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, 1998-2014 Free Software
34 This file is not considered part of GNU Emacs.
36 This program is free software: you can redistribute it and/or modify
37 it under the terms of the GNU General Public License as published by
38 the Free Software Foundation, either version 3 of the License, or
39 (at your option) any later version.
41 This program is distributed in the hope that it will be useful,
42 but WITHOUT ANY WARRANTY; without even the implied warranty of
43 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
44 GNU General Public License for more details.
46 You should have received a copy of the GNU General Public License
47 along with this program. If not, see <http://www.gnu.org/licenses/>. */
50 /* NB To comply with the above BSD license, copyright information is
51 reproduced in etc/ETAGS.README. That file should be updated when the
54 To the best of our knowledge, this code was originally based on the
55 ctags.c distributed with BSD4.2, which was copyrighted by the
56 University of California, as described above. */
61 * 1983 Ctags originally by Ken Arnold.
62 * 1984 Fortran added by Jim Kleckner.
63 * 1984 Ed Pelegri-Llopart added C typedefs.
64 * 1985 Emacs TAGS format by Richard Stallman.
65 * 1989 Sam Kendall added C++.
66 * 1992 Joseph B. Wells improved C and C++ parsing.
67 * 1993 Francesco Potortì reorganized C and C++.
68 * 1994 Line-by-line regexp tags by Tom Tromey.
69 * 2001 Nested classes by Francesco Potortì (concept by Mykola Dzyuba).
70 * 2002 #line directives by Francesco Potortì.
72 * Francesco Potortì <pot@gnu.org> has maintained and improved it since 1993.
76 * If you want to add support for a new language, start by looking at the LUA
77 * language, which is the simplest. Alternatively, consider distributing etags
78 * together with a configuration file containing regexp definitions for etags.
81 char pot_etags_version
[] = "@(#) pot revision number is 17.38.1.4";
88 # define NDEBUG /* disable assert */
94 # define _GNU_SOURCE 1 /* enables some compiler checks on GNU */
97 /* WIN32_NATIVE is for XEmacs.
98 MSDOS, WINDOWSNT, DOS_NT are for Emacs. */
103 #endif /* WIN32_NATIVE */
108 # include <sys/param.h>
115 # define MAXPATHLEN _MAX_PATH
119 #endif /* WINDOWSNT */
128 #include <sys/types.h>
129 #include <sys/stat.h>
130 #include <binary-io.h>
131 #include <c-strcase.h>
135 # undef assert /* some systems have a buggy assert.h */
136 # define assert(x) ((void) 0)
142 /* Define CTAGS to make the program "ctags" compatible with the usual one.
143 Leave it undefined to make the program "etags", which makes emacs-style
144 tag tables and tags typedefs, #defines and struct/union/enum by default. */
152 #define streq(s,t) (assert ((s)!=NULL || (t)!=NULL), !strcmp (s, t))
153 #define strcaseeq(s,t) (assert ((s)!=NULL && (t)!=NULL), !c_strcasecmp (s, t))
154 #define strneq(s,t,n) (assert ((s)!=NULL || (t)!=NULL), !strncmp (s, t, n))
155 #define strncaseeq(s,t,n) (assert ((s)!=NULL && (t)!=NULL), !c_strncasecmp (s, t, n))
157 #define CHARS 256 /* 2^sizeof(char) */
158 #define CHAR(x) ((unsigned int)(x) & (CHARS - 1))
159 #define iswhite(c) (_wht[CHAR (c)]) /* c is white (see white) */
160 #define notinname(c) (_nin[CHAR (c)]) /* c is not in a name (see nonam) */
161 #define begtoken(c) (_btk[CHAR (c)]) /* c can start token (see begtk) */
162 #define intoken(c) (_itk[CHAR (c)]) /* c can be in token (see midtk) */
163 #define endtoken(c) (_etk[CHAR (c)]) /* c ends tokens (see endtk) */
165 #define ISALNUM(c) isalnum (CHAR (c))
166 #define ISALPHA(c) isalpha (CHAR (c))
167 #define ISDIGIT(c) isdigit (CHAR (c))
168 #define ISLOWER(c) islower (CHAR (c))
170 #define lowcase(c) tolower (CHAR (c))
174 * xnew, xrnew -- allocate, reallocate storage
176 * SYNOPSIS: Type *xnew (int n, Type);
177 * void xrnew (OldPointer, int n, Type);
180 # include "chkmalloc.h"
181 # define xnew(n,Type) ((Type *) trace_malloc (__FILE__, __LINE__, \
182 (n) * sizeof (Type)))
183 # define xrnew(op,n,Type) ((op) = (Type *) trace_realloc (__FILE__, __LINE__, \
184 (char *) (op), (n) * sizeof (Type)))
186 # define xnew(n,Type) ((Type *) xmalloc ((n) * sizeof (Type)))
187 # define xrnew(op,n,Type) ((op) = (Type *) xrealloc ( \
188 (char *) (op), (n) * sizeof (Type)))
191 typedef void Lang_function (FILE *);
195 const char *suffix
; /* file name suffix for this compressor */
196 const char *command
; /* takes one arg and decompresses to stdout */
201 const char *name
; /* language name */
202 const char *help
; /* detailed help for the language */
203 Lang_function
*function
; /* parse function */
204 const char **suffixes
; /* name suffixes of this language's files */
205 const char **filenames
; /* names of this language's files */
206 const char **interpreters
; /* interpreters for this language */
207 bool metasource
; /* source used to generate other sources */
212 struct fdesc
*next
; /* for the linked list */
213 char *infname
; /* uncompressed input file name */
214 char *infabsname
; /* absolute uncompressed input file name */
215 char *infabsdir
; /* absolute dir of input file */
216 char *taggedfname
; /* file name to write in tagfile */
217 language
*lang
; /* language of file */
218 char *prop
; /* file properties to write in tagfile */
219 bool usecharno
; /* etags tags shall contain char number */
220 bool written
; /* entry written in the tags file */
223 typedef struct node_st
224 { /* sorting structure */
225 struct node_st
*left
, *right
; /* left and right sons */
226 fdesc
*fdp
; /* description of file to whom tag belongs */
227 char *name
; /* tag name */
228 char *regex
; /* search regexp */
229 bool valid
; /* write this tag on the tag file */
230 bool is_func
; /* function tag: use regexp in CTAGS mode */
231 bool been_warned
; /* warning already given for duplicated tag */
232 int lno
; /* line number tag is on */
233 long cno
; /* character number line starts on */
237 * A `linebuffer' is a structure which holds a line of text.
238 * `readline_internal' reads a line from a stream into a linebuffer
239 * and works regardless of the length of the line.
240 * SIZE is the size of BUFFER, LEN is the length of the string in
241 * BUFFER after readline reads it.
250 /* Used to support mixing of --lang and file names. */
254 at_language
, /* a language specification */
255 at_regexp
, /* a regular expression */
256 at_filename
, /* a file name */
257 at_stdin
, /* read from stdin here */
258 at_end
/* stop parsing the list */
259 } arg_type
; /* argument type */
260 language
*lang
; /* language associated with the argument */
261 char *what
; /* the argument itself */
264 /* Structure defining a regular expression. */
265 typedef struct regexp
267 struct regexp
*p_next
; /* pointer to next in list */
268 language
*lang
; /* if set, use only for this language */
269 char *pattern
; /* the regexp pattern */
270 char *name
; /* tag name */
271 struct re_pattern_buffer
*pat
; /* the compiled pattern */
272 struct re_registers regs
; /* re registers */
273 bool error_signaled
; /* already signaled for this regexp */
274 bool force_explicit_name
; /* do not allow implicit tag name */
275 bool ignore_case
; /* ignore case when matching */
276 bool multi_line
; /* do a multi-line match on the whole file */
280 /* Many compilers barf on this:
281 Lang_function Ada_funcs;
282 so let's write it this way */
283 static void Ada_funcs (FILE *);
284 static void Asm_labels (FILE *);
285 static void C_entries (int c_ext
, FILE *);
286 static void default_C_entries (FILE *);
287 static void plain_C_entries (FILE *);
288 static void Cjava_entries (FILE *);
289 static void Cobol_paragraphs (FILE *);
290 static void Cplusplus_entries (FILE *);
291 static void Cstar_entries (FILE *);
292 static void Erlang_functions (FILE *);
293 static void Forth_words (FILE *);
294 static void Fortran_functions (FILE *);
295 static void HTML_labels (FILE *);
296 static void Lisp_functions (FILE *);
297 static void Lua_functions (FILE *);
298 static void Makefile_targets (FILE *);
299 static void Pascal_functions (FILE *);
300 static void Perl_functions (FILE *);
301 static void PHP_functions (FILE *);
302 static void PS_functions (FILE *);
303 static void Prolog_functions (FILE *);
304 static void Python_functions (FILE *);
305 static void Scheme_functions (FILE *);
306 static void TeX_commands (FILE *);
307 static void Texinfo_nodes (FILE *);
308 static void Yacc_entries (FILE *);
309 static void just_read_file (FILE *);
311 static language
*get_language_from_langname (const char *);
312 static void readline (linebuffer
*, FILE *);
313 static long readline_internal (linebuffer
*, FILE *);
314 static bool nocase_tail (const char *);
315 static void get_tag (char *, char **);
317 static void analyse_regex (char *);
318 static void free_regexps (void);
319 static void regex_tag_multiline (void);
320 static void error (const char *, ...) ATTRIBUTE_FORMAT_PRINTF (1, 2);
321 static _Noreturn
void suggest_asking_for_help (void);
322 _Noreturn
void fatal (const char *, const char *);
323 static _Noreturn
void pfatal (const char *);
324 static void add_node (node
*, node
**);
326 static void init (void);
327 static void process_file_name (char *, language
*);
328 static void process_file (FILE *, char *, language
*);
329 static void find_entries (FILE *);
330 static void free_tree (node
*);
331 static void free_fdesc (fdesc
*);
332 static void pfnote (char *, bool, char *, int, int, long);
333 static void invalidate_nodes (fdesc
*, node
**);
334 static void put_entries (node
*);
336 static char *concat (const char *, const char *, const char *);
337 static char *skip_spaces (char *);
338 static char *skip_non_spaces (char *);
339 static char *skip_name (char *);
340 static char *savenstr (const char *, int);
341 static char *savestr (const char *);
342 static char *etags_strchr (const char *, int);
343 static char *etags_strrchr (const char *, int);
344 static char *etags_getcwd (void);
345 static char *relative_filename (char *, char *);
346 static char *absolute_filename (char *, char *);
347 static char *absolute_dirname (char *, char *);
348 static bool filename_is_absolute (char *f
);
349 static void canonicalize_filename (char *);
350 static void linebuffer_init (linebuffer
*);
351 static void linebuffer_setlen (linebuffer
*, int);
352 static void *xmalloc (size_t);
353 static void *xrealloc (char *, size_t);
356 static char searchar
= '/'; /* use /.../ searches */
358 static char *tagfile
; /* output file */
359 static char *progname
; /* name this program was invoked with */
360 static char *cwd
; /* current working directory */
361 static char *tagfiledir
; /* directory of tagfile */
362 static FILE *tagf
; /* ioptr for tags file */
363 static ptrdiff_t whatlen_max
; /* maximum length of any 'what' member */
365 static fdesc
*fdhead
; /* head of file description list */
366 static fdesc
*curfdp
; /* current file description */
367 static int lineno
; /* line number of current line */
368 static long charno
; /* current character number */
369 static long linecharno
; /* charno of start of current line */
370 static char *dbp
; /* pointer to start of current tag */
372 static const int invalidcharno
= -1;
374 static node
*nodehead
; /* the head of the binary tree of tags */
375 static node
*last_node
; /* the last node created */
377 static linebuffer lb
; /* the current line */
378 static linebuffer filebuf
; /* a buffer containing the whole file */
379 static linebuffer token_name
; /* a buffer containing a tag name */
381 /* boolean "functions" (see init) */
382 static bool _wht
[CHARS
], _nin
[CHARS
], _itk
[CHARS
], _btk
[CHARS
], _etk
[CHARS
];
385 *white
= " \f\t\n\r\v",
387 *nonam
= " \f\t\n\r()=,;", /* look at make_tag before modifying! */
388 /* token ending chars */
389 *endtk
= " \t\n\r\"'#()[]{}=-+%*/&|^~!<>;,.:?",
390 /* token starting chars */
391 *begtk
= "ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz$~@",
392 /* valid in-token chars */
393 *midtk
= "ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz$0123456789";
395 static bool append_to_tagfile
; /* -a: append to tags */
396 /* The next five default to true in C and derived languages. */
397 static bool typedefs
; /* -t: create tags for C and Ada typedefs */
398 static bool typedefs_or_cplusplus
; /* -T: create tags for C typedefs, level */
399 /* 0 struct/enum/union decls, and C++ */
400 /* member functions. */
401 static bool constantypedefs
; /* -d: create tags for C #define, enum */
402 /* constants and variables. */
403 /* -D: opposite of -d. Default under ctags. */
404 static int globals
; /* create tags for global variables */
405 static int members
; /* create tags for C member variables */
406 static int declarations
; /* --declarations: tag them and extern in C&Co*/
407 static int no_line_directive
; /* ignore #line directives (undocumented) */
408 static int no_duplicates
; /* no duplicate tags for ctags (undocumented) */
409 static bool update
; /* -u: update tags */
410 static bool vgrind_style
; /* -v: create vgrind style index output */
411 static bool no_warnings
; /* -w: suppress warnings (undocumented) */
412 static bool cxref_style
; /* -x: create cxref style output */
413 static bool cplusplus
; /* .[hc] means C++, not C (undocumented) */
414 static bool ignoreindent
; /* -I: ignore indentation in C */
415 static int packages_only
; /* --packages-only: in Ada, only tag packages*/
417 /* STDIN is defined in LynxOS system headers */
422 #define STDIN 0x1001 /* returned by getopt_long on --parse-stdin */
423 static bool parsing_stdin
; /* --parse-stdin used */
425 static regexp
*p_head
; /* list of all regexps */
426 static bool need_filebuf
; /* some regexes are multi-line */
428 static struct option longopts
[] =
430 { "append", no_argument
, NULL
, 'a' },
431 { "packages-only", no_argument
, &packages_only
, 1 },
432 { "c++", no_argument
, NULL
, 'C' },
433 { "declarations", no_argument
, &declarations
, 1 },
434 { "no-line-directive", no_argument
, &no_line_directive
, 1 },
435 { "no-duplicates", no_argument
, &no_duplicates
, 1 },
436 { "help", no_argument
, NULL
, 'h' },
437 { "help", no_argument
, NULL
, 'H' },
438 { "ignore-indentation", no_argument
, NULL
, 'I' },
439 { "language", required_argument
, NULL
, 'l' },
440 { "members", no_argument
, &members
, 1 },
441 { "no-members", no_argument
, &members
, 0 },
442 { "output", required_argument
, NULL
, 'o' },
443 { "regex", required_argument
, NULL
, 'r' },
444 { "no-regex", no_argument
, NULL
, 'R' },
445 { "ignore-case-regex", required_argument
, NULL
, 'c' },
446 { "parse-stdin", required_argument
, NULL
, STDIN
},
447 { "version", no_argument
, NULL
, 'V' },
449 #if CTAGS /* Ctags options */
450 { "backward-search", no_argument
, NULL
, 'B' },
451 { "cxref", no_argument
, NULL
, 'x' },
452 { "defines", no_argument
, NULL
, 'd' },
453 { "globals", no_argument
, &globals
, 1 },
454 { "typedefs", no_argument
, NULL
, 't' },
455 { "typedefs-and-c++", no_argument
, NULL
, 'T' },
456 { "update", no_argument
, NULL
, 'u' },
457 { "vgrind", no_argument
, NULL
, 'v' },
458 { "no-warn", no_argument
, NULL
, 'w' },
460 #else /* Etags options */
461 { "no-defines", no_argument
, NULL
, 'D' },
462 { "no-globals", no_argument
, &globals
, 0 },
463 { "include", required_argument
, NULL
, 'i' },
468 static compressor compressors
[] =
470 { "z", "gzip -d -c"},
471 { "Z", "gzip -d -c"},
472 { "gz", "gzip -d -c"},
473 { "GZ", "gzip -d -c"},
474 { "bz2", "bzip2 -d -c" },
475 { "xz", "xz -d -c" },
484 static const char *Ada_suffixes
[] =
485 { "ads", "adb", "ada", NULL
};
486 static const char Ada_help
[] =
487 "In Ada code, functions, procedures, packages, tasks and types are\n\
488 tags. Use the `--packages-only' option to create tags for\n\
490 Ada tag names have suffixes indicating the type of entity:\n\
491 Entity type: Qualifier:\n\
492 ------------ ----------\n\
499 Thus, `M-x find-tag <RET> bidule/b <RET>' will go directly to the\n\
500 body of the package `bidule', while `M-x find-tag <RET> bidule <RET>'\n\
501 will just search for any tag `bidule'.";
504 static const char *Asm_suffixes
[] =
505 { "a", /* Unix assembler */
506 "asm", /* Microcontroller assembly */
507 "def", /* BSO/Tasking definition includes */
508 "inc", /* Microcontroller include files */
509 "ins", /* Microcontroller include files */
510 "s", "sa", /* Unix assembler */
511 "S", /* cpp-processed Unix assembler */
512 "src", /* BSO/Tasking C compiler output */
515 static const char Asm_help
[] =
516 "In assembler code, labels appearing at the beginning of a line,\n\
517 followed by a colon, are tags.";
520 /* Note that .c and .h can be considered C++, if the --c++ flag was
521 given, or if the `class' or `template' keywords are met inside the file.
