1 /* Tags file maker to go with GNU Emacs
2 Copyright (C) 1984, 87, 88, 89, 93, 94, 95, 98, 99
3 Free Software Foundation, Inc. and Ken Arnold
5 This file is not considered part of GNU Emacs.
7 This program is free software; you can redistribute it and/or modify
8 it under the terms of the GNU General Public License as published by
9 the Free Software Foundation; either version 2 of the License, or
10 (at your option) any later version.
12 This program is distributed in the hope that it will be useful,
13 but WITHOUT ANY WARRANTY; without even the implied warranty of
14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 GNU General Public License for more details.
17 You should have received a copy of the GNU General Public License
18 along with this program; if not, write to the Free Software Foundation,
19 Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
23 * Ctags originally by Ken Arnold.
24 * Fortran added by Jim Kleckner.
25 * Ed Pelegri-Llopart added C typedefs.
26 * Gnu Emacs TAGS format and modifications by RMS?
27 * Sam Kendall added C++.
28 * Francesco Potorti` reorganised C and C++ based on work by Joe Wells.
29 * Regexp tags by Tom Tromey.
31 * Francesco Potorti` (pot@gnu.org) is the current maintainer.
34 char pot_etags_version
[] = "@(#) pot revision number is 13.33";
40 # define _GNU_SOURCE /* enables some compiler checks on GNU */
48 /* On some systems, Emacs defines static as nothing for the sake
49 of unexec. We don't want that here since we don't use unexec. */
51 # define ETAGS_REGEXPS /* use the regexp features */
52 # define LONG_OPTIONS /* accept long options */
53 #endif /* HAVE_CONFIG_H */
57 # include <sys/param.h>
59 # ifndef HAVE_CONFIG_H
61 # include <sys/config.h>
70 # define MAXPATHLEN _MAX_PATH
75 # endif /* not HAVE_CONFIG_H */
78 # endif /* undef HAVE_GETCWD */
79 #endif /* WINDOWSNT */
81 #if !defined (WINDOWSNT) && defined (STDC_HEADERS)
90 extern char *getcwd ();
92 #endif /* HAVE_UNISTD_H */
100 #include <sys/types.h>
101 #include <sys/stat.h>
103 #if !defined (S_ISREG) && defined (S_IFREG)
104 # define S_ISREG(m) (((m) & S_IFMT) == S_IFREG)
110 # define getopt_long(argc,argv,optstr,lopts,lind) getopt (argc, argv, optstr)
112 extern int optind
, opterr
;
113 #endif /* LONG_OPTIONS */
117 #endif /* ETAGS_REGEXPS */
119 /* Define CTAGS to make the program "ctags" compatible with the usual one.
120 Leave it undefined to make the program "etags", which makes emacs-style
121 tag tables and tags typedefs, #defines and struct/union/enum by default. */
129 /* Exit codes for success and failure. */
139 #define C_PLPL 0x00001 /* C++ */
140 #define C_STAR 0x00003 /* C* */
141 #define C_JAVA 0x00005 /* JAVA */
142 #define YACC 0x10000 /* yacc file */
144 #define streq(s,t) ((DEBUG && (s) == NULL && (t) == NULL \
145 && (abort (), 1)) || !strcmp (s, t))
146 #define strneq(s,t,n) ((DEBUG && (s) == NULL && (t) == NULL \
147 && (abort (), 1)) || !strncmp (s, t, n))
149 #define lowcase(c) tolower ((char)c)
151 #define CHARS 256 /* 2^sizeof(char) */
152 #define CHAR(x) ((unsigned int)x & (CHARS - 1))
153 #define iswhite(c) (_wht[CHAR(c)]) /* c is white */
154 #define notinname(c) (_nin[CHAR(c)]) /* c is not in a name */
155 #define begtoken(c) (_btk[CHAR(c)]) /* c can start token */
156 #define intoken(c) (_itk[CHAR(c)]) /* c can be in token */
157 #define endtoken(c) (_etk[CHAR(c)]) /* c ends tokens */
161 * xnew, xrnew -- allocate, reallocate storage
163 * SYNOPSIS: Type *xnew (int n, Type);
164 * Type *xrnew (OldPointer, int n, Type);
167 # include "chkmalloc.h"
168 # define xnew(n,Type) ((Type *) trace_malloc (__FILE__, __LINE__, \
169 (n) * sizeof (Type)))
170 # define xrnew(op,n,Type) ((Type *) trace_realloc (__FILE__, __LINE__, \
171 (op), (n) * sizeof (Type)))
173 # define xnew(n,Type) ((Type *) xmalloc ((n) * sizeof (Type)))
174 # define xrnew(op,n,Type) ((Type *) xrealloc ((op), (n) * sizeof (Type)))
179 typedef void Lang_function ();
184 char *command
; /* Takes one arg and decompresses to stdout */
190 Lang_function
*function
;
195 extern char *getenv ();
197 /* Many compilers barf on this:
198 Lang_function Ada_funcs;
199 so let's write it this way */
203 void default_C_entries ();
204 void plain_C_entries ();
205 void Cjava_entries ();
206 void Cobol_paragraphs ();
207 void Cplusplus_entries ();
208 void Cstar_entries ();
209 void Erlang_functions ();
210 void Fortran_functions ();
211 void Yacc_entries ();
212 void Lisp_functions ();
213 void Pascal_functions ();
214 void Perl_functions ();
215 void Postscript_functions ();
216 void Prolog_functions ();
217 void Python_functions ();
218 void Scheme_functions ();
219 void TeX_functions ();
220 void just_read_file ();
222 compressor
*get_compressor_from_suffix ();
223 language
*get_language_from_name ();
224 language
*get_language_from_interpreter ();
225 language
*get_language_from_suffix ();
226 int total_size_of_entries ();
227 long readline (), readline_internal ();
231 void analyse_regex ();
233 void free_patterns ();
234 #endif /* ETAGS_REGEXPS */
236 void suggest_asking_for_help ();
237 void fatal (), pfatal ();
242 void find_entries ();
244 void pfnote (), new_pfnote ();
245 void process_file ();
250 char *skip_spaces (), *skip_non_spaces ();
251 char *savenstr (), *savestr ();
252 char *etags_strchr (), *etags_strrchr ();
253 char *etags_getcwd ();
254 char *relative_filename (), *absolute_filename (), *absolute_dirname ();
255 bool filename_is_absolute ();
256 void canonicalize_filename ();
257 void grow_linebuffer ();
258 long *xmalloc (), *xrealloc ();
261 char searchar
= '/'; /* use /.../ searches */
263 char *tagfile
; /* output file */
264 char *progname
; /* name this program was invoked with */
265 char *cwd
; /* current working directory */
266 char *tagfiledir
; /* directory of tagfile */
267 FILE *tagf
; /* ioptr for tags file */
269 char *curfile
; /* current input file name */
270 language
*curlang
; /* current language */
272 int lineno
; /* line number of current line */
273 long charno
; /* current character number */
274 long linecharno
; /* charno of start of current line */
275 char *dbp
; /* pointer to start of current tag */
277 typedef struct node_st
278 { /* sorting structure */
279 char *name
; /* function or type name */
280 char *file
; /* file name */
281 bool is_func
; /* use pattern or line no */
282 bool been_warned
; /* set if noticed dup */
283 int lno
; /* line number tag is on */
284 long cno
; /* character number line starts on */
285 char *pat
; /* search pattern */
286 struct node_st
*left
, *right
; /* left and right sons */
289 node
*head
; /* the head of the binary tree of tags */
292 * A `linebuffer' is a structure which holds a line of text.
293 * `readline_internal' reads a line from a stream into a linebuffer
294 * and works regardless of the length of the line.
295 * SIZE is the size of BUFFER, LEN is the length of the string in
296 * BUFFER after readline reads it.
