
3902 if (structdef
== stagseen
)
3906 if (definedef
!= dnone
|| inattribute
)
3908 if (structdef
== stagseen
)
3915 make_C_tag (false); /* a typedef */
3927 if ((members
&& bracelev
== 1)
3928 || (globals
&& bracelev
== 0
3929 && (!fvextern
|| declarations
)))
3930 make_C_tag (false); /* a variable */
3946 if (definedef
!= dnone
)
3948 if (objdef
== otagseen
&& parlev
== 0)
3949 objdef
= oparenseen
;
3953 if (typdef
== ttypeseen
3957 /* This handles constructs like:
3958 typedef void OperatorFun (int fun); */
3979 if (--attrparlev
== 0)
3980 inattribute
= false;
3989 if (definedef
!= dnone
)
3991 if (objdef
== ocatseen
&& parlev
== 1)
3993 make_C_tag (true); /* an Objective C category */
4009 || typdef
== ttypeseen
))
4012 make_C_tag (false); /* a typedef */
4015 else if (parlev
< 0) /* can happen due to ill-conceived #if's. */
4019 if (definedef
!= dnone
)
4021 if (typdef
== ttypeseen
)
4023 /* Whenever typdef is set to tinbody (currently only
4024 here), typdefbracelev should be set to bracelev. */
4026 typdefbracelev
= bracelev
;
4031 if (cplpl
&& !class_qualify
)
4033 /* Remove class and namespace qualifiers from the token,
4034 leaving only the method/member name. */
4035 char *cc
, *uqname
= token_name
.buffer
;
4036 char *tok_end
= token_name
.buffer
+ token_name
.len
;
4038 for (cc
= token_name
.buffer
; cc
< tok_end
; cc
++)
4040 if (*cc
== ':' && cc
[1] == ':')
4046 if (uqname
> token_name
.buffer
)
4048 int uqlen
= strlen (uqname
);
4049 linebuffer_setlen (&token_name
, uqlen
);
4050 memmove (token_name
.buffer
, uqname
, uqlen
+ 1);
4053 make_C_tag (true); /* a function */
4062 make_C_tag (true); /* an Objective C class */
4067 make_C_tag (true); /* an Objective C method */
4071 /* Neutralize `extern "C" {' grot. */
4072 if (bracelev
== 0 && structdef
== snone
&& nestlev
== 0
4082 case skeyseen
: /* unnamed struct */
4083 pushclass_above (bracelev
, NULL
, 0);
4086 case stagseen
: /* named struct or enum */
4087 case scolonseen
: /* a class */
4088 pushclass_above (bracelev
,token
.line
+token
.offset
, token
.length
);
4090 make_C_tag (false); /* a struct or enum */
4098 if (definedef
!= dnone
)
4100 if (fvdef
== fstartlist
)
4102 fvdef
= fvnone
; /* avoid tagging `foo' in `foo (*bar()) ()' */
4103 token
.valid
= false;
4107 if (definedef
!= dnone
)
4110 if (!ignoreindent
&& lp
== newlb
.buffer
+ 1)
4113 token
.valid
= false; /* unexpected value, token unreliable */
4114 bracelev
= 0; /* reset brace level if first column */
4115 parlev
= 0; /* also reset paren level, just in case... */
4117 else if (bracelev
< 0)
4119 token
.valid
= false; /* something gone amiss, token unreliable */
4122 if (bracelev
== 0 && fvdef
== vignore
)
4123 fvdef
= fvnone
; /* end of function */
4124 popclass_above (bracelev
);
4126 /* Only if typdef == tinbody is typdefbracelev significant. */
4127 if (typdef
== tinbody
&& bracelev
<= typdefbracelev
)
4129 assert (bracelev
== typdefbracelev
);
4134 if (definedef
!= dnone
)
4144 if ((members
&& bracelev
== 1)
4145 || (globals
&& bracelev
== 0 && (!fvextern
|| declarations
)))
4146 make_C_tag (false); /* a variable */
4154 && (structdef
== stagseen
|| fvdef
== fvnameseen
))
4161 if (templatelev
> 0)
4169 if (objdef
== oinbody
&& bracelev
== 0)
4171 objdef
= omethodsign
;
4176 case '#': case '~': case '&': case '%': case '/':
4177 case '|': case '^': case '!': case '.': case '?':
4178 if (definedef
!= dnone
)
4180 /* These surely cannot follow a function tag in C. */
4193 if (objdef
== otagseen
)
4195 make_C_tag (true); /* an Objective C class */
4198 /* If a macro spans multiple lines don't reset its state. */
4200 CNL_SAVE_DEFINEDEF ();
4206 } /* while not eof */
4208 free (lbs
[0].lb
.buffer
);
4209 free (lbs
[1].lb
.buffer
);
4213 * Process either a C++ file or a C file depending on the setting
4217 default_C_entries (FILE *inf
)
4219 C_entries (cplusplus
? C_PLPL
: C_AUTO
, inf
);
4222 /* Always do plain C. */
4224 plain_C_entries (FILE *inf
)
4229 /* Always do C++. */
4231 Cplusplus_entries (FILE *inf
)
4233 C_entries (C_PLPL
, inf
);
4236 /* Always do Java. */
4238 Cjava_entries (FILE *inf
)
4240 C_entries (C_JAVA
, inf
);
4245 Cstar_entries (FILE *inf
)
4247 C_entries (C_STAR
, inf
);
4250 /* Always do Yacc. */
4252 Yacc_entries (FILE *inf
)
4254 C_entries (YACC
, inf
);
4258 /* Useful macros. */
4259 #define LOOP_ON_INPUT_LINES(file_pointer, line_buffer, char_pointer) \
4260 while (perhaps_more_input (file_pointer) \
4261 && (readline (&(line_buffer), file_pointer), \
4262 (char_pointer) = (line_buffer).buffer, \
4265 #define LOOKING_AT(cp, kw) /* kw is the keyword, a literal string */ \
4266 ((assert ("" kw), true) /* syntax error if not a literal string */ \
4267 && strneq ((cp), kw, sizeof (kw)-1) /* cp points at kw */ \
4268 && notinname ((cp)[sizeof (kw)-1]) /* end of kw */ \
4269 && ((cp) = skip_spaces ((cp) + sizeof (kw) - 1), true)) /* skip spaces */
4271 /* Similar to LOOKING_AT but does not use notinname, does not skip */
4272 #define LOOKING_AT_NOCASE(cp, kw) /* the keyword is a literal string */ \
4273 ((assert ("" kw), true) /* syntax error if not a literal string */ \
4274 && strncaseeq ((cp), kw, sizeof (kw)-1) /* cp points at kw */ \
4275 && ((cp) += sizeof (kw) - 1, true)) /* skip spaces */
4278 * Read a file, but do no processing. This is used to do regexp
4279 * matching on files that have no language defined.
4282 just_read_file (FILE *inf
)
4284 while (perhaps_more_input (inf
))
4285 readline (&lb
, inf
);
4289 /* Fortran parsing */
4291 static void F_takeprec (void);
4292 static void F_getit (FILE *);
4297 dbp
= skip_spaces (dbp
);
4301 dbp
= skip_spaces (dbp
);
4302 if (strneq (dbp
, "(*)", 3))
4307 if (!c_isdigit (*dbp
))
4309 --dbp
; /* force failure */
4314 while (c_isdigit (*dbp
));
4322 dbp
= skip_spaces (dbp
);
4325 readline (&lb
, inf
);
4330 dbp
= skip_spaces (dbp
);
4332 if (!c_isalpha (*dbp
) && *dbp
!= '_' && *dbp
!= '$')
4334 for (cp
= dbp
+ 1; *cp
!= '\0' && intoken (*cp
); cp
++)
4336 make_tag (dbp
, cp
-dbp
, true,
4337 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4342 Fortran_functions (FILE *inf
)
4344 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
4347 dbp
++; /* Ratfor escape to fortran */
4348 dbp
= skip_spaces (dbp
);
4352 if (LOOKING_AT_NOCASE (dbp
, "recursive"))
4353 dbp
= skip_spaces (dbp
);
4355 if (LOOKING_AT_NOCASE (dbp
, "pure"))
4356 dbp
= skip_spaces (dbp
);
4358 if (LOOKING_AT_NOCASE (dbp
, "elemental"))
4359 dbp
= skip_spaces (dbp
);
4361 switch (c_tolower (*dbp
))
4364 if (nocase_tail ("integer"))
4368 if (nocase_tail ("real"))
4372 if (nocase_tail ("logical"))
4376 if (nocase_tail ("complex") || nocase_tail ("character"))
4380 if (nocase_tail ("double"))
4382 dbp
= skip_spaces (dbp
);
4385 if (nocase_tail ("precision"))
4391 dbp
= skip_spaces (dbp
);
4394 switch (c_tolower (*dbp
))
4397 if (nocase_tail ("function"))
4401 if (nocase_tail ("subroutine"))
4405 if (nocase_tail ("entry"))
4409 if (nocase_tail ("blockdata") || nocase_tail ("block data"))
4411 dbp
= skip_spaces (dbp
);
4412 if (*dbp
== '\0') /* assume un-named */
4413 make_tag ("blockdata", 9, true,
4414 lb
.buffer
, dbp
- lb
.buffer
, lineno
, linecharno
);
4416 F_getit (inf
); /* look for name */
4425 * Go language support
4426 * Original code by Xi Lu <lx@shellcodes.org> (2016)
4429 Go_functions(FILE *inf
)
4433 LOOP_ON_INPUT_LINES(inf
, lb
, cp
)
4435 cp
= skip_spaces (cp
);
4437 if (LOOKING_AT (cp
, "package"))
4440 while (!notinname (*cp
) && *cp
!= '\0')
4442 make_tag (name
, cp
- name
, false, lb
.buffer
,
4443 cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4445 else if (LOOKING_AT (cp
, "func"))
4447 /* Go implementation of interface, such as:
4448 func (n *Integer) Add(m Integer) ...