522 That is why default_C_entries is called for these. */
523 static const char *default_C_suffixes
[] =
525 #if CTAGS /* C help for Ctags */
526 static const char default_C_help
[] =
527 "In C code, any C function is a tag. Use -t to tag typedefs.\n\
528 Use -T to tag definitions of `struct', `union' and `enum'.\n\
529 Use -d to tag `#define' macro definitions and `enum' constants.\n\
530 Use --globals to tag global variables.\n\
531 You can tag function declarations and external variables by\n\
532 using `--declarations', and struct members by using `--members'.";
533 #else /* C help for Etags */
534 static const char default_C_help
[] =
535 "In C code, any C function or typedef is a tag, and so are\n\
536 definitions of `struct', `union' and `enum'. `#define' macro\n\
537 definitions and `enum' constants are tags unless you specify\n\
538 `--no-defines'. Global variables are tags unless you specify\n\
539 `--no-globals' and so are struct members unless you specify\n\
540 `--no-members'. Use of `--no-globals', `--no-defines' and\n\
541 `--no-members' can make the tags table file much smaller.\n\
542 You can tag function declarations and external variables by\n\
543 using `--declarations'.";
544 #endif /* C help for Ctags and Etags */
546 static const char *Cplusplus_suffixes
[] =
547 { "C", "c++", "cc", "cpp", "cxx", "H", "h++", "hh", "hpp", "hxx",
548 "M", /* Objective C++ */
549 "pdb", /* PostScript with C syntax */
551 static const char Cplusplus_help
[] =
552 "In C++ code, all the tag constructs of C code are tagged. (Use\n\
553 --help --lang=c --lang=c++ for full help.)\n\
554 In addition to C tags, member functions are also recognized. Member\n\
555 variables are recognized unless you use the `--no-members' option.\n\
556 Tags for variables and functions in classes are named `CLASS::VARIABLE'\n\
557 and `CLASS::FUNCTION'. `operator' definitions have tag names like\n\
560 static const char *Cjava_suffixes
[] =
562 static char Cjava_help
[] =
563 "In Java code, all the tags constructs of C and C++ code are\n\
564 tagged. (Use --help --lang=c --lang=c++ --lang=java for full help.)";
567 static const char *Cobol_suffixes
[] =
568 { "COB", "cob", NULL
};
569 static char Cobol_help
[] =
570 "In Cobol code, tags are paragraph names; that is, any word\n\
571 starting in column 8 and followed by a period.";
573 static const char *Cstar_suffixes
[] =
574 { "cs", "hs", NULL
};
576 static const char *Erlang_suffixes
[] =
577 { "erl", "hrl", NULL
};
578 static const char Erlang_help
[] =
579 "In Erlang code, the tags are the functions, records and macros\n\
580 defined in the file.";
582 const char *Forth_suffixes
[] =
583 { "fth", "tok", NULL
};
584 static const char Forth_help
[] =
585 "In Forth code, tags are words defined by `:',\n\
586 constant, code, create, defer, value, variable, buffer:, field.";
588 static const char *Fortran_suffixes
[] =
589 { "F", "f", "f90", "for", NULL
};
590 static const char Fortran_help
[] =
591 "In Fortran code, functions, subroutines and block data are tags.";
593 static const char *HTML_suffixes
[] =
594 { "htm", "html", "shtml", NULL
};
595 static const char HTML_help
[] =
596 "In HTML input files, the tags are the `title' and the `h1', `h2',\n\
597 `h3' headers. Also, tags are `name=' in anchors and all\n\
598 occurrences of `id='.";
600 static const char *Lisp_suffixes
[] =
601 { "cl", "clisp", "el", "l", "lisp", "LSP", "lsp", "ml", NULL
};
602 static const char Lisp_help
[] =
603 "In Lisp code, any function defined with `defun', any variable\n\
604 defined with `defvar' or `defconst', and in general the first\n\
605 argument of any expression that starts with `(def' in column zero\n\
607 The `--declarations' option tags \"(defvar foo)\" constructs too.";
609 static const char *Lua_suffixes
[] =
610 { "lua", "LUA", NULL
};
611 static const char Lua_help
[] =
612 "In Lua scripts, all functions are tags.";
614 static const char *Makefile_filenames
[] =
615 { "Makefile", "makefile", "GNUMakefile", "Makefile.in", "Makefile.am", NULL
};
616 static const char Makefile_help
[] =
617 "In makefiles, targets are tags; additionally, variables are tags\n\
618 unless you specify `--no-globals'.";
620 static const char *Objc_suffixes
[] =
621 { "lm", /* Objective lex file */
622 "m", /* Objective C file */
624 static const char Objc_help
[] =
625 "In Objective C code, tags include Objective C definitions for classes,\n\
626 class categories, methods and protocols. Tags for variables and\n\
627 functions in classes are named `CLASS::VARIABLE' and `CLASS::FUNCTION'.\n\
628 (Use --help --lang=c --lang=objc --lang=java for full help.)";
630 static const char *Pascal_suffixes
[] =
631 { "p", "pas", NULL
};
632 static const char Pascal_help
[] =
633 "In Pascal code, the tags are the functions and procedures defined\n\
635 /* " // this is for working around an Emacs highlighting bug... */
637 static const char *Perl_suffixes
[] =
638 { "pl", "pm", NULL
};
639 static const char *Perl_interpreters
[] =
640 { "perl", "@PERL@", NULL
};
641 static const char Perl_help
[] =
642 "In Perl code, the tags are the packages, subroutines and variables\n\
643 defined by the `package', `sub', `my' and `local' keywords. Use\n\
644 `--globals' if you want to tag global variables. Tags for\n\
645 subroutines are named `PACKAGE::SUB'. The name for subroutines\n\
646 defined in the default package is `main::SUB'.";
648 static const char *PHP_suffixes
[] =
649 { "php", "php3", "php4", NULL
};
650 static const char PHP_help
[] =
651 "In PHP code, tags are functions, classes and defines. Unless you use\n\
652 the `--no-members' option, vars are tags too.";
654 static const char *plain_C_suffixes
[] =
655 { "pc", /* Pro*C file */
658 static const char *PS_suffixes
[] =
659 { "ps", "psw", NULL
}; /* .psw is for PSWrap */
660 static const char PS_help
[] =
661 "In PostScript code, the tags are the functions.";
663 static const char *Prolog_suffixes
[] =
665 static const char Prolog_help
[] =
666 "In Prolog code, tags are predicates and rules at the beginning of\n\
669 static const char *Python_suffixes
[] =
671 static const char Python_help
[] =
672 "In Python code, `def' or `class' at the beginning of a line\n\
675 /* Can't do the `SCM' or `scm' prefix with a version number. */
676 static const char *Scheme_suffixes
[] =
677 { "oak", "sch", "scheme", "SCM", "scm", "SM", "sm", "ss", "t", NULL
};
678 static const char Scheme_help
[] =
679 "In Scheme code, tags include anything defined with `def' or with a\n\
680 construct whose name starts with `def'. They also include\n\
681 variables set with `set!' at top level in the file.";
683 static const char *TeX_suffixes
[] =
684 { "bib", "clo", "cls", "ltx", "sty", "TeX", "tex", NULL
};
685 static const char TeX_help
[] =
686 "In LaTeX text, the argument of any of the commands `\\chapter',\n\
687 `\\section', `\\subsection', `\\subsubsection', `\\eqno', `\\label',\n\
688 `\\ref', `\\cite', `\\bibitem', `\\part', `\\appendix', `\\entry',\n\
689 `\\index', `\\def', `\\newcommand', `\\renewcommand',\n\
690 `\\newenvironment' or `\\renewenvironment' is a tag.\n\
692 Other commands can be specified by setting the environment variable\n\
693 `TEXTAGS' to a colon-separated list like, for example,\n\
694 TEXTAGS=\"mycommand:myothercommand\".";
697 static const char *Texinfo_suffixes
[] =
698 { "texi", "texinfo", "txi", NULL
};
699 static const char Texinfo_help
[] =
700 "for texinfo files, lines starting with @node are tagged.";
702 static const char *Yacc_suffixes
[] =
703 { "y", "y++", "ym", "yxx", "yy", NULL
}; /* .ym is Objective yacc file */
704 static const char Yacc_help
[] =
705 "In Bison or Yacc input files, each rule defines as a tag the\n\
706 nonterminal it constructs. The portions of the file that contain\n\
707 C code are parsed as C code (use --help --lang=c --lang=yacc\n\
710 static const char auto_help
[] =
711 "`auto' is not a real language, it indicates to use\n\
712 a default language for files base on file name suffix and file contents.";
714 static const char none_help
[] =
715 "`none' is not a real language, it indicates to only do\n\
716 regexp processing on files.";
718 static const char no_lang_help
[] =
719 "No detailed help available for this language.";
723 * Table of languages.
725 * It is ok for a given function to be listed under more than one
726 * name. I just didn't.
729 static language lang_names
[] =
731 { "ada", Ada_help
, Ada_funcs
, Ada_suffixes
},
732 { "asm", Asm_help
, Asm_labels
, Asm_suffixes
},
733 { "c", default_C_help
, default_C_entries
, default_C_suffixes
},
734 { "c++", Cplusplus_help
, Cplusplus_entries
, Cplusplus_suffixes
},
735 { "c*", no_lang_help
, Cstar_entries
, Cstar_suffixes
},
736 { "cobol", Cobol_help
, Cobol_paragraphs
, Cobol_suffixes
},
737 { "erlang", Erlang_help
, Erlang_functions
, Erlang_suffixes
},
738 { "forth", Forth_help
, Forth_words
, Forth_suffixes
},
739 { "fortran", Fortran_help
, Fortran_functions
, Fortran_suffixes
},
740 { "html", HTML_help
, HTML_labels
, HTML_suffixes
},
741 { "java", Cjava_help
, Cjava_entries
, Cjava_suffixes
},
742 { "lisp", Lisp_help
, Lisp_functions
, Lisp_suffixes
},
743 { "lua", Lua_help
, Lua_functions
, Lua_suffixes
},
744 { "makefile", Makefile_help
,Makefile_targets
,NULL
,Makefile_filenames
},
745 { "objc", Objc_help
, plain_C_entries
, Objc_suffixes
},
746 { "pascal", Pascal_help
, Pascal_functions
, Pascal_suffixes
},
747 { "perl",Perl_help
,Perl_functions
,Perl_suffixes
,NULL
,Perl_interpreters
},
748 { "php", PHP_help
, PHP_functions
, PHP_suffixes
},
749 { "postscript",PS_help
, PS_functions
, PS_suffixes
},
750 { "proc", no_lang_help
, plain_C_entries
, plain_C_suffixes
},
751 { "prolog", Prolog_help
, Prolog_functions
, Prolog_suffixes
},
752 { "python", Python_help
, Python_functions
, Python_suffixes
},
753 { "scheme", Scheme_help
, Scheme_functions
, Scheme_suffixes
},
754 { "tex", TeX_help
, TeX_commands
, TeX_suffixes
},
755 { "texinfo", Texinfo_help
, Texinfo_nodes
, Texinfo_suffixes
},
756 { "yacc", Yacc_help
,Yacc_entries
,Yacc_suffixes
,NULL
,NULL
,true},
757 { "auto", auto_help
}, /* default guessing scheme */
758 { "none", none_help
, just_read_file
}, /* regexp matching only */
759 { NULL
} /* end of list */
764 print_language_names (void)
767 const char **name
, **ext
;
769 puts ("\nThese are the currently supported languages, along with the\n\
770 default file names and dot suffixes:");
771 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
773 printf (" %-*s", 10, lang
->name
);
774 if (lang
->filenames
!= NULL
)
775 for (name
= lang
->filenames
; *name
!= NULL
; name
++)
776 printf (" %s", *name
);
777 if (lang
->suffixes
!= NULL
)
778 for (ext
= lang
->suffixes
; *ext
!= NULL
; ext
++)
779 printf (" .%s", *ext
);
782 puts ("where `auto' means use default language for files based on file\n\
783 name suffix, and `none' means only do regexp processing on files.\n\
784 If no language is specified and no matching suffix is found,\n\
785 the first line of the file is read for a sharp-bang (#!) sequence\n\
786 followed by the name of an interpreter. If no such sequence is found,\n\
787 Fortran is tried first; if no tags are found, C is tried next.\n\
788 When parsing any C file, a \"class\" or \"template\" keyword\n\
790 puts ("Compressed files are supported using gzip, bzip2, and xz.\n\
792 For detailed help on a given language use, for example,\n\
793 etags --help --lang=ada.");
797 # define EMACS_NAME "standalone"
800 # define VERSION "17.38.1.4"
802 static _Noreturn
void
805 char emacs_copyright
[] = COPYRIGHT
;
807 printf ("%s (%s %s)\n", (CTAGS
) ? "ctags" : "etags", EMACS_NAME
, VERSION
);
808 puts (emacs_copyright
);
809 puts ("This program is distributed under the terms in ETAGS.README");
814 #ifndef PRINT_UNDOCUMENTED_OPTIONS_HELP
815 # define PRINT_UNDOCUMENTED_OPTIONS_HELP false
818 static _Noreturn
void
819 print_help (argument
*argbuffer
)
821 bool help_for_lang
= false;
823 for (; argbuffer
->arg_type
!= at_end
; argbuffer
++)
824 if (argbuffer
->arg_type
== at_language
)
828 puts (argbuffer
->lang
->help
);
829 help_for_lang
= true;
835 printf ("Usage: %s [options] [[regex-option ...] file-name] ...\n\
837 These are the options accepted by %s.\n", progname
, progname
);
838 puts ("You may use unambiguous abbreviations for the long option names.");
839 puts (" A - as file name means read names from stdin (one per line).\n\
840 Absolute names are stored in the output file as they are.\n\
841 Relative ones are stored relative to the output file's directory.\n");
843 puts ("-a, --append\n\
844 Append tag entries to existing tags file.");
846 puts ("--packages-only\n\
847 For Ada files, only generate tags for packages.");
850 puts ("-B, --backward-search\n\
851 Write the search commands for the tag entries using '?', the\n\
852 backward-search command instead of '/', the forward-search command.");
854 /* This option is mostly obsolete, because etags can now automatically
855 detect C++. Retained for backward compatibility and for debugging and
856 experimentation. In principle, we could want to tag as C++ even
857 before any "class" or "template" keyword.
859 Treat files whose name suffix defaults to C language as C++ files.");
862 puts ("--declarations\n\
863 In C and derived languages, create tags for function declarations,");
865 puts ("\tand create tags for extern variables if --globals is used.");
868 ("\tand create tags for extern variables unless --no-globals is used.");
871 puts ("-d, --defines\n\
872 Create tag entries for C #define constants and enum constants, too.");
874 puts ("-D, --no-defines\n\
875 Don't create tag entries for C #define constants and enum constants.\n\
876 This makes the tags file smaller.");
879 puts ("-i FILE, --include=FILE\n\
880 Include a note in tag file indicating that, when searching for\n\
881 a tag, one should also consult the tags file FILE after\n\
882 checking the current file.");
884 puts ("-l LANG, --language=LANG\n\
885 Force the following files to be considered as written in the\n\
886 named language up to the next --language=LANG option.");
890 Create tag entries for global variables in some languages.");
892 puts ("--no-globals\n\
893 Do not create tag entries for global variables in some\n\
894 languages. This makes the tags file smaller.");
896 if (PRINT_UNDOCUMENTED_OPTIONS_HELP
)
897 puts ("--no-line-directive\n\
898 Ignore #line preprocessor directives in C and derived languages.");
902 Create tag entries for members of structures in some languages.");
904 puts ("--no-members\n\
905 Do not create tag entries for members of structures\n\
906 in some languages.");
908 puts ("-r REGEXP, --regex=REGEXP or --regex=@regexfile\n\
909 Make a tag for each line matching a regular expression pattern\n\
910 in the following files. {LANGUAGE}REGEXP uses REGEXP for LANGUAGE\n\
911 files only. REGEXFILE is a file containing one REGEXP per line.\n\
912 REGEXP takes the form /TAGREGEXP/TAGNAME/MODS, where TAGNAME/ is\n\
913 optional. The TAGREGEXP pattern is anchored (as if preceded by ^).");
914 puts (" If TAGNAME/ is present, the tags created are named.\n\
915 For example Tcl named tags can be created with:\n\
916 --regex=\"/proc[ \\t]+\\([^ \\t]+\\)/\\1/.\".\n\
917 MODS are optional one-letter modifiers: `i' means to ignore case,\n\
918 `m' means to allow multi-line matches, `s' implies `m' and\n\
919 causes dot to match any character, including newline.");
921 puts ("-R, --no-regex\n\
922 Don't create tags from regexps for the following files.");
924 puts ("-I, --ignore-indentation\n\
925 In C and C++ do not assume that a closing brace in the first\n\
926 column is the final brace of a function or structure definition.");
928 puts ("-o FILE, --output=FILE\n\
929 Write the tags to FILE.");
931 puts ("--parse-stdin=NAME\n\
932 Read from standard input and record tags as belonging to file NAME.");
936 puts ("-t, --typedefs\n\
937 Generate tag entries for C and Ada typedefs.");
938 puts ("-T, --typedefs-and-c++\n\
939 Generate tag entries for C typedefs, C struct/enum/union tags,\n\
940 and C++ member functions.");
944 puts ("-u, --update\n\
945 Update the tag entries for the given files, leaving tag\n\
946 entries for other files in place. Currently, this is\n\
947 implemented by deleting the existing entries for the given\n\
948 files and then rewriting the new entries at the end of the\n\
949 tags file. It is often faster to simply rebuild the entire\n\
950 tag file than to use this.");
954 puts ("-v, --vgrind\n\
955 Print on the standard output an index of items intended for\n\
956 human consumption, similar to the output of vgrind. The index\n\
957 is sorted, and gives the page number of each item.");
959 if (PRINT_UNDOCUMENTED_OPTIONS_HELP
)
960 puts ("-w, --no-duplicates\n\
961 Do not create duplicate tag entries, for compatibility with\n\
962 traditional ctags.");
964 if (PRINT_UNDOCUMENTED_OPTIONS_HELP
)
965 puts ("-w, --no-warn\n\
966 Suppress warning messages about duplicate tag entries.");
968 puts ("-x, --cxref\n\
969 Like --vgrind, but in the style of cxref, rather than vgrind.\n\
970 The output uses line numbers instead of page numbers, but\n\
971 beyond that the differences are cosmetic; try both to see\n\
975 puts ("-V, --version\n\
976 Print the version of the program.\n\
978 Print this help message.\n\
979 Followed by one or more `--language' options prints detailed\n\
980 help about tag generation for the specified languages.");
982 print_language_names ();
985 puts ("Report bugs to bug-gnu-emacs@gnu.org");
992 main (int argc
, char **argv
)
995 unsigned int nincluded_files
;
996 char **included_files
;
998 int current_arg
, file_count
;
999 linebuffer filename_lb
;
1000 bool help_asked
= false;
1006 nincluded_files
= 0;
1007 included_files
= xnew (argc
, char *);
1011 /* Allocate enough no matter what happens. Overkill, but each one
1013 argbuffer
= xnew (argc
, argument
);
1016 * Always find typedefs and structure tags.
1017 * Also default to find macro constants, enum constants, struct
1018 * members and global variables. Do it for both etags and ctags.
1020 typedefs
= typedefs_or_cplusplus
= constantypedefs
= true;
1021 globals
= members
= true;
1023 /* When the optstring begins with a '-' getopt_long does not rearrange the
1024 non-options arguments to be at the end, but leaves them alone. */
1025 optstring
= concat ("-ac:Cf:Il:o:r:RSVhH",
1026 (CTAGS
) ? "BxdtTuvw" : "Di:",
1029 while ((opt
= getopt_long (argc
, argv
, optstring
, longopts
, NULL
)) != EOF
)
1033 /* If getopt returns 0, then it has already processed a
1034 long-named option. We should do nothing. */
1038 /* This means that a file name has been seen. Record it. */
1039 argbuffer
[current_arg
].arg_type
= at_filename
;
1040 argbuffer
[current_arg
].what
= optarg
;
1041 len
= strlen (optarg
);
1042 if (whatlen_max
< len
)
1049 /* Parse standard input. Idea by Vivek <vivek@etla.org>. */
1050 argbuffer
[current_arg
].arg_type
= at_stdin
;
1051 argbuffer
[current_arg
].what
= optarg
;
1052 len
= strlen (optarg
);
1053 if (whatlen_max
< len
)
1058 fatal ("cannot parse standard input more than once", (char *)NULL
);
1059 parsing_stdin
= true;
1062 /* Common options. */
1063 case 'a': append_to_tagfile
= true; break;
1064 case 'C': cplusplus
= true; break;
1065 case 'f': /* for compatibility with old makefiles */
1069 error ("-o option may only be given once.");
1070 suggest_asking_for_help ();
1076 case 'S': /* for backward compatibility */
1077 ignoreindent
= true;
1081 language
*lang
= get_language_from_langname (optarg
);
1084 argbuffer
[current_arg
].lang
= lang
;
1085 argbuffer
[current_arg
].arg_type
= at_language
;
1091 /* Backward compatibility: support obsolete --ignore-case-regexp. */
1092 optarg
= concat (optarg
, "i", ""); /* memory leak here */
1095 argbuffer
[current_arg
].arg_type
= at_regexp
;
1096 argbuffer
[current_arg
].what
= optarg
;
1097 len
= strlen (optarg
);
1098 if (whatlen_max
< len
)
1103 argbuffer
[current_arg
].arg_type
= at_regexp
;
1104 argbuffer
[current_arg
].what
= NULL
;
1116 case 'D': constantypedefs
= false; break;
1117 case 'i': included_files
[nincluded_files
++] = optarg
; break;
1119 /* Ctags options. */
1120 case 'B': searchar
= '?'; break;
1121 case 'd': constantypedefs
= true; break;
1122 case 't': typedefs
= true; break;
1123 case 'T': typedefs
= typedefs_or_cplusplus
= true; break;
1124 case 'u': update
= true; break;
1125 case 'v': vgrind_style
= true; /*FALLTHRU*/
1126 case 'x': cxref_style
= true; break;
1127 case 'w': no_warnings
= true; break;
1129 suggest_asking_for_help ();
1133 /* No more options. Store the rest of arguments. */
1134 for (; optind
< argc
; optind
++)
1136 argbuffer
[current_arg
].arg_type
= at_filename
;
1137 argbuffer
[current_arg
].what
= argv
[optind
];
1138 len
= strlen (argv
[optind
]);
1139 if (whatlen_max
< len
)
1145 argbuffer
[current_arg
].arg_type
= at_end
;
1148 print_help (argbuffer
);
1151 if (nincluded_files
== 0 && file_count
== 0)
1153 error ("no input files specified.");
1154 suggest_asking_for_help ();
1158 if (tagfile
== NULL
)
1159 tagfile
= savestr (CTAGS
? "tags" : "TAGS");
1160 cwd
= etags_getcwd (); /* the current working directory */
1161 if (cwd
[strlen (cwd
) - 1] != '/')
1164 cwd
= concat (oldcwd
, "/", "");
1168 /* Compute base directory for relative file names. */
1169 if (streq (tagfile
, "-")
1170 || strneq (tagfile
, "/dev/", 5))
1171 tagfiledir
= cwd
; /* relative file names are relative to cwd */
1174 canonicalize_filename (tagfile
);
1175 tagfiledir
= absolute_dirname (tagfile
, cwd
);
1178 init (); /* set up boolean "functions" */
1180 linebuffer_init (&lb
);
1181 linebuffer_init (&filename_lb
);
1182 linebuffer_init (&filebuf
);
1183 linebuffer_init (&token_name
);
1187 if (streq (tagfile
, "-"))
1190 SET_BINARY (fileno (stdout
));
1193 tagf
= fopen (tagfile
, append_to_tagfile
? "ab" : "wb");
1199 * Loop through files finding functions.