305 linebuffer lb
; /* the current line */
306 linebuffer token_name
; /* used by C_entries as a temporary area */
310 linebuffer lb
; /* used by C_entries instead of lb */
313 /* boolean "functions" (see init) */
314 bool _wht
[CHARS
], _nin
[CHARS
], _itk
[CHARS
], _btk
[CHARS
], _etk
[CHARS
];
317 *white
= " \f\t\n\r",
319 *nonam
= " \f\t\n\r(=,[;",
320 /* token ending chars */
321 *endtk
= " \t\n\r\"'#()[]{}=-+%*/&|^~!<>;,.:?",
322 /* token starting chars */
323 *begtk
= "ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz$~@",
324 /* valid in-token chars */
325 *midtk
= "ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz$0123456789";
327 bool append_to_tagfile
; /* -a: append to tags */
328 /* The following four default to TRUE for etags, but to FALSE for ctags. */
329 bool typedefs
; /* -t: create tags for C and Ada typedefs */
330 bool typedefs_and_cplusplus
; /* -T: create tags for C typedefs, level */
331 /* 0 struct/enum/union decls, and C++ */
332 /* member functions. */
333 bool constantypedefs
; /* -d: create tags for C #define, enum */
334 /* constants and variables. */
335 /* -D: opposite of -d. Default under ctags. */
336 bool declarations
; /* --declarations: tag them and extern in C&Co*/
337 bool globals
; /* create tags for global variables */
338 bool members
; /* create tags for C member variables */
339 bool update
; /* -u: update tags */
340 bool vgrind_style
; /* -v: create vgrind style index output */
341 bool no_warnings
; /* -w: suppress warnings */
342 bool cxref_style
; /* -x: create cxref style output */
343 bool cplusplus
; /* .[hc] means C++, not C */
344 bool noindentypedefs
; /* -I: ignore indentation in C */
345 bool packages_only
; /* --packages-only: in Ada, only tag packages*/
348 struct option longopts
[] =
350 { "packages-only", no_argument
, &packages_only
, TRUE
},
351 { "append", no_argument
, NULL
, 'a' },
352 { "backward-search", no_argument
, NULL
, 'B' },
353 { "c++", no_argument
, NULL
, 'C' },
354 { "cxref", no_argument
, NULL
, 'x' },
355 { "defines", no_argument
, NULL
, 'd' },
356 { "declarations", no_argument
, &declarations
, TRUE
},
357 { "no-defines", no_argument
, NULL
, 'D' },
358 { "globals", no_argument
, &globals
, TRUE
},
359 { "no-globals", no_argument
, &globals
, FALSE
},
360 { "help", no_argument
, NULL
, 'h' },
361 { "help", no_argument
, NULL
, 'H' },
362 { "ignore-indentation", no_argument
, NULL
, 'I' },
363 { "include", required_argument
, NULL
, 'i' },
364 { "language", required_argument
, NULL
, 'l' },
365 { "members", no_argument
, &members
, TRUE
},
366 { "no-members", no_argument
, &members
, FALSE
},
367 { "no-warn", no_argument
, NULL
, 'w' },
368 { "output", required_argument
, NULL
, 'o' },
370 { "regex", required_argument
, NULL
, 'r' },
371 { "no-regex", no_argument
, NULL
, 'R' },
372 { "ignore-case-regex", required_argument
, NULL
, 'c' },
373 #endif /* ETAGS_REGEXPS */
374 { "typedefs", no_argument
, NULL
, 't' },
375 { "typedefs-and-c++", no_argument
, NULL
, 'T' },
376 { "update", no_argument
, NULL
, 'u' },
377 { "version", no_argument
, NULL
, 'V' },
378 { "vgrind", no_argument
, NULL
, 'v' },
381 #endif /* LONG_OPTIONS */
384 /* Structure defining a regular expression. Elements are
385 the compiled pattern, and the name string. */
386 typedef struct pattern
388 struct pattern
*p_next
;
391 struct re_pattern_buffer
*pattern
;
392 struct re_registers regs
;
397 /* List of all regexps. */
398 pattern
*p_head
= NULL
;
400 /* How many characters in the character set. (From regex.c.) */
401 #define CHAR_SET_SIZE 256
402 /* Translation table for case-insensitive matching. */
403 char lc_trans
[CHAR_SET_SIZE
];
404 #endif /* ETAGS_REGEXPS */
406 compressor compressors
[] =
408 { "z", "gzip -d -c"},
409 { "Z", "gzip -d -c"},
410 { "gz", "gzip -d -c"},
411 { "GZ", "gzip -d -c"},
412 { "bz2", "bzip2 -d -c" },
420 /* Non-NULL if language fixed. */
421 language
*forced_lang
= NULL
;
424 char *Ada_suffixes
[] =
425 { "ads", "adb", "ada", NULL
};
428 char *Asm_suffixes
[] = { "a", /* Unix assembler */
429 "asm", /* Microcontroller assembly */
430 "def", /* BSO/Tasking definition includes */
431 "inc", /* Microcontroller include files */
432 "ins", /* Microcontroller include files */
433 "s", "sa", /* Unix assembler */
434 "S", /* cpp-processed Unix assembler */
435 "src", /* BSO/Tasking C compiler output */
439 /* Note that .c and .h can be considered C++, if the --c++ flag was
440 given. That is why default_C_entries is called here. */
441 char *default_C_suffixes
[] =
444 char *Cplusplus_suffixes
[] =
445 { "C", "H", "c++", "cc", "cpp", "cxx", "h++", "hh", "hpp", "hxx",
446 "M", /* Objective C++ */
447 "pdb", /* Postscript with C syntax */
450 char *Cjava_suffixes
[] =
453 char *Cobol_suffixes
[] =
454 { "COB", "cob", NULL
};
456 char *Cstar_suffixes
[] =
457 { "cs", "hs", NULL
};
459 char *Erlang_suffixes
[] =
460 { "erl", "hrl", NULL
};
462 char *Fortran_suffixes
[] =
463 { "F", "f", "f90", "for", NULL
};
465 char *Lisp_suffixes
[] =
466 { "cl", "clisp", "el", "l", "lisp", "lsp", "ml", NULL
};
468 char *Pascal_suffixes
[] =
469 { "p", "pas", NULL
};
471 char *Perl_suffixes
[] =
472 { "pl", "pm", NULL
};
473 char *Perl_interpreters
[] =
474 { "perl", "@PERL@", NULL
};
476 char *plain_C_suffixes
[] =
477 { "pc", /* Pro*C file */
478 "m", /* Objective C file */
479 "lm", /* Objective lex file */
482 char *Postscript_suffixes
[] =
483 { "ps", "psw", NULL
}; /* .psw is for PSWrap */
485 char *Prolog_suffixes
[] =
488 char *Python_suffixes
[] =
491 /* Can't do the `SCM' or `scm' prefix with a version number. */
492 char *Scheme_suffixes
[] =
493 { "SCM", "SM", "oak", "sch", "scheme", "scm", "sm", "ss", "t", NULL
};
495 char *TeX_suffixes
[] =
496 { "TeX", "bib", "clo", "cls", "ltx", "sty", "tex", NULL
};
498 char *Yacc_suffixes
[] =
499 { "y", "ym", "yy", "yxx", "y++", NULL
}; /* .ym is Objective yacc file */
502 * Table of languages.
504 * It is ok for a given function to be listed under more than one
505 * name. I just didn't.
508 language lang_names
[] =
510 { "ada", Ada_funcs
, Ada_suffixes
, NULL
},
511 { "asm", Asm_labels
, Asm_suffixes
, NULL
},
512 { "c", default_C_entries
, default_C_suffixes
, NULL
},
513 { "c++", Cplusplus_entries
, Cplusplus_suffixes
, NULL
},
514 { "c*", Cstar_entries
, Cstar_suffixes
, NULL
},
515 { "cobol", Cobol_paragraphs
, Cobol_suffixes
, NULL
},
516 { "erlang", Erlang_functions
, Erlang_suffixes
, NULL
},
517 { "fortran", Fortran_functions
, Fortran_suffixes
, NULL
},
518 { "java", Cjava_entries
, Cjava_suffixes
, NULL
},
519 { "lisp", Lisp_functions
, Lisp_suffixes
, NULL
},
520 { "pascal", Pascal_functions
, Pascal_suffixes
, NULL
},
521 { "perl", Perl_functions
, Perl_suffixes
, Perl_interpreters
},
522 { "postscript", Postscript_functions
, Postscript_suffixes
, NULL
},
523 { "proc", plain_C_entries
, plain_C_suffixes
, NULL
},
524 { "prolog", Prolog_functions
, Prolog_suffixes
, NULL
},
525 { "python", Python_functions
, Python_suffixes
, NULL
},
526 { "scheme", Scheme_functions
, Scheme_suffixes
, NULL
},
527 { "tex", TeX_functions
, TeX_suffixes
, NULL
},
528 { "yacc", Yacc_entries
, Yacc_suffixes
, NULL
},
529 { "auto", NULL
}, /* default guessing scheme */
530 { "none", just_read_file
}, /* regexp matching only */
531 { NULL
, NULL
} /* end of list */
535 print_language_names ()
540 puts ("\nThese are the currently supported languages, along with the\n\
541 default file name suffixes:");
542 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
544 printf ("\t%s\t", lang
->name
);
545 if (lang
->suffixes
!= NULL
)
546 for (ext
= lang
->suffixes
; *ext
!= NULL
; ext
++)
547 printf (" .%s", *ext
);
550 puts ("Where `auto' means use default language for files based on file\n\
551 name suffix, and `none' means only do regexp processing on files.\n\
552 If no language is specified and no matching suffix is found,\n\
553 the first line of the file is read for a sharp-bang (#!) sequence\n\
554 followed by the name of an interpreter. If no such sequence is found,\n\
555 Fortran is tried first; if no tags are found, C is tried next.\n\
556 Compressed files are supported using gzip and bzip2.");
560 # define VERSION "20"
565 printf ("%s (GNU Emacs %s)\n", (CTAGS
) ? "ctags" : "etags", VERSION
);
566 puts ("Copyright (C) 1999 Free Software Foundation, Inc. and Ken Arnold");
567 puts ("This program is distributed under the same terms as Emacs");
575 printf ("Usage: %s [options] [[regex-option ...] file-name] ...\n\
577 These are the options accepted by %s.\n", progname
, progname
);
579 puts ("You may use unambiguous abbreviations for the long option names.");
581 puts ("Long option names do not work with this executable, as it is not\n\
582 linked with GNU getopt.");
583 #endif /* LONG_OPTIONS */
584 puts ("A - as file name means read names from stdin (one per line).");
586 printf (" Absolute names are stored in the output file as they are.\n\
587 Relative ones are stored relative to the output file's directory.");
590 puts ("-a, --append\n\
591 Append tag entries to existing tags file.");
593 puts ("--packages-only\n\
594 For Ada files, only generate tags for packages .");
597 puts ("-B, --backward-search\n\
598 Write the search commands for the tag entries using '?', the\n\
599 backward-search command instead of '/', the forward-search command.");
602 Treat files whose name suffix defaults to C language as C++ files.");
604 puts ("--declarations\n\
605 In C and derived languages, create tags for function declarations,");
607 puts ("\tand create tags for extern variables if --globals is used.");
610 ("\tand create tags for extern variables unless --no-globals is used.");
613 puts ("-d, --defines\n\
614 Create tag entries for C #define constants and enum constants, too.");
616 puts ("-D, --no-defines\n\
617 Don't create tag entries for C #define constants and enum constants.\n\
618 This makes the tags file smaller.");
622 puts ("-i FILE, --include=FILE\n\
623 Include a note in tag file indicating that, when searching for\n\
624 a tag, one should also consult the tags file FILE after\n\
625 checking the current file.");
626 puts ("-l LANG, --language=LANG\n\
627 Force the following files to be considered as written in the\n\
628 named language up to the next --language=LANG option.");
633 Create tag entries for global variables in some languages.");
635 puts ("--no-globals\n\
636 Do not create tag entries for global variables in some\n\
637 languages. This makes the tags file smaller.");
639 Create tag entries for member variables in C and derived languages.");
642 puts ("-r /REGEXP/, --regex=/REGEXP/ or --regex=@regexfile\n\
643 Make a tag for each line matching pattern REGEXP in the following\n\
644 files. {LANGUAGE}/REGEXP/ uses REGEXP for LANGUAGE files only.\n\
645 regexfile is a file containing one REGEXP per line.\n\
646 REGEXP is anchored (as if preceded by ^).\n\
647 The form /REGEXP/NAME/ creates a named tag.\n\
648 For example Tcl named tags can be created with:\n\
649 --regex=/proc[ \\t]+\\([^ \\t]+\\)/\\1/.");
650 puts ("-c /REGEXP/, --ignore-case-regex=/REGEXP/ or --ignore-case-regex=@regexfile\n\
651 Like -r, --regex but ignore case when matching expressions.");
652 puts ("-R, --no-regex\n\
653 Don't create tags from regexps for the following files.");
654 #endif /* ETAGS_REGEXPS */
655 puts ("-o FILE, --output=FILE\n\
656 Write the tags to FILE.");
657 puts ("-I, --ignore-indentation\n\
658 Don't rely on indentation quite as much as normal. Currently,\n\
659 this means not to assume that a closing brace in the first\n\
660 column is the final brace of a function or structure\n\
661 definition in C and C++.");
665 puts ("-t, --typedefs\n\
666 Generate tag entries for C and Ada typedefs.");
667 puts ("-T, --typedefs-and-c++\n\
668 Generate tag entries for C typedefs, C struct/enum/union tags,\n\
669 and C++ member functions.");
670 puts ("-u, --update\n\
671 Update the tag entries for the given files, leaving tag\n\
672 entries for other files in place. Currently, this is\n\
673 implemented by deleting the existing entries for the given\n\
674 files and then rewriting the new entries at the end of the\n\
675 tags file. It is often faster to simply rebuild the entire\n\
676 tag file than to use this.");
677 puts ("-v, --vgrind\n\
678 Generates an index of items intended for human consumption,\n\
679 similar to the output of vgrind. The index is sorted, and\n\
680 gives the page number of each item.");
681 puts ("-w, --no-warn\n\
682 Suppress warning messages about entries defined in multiple\n\
684 puts ("-x, --cxref\n\
685 Like --vgrind, but in the style of cxref, rather than vgrind.\n\
686 The output uses line numbers instead of page numbers, but\n\
687 beyond that the differences are cosmetic; try both to see\n\
691 puts ("-V, --version\n\
692 Print the version of the program.\n\
694 Print this help message.");
696 print_language_names ();
699 puts ("Report bugs to bug-gnu-emacs@gnu.org");
713 /* This structure helps us allow mixing of --lang and file names. */
716 enum argument_type arg_type
;
718 language
*lang
; /* language of the regexp */
721 #ifdef VMS /* VMS specific functions */
725 /* This is a BUG! ANY arbitrary limit is a BUG!