4449 skip `(n *Integer)` part.
4455 cp
= skip_spaces (cp
+1);
4462 while (!notinname (*cp
))
4465 make_tag (name
, cp
- name
, true, lb
.buffer
,
4466 cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4469 else if (members
&& LOOKING_AT (cp
, "type"))
4473 /* Ignore the likes of the following:
4481 while (!notinname (*cp
) && *cp
!= '\0')
4484 make_tag (name
, cp
- name
, false, lb
.buffer
,
4485 cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4494 * Philippe Waroquiers (1998)
4497 /* Once we are positioned after an "interesting" keyword, let's get
4498 the real tag value necessary. */
4500 Ada_getit (FILE *inf
, const char *name_qualifier
)
4506 while (perhaps_more_input (inf
))
4508 dbp
= skip_spaces (dbp
);
4510 || (dbp
[0] == '-' && dbp
[1] == '-'))
4512 readline (&lb
, inf
);
4515 switch (c_tolower (*dbp
))
4518 if (nocase_tail ("body"))
4520 /* Skipping body of procedure body or package body or ....
4521 resetting qualifier to body instead of spec. */
4522 name_qualifier
= "/b";
4527 /* Skipping type of task type or protected type ... */
4528 if (nocase_tail ("type"))
4535 for (cp
= dbp
; *cp
!= '\0' && *cp
!= '"'; cp
++)
4540 dbp
= skip_spaces (dbp
);
4542 c_isalnum (*cp
) || *cp
== '_' || *cp
== '.';
4550 name
= concat (dbp
, name_qualifier
, "");
4552 make_tag (name
, strlen (name
), true,
4553 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4562 Ada_funcs (FILE *inf
)
4564 bool inquote
= false;
4565 bool skip_till_semicolumn
= false;
4567 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
4569 while (*dbp
!= '\0')
4571 /* Skip a string i.e. "abcd". */
4572 if (inquote
|| (*dbp
== '"'))
4574 dbp
= strchr (dbp
+ !inquote
, '"');
4579 continue; /* advance char */
4584 break; /* advance line */
4588 /* Skip comments. */
4589 if (dbp
[0] == '-' && dbp
[1] == '-')
4590 break; /* advance line */
4592 /* Skip character enclosed in single quote i.e. 'a'
4593 and skip single quote starting an attribute i.e. 'Image. */
4602 if (skip_till_semicolumn
)
4605 skip_till_semicolumn
= false;
4607 continue; /* advance char */
4610 /* Search for beginning of a token. */
4611 if (!begtoken (*dbp
))
4614 continue; /* advance char */
4617 /* We are at the beginning of a token. */
4618 switch (c_tolower (*dbp
))
4621 if (!packages_only
&& nocase_tail ("function"))
4622 Ada_getit (inf
, "/f");
4624 break; /* from switch */
4625 continue; /* advance char */
4627 if (!packages_only
&& nocase_tail ("procedure"))
4628 Ada_getit (inf
, "/p");
4629 else if (nocase_tail ("package"))
4630 Ada_getit (inf
, "/s");
4631 else if (nocase_tail ("protected")) /* protected type */
4632 Ada_getit (inf
, "/t");
4634 break; /* from switch */
4635 continue; /* advance char */
4638 if (typedefs
&& !packages_only
&& nocase_tail ("use"))
4640 /* when tagging types, avoid tagging use type Pack.Typename;
4641 for this, we will skip everything till a ; */
4642 skip_till_semicolumn
= true;
4643 continue; /* advance char */
4647 if (!packages_only
&& nocase_tail ("task"))
4648 Ada_getit (inf
, "/k");
4649 else if (typedefs
&& !packages_only
&& nocase_tail ("type"))
4651 Ada_getit (inf
, "/t");
4652 while (*dbp
!= '\0')
4656 break; /* from switch */
4657 continue; /* advance char */
4660 /* Look for the end of the token. */
4661 while (!endtoken (*dbp
))
4664 } /* advance char */
4665 } /* advance line */
4670 * Unix and microcontroller assembly tag handling
4671 * Labels: /^[a-zA-Z_.$][a-zA_Z0-9_.$]*[: ^I^J]/
4672 * Idea by Bob Weiner, Motorola Inc. (1994)
4675 Asm_labels (FILE *inf
)
4679 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4681 /* If first char is alphabetic or one of [_.$], test for colon
4682 following identifier. */
4683 if (c_isalpha (*cp
) || *cp
== '_' || *cp
== '.' || *cp
== '$')
4685 /* Read past label. */
4687 while (c_isalnum (*cp
) || *cp
== '_' || *cp
== '.' || *cp
== '$')
4689 if (*cp
== ':' || c_isspace (*cp
))
4690 /* Found end of label, so copy it and add it to the table. */
4691 make_tag (lb
.buffer
, cp
- lb
.buffer
, true,
4692 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4700 * Perl sub names: /^sub[ \t\n]+[^ \t\n{]+/
4701 * /^use constant[ \t\n]+[^ \t\n{=,;]+/
4702 * Perl variable names: /^(my|local).../
4703 * Original code by Bart Robinson <lomew@cs.utah.edu> (1995)
4704 * Additions by Michael Ernst <mernst@alum.mit.edu> (1997)
4705 * Ideas by Kai Großjohann <Kai.Grossjohann@CS.Uni-Dortmund.DE> (2001)
4708 Perl_functions (FILE *inf
)
4710 char *package
= savestr ("main"); /* current package name */
4713 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4715 cp
= skip_spaces (cp
);
4717 if (LOOKING_AT (cp
, "package"))
4720 get_tag (cp
, &package
);
4722 else if (LOOKING_AT (cp
, "sub"))
4728 while (!notinname (*cp
))
4731 continue; /* nothing found */
4732 pos
= strchr (sp
, ':');
4733 if (pos
&& pos
< cp
&& pos
[1] == ':')
4735 /* The name is already qualified. */
4738 char *q
= pos
+ 2, *qpos
;
4739 while ((qpos
= strchr (q
, ':')) != NULL
4745 make_tag (sp
, cp
- sp
, true,
4746 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4748 else if (class_qualify
)
4751 char savechar
, *name
;
4755 name
= concat (package
, "::", sp
);
4757 make_tag (name
, strlen (name
), true,
4758 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4762 make_tag (sp
, cp
- sp
, true,
4763 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4765 else if (LOOKING_AT (cp
, "use constant")
4766 || LOOKING_AT (cp
, "use constant::defer"))
4768 /* For hash style multi-constant like
4769 use constant { FOO => 123,
4771 only the first FOO is picked up. Parsing across the value
4772 expressions would be difficult in general, due to possible nested
4773 hashes, here-documents, etc. */
4775 cp
= skip_spaces (cp
+1);
4778 else if (globals
) /* only if we are tagging global vars */
4780 /* Skip a qualifier, if any. */
4781 bool qual
= LOOKING_AT (cp
, "my") || LOOKING_AT (cp
, "local");
4782 /* After "my" or "local", but before any following paren or space. */
4783 char *varstart
= cp
;
4785 if (qual
/* should this be removed? If yes, how? */
4786 && (*cp
== '$' || *cp
== '@' || *cp
== '%'))
4791 while (c_isalnum (*cp
) || *cp
== '_');
4795 /* Should be examining a variable list at this point;
4796 could insist on seeing an open parenthesis. */
4797 while (*cp
!= '\0' && *cp
!= ';' && *cp
!= '=' && *cp
!= ')')
4803 make_tag (varstart
, cp
- varstart
, false,
4804 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4813 * Look for /^[\t]*def[ \t\n]+[^ \t\n(:]+/ or /^class[ \t\n]+[^ \t\n(:]+/
4814 * Idea by Eric S. Raymond <esr@thyrsus.com> (1997)
4815 * More ideas by seb bacon <seb@jamkit.com> (2002)
4818 Python_functions (FILE *inf
)
4822 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4824 cp
= skip_spaces (cp
);
4825 if (LOOKING_AT (cp
, "def") || LOOKING_AT (cp
, "class"))
4828 while (!notinname (*cp
) && *cp
!= ':')
4830 make_tag (name
, cp
- name
, true,
4831 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4838 * Original code by Xi Lu <lx@shellcodes.org> (2015)
4841 Ruby_functions (FILE *inf
)
4844 bool reader
= false, writer
= false, alias
= false, continuation
= false;
4846 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4848 bool is_class
= false;
4849 bool is_method
= false;
4852 cp
= skip_spaces (cp
);
4855 && c_isalpha (*cp
) && c_isupper (*cp
))
4857 char *bp
, *colon
= NULL
;
4861 for (cp
++; c_isalnum (*cp
) || *cp
== '_' || *cp
== ':'; cp
++)
4868 bp
= skip_spaces (cp
);
4869 if (*bp
== '=' && !(bp
[1] == '=' || bp
[1] == '>'))
4871 if (colon
&& !c_isspace (colon
[1]))
4873 make_tag (name
, cp
- name
, false,
4874 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4878 else if (!continuation
4879 /* Modules, classes, methods. */
4880 && ((is_method
= LOOKING_AT (cp
, "def"))
4881 || (is_class
= LOOKING_AT (cp
, "class"))
4882 || LOOKING_AT (cp
, "module")))
4884 const char self_name
[] = "self.";
4885 const size_t self_size1
= sizeof (self_name
) - 1;
4889 /* Ruby method names can end in a '='. Also, operator overloading can
4890 define operators whose names include '='. */
4891 while (!notinname (*cp
) || *cp
== '=')
4894 /* Remove "self." from the method name. */
4895 if (cp
- name
> self_size1
4896 && strneq (name
, self_name
, self_size1
))
4899 /* Remove the class/module qualifiers from method names. */
4904 for (q
= name
; q
< cp
&& *q
!= '.'; q
++)
4906 if (q
< cp
- 1) /* punt if we see just "FOO." */
4910 /* Don't tag singleton classes. */
4911 if (is_class
&& strneq (name
, "<<", 2) && cp
== name
+ 2)
4914 make_tag (name
, cp
- name
, true,
4915 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4919 /* Tag accessors and aliases. */
4922 reader
= writer
= alias
= false;