1201 for (i
= 0; i
< current_arg
; i
++)
1203 static language
*lang
; /* non-NULL if language is forced */
1206 switch (argbuffer
[i
].arg_type
)
1209 lang
= argbuffer
[i
].lang
;
1212 analyse_regex (argbuffer
[i
].what
);
1215 this_file
= argbuffer
[i
].what
;
1216 /* Input file named "-" means read file names from stdin
1217 (one per line) and use them. */
1218 if (streq (this_file
, "-"))
1221 fatal ("cannot parse standard input AND read file names from it",
1223 while (readline_internal (&filename_lb
, stdin
) > 0)
1224 process_file_name (filename_lb
.buffer
, lang
);
1227 process_file_name (this_file
, lang
);
1230 this_file
= argbuffer
[i
].what
;
1231 process_file (stdin
, this_file
, lang
);
1238 free (filebuf
.buffer
);
1239 free (token_name
.buffer
);
1241 if (!CTAGS
|| cxref_style
)
1243 /* Write the remaining tags to tagf (ETAGS) or stdout (CXREF). */
1244 put_entries (nodehead
);
1245 free_tree (nodehead
);
1251 /* Output file entries that have no tags. */
1252 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
1254 fprintf (tagf
, "\f\n%s,0\n", fdp
->taggedfname
);
1256 while (nincluded_files
-- > 0)
1257 fprintf (tagf
, "\f\n%s,include\n", *included_files
++);
1259 if (fclose (tagf
) == EOF
)
1263 exit (EXIT_SUCCESS
);
1266 /* From here on, we are in (CTAGS && !cxref_style) */
1270 xmalloc (strlen (tagfile
) + whatlen_max
+
1271 sizeof "mv..OTAGS;fgrep -v '\t\t' OTAGS >;rm OTAGS");
1272 for (i
= 0; i
< current_arg
; ++i
)
1274 switch (argbuffer
[i
].arg_type
)
1280 continue; /* the for loop */
1282 strcpy (cmd
, "mv ");
1283 strcat (cmd
, tagfile
);
1284 strcat (cmd
, " OTAGS;fgrep -v '\t");
1285 strcat (cmd
, argbuffer
[i
].what
);
1286 strcat (cmd
, "\t' OTAGS >");
1287 strcat (cmd
, tagfile
);
1288 strcat (cmd
, ";rm OTAGS");
1289 if (system (cmd
) != EXIT_SUCCESS
)
1290 fatal ("failed to execute shell command", (char *)NULL
);
1293 append_to_tagfile
= true;
1296 tagf
= fopen (tagfile
, append_to_tagfile
? "ab" : "wb");
1299 put_entries (nodehead
); /* write all the tags (CTAGS) */
1300 free_tree (nodehead
);
1302 if (fclose (tagf
) == EOF
)
1306 if (append_to_tagfile
|| update
)
1308 char *cmd
= xmalloc (2 * strlen (tagfile
) + sizeof "sort -u -o..");
1309 /* Maybe these should be used:
1310 setenv ("LC_COLLATE", "C", 1);
1311 setenv ("LC_ALL", "C", 1); */
1312 strcpy (cmd
, "sort -u -o ");
1313 strcat (cmd
, tagfile
);
1315 strcat (cmd
, tagfile
);
1316 exit (system (cmd
));
1318 return EXIT_SUCCESS
;
1323 * Return a compressor given the file name. If EXTPTR is non-zero,
1324 * return a pointer into FILE where the compressor-specific
1325 * extension begins. If no compressor is found, NULL is returned
1326 * and EXTPTR is not significant.
1327 * Idea by Vladimir Alexiev <vladimir@cs.ualberta.ca> (1998)
1330 get_compressor_from_suffix (char *file
, char **extptr
)
1333 char *slash
, *suffix
;
1335 /* File has been processed by canonicalize_filename,
1336 so we don't need to consider backslashes on DOS_NT. */
1337 slash
= etags_strrchr (file
, '/');
1338 suffix
= etags_strrchr (file
, '.');
1339 if (suffix
== NULL
|| suffix
< slash
)
1344 /* Let those poor souls who live with DOS 8+3 file name limits get
1345 some solace by treating foo.cgz as if it were foo.c.gz, etc.
1346 Only the first do loop is run if not MSDOS */
1349 for (compr
= compressors
; compr
->suffix
!= NULL
; compr
++)
1350 if (streq (compr
->suffix
, suffix
))
1353 break; /* do it only once: not really a loop */
1356 } while (*suffix
!= '\0');
1363 * Return a language given the name.
1366 get_language_from_langname (const char *name
)
1371 error ("empty language name");
1374 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1375 if (streq (name
, lang
->name
))
1377 error ("unknown language \"%s\"", name
);
1385 * Return a language given the interpreter name.
1388 get_language_from_interpreter (char *interpreter
)
1393 if (interpreter
== NULL
)
1395 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1396 if (lang
->interpreters
!= NULL
)
1397 for (iname
= lang
->interpreters
; *iname
!= NULL
; iname
++)
1398 if (streq (*iname
, interpreter
))
1407 * Return a language given the file name.
1410 get_language_from_filename (char *file
, int case_sensitive
)
1413 const char **name
, **ext
, *suffix
;
1415 /* Try whole file name first. */
1416 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1417 if (lang
->filenames
!= NULL
)
1418 for (name
= lang
->filenames
; *name
!= NULL
; name
++)
1419 if ((case_sensitive
)
1420 ? streq (*name
, file
)
1421 : strcaseeq (*name
, file
))
1424 /* If not found, try suffix after last dot. */
1425 suffix
= etags_strrchr (file
, '.');
1429 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1430 if (lang
->suffixes
!= NULL
)
1431 for (ext
= lang
->suffixes
; *ext
!= NULL
; ext
++)
1432 if ((case_sensitive
)
1433 ? streq (*ext
, suffix
)
1434 : strcaseeq (*ext
, suffix
))
1441 * This routine is called on each file argument.
1444 process_file_name (char *file
, language
*lang
)
1446 struct stat stat_buf
;
1450 char *compressed_name
, *uncompressed_name
;
1451 char *ext
, *real_name
;
1454 canonicalize_filename (file
);
1455 if (streq (file
, tagfile
) && !streq (tagfile
, "-"))
1457 error ("skipping inclusion of %s in self.", file
);
1460 if ((compr
= get_compressor_from_suffix (file
, &ext
)) == NULL
)
1462 compressed_name
= NULL
;
1463 real_name
= uncompressed_name
= savestr (file
);
1467 real_name
= compressed_name
= savestr (file
);
1468 uncompressed_name
= savenstr (file
, ext
- file
);
1471 /* If the canonicalized uncompressed name
1472 has already been dealt with, skip it silently. */
1473 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
1475 assert (fdp
->infname
!= NULL
);
1476 if (streq (uncompressed_name
, fdp
->infname
))
1480 if (stat (real_name
, &stat_buf
) != 0)
1482 /* Reset real_name and try with a different name. */
1484 if (compressed_name
!= NULL
) /* try with the given suffix */
1486 if (stat (uncompressed_name
, &stat_buf
) == 0)
1487 real_name
= uncompressed_name
;
1489 else /* try all possible suffixes */
1491 for (compr
= compressors
; compr
->suffix
!= NULL
; compr
++)
1493 compressed_name
= concat (file
, ".", compr
->suffix
);
1494 if (stat (compressed_name
, &stat_buf
) != 0)
1498 char *suf
= compressed_name
+ strlen (file
);
1499 size_t suflen
= strlen (compr
->suffix
) + 1;
1500 for ( ; suf
[1]; suf
++, suflen
--)
1502 memmove (suf
, suf
+ 1, suflen
);
1503 if (stat (compressed_name
, &stat_buf
) == 0)
1505 real_name
= compressed_name
;
1509 if (real_name
!= NULL
)
1512 free (compressed_name
);
1513 compressed_name
= NULL
;
1517 real_name
= compressed_name
;
1522 if (real_name
== NULL
)
1527 } /* try with a different name */
1529 if (!S_ISREG (stat_buf
.st_mode
))
1531 error ("skipping %s: it is not a regular file.", real_name
);
1534 if (real_name
== compressed_name
)
1536 char *cmd
= concat (compr
->command
, " ", real_name
);
1537 inf
= popen (cmd
, "rb");
1541 inf
= fopen (real_name
, "rb");
1548 process_file (inf
, uncompressed_name
, lang
);
1550 if (real_name
== compressed_name
)
1551 retval
= pclose (inf
);
1553 retval
= fclose (inf
);
1558 free (compressed_name
);
1559 free (uncompressed_name
);
1566 process_file (FILE *fh
, char *fn
, language
*lang
)
1568 static const fdesc emptyfdesc
;
1571 /* Create a new input file description entry. */
1572 fdp
= xnew (1, fdesc
);
1575 fdp
->infname
= savestr (fn
);
1577 fdp
->infabsname
= absolute_filename (fn
, cwd
);
1578 fdp
->infabsdir
= absolute_dirname (fn
, cwd
);
1579 if (filename_is_absolute (fn
))
1581 /* An absolute file name. Canonicalize it. */
1582 fdp
->taggedfname
= absolute_filename (fn
, NULL
);
1586 /* A file name relative to cwd. Make it relative
1587 to the directory of the tags file. */
1588 fdp
->taggedfname
= relative_filename (fn
, tagfiledir
);
1590 fdp
->usecharno
= true; /* use char position when making tags */
1592 fdp
->written
= false; /* not written on tags file yet */
1595 curfdp
= fdhead
; /* the current file description */
1599 /* If not Ctags, and if this is not metasource and if it contained no #line
1600 directives, we can write the tags and free all nodes pointing to
1603 && curfdp
->usecharno
/* no #line directives in this file */
1604 && !curfdp
->lang
->metasource
)
1608 /* Look for the head of the sublist relative to this file. See add_node
1609 for the structure of the node tree. */
1611 for (np
= nodehead
; np
!= NULL
; prev
= np
, np
= np
->left
)
1612 if (np
->fdp
== curfdp
)
1615 /* If we generated tags for this file, write and delete them. */
1618 /* This is the head of the last sublist, if any. The following
1619 instructions depend on this being true. */
1620 assert (np
->left
== NULL
);
1622 assert (fdhead
== curfdp
);
1623 assert (last_node
->fdp
== curfdp
);
1624 put_entries (np
); /* write tags for file curfdp->taggedfname */
1625 free_tree (np
); /* remove the written nodes */
1627 nodehead
= NULL
; /* no nodes left */
1629 prev
->left
= NULL
; /* delete the pointer to the sublist */
1635 * This routine sets up the boolean pseudo-functions which work
1636 * by setting boolean flags dependent upon the corresponding character.
1637 * Every char which is NOT in that string is not a white char. Therefore,
1638 * all of the array "_wht" is set to false, and then the elements
1639 * subscripted by the chars in "white" are set to true. Thus "_wht"
1640 * of a char is true if it is the string "white", else false.
1648 for (i
= 0; i
< CHARS
; i
++)
1649 iswhite (i
) = notinname (i
) = begtoken (i
) = intoken (i
) = endtoken (i
)
1651 for (sp
= white
; *sp
!= '\0'; sp
++) iswhite (*sp
) = true;
1652 for (sp
= nonam
; *sp
!= '\0'; sp
++) notinname (*sp
) = true;
1653 notinname ('\0') = notinname ('\n');
1654 for (sp
= begtk
; *sp
!= '\0'; sp
++) begtoken (*sp
) = true;
1655 begtoken ('\0') = begtoken ('\n');
1656 for (sp
= midtk
; *sp
!= '\0'; sp
++) intoken (*sp
) = true;
1657 intoken ('\0') = intoken ('\n');
1658 for (sp
= endtk
; *sp
!= '\0'; sp
++) endtoken (*sp
) = true;
1659 endtoken ('\0') = endtoken ('\n');
1663 * This routine opens the specified file and calls the function
1664 * which finds the function and type definitions.
1667 find_entries (FILE *inf
)
1670 language
*lang
= curfdp
->lang
;
1671 Lang_function
*parser
= NULL
;
1673 /* If user specified a language, use it. */
1674 if (lang
!= NULL
&& lang
->function
!= NULL
)
1676 parser
= lang
->function
;
1679 /* Else try to guess the language given the file name. */
1682 lang
= get_language_from_filename (curfdp
->infname
, true);
1683 if (lang
!= NULL
&& lang
->function
!= NULL
)
1685 curfdp
->lang
= lang
;
1686 parser
= lang
->function
;
1690 /* Else look for sharp-bang as the first two characters. */
1692 && readline_internal (&lb
, inf
) > 0
1694 && lb
.buffer
[0] == '#'
1695 && lb
.buffer
[1] == '!')
1699 /* Set lp to point at the first char after the last slash in the
1700 line or, if no slashes, at the first nonblank. Then set cp to
1701 the first successive blank and terminate the string. */
1702 lp
= etags_strrchr (lb
.buffer
+2, '/');
1706 lp
= skip_spaces (lb
.buffer
+ 2);
1707 cp
= skip_non_spaces (lp
);
1710 if (strlen (lp
) > 0)
1712 lang
= get_language_from_interpreter (lp
);
1713 if (lang
!= NULL
&& lang
->function
!= NULL
)
1715 curfdp
->lang
= lang
;
1716 parser
= lang
->function
;
1721 /* We rewind here, even if inf may be a pipe. We fail if the
1722 length of the first line is longer than the pipe block size,
1723 which is unlikely. */
1726 /* Else try to guess the language given the case insensitive file name. */
1729 lang
= get_language_from_filename (curfdp
->infname
, false);
1730 if (lang
!= NULL
&& lang
->function
!= NULL
)
1732 curfdp
->lang
= lang
;
1733 parser
= lang
->function
;
1737 /* Else try Fortran or C. */
1740 node
*old_last_node
= last_node
;
1742 curfdp
->lang
= get_language_from_langname ("fortran");
1745 if (old_last_node
== last_node
)
1746 /* No Fortran entries found. Try C. */
1748 /* We do not tag if rewind fails.
1749 Only the file name will be recorded in the tags file. */
1751 curfdp
->lang
= get_language_from_langname (cplusplus
? "c++" : "c");
1757 if (!no_line_directive
1758 && curfdp
->lang
!= NULL
&& curfdp
->lang
->metasource
)
1759 /* It may be that this is a bingo.y file, and we already parsed a bingo.c
1760 file, or anyway we parsed a file that is automatically generated from
1761 this one. If this is the case, the bingo.c file contained #line
1762 directives that generated tags pointing to this file. Let's delete
1763 them all before parsing this file, which is the real source. */
1765 fdesc
**fdpp
= &fdhead
;
1766 while (*fdpp
!= NULL
)
1768 && streq ((*fdpp
)->taggedfname
, curfdp
->taggedfname
))
1769 /* We found one of those! We must delete both the file description
1770 and all tags referring to it. */
1772 fdesc
*badfdp
= *fdpp
;
1774 /* Delete the tags referring to badfdp->taggedfname
1775 that were obtained from badfdp->infname. */
1776 invalidate_nodes (badfdp
, &nodehead
);
1778 *fdpp
= badfdp
->next
; /* remove the bad description from the list */
1779 free_fdesc (badfdp
);
1782 fdpp
= &(*fdpp
)->next
; /* advance the list pointer */
1785 assert (parser
!= NULL
);
1787 /* Generic initializations before reading from file. */
1788 linebuffer_setlen (&filebuf
, 0); /* reset the file buffer */
1790 /* Generic initializations before parsing file with readline. */
1791 lineno
= 0; /* reset global line number */
1792 charno
= 0; /* reset global char number */
1793 linecharno
= 0; /* reset global char number of line start */
1797 regex_tag_multiline ();
1802 * Check whether an implicitly named tag should be created,
1803 * then call `pfnote'.
1804 * NAME is a string that is internally copied by this function.
1806 * TAGS format specification
1807 * Idea by Sam Kendall <kendall@mv.mv.com> (1997)
1808 * The following is explained in some more detail in etc/ETAGS.EBNF.
1810 * make_tag creates tags with "implicit tag names" (unnamed tags)
1811 * if the following are all true, assuming NONAM=" \f\t\n\r()=,;":
1812 * 1. NAME does not contain any of the characters in NONAM;
1813 * 2. LINESTART contains name as either a rightmost, or rightmost but
1814 * one character, substring;
1815 * 3. the character, if any, immediately before NAME in LINESTART must
1816 * be a character in NONAM;
1817 * 4. the character, if any, immediately after NAME in LINESTART must
1818 * also be a character in NONAM.
1820 * The implementation uses the notinname() macro, which recognizes the
1821 * characters stored in the string `nonam'.
1822 * etags.el needs to use the same characters that are in NONAM.
1825 make_tag (const char *name
, /* tag name, or NULL if unnamed */
1826 int namelen
, /* tag length */
1827 bool is_func
, /* tag is a function */
1828 char *linestart
, /* start of the line where tag is */
1829 int linelen
, /* length of the line where tag is */
1830 int lno
, /* line number */
1831 long int cno
) /* character number */
1833 bool named
= (name
!= NULL
&& namelen
> 0);
1836 if (!CTAGS
&& named
) /* maybe set named to false */
1837 /* Let's try to make an implicit tag name, that is, create an unnamed tag
1838 such that etags.el can guess a name from it. */
1841 register const char *cp
= name
;
1843 for (i
= 0; i
< namelen
; i
++)
1844 if (notinname (*cp
++))
1846 if (i
== namelen
) /* rule #1 */
1848 cp
= linestart
+ linelen
- namelen
;
1849 if (notinname (linestart
[linelen
-1]))
1850 cp
-= 1; /* rule #4 */
1851 if (cp
>= linestart
/* rule #2 */
1853 || notinname (cp
[-1])) /* rule #3 */
1854 && strneq (name
, cp
, namelen
)) /* rule #2 */
1855 named
= false; /* use implicit tag name */
1860 nname
= savenstr (name
, namelen
);
1862 pfnote (nname
, is_func
, linestart
, linelen
, lno
, cno
);
1867 pfnote (char *name
, bool is_func
, char *linestart
, int linelen
, int lno
,
1869 /* tag name, or NULL if unnamed */
1870 /* tag is a function */
1871 /* start of the line where tag is */
1872 /* length of the line where tag is */
1874 /* character number */
1878 assert (name
== NULL
|| name
[0] != '\0');
1879 if (CTAGS
&& name
== NULL
)
1882 np
= xnew (1, node
);
1884 /* If ctags mode, change name "main" to M<thisfilename>. */
1885 if (CTAGS
&& !cxref_style
&& streq (name
, "main"))
1887 register char *fp
= etags_strrchr (curfdp
->taggedfname
, '/');
1888 np
->name
= concat ("M", fp
== NULL
? curfdp
->taggedfname
: fp
+ 1, "");
1889 fp
= etags_strrchr (np
->name
, '.');
1890 if (fp
!= NULL
&& fp
[1] != '\0' && fp
[2] == '\0')
1896 np
->been_warned
= false;
1898 np
->is_func
= is_func
;
1900 if (np
->fdp
->usecharno
)
1901 /* Our char numbers are 0-base, because of C language tradition?
1902 ctags compatibility? old versions compatibility? I don't know.
1903 Anyway, since emacs's are 1-base we expect etags.el to take care
1904 of the difference. If we wanted to have 1-based numbers, we would
1905 uncomment the +1 below. */
1906 np
->cno
= cno
/* + 1 */ ;
1908 np
->cno
= invalidcharno
;
1909 np
->left
= np
->right
= NULL
;
1910 if (CTAGS
&& !cxref_style
)
1912 if (strlen (linestart
) < 50)
1913 np
->regex
= concat (linestart
, "$", "");
1915 np
->regex
= savenstr (linestart
, 50);
1918 np
->regex
= savenstr (linestart
, linelen
);
1920 add_node (np
, &nodehead
);
1925 * recurse on left children, iterate on right children.