726 Won't someone please fix this? */
727 #define MAX_FILE_SPEC_LEN 255
730 char body
[MAX_FILE_SPEC_LEN
+ 1];
734 v1.05 nmm 26-Jun-86 fn_exp - expand specification of list of file names
735 returning in each successive call the next file name matching the input
736 spec. The function expects that each in_spec passed
737 to it will be processed to completion; in particular, up to and
738 including the call following that in which the last matching name
739 is returned, the function ignores the value of in_spec, and will
740 only start processing a new spec with the following call.
741 If an error occurs, on return out_spec contains the value
742 of in_spec when the error occurred.
744 With each successive file name returned in out_spec, the
745 function's return value is one. When there are no more matching
746 names the function returns zero. If on the first call no file
747 matches in_spec, or there is any other error, -1 is returned.
752 #define OUTSIZE MAX_FILE_SPEC_LEN
758 static long context
= 0;
759 static struct dsc$descriptor_s o
;
760 static struct dsc$descriptor_s i
;
761 static bool pass1
= TRUE
;
768 o
.dsc$a_pointer
= (char *) out
;
769 o
.dsc$w_length
= (short)OUTSIZE
;
770 i
.dsc$a_pointer
= in
;
771 i
.dsc$w_length
= (short)strlen(in
);
772 i
.dsc$b_dtype
= DSC$K_DTYPE_T
;
773 i
.dsc$b_class
= DSC$K_CLASS_S
;
774 o
.dsc$b_dtype
= DSC$K_DTYPE_VT
;
775 o
.dsc$b_class
= DSC$K_CLASS_VS
;
777 if ((status
= lib$
find_file(&i
, &o
, &context
, 0, 0)) == RMS$_NORMAL
)
779 out
->body
[out
->curlen
] = EOS
;
782 else if (status
== RMS$_NMF
)
786 strcpy(out
->body
, in
);
789 lib$
find_file_end(&context
);
795 v1.01 nmm 19-Aug-85 gfnames - return in successive calls the
796 name of each file specified by the provided arg expanding wildcards.
799 gfnames (arg
, p_error
)
803 static vspec filename
= {MAX_FILE_SPEC_LEN
, "\0"};
805 switch (fn_exp (&filename
, arg
))
809 return filename
.body
;
815 return filename
.body
;
819 #ifndef OLD /* Newer versions of VMS do provide `system'. */
823 error ("%s", "system() function not implemented under VMS");
827 #define VERSION_DELIM ';'
828 char *massage_name (s
)
834 if (*s
== VERSION_DELIM
)
852 unsigned int nincluded_files
;
853 char **included_files
;
856 int current_arg
, file_count
;
857 linebuffer filename_lb
;
863 _fmode
= O_BINARY
; /* all of files are treated as binary files */
868 included_files
= xnew (argc
, char *);
872 /* Allocate enough no matter what happens. Overkill, but each one
874 argbuffer
= xnew (argc
, argument
);
877 /* Set syntax for regular expression routines. */
878 re_set_syntax (RE_SYNTAX_EMACS
| RE_INTERVALS
);
879 /* Translation table for case-insensitive search. */
880 for (i
= 0; i
< CHAR_SET_SIZE
; i
++)
881 lc_trans
[i
] = lowcase (i
);
882 #endif /* ETAGS_REGEXPS */
885 * If etags, always find typedefs and structure tags. Why not?
886 * Also default is to find macro constants, enum constants and
891 typedefs
= typedefs_and_cplusplus
= constantypedefs
= TRUE
;
902 optstring
= "-aCdDf:Il:o:r:c:RStTi:BuvxwVhH";
904 optstring
= "-aCdDf:Il:o:StTi:BuvxwVhH";
905 #endif /* ETAGS_REGEXPS */
908 optstring
= optstring
+ 1;
909 #endif /* LONG_OPTIONS */
911 opt
= getopt_long (argc
, argv
, optstring
, longopts
, 0);
918 /* If getopt returns 0, then it has already processed a
919 long-named option. We should do nothing. */
923 /* This means that a file name has been seen. Record it. */
924 argbuffer
[current_arg
].arg_type
= at_filename
;
925 argbuffer
[current_arg
].what
= optarg
;
930 /* Common options. */
931 case 'a': append_to_tagfile
= TRUE
; break;
932 case 'C': cplusplus
= TRUE
; break;
933 case 'd': constantypedefs
= TRUE
; break;
934 case 'D': constantypedefs
= FALSE
; break;
935 case 'f': /* for compatibility with old makefiles */
939 error ("-o option may only be given once.", (char *)NULL
);
940 suggest_asking_for_help ();
945 case 'S': /* for backward compatibility */
946 noindentypedefs
= TRUE
;
950 language
*lang
= get_language_from_name (optarg
);
953 argbuffer
[current_arg
].lang
= lang
;
954 argbuffer
[current_arg
].arg_type
= at_language
;
961 argbuffer
[current_arg
].arg_type
= at_regexp
;
962 argbuffer
[current_arg
].what
= optarg
;
966 argbuffer
[current_arg
].arg_type
= at_regexp
;
967 argbuffer
[current_arg
].what
= NULL
;
971 argbuffer
[current_arg
].arg_type
= at_icregexp
;
972 argbuffer
[current_arg
].what
= optarg
;
975 #endif /* ETAGS_REGEXPS */
987 typedefs
= typedefs_and_cplusplus
= TRUE
;
992 included_files
[nincluded_files
++] = optarg
;
996 case 'B': searchar
= '?'; break;
997 case 'u': update
= TRUE
; break;
998 case 'v': vgrind_style
= TRUE
; /*FALLTHRU*/
999 case 'x': cxref_style
= TRUE
; break;
1000 case 'w': no_warnings
= TRUE
; break;
1003 suggest_asking_for_help ();
1007 for (; optind
< argc
; ++optind
)
1009 argbuffer
[current_arg
].arg_type
= at_filename
;
1010 argbuffer
[current_arg
].what
= argv
[optind
];
1015 if (nincluded_files
== 0 && file_count
== 0)
1017 error ("no input files specified.", (char *)NULL
);
1018 suggest_asking_for_help ();
1021 if (tagfile
== NULL
)
1022 tagfile
= CTAGS
? "tags" : "TAGS";
1023 cwd
= etags_getcwd (); /* the current working directory */
1024 if (cwd
[strlen (cwd
) - 1] != '/')
1027 cwd
= concat (oldcwd
, "/", "");
1030 if (streq (tagfile
, "-"))
1033 tagfiledir
= absolute_dirname (tagfile
, cwd
);
1035 init (); /* set up boolean "functions" */
1038 initbuffer (&token_name
);
1039 initbuffer (&lbs
[0].lb
);
1040 initbuffer (&lbs
[1].lb
);
1041 initbuffer (&filename_lb
);
1045 if (streq (tagfile
, "-"))
1049 /* Switch redirected `stdout' to binary mode (setting `_fmode'
1050 doesn't take effect until after `stdout' is already open). */
1051 if (!isatty (fileno (stdout
)))
1052 setmode (fileno (stdout
), O_BINARY
);
1056 tagf
= fopen (tagfile
, append_to_tagfile
? "a" : "w");
1062 * Loop through files finding functions.