4924 while (*cp
&& *cp
!= '#')
4928 reader
= writer
= alias
= false;
4929 if (LOOKING_AT (cp
, "attr_reader"))
4931 else if (LOOKING_AT (cp
, "attr_writer"))
4933 else if (LOOKING_AT (cp
, "attr_accessor"))
4938 else if (LOOKING_AT (cp
, "alias_method"))
4941 if (reader
|| writer
|| alias
)
4946 cp
= skip_spaces (cp
);
4948 cp
= skip_spaces (cp
+ 1);
4950 cp
= skip_name (cp
);
4956 make_tag (np
, cp
- np
, true,
4957 lb
.buffer
, cp
- lb
.buffer
+ 1,
4958 lineno
, linecharno
);
4959 continuation
= false;
4963 size_t name_len
= cp
- np
+ 1;
4964 char *wr_name
= xnew (name_len
+ 1, char);
4966 memcpy (wr_name
, np
, name_len
- 1);
4967 memcpy (wr_name
+ name_len
- 1, "=", 2);
4968 pfnote (wr_name
, true, lb
.buffer
, cp
- lb
.buffer
+ 1,
4969 lineno
, linecharno
);
4970 continuation
= false;
4975 make_tag (np
, cp
- np
, true,
4976 lb
.buffer
, cp
- lb
.buffer
+ 1,
4977 lineno
, linecharno
);
4978 continuation
= false;
4979 while (*cp
&& *cp
!= '#' && *cp
!= ';')
4982 continuation
= true;
4983 else if (!c_isspace (*cp
))
4984 continuation
= false;
4988 continuation
= false;
4990 cp
= skip_spaces (cp
);
4993 : (continuation
= (*cp
== ',')))
4994 && (cp
= skip_spaces (cp
+ 1), *cp
&& *cp
!= '#'));
4997 cp
= skip_name (cp
);
4998 while (*cp
&& *cp
!= '#' && notinname (*cp
))
5009 * - /^[ \t]*function[ \t\n]+[^ \t\n(]+/
5010 * - /^[ \t]*class[ \t\n]+[^ \t\n]+/
5011 * - /^[ \t]*define\(\"[^\"]+/
5012 * Only with --members:
5013 * - /^[ \t]*var[ \t\n]+\$[^ \t\n=;]/
5014 * Idea by Diez B. Roggisch (2001)
5017 PHP_functions (FILE *inf
)
5020 bool search_identifier
= false;
5022 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5024 cp
= skip_spaces (cp
);
5026 if (search_identifier
5029 while (!notinname (*cp
))
5031 make_tag (name
, cp
- name
, true,
5032 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
5033 search_identifier
= false;
5035 else if (LOOKING_AT (cp
, "function"))
5038 cp
= skip_spaces (cp
+1);
5042 while (!notinname (*cp
))
5044 make_tag (name
, cp
- name
, true,
5045 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
5048 search_identifier
= true;
5050 else if (LOOKING_AT (cp
, "class"))
5055 while (*cp
!= '\0' && !c_isspace (*cp
))
5057 make_tag (name
, cp
- name
, false,
5058 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
5061 search_identifier
= true;
5063 else if (strneq (cp
, "define", 6)
5064 && (cp
= skip_spaces (cp
+6))
5066 && (*cp
== '"' || *cp
== '\''))
5070 while (*cp
!= quote
&& *cp
!= '\0')
5072 make_tag (name
, cp
- name
, false,
5073 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
5076 && LOOKING_AT (cp
, "var")
5080 while (!notinname (*cp
))
5082 make_tag (name
, cp
- name
, false,
5083 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
5090 * Cobol tag functions
5091 * We could look for anything that could be a paragraph name.
5092 * i.e. anything that starts in column 8 is one word and ends in a full stop.
5093 * Idea by Corny de Souza (1993)
5096 Cobol_paragraphs (FILE *inf
)
5098 register char *bp
, *ep
;
5100 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
5106 /* If eoln, compiler option or comment ignore whole line. */
5107 if (bp
[-1] != ' ' || !c_isalnum (bp
[0]))
5110 for (ep
= bp
; c_isalnum (*ep
) || *ep
== '-'; ep
++)
5113 make_tag (bp
, ep
- bp
, true,
5114 lb
.buffer
, ep
- lb
.buffer
+ 1, lineno
, linecharno
);
5121 * Ideas by Assar Westerlund <assar@sics.se> (2001)
5124 Makefile_targets (FILE *inf
)
5128 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
5130 if (*bp
== '\t' || *bp
== '#')
5132 while (*bp
!= '\0' && *bp
!= '=' && *bp
!= ':')
5134 if (*bp
== ':' || (globals
&& *bp
== '='))
5136 /* We should detect if there is more than one tag, but we do not.
5137 We just skip initial and final spaces. */
5138 char * namestart
= skip_spaces (lb
.buffer
);
5139 while (--bp
> namestart
)
5140 if (!notinname (*bp
))
5142 make_tag (namestart
, bp
- namestart
+ 1, true,
5143 lb
.buffer
, bp
- lb
.buffer
+ 2, lineno
, linecharno
);
5151 * Original code by Mosur K. Mohan (1989)
5153 * Locates tags for procedures & functions. Doesn't do any type- or
5154 * var-definitions. It does look for the keyword "extern" or
5155 * "forward" immediately following the procedure statement; if found,
5156 * the tag is skipped.
5159 Pascal_functions (FILE *inf
)
5161 linebuffer tline
; /* mostly copied from C_entries */
5163 int save_lineno
, namelen
, taglen
;
5166 bool /* each of these flags is true if: */
5167 incomment
, /* point is inside a comment */
5168 inquote
, /* point is inside '..' string */
5169 get_tagname
, /* point is after PROCEDURE/FUNCTION
5170 keyword, so next item = potential tag */
5171 found_tag
, /* point is after a potential tag */
5172 inparms
, /* point is within parameter-list */
5173 verify_tag
; /* point has passed the parm-list, so the
5174 next token will determine whether this
5175 is a FORWARD/EXTERN to be ignored, or
5176 whether it is a real tag */
5178 save_lcno
= save_lineno
= namelen
= taglen
= 0; /* keep compiler quiet */
5179 name
= NULL
; /* keep compiler quiet */
5182 linebuffer_init (&tline
);
5184 incomment
= inquote
= false;
5185 found_tag
= false; /* have a proc name; check if extern */
5186 get_tagname
= false; /* found "procedure" keyword */
5187 inparms
= false; /* found '(' after "proc" */
5188 verify_tag
= false; /* check if "extern" is ahead */
5191 while (perhaps_more_input (inf
)) /* long main loop to get next char */
5194 if (c
== '\0') /* if end of line */
5196 readline (&lb
, inf
);
5200 if (!((found_tag
&& verify_tag
)
5202 c
= *dbp
++; /* only if don't need *dbp pointing
5203 to the beginning of the name of
5204 the procedure or function */
5208 if (c
== '}') /* within { } comments */
5210 else if (c
== '*' && *dbp
== ')') /* within (* *) comments */
5227 inquote
= true; /* found first quote */
5229 case '{': /* found open { comment */
5233 if (*dbp
== '*') /* found open (* comment */
5238 else if (found_tag
) /* found '(' after tag, i.e., parm-list */
5241 case ')': /* end of parms list */
5246 if (found_tag
&& !inparms
) /* end of proc or fn stmt */
5253 if (found_tag
&& verify_tag
&& (*dbp
!= ' '))
5255 /* Check if this is an "extern" declaration. */
5258 if (c_tolower (*dbp
) == 'e')
5260 if (nocase_tail ("extern")) /* superfluous, really! */
5266 else if (c_tolower (*dbp
) == 'f')
5268 if (nocase_tail ("forward")) /* check for forward reference */
5274 if (found_tag
&& verify_tag
) /* not external proc, so make tag */
5278 make_tag (name
, namelen
, true,
5279 tline
.buffer
, taglen
, save_lineno
, save_lcno
);
5283 if (get_tagname
) /* grab name of proc or fn */
5290 /* Find block name. */
5291 for (cp
= dbp
+ 1; *cp
!= '\0' && !endtoken (*cp
); cp
++)
5294 /* Save all values for later tagging. */
5295 linebuffer_setlen (&tline
, lb
.len
);
5296 strcpy (tline
.buffer
, lb
.buffer
);
5297 save_lineno
= lineno
;
5298 save_lcno
= linecharno
;
5299 name
= tline
.buffer
+ (dbp
- lb
.buffer
);
5301 taglen
= cp
- lb
.buffer
+ 1;
5303 dbp
= cp
; /* set dbp to e-o-token */
5304 get_tagname
= false;
5308 /* And proceed to check for "extern". */
5310 else if (!incomment
&& !inquote
&& !found_tag
)
5312 /* Check for proc/fn keywords. */
5313 switch (c_tolower (c
))
5316 if (nocase_tail ("rocedure")) /* c = 'p', dbp has advanced */
5320 if (nocase_tail ("unction"))
5325 } /* while not eof */
5327 free (tline
.buffer
);
5332 * Lisp tag functions
5333 * look for (def or (DEF, quote or QUOTE
5336 static void L_getit (void);
5341 if (*dbp
== '\'') /* Skip prefix quote */
5343 else if (*dbp
== '(')
5346 /* Try to skip "(quote " */
5347 if (!LOOKING_AT (dbp
, "quote") && !LOOKING_AT (dbp
, "QUOTE"))
5348 /* Ok, then skip "(" before name in (defstruct (foo)) */
5349 dbp
= skip_spaces (dbp
);
5351 get_lispy_tag (dbp
);
5355 Lisp_functions (FILE *inf
)
5357 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
5362 /* "(defvar foo)" is a declaration rather than a definition. */
5366 if (LOOKING_AT (p
, "defvar"))
5368 p
= skip_name (p
); /* past var name */
5369 p
= skip_spaces (p
);
5375 if (strneq (dbp
+ 1, "cl-", 3) || strneq (dbp
+ 1, "CL-", 3))
5378 if (strneq (dbp
+1, "def", 3) || strneq (dbp
+1, "DEF", 3))
5380 dbp
= skip_non_spaces (dbp
);
5381 dbp
= skip_spaces (dbp
);
5386 /* Check for (foo::defmumble name-defined ... */
5389 while (!notinname (*dbp
) && *dbp
!= ':');
5394 while (*dbp
== ':');
5396 if (strneq (dbp
, "def", 3) || strneq (dbp
, "DEF", 3))
5398 dbp
= skip_non_spaces (dbp
);
5399 dbp
= skip_spaces (dbp
);
5409 * Lua script language parsing
5410 * Original code by David A. Capello <dacap@users.sourceforge.net> (2004)
5412 * "function" and "local function" are tags if they start at column 1.