1928 free_tree (register node
*np
)
1932 register node
*node_right
= np
->right
;
1933 free_tree (np
->left
);
1943 * delete a file description
1946 free_fdesc (register fdesc
*fdp
)
1948 free (fdp
->infname
);
1949 free (fdp
->infabsname
);
1950 free (fdp
->infabsdir
);
1951 free (fdp
->taggedfname
);
1958 * Adds a node to the tree of nodes. In etags mode, sort by file
1959 * name. In ctags mode, sort by tag name. Make no attempt at
1962 * add_node is the only function allowed to add nodes, so it can
1966 add_node (node
*np
, node
**cur_node_p
)
1969 register node
*cur_node
= *cur_node_p
;
1971 if (cur_node
== NULL
)
1981 /* For each file name, tags are in a linked sublist on the right
1982 pointer. The first tags of different files are a linked list
1983 on the left pointer. last_node points to the end of the last
1985 if (last_node
!= NULL
&& last_node
->fdp
== np
->fdp
)
1987 /* Let's use the same sublist as the last added node. */
1988 assert (last_node
->right
== NULL
);
1989 last_node
->right
= np
;
1992 else if (cur_node
->fdp
== np
->fdp
)
1994 /* Scanning the list we found the head of a sublist which is
1995 good for us. Let's scan this sublist. */
1996 add_node (np
, &cur_node
->right
);
1999 /* The head of this sublist is not good for us. Let's try the
2001 add_node (np
, &cur_node
->left
);
2002 } /* if ETAGS mode */
2007 dif
= strcmp (np
->name
, cur_node
->name
);
2010 * If this tag name matches an existing one, then
2011 * do not add the node, but maybe print a warning.
2013 if (no_duplicates
&& !dif
)
2015 if (np
->fdp
== cur_node
->fdp
)
2019 fprintf (stderr
, "Duplicate entry in file %s, line %d: %s\n",
2020 np
->fdp
->infname
, lineno
, np
->name
);
2021 fprintf (stderr
, "Second entry ignored\n");
2024 else if (!cur_node
->been_warned
&& !no_warnings
)
2028 "Duplicate entry in files %s and %s: %s (Warning only)\n",
2029 np
->fdp
->infname
, cur_node
->fdp
->infname
, np
->name
);
2030 cur_node
->been_warned
= true;
2035 /* Actually add the node */
2036 add_node (np
, dif
< 0 ? &cur_node
->left
: &cur_node
->right
);
2037 } /* if CTAGS mode */
2041 * invalidate_nodes ()
2042 * Scan the node tree and invalidate all nodes pointing to the
2043 * given file description (CTAGS case) or free them (ETAGS case).
2046 invalidate_nodes (fdesc
*badfdp
, node
**npp
)
2055 if (np
->left
!= NULL
)
2056 invalidate_nodes (badfdp
, &np
->left
);
2057 if (np
->fdp
== badfdp
)
2059 if (np
->right
!= NULL
)
2060 invalidate_nodes (badfdp
, &np
->right
);
2064 assert (np
->fdp
!= NULL
);
2065 if (np
->fdp
== badfdp
)
2067 *npp
= np
->left
; /* detach the sublist from the list */
2068 np
->left
= NULL
; /* isolate it */
2069 free_tree (np
); /* free it */
2070 invalidate_nodes (badfdp
, npp
);
2073 invalidate_nodes (badfdp
, &np
->left
);
2078 static int total_size_of_entries (node
*);
2079 static int number_len (long) ATTRIBUTE_CONST
;
2081 /* Length of a non-negative number's decimal representation. */
2083 number_len (long int num
)
2086 while ((num
/= 10) > 0)
2092 * Return total number of characters that put_entries will output for
2093 * the nodes in the linked list at the right of the specified node.
2094 * This count is irrelevant with etags.el since emacs 19.34 at least,
2095 * but is still supplied for backward compatibility.
2098 total_size_of_entries (register node
*np
)
2100 register int total
= 0;
2102 for (; np
!= NULL
; np
= np
->right
)
2105 total
+= strlen (np
->regex
) + 1; /* pat\177 */
2106 if (np
->name
!= NULL
)
2107 total
+= strlen (np
->name
) + 1; /* name\001 */
2108 total
+= number_len ((long) np
->lno
) + 1; /* lno, */
2109 if (np
->cno
!= invalidcharno
) /* cno */
2110 total
+= number_len (np
->cno
);
2111 total
+= 1; /* newline */
2118 put_entries (register node
*np
)
2121 static fdesc
*fdp
= NULL
;
2126 /* Output subentries that precede this one */
2128 put_entries (np
->left
);
2130 /* Output this entry */
2139 fprintf (tagf
, "\f\n%s,%d\n",
2140 fdp
->taggedfname
, total_size_of_entries (np
));
2141 fdp
->written
= true;
2143 fputs (np
->regex
, tagf
);
2144 fputc ('\177', tagf
);
2145 if (np
->name
!= NULL
)
2147 fputs (np
->name
, tagf
);
2148 fputc ('\001', tagf
);
2150 fprintf (tagf
, "%d,", np
->lno
);
2151 if (np
->cno
!= invalidcharno
)
2152 fprintf (tagf
, "%ld", np
->cno
);
2158 if (np
->name
== NULL
)
2159 error ("internal error: NULL name in ctags mode.");
2164 fprintf (stdout
, "%s %s %d\n",
2165 np
->name
, np
->fdp
->taggedfname
, (np
->lno
+ 63) / 64);
2167 fprintf (stdout
, "%-16s %3d %-16s %s\n",
2168 np
->name
, np
->lno
, np
->fdp
->taggedfname
, np
->regex
);
2172 fprintf (tagf
, "%s\t%s\t", np
->name
, np
->fdp
->taggedfname
);
2175 { /* function or #define macro with args */
2176 putc (searchar
, tagf
);
2179 for (sp
= np
->regex
; *sp
; sp
++)
2181 if (*sp
== '\\' || *sp
== searchar
)
2185 putc (searchar
, tagf
);
2188 { /* anything else; text pattern inadequate */
2189 fprintf (tagf
, "%d", np
->lno
);
2194 } /* if this node contains a valid tag */
2196 /* Output subentries that follow this one */
2197 put_entries (np
->right
);
2199 put_entries (np
->left
);
2204 #define C_EXT 0x00fff /* C extensions */
2205 #define C_PLAIN 0x00000 /* C */
2206 #define C_PLPL 0x00001 /* C++ */
2207 #define C_STAR 0x00003 /* C* */
2208 #define C_JAVA 0x00005 /* JAVA */
2209 #define C_AUTO 0x01000 /* C, but switch to C++ if `class' is met */
2210 #define YACC 0x10000 /* yacc file */
2213 * The C symbol tables.
2218 st_C_objprot
, st_C_objimpl
, st_C_objend
,
2220 st_C_ignore
, st_C_attribute
,
2223 st_C_class
, st_C_template
,
2224 st_C_struct
, st_C_extern
, st_C_enum
, st_C_define
, st_C_typedef
2227 /* Feed stuff between (but not including) %[ and %] lines to:
2233 struct C_stab_entry { char *name; int c_ext; enum sym_type type; }
2237 while, 0, st_C_ignore
2238 switch, 0, st_C_ignore
2239 return, 0, st_C_ignore
2240 __attribute__, 0, st_C_attribute
2241 GTY, 0, st_C_attribute
2242 @interface, 0, st_C_objprot
2243 @protocol, 0, st_C_objprot
2244 @implementation,0, st_C_objimpl
2245 @end, 0, st_C_objend
2246 import, (C_JAVA & ~C_PLPL), st_C_ignore
2247 package, (C_JAVA & ~C_PLPL), st_C_ignore
2248 friend, C_PLPL, st_C_ignore
2249 extends, (C_JAVA & ~C_PLPL), st_C_javastruct
2250 implements, (C_JAVA & ~C_PLPL), st_C_javastruct
2251 interface, (C_JAVA & ~C_PLPL), st_C_struct
2252 class, 0, st_C_class
2253 namespace, C_PLPL, st_C_struct
2254 domain, C_STAR, st_C_struct
2255 union, 0, st_C_struct
2256 struct, 0, st_C_struct
2257 extern, 0, st_C_extern
2259 typedef, 0, st_C_typedef
2260 define, 0, st_C_define
2261 undef, 0, st_C_define
2262 operator, C_PLPL, st_C_operator
2263 template, 0, st_C_template
2264 # DEFUN used in emacs, the next three used in glibc (SYSCALL only for mach).
2265 DEFUN, 0, st_C_gnumacro
2266 SYSCALL, 0, st_C_gnumacro
2267 ENTRY, 0, st_C_gnumacro
2268 PSEUDO, 0, st_C_gnumacro
2269 # These are defined inside C functions, so currently they are not met.
2270 # EXFUN used in glibc, DEFVAR_* in emacs.
2271 #EXFUN, 0, st_C_gnumacro
2272 #DEFVAR_, 0, st_C_gnumacro
2274 and replace lines between %< and %> with its output, then:
2275 - remove the #if characterset check
2276 - make in_word_set static and not inline. */
2278 /* C code produced by gperf version 3.0.1 */
2279 /* Command-line: gperf -m 5 */
2280 /* Computed positions: -k'2-3' */
2282 struct C_stab_entry
{ const char *name
; int c_ext
; enum sym_type type
; };
2283 /* maximum key range = 33, duplicates = 0 */
2286 hash (const char *str
, int len
)
2288 static char const asso_values
[] =
2290 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2291 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2292 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2293 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2294 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2295 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2296 35, 35, 35, 35, 35, 35, 35, 35, 35, 3,
2297 26, 35, 35, 35, 35, 35, 35, 35, 27, 35,
2298 35, 35, 35, 24, 0, 35, 35, 35, 35, 0,
2299 35, 35, 35, 35, 35, 1, 35, 16, 35, 6,
2300 23, 0, 0, 35, 22, 0, 35, 35, 5, 0,
2301 0, 15, 1, 35, 6, 35, 8, 19, 35, 16,
2302 4, 5, 35, 35, 35, 35, 35, 35, 35, 35,
2303 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2304 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2305 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2306 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2307 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2308 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2309 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2310 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2311 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2312 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2313 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2314 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
2315 35, 35, 35, 35, 35, 35
2322 hval
+= asso_values
[(unsigned char) str
[2]];
2325 hval
+= asso_values
[(unsigned char) str
[1]];
2331 static struct C_stab_entry
*
2332 in_word_set (register const char *str
, register unsigned int len
)
2336 TOTAL_KEYWORDS
= 33,
2337 MIN_WORD_LENGTH
= 2,
2338 MAX_WORD_LENGTH
= 15,
2343 static struct C_stab_entry wordlist
[] =
2346 {"if", 0, st_C_ignore
},
2347 {"GTY", 0, st_C_attribute
},
2348 {"@end", 0, st_C_objend
},
2349 {"union", 0, st_C_struct
},
2350 {"define", 0, st_C_define
},
2351 {"import", (C_JAVA
& ~C_PLPL
), st_C_ignore
},
2352 {"template", 0, st_C_template
},
2353 {"operator", C_PLPL
, st_C_operator
},
2354 {"@interface", 0, st_C_objprot
},
2355 {"implements", (C_JAVA
& ~C_PLPL
), st_C_javastruct
},
2356 {"friend", C_PLPL
, st_C_ignore
},
2357 {"typedef", 0, st_C_typedef
},
2358 {"return", 0, st_C_ignore
},
2359 {"@implementation",0, st_C_objimpl
},
2360 {"@protocol", 0, st_C_objprot
},
2361 {"interface", (C_JAVA
& ~C_PLPL
), st_C_struct
},
2362 {"extern", 0, st_C_extern
},
2363 {"extends", (C_JAVA
& ~C_PLPL
), st_C_javastruct
},
2364 {"struct", 0, st_C_struct
},
2365 {"domain", C_STAR
, st_C_struct
},
2366 {"switch", 0, st_C_ignore
},
2367 {"enum", 0, st_C_enum
},
2368 {"for", 0, st_C_ignore
},
2369 {"namespace", C_PLPL
, st_C_struct
},
2370 {"class", 0, st_C_class
},
2371 {"while", 0, st_C_ignore
},
2372 {"undef", 0, st_C_define
},
2373 {"package", (C_JAVA
& ~C_PLPL
), st_C_ignore
},
2374 {"__attribute__", 0, st_C_attribute
},
2375 {"SYSCALL", 0, st_C_gnumacro
},
2376 {"ENTRY", 0, st_C_gnumacro
},
2377 {"PSEUDO", 0, st_C_gnumacro
},
2378 {"DEFUN", 0, st_C_gnumacro
}
2381 if (len
<= MAX_WORD_LENGTH
&& len
>= MIN_WORD_LENGTH
)
2383 int key
= hash (str
, len
);
2385 if (key
<= MAX_HASH_VALUE
&& key
>= 0)
2387 const char *s
= wordlist
[key
].name
;
2389 if (*str
== *s
&& !strncmp (str
+ 1, s
+ 1, len
- 1) && s
[len
] == '\0')
2390 return &wordlist
[key
];
2397 static enum sym_type
2398 C_symtype (char *str
, int len
, int c_ext
)
2400 register struct C_stab_entry
*se
= in_word_set (str
, len
);
2402 if (se
== NULL
|| (se
->c_ext
&& !(c_ext
& se
->c_ext
)))
2409 * Ignoring __attribute__ ((list))
2411 static bool inattribute
; /* looking at an __attribute__ construct */
2414 * C functions and variables are recognized using a simple
2415 * finite automaton. fvdef is its state variable.
2419 fvnone
, /* nothing seen */
2420 fdefunkey
, /* Emacs DEFUN keyword seen */
2421 fdefunname
, /* Emacs DEFUN name seen */
2422 foperator
, /* func: operator keyword seen (cplpl) */
2423 fvnameseen
, /* function or variable name seen */
2424 fstartlist
, /* func: just after open parenthesis */
2425 finlist
, /* func: in parameter list */
2426 flistseen
, /* func: after parameter list */
2427 fignore
, /* func: before open brace */
2428 vignore
/* var-like: ignore until ';' */
2431 static bool fvextern
; /* func or var: extern keyword seen; */
2434 * typedefs are recognized using a simple finite automaton.
2435 * typdef is its state variable.
2439 tnone
, /* nothing seen */
2440 tkeyseen
, /* typedef keyword seen */
2441 ttypeseen
, /* defined type seen */
2442 tinbody
, /* inside typedef body */
2443 tend
, /* just before typedef tag */
2444 tignore
/* junk after typedef tag */
2448 * struct-like structures (enum, struct and union) are recognized
2449 * using another simple finite automaton. `structdef' is its state
2454 snone
, /* nothing seen yet,
2455 or in struct body if bracelev > 0 */
2456 skeyseen
, /* struct-like keyword seen */
2457 stagseen
, /* struct-like tag seen */
2458 scolonseen
/* colon seen after struct-like tag */
2462 * When objdef is different from onone, objtag is the name of the class.
2464 static const char *objtag
= "<uninited>";
2467 * Yet another little state machine to deal with preprocessor lines.
2471 dnone
, /* nothing seen */
2472 dsharpseen
, /* '#' seen as first char on line */
2473 ddefineseen
, /* '#' and 'define' seen */
2474 dignorerest
/* ignore rest of line */
2478 * State machine for Objective C protocols and implementations.
2479 * Idea by Tom R.Hageman <tom@basil.icce.rug.nl> (1995)
2483 onone
, /* nothing seen */
2484 oprotocol
, /* @interface or @protocol seen */
2485 oimplementation
, /* @implementations seen */
2486 otagseen
, /* class name seen */
2487 oparenseen
, /* parenthesis before category seen */
2488 ocatseen
, /* category name seen */
2489 oinbody
, /* in @implementation body */
2490 omethodsign
, /* in @implementation body, after +/- */
2491 omethodtag
, /* after method name */
2492 omethodcolon
, /* after method colon */
2493 omethodparm
, /* after method parameter */
2494 oignore
/* wait for @end */
2499 * Use this structure to keep info about the token read, and how it
2500 * should be tagged. Used by the make_C_tag function to build a tag.
2504 char *line
; /* string containing the token */
2505 int offset
; /* where the token starts in LINE */
2506 int length
; /* token length */
2508 The previous members can be used to pass strings around for generic
2509 purposes. The following ones specifically refer to creating tags. In this
2510 case the token contained here is the pattern that will be used to create a
2513 bool valid
; /* do not create a tag; the token should be
2514 invalidated whenever a state machine is
2515 reset prematurely */
2516 bool named
; /* create a named tag */
2517 int lineno
; /* source line number of tag */
2518 long linepos
; /* source char number of tag */
2519 } token
; /* latest token read */
2522 * Variables and functions for dealing with nested structures.
2523 * Idea by Mykola Dzyuba <mdzyuba@yahoo.com> (2001)
2525 static void pushclass_above (int, char *, int);
2526 static void popclass_above (int);
2527 static void write_classname (linebuffer
*, const char *qualifier
);
2530 char **cname
; /* nested class names */
2531 int *bracelev
; /* nested class brace level */
2532 int nl
; /* class nesting level (elements used) */
2533 int size
; /* length of the array */
2534 } cstack
; /* stack for nested declaration tags */
2535 /* Current struct nesting depth (namespace, class, struct, union, enum). */
2536 #define nestlev (cstack.nl)
2537 /* After struct keyword or in struct body, not inside a nested function. */
2538 #define instruct (structdef == snone && nestlev > 0 \
2539 && bracelev == cstack.bracelev[nestlev-1] + 1)
2542 pushclass_above (int bracelev
, char *str
, int len
)
2546 popclass_above (bracelev
);
2548 if (nl
>= cstack
.size
)
2550 int size
= cstack
.size
*= 2;
2551 xrnew (cstack
.cname
, size
, char *);
2552 xrnew (cstack
.bracelev
, size
, int);
2554 assert (nl
== 0 || cstack
.bracelev
[nl
-1] < bracelev
);
2555 cstack
.cname
[nl
] = (str
== NULL
) ? NULL
: savenstr (str
, len
);
2556 cstack
.bracelev
[nl
] = bracelev
;
2561 popclass_above (int bracelev
)
2565 for (nl
= cstack
.nl
- 1;
2566 nl
>= 0 && cstack
.bracelev
[nl
] >= bracelev
;
2569 free (cstack
.cname
[nl
]);
2575 write_classname (linebuffer
*cn
, const char *qualifier
)
2578 int qlen
= strlen (qualifier
);
2580 if (cstack
.nl
== 0 || cstack
.cname
[0] == NULL
)
2584 cn
->buffer
[0] = '\0';
2588 len
= strlen (cstack
.cname
[0]);
2589 linebuffer_setlen (cn
, len
);
2590 strcpy (cn
->buffer
, cstack
.cname
[0]);
2592 for (i
= 1; i
< cstack
.nl
; i
++)
2594 char *s
= cstack
.cname
[i
];
2597 linebuffer_setlen (cn
, len
+ qlen
+ strlen (s
));
2598 len
+= sprintf (cn
->buffer
+ len
, "%s%s", qualifier
, s
);
2603 static bool consider_token (char *, int, int, int *, int, int, bool *);
2604 static void make_C_tag (bool);
2608 * checks to see if the current token is at the start of a
2609 * function or variable, or corresponds to a typedef, or
2610 * is a struct/union/enum tag, or #define, or an enum constant.
2612 * *IS_FUNC_OR_VAR gets true if the token is a function or #define macro
2613 * with args. C_EXTP points to which language we are looking at.
2624 consider_token (char *str
, int len
, int c
, int *c_extp
,
2625 int bracelev
, int parlev
, bool *is_func_or_var
)
2626 /* IN: token pointer */
2627 /* IN: token length */
2628 /* IN: first char after the token */
2629 /* IN, OUT: C extensions mask */
2630 /* IN: brace level */
2631 /* IN: parenthesis level */
2632 /* OUT: function or variable found */
2634 /* When structdef is stagseen, scolonseen, or snone with bracelev > 0,
2635 structtype is the type of the preceding struct-like keyword, and
2636 structbracelev is the brace level where it has been seen. */
2637 static enum sym_type structtype
;
2638 static int structbracelev
;
2639 static enum sym_type toktype
;
2642 toktype
= C_symtype (str
, len
, *c_extp
);
2645 * Skip __attribute__
2647 if (toktype
== st_C_attribute
)
2654 * Advance the definedef state machine.
2659 /* We're not on a preprocessor line. */
2660 if (toktype
== st_C_gnumacro
)
2667 if (toktype
== st_C_define
)
2669 definedef
= ddefineseen
;
2673 definedef
= dignorerest
;
2678 * Make a tag for any macro, unless it is a constant
2679 * and constantypedefs is false.