1064 for (i
= 0; i
< current_arg
; ++i
)
1066 switch (argbuffer
[i
].arg_type
)
1069 forced_lang
= argbuffer
[i
].lang
;
1071 #ifdef ETAGS_REGEXPS
1073 analyse_regex (argbuffer
[i
].what
, FALSE
);
1076 analyse_regex (argbuffer
[i
].what
, TRUE
);
1081 while ((this_file
= gfnames (argbuffer
[i
].what
, &got_err
)) != NULL
)
1085 error ("can't find file %s\n", this_file
);
1090 this_file
= massage_name (this_file
);
1093 this_file
= argbuffer
[i
].what
;
1095 /* Input file named "-" means read file names from stdin
1096 (one per line) and use them. */
1097 if (streq (this_file
, "-"))
1098 while (readline_internal (&filename_lb
, stdin
) > 0)
1099 process_file (filename_lb
.buffer
);
1101 process_file (this_file
);
1109 #ifdef ETAGS_REGEXPS
1111 #endif /* ETAGS_REGEXPS */
1115 while (nincluded_files
-- > 0)
1116 fprintf (tagf
, "\f\n%s,include\n", *included_files
++);
1122 /* If CTAGS, we are here. process_file did not write the tags yet,
1123 because we want them ordered. Let's do it now. */
1135 for (i
= 0; i
< current_arg
; ++i
)
1137 if (argbuffer
[i
].arg_type
!= at_filename
)
1140 "mv %s OTAGS;fgrep -v '\t%s\t' OTAGS >%s;rm OTAGS",
1141 tagfile
, argbuffer
[i
].what
, tagfile
);
1142 if (system (cmd
) != GOOD
)
1143 fatal ("failed to execute shell command", (char *)NULL
);
1145 append_to_tagfile
= TRUE
;
1148 tagf
= fopen (tagfile
, append_to_tagfile
? "a" : "w");
1159 sprintf (cmd
, "sort %s -o %s", tagfile
, tagfile
);
1160 exit (system (cmd
));
1168 * Return a compressor given the file name. If EXTPTR is non-zero,
1169 * return a pointer into FILE where the compressor-specific
1170 * extension begins. If no compressor is found, NULL is returned
1171 * and EXTPTR is not significant.
1172 * Idea by Vladimir Alexiev <vladimir@cs.ualberta.ca>
1175 get_compressor_from_suffix (file
, extptr
)
1180 char *slash
, *suffix
;
1182 /* This relies on FN to be after canonicalize_filename,
1183 so we don't need to consider backslashes on DOS_NT. */
1184 slash
= etags_strrchr (file
, '/');
1185 suffix
= etags_strrchr (file
, '.');
1186 if (suffix
== NULL
|| suffix
< slash
)
1191 /* Let those poor souls who live with DOS 8+3 file name limits get
1192 some solace by treating foo.cgz as if it were foo.c.gz, etc.
1193 Only the first do loop is run if not MSDOS */
1196 for (compr
= compressors
; compr
->suffix
!= NULL
; compr
++)
1197 if (streq (compr
->suffix
, suffix
))
1204 } while (*suffix
!= '\0');
1211 * Return a language given the name.
1214 get_language_from_name (name
)
1220 error ("empty language name", (char *)NULL
);
1223 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1224 if (streq (name
, lang
->name
))
1226 error ("unknown language \"%s\"", name
);
1234 * Return a language given the interpreter name.
1237 get_language_from_interpreter (interpreter
)
1243 if (interpreter
== NULL
)
1245 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1246 if (lang
->interpreters
!= NULL
)
1247 for (iname
= lang
->interpreters
; *iname
!= NULL
; iname
++)
1248 if (streq (*iname
, interpreter
))
1257 * Return a language given the file name.
1260 get_language_from_suffix (file
)
1264 char **ext
, *suffix
;
1266 suffix
= etags_strrchr (file
, '.');
1270 for (lang
= lang_names
; lang
->name
!= NULL
; lang
++)
1271 if (lang
->suffixes
!= NULL
)
1272 for (ext
= lang
->suffixes
; *ext
!= NULL
; ext
++)
1273 if (streq (*ext
, suffix
))
1281 * This routine is called on each file argument.
1287 struct stat stat_buf
;
1290 char *compressed_name
, *uncompressed_name
;
1291 char *ext
, *real_name
;
1294 canonicalize_filename (file
);
1295 if (streq (file
, tagfile
) && !streq (tagfile
, "-"))
1297 error ("skipping inclusion of %s in self.", file
);
1300 if ((compr
= get_compressor_from_suffix (file
, &ext
)) == NULL
)
1302 compressed_name
= NULL
;
1303 real_name
= uncompressed_name
= savestr (file
);
1307 real_name
= compressed_name
= savestr (file
);
1308 uncompressed_name
= savenstr (file
, ext
- file
);
1311 /* If the canonicalised uncompressed name has already be dealt with,
1312 skip it silently, else add it to the list. */
1314 typedef struct processed_file
1317 struct processed_file
*next
;
1319 static processed_file
*pf_head
= NULL
;
1320 register processed_file
*fnp
;
1322 for (fnp
= pf_head
; fnp
!= NULL
; fnp
= fnp
->next
)
1323 if (streq (uncompressed_name
, fnp
->filename
))
1326 pf_head
= xnew (1, struct processed_file
);
1327 pf_head
->filename
= savestr (uncompressed_name
);
1328 pf_head
->next
= fnp
;
1331 if (stat (real_name
, &stat_buf
) != 0)
1333 /* Reset real_name and try with a different name. */
1335 if (compressed_name
!= NULL
) /* try with the given suffix */
1337 if (stat (uncompressed_name
, &stat_buf
) == 0)
1338 real_name
= uncompressed_name
;
1340 else /* try all possible suffixes */
1342 for (compr
= compressors
; compr
->suffix
!= NULL
; compr
++)
1344 compressed_name
= concat (file
, ".", compr
->suffix
);
1345 if (stat (compressed_name
, &stat_buf
) != 0)
1348 char *suf
= compressed_name
+ strlen (file
);
1349 size_t suflen
= strlen (compr
->suffix
) + 1;
1350 for ( ; suf
[1]; suf
++, suflen
--)
1352 memmove (suf
, suf
+ 1, suflen
);
1353 if (stat (compressed_name
, &stat_buf
) == 0)
1355 real_name
= compressed_name
;
1359 if (real_name
!= NULL
)
1362 free (compressed_name
);
1363 compressed_name
= NULL
;
1367 real_name
= compressed_name
;
1372 if (real_name
== NULL
)
1377 } /* try with a different name */
1379 if (!S_ISREG (stat_buf
.st_mode
))
1381 error ("skipping %s: it is not a regular file.", real_name
);
1384 if (real_name
== compressed_name
)
1386 char *cmd
= concat (compr
->command
, " ", real_name
);
1387 inf
= popen (cmd
, "r");
1391 inf
= fopen (real_name
, "r");
1398 find_entries (uncompressed_name
, inf
);
1400 if (real_name
== compressed_name
)
1409 if (filename_is_absolute (uncompressed_name
))
1411 /* file is an absolute file name. Canonicalise it. */
1412 filename
= absolute_filename (uncompressed_name
, cwd
);
1416 /* file is a file name relative to cwd. Make it relative
1417 to the directory of the tags file. */
1418 filename
= relative_filename (uncompressed_name
, tagfiledir
);
1420 fprintf (tagf
, "\f\n%s,%d\n", filename
, total_size_of_entries (head
));
1428 if (compressed_name
) free(compressed_name
);
1429 if (uncompressed_name
) free(uncompressed_name
);
1434 * This routine sets up the boolean pseudo-functions which work
1435 * by setting boolean flags dependent upon the corresponding character.
1436 * Every char which is NOT in that string is not a white char. Therefore,
1437 * all of the array "_wht" is set to FALSE, and then the elements
1438 * subscripted by the chars in "white" are set to TRUE. Thus "_wht"
1439 * of a char is TRUE if it is the string "white", else FALSE.
1447 for (i
= 0; i
< CHARS
; i
++)
1448 iswhite(i
) = notinname(i
) = begtoken(i
) = intoken(i
) = endtoken(i
) = FALSE
;
1449 for (sp
= white
; *sp
!= '\0'; sp
++) iswhite (*sp
) = TRUE
;
1450 for (sp
= nonam
; *sp
!= '\0'; sp
++) notinname (*sp
) = TRUE
;
1451 for (sp
= begtk
; *sp
!= '\0'; sp
++) begtoken (*sp
) = TRUE
;
1452 for (sp
= midtk
; *sp
!= '\0'; sp
++) intoken (*sp
) = TRUE
;
1453 for (sp
= endtk
; *sp
!= '\0'; sp
++) endtoken (*sp
) = TRUE
;
1454 iswhite('\0') = iswhite('\n');
1455 notinname('\0') = notinname('\n');
1456 begtoken('\0') = begtoken('\n');
1457 intoken('\0') = intoken('\n');
1458 endtoken('\0') = endtoken('\n');
1462 * This routine opens the specified file and calls the function
1463 * which finds the function and type definitions.
1465 node
*last_node
= NULL
;
1468 find_entries (file
, inf
)
1474 node
*old_last_node
;
1476 /* Memory leakage here: the string pointed by curfile is
1477 never released, because curfile is copied into np->file
1478 for each node, to be used in CTAGS mode. The amount of
1479 memory leaked here is the sum of the lengths of the
1481 curfile
= savestr (file
);
1483 /* If user specified a language, use it. */
1485 if (lang
!= NULL
&& lang
->function
!= NULL
)
1488 lang
->function (inf
);
1492 /* Try to guess the language given the file name. */
1493 lang
= get_language_from_suffix (file
);
1494 if (lang
!= NULL
&& lang
->function
!= NULL
)
1497 lang
->function (inf
);
1501 /* Look for sharp-bang as the first two characters. */
1502 if (readline_internal (&lb
, inf
) > 0
1504 && lb
.buffer
[0] == '#'
1505 && lb
.buffer
[1] == '!')
1509 /* Set lp to point at the first char after the last slash in the
1510 line or, if no slashes, at the first nonblank. Then set cp to
1511 the first successive blank and terminate the string. */
1512 lp
= etags_strrchr (lb
.buffer
+2, '/');
1516 lp
= skip_spaces (lb
.buffer
+ 2);
1517 cp
= skip_non_spaces (lp
);
1520 if (strlen (lp
) > 0)
1522 lang
= get_language_from_interpreter (lp
);
1523 if (lang
!= NULL
&& lang
->function
!= NULL
)
1526 lang
->function (inf
);
1531 /* We rewind here, even if inf may be a pipe. We fail if the
1532 length of the first line is longer than the pipe block size,
1533 which is unlikely. */
1537 old_last_node
= last_node
;
1538 curlang
= get_language_from_name ("fortran");
1539 Fortran_functions (inf
);
1541 /* No Fortran entries found. Try C. */
1542 if (old_last_node
== last_node
)
1544 /* We do not tag if rewind fails.