5415 Lua_functions (FILE *inf
)
5419 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
5421 bp
= skip_spaces (bp
);
5422 if (bp
[0] != 'f' && bp
[0] != 'l')
5425 (void)LOOKING_AT (bp
, "local"); /* skip possible "local" */
5427 if (LOOKING_AT (bp
, "function"))
5429 char *tag_name
, *tp_dot
, *tp_colon
;
5431 get_tag (bp
, &tag_name
);
5432 /* If the tag ends with ".foo" or ":foo", make an additional tag for
5434 tp_dot
= strrchr (tag_name
, '.');
5435 tp_colon
= strrchr (tag_name
, ':');
5436 if (tp_dot
|| tp_colon
)
5438 char *p
= tp_dot
> tp_colon
? tp_dot
: tp_colon
;
5439 int len_add
= p
- tag_name
+ 1;
5441 get_tag (bp
+ len_add
, NULL
);
5450 * Just look for lines where the first character is '/'
5451 * Also look at "defineps" for PSWrap
5453 * Richard Mlynarik <mly@adoc.xerox.com> (1997)
5454 * Masatake Yamato <masata-y@is.aist-nara.ac.jp> (1999)
5457 PS_functions (FILE *inf
)
5459 register char *bp
, *ep
;
5461 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
5466 *ep
!= '\0' && *ep
!= ' ' && *ep
!= '{';
5469 make_tag (bp
, ep
- bp
, true,
5470 lb
.buffer
, ep
- lb
.buffer
+ 1, lineno
, linecharno
);
5472 else if (LOOKING_AT (bp
, "defineps"))
5480 * Ignore anything after \ followed by space or in ( )
5481 * Look for words defined by :
5482 * Look for constant, code, create, defer, value, and variable
5483 * OBP extensions: Look for buffer:, field,
5484 * Ideas by Eduardo Horvath <eeh@netbsd.org> (2004)
5487 Forth_words (FILE *inf
)
5491 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
5492 while ((bp
= skip_spaces (bp
))[0] != '\0')
5493 if (bp
[0] == '\\' && c_isspace (bp
[1]))
5494 break; /* read next line */
5495 else if (bp
[0] == '(' && c_isspace (bp
[1]))
5496 do /* skip to ) or eol */
5498 while (*bp
!= ')' && *bp
!= '\0');
5499 else if (((bp
[0] == ':' && c_isspace (bp
[1]) && bp
++)
5500 || LOOKING_AT_NOCASE (bp
, "constant")
5501 || LOOKING_AT_NOCASE (bp
, "2constant")
5502 || LOOKING_AT_NOCASE (bp
, "fconstant")
5503 || LOOKING_AT_NOCASE (bp
, "code")
5504 || LOOKING_AT_NOCASE (bp
, "create")
5505 || LOOKING_AT_NOCASE (bp
, "defer")
5506 || LOOKING_AT_NOCASE (bp
, "value")
5507 || LOOKING_AT_NOCASE (bp
, "2value")
5508 || LOOKING_AT_NOCASE (bp
, "fvalue")
5509 || LOOKING_AT_NOCASE (bp
, "variable")
5510 || LOOKING_AT_NOCASE (bp
, "2variable")
5511 || LOOKING_AT_NOCASE (bp
, "fvariable")
5512 || LOOKING_AT_NOCASE (bp
, "buffer:")
5513 || LOOKING_AT_NOCASE (bp
, "field:")
5514 || LOOKING_AT_NOCASE (bp
, "+field")
5515 || LOOKING_AT_NOCASE (bp
, "field") /* not standard? */
5516 || LOOKING_AT_NOCASE (bp
, "begin-structure")
5517 || LOOKING_AT_NOCASE (bp
, "synonym")
5519 && c_isspace (bp
[0]))
5521 /* Yay! A definition! */
5522 char* name_start
= skip_spaces (bp
);
5523 char* name_end
= skip_non_spaces (name_start
);
5524 if (name_start
< name_end
)
5525 make_tag (name_start
, name_end
- name_start
,
5526 true, lb
.buffer
, name_end
- lb
.buffer
,
5527 lineno
, linecharno
);
5531 bp
= skip_non_spaces (bp
);
5536 * Scheme tag functions
5537 * look for (def... xyzzy
5539 * (def ... ((...(xyzzy ....
5541 * Original code by Ken Haase (1985?)
5544 Scheme_functions (FILE *inf
)
5548 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
5550 if (strneq (bp
, "(def", 4) || strneq (bp
, "(DEF", 4))
5552 bp
= skip_non_spaces (bp
+4);
5553 /* Skip over open parens and white space.
5554 Don't continue past '\0' or '='. */
5555 while (*bp
&& notinname (*bp
) && *bp
!= '=')
5559 if (LOOKING_AT (bp
, "(SET!") || LOOKING_AT (bp
, "(set!"))
5565 /* Find tags in TeX and LaTeX input files. */
5567 /* TEX_toktab is a table of TeX control sequences that define tags.
5568 * Each entry records one such control sequence.
5570 * Original code from who knows whom.
5572 * Stefan Monnier (2002)
5575 static linebuffer
*TEX_toktab
= NULL
; /* Table with tag tokens */
5577 /* Default set of control sequences to put into TEX_toktab.
5578 The value of environment var TEXTAGS is prepended to this. */
5579 static const char *TEX_defenv
= "\
5580 :chapter:section:subsection:subsubsection:eqno:label:ref:cite:bibitem\
5581 :part:appendix:entry:index:def\
5582 :newcommand:renewcommand:newenvironment:renewenvironment";
5584 static void TEX_decode_env (const char *, const char *);
5587 * TeX/LaTeX scanning loop.
5590 TeX_commands (FILE *inf
)
5595 char TEX_esc
= '\0';
5596 char TEX_opgrp
, TEX_clgrp
;
5598 /* Initialize token table once from environment. */
5599 if (TEX_toktab
== NULL
)
5600 TEX_decode_env ("TEXTAGS", TEX_defenv
);
5602 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5604 /* Look at each TEX keyword in line. */
5607 /* Look for a TEX escape. */
5611 if (c
== '\0' || c
== '%')
5614 /* Select either \ or ! as escape character, whichever comes
5615 first outside a comment. */
5636 for (key
= TEX_toktab
; key
->buffer
!= NULL
; key
++)
5637 if (strneq (cp
, key
->buffer
, key
->len
))
5640 int namelen
, linelen
;
5643 cp
= skip_spaces (cp
+ key
->len
);
5644 if (*cp
== TEX_opgrp
)
5650 (!c_isspace (*p
) && *p
!= '#' &&
5651 *p
!= TEX_opgrp
&& *p
!= TEX_clgrp
);
5656 if (!opgrp
|| *p
== TEX_clgrp
)
5658 while (*p
!= '\0' && *p
!= TEX_opgrp
&& *p
!= TEX_clgrp
)
5660 linelen
= p
- lb
.buffer
+ 1;
5662 make_tag (cp
, namelen
, true,
5663 lb
.buffer
, linelen
, lineno
, linecharno
);
5664 goto tex_next_line
; /* We only tag a line once */
5672 /* Read environment and prepend it to the default string.