2681 definedef
= dignorerest
;
2682 *is_func_or_var
= (c
== '(');
2683 if (!*is_func_or_var
&& !constantypedefs
)
2690 error ("internal error: definedef value.");
2699 if (toktype
== st_C_typedef
)
2719 if (structdef
== snone
&& fvdef
== fvnone
)
2738 case st_C_javastruct
:
2739 if (structdef
== stagseen
)
2740 structdef
= scolonseen
;
2744 if ((*c_extp
& C_AUTO
) /* automatic detection of C++ language */
2746 && definedef
== dnone
&& structdef
== snone
2747 && typdef
== tnone
&& fvdef
== fvnone
)
2748 *c_extp
= (*c_extp
| C_PLPL
) & ~C_AUTO
;
2749 if (toktype
== st_C_template
)
2756 && (typdef
== tkeyseen
2757 || (typedefs_or_cplusplus
&& structdef
== snone
)))
2759 structdef
= skeyseen
;
2760 structtype
= toktype
;
2761 structbracelev
= bracelev
;
2762 if (fvdef
== fvnameseen
)
2768 if (structdef
== skeyseen
)
2770 structdef
= stagseen
;
2774 if (typdef
!= tnone
)
2777 /* Detect Objective C constructs. */
2787 objdef
= oimplementation
;
2791 case oimplementation
:
2792 /* Save the class tag for functions or variables defined inside. */
2793 objtag
= savenstr (str
, len
);
2797 /* Save the class tag for categories. */
2798 objtag
= savenstr (str
, len
);
2800 *is_func_or_var
= true;
2804 *is_func_or_var
= true;
2812 objdef
= omethodtag
;
2813 linebuffer_setlen (&token_name
, len
);
2814 memcpy (token_name
.buffer
, str
, len
);
2815 token_name
.buffer
[len
] = '\0';
2821 objdef
= omethodparm
;
2826 int oldlen
= token_name
.len
;
2828 objdef
= omethodtag
;
2829 linebuffer_setlen (&token_name
, oldlen
+ len
);
2830 memcpy (token_name
.buffer
+ oldlen
, str
, len
);
2831 token_name
.buffer
[oldlen
+ len
] = '\0';
2836 if (toktype
== st_C_objend
)
2838 /* Memory leakage here: the string pointed by objtag is
2839 never released, because many tests would be needed to
2840 avoid breaking on incorrect input code. The amount of
2841 memory leaked here is the sum of the lengths of the
2849 /* A function, variable or enum constant? */
2871 *is_func_or_var
= true;
2875 && structdef
== snone
2876 && structtype
== st_C_enum
&& bracelev
> structbracelev
)
2877 return true; /* enum constant */
2883 fvdef
= fdefunname
; /* GNU macro */
2884 *is_func_or_var
= true;
2892 if ((strneq (str
, "asm", 3) && endtoken (str
[3]))
2893 || (strneq (str
, "__asm__", 7) && endtoken (str
[7])))
2902 if (len
>= 10 && strneq (str
+len
-10, "::operator", 10))
2904 if (*c_extp
& C_AUTO
) /* automatic detection of C++ */
2905 *c_extp
= (*c_extp
| C_PLPL
) & ~C_AUTO
;
2907 *is_func_or_var
= true;
2910 if (bracelev
> 0 && !instruct
)
2912 fvdef
= fvnameseen
; /* function or variable */
2913 *is_func_or_var
= true;
2924 * C_entries often keeps pointers to tokens or lines which are older than
2925 * the line currently read. By keeping two line buffers, and switching
2926 * them at end of line, it is possible to use those pointers.
2934 #define current_lb_is_new (newndx == curndx)
2935 #define switch_line_buffers() (curndx = 1 - curndx)
2937 #define curlb (lbs[curndx].lb)
2938 #define newlb (lbs[newndx].lb)
2939 #define curlinepos (lbs[curndx].linepos)
2940 #define newlinepos (lbs[newndx].linepos)
2942 #define plainc ((c_ext & C_EXT) == C_PLAIN)
2943 #define cplpl (c_ext & C_PLPL)
2944 #define cjava ((c_ext & C_JAVA) == C_JAVA)
2946 #define CNL_SAVE_DEFINEDEF() \
2948 curlinepos = charno; \
2949 readline (&curlb, inf); \
2950 lp = curlb.buffer; \
2957 CNL_SAVE_DEFINEDEF(); \
2958 if (savetoken.valid) \
2960 token = savetoken; \
2961 savetoken.valid = false; \
2963 definedef = dnone; \
2968 make_C_tag (bool isfun
)
2970 /* This function is never called when token.valid is false, but
2971 we must protect against invalid input or internal errors. */
2973 make_tag (token_name
.buffer
, token_name
.len
, isfun
, token
.line
,
2974 token
.offset
+token
.length
+1, token
.lineno
, token
.linepos
);
2976 { /* this branch is optimized away if !DEBUG */
2977 make_tag (concat ("INVALID TOKEN:-->", token_name
.buffer
, ""),
2978 token_name
.len
+ 17, isfun
, token
.line
,
2979 token
.offset
+token
.length
+1, token
.lineno
, token
.linepos
);
2980 error ("INVALID TOKEN");
2983 token
.valid
= false;
2989 * This routine finds functions, variables, typedefs,
2990 * #define's, enum constants and struct/union/enum definitions in
2991 * C syntax and adds them to the list.
2994 C_entries (int c_ext
, FILE *inf
)
2995 /* extension of C */
2998 register char c
; /* latest char read; '\0' for end of line */
2999 register char *lp
; /* pointer one beyond the character `c' */
3000 int curndx
, newndx
; /* indices for current and new lb */
3001 register int tokoff
; /* offset in line of start of current token */
3002 register int toklen
; /* length of current token */
3003 const char *qualifier
; /* string used to qualify names */
3004 int qlen
; /* length of qualifier */
3005 int bracelev
; /* current brace level */
3006 int bracketlev
; /* current bracket level */
3007 int parlev
; /* current parenthesis level */
3008 int attrparlev
; /* __attribute__ parenthesis level */
3009 int templatelev
; /* current template level */
3010 int typdefbracelev
; /* bracelev where a typedef struct body begun */
3011 bool incomm
, inquote
, inchar
, quotednl
, midtoken
;
3012 bool yacc_rules
; /* in the rules part of a yacc file */
3013 struct tok savetoken
= {0}; /* token saved during preprocessor handling */
3016 linebuffer_init (&lbs
[0].lb
);
3017 linebuffer_init (&lbs
[1].lb
);
3018 if (cstack
.size
== 0)
3020 cstack
.size
= (DEBUG
) ? 1 : 4;
3022 cstack
.cname
= xnew (cstack
.size
, char *);
3023 cstack
.bracelev
= xnew (cstack
.size
, int);
3026 tokoff
= toklen
= typdefbracelev
= 0; /* keep compiler quiet */
3027 curndx
= newndx
= 0;
3031 fvdef
= fvnone
; fvextern
= false; typdef
= tnone
;
3032 structdef
= snone
; definedef
= dnone
; objdef
= onone
;
3034 midtoken
= inquote
= inchar
= incomm
= quotednl
= false;
3035 token
.valid
= savetoken
.valid
= false;
3036 bracelev
= bracketlev
= parlev
= attrparlev
= templatelev
= 0;
3038 { qualifier
= "."; qlen
= 1; }
3040 { qualifier
= "::"; qlen
= 2; }
3048 /* If we are at the end of the line, the next character is a
3049 '\0'; do not skip it, because it is what tells us
3050 to read the next line. */
3071 /* Newlines inside comments do not end macro definitions in
3073 CNL_SAVE_DEFINEDEF ();
3086 /* Newlines inside strings do not end macro definitions
3087 in traditional cpp, even though compilers don't
3088 usually accept them. */
3089 CNL_SAVE_DEFINEDEF ();
3099 /* Hmmm, something went wrong. */
3135 if (fvdef
!= finlist
&& fvdef
!= fignore
&& fvdef
!= vignore
)
3150 else if (/* cplpl && */ *lp
== '/')
3156 if ((c_ext
& YACC
) && *lp
== '%')
3158 /* Entering or exiting rules section in yacc file. */
3160 definedef
= dnone
; fvdef
= fvnone
; fvextern
= false;
3161 typdef
= tnone
; structdef
= snone
;
3162 midtoken
= inquote
= inchar
= incomm
= quotednl
= false;
3164 yacc_rules
= !yacc_rules
;
3170 if (definedef
== dnone
)
3173 bool cpptoken
= true;
3175 /* Look back on this line. If all blanks, or nonblanks
3176 followed by an end of comment, this is a preprocessor
3178 for (cp
= newlb
.buffer
; cp
< lp
-1; cp
++)
3181 if (*cp
== '*' && cp
[1] == '/')
3190 definedef
= dsharpseen
;
3191 } /* if (definedef == dnone) */
3202 CNL_SAVE_DEFINEDEF ();
3209 /* Consider token only if some involved conditions are satisfied. */
3210 if (typdef
!= tignore
3211 && definedef
!= dignorerest
3214 && (definedef
!= dnone
3215 || structdef
!= scolonseen
)
3222 if (c
== ':' && *lp
== ':' && begtoken (lp
[1]))
3223 /* This handles :: in the middle,
3224 but not at the beginning of an identifier.
3225 Also, space-separated :: is not recognized. */
3227 if (c_ext
& C_AUTO
) /* automatic detection of C++ */
3228 c_ext
= (c_ext
| C_PLPL
) & ~C_AUTO
;
3232 goto still_in_token
;
3236 bool funorvar
= false;
3239 || consider_token (newlb
.buffer
+ tokoff
, toklen
, c
,
3240 &c_ext
, bracelev
, parlev
,
3243 if (fvdef
== foperator
)
3246 lp
= skip_spaces (lp
-1);
3250 && !iswhite (*lp
) && *lp
!= '(')
3253 toklen
+= lp
- oldlp
;
3255 token
.named
= false;
3257 && nestlev
> 0 && definedef
== dnone
)
3258 /* in struct body */
3261 write_classname (&token_name
, qualifier
);
3262 len
= token_name
.len
;
3263 linebuffer_setlen (&token_name
, len
+qlen
+toklen
);
3264 sprintf (token_name
.buffer
+ len
, "%s%.*s",
3265 qualifier
, toklen
, newlb
.buffer
+ tokoff
);
3268 else if (objdef
== ocatseen
)
3269 /* Objective C category */
3271 int len
= strlen (objtag
) + 2 + toklen
;
3272 linebuffer_setlen (&token_name
, len
);
3273 sprintf (token_name
.buffer
, "%s(%.*s)",
3274 objtag
, toklen
, newlb
.buffer
+ tokoff
);
3277 else if (objdef
== omethodtag
3278 || objdef
== omethodparm
)
3279 /* Objective C method */
3283 else if (fvdef
== fdefunname
)
3284 /* GNU DEFUN and similar macros */
3286 bool defun
= (newlb
.buffer
[tokoff
] == 'F');
3290 /* Rewrite the tag so that emacs lisp DEFUNs
3291 can be found by their elisp name */
3297 linebuffer_setlen (&token_name
, len
);
3298 memcpy (token_name
.buffer
,
3299 newlb
.buffer
+ off
, len
);
3300 token_name
.buffer
[len
] = '\0';
3303 if (token_name
.buffer
[len
] == '_')
3304 token_name
.buffer
[len
] = '-';
3305 token
.named
= defun
;
3309 linebuffer_setlen (&token_name
, toklen
);
3310 memcpy (token_name
.buffer
,
3311 newlb
.buffer
+ tokoff
, toklen
);
3312 token_name
.buffer
[toklen
] = '\0';
3313 /* Name macros and members. */
3314 token
.named
= (structdef
== stagseen
3315 || typdef
== ttypeseen
3318 && definedef
== dignorerest
)
3320 && definedef
== dnone
3321 && structdef
== snone
3324 token
.lineno
= lineno
;
3325 token
.offset
= tokoff
;
3326 token
.length
= toklen
;
3327 token
.line
= newlb
.buffer
;
3328 token
.linepos
= newlinepos
;
3331 if (definedef
== dnone
3332 && (fvdef
== fvnameseen
3333 || fvdef
== foperator
3334 || structdef
== stagseen
3336 || typdef
== ttypeseen
3337 || objdef
!= onone
))
3339 if (current_lb_is_new
)
3340 switch_line_buffers ();
3342 else if (definedef
!= dnone
3343 || fvdef
== fdefunname
3345 make_C_tag (funorvar
);
3347 else /* not yacc and consider_token failed */
3349 if (inattribute
&& fvdef
== fignore
)
3351 /* We have just met __attribute__ after a
3352 function parameter list: do not tag the
3359 } /* if (endtoken (c)) */
3360 else if (intoken (c
))
3366 } /* if (midtoken) */
3367 else if (begtoken (c
))
3375 /* This prevents tagging fb in
3376 void (__attribute__((noreturn)) *fb) (void);
3377 Fixing this is not easy and not very important. */
3381 if (plainc
|| declarations
)
3383 make_C_tag (true); /* a function */
3388 if (structdef
== stagseen
&& !cjava
)
3390 popclass_above (bracelev
);
3398 if (!yacc_rules
|| lp
== newlb
.buffer
+ 1)
3400 tokoff
= lp
- 1 - newlb
.buffer
;
3405 } /* if (begtoken) */
3406 } /* if must look at token */
3409 /* Detect end of line, colon, comma, semicolon and various braces
3410 after having handled a token.*/
3416 if (yacc_rules
&& token
.offset
== 0 && token
.valid
)
3418 make_C_tag (false); /* a yacc function */
3421 if (definedef
!= dnone
)
3427 make_C_tag (true); /* an Objective C class */
3431 objdef
= omethodcolon
;
3432 linebuffer_setlen (&token_name
, token_name
.len
+ 1);
3433 strcat (token_name
.buffer
, ":");
3436 if (structdef
== stagseen
)
3438 structdef
= scolonseen
;
3441 /* Should be useless, but may be work as a safety net. */
3442 if (cplpl
&& fvdef
== flistseen
)
3444 make_C_tag (true); /* a function */
3450 if (definedef
!= dnone
|| inattribute
)
3456 make_C_tag (false); /* a typedef */
3466 if (typdef
== tignore
|| cplpl
)
3470 if ((globals
&& bracelev
== 0 && (!fvextern
|| declarations
))
3471 || (members
&& instruct
))
3472 make_C_tag (false); /* a variable */
3475 token
.valid
= false;
3479 && (cplpl
|| !instruct
)
3480 && (typdef
== tnone
|| (typdef
!= tignore
&& instruct
)))
3482 && plainc
&& instruct
))
3483 make_C_tag (true); /* a function */
3489 && cplpl
&& structdef
== stagseen
)
3490 make_C_tag (false); /* forward declaration */
3492 token
.valid
= false;
3493 } /* switch (fvdef) */
3499 if (structdef
== stagseen
)
3503 if (definedef
!= dnone
|| inattribute
)
3509 make_C_tag (true); /* an Objective C method */
3530 && (!fvextern
|| declarations
))
3531 || (members
&& instruct
)))
3532 make_C_tag (false); /* a variable */
3535 if ((declarations
&& typdef
== tnone
&& !instruct
)
3536 || (members
&& typdef
!= tignore
&& instruct
))
3538 make_C_tag (true); /* a function */
3541 else if (!declarations
)
3543 token
.valid
= false;
3548 if (structdef
== stagseen
)
3552 if (definedef
!= dnone
|| inattribute
)
3554 if (structdef
== stagseen
)
3561 make_C_tag (false); /* a typedef */
3573 if ((members
&& bracelev
== 1)
3574 || (globals
&& bracelev
== 0
3575 && (!fvextern
|| declarations
)))
3576 make_C_tag (false); /* a variable */
3590 if (definedef
!= dnone
)
3592 if (objdef
== otagseen
&& parlev
== 0)
3593 objdef
= oparenseen
;
3597 if (typdef
== ttypeseen
3601 /* This handles constructs like:
3602 typedef void OperatorFun (int fun); */
3621 if (--attrparlev
== 0)
3622 inattribute
= false;
3625 if (definedef
!= dnone
)
3627 if (objdef
== ocatseen
&& parlev
== 1)
3629 make_C_tag (true); /* an Objective C category */
3643 || typdef
== ttypeseen
))
3646 make_C_tag (false); /* a typedef */
3649 else if (parlev
< 0) /* can happen due to ill-conceived #if's. */
3653 if (definedef
!= dnone
)
3655 if (typdef
== ttypeseen
)
3657 /* Whenever typdef is set to tinbody (currently only
3658 here), typdefbracelev should be set to bracelev. */
3660 typdefbracelev
= bracelev
;
3665 make_C_tag (true); /* a function */
3674 make_C_tag (true); /* an Objective C class */
3679 make_C_tag (true); /* an Objective C method */
3683 /* Neutralize `extern "C" {' grot. */
3684 if (bracelev
== 0 && structdef
== snone
&& nestlev
== 0
3692 case skeyseen
: /* unnamed struct */
3693 pushclass_above (bracelev
, NULL
, 0);
3696 case stagseen
: /* named struct or enum */
3697 case scolonseen
: /* a class */
3698 pushclass_above (bracelev
,token
.line
+token
.offset
, token
.length
);
3700 make_C_tag (false); /* a struct or enum */
3706 if (definedef
!= dnone
)
3708 if (fvdef
== fstartlist
)
3710 fvdef
= fvnone
; /* avoid tagging `foo' in `foo (*bar()) ()' */
3711 token
.valid
= false;
3715 if (definedef
!= dnone
)
3718 if (!ignoreindent
&& lp
== newlb
.buffer
+ 1)
3721 token
.valid
= false; /* unexpected value, token unreliable */
3722 bracelev
= 0; /* reset brace level if first column */
3723 parlev
= 0; /* also reset paren level, just in case... */
3725 else if (bracelev
< 0)
3727 token
.valid
= false; /* something gone amiss, token unreliable */
3730 if (bracelev
== 0 && fvdef
== vignore
)
3731 fvdef
= fvnone
; /* end of function */
3732 popclass_above (bracelev
);
3734 /* Only if typdef == tinbody is typdefbracelev significant. */
3735 if (typdef
== tinbody
&& bracelev
<= typdefbracelev
)
3737 assert (bracelev
== typdefbracelev
);
3742 if (definedef
!= dnone
)
3752 if ((members
&& bracelev
== 1)
3753 || (globals
&& bracelev
== 0 && (!fvextern
|| declarations
)))
3754 make_C_tag (false); /* a variable */
3762 && (structdef
== stagseen
|| fvdef
== fvnameseen
))
3769 if (templatelev
> 0)
3777 if (objdef
== oinbody
&& bracelev
== 0)
3779 objdef
= omethodsign
;
3784 case '#': case '~': case '&': case '%': case '/':
3785 case '|': case '^': case '!': case '.': case '?':
3786 if (definedef
!= dnone
)
3788 /* These surely cannot follow a function tag in C. */
3801 if (objdef
== otagseen
)
3803 make_C_tag (true); /* an Objective C class */
3806 /* If a macro spans multiple lines don't reset its state. */
3808 CNL_SAVE_DEFINEDEF ();
3814 } /* while not eof */
3816 free (lbs
[0].lb
.buffer
);
3817 free (lbs
[1].lb
.buffer
);
3821 * Process either a C++ file or a C file depending on the setting
3825 default_C_entries (FILE *inf
)
3827 C_entries (cplusplus
? C_PLPL
: C_AUTO
, inf
);
3830 /* Always do plain C. */
3832 plain_C_entries (FILE *inf
)
3837 /* Always do C++. */
3839 Cplusplus_entries (FILE *inf
)
3841 C_entries (C_PLPL
, inf
);
3844 /* Always do Java. */
3846 Cjava_entries (FILE *inf
)
3848 C_entries (C_JAVA
, inf
);
3853 Cstar_entries (FILE *inf
)
3855 C_entries (C_STAR
, inf
);
3858 /* Always do Yacc. */
3860 Yacc_entries (FILE *inf
)
3862 C_entries (YACC
, inf
);
3866 /* Useful macros. */
3867 #define LOOP_ON_INPUT_LINES(file_pointer, line_buffer, char_pointer) \
3868 for (; /* loop initialization */ \
3869 !feof (file_pointer) /* loop test */ \
3870 && /* instructions at start of loop */ \
3871 (readline (&line_buffer, file_pointer), \
3872 char_pointer = line_buffer.buffer, \
3876 #define LOOKING_AT(cp, kw) /* kw is the keyword, a literal string */ \
3877 ((assert ("" kw), true) /* syntax error if not a literal string */ \
3878 && strneq ((cp), kw, sizeof (kw)-1) /* cp points at kw */ \
3879 && notinname ((cp)[sizeof (kw)-1]) /* end of kw */ \
3880 && ((cp) = skip_spaces ((cp)+sizeof (kw)-1))) /* skip spaces */
3882 /* Similar to LOOKING_AT but does not use notinname, does not skip */
3883 #define LOOKING_AT_NOCASE(cp, kw) /* the keyword is a literal string */ \
3884 ((assert ("" kw), true) /* syntax error if not a literal string */ \
3885 && strncaseeq ((cp), kw, sizeof (kw)-1) /* cp points at kw */ \
3886 && ((cp) += sizeof (kw)-1)) /* skip spaces */
3889 * Read a file, but do no processing. This is used to do regexp
3890 * matching on files that have no language defined.