1545 Only the file name will be recorded in the tags file. */
1547 curlang
= get_language_from_name (cplusplus
? "c++" : "c");
1548 default_C_entries (inf
);
1555 pfnote (name
, is_func
, linestart
, linelen
, lno
, cno
)
1556 char *name
; /* tag name, or NULL if unnamed */
1557 bool is_func
; /* tag is a function */
1558 char *linestart
; /* start of the line where tag is */
1559 int linelen
; /* length of the line where tag is */
1560 int lno
; /* line number */
1561 long cno
; /* character number */
1565 if (CTAGS
&& name
== NULL
)
1568 np
= xnew (1, node
);
1570 /* If ctags mode, change name "main" to M<thisfilename>. */
1571 if (CTAGS
&& !cxref_style
&& streq (name
, "main"))
1573 register char *fp
= etags_strrchr (curfile
, '/');
1574 np
->name
= concat ("M", fp
== NULL
? curfile
: fp
+ 1, "");
1575 fp
= etags_strrchr (np
->name
, '.');
1576 if (fp
!= NULL
&& fp
[1] != '\0' && fp
[2] == '\0')
1581 np
->been_warned
= FALSE
;
1583 np
->is_func
= is_func
;
1585 /* Our char numbers are 0-base, because of C language tradition?
1586 ctags compatibility? old versions compatibility? I don't know.
1587 Anyway, since emacs's are 1-base we expect etags.el to take care
1588 of the difference. If we wanted to have 1-based numbers, we would
1589 uncomment the +1 below. */
1590 np
->cno
= cno
/* + 1 */ ;
1591 np
->left
= np
->right
= NULL
;
1592 if (CTAGS
&& !cxref_style
)
1594 if (strlen (linestart
) < 50)
1595 np
->pat
= concat (linestart
, "$", "");
1597 np
->pat
= savenstr (linestart
, 50);
1600 np
->pat
= savenstr (linestart
, linelen
);
1602 add_node (np
, &head
);
1605 /* Date: Wed, 22 Jan 1997 02:56:31 -0500 [last amended 18 Sep 1997]
1606 * From: Sam Kendall <kendall@mv.mv.com>
1607 * Subject: Proposal for firming up the TAGS format specification
1608 * To: F.Potorti@cnuce.cnr.it
1610 * pfnote should emit the optimized form [unnamed tag] only if:
1611 * 1. name does not contain any of the characters " \t\r\n(),;";
1612 * 2. linestart contains name as either a rightmost, or rightmost but
1613 * one character, substring;
1614 * 3. the character, if any, immediately before name in linestart must
1615 * be one of the characters " \t(),;";
1616 * 4. the character, if any, immediately after name in linestart must
1617 * also be one of the characters " \t(),;".
1619 * The real implementation uses the notinname() macro, which recognises
1620 * characters slightly different form " \t\r\n(),;". See the variable
1623 #define traditional_tag_style TRUE
1625 new_pfnote (name
, namelen
, is_func
, linestart
, linelen
, lno
, cno
)
1626 char *name
; /* tag name, or NULL if unnamed */
1627 int namelen
; /* tag length */
1628 bool is_func
; /* tag is a function */
1629 char *linestart
; /* start of the line where tag is */
1630 int linelen
; /* length of the line where tag is */
1631 int lno
; /* line number */
1632 long cno
; /* character number */
1640 for (cp
= name
; !notinname (*cp
); cp
++)
1642 if (*cp
== '\0') /* rule #1 */
1644 cp
= linestart
+ linelen
- namelen
;
1645 if (notinname (linestart
[linelen
-1]))
1646 cp
-= 1; /* rule #4 */
1647 if (cp
>= linestart
/* rule #2 */
1649 || notinname (cp
[-1])) /* rule #3 */
1650 && strneq (name
, cp
, namelen
)) /* rule #2 */
1651 named
= FALSE
; /* use unnamed tag */
1656 name
= savenstr (name
, namelen
);
1659 pfnote (name
, is_func
, linestart
, linelen
, lno
, cno
);
1664 * recurse on left children, iterate on right children.
1672 register node
*node_right
= np
->right
;
1673 free_tree (np
->left
);
1674 if (np
->name
!= NULL
)
1684 * Adds a node to the tree of nodes. In etags mode, we don't keep
1685 * it sorted; we just keep a linear list. In ctags mode, maintain
1686 * an ordered tree, with no attempt at balancing.
1688 * add_node is the only function allowed to add nodes, so it can
1692 add_node (np
, cur_node_p
)
1693 node
*np
, **cur_node_p
;
1696 register node
*cur_node
= *cur_node_p
;
1698 if (cur_node
== NULL
)
1708 if (last_node
== NULL
)
1709 fatal ("internal error in add_node", (char *)NULL
);
1710 last_node
->right
= np
;
1716 dif
= strcmp (np
->name
, cur_node
->name
);
1719 * If this tag name matches an existing one, then
1720 * do not add the node, but maybe print a warning.
1724 if (streq (np
->file
, cur_node
->file
))
1728 fprintf (stderr
, "Duplicate entry in file %s, line %d: %s\n",
1729 np
->file
, lineno
, np
->name
);
1730 fprintf (stderr
, "Second entry ignored\n");
1733 else if (!cur_node
->been_warned
&& !no_warnings
)
1737 "Duplicate entry in files %s and %s: %s (Warning only)\n",
1738 np
->file
, cur_node
->file
, np
->name
);
1739 cur_node
->been_warned
= TRUE
;
1744 /* Actually add the node */
1745 add_node (np
, dif
< 0 ? &cur_node
->left
: &cur_node
->right
);
1758 /* Output subentries that precede this one */
1759 put_entries (np
->left
);
1761 /* Output this entry */
1765 if (np
->name
!= NULL
)
1766 fprintf (tagf
, "%s\177%s\001%d,%ld\n",
1767 np
->pat
, np
->name
, np
->lno
, np
->cno
);
1769 fprintf (tagf
, "%s\177%d,%ld\n",
1770 np
->pat
, np
->lno
, np
->cno
);
1774 if (np
->name
== NULL
)
1775 error ("internal error: NULL name in ctags mode.", (char *)NULL
);
1780 fprintf (stdout
, "%s %s %d\n",
1781 np
->name
, np
->file
, (np
->lno
+ 63) / 64);
1783 fprintf (stdout
, "%-16s %3d %-16s %s\n",
1784 np
->name
, np
->lno
, np
->file
, np
->pat
);
1788 fprintf (tagf
, "%s\t%s\t", np
->name
, np
->file
);
1792 putc (searchar
, tagf
);
1795 for (sp
= np
->pat
; *sp
; sp
++)
1797 if (*sp
== '\\' || *sp
== searchar
)
1801 putc (searchar
, tagf
);
1804 { /* a typedef; text pattern inadequate */
1805 fprintf (tagf
, "%d", np
->lno
);
1811 /* Output subentries that follow this one */
1812 put_entries (np
->right
);
1815 /* Length of a number's decimal representation. */
1821 while ((num
/= 10) > 0)
1827 * Return total number of characters that put_entries will output for
1828 * the nodes in the subtree of the specified node. Works only if
1829 * we are not ctags, but called only in that case. This count
1830 * is irrelevant with the new tags.el, but is still supplied for
1831 * backward compatibility.
1834 total_size_of_entries (np
)
1842 for (total
= 0; np
!= NULL
; np
= np
->right
)
1844 /* Count left subentries. */
1845 total
+= total_size_of_entries (np
->left
);
1847 /* Count this entry */
1848 total
+= strlen (np
->pat
) + 1;
1849 total
+= number_len ((long) np
->lno
) + 1 + number_len (np
->cno
) + 1;
1850 if (np
->name
!= NULL
)
1851 total
+= 1 + strlen (np
->name
); /* \001name */
1858 * The C symbol tables.
1863 st_C_objprot
, st_C_objimpl
, st_C_objend
,
1868 st_C_struct
, st_C_extern
, st_C_enum
, st_C_define
, st_C_typedef
, st_C_typespec
1871 /* Feed stuff between (but not including) %[ and %] lines to:
1872 gperf -c -k 1,3 -o -p -r -t
1874 struct C_stab_entry { char *name; int c_ext; enum sym_type type; }
1878 while, 0, st_C_ignore
1879 switch, 0, st_C_ignore
1880 return, 0, st_C_ignore
1881 @interface, 0, st_C_objprot
1882 @protocol, 0, st_C_objprot
1883 @implementation,0, st_C_objimpl
1884 @end, 0, st_C_objend
1885 import, C_JAVA, st_C_ignore
1886 package, C_JAVA, st_C_ignore
1887 friend, C_PLPL, st_C_ignore
1888 extends, C_JAVA, st_C_javastruct
1889 implements, C_JAVA, st_C_javastruct
1890 interface, C_JAVA, st_C_struct
1891 class, C_PLPL, st_C_struct
1892 namespace, C_PLPL, st_C_struct
1893 domain, C_STAR, st_C_struct
1894 union, 0, st_C_struct
1895 struct, 0, st_C_struct
1896 extern, 0, st_C_extern
1898 typedef, 0, st_C_typedef
1899 define, 0, st_C_define
1900 operator, C_PLPL, st_C_operator
1901 bool, C_PLPL, st_C_typespec
1902 long, 0, st_C_typespec
1903 short, 0, st_C_typespec
1904 int, 0, st_C_typespec
1905 char, 0, st_C_typespec
1906 float, 0, st_C_typespec
1907 double, 0, st_C_typespec
1908 signed, 0, st_C_typespec
1909 unsigned, 0, st_C_typespec
1910 auto, 0, st_C_typespec
1911 void, 0, st_C_typespec
1912 static, 0, st_C_typespec
1913 const, 0, st_C_typespec
1914 volatile, 0, st_C_typespec
1915 explicit, C_PLPL, st_C_typespec
1916 mutable, C_PLPL, st_C_typespec
1917 typename, C_PLPL, st_C_typespec
1918 # DEFUN used in emacs, the next three used in glibc (SYSCALL only for mach).