5673 Build token table. */
5675 TEX_decode_env (const char *evarname
, const char *defenv
)
5677 register const char *env
, *p
;
5680 /* Append default string to environment. */
5681 env
= getenv (evarname
);
5685 env
= concat (env
, defenv
, "");
5687 /* Allocate a token table */
5688 for (len
= 1, p
= env
; (p
= strchr (p
, ':')); )
5691 TEX_toktab
= xnew (len
, linebuffer
);
5693 /* Unpack environment string into token table. Be careful about */
5694 /* zero-length strings (leading ':', "::" and trailing ':') */
5695 for (i
= 0; *env
!= '\0';)
5697 p
= strchr (env
, ':');
5698 if (!p
) /* End of environment string. */
5699 p
= env
+ strlen (env
);
5701 { /* Only non-zero strings. */
5702 TEX_toktab
[i
].buffer
= savenstr (env
, p
- env
);
5703 TEX_toktab
[i
].len
= p
- env
;
5710 TEX_toktab
[i
].buffer
= NULL
; /* Mark end of table. */
5711 TEX_toktab
[i
].len
= 0;
5718 /* Texinfo support. Dave Love, Mar. 2000. */
5720 Texinfo_nodes (FILE *inf
)
5723 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5724 if (LOOKING_AT (cp
, "@node"))
5727 while (*cp
!= '\0' && *cp
!= ',')
5729 make_tag (start
, cp
- start
, true,
5730 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
5737 * Contents of <title>, <h1>, <h2>, <h3> are tags.
5738 * Contents of <a name=xxx> are tags with name xxx.
5740 * Francesco Potortì, 2002.
5743 HTML_labels (FILE *inf
)
5745 bool getnext
= false; /* next text outside of HTML tags is a tag */
5746 bool skiptag
= false; /* skip to the end of the current HTML tag */
5747 bool intag
= false; /* inside an html tag, looking for ID= */
5748 bool inanchor
= false; /* when INTAG, is an anchor, look for NAME= */
5752 linebuffer_setlen (&token_name
, 0); /* no name in buffer */
5754 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
5755 for (;;) /* loop on the same line */
5757 if (skiptag
) /* skip HTML tag */
5759 while (*dbp
!= '\0' && *dbp
!= '>')
5765 continue; /* look on the same line */
5767 break; /* go to next line */
5770 else if (intag
) /* look for "name=" or "id=" */
5772 while (*dbp
!= '\0' && *dbp
!= '>'
5773 && c_tolower (*dbp
) != 'n' && c_tolower (*dbp
) != 'i')
5776 break; /* go to next line */
5781 continue; /* look on the same line */
5783 if ((inanchor
&& LOOKING_AT_NOCASE (dbp
, "name="))
5784 || LOOKING_AT_NOCASE (dbp
, "id="))
5786 bool quoted
= (dbp
[0] == '"');
5789 for (end
= ++dbp
; *end
!= '\0' && *end
!= '"'; end
++)
5792 for (end
= dbp
; *end
!= '\0' && intoken (*end
); end
++)
5794 linebuffer_setlen (&token_name
, end
- dbp
);
5795 memcpy (token_name
.buffer
, dbp
, end
- dbp
);
5796 token_name
.buffer
[end
- dbp
] = '\0';
5799 intag
= false; /* we found what we looked for */
5800 skiptag
= true; /* skip to the end of the tag */
5801 getnext
= true; /* then grab the text */
5802 continue; /* look on the same line */
5807 else if (getnext
) /* grab next tokens and tag them */
5809 dbp
= skip_spaces (dbp
);
5811 break; /* go to next line */
5815 inanchor
= (c_tolower (dbp
[1]) == 'a' && !intoken (dbp
[2]));
5816 continue; /* look on the same line */
5819 for (end
= dbp
+ 1; *end
!= '\0' && *end
!= '<'; end
++)
5821 make_tag (token_name
.buffer
, token_name
.len
, true,
5822 dbp
, end
- dbp
, lineno
, linecharno
);
5823 linebuffer_setlen (&token_name
, 0); /* no name in buffer */
5825 break; /* go to next line */
5828 else /* look for an interesting HTML tag */
5830 while (*dbp
!= '\0' && *dbp
!= '<')
5833 break; /* go to next line */
5835 if (c_tolower (dbp
[1]) == 'a' && !intoken (dbp
[2]))
5838 continue; /* look on the same line */
5840 else if (LOOKING_AT_NOCASE (dbp
, "<title>")
5841 || LOOKING_AT_NOCASE (dbp
, "<h1>")
5842 || LOOKING_AT_NOCASE (dbp
, "<h2>")
5843 || LOOKING_AT_NOCASE (dbp
, "<h3>"))
5847 continue; /* look on the same line */
5858 * Assumes that the predicate or rule starts at column 0.
5859 * Only the first clause of a predicate or rule is added.
5860 * Original code by Sunichirou Sugou (1989)
5861 * Rewritten by Anders Lindgren (1996)
5863 static size_t prolog_pr (char *, char *);
5864 static void prolog_skip_comment (linebuffer
*, FILE *);
5865 static size_t prolog_atom (char *, size_t);
5868 Prolog_functions (FILE *inf
)
5878 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5880 if (cp
[0] == '\0') /* Empty line */
5882 else if (c_isspace (cp
[0])) /* Not a predicate */
5884 else if (cp
[0] == '/' && cp
[1] == '*') /* comment. */
5885 prolog_skip_comment (&lb
, inf
);
5886 else if ((len
= prolog_pr (cp
, last
)) > 0)
5888 /* Predicate or rule. Store the function name so that we
5889 only generate a tag for the first clause. */
5891 last
= xnew (len
+ 1, char);
5892 else if (len
+ 1 > allocated
)
5893 xrnew (last
, len
+ 1, char);
5894 allocated
= len
+ 1;
5895 memcpy (last
, cp
, len
);
5904 prolog_skip_comment (linebuffer
*plb
, FILE *inf
)
5910 for (cp
= plb
->buffer
; *cp
!= '\0'; cp
++)
5911 if (cp
[0] == '*' && cp
[1] == '/')
5913 readline (plb
, inf
);
5915 while (perhaps_more_input (inf
));
5919 * A predicate or rule definition is added if it matches:
5920 * <beginning of line><Prolog Atom><whitespace>(
5921 * or <beginning of line><Prolog Atom><whitespace>:-
5923 * It is added to the tags database if it doesn't match the
5924 * name of the previous clause header.
5926 * Return the size of the name of the predicate or rule, or 0 if no
5930 prolog_pr (char *s
, char *last
)
5932 /* Name of last clause. */
5937 pos
= prolog_atom (s
, 0);
5942 pos
= skip_spaces (s
+ pos
) - s
;
5945 || (s
[pos
] == '(' && (pos
+= 1))
5946 || (s
[pos
] == ':' && s
[pos
+ 1] == '-' && (pos
+= 2)))
5947 && (last
== NULL
/* save only the first clause */
5948 || len
!= strlen (last
)
5949 || !strneq (s
, last
, len
)))
5951 make_tag (s
, len
, true, s
, pos
, lineno
, linecharno
);
5959 * Consume a Prolog atom.
5960 * Return the number of bytes consumed, or 0 if there was an error.
5962 * A prolog atom, in this context, could be one of:
5963 * - An alphanumeric sequence, starting with a lower case letter.
5964 * - A quoted arbitrary string. Single quotes can escape themselves.
5965 * Backslash quotes everything.
5968 prolog_atom (char *s
, size_t pos
)
5974 if (c_islower (s
[pos
]) || s
[pos
] == '_')
5976 /* The atom is unquoted. */
5978 while (c_isalnum (s
[pos
]) || s
[pos
] == '_')
5982 return pos
- origpos
;
5984 else if (s
[pos
] == '\'')
5995 pos
++; /* A double quote */
5997 else if (s
[pos
] == '\0')
5998 /* Multiline quoted atoms are ignored. */
6000 else if (s
[pos
] == '\\')
6002 if (s
[pos
+1] == '\0')
6009 return pos
- origpos
;
6017 * Support for Erlang
6019 * Generates tags for functions, defines, and records.
6020 * Assumes that Erlang functions start at column 0.
6021 * Original code by Anders Lindgren (1996)
6023 static int erlang_func (char *, char *);
6024 static void erlang_attribute (char *);
6025 static int erlang_atom (char *);
6028 Erlang_functions (FILE *inf
)
6038 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
6040 if (cp
[0] == '\0') /* Empty line */
6042 else if (c_isspace (cp
[0])) /* Not function nor attribute */
6044 else if (cp
[0] == '%') /* comment */
6046 else if (cp
[0] == '"') /* Sometimes, strings start in column one */
6048 else if (cp
[0] == '-') /* attribute, e.g. "-define" */
6050 erlang_attribute (cp
);
6057 else if ((len
= erlang_func (cp
, last
)) > 0)
6060 * Function. Store the function name so that we only
6061 * generates a tag for the first clause.
6064 last
= xnew (len
+ 1, char);
6065 else if (len
+ 1 > allocated
)
6066 xrnew (last
, len
+ 1, char);
6067 allocated
= len
+ 1;
6068 memcpy (last
, cp
, len
);
6077 * A function definition is added if it matches:
6078 * <beginning of line><Erlang Atom><whitespace>(
6080 * It is added to the tags database if it doesn't match the
6081 * name of the previous clause header.