3893 just_read_file (FILE *inf
)
3896 readline (&lb
, inf
);
3900 /* Fortran parsing */
3902 static void F_takeprec (void);
3903 static void F_getit (FILE *);
3908 dbp
= skip_spaces (dbp
);
3912 dbp
= skip_spaces (dbp
);
3913 if (strneq (dbp
, "(*)", 3))
3918 if (!ISDIGIT (*dbp
))
3920 --dbp
; /* force failure */
3925 while (ISDIGIT (*dbp
));
3933 dbp
= skip_spaces (dbp
);
3936 readline (&lb
, inf
);
3941 dbp
= skip_spaces (dbp
);
3943 if (!ISALPHA (*dbp
) && *dbp
!= '_' && *dbp
!= '$')
3945 for (cp
= dbp
+ 1; *cp
!= '\0' && intoken (*cp
); cp
++)
3947 make_tag (dbp
, cp
-dbp
, true,
3948 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
3953 Fortran_functions (FILE *inf
)
3955 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
3958 dbp
++; /* Ratfor escape to fortran */
3959 dbp
= skip_spaces (dbp
);
3963 if (LOOKING_AT_NOCASE (dbp
, "recursive"))
3964 dbp
= skip_spaces (dbp
);
3966 if (LOOKING_AT_NOCASE (dbp
, "pure"))
3967 dbp
= skip_spaces (dbp
);
3969 if (LOOKING_AT_NOCASE (dbp
, "elemental"))
3970 dbp
= skip_spaces (dbp
);
3972 switch (lowcase (*dbp
))
3975 if (nocase_tail ("integer"))
3979 if (nocase_tail ("real"))
3983 if (nocase_tail ("logical"))
3987 if (nocase_tail ("complex") || nocase_tail ("character"))
3991 if (nocase_tail ("double"))
3993 dbp
= skip_spaces (dbp
);
3996 if (nocase_tail ("precision"))
4002 dbp
= skip_spaces (dbp
);
4005 switch (lowcase (*dbp
))
4008 if (nocase_tail ("function"))
4012 if (nocase_tail ("subroutine"))
4016 if (nocase_tail ("entry"))
4020 if (nocase_tail ("blockdata") || nocase_tail ("block data"))
4022 dbp
= skip_spaces (dbp
);
4023 if (*dbp
== '\0') /* assume un-named */
4024 make_tag ("blockdata", 9, true,
4025 lb
.buffer
, dbp
- lb
.buffer
, lineno
, linecharno
);
4027 F_getit (inf
); /* look for name */
4038 * Philippe Waroquiers (1998)
4041 /* Once we are positioned after an "interesting" keyword, let's get
4042 the real tag value necessary. */
4044 Ada_getit (FILE *inf
, const char *name_qualifier
)
4052 dbp
= skip_spaces (dbp
);
4054 || (dbp
[0] == '-' && dbp
[1] == '-'))
4056 readline (&lb
, inf
);
4059 switch (lowcase (*dbp
))
4062 if (nocase_tail ("body"))
4064 /* Skipping body of procedure body or package body or ....
4065 resetting qualifier to body instead of spec. */
4066 name_qualifier
= "/b";
4071 /* Skipping type of task type or protected type ... */
4072 if (nocase_tail ("type"))
4079 for (cp
= dbp
; *cp
!= '\0' && *cp
!= '"'; cp
++)
4084 dbp
= skip_spaces (dbp
);
4087 && (ISALPHA (*cp
) || ISDIGIT (*cp
) || *cp
== '_' || *cp
== '.'));
4095 name
= concat (dbp
, name_qualifier
, "");
4097 make_tag (name
, strlen (name
), true,
4098 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4107 Ada_funcs (FILE *inf
)
4109 bool inquote
= false;
4110 bool skip_till_semicolumn
= false;
4112 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
4114 while (*dbp
!= '\0')
4116 /* Skip a string i.e. "abcd". */
4117 if (inquote
|| (*dbp
== '"'))
4119 dbp
= etags_strchr (dbp
+ !inquote
, '"');
4124 continue; /* advance char */
4129 break; /* advance line */
4133 /* Skip comments. */
4134 if (dbp
[0] == '-' && dbp
[1] == '-')
4135 break; /* advance line */
4137 /* Skip character enclosed in single quote i.e. 'a'
4138 and skip single quote starting an attribute i.e. 'Image. */
4147 if (skip_till_semicolumn
)
4150 skip_till_semicolumn
= false;
4152 continue; /* advance char */
4155 /* Search for beginning of a token. */
4156 if (!begtoken (*dbp
))
4159 continue; /* advance char */
4162 /* We are at the beginning of a token. */
4163 switch (lowcase (*dbp
))
4166 if (!packages_only
&& nocase_tail ("function"))
4167 Ada_getit (inf
, "/f");
4169 break; /* from switch */
4170 continue; /* advance char */
4172 if (!packages_only
&& nocase_tail ("procedure"))
4173 Ada_getit (inf
, "/p");
4174 else if (nocase_tail ("package"))
4175 Ada_getit (inf
, "/s");
4176 else if (nocase_tail ("protected")) /* protected type */
4177 Ada_getit (inf
, "/t");
4179 break; /* from switch */
4180 continue; /* advance char */
4183 if (typedefs
&& !packages_only
&& nocase_tail ("use"))
4185 /* when tagging types, avoid tagging use type Pack.Typename;
4186 for this, we will skip everything till a ; */
4187 skip_till_semicolumn
= true;
4188 continue; /* advance char */
4192 if (!packages_only
&& nocase_tail ("task"))
4193 Ada_getit (inf
, "/k");
4194 else if (typedefs
&& !packages_only
&& nocase_tail ("type"))
4196 Ada_getit (inf
, "/t");
4197 while (*dbp
!= '\0')
4201 break; /* from switch */
4202 continue; /* advance char */
4205 /* Look for the end of the token. */
4206 while (!endtoken (*dbp
))
4209 } /* advance char */
4210 } /* advance line */
4215 * Unix and microcontroller assembly tag handling
4216 * Labels: /^[a-zA-Z_.$][a-zA_Z0-9_.$]*[: ^I^J]/
4217 * Idea by Bob Weiner, Motorola Inc. (1994)
4220 Asm_labels (FILE *inf
)
4224 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4226 /* If first char is alphabetic or one of [_.$], test for colon
4227 following identifier. */
4228 if (ISALPHA (*cp
) || *cp
== '_' || *cp
== '.' || *cp
== '$')
4230 /* Read past label. */
4232 while (ISALNUM (*cp
) || *cp
== '_' || *cp
== '.' || *cp
== '$')
4234 if (*cp
== ':' || iswhite (*cp
))
4235 /* Found end of label, so copy it and add it to the table. */
4236 make_tag (lb
.buffer
, cp
- lb
.buffer
, true,
4237 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4245 * Perl sub names: /^sub[ \t\n]+[^ \t\n{]+/
4246 * /^use constant[ \t\n]+[^ \t\n{=,;]+/
4247 * Perl variable names: /^(my|local).../
4248 * Original code by Bart Robinson <lomew@cs.utah.edu> (1995)
4249 * Additions by Michael Ernst <mernst@alum.mit.edu> (1997)
4250 * Ideas by Kai Großjohann <Kai.Grossjohann@CS.Uni-Dortmund.DE> (2001)
4253 Perl_functions (FILE *inf
)
4255 char *package
= savestr ("main"); /* current package name */
4258 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4260 cp
= skip_spaces (cp
);
4262 if (LOOKING_AT (cp
, "package"))
4265 get_tag (cp
, &package
);
4267 else if (LOOKING_AT (cp
, "sub"))
4273 while (!notinname (*cp
))
4276 continue; /* nothing found */
4277 if ((pos
= etags_strchr (sp
, ':')) != NULL
4278 && pos
< cp
&& pos
[1] == ':')
4279 /* The name is already qualified. */
4280 make_tag (sp
, cp
- sp
, true,
4281 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4285 char savechar
, *name
;
4289 name
= concat (package
, "::", sp
);
4291 make_tag (name
, strlen (name
), true,
4292 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4296 else if (LOOKING_AT (cp
, "use constant")
4297 || LOOKING_AT (cp
, "use constant::defer"))
4299 /* For hash style multi-constant like
4300 use constant { FOO => 123,
4302 only the first FOO is picked up. Parsing across the value
4303 expressions would be difficult in general, due to possible nested
4304 hashes, here-documents, etc. */
4306 cp
= skip_spaces (cp
+1);
4309 else if (globals
) /* only if we are tagging global vars */
4311 /* Skip a qualifier, if any. */
4312 bool qual
= LOOKING_AT (cp
, "my") || LOOKING_AT (cp
, "local");
4313 /* After "my" or "local", but before any following paren or space. */
4314 char *varstart
= cp
;
4316 if (qual
/* should this be removed? If yes, how? */
4317 && (*cp
== '$' || *cp
== '@' || *cp
== '%'))
4322 while (ISALNUM (*cp
) || *cp
== '_');
4326 /* Should be examining a variable list at this point;
4327 could insist on seeing an open parenthesis. */
4328 while (*cp
!= '\0' && *cp
!= ';' && *cp
!= '=' && *cp
!= ')')
4334 make_tag (varstart
, cp
- varstart
, false,
4335 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4344 * Look for /^[\t]*def[ \t\n]+[^ \t\n(:]+/ or /^class[ \t\n]+[^ \t\n(:]+/
4345 * Idea by Eric S. Raymond <esr@thyrsus.com> (1997)
4346 * More ideas by seb bacon <seb@jamkit.com> (2002)
4349 Python_functions (FILE *inf
)
4353 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4355 cp
= skip_spaces (cp
);
4356 if (LOOKING_AT (cp
, "def") || LOOKING_AT (cp
, "class"))
4359 while (!notinname (*cp
) && *cp
!= ':')
4361 make_tag (name
, cp
- name
, true,
4362 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4371 * - /^[ \t]*function[ \t\n]+[^ \t\n(]+/
4372 * - /^[ \t]*class[ \t\n]+[^ \t\n]+/
4373 * - /^[ \t]*define\(\"[^\"]+/
4374 * Only with --members:
4375 * - /^[ \t]*var[ \t\n]+\$[^ \t\n=;]/
4376 * Idea by Diez B. Roggisch (2001)
4379 PHP_functions (FILE *inf
)
4382 bool search_identifier
= false;
4384 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4386 cp
= skip_spaces (cp
);
4388 if (search_identifier
4391 while (!notinname (*cp
))
4393 make_tag (name
, cp
- name
, true,
4394 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4395 search_identifier
= false;
4397 else if (LOOKING_AT (cp
, "function"))
4400 cp
= skip_spaces (cp
+1);
4404 while (!notinname (*cp
))
4406 make_tag (name
, cp
- name
, true,
4407 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4410 search_identifier
= true;
4412 else if (LOOKING_AT (cp
, "class"))
4417 while (*cp
!= '\0' && !iswhite (*cp
))
4419 make_tag (name
, cp
- name
, false,
4420 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4423 search_identifier
= true;
4425 else if (strneq (cp
, "define", 6)
4426 && (cp
= skip_spaces (cp
+6))
4428 && (*cp
== '"' || *cp
== '\''))
4432 while (*cp
!= quote
&& *cp
!= '\0')
4434 make_tag (name
, cp
- name
, false,
4435 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4438 && LOOKING_AT (cp
, "var")
4442 while (!notinname (*cp
))
4444 make_tag (name
, cp
- name
, false,
4445 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4452 * Cobol tag functions
4453 * We could look for anything that could be a paragraph name.
4454 * i.e. anything that starts in column 8 is one word and ends in a full stop.
4455 * Idea by Corny de Souza (1993)
4458 Cobol_paragraphs (FILE *inf
)
4460 register char *bp
, *ep
;
4462 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4468 /* If eoln, compiler option or comment ignore whole line. */
4469 if (bp
[-1] != ' ' || !ISALNUM (bp
[0]))
4472 for (ep
= bp
; ISALNUM (*ep
) || *ep
== '-'; ep
++)
4475 make_tag (bp
, ep
- bp
, true,
4476 lb
.buffer
, ep
- lb
.buffer
+ 1, lineno
, linecharno
);
4483 * Ideas by Assar Westerlund <assar@sics.se> (2001)
4486 Makefile_targets (FILE *inf
)
4490 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4492 if (*bp
== '\t' || *bp
== '#')
4494 while (*bp
!= '\0' && *bp
!= '=' && *bp
!= ':')
4496 if (*bp
== ':' || (globals
&& *bp
== '='))
4498 /* We should detect if there is more than one tag, but we do not.
4499 We just skip initial and final spaces. */
4500 char * namestart
= skip_spaces (lb
.buffer
);
4501 while (--bp
> namestart
)
4502 if (!notinname (*bp
))
4504 make_tag (namestart
, bp
- namestart
+ 1, true,
4505 lb
.buffer
, bp
- lb
.buffer
+ 2, lineno
, linecharno
);
4513 * Original code by Mosur K. Mohan (1989)
4515 * Locates tags for procedures & functions. Doesn't do any type- or
4516 * var-definitions. It does look for the keyword "extern" or
4517 * "forward" immediately following the procedure statement; if found,
4518 * the tag is skipped.
4521 Pascal_functions (FILE *inf
)
4523 linebuffer tline
; /* mostly copied from C_entries */
4525 int save_lineno
, namelen
, taglen
;
4528 bool /* each of these flags is true if: */
4529 incomment
, /* point is inside a comment */
4530 inquote
, /* point is inside '..' string */
4531 get_tagname
, /* point is after PROCEDURE/FUNCTION
4532 keyword, so next item = potential tag */
4533 found_tag
, /* point is after a potential tag */
4534 inparms
, /* point is within parameter-list */
4535 verify_tag
; /* point has passed the parm-list, so the
4536 next token will determine whether this
4537 is a FORWARD/EXTERN to be ignored, or
4538 whether it is a real tag */
4540 save_lcno
= save_lineno
= namelen
= taglen
= 0; /* keep compiler quiet */
4541 name
= NULL
; /* keep compiler quiet */
4544 linebuffer_init (&tline
);
4546 incomment
= inquote
= false;
4547 found_tag
= false; /* have a proc name; check if extern */
4548 get_tagname
= false; /* found "procedure" keyword */
4549 inparms
= false; /* found '(' after "proc" */
4550 verify_tag
= false; /* check if "extern" is ahead */
4553 while (!feof (inf
)) /* long main loop to get next char */
4556 if (c
== '\0') /* if end of line */
4558 readline (&lb
, inf
);
4562 if (!((found_tag
&& verify_tag
)
4564 c
= *dbp
++; /* only if don't need *dbp pointing
4565 to the beginning of the name of
4566 the procedure or function */
4570 if (c
== '}') /* within { } comments */
4572 else if (c
== '*' && *dbp
== ')') /* within (* *) comments */
4589 inquote
= true; /* found first quote */
4591 case '{': /* found open { comment */
4595 if (*dbp
== '*') /* found open (* comment */
4600 else if (found_tag
) /* found '(' after tag, i.e., parm-list */
4603 case ')': /* end of parms list */
4608 if (found_tag
&& !inparms
) /* end of proc or fn stmt */
4615 if (found_tag
&& verify_tag
&& (*dbp
!= ' '))
4617 /* Check if this is an "extern" declaration. */
4620 if (lowcase (*dbp
) == 'e')
4622 if (nocase_tail ("extern")) /* superfluous, really! */
4628 else if (lowcase (*dbp
) == 'f')
4630 if (nocase_tail ("forward")) /* check for forward reference */
4636 if (found_tag
&& verify_tag
) /* not external proc, so make tag */
4640 make_tag (name
, namelen
, true,
4641 tline
.buffer
, taglen
, save_lineno
, save_lcno
);
4645 if (get_tagname
) /* grab name of proc or fn */
4652 /* Find block name. */
4653 for (cp
= dbp
+ 1; *cp
!= '\0' && !endtoken (*cp
); cp
++)
4656 /* Save all values for later tagging. */
4657 linebuffer_setlen (&tline
, lb
.len
);
4658 strcpy (tline
.buffer
, lb
.buffer
);
4659 save_lineno
= lineno
;
4660 save_lcno
= linecharno
;
4661 name
= tline
.buffer
+ (dbp
- lb
.buffer
);
4663 taglen
= cp
- lb
.buffer
+ 1;
4665 dbp
= cp
; /* set dbp to e-o-token */
4666 get_tagname
= false;
4670 /* And proceed to check for "extern". */
4672 else if (!incomment
&& !inquote
&& !found_tag
)
4674 /* Check for proc/fn keywords. */
4675 switch (lowcase (c
))
4678 if (nocase_tail ("rocedure")) /* c = 'p', dbp has advanced */
4682 if (nocase_tail ("unction"))
4687 } /* while not eof */
4689 free (tline
.buffer
);
4694 * Lisp tag functions
4695 * look for (def or (DEF, quote or QUOTE
4698 static void L_getit (void);
4703 if (*dbp
== '\'') /* Skip prefix quote */
4705 else if (*dbp
== '(')
4708 /* Try to skip "(quote " */
4709 if (!LOOKING_AT (dbp
, "quote") && !LOOKING_AT (dbp
, "QUOTE"))
4710 /* Ok, then skip "(" before name in (defstruct (foo)) */
4711 dbp
= skip_spaces (dbp
);
4713 get_tag (dbp
, NULL
);
4717 Lisp_functions (FILE *inf
)
4719 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
4724 /* "(defvar foo)" is a declaration rather than a definition. */
4728 if (LOOKING_AT (p
, "defvar"))
4730 p
= skip_name (p
); /* past var name */
4731 p
= skip_spaces (p
);
4737 if (strneq (dbp
+ 1, "cl-", 3) || strneq (dbp
+ 1, "CL-", 3))
4740 if (strneq (dbp
+1, "def", 3) || strneq (dbp
+1, "DEF", 3))
4742 dbp
= skip_non_spaces (dbp
);
4743 dbp
= skip_spaces (dbp
);
4748 /* Check for (foo::defmumble name-defined ... */
4751 while (!notinname (*dbp
) && *dbp
!= ':');
4756 while (*dbp
== ':');
4758 if (strneq (dbp
, "def", 3) || strneq (dbp
, "DEF", 3))
4760 dbp
= skip_non_spaces (dbp
);
4761 dbp
= skip_spaces (dbp
);
4771 * Lua script language parsing
4772 * Original code by David A. Capello <dacap@users.sourceforge.net> (2004)
4774 * "function" and "local function" are tags if they start at column 1.
4777 Lua_functions (FILE *inf
)
4781 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4783 if (bp
[0] != 'f' && bp
[0] != 'l')
4786 (void)LOOKING_AT (bp
, "local"); /* skip possible "local" */
4788 if (LOOKING_AT (bp
, "function"))
4796 * Just look for lines where the first character is '/'
4797 * Also look at "defineps" for PSWrap
4799 * Richard Mlynarik <mly@adoc.xerox.com> (1997)
4800 * Masatake Yamato <masata-y@is.aist-nara.ac.jp> (1999)
4803 PS_functions (FILE *inf
)
4805 register char *bp
, *ep
;
4807 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4812 *ep
!= '\0' && *ep
!= ' ' && *ep
!= '{';
4815 make_tag (bp
, ep
- bp
, true,
4816 lb
.buffer
, ep
- lb
.buffer
+ 1, lineno
, linecharno
);
4818 else if (LOOKING_AT (bp
, "defineps"))
4826 * Ignore anything after \ followed by space or in ( )
4827 * Look for words defined by :
4828 * Look for constant, code, create, defer, value, and variable
4829 * OBP extensions: Look for buffer:, field,
4830 * Ideas by Eduardo Horvath <eeh@netbsd.org> (2004)
4833 Forth_words (FILE *inf
)
4837 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4838 while ((bp
= skip_spaces (bp
))[0] != '\0')
4839 if (bp
[0] == '\\' && iswhite (bp
[1]))
4840 break; /* read next line */
4841 else if (bp
[0] == '(' && iswhite (bp
[1]))
4842 do /* skip to ) or eol */
4844 while (*bp
!= ')' && *bp
!= '\0');
4845 else if ((bp
[0] == ':' && iswhite (bp
[1]) && bp
++)
4846 || LOOKING_AT_NOCASE (bp
, "constant")
4847 || LOOKING_AT_NOCASE (bp
, "code")
4848 || LOOKING_AT_NOCASE (bp
, "create")
4849 || LOOKING_AT_NOCASE (bp
, "defer")
4850 || LOOKING_AT_NOCASE (bp
, "value")
4851 || LOOKING_AT_NOCASE (bp
, "variable")
4852 || LOOKING_AT_NOCASE (bp
, "buffer:")
4853 || LOOKING_AT_NOCASE (bp
, "field"))
4854 get_tag (skip_spaces (bp
), NULL
); /* Yay! A definition! */
4856 bp
= skip_non_spaces (bp
);
4861 * Scheme tag functions
4862 * look for (def... xyzzy
4864 * (def ... ((...(xyzzy ....