1919 DEFUN, 0, st_C_gnumacro
1920 SYSCALL, 0, st_C_gnumacro
1921 ENTRY, 0, st_C_gnumacro
1922 PSEUDO, 0, st_C_gnumacro
1923 # These are defined inside C functions, so currently they are not met.
1924 # EXFUN used in glibc, DEFVAR_* in emacs.
1925 #EXFUN, 0, st_C_gnumacro
1926 #DEFVAR_, 0, st_C_gnumacro
1928 and replace lines between %< and %> with its output. */
1930 /* C code produced by gperf version 2.7.1 (19981006 egcs) */
1931 /* Command-line: gperf -c -k 1,3 -o -p -r -t */
1932 struct C_stab_entry
{ char *name
; int c_ext
; enum sym_type type
; };
1934 #define TOTAL_KEYWORDS 46
1935 #define MIN_WORD_LENGTH 2
1936 #define MAX_WORD_LENGTH 15
1937 #define MIN_HASH_VALUE 13
1938 #define MAX_HASH_VALUE 123
1939 /* maximum key range = 111, duplicates = 0 */
1946 register const char *str
;
1947 register unsigned int len
;
1949 static unsigned char asso_values
[] =
1951 124, 124, 124, 124, 124, 124, 124, 124, 124, 124,
1952 124, 124, 124, 124, 124, 124, 124, 124, 124, 124,
1953 124, 124, 124, 124, 124, 124, 124, 124, 124, 124,
1954 124, 124, 124, 124, 124, 124, 124, 124, 124, 124,
1955 124, 124, 124, 124, 124, 124, 124, 124, 124, 124,
1956 124, 124, 124, 124, 124, 124, 124, 124, 124, 124,
1957 124, 124, 124, 124, 3, 124, 124, 124, 43, 6,
1958 11, 124, 124, 124, 124, 124, 124, 124, 124, 124,
1959 11, 124, 124, 58, 7, 124, 124, 124, 124, 124,
1960 124, 124, 124, 124, 124, 124, 124, 57, 7, 42,
1961 4, 14, 52, 0, 124, 53, 124, 124, 29, 11,
1962 6, 35, 32, 124, 29, 34, 59, 58, 51, 24,
1963 124, 124, 124, 124, 124, 124, 124, 124, 124, 124,
1964 124, 124, 124, 124, 124, 124, 124, 124, 124, 124,
1965 124, 124, 124, 124, 124, 124, 124, 124, 124, 124,
1966 124, 124, 124, 124, 124, 124, 124, 124, 124, 124,
1967 124, 124, 124, 124, 124, 124, 124, 124, 124, 124,
1968 124, 124, 124, 124, 124, 124, 124, 124, 124, 124,
1969 124, 124, 124, 124, 124, 124, 124, 124, 124, 124,
1970 124, 124, 124, 124, 124, 124, 124, 124, 124, 124,
1971 124, 124, 124, 124, 124, 124, 124, 124, 124, 124,
1972 124, 124, 124, 124, 124, 124, 124, 124, 124, 124,
1973 124, 124, 124, 124, 124, 124, 124, 124, 124, 124,
1974 124, 124, 124, 124, 124, 124, 124, 124, 124, 124,
1975 124, 124, 124, 124, 124, 124, 124, 124, 124, 124,
1976 124, 124, 124, 124, 124, 124
1978 register int hval
= len
;
1984 hval
+= asso_values
[(unsigned char)str
[2]];
1987 hval
+= asso_values
[(unsigned char)str
[0]];
1996 struct C_stab_entry
*
1997 in_word_set (str
, len
)
1998 register const char *str
;
1999 register unsigned int len
;
2001 static struct C_stab_entry wordlist
[] =
2003 {""}, {""}, {""}, {""}, {""}, {""}, {""}, {""}, {""},
2004 {""}, {""}, {""}, {""},
2005 {"@end", 0, st_C_objend
},
2006 {""}, {""}, {""}, {""},
2007 {"ENTRY", 0, st_C_gnumacro
},
2008 {"@interface", 0, st_C_objprot
},
2010 {"domain", C_STAR
, st_C_struct
},
2012 {"PSEUDO", 0, st_C_gnumacro
},
2014 {"namespace", C_PLPL
, st_C_struct
},
2016 {"@implementation",0, st_C_objimpl
},
2017 {""}, {""}, {""}, {""}, {""}, {""}, {""}, {""}, {""},
2018 {"long", 0, st_C_typespec
},
2019 {"signed", 0, st_C_typespec
},
2020 {"@protocol", 0, st_C_objprot
},
2021 {""}, {""}, {""}, {""},
2022 {"bool", C_PLPL
, st_C_typespec
},
2023 {""}, {""}, {""}, {""}, {""}, {""},
2024 {"const", 0, st_C_typespec
},
2025 {"explicit", C_PLPL
, st_C_typespec
},
2026 {"if", 0, st_C_ignore
},
2028 {"operator", C_PLPL
, st_C_operator
},
2030 {"DEFUN", 0, st_C_gnumacro
},
2032 {"define", 0, st_C_define
},
2033 {""}, {""}, {""}, {""}, {""},
2034 {"double", 0, st_C_typespec
},
2035 {"struct", 0, st_C_struct
},
2036 {""}, {""}, {""}, {""},
2037 {"short", 0, st_C_typespec
},
2039 {"enum", 0, st_C_enum
},
2040 {"mutable", C_PLPL
, st_C_typespec
},
2042 {"extern", 0, st_C_extern
},
2043 {"extends", C_JAVA
, st_C_javastruct
},
2044 {"package", C_JAVA
, st_C_ignore
},
2045 {"while", 0, st_C_ignore
},
2047 {"for", 0, st_C_ignore
},
2049 {"volatile", 0, st_C_typespec
},
2051 {"import", C_JAVA
, st_C_ignore
},
2052 {"float", 0, st_C_typespec
},
2053 {"switch", 0, st_C_ignore
},
2054 {"return", 0, st_C_ignore
},
2055 {"implements", C_JAVA
, st_C_javastruct
},
2057 {"static", 0, st_C_typespec
},
2058 {"typedef", 0, st_C_typedef
},
2059 {"typename", C_PLPL
, st_C_typespec
},
2060 {"unsigned", 0, st_C_typespec
},
2062 {"char", 0, st_C_typespec
},
2063 {"class", C_PLPL
, st_C_struct
},
2065 {"void", 0, st_C_typespec
},
2067 {"friend", C_PLPL
, st_C_ignore
},
2069 {"int", 0, st_C_typespec
},
2070 {"union", 0, st_C_struct
},
2072 {"auto", 0, st_C_typespec
},
2073 {"interface", C_JAVA
, st_C_struct
},
2075 {"SYSCALL", 0, st_C_gnumacro
}
2078 if (len
<= MAX_WORD_LENGTH
&& len
>= MIN_WORD_LENGTH
)
2080 register int key
= hash (str
, len
);
2082 if (key
<= MAX_HASH_VALUE
&& key
>= 0)
2084 register const char *s
= wordlist
[key
].name
;
2086 if (*str
== *s
&& !strncmp (str
+ 1, s
+ 1, len
- 1))
2087 return &wordlist
[key
];
2095 C_symtype (str
, len
, c_ext
)
2100 register struct C_stab_entry
*se
= in_word_set (str
, len
);
2102 if (se
== NULL
|| (se
->c_ext
&& !(c_ext
& se
->c_ext
)))
2108 * C functions and variables are recognized using a simple
2109 * finite automaton. fvdef is its state variable.
2113 fvnone
, /* nothing seen */
2114 foperator
, /* func: operator keyword seen (cplpl) */
2115 fvnameseen
, /* function or variable name seen */
2116 fstartlist
, /* func: just after open parenthesis */
2117 finlist
, /* func: in parameter list */
2118 flistseen
, /* func: after parameter list */
2119 fignore
, /* func: before open brace */
2120 vignore
/* var-like: ignore until ';' */
2123 bool fvextern
; /* func or var: extern keyword seen; */
2126 * typedefs are recognized using a simple finite automaton.
2127 * typdef is its state variable.
2131 tnone
, /* nothing seen */
2132 tkeyseen
, /* typedef keyword seen */
2133 ttypeseen
, /* defined type seen */
2134 tinbody
, /* inside typedef body */
2135 tend
, /* just before typedef tag */
2136 tignore
/* junk after typedef tag */
2141 * struct-like structures (enum, struct and union) are recognized
2142 * using another simple finite automaton. `structdef' is its state
2147 snone
, /* nothing seen yet */
2148 skeyseen
, /* struct-like keyword seen */
2149 stagseen
, /* struct-like tag seen */
2150 scolonseen
, /* colon seen after struct-like tag */
2151 sinbody
/* in struct body: recognize member func defs*/
2155 * When structdef is stagseen, scolonseen, or sinbody, structtag is the
2156 * struct tag, and structtype is the type of the preceding struct-like
2159 char *structtag
= "<uninited>";
2160 enum sym_type structtype
;
2163 * When objdef is different from onone, objtag is the name of the class.
2165 char *objtag
= "<uninited>";
2168 * Yet another little state machine to deal with preprocessor lines.
2172 dnone
, /* nothing seen */
2173 dsharpseen
, /* '#' seen as first char on line */
2174 ddefineseen
, /* '#' and 'define' seen */
2175 dignorerest
/* ignore rest of line */
2179 * State machine for Objective C protocols and implementations.
2180 * Tom R.Hageman <tom@basil.icce.rug.nl>
2184 onone
, /* nothing seen */
2185 oprotocol
, /* @interface or @protocol seen */
2186 oimplementation
, /* @implementations seen */
2187 otagseen
, /* class name seen */
2188 oparenseen
, /* parenthesis before category seen */
2189 ocatseen
, /* category name seen */
2190 oinbody
, /* in @implementation body */
2191 omethodsign
, /* in @implementation body, after +/- */
2192 omethodtag
, /* after method name */
2193 omethodcolon
, /* after method colon */
2194 omethodparm
, /* after method parameter */
2195 oignore
/* wait for @end */
2200 * Use this structure to keep info about the token read, and how it
2201 * should be tagged. Used by the make_C_tag function to build a tag.