6083 * Return the size of the name of the function, or 0 if no function
6087 erlang_func (char *s
, char *last
)
6089 /* Name of last clause. */
6094 pos
= erlang_atom (s
);
6099 pos
= skip_spaces (s
+ pos
) - s
;
6101 /* Save only the first clause. */
6104 || len
!= (int)strlen (last
)
6105 || !strneq (s
, last
, len
)))
6107 make_tag (s
, len
, true, s
, pos
, lineno
, linecharno
);
6116 * Handle attributes. Currently, tags are generated for defines
6119 * They are on the form:
6120 * -define(foo, bar).
6121 * -define(Foo(M, N), M+N).
6122 * -record(graph, {vtab = notable, cyclic = true}).
6125 erlang_attribute (char *s
)
6129 if ((LOOKING_AT (cp
, "-define") || LOOKING_AT (cp
, "-record"))
6132 int len
= erlang_atom (skip_spaces (cp
));
6134 make_tag (cp
, len
, true, s
, cp
+ len
- s
, lineno
, linecharno
);
6141 * Consume an Erlang atom (or variable).
6142 * Return the number of bytes consumed, or -1 if there was an error.
6145 erlang_atom (char *s
)
6149 if (c_isalpha (s
[pos
]) || s
[pos
] == '_')
6151 /* The atom is unquoted. */
6154 while (c_isalnum (s
[pos
]) || s
[pos
] == '_');
6156 else if (s
[pos
] == '\'')
6158 for (pos
++; s
[pos
] != '\''; pos
++)
6159 if (s
[pos
] == '\0' /* multiline quoted atoms are ignored */
6160 || (s
[pos
] == '\\' && s
[++pos
] == '\0'))
6169 static char *scan_separators (char *);
6170 static void add_regex (char *, language
*);
6171 static char *substitute (char *, char *, struct re_registers
*);
6174 * Take a string like "/blah/" and turn it into "blah", verifying
6175 * that the first and last characters are the same, and handling
6176 * quoted separator characters. Actually, stops on the occurrence of
6177 * an unquoted separator. Also process \t, \n, etc. and turn into
6178 * appropriate characters. Works in place. Null terminates name string.
6179 * Returns pointer to terminating separator, or NULL for
6180 * unterminated regexps.
6183 scan_separators (char *name
)
6186 char *copyto
= name
;
6187 bool quoted
= false;
6189 for (++name
; *name
!= '\0'; ++name
)
6195 case 'a': *copyto
++ = '\007'; break; /* BEL (bell) */
6196 case 'b': *copyto
++ = '\b'; break; /* BS (back space) */
6197 case 'd': *copyto
++ = 0177; break; /* DEL (delete) */
6198 case 'e': *copyto
++ = 033; break; /* ESC (delete) */
6199 case 'f': *copyto
++ = '\f'; break; /* FF (form feed) */
6200 case 'n': *copyto
++ = '\n'; break; /* NL (new line) */
6201 case 'r': *copyto
++ = '\r'; break; /* CR (carriage return) */
6202 case 't': *copyto
++ = '\t'; break; /* TAB (horizontal tab) */
6203 case 'v': *copyto
++ = '\v'; break; /* VT (vertical tab) */
6209 /* Something else is quoted, so preserve the quote. */
6217 else if (*name
== '\\')
6219 else if (*name
== sep
)
6225 name
= NULL
; /* signal unterminated regexp */
6227 /* Terminate copied string. */
6232 /* Look at the argument of --regex or --no-regex and do the right
6233 thing. Same for each line of a regexp file. */
6235 analyze_regex (char *regex_arg
)
6237 if (regex_arg
== NULL
)
6239 free_regexps (); /* --no-regex: remove existing regexps */
6243 /* A real --regexp option or a line in a regexp file. */
6244 switch (regex_arg
[0])
6246 /* Comments in regexp file or null arg to --regex. */
6252 /* Read a regex file. This is recursive and may result in a
6253 loop, which will stop when the file descriptors are exhausted. */
6257 linebuffer regexbuf
;
6258 char *regexfile
= regex_arg
+ 1;
6260 /* regexfile is a file containing regexps, one per line. */
6261 regexfp
= fopen (regexfile
, "r" FOPEN_BINARY
);
6262 if (regexfp
== NULL
)
6264 linebuffer_init (®exbuf
);
6265 while (readline_internal (®exbuf
, regexfp
, regexfile
) > 0)
6266 analyze_regex (regexbuf
.buffer
);
6267 free (regexbuf
.buffer
);
6268 if (fclose (regexfp
) != 0)
6273 /* Regexp to be used for a specific language only. */
6277 char *lang_name
= regex_arg
+ 1;
6280 for (cp
= lang_name
; *cp
!= '}'; cp
++)
6283 error ("unterminated language name in regex: %s", regex_arg
);
6287 lang
= get_language_from_langname (lang_name
);
6290 add_regex (cp
, lang
);
6294 /* Regexp to be used for any language. */
6296 add_regex (regex_arg
, NULL
);
6301 /* Separate the regexp pattern, compile it,
6302 and care for optional name and modifiers. */
6304 add_regex (char *regexp_pattern
, language
*lang
)
6306 static struct re_pattern_buffer zeropattern
;
6307 char sep
, *pat
, *name
, *modifiers
;
6310 struct re_pattern_buffer
*patbuf
;
6313 force_explicit_name
= true, /* do not use implicit tag names */
6314 ignore_case
= false, /* case is significant */
6315 multi_line
= false, /* matches are done one line at a time */
6316 single_line
= false; /* dot does not match newline */
6319 if (strlen (regexp_pattern
) < 3)
6321 error ("null regexp");
6324 sep
= regexp_pattern
[0];
6325 name
= scan_separators (regexp_pattern
);
6328 error ("%s: unterminated regexp", regexp_pattern
);
6333 error ("null name for regexp \"%s\"", regexp_pattern
);
6336 modifiers
= scan_separators (name
);
6337 if (modifiers
== NULL
) /* no terminating separator --> no name */
6343 modifiers
+= 1; /* skip separator */
6345 /* Parse regex modifiers. */
6346 for (; modifiers
[0] != '\0'; modifiers
++)
6347 switch (modifiers
[0])
6350 if (modifiers
== name
)
6351 error ("forcing explicit tag name but no name, ignoring");
6352 force_explicit_name
= true;
6362 need_filebuf
= true;
6365 error ("invalid regexp modifier '%c', ignoring", modifiers
[0]);
6369 patbuf
= xnew (1, struct re_pattern_buffer
);
6370 *patbuf
= zeropattern
;
6373 static char lc_trans
[UCHAR_MAX
+ 1];
6375 for (i
= 0; i
< UCHAR_MAX
+ 1; i
++)
6376 lc_trans
[i
] = c_tolower (i
);
6377 patbuf
->translate
= lc_trans
; /* translation table to fold case */
6381 pat
= concat ("^", regexp_pattern
, ""); /* anchor to beginning of line */
6383 pat
= regexp_pattern
;
6386 re_set_syntax (RE_SYNTAX_EMACS
| RE_DOT_NEWLINE
);
6388 re_set_syntax (RE_SYNTAX_EMACS
);
6390 err
= re_compile_pattern (pat
, strlen (pat
), patbuf
);
6395 error ("%s while compiling pattern", err
);
6400 p_head
= xnew (1, regexp
);
6401 p_head
->pattern
= savestr (regexp_pattern
);
6402 p_head
->p_next
= rp
;
6403 p_head
->lang
= lang
;
6404 p_head
->pat
= patbuf
;
6405 p_head
->name
= savestr (name
);
6406 p_head
->error_signaled
= false;
6407 p_head
->force_explicit_name
= force_explicit_name
;
6408 p_head
->ignore_case
= ignore_case
;
6409 p_head
->multi_line
= multi_line
;
6413 * Do the substitutions indicated by the regular expression and
6417 substitute (char *in
, char *out
, struct re_registers
*regs
)
6420 int size
, dig
, diglen
;
6423 size
= strlen (out
);
6425 /* Pass 1: figure out how much to allocate by finding all \N strings. */
6426 if (out
[size
- 1] == '\\')
6427 fatal ("pattern error in \"%s\"", out
);
6428 for (t
= strchr (out
, '\\');
6430 t
= strchr (t
+ 2, '\\'))
6431 if (c_isdigit (t
[1]))
6434 diglen
= regs
->end
[dig
] - regs
->start
[dig
];
6440 /* Allocate space and do the substitutions. */
6442 result
= xnew (size
+ 1, char);
6444 for (t
= result
; *out
!= '\0'; out
++)
6445 if (*out
== '\\' && c_isdigit (*++out
))
6448 diglen
= regs
->end
[dig
] - regs
->start
[dig
];
6449 memcpy (t
, in
+ regs
->start
[dig
], diglen
);
6456 assert (t
<= result
+ size
);
6457 assert (t
- result
== (int)strlen (result
));
6462 /* Deallocate all regexps. */
6467 while (p_head
!= NULL
)
6469 rp
= p_head
->p_next
;
6470 free (p_head
->pattern
);
6471 free (p_head
->name
);
6479 * Reads the whole file as a single string from `filebuf' and looks for
6480 * multi-line regular expressions, creating tags on matches.
6481 * readline already dealt with normal regexps.
6483 * Idea by Ben Wing <ben@666.com> (2002).