4866 * Original code by Ken Haase (1985?)
4869 Scheme_functions (FILE *inf
)
4873 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4875 if (strneq (bp
, "(def", 4) || strneq (bp
, "(DEF", 4))
4877 bp
= skip_non_spaces (bp
+4);
4878 /* Skip over open parens and white space. Don't continue past
4880 while (*bp
&& notinname (*bp
))
4884 if (LOOKING_AT (bp
, "(SET!") || LOOKING_AT (bp
, "(set!"))
4890 /* Find tags in TeX and LaTeX input files. */
4892 /* TEX_toktab is a table of TeX control sequences that define tags.
4893 * Each entry records one such control sequence.
4895 * Original code from who knows whom.
4897 * Stefan Monnier (2002)
4900 static linebuffer
*TEX_toktab
= NULL
; /* Table with tag tokens */
4902 /* Default set of control sequences to put into TEX_toktab.
4903 The value of environment var TEXTAGS is prepended to this. */
4904 static const char *TEX_defenv
= "\
4905 :chapter:section:subsection:subsubsection:eqno:label:ref:cite:bibitem\
4906 :part:appendix:entry:index:def\
4907 :newcommand:renewcommand:newenvironment:renewenvironment";
4909 static void TEX_mode (FILE *);
4910 static void TEX_decode_env (const char *, const char *);
4912 static char TEX_esc
= '\\';
4913 static char TEX_opgrp
= '{';
4914 static char TEX_clgrp
= '}';
4917 * TeX/LaTeX scanning loop.
4920 TeX_commands (FILE *inf
)
4925 /* Select either \ or ! as escape character. */
4928 /* Initialize token table once from environment. */
4929 if (TEX_toktab
== NULL
)
4930 TEX_decode_env ("TEXTAGS", TEX_defenv
);
4932 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4934 /* Look at each TEX keyword in line. */
4937 /* Look for a TEX escape. */
4938 while (*cp
++ != TEX_esc
)
4939 if (cp
[-1] == '\0' || cp
[-1] == '%')
4942 for (key
= TEX_toktab
; key
->buffer
!= NULL
; key
++)
4943 if (strneq (cp
, key
->buffer
, key
->len
))
4946 int namelen
, linelen
;
4949 cp
= skip_spaces (cp
+ key
->len
);
4950 if (*cp
== TEX_opgrp
)
4956 (!iswhite (*p
) && *p
!= '#' &&
4957 *p
!= TEX_opgrp
&& *p
!= TEX_clgrp
);
4962 if (!opgrp
|| *p
== TEX_clgrp
)
4964 while (*p
!= '\0' && *p
!= TEX_opgrp
&& *p
!= TEX_clgrp
)
4966 linelen
= p
- lb
.buffer
+ 1;
4968 make_tag (cp
, namelen
, true,
4969 lb
.buffer
, linelen
, lineno
, linecharno
);
4970 goto tex_next_line
; /* We only tag a line once */
4978 #define TEX_LESC '\\'
4979 #define TEX_SESC '!'
4981 /* Figure out whether TeX's escapechar is '\\' or '!' and set grouping
4982 chars accordingly. */
4984 TEX_mode (FILE *inf
)
4988 while ((c
= getc (inf
)) != EOF
)
4990 /* Skip to next line if we hit the TeX comment char. */
4992 while (c
!= '\n' && c
!= EOF
)
4994 else if (c
== TEX_LESC
|| c
== TEX_SESC
)
5010 /* If the input file is compressed, inf is a pipe, and rewind may fail.
5011 No attempt is made to correct the situation. */
5015 /* Read environment and prepend it to the default string.
5016 Build token table. */
5018 TEX_decode_env (const char *evarname
, const char *defenv
)
5020 register const char *env
, *p
;
5023 /* Append default string to environment. */
5024 env
= getenv (evarname
);
5028 env
= concat (env
, defenv
, "");
5030 /* Allocate a token table */
5031 for (len
= 1, p
= env
; p
;)
5032 if ((p
= etags_strchr (p
, ':')) && *++p
!= '\0')
5034 TEX_toktab
= xnew (len
, linebuffer
);
5036 /* Unpack environment string into token table. Be careful about */
5037 /* zero-length strings (leading ':', "::" and trailing ':') */
5038 for (i
= 0; *env
!= '\0';)
5040 p
= etags_strchr (env
, ':');
5041 if (!p
) /* End of environment string. */
5042 p
= env
+ strlen (env
);
5044 { /* Only non-zero strings. */
5045 TEX_toktab
[i
].buffer
= savenstr (env
, p
- env
);
5046 TEX_toktab
[i
].len
= p
- env
;
5053 TEX_toktab
[i
].buffer
= NULL
; /* Mark end of table. */
5054 TEX_toktab
[i
].len
= 0;
5061 /* Texinfo support. Dave Love, Mar. 2000. */
5063 Texinfo_nodes (FILE *inf
)
5066 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5067 if (LOOKING_AT (cp
, "@node"))
5070 while (*cp
!= '\0' && *cp
!= ',')
5072 make_tag (start
, cp
- start
, true,
5073 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
5080 * Contents of <title>, <h1>, <h2>, <h3> are tags.
5081 * Contents of <a name=xxx> are tags with name xxx.
5083 * Francesco Potortì, 2002.
5086 HTML_labels (FILE *inf
)
5088 bool getnext
= false; /* next text outside of HTML tags is a tag */
5089 bool skiptag
= false; /* skip to the end of the current HTML tag */
5090 bool intag
= false; /* inside an html tag, looking for ID= */
5091 bool inanchor
= false; /* when INTAG, is an anchor, look for NAME= */
5095 linebuffer_setlen (&token_name
, 0); /* no name in buffer */
5097 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
5098 for (;;) /* loop on the same line */
5100 if (skiptag
) /* skip HTML tag */
5102 while (*dbp
!= '\0' && *dbp
!= '>')
5108 continue; /* look on the same line */
5110 break; /* go to next line */
5113 else if (intag
) /* look for "name=" or "id=" */
5115 while (*dbp
!= '\0' && *dbp
!= '>'
5116 && lowcase (*dbp
) != 'n' && lowcase (*dbp
) != 'i')
5119 break; /* go to next line */
5124 continue; /* look on the same line */
5126 if ((inanchor
&& LOOKING_AT_NOCASE (dbp
, "name="))
5127 || LOOKING_AT_NOCASE (dbp
, "id="))
5129 bool quoted
= (dbp
[0] == '"');
5132 for (end
= ++dbp
; *end
!= '\0' && *end
!= '"'; end
++)
5135 for (end
= dbp
; *end
!= '\0' && intoken (*end
); end
++)
5137 linebuffer_setlen (&token_name
, end
- dbp
);
5138 memcpy (token_name
.buffer
, dbp
, end
- dbp
);
5139 token_name
.buffer
[end
- dbp
] = '\0';
5142 intag
= false; /* we found what we looked for */
5143 skiptag
= true; /* skip to the end of the tag */
5144 getnext
= true; /* then grab the text */
5145 continue; /* look on the same line */
5150 else if (getnext
) /* grab next tokens and tag them */
5152 dbp
= skip_spaces (dbp
);
5154 break; /* go to next line */
5158 inanchor
= (lowcase (dbp
[1]) == 'a' && !intoken (dbp
[2]));
5159 continue; /* look on the same line */
5162 for (end
= dbp
+ 1; *end
!= '\0' && *end
!= '<'; end
++)
5164 make_tag (token_name
.buffer
, token_name
.len
, true,
5165 dbp
, end
- dbp
, lineno
, linecharno
);
5166 linebuffer_setlen (&token_name
, 0); /* no name in buffer */
5168 break; /* go to next line */
5171 else /* look for an interesting HTML tag */
5173 while (*dbp
!= '\0' && *dbp
!= '<')
5176 break; /* go to next line */
5178 if (lowcase (dbp
[1]) == 'a' && !intoken (dbp
[2]))
5181 continue; /* look on the same line */
5183 else if (LOOKING_AT_NOCASE (dbp
, "<title>")
5184 || LOOKING_AT_NOCASE (dbp
, "<h1>")
5185 || LOOKING_AT_NOCASE (dbp
, "<h2>")
5186 || LOOKING_AT_NOCASE (dbp
, "<h3>"))
5190 continue; /* look on the same line */
5201 * Assumes that the predicate or rule starts at column 0.
5202 * Only the first clause of a predicate or rule is added.
5203 * Original code by Sunichirou Sugou (1989)
5204 * Rewritten by Anders Lindgren (1996)
5206 static size_t prolog_pr (char *, char *);
5207 static void prolog_skip_comment (linebuffer
*, FILE *);
5208 static size_t prolog_atom (char *, size_t);
5211 Prolog_functions (FILE *inf
)
5221 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5223 if (cp
[0] == '\0') /* Empty line */
5225 else if (iswhite (cp
[0])) /* Not a predicate */
5227 else if (cp
[0] == '/' && cp
[1] == '*') /* comment. */
5228 prolog_skip_comment (&lb
, inf
);
5229 else if ((len
= prolog_pr (cp
, last
)) > 0)
5231 /* Predicate or rule. Store the function name so that we
5232 only generate a tag for the first clause. */
5234 last
= xnew (len
+ 1, char);
5235 else if (len
+ 1 > allocated
)
5236 xrnew (last
, len
+ 1, char);
5237 allocated
= len
+ 1;
5238 memcpy (last
, cp
, len
);
5247 prolog_skip_comment (linebuffer
*plb
, FILE *inf
)
5253 for (cp
= plb
->buffer
; *cp
!= '\0'; cp
++)
5254 if (cp
[0] == '*' && cp
[1] == '/')
5256 readline (plb
, inf
);
5258 while (!feof (inf
));
5262 * A predicate or rule definition is added if it matches:
5263 * <beginning of line><Prolog Atom><whitespace>(
5264 * or <beginning of line><Prolog Atom><whitespace>:-
5266 * It is added to the tags database if it doesn't match the
5267 * name of the previous clause header.
5269 * Return the size of the name of the predicate or rule, or 0 if no
5273 prolog_pr (char *s
, char *last
)
5275 /* Name of last clause. */
5280 pos
= prolog_atom (s
, 0);
5285 pos
= skip_spaces (s
+ pos
) - s
;
5288 || (s
[pos
] == '(' && (pos
+= 1))
5289 || (s
[pos
] == ':' && s
[pos
+ 1] == '-' && (pos
+= 2)))
5290 && (last
== NULL
/* save only the first clause */
5291 || len
!= strlen (last
)
5292 || !strneq (s
, last
, len
)))
5294 make_tag (s
, len
, true, s
, pos
, lineno
, linecharno
);
5302 * Consume a Prolog atom.
5303 * Return the number of bytes consumed, or 0 if there was an error.
5305 * A prolog atom, in this context, could be one of:
5306 * - An alphanumeric sequence, starting with a lower case letter.
5307 * - A quoted arbitrary string. Single quotes can escape themselves.
5308 * Backslash quotes everything.
5311 prolog_atom (char *s
, size_t pos
)
5317 if (ISLOWER (s
[pos
]) || (s
[pos
] == '_'))
5319 /* The atom is unquoted. */
5321 while (ISALNUM (s
[pos
]) || (s
[pos
] == '_'))
5325 return pos
- origpos
;
5327 else if (s
[pos
] == '\'')
5338 pos
++; /* A double quote */
5340 else if (s
[pos
] == '\0')
5341 /* Multiline quoted atoms are ignored. */
5343 else if (s
[pos
] == '\\')
5345 if (s
[pos
+1] == '\0')
5352 return pos
- origpos
;
5360 * Support for Erlang
5362 * Generates tags for functions, defines, and records.
5363 * Assumes that Erlang functions start at column 0.
5364 * Original code by Anders Lindgren (1996)
5366 static int erlang_func (char *, char *);
5367 static void erlang_attribute (char *);
5368 static int erlang_atom (char *);
5371 Erlang_functions (FILE *inf
)
5381 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
5383 if (cp
[0] == '\0') /* Empty line */
5385 else if (iswhite (cp
[0])) /* Not function nor attribute */
5387 else if (cp
[0] == '%') /* comment */
5389 else if (cp
[0] == '"') /* Sometimes, strings start in column one */
5391 else if (cp
[0] == '-') /* attribute, e.g. "-define" */
5393 erlang_attribute (cp
);
5400 else if ((len
= erlang_func (cp
, last
)) > 0)
5403 * Function. Store the function name so that we only
5404 * generates a tag for the first clause.
5407 last
= xnew (len
+ 1, char);
5408 else if (len
+ 1 > allocated
)
5409 xrnew (last
, len
+ 1, char);
5410 allocated
= len
+ 1;
5411 memcpy (last
, cp
, len
);
5420 * A function definition is added if it matches:
5421 * <beginning of line><Erlang Atom><whitespace>(
5423 * It is added to the tags database if it doesn't match the
5424 * name of the previous clause header.
5426 * Return the size of the name of the function, or 0 if no function
5430 erlang_func (char *s
, char *last
)
5432 /* Name of last clause. */
5437 pos
= erlang_atom (s
);
5442 pos
= skip_spaces (s
+ pos
) - s
;
5444 /* Save only the first clause. */
5447 || len
!= (int)strlen (last
)
5448 || !strneq (s
, last
, len
)))
5450 make_tag (s
, len
, true, s
, pos
, lineno
, linecharno
);
5459 * Handle attributes. Currently, tags are generated for defines
5462 * They are on the form:
5463 * -define(foo, bar).
5464 * -define(Foo(M, N), M+N).
5465 * -record(graph, {vtab = notable, cyclic = true}).
5468 erlang_attribute (char *s
)
5472 if ((LOOKING_AT (cp
, "-define") || LOOKING_AT (cp
, "-record"))
5475 int len
= erlang_atom (skip_spaces (cp
));
5477 make_tag (cp
, len
, true, s
, cp
+ len
- s
, lineno
, linecharno
);
5484 * Consume an Erlang atom (or variable).
5485 * Return the number of bytes consumed, or -1 if there was an error.
5488 erlang_atom (char *s
)
5492 if (ISALPHA (s
[pos
]) || s
[pos
] == '_')
5494 /* The atom is unquoted. */
5497 while (ISALNUM (s
[pos
]) || s
[pos
] == '_');
5499 else if (s
[pos
] == '\'')
5501 for (pos
++; s
[pos
] != '\''; pos
++)
5502 if (s
[pos
] == '\0' /* multiline quoted atoms are ignored */
5503 || (s
[pos
] == '\\' && s
[++pos
] == '\0'))
5512 static char *scan_separators (char *);
5513 static void add_regex (char *, language
*);
5514 static char *substitute (char *, char *, struct re_registers
*);
5517 * Take a string like "/blah/" and turn it into "blah", verifying
5518 * that the first and last characters are the same, and handling
5519 * quoted separator characters. Actually, stops on the occurrence of
5520 * an unquoted separator. Also process \t, \n, etc. and turn into
5521 * appropriate characters. Works in place. Null terminates name string.
5522 * Returns pointer to terminating separator, or NULL for
5523 * unterminated regexps.
5526 scan_separators (char *name
)
5529 char *copyto
= name
;
5530 bool quoted
= false;
5532 for (++name
; *name
!= '\0'; ++name
)
5538 case 'a': *copyto
++ = '\007'; break; /* BEL (bell) */
5539 case 'b': *copyto
++ = '\b'; break; /* BS (back space) */
5540 case 'd': *copyto
++ = 0177; break; /* DEL (delete) */
5541 case 'e': *copyto
++ = 033; break; /* ESC (delete) */
5542 case 'f': *copyto
++ = '\f'; break; /* FF (form feed) */
5543 case 'n': *copyto
++ = '\n'; break; /* NL (new line) */
5544 case 'r': *copyto
++ = '\r'; break; /* CR (carriage return) */
5545 case 't': *copyto
++ = '\t'; break; /* TAB (horizontal tab) */
5546 case 'v': *copyto
++ = '\v'; break; /* VT (vertical tab) */
5552 /* Something else is quoted, so preserve the quote. */
5560 else if (*name
== '\\')
5562 else if (*name
== sep
)
5568 name
= NULL
; /* signal unterminated regexp */
5570 /* Terminate copied string. */
5575 /* Look at the argument of --regex or --no-regex and do the right
5576 thing. Same for each line of a regexp file. */
5578 analyse_regex (char *regex_arg
)
5580 if (regex_arg
== NULL
)
5582 free_regexps (); /* --no-regex: remove existing regexps */
5586 /* A real --regexp option or a line in a regexp file. */
5587 switch (regex_arg
[0])
5589 /* Comments in regexp file or null arg to --regex. */
5595 /* Read a regex file. This is recursive and may result in a
5596 loop, which will stop when the file descriptors are exhausted. */
5600 linebuffer regexbuf
;
5601 char *regexfile
= regex_arg
+ 1;
5603 /* regexfile is a file containing regexps, one per line. */
5604 regexfp
= fopen (regexfile
, "rb");
5605 if (regexfp
== NULL
)
5607 linebuffer_init (®exbuf
);
5608 while (readline_internal (®exbuf
, regexfp
) > 0)
5609 analyse_regex (regexbuf
.buffer
);
5610 free (regexbuf
.buffer
);
5615 /* Regexp to be used for a specific language only. */
5619 char *lang_name
= regex_arg
+ 1;
5622 for (cp
= lang_name
; *cp
!= '}'; cp
++)
5625 error ("unterminated language name in regex: %s", regex_arg
);
5629 lang
= get_language_from_langname (lang_name
);
5632 add_regex (cp
, lang
);
5636 /* Regexp to be used for any language. */
5638 add_regex (regex_arg
, NULL
);
5643 /* Separate the regexp pattern, compile it,
5644 and care for optional name and modifiers. */
5646 add_regex (char *regexp_pattern
, language
*lang
)
5648 static struct re_pattern_buffer zeropattern
;
5649 char sep
, *pat
, *name
, *modifiers
;
5652 struct re_pattern_buffer
*patbuf
;
5655 force_explicit_name
= true, /* do not use implicit tag names */
5656 ignore_case
= false, /* case is significant */
5657 multi_line
= false, /* matches are done one line at a time */
5658 single_line
= false; /* dot does not match newline */
5661 if (strlen (regexp_pattern
) < 3)
5663 error ("null regexp");
5666 sep
= regexp_pattern
[0];
5667 name
= scan_separators (regexp_pattern
);
5670 error ("%s: unterminated regexp", regexp_pattern
);
5675 error ("null name for regexp \"%s\"", regexp_pattern
);
5678 modifiers
= scan_separators (name
);
5679 if (modifiers
== NULL
) /* no terminating separator --> no name */
5685 modifiers
+= 1; /* skip separator */
5687 /* Parse regex modifiers. */
5688 for (; modifiers
[0] != '\0'; modifiers
++)
5689 switch (modifiers
[0])
5692 if (modifiers
== name
)
5693 error ("forcing explicit tag name but no name, ignoring");
5694 force_explicit_name
= true;
5704 need_filebuf
= true;
5707 error ("invalid regexp modifier `%c', ignoring", modifiers
[0]);
5711 patbuf
= xnew (1, struct re_pattern_buffer
);
5712 *patbuf
= zeropattern
;
5715 static char lc_trans
[CHARS
];
5717 for (i
= 0; i
< CHARS
; i
++)
5718 lc_trans
[i
] = lowcase (i
);
5719 patbuf
->translate
= lc_trans
; /* translation table to fold case */
5723 pat
= concat ("^", regexp_pattern
, ""); /* anchor to beginning of line */
5725 pat
= regexp_pattern
;
5728 re_set_syntax (RE_SYNTAX_EMACS
| RE_DOT_NEWLINE
);
5730 re_set_syntax (RE_SYNTAX_EMACS
);
5732 err
= re_compile_pattern (pat
, strlen (pat
), patbuf
);
5737 error ("%s while compiling pattern", err
);
5742 p_head
= xnew (1, regexp
);
5743 p_head
->pattern
= savestr (regexp_pattern
);
5744 p_head
->p_next
= rp
;
5745 p_head
->lang
= lang
;
5746 p_head
->pat
= patbuf
;
5747 p_head
->name
= savestr (name
);
5748 p_head
->error_signaled
= false;
5749 p_head
->force_explicit_name
= force_explicit_name
;
5750 p_head
->ignore_case
= ignore_case
;
5751 p_head
->multi_line
= multi_line
;
5755 * Do the substitutions indicated by the regular expression and
5759 substitute (char *in
, char *out
, struct re_registers
*regs
)
5762 int size
, dig
, diglen
;
5765 size
= strlen (out
);
5767 /* Pass 1: figure out how much to allocate by finding all \N strings. */
5768 if (out
[size
- 1] == '\\')
5769 fatal ("pattern error in \"%s\"", out
);
5770 for (t
= etags_strchr (out
, '\\');
5772 t
= etags_strchr (t
+ 2, '\\'))
5776 diglen
= regs
->end
[dig
] - regs
->start
[dig
];
5782 /* Allocate space and do the substitutions. */
5784 result
= xnew (size
+ 1, char);
5786 for (t
= result
; *out
!= '\0'; out
++)
5787 if (*out
== '\\' && ISDIGIT (*++out
))
5790 diglen
= regs
->end
[dig
] - regs
->start
[dig
];
5791 memcpy (t
, in
+ regs
->start
[dig
], diglen
);
5798 assert (t
<= result
+ size
);
5799 assert (t
- result
== (int)strlen (result
));
5804 /* Deallocate all regexps. */
5809 while (p_head
!= NULL
)
5811 rp
= p_head
->p_next
;
5812 free (p_head
->pattern
);
5813 free (p_head
->name
);
5821 * Reads the whole file as a single string from `filebuf' and looks for
5822 * multi-line regular expressions, creating tags on matches.