2214 token tok
; /* latest token read */
2217 * Set this to TRUE, and the next token considered is called a function.
2218 * Used only for GNU emacs's function-defining macros.
2220 bool next_token_is_func
;
2223 * TRUE in the rules part of a yacc file, FALSE outside (parse as C).
2228 * methodlen is the length of the method name stored in token_name.
2234 * checks to see if the current token is at the start of a
2235 * function or variable, or corresponds to a typedef, or
2236 * is a struct/union/enum tag, or #define, or an enum constant.
2238 * *IS_FUNC gets TRUE iff the token is a function or #define macro
2239 * with args. C_EXT is which language we are looking at.
2247 * next_token_is_func IN OUT
2251 consider_token (str
, len
, c
, c_ext
, cblev
, parlev
, is_func_or_var
)
2252 register char *str
; /* IN: token pointer */
2253 register int len
; /* IN: token length */
2254 register char c
; /* IN: first char after the token */
2255 int c_ext
; /* IN: C extensions mask */
2256 int cblev
; /* IN: curly brace level */
2257 int parlev
; /* IN: parenthesis level */
2258 bool *is_func_or_var
; /* OUT: function or variable found */
2260 enum sym_type toktype
= C_symtype (str
, len
, c_ext
);
2263 * Advance the definedef state machine.
2268 /* We're not on a preprocessor line. */
2271 if (toktype
== st_C_define
)
2273 definedef
= ddefineseen
;
2277 definedef
= dignorerest
;
2282 * Make a tag for any macro, unless it is a constant
2283 * and constantypedefs is FALSE.
2285 definedef
= dignorerest
;
2286 *is_func_or_var
= (c
== '(');
2287 if (!*is_func_or_var
&& !constantypedefs
)
2294 error ("internal error: definedef value.", (char *)NULL
);
2303 if (toktype
== st_C_typedef
)
2322 /* Do not return here, so the structdef stuff has a chance. */
2336 * This structdef business is currently only invoked when cblev==0.
2337 * It should be recursively invoked whatever the curly brace level,
2338 * and a stack of states kept, to allow for definitions of structs
2341 * This structdef business is NOT invoked when we are ctags and the
2342 * file is plain C. This is because a struct tag may have the same
2343 * name as another tag, and this loses with ctags.
2347 case st_C_javastruct
:
2348 if (structdef
== stagseen
)
2349 structdef
= scolonseen
;
2353 if (typdef
== tkeyseen
2354 || (typedefs_and_cplusplus
&& cblev
== 0 && structdef
== snone
))
2356 structdef
= skeyseen
;
2357 structtype
= toktype
;
2362 if (structdef
== skeyseen
)
2364 /* Save the tag for struct/union/class, for functions and variables
2365 that may be defined inside. */
2366 if (structtype
== st_C_struct
)
2367 structtag
= savenstr (str
, len
);
2369 structtag
= "<enum>";
2370 structdef
= stagseen
;
2374 if (typdef
!= tnone
)
2377 /* Detect GNU macros.
2379 Writers of emacs code are recommended to put the
2380 first two args of a DEFUN on the same line.
2382 The DEFUN macro, used in emacs C source code, has a first arg
2383 that is a string (the lisp function name), and a second arg that
2384 is a C function name. Since etags skips strings, the second arg
2385 is tagged. This is unfortunate, as it would be better to tag the
2386 first arg. The simplest way to deal with this problem would be
2387 to name the tag with a name built from the function name, by
2388 removing the initial 'F' character and substituting '-' for '_'.
2389 Anyway, this assumes that the conventions of naming lisp
2390 functions will never change. Currently, this method is not
2392 if (definedef
== dnone
&& toktype
== st_C_gnumacro
)
2394 next_token_is_func
= TRUE
;
2397 if (next_token_is_func
)
2399 next_token_is_func
= FALSE
;
2401 *is_func_or_var
= TRUE
;
2405 /* Detect Objective C constructs. */
2415 objdef
= oimplementation
;
2419 case oimplementation
:
2420 /* Save the class tag for functions or variables defined inside. */
2421 objtag
= savenstr (str
, len
);
2425 /* Save the class tag for categories. */
2426 objtag
= savenstr (str
, len
);
2428 *is_func_or_var
= TRUE
;
2432 *is_func_or_var
= TRUE
;
2439 objdef
= omethodtag
;
2441 grow_linebuffer (&token_name
, methodlen
+ 1);
2442 strncpy (token_name
.buffer
, str
, len
);
2443 token_name
.buffer
[methodlen
] = '\0';
2444 token_name
.len
= methodlen
;
2450 objdef
= omethodparm
;
2455 objdef
= omethodtag
;
2457 grow_linebuffer (&token_name
, methodlen
+ 1);
2458 strncat (token_name
.buffer
, str
, len
);
2459 token_name
.len
= methodlen
;
2464 if (toktype
== st_C_objend
)
2466 /* Memory leakage here: the string pointed by objtag is
2467 never released, because many tests would be needed to
2468 avoid breaking on incorrect input code. The amount of
2469 memory leaked here is the sum of the lengths of the
2477 /* A function, variable or enum constant? */
2484 if (fvdef
!= finlist
&& fvdef
!= fignore
&& fvdef
!= vignore
)
2485 fvdef
= fvnone
; /* should be useless */
2493 *is_func_or_var
= TRUE
;
2496 if ((c_ext
& C_PLPL
) && strneq (str
+len
-10, "::operator", 10))
2499 *is_func_or_var
= TRUE
;
2502 if (constantypedefs
&& structdef
== sinbody
&& structtype
== st_C_enum
)
2504 if (fvdef
== fvnone
)
2506 fvdef
= fvnameseen
; /* function or variable */
2507 *is_func_or_var
= TRUE
;
2518 * This routine finds functions, variables, typedefs,
2519 * #define's, enum constants and struct/union/enum definitions in
2520 * C syntax and adds them to the list.
2522 #define current_lb_is_new (newndx == curndx)
2523 #define switch_line_buffers() (curndx = 1 - curndx)
2525 #define curlb (lbs[curndx].lb)
2526 #define othlb (lbs[1-curndx].lb)
2527 #define newlb (lbs[newndx].lb)
2528 #define curlinepos (lbs[curndx].linepos)
2529 #define othlinepos (lbs[1-curndx].linepos)
2530 #define newlinepos (lbs[newndx].linepos)
2532 #define CNL_SAVE_DEFINEDEF() \
2534 curlinepos = charno; \
2536 linecharno = charno; \
2537 charno += readline (&curlb, inf); \
2538 lp = curlb.buffer; \
2545 CNL_SAVE_DEFINEDEF(); \
2546 if (savetok.valid) \
2549 savetok.valid = FALSE; \
2551 definedef = dnone; \
2559 /* This function should never be called when tok.valid is FALSE, but
2560 we must protect against invalid input or internal errors. */
2563 if (traditional_tag_style
)
2565 /* This was the original code. Now we call new_pfnote instead,
2566 which uses the new method for naming tags (see new_pfnote). */
2569 if (CTAGS
|| tok
.named
)
2570 name
= savestr (token_name
.buffer
);
2571 pfnote (name
, isfun
,
2572 tok
.buffer
, tok
.linelen
, tok
.lineno
, tok
.linepos
);
2575 new_pfnote (token_name
.buffer
, token_name
.len
, isfun
,
2576 tok
.buffer
, tok
.linelen
, tok
.lineno
, tok
.linepos
);
2585 C_entries (c_ext
, inf
)
2586 int c_ext
; /* extension of C */
2587 FILE *inf
; /* input file */
2589 register char c
; /* latest char read; '\0' for end of line */
2590 register char *lp
; /* pointer one beyond the character `c' */
2591 int curndx
, newndx
; /* indices for current and new lb */
2592 register int tokoff
; /* offset in line of start of current token */
2593 register int toklen
; /* length of current token */
2594 char *qualifier
; /* string used to qualify names */
2595 int qlen
; /* length of qualifier */
2596 int cblev
; /* current curly brace level */
2597 int parlev
; /* current parenthesis level */
2598 bool incomm
, inquote
, inchar
, quotednl
, midtoken
;
2599 bool purec
, cplpl
, cjava
;
2600 token savetok
; /* token saved during preprocessor handling */
2603 tokoff
= toklen
= 0; /* keep compiler quiet */
2604 curndx
= newndx
= 0;
2610 fvdef
= fvnone
; fvextern
= FALSE
; typdef
= tnone
;
2611 structdef
= snone
; definedef
= dnone
; objdef
= onone
;
2612 next_token_is_func
= yacc_rules
= FALSE
;
2613 midtoken
= inquote
= inchar
= incomm
= quotednl
= FALSE
;
2614 tok
.valid
= savetok
.valid
= FALSE
;
2617 purec
= !(c_ext
& ~YACC
); /* no extensions (apart from possibly yacc) */
2618 cplpl
= (c_ext
& C_PLPL
) == C_PLPL
;
2619 cjava
= (c_ext
& C_JAVA
) == C_JAVA
;
2621 { qualifier
= "."; qlen
= 1; }
2623 { qualifier
= "::"; qlen
= 2; }
2630 /* If we're at the end of the line, the next character is a
2631 '\0'; don't skip it, because it's the thing that tells us
2632 to read the next line. */
2653 /* Newlines inside comments do not end macro definitions in
2655 CNL_SAVE_DEFINEDEF ();
2668 /* Newlines inside strings do not end macro definitions
2669 in traditional cpp, even though compilers don't
2670 usually accept them. */
2671 CNL_SAVE_DEFINEDEF ();
2681 /* Hmmm, something went wrong. */
2695 if (fvdef
!= finlist
&& fvdef
!= fignore
&& fvdef
!=vignore
)
2703 if (fvdef
!= finlist
&& fvdef
!= fignore
&& fvdef
!=vignore
)
2716 else if (/* cplpl && */ *lp
== '/')
2724 if ((c_ext
& YACC
) && *lp
== '%')
2726 /* entering or exiting rules section in yacc file */
2728 definedef
= dnone
; fvdef
= fvnone
; fvextern
= FALSE
;
2729 typdef
= tnone
; structdef
= snone
;
2730 next_token_is_func
= FALSE
;
2731 midtoken
= inquote
= inchar
= incomm
= quotednl
= FALSE
;
2733 yacc_rules
= !yacc_rules
;
2739 if (definedef
== dnone
)
2742 bool cpptoken
= TRUE
;
2744 /* Look back on this line. If all blanks, or nonblanks
2745 followed by an end of comment, this is a preprocessor
2747 for (cp
= newlb
.buffer
; cp
< lp
-1; cp
++)
2750 if (*cp
== '*' && *(cp
+1) == '/')
2759 definedef
= dsharpseen
;
2760 } /* if (definedef == dnone) */
2766 /* Consider token only if some complicated conditions are satisfied. */
2767 if ((definedef
!= dnone
2768 || (cblev
== 0 && structdef
!= scolonseen
)
2769 || (cblev
== 1 && cplpl
&& structdef
== sinbody
)
2770 || (structdef
== sinbody
&& purec
))
2771 && typdef
!= tignore
2772 && definedef
!= dignorerest
2773 && fvdef
!= finlist
)
2779 bool funorvar
= FALSE
;
2781 if (c
== ':' && cplpl
&& *lp
== ':' && begtoken (lp
[1]))
2784 * This handles :: in the middle, but not at the
2785 * beginning of an identifier. Also, space-separated
2786 * :: is not recognised.