6486 regex_tag_multiline (void)
6488 char *buffer
= filebuf
.buffer
;
6492 for (rp
= p_head
; rp
!= NULL
; rp
= rp
->p_next
)
6496 if (!rp
->multi_line
)
6497 continue; /* skip normal regexps */
6499 /* Generic initializations before parsing file from memory. */
6500 lineno
= 1; /* reset global line number */
6501 charno
= 0; /* reset global char number */
6502 linecharno
= 0; /* reset global char number of line start */
6504 /* Only use generic regexps or those for the current language. */
6505 if (rp
->lang
!= NULL
&& rp
->lang
!= curfdp
->lang
)
6508 while (match
>= 0 && match
< filebuf
.len
)
6510 match
= re_search (rp
->pat
, buffer
, filebuf
.len
, charno
,
6511 filebuf
.len
- match
, &rp
->regs
);
6516 if (!rp
->error_signaled
)
6518 error ("regexp stack overflow while matching \"%s\"",
6520 rp
->error_signaled
= true;
6527 if (match
== rp
->regs
.end
[0])
6529 if (!rp
->error_signaled
)
6531 error ("regexp matches the empty string: \"%s\"",
6533 rp
->error_signaled
= true;
6535 match
= -3; /* exit from while loop */
6539 /* Match occurred. Construct a tag. */
6540 while (charno
< rp
->regs
.end
[0])
6541 if (buffer
[charno
++] == '\n')
6542 lineno
++, linecharno
= charno
;
6544 if (name
[0] == '\0')
6546 else /* make a named tag */
6547 name
= substitute (buffer
, rp
->name
, &rp
->regs
);
6548 if (rp
->force_explicit_name
)
6549 /* Force explicit tag name, if a name is there. */
6550 pfnote (name
, true, buffer
+ linecharno
,
6551 charno
- linecharno
+ 1, lineno
, linecharno
);
6553 make_tag (name
, strlen (name
), true, buffer
+ linecharno
,
6554 charno
- linecharno
+ 1, lineno
, linecharno
);
6563 nocase_tail (const char *cp
)
6567 while (*cp
!= '\0' && c_tolower (*cp
) == c_tolower (dbp
[len
]))
6569 if (*cp
== '\0' && !intoken (dbp
[len
]))
6578 get_tag (register char *bp
, char **namepp
)
6580 register char *cp
= bp
;
6584 /* Go till you get to white space or a syntactic break */
6585 for (cp
= bp
+ 1; !notinname (*cp
); cp
++)
6587 make_tag (bp
, cp
- bp
, true,
6588 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
6592 *namepp
= savenstr (bp
, cp
- bp
);
6595 /* Similar to get_tag, but include '=' as part of the tag. */
6597 get_lispy_tag (register char *bp
)
6599 register char *cp
= bp
;
6603 /* Go till you get to white space or a syntactic break */
6604 for (cp
= bp
+ 1; !notinname (*cp
) || *cp
== '='; cp
++)
6606 make_tag (bp
, cp
- bp
, true,
6607 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
6612 * Read a line of text from `stream' into `lbp', excluding the
6613 * newline or CR-NL, if any. Return the number of characters read from
6614 * `stream', which is the length of the line including the newline.
6616 * On DOS or Windows we do not count the CR character, if any before the
6617 * NL, in the returned length; this mirrors the behavior of Emacs on those
6618 * platforms (for text files, it translates CR-NL to NL as it reads in the
6621 * If multi-line regular expressions are requested, each line read is
6622 * appended to `filebuf'.
6625 readline_internal (linebuffer
*lbp
, FILE *stream
, char const *filename
)
6627 char *buffer
= lbp
->buffer
;
6628 char *p
= lbp
->buffer
;
6632 pend
= p
+ lbp
->size
; /* Separate to avoid 386/IX compiler bug. */
6636 register int c
= getc (stream
);
6639 /* We're at the end of linebuffer: expand it. */
6641 xrnew (buffer
, lbp
->size
, char);
6642 p
+= buffer
- lbp
->buffer
;
6643 pend
= buffer
+ lbp
->size
;
6644 lbp
->buffer
= buffer
;
6648 if (ferror (stream
))
6656 if (p
> buffer
&& p
[-1] == '\r')
6670 lbp
->len
= p
- buffer
;
6672 if (need_filebuf
/* we need filebuf for multi-line regexps */
6673 && chars_deleted
> 0) /* not at EOF */
6675 while (filebuf
.size
<= filebuf
.len
+ lbp
->len
+ 1) /* +1 for \n */
6677 /* Expand filebuf. */
6679 xrnew (filebuf
.buffer
, filebuf
.size
, char);
6681 memcpy (filebuf
.buffer
+ filebuf
.len
, lbp
->buffer
, lbp
->len
);
6682 filebuf
.len
+= lbp
->len
;
6683 filebuf
.buffer
[filebuf
.len
++] = '\n';
6684 filebuf
.buffer
[filebuf
.len
] = '\0';
6687 return lbp
->len
+ chars_deleted
;
6691 * Like readline_internal, above, but in addition try to match the
6692 * input line against relevant regular expressions and manage #line
6696 readline (linebuffer
*lbp
, FILE *stream
)
6700 linecharno
= charno
; /* update global char number of line start */
6701 result
= readline_internal (lbp
, stream
, infilename
); /* read line */
6702 lineno
+= 1; /* increment global line number */
6703 charno
+= result
; /* increment global char number */
6705 /* Honor #line directives. */
6706 if (!no_line_directive
)
6708 static bool discard_until_line_directive
;
6710 /* Check whether this is a #line directive. */
6711 if (result
> 12 && strneq (lbp
->buffer
, "#line ", 6))
6716 if (sscanf (lbp
->buffer
, "#line %u \"%n", &lno
, &start
) >= 1
6717 && start
> 0) /* double quote character found */
6719 char *endp
= lbp
->buffer
+ start
;
6721 while ((endp
= strchr (endp
, '"')) != NULL
6722 && endp
[-1] == '\\')
6725 /* Ok, this is a real #line directive. Let's deal with it. */
6727 char *taggedabsname
; /* absolute name of original file */
6728 char *taggedfname
; /* name of original file as given */
6729 char *name
; /* temp var */
6731 discard_until_line_directive
= false; /* found it */
6732 name
= lbp
->buffer
+ start
;
6734 canonicalize_filename (name
);
6735 taggedabsname
= absolute_filename (name
, tagfiledir
);
6736 if (filename_is_absolute (name
)
6737 || filename_is_absolute (curfdp
->infname
))
6738 taggedfname
= savestr (taggedabsname
);
6740 taggedfname
= relative_filename (taggedabsname
,tagfiledir
);
6742 if (streq (curfdp
->taggedfname
, taggedfname
))
6743 /* The #line directive is only a line number change. We
6744 deal with this afterwards. */
6747 /* The tags following this #line directive should be
6748 attributed to taggedfname. In order to do this, set
6749 curfdp accordingly. */
6751 fdesc
*fdp
; /* file description pointer */
6753 /* Go look for a file description already set up for the
6754 file indicated in the #line directive. If there is
6755 one, use it from now until the next #line
6757 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
6758 if (streq (fdp
->infname
, curfdp
->infname
)
6759 && streq (fdp
->taggedfname
, taggedfname
))
6760 /* If we remove the second test above (after the &&)
6761 then all entries pertaining to the same file are
6762 coalesced in the tags file. If we use it, then
6763 entries pertaining to the same file but generated
6764 from different files (via #line directives) will
6765 go into separate sections in the tags file. These
6766 alternatives look equivalent. The first one
6767 destroys some apparently useless information. */
6773 /* Else, if we already tagged the real file, skip all
6774 input lines until the next #line directive. */
6775 if (fdp
== NULL
) /* not found */
6776 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
6777 if (streq (fdp
->infabsname
, taggedabsname
))
6779 discard_until_line_directive
= true;
6783 /* Else create a new file description and use that from
6784 now on, until the next #line directive. */
6785 if (fdp
== NULL
) /* not found */
6788 fdhead
= xnew (1, fdesc
);
6789 *fdhead
= *curfdp
; /* copy curr. file description */
6791 fdhead
->infname
= savestr (curfdp
->infname
);
6792 fdhead
->infabsname
= savestr (curfdp
->infabsname
);
6793 fdhead
->infabsdir
= savestr (curfdp
->infabsdir
);
6794 fdhead
->taggedfname
= taggedfname
;
6795 fdhead
->usecharno
= false;
6796 fdhead
->prop
= NULL
;
6797 fdhead
->written
= false;
6801 free (taggedabsname
);
6803 readline (lbp
, stream
);
6805 } /* if a real #line directive */
6806 } /* if #line is followed by a number */
6807 } /* if line begins with "#line " */
6809 /* If we are here, no #line directive was found. */
6810 if (discard_until_line_directive
)
6814 /* Do a tail recursion on ourselves, thus discarding the contents
6815 of the line buffer. */
6816 readline (lbp
, stream
);
6820 discard_until_line_directive
= false;
6823 } /* if #line directives should be considered */
6830 /* Match against relevant regexps. */
6832 for (rp
= p_head
; rp
!= NULL
; rp
= rp
->p_next
)
6834 /* Only use generic regexps or those for the current language.
6835 Also do not use multiline regexps, which is the job of
6836 regex_tag_multiline. */
6837 if ((rp
->lang
!= NULL
&& rp
->lang
!= fdhead
->lang
)
6841 match
= re_match (rp
->pat
, lbp
->buffer
, lbp
->len
, 0, &rp
->regs
);
6846 if (!rp
->error_signaled
)
6848 error ("regexp stack overflow while matching \"%s\"",
6850 rp
->error_signaled
= true;
6857 /* Empty string matched. */
6858 if (!rp
->error_signaled
)
6860 error ("regexp matches the empty string: \"%s\"", rp
->pattern
);
6861 rp
->error_signaled
= true;
6865 /* Match occurred. Construct a tag. */
6867 if (name
[0] == '\0')
6869 else /* make a named tag */
6870 name
= substitute (lbp
->buffer
, rp
->name
, &rp
->regs
);
6871 if (rp
->force_explicit_name
)
6872 /* Force explicit tag name, if a name is there. */
6873 pfnote (name
, true, lbp
->buffer
, match
, lineno
, linecharno
);
6875 make_tag (name
, strlen (name
), true,
6876 lbp
->buffer
, match
, lineno
, linecharno
);
6885 * Return a pointer to a space of size strlen(cp)+1 allocated
6886 * with xnew where the string CP has been copied.