5823 * readline already dealt with normal regexps.
5825 * Idea by Ben Wing <ben@666.com> (2002).
5828 regex_tag_multiline (void)
5830 char *buffer
= filebuf
.buffer
;
5834 for (rp
= p_head
; rp
!= NULL
; rp
= rp
->p_next
)
5838 if (!rp
->multi_line
)
5839 continue; /* skip normal regexps */
5841 /* Generic initializations before parsing file from memory. */
5842 lineno
= 1; /* reset global line number */
5843 charno
= 0; /* reset global char number */
5844 linecharno
= 0; /* reset global char number of line start */
5846 /* Only use generic regexps or those for the current language. */
5847 if (rp
->lang
!= NULL
&& rp
->lang
!= curfdp
->lang
)
5850 while (match
>= 0 && match
< filebuf
.len
)
5852 match
= re_search (rp
->pat
, buffer
, filebuf
.len
, charno
,
5853 filebuf
.len
- match
, &rp
->regs
);
5858 if (!rp
->error_signaled
)
5860 error ("regexp stack overflow while matching \"%s\"",
5862 rp
->error_signaled
= true;
5869 if (match
== rp
->regs
.end
[0])
5871 if (!rp
->error_signaled
)
5873 error ("regexp matches the empty string: \"%s\"",
5875 rp
->error_signaled
= true;
5877 match
= -3; /* exit from while loop */
5881 /* Match occurred. Construct a tag. */
5882 while (charno
< rp
->regs
.end
[0])
5883 if (buffer
[charno
++] == '\n')
5884 lineno
++, linecharno
= charno
;
5886 if (name
[0] == '\0')
5888 else /* make a named tag */
5889 name
= substitute (buffer
, rp
->name
, &rp
->regs
);
5890 if (rp
->force_explicit_name
)
5891 /* Force explicit tag name, if a name is there. */
5892 pfnote (name
, true, buffer
+ linecharno
,
5893 charno
- linecharno
+ 1, lineno
, linecharno
);
5895 make_tag (name
, strlen (name
), true, buffer
+ linecharno
,
5896 charno
- linecharno
+ 1, lineno
, linecharno
);
5905 nocase_tail (const char *cp
)
5907 register int len
= 0;
5909 while (*cp
!= '\0' && lowcase (*cp
) == lowcase (dbp
[len
]))
5911 if (*cp
== '\0' && !intoken (dbp
[len
]))
5920 get_tag (register char *bp
, char **namepp
)
5922 register char *cp
= bp
;
5926 /* Go till you get to white space or a syntactic break */
5927 for (cp
= bp
+ 1; !notinname (*cp
); cp
++)
5929 make_tag (bp
, cp
- bp
, true,
5930 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
5934 *namepp
= savenstr (bp
, cp
- bp
);
5938 * Read a line of text from `stream' into `lbp', excluding the
5939 * newline or CR-NL, if any. Return the number of characters read from
5940 * `stream', which is the length of the line including the newline.
5942 * On DOS or Windows we do not count the CR character, if any before the
5943 * NL, in the returned length; this mirrors the behavior of Emacs on those
5944 * platforms (for text files, it translates CR-NL to NL as it reads in the
5947 * If multi-line regular expressions are requested, each line read is
5948 * appended to `filebuf'.
5951 readline_internal (linebuffer
*lbp
, register FILE *stream
)
5953 char *buffer
= lbp
->buffer
;
5954 register char *p
= lbp
->buffer
;
5955 register char *pend
;
5958 pend
= p
+ lbp
->size
; /* Separate to avoid 386/IX compiler bug. */
5962 register int c
= getc (stream
);
5965 /* We're at the end of linebuffer: expand it. */
5967 xrnew (buffer
, lbp
->size
, char);
5968 p
+= buffer
- lbp
->buffer
;
5969 pend
= buffer
+ lbp
->size
;
5970 lbp
->buffer
= buffer
;
5980 if (p
> buffer
&& p
[-1] == '\r')
5984 /* Assume CRLF->LF translation will be performed by Emacs
5985 when loading this file, so CRs won't appear in the buffer.
5986 It would be cleaner to compensate within Emacs;
5987 however, Emacs does not know how many CRs were deleted
5988 before any given point in the file. */
6003 lbp
->len
= p
- buffer
;
6005 if (need_filebuf
/* we need filebuf for multi-line regexps */
6006 && chars_deleted
> 0) /* not at EOF */
6008 while (filebuf
.size
<= filebuf
.len
+ lbp
->len
+ 1) /* +1 for \n */
6010 /* Expand filebuf. */
6012 xrnew (filebuf
.buffer
, filebuf
.size
, char);
6014 memcpy (filebuf
.buffer
+ filebuf
.len
, lbp
->buffer
, lbp
->len
);
6015 filebuf
.len
+= lbp
->len
;
6016 filebuf
.buffer
[filebuf
.len
++] = '\n';
6017 filebuf
.buffer
[filebuf
.len
] = '\0';
6020 return lbp
->len
+ chars_deleted
;
6024 * Like readline_internal, above, but in addition try to match the
6025 * input line against relevant regular expressions and manage #line
6029 readline (linebuffer
*lbp
, FILE *stream
)
6033 linecharno
= charno
; /* update global char number of line start */
6034 result
= readline_internal (lbp
, stream
); /* read line */
6035 lineno
+= 1; /* increment global line number */
6036 charno
+= result
; /* increment global char number */
6038 /* Honor #line directives. */
6039 if (!no_line_directive
)
6041 static bool discard_until_line_directive
;
6043 /* Check whether this is a #line directive. */
6044 if (result
> 12 && strneq (lbp
->buffer
, "#line ", 6))
6049 if (sscanf (lbp
->buffer
, "#line %u \"%n", &lno
, &start
) >= 1
6050 && start
> 0) /* double quote character found */
6052 char *endp
= lbp
->buffer
+ start
;
6054 while ((endp
= etags_strchr (endp
, '"')) != NULL
6055 && endp
[-1] == '\\')
6058 /* Ok, this is a real #line directive. Let's deal with it. */
6060 char *taggedabsname
; /* absolute name of original file */
6061 char *taggedfname
; /* name of original file as given */
6062 char *name
; /* temp var */
6064 discard_until_line_directive
= false; /* found it */
6065 name
= lbp
->buffer
+ start
;
6067 canonicalize_filename (name
);
6068 taggedabsname
= absolute_filename (name
, tagfiledir
);
6069 if (filename_is_absolute (name
)
6070 || filename_is_absolute (curfdp
->infname
))
6071 taggedfname
= savestr (taggedabsname
);
6073 taggedfname
= relative_filename (taggedabsname
,tagfiledir
);
6075 if (streq (curfdp
->taggedfname
, taggedfname
))
6076 /* The #line directive is only a line number change. We
6077 deal with this afterwards. */
6080 /* The tags following this #line directive should be
6081 attributed to taggedfname. In order to do this, set
6082 curfdp accordingly. */
6084 fdesc
*fdp
; /* file description pointer */
6086 /* Go look for a file description already set up for the
6087 file indicated in the #line directive. If there is
6088 one, use it from now until the next #line
6090 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
6091 if (streq (fdp
->infname
, curfdp
->infname
)
6092 && streq (fdp
->taggedfname
, taggedfname
))
6093 /* If we remove the second test above (after the &&)
6094 then all entries pertaining to the same file are
6095 coalesced in the tags file. If we use it, then
6096 entries pertaining to the same file but generated
6097 from different files (via #line directives) will
6098 go into separate sections in the tags file. These
6099 alternatives look equivalent. The first one
6100 destroys some apparently useless information. */
6106 /* Else, if we already tagged the real file, skip all
6107 input lines until the next #line directive. */
6108 if (fdp
== NULL
) /* not found */
6109 for (fdp
= fdhead
; fdp
!= NULL
; fdp
= fdp
->next
)
6110 if (streq (fdp
->infabsname
, taggedabsname
))
6112 discard_until_line_directive
= true;
6116 /* Else create a new file description and use that from
6117 now on, until the next #line directive. */
6118 if (fdp
== NULL
) /* not found */
6121 fdhead
= xnew (1, fdesc
);
6122 *fdhead
= *curfdp
; /* copy curr. file description */
6124 fdhead
->infname
= savestr (curfdp
->infname
);
6125 fdhead
->infabsname
= savestr (curfdp
->infabsname
);
6126 fdhead
->infabsdir
= savestr (curfdp
->infabsdir
);
6127 fdhead
->taggedfname
= taggedfname
;
6128 fdhead
->usecharno
= false;
6129 fdhead
->prop
= NULL
;
6130 fdhead
->written
= false;
6134 free (taggedabsname
);
6136 readline (lbp
, stream
);
6138 } /* if a real #line directive */
6139 } /* if #line is followed by a number */
6140 } /* if line begins with "#line " */
6142 /* If we are here, no #line directive was found. */
6143 if (discard_until_line_directive
)
6147 /* Do a tail recursion on ourselves, thus discarding the contents
6148 of the line buffer. */
6149 readline (lbp
, stream
);
6153 discard_until_line_directive
= false;
6156 } /* if #line directives should be considered */
6163 /* Match against relevant regexps. */
6165 for (rp
= p_head
; rp
!= NULL
; rp
= rp
->p_next
)
6167 /* Only use generic regexps or those for the current language.
6168 Also do not use multiline regexps, which is the job of
6169 regex_tag_multiline. */
6170 if ((rp
->lang
!= NULL
&& rp
->lang
!= fdhead
->lang
)
6174 match
= re_match (rp
->pat
, lbp
->buffer
, lbp
->len
, 0, &rp
->regs
);
6179 if (!rp
->error_signaled
)
6181 error ("regexp stack overflow while matching \"%s\"",
6183 rp
->error_signaled
= true;
6190 /* Empty string matched. */
6191 if (!rp
->error_signaled
)
6193 error ("regexp matches the empty string: \"%s\"", rp
->pattern
);
6194 rp
->error_signaled
= true;
6198 /* Match occurred. Construct a tag. */
6200 if (name
[0] == '\0')
6202 else /* make a named tag */
6203 name
= substitute (lbp
->buffer
, rp
->name
, &rp
->regs
);
6204 if (rp
->force_explicit_name
)
6205 /* Force explicit tag name, if a name is there. */
6206 pfnote (name
, true, lbp
->buffer
, match
, lineno
, linecharno
);
6208 make_tag (name
, strlen (name
), true,
6209 lbp
->buffer
, match
, lineno
, linecharno
);
6218 * Return a pointer to a space of size strlen(cp)+1 allocated
6219 * with xnew where the string CP has been copied.
6222 savestr (const char *cp
)
6224 return savenstr (cp
, strlen (cp
));
6228 * Return a pointer to a space of size LEN+1 allocated with xnew where
6229 * the string CP has been copied for at most the first LEN characters.
6232 savenstr (const char *cp
, int len
)
6234 char *dp
= xnew (len
+ 1, char);
6236 return memcpy (dp
, cp
, len
);
6240 * Return the ptr in sp at which the character c last
6241 * appears; NULL if not found
6243 * Identical to POSIX strrchr, included for portability.
6246 etags_strrchr (register const char *sp
, register int c
)
6248 register const char *r
;
6260 * Return the ptr in sp at which the character c first
6261 * appears; NULL if not found
6263 * Identical to POSIX strchr, included for portability.
6266 etags_strchr (register const char *sp
, register int c
)
6276 /* Skip spaces (end of string is not space), return new pointer. */
6278 skip_spaces (char *cp
)
6280 while (iswhite (*cp
))
6285 /* Skip non spaces, except end of string, return new pointer. */
6287 skip_non_spaces (char *cp
)
6289 while (*cp
!= '\0' && !iswhite (*cp
))
6294 /* Skip any chars in the "name" class.*/
6296 skip_name (char *cp
)
6298 /* '\0' is a notinname() so loop stops there too */
6299 while (! notinname (*cp
))
6304 /* Print error message and exit. */
6306 fatal (const char *s1
, const char *s2
)
6309 exit (EXIT_FAILURE
);
6313 pfatal (const char *s1
)
6316 exit (EXIT_FAILURE
);
6320 suggest_asking_for_help (void)
6322 fprintf (stderr
, "\tTry `%s --help' for a complete list of options.\n",
6324 exit (EXIT_FAILURE
);
6327 /* Output a diagnostic with printf-style FORMAT and args. */
6329 error (const char *format
, ...)
6332 va_start (ap
, format
);
6333 fprintf (stderr
, "%s: ", progname
);
6334 vfprintf (stderr
, format
, ap
);
6335 fprintf (stderr
, "\n");
6339 /* Return a newly-allocated string whose contents
6340 concatenate those of s1, s2, s3. */
6342 concat (const char *s1
, const char *s2
, const char *s3
)
6344 int len1
= strlen (s1
), len2
= strlen (s2
), len3
= strlen (s3
);
6345 char *result
= xnew (len1
+ len2
+ len3
+ 1, char);
6347 strcpy (result
, s1
);
6348 strcpy (result
+ len1
, s2
);
6349 strcpy (result
+ len1
+ len2
, s3
);
6355 /* Does the same work as the system V getcwd, but does not need to
6356 guess the buffer size in advance. */
6361 char *path
= xnew (bufsize
, char);
6363 while (getcwd (path
, bufsize
) == NULL
)
6365 if (errno
!= ERANGE
)
6369 path
= xnew (bufsize
, char);
6372 canonicalize_filename (path
);
6376 /* Return a newly allocated string containing the file name of FILE
6377 relative to the absolute directory DIR (which should end with a slash). */
6379 relative_filename (char *file
, char *dir
)
6381 char *fp
, *dp
, *afn
, *res
;
6384 /* Find the common root of file and dir (with a trailing slash). */
6385 afn
= absolute_filename (file
, cwd
);
6388 while (*fp
++ == *dp
++)
6390 fp
--, dp
--; /* back to the first differing char */
6392 if (fp
== afn
&& afn
[0] != '/') /* cannot build a relative name */
6395 do /* look at the equal chars until '/' */
6399 /* Build a sequence of "../" strings for the resulting relative file name. */
6401 while ((dp
= etags_strchr (dp
+ 1, '/')) != NULL
)
6403 res
= xnew (3*i
+ strlen (fp
+ 1) + 1, char);
6406 strcat (res
, "../");
6408 /* Add the file name relative to the common root of file and dir. */
6409 strcat (res
, fp
+ 1);
6415 /* Return a newly allocated string containing the absolute file name
6416 of FILE given DIR (which should end with a slash). */
6418 absolute_filename (char *file
, char *dir
)
6420 char *slashp
, *cp
, *res
;
6422 if (filename_is_absolute (file
))
6423 res
= savestr (file
);
6425 /* We don't support non-absolute file names with a drive
6426 letter, like `d:NAME' (it's too much hassle). */
6427 else if (file
[1] == ':')
6428 fatal ("%s: relative file names with drive letters not supported", file
);
6431 res
= concat (dir
, file
, "");
6433 /* Delete the "/dirname/.." and "/." substrings. */
6434 slashp
= etags_strchr (res
, '/');
6435 while (slashp
!= NULL
&& slashp
[0] != '\0')
6437 if (slashp
[1] == '.')
6439 if (slashp
[2] == '.'
6440 && (slashp
[3] == '/' || slashp
[3] == '\0'))
6445 while (cp
>= res
&& !filename_is_absolute (cp
));
6447 cp
= slashp
; /* the absolute name begins with "/.." */
6449 /* Under MSDOS and NT we get `d:/NAME' as absolute
6450 file name, so the luser could say `d:/../NAME'.
6451 We silently treat this as `d:/NAME'. */
6452 else if (cp
[0] != '/')
6455 memmove (cp
, slashp
+ 3, strlen (slashp
+ 2));
6459 else if (slashp
[2] == '/' || slashp
[2] == '\0')
6461 memmove (slashp
, slashp
+ 2, strlen (slashp
+ 1));
6466 slashp
= etags_strchr (slashp
+ 1, '/');
6469 if (res
[0] == '\0') /* just a safety net: should never happen */
6472 return savestr ("/");
6478 /* Return a newly allocated string containing the absolute
6479 file name of dir where FILE resides given DIR (which should
6480 end with a slash). */
6482 absolute_dirname (char *file
, char *dir
)
6487 slashp
= etags_strrchr (file
, '/');
6489 return savestr (dir
);
6492 res
= absolute_filename (file
, dir
);
6498 /* Whether the argument string is an absolute file name. The argument
6499 string must have been canonicalized with canonicalize_filename. */
6501 filename_is_absolute (char *fn
)
6503 return (fn
[0] == '/'
6505 || (ISALPHA (fn
[0]) && fn
[1] == ':' && fn
[2] == '/')
6510 /* Downcase DOS drive letter and collapse separators into single slashes.
6513 canonicalize_filename (register char *fn
)
6519 /* Canonicalize drive letter case. */
6520 # define ISUPPER(c) isupper (CHAR (c))
6521 if (fn
[0] != '\0' && fn
[1] == ':' && ISUPPER (fn
[0]))
6522 fn
[0] = lowcase (fn
[0]);
6527 /* Collapse multiple separators into a single slash. */
6528 for (cp
= fn
; *cp
!= '\0'; cp
++, fn
++)
6532 while (cp
[1] == sep
)
6541 /* Initialize a linebuffer for use. */
6543 linebuffer_init (linebuffer
*lbp
)
6545 lbp
->size
= (DEBUG
) ? 3 : 200;
6546 lbp
->buffer
= xnew (lbp
->size
, char);
6547 lbp
->buffer
[0] = '\0';
6551 /* Set the minimum size of a string contained in a linebuffer. */
6553 linebuffer_setlen (linebuffer
*lbp
, int toksize
)
6555 while (lbp
->size
<= toksize
)
6558 xrnew (lbp
->buffer
, lbp
->size
, char);
6563 /* Like malloc but get fatal error if memory is exhausted. */
6565 xmalloc (size_t size
)
6567 void *result
= malloc (size
);
6569 fatal ("virtual memory exhausted", (char *)NULL
);
6574 xrealloc (char *ptr
, size_t size
)
6576 void *result
= realloc (ptr
, size
);
6578 fatal ("virtual memory exhausted", (char *)NULL
);
6584 * indent-tabs-mode: t
6587 * c-font-lock-extra-types: ("FILE" "bool" "language" "linebuffer" "fdesc" "node" "regexp")
6588 * c-file-style: "gnu"
6592 /* etags.c ends here */