6889 savestr (const char *cp
)
6891 return savenstr (cp
, strlen (cp
));
6895 * Return a pointer to a space of size LEN+1 allocated with xnew where
6896 * the string CP has been copied for at most the first LEN characters.
6899 savenstr (const char *cp
, int len
)
6901 char *dp
= xnew (len
+ 1, char);
6903 return memcpy (dp
, cp
, len
);
6906 /* Skip spaces (end of string is not space), return new pointer. */
6908 skip_spaces (char *cp
)
6910 while (c_isspace (*cp
))
6915 /* Skip non spaces, except end of string, return new pointer. */
6917 skip_non_spaces (char *cp
)
6919 while (*cp
!= '\0' && !c_isspace (*cp
))
6924 /* Skip any chars in the "name" class.*/
6926 skip_name (char *cp
)
6928 /* '\0' is a notinname() so loop stops there too */
6929 while (! notinname (*cp
))
6934 /* Print error message and exit. */
6936 fatal (char const *format
, ...)
6939 va_start (ap
, format
);
6940 verror (format
, ap
);
6942 exit (EXIT_FAILURE
);
6946 pfatal (const char *s1
)
6949 exit (EXIT_FAILURE
);
6953 suggest_asking_for_help (void)
6955 fprintf (stderr
, "\tTry '%s --help' for a complete list of options.\n",
6957 exit (EXIT_FAILURE
);
6960 /* Output a diagnostic with printf-style FORMAT and args. */
6962 error (const char *format
, ...)
6965 va_start (ap
, format
);
6966 verror (format
, ap
);
6971 verror (char const *format
, va_list ap
)
6973 fprintf (stderr
, "%s: ", progname
);
6974 vfprintf (stderr
, format
, ap
);
6975 fprintf (stderr
, "\n");
6978 /* Return a newly-allocated string whose contents
6979 concatenate those of s1, s2, s3. */
6981 concat (const char *s1
, const char *s2
, const char *s3
)
6983 int len1
= strlen (s1
), len2
= strlen (s2
), len3
= strlen (s3
);
6984 char *result
= xnew (len1
+ len2
+ len3
+ 1, char);
6986 strcpy (result
, s1
);
6987 strcpy (result
+ len1
, s2
);
6988 strcpy (result
+ len1
+ len2
, s3
);
6994 /* Does the same work as the system V getcwd, but does not need to
6995 guess the buffer size in advance. */
7000 char *path
= xnew (bufsize
, char);
7002 while (getcwd (path
, bufsize
) == NULL
)
7004 if (errno
!= ERANGE
)
7008 path
= xnew (bufsize
, char);
7011 canonicalize_filename (path
);
7015 /* Return a newly allocated string containing a name of a temporary file. */
7019 const char *tmpdir
= getenv ("TMPDIR");
7020 const char *slash
= "/";
7022 #if MSDOS || defined (DOS_NT)
7024 tmpdir
= getenv ("TEMP");
7026 tmpdir
= getenv ("TMP");
7029 if (tmpdir
[strlen (tmpdir
) - 1] == '/'
7030 || tmpdir
[strlen (tmpdir
) - 1] == '\\')
7035 if (tmpdir
[strlen (tmpdir
) - 1] == '/')
7039 char *templt
= concat (tmpdir
, slash
, "etXXXXXX");
7040 int fd
= mkostemp (templt
, O_CLOEXEC
);
7041 if (fd
< 0 || close (fd
) != 0)
7043 int temp_errno
= errno
;
7049 #if defined (DOS_NT)
7050 /* The file name will be used in shell redirection, so it needs to have
7051 DOS-style backslashes, or else the Windows shell will barf. */
7053 for (p
= templt
; *p
; p
++)
7061 /* Return a newly allocated string containing the file name of FILE
7062 relative to the absolute directory DIR (which should end with a slash). */
7064 relative_filename (char *file
, char *dir
)
7066 char *fp
, *dp
, *afn
, *res
;
7069 /* Find the common root of file and dir (with a trailing slash). */
7070 afn
= absolute_filename (file
, cwd
);
7073 while (*fp
++ == *dp
++)
7075 fp
--, dp
--; /* back to the first differing char */
7077 if (fp
== afn
&& afn
[0] != '/') /* cannot build a relative name */
7080 do /* look at the equal chars until '/' */
7084 /* Build a sequence of "../" strings for the resulting relative file name. */
7086 while ((dp
= strchr (dp
+ 1, '/')) != NULL
)
7088 res
= xnew (3*i
+ strlen (fp
+ 1) + 1, char);
7091 z
= stpcpy (z
, "../");
7093 /* Add the file name relative to the common root of file and dir. */
7100 /* Return a newly allocated string containing the absolute file name
7101 of FILE given DIR (which should end with a slash). */
7103 absolute_filename (char *file
, char *dir
)
7105 char *slashp
, *cp
, *res
;
7107 if (filename_is_absolute (file
))
7108 res
= savestr (file
);
7110 /* We don't support non-absolute file names with a drive
7111 letter, like `d:NAME' (it's too much hassle). */
7112 else if (file
[1] == ':')
7113 fatal ("%s: relative file names with drive letters not supported", file
);
7116 res
= concat (dir
, file
, "");
7118 /* Delete the "/dirname/.." and "/." substrings. */
7119 slashp
= strchr (res
, '/');
7120 while (slashp
!= NULL
&& slashp
[0] != '\0')
7122 if (slashp
[1] == '.')
7124 if (slashp
[2] == '.'
7125 && (slashp
[3] == '/' || slashp
[3] == '\0'))
7130 while (cp
>= res
&& !filename_is_absolute (cp
));
7132 cp
= slashp
; /* the absolute name begins with "/.." */
7134 /* Under MSDOS and NT we get `d:/NAME' as absolute
7135 file name, so the luser could say `d:/../NAME'.
7136 We silently treat this as `d:/NAME'. */
7137 else if (cp
[0] != '/')
7140 memmove (cp
, slashp
+ 3, strlen (slashp
+ 2));
7144 else if (slashp
[2] == '/' || slashp
[2] == '\0')
7146 memmove (slashp
, slashp
+ 2, strlen (slashp
+ 1));
7151 slashp
= strchr (slashp
+ 1, '/');
7154 if (res
[0] == '\0') /* just a safety net: should never happen */
7157 return savestr ("/");
7163 /* Return a newly allocated string containing the absolute
7164 file name of dir where FILE resides given DIR (which should
7165 end with a slash). */
7167 absolute_dirname (char *file
, char *dir
)
7172 slashp
= strrchr (file
, '/');
7174 return savestr (dir
);
7177 res
= absolute_filename (file
, dir
);
7183 /* Whether the argument string is an absolute file name. The argument
7184 string must have been canonicalized with canonicalize_filename. */
7186 filename_is_absolute (char *fn
)
7188 return (fn
[0] == '/'
7190 || (c_isalpha (fn
[0]) && fn
[1] == ':' && fn
[2] == '/')
7195 /* Downcase DOS drive letter and collapse separators into single slashes.
7198 canonicalize_filename (register char *fn
)
7203 /* Canonicalize drive letter case. */
7204 if (c_isupper (fn
[0]) && fn
[1] == ':')
7205 fn
[0] = c_tolower (fn
[0]);
7207 /* Collapse multiple forward- and back-slashes into a single forward
7209 for (cp
= fn
; *cp
!= '\0'; cp
++, fn
++)
7210 if (*cp
== '/' || *cp
== '\\')
7213 while (cp
[1] == '/' || cp
[1] == '\\')
7221 /* Collapse multiple slashes into a single slash. */
7222 for (cp
= fn
; *cp
!= '\0'; cp
++, fn
++)
7226 while (cp
[1] == '/')
7232 #endif /* !DOS_NT */
7238 /* Initialize a linebuffer for use. */
7240 linebuffer_init (linebuffer
*lbp
)
7242 lbp
->size
= (DEBUG
) ? 3 : 200;
7243 lbp
->buffer
= xnew (lbp
->size
, char);
7244 lbp
->buffer
[0] = '\0';
7248 /* Set the minimum size of a string contained in a linebuffer. */
7250 linebuffer_setlen (linebuffer
*lbp
, int toksize
)
7252 while (lbp
->size
<= toksize
)
7255 xrnew (lbp
->buffer
, lbp
->size
, char);
7260 /* Like malloc but get fatal error if memory is exhausted. */
7262 xmalloc (size_t size
)
7264 void *result
= malloc (size
);
7266 fatal ("virtual memory exhausted");
7271 xrealloc (void *ptr
, size_t size
)
7273 void *result
= realloc (ptr
, size
);
7275 fatal ("virtual memory exhausted");
7281 * indent-tabs-mode: t
7284 * c-font-lock-extra-types: ("FILE" "bool" "language" "linebuffer" "fdesc" "node" "regexp")
7285 * c-file-style: "gnu"
7289 /* etags.c ends here */