2796 || consider_token (newlb
.buffer
+ tokoff
, toklen
, c
,
2797 c_ext
, cblev
, parlev
, &funorvar
))
2799 if (fvdef
== foperator
)
2802 lp
= skip_spaces (lp
-1);
2806 && !isspace (*lp
) && *lp
!= '(')
2809 toklen
+= lp
- oldlp
;
2814 && definedef
== dnone
2815 && structdef
== sinbody
)
2816 /* function or var defined in C++ class body */
2818 int len
= strlen (structtag
) + qlen
+ toklen
;
2819 grow_linebuffer (&token_name
, len
+ 1);
2820 strcpy (token_name
.buffer
, structtag
);
2821 strcat (token_name
.buffer
, qualifier
);
2822 strncat (token_name
.buffer
,
2823 newlb
.buffer
+ tokoff
, toklen
);
2824 token_name
.len
= len
;
2827 else if (objdef
== ocatseen
)
2828 /* Objective C category */
2830 int len
= strlen (objtag
) + 2 + toklen
;
2831 grow_linebuffer (&token_name
, len
+ 1);
2832 strcpy (token_name
.buffer
, objtag
);
2833 strcat (token_name
.buffer
, "(");
2834 strncat (token_name
.buffer
,
2835 newlb
.buffer
+ tokoff
, toklen
);
2836 strcat (token_name
.buffer
, ")");
2837 token_name
.len
= len
;
2840 else if (objdef
== omethodtag
2841 || objdef
== omethodparm
)
2842 /* Objective C method */
2848 grow_linebuffer (&token_name
, toklen
+ 1);
2849 strncpy (token_name
.buffer
,
2850 newlb
.buffer
+ tokoff
, toklen
);
2851 token_name
.buffer
[toklen
] = '\0';
2852 token_name
.len
= toklen
;
2853 /* Name macros and members. */
2854 tok
.named
= (structdef
== stagseen
2855 || typdef
== ttypeseen
2858 && definedef
== dignorerest
)
2860 && definedef
== dnone
2861 && structdef
== sinbody
));
2863 tok
.lineno
= lineno
;
2864 tok
.linelen
= tokoff
+ toklen
+ 1;
2865 tok
.buffer
= newlb
.buffer
;
2866 tok
.linepos
= newlinepos
;
2869 if (definedef
== dnone
2870 && (fvdef
== fvnameseen
2871 || fvdef
== foperator
2872 || structdef
== stagseen
2874 || objdef
!= onone
))
2876 if (current_lb_is_new
)
2877 switch_line_buffers ();
2880 make_C_tag (funorvar
);
2884 } /* if (endtoken (c)) */
2885 else if (intoken (c
))
2891 } /* if (midtoken) */
2892 else if (begtoken (c
))
2903 make_C_tag (TRUE
); /* a function */
2910 if (structdef
== stagseen
&& !cjava
)
2916 if (!yacc_rules
|| lp
== newlb
.buffer
+ 1)
2918 tokoff
= lp
- 1 - newlb
.buffer
;
2923 } /* if (begtoken) */
2924 } /* if must look at token */
2927 /* Detect end of line, colon, comma, semicolon and various braces
2928 after having handled a token.*/
2932 if (definedef
!= dnone
)
2938 make_C_tag (TRUE
); /* an Objective C class */
2942 objdef
= omethodcolon
;
2944 grow_linebuffer (&token_name
, methodlen
+ 1);
2945 strcat (token_name
.buffer
, ":");
2946 token_name
.len
= methodlen
;
2949 if (structdef
== stagseen
)
2950 structdef
= scolonseen
;
2957 make_C_tag (FALSE
); /* a yacc function */
2968 if (definedef
!= dnone
)
2974 make_C_tag (FALSE
); /* a typedef */
2984 if ((members
&& cblev
== 1)
2985 || (globals
&& cblev
== 0 && (!fvextern
|| declarations
)))
2986 make_C_tag (FALSE
); /* a variable */
2992 if (declarations
&& (cblev
== 0 || cblev
== 1))
2993 make_C_tag (TRUE
); /* a function declaration */
2998 /* The following instruction invalidates the token.
2999 Probably the token should be invalidated in all
3000 other cases where some state machine is reset. */
3003 if (structdef
== stagseen
)
3007 if (definedef
!= dnone
)
3013 make_C_tag (TRUE
); /* an Objective C method */
3025 if ((members
&& cblev
== 1)
3026 || (globals
&& cblev
== 0 && (!fvextern
|| declarations
)))
3027 make_C_tag (FALSE
); /* a variable */
3032 if (structdef
== stagseen
)
3036 if (definedef
!= dnone
)
3038 if (cblev
== 0 && typdef
== tend
)
3041 make_C_tag (FALSE
); /* a typedef */
3052 if ((members
&& cblev
== 1)
3053 || (globals
&& cblev
== 0 && (!fvextern
|| declarations
)))
3054 make_C_tag (FALSE
); /* a variable */
3059 if (structdef
== stagseen
)
3063 if (definedef
!= dnone
)
3065 if (objdef
== otagseen
&& parlev
== 0)
3066 objdef
= oparenseen
;
3070 if (typdef
== ttypeseen
3073 && structdef
!= sinbody
)
3075 /* This handles constructs like:
3076 typedef void OperatorFun (int fun); */
3091 if (definedef
!= dnone
)
3093 if (objdef
== ocatseen
&& parlev
== 1)
3095 make_C_tag (TRUE
); /* an Objective C category */
3107 if (cblev
== 0 && (typdef
== tend
))
3110 make_C_tag (FALSE
); /* a typedef */
3113 else if (parlev
< 0) /* can happen due to ill-conceived #if's. */
3117 if (definedef
!= dnone
)
3119 if (typdef
== ttypeseen
)
3123 case skeyseen
: /* unnamed struct */
3124 structdef
= sinbody
;
3125 structtag
= "_anonymous_";
3128 case scolonseen
: /* named struct */
3129 structdef
= sinbody
;
3130 make_C_tag (FALSE
); /* a struct */
3136 make_C_tag (TRUE
); /* a function */
3145 make_C_tag (TRUE
); /* an Objective C class */
3150 make_C_tag (TRUE
); /* an Objective C method */
3154 /* Neutralize `extern "C" {' grot. */
3155 if (cblev
== 0 && structdef
== snone
&& typdef
== tnone
)
3162 if (definedef
!= dnone
)
3164 if (fvdef
== fstartlist
)
3165 fvdef
= fvnone
; /* avoid tagging `foo' in `foo (*bar()) ()' */
3168 if (definedef
!= dnone
)
3170 if (!noindentypedefs
&& lp
== newlb
.buffer
+ 1)
3172 cblev
= 0; /* reset curly brace level if first column */
3173 parlev
= 0; /* also reset paren level, just in case... */
3179 if (typdef
== tinbody
)
3181 /* Memory leakage here: the string pointed by structtag is
3182 never released, because I fear to miss something and
3183 break things while freeing the area. The amount of
3184 memory leaked here is the sum of the lengths of the
3186 if (structdef == sinbody)
3187 free (structtag); */
3190 structtag
= "<error>";
3194 if (definedef
!= dnone
)
3204 if ((members
&& cblev
== 1)
3205 || (globals
&& cblev
== 0 && (!fvextern
|| declarations
)))
3206 make_C_tag (FALSE
); /* a variable */
3214 if (objdef
== oinbody
&& cblev
== 0)
3216 objdef
= omethodsign
;
3220 case '#': case '~': case '&': case '%': case '/': case '|':
3221 case '^': case '!': case '<': case '>': case '.': case '?': case ']':
3222 if (definedef
!= dnone
)
3224 /* These surely cannot follow a function tag in C. */
3237 if (objdef
== otagseen
)
3239 make_C_tag (TRUE
); /* an Objective C class */
3242 /* If a macro spans multiple lines don't reset its state. */
3244 CNL_SAVE_DEFINEDEF ();
3250 } /* while not eof */
3254 * Process either a C++ file or a C file depending on the setting
3258 default_C_entries (inf
)
3261 C_entries (cplusplus
? C_PLPL
: 0, inf
);
3264 /* Always do plain ANSI C. */
3266 plain_C_entries (inf
)
3272 /* Always do C++. */
3274 Cplusplus_entries (inf
)
3277 C_entries (C_PLPL
, inf
);
3280 /* Always do Java. */
3285 C_entries (C_JAVA
, inf
);
3293 C_entries (C_STAR
, inf
);
3296 /* Always do Yacc. */
3301 C_entries (YACC
, inf
);
3304 /* A useful macro. */
3305 #define LOOP_ON_INPUT_LINES(file_pointer, line_buffer, char_pointer) \
3306 for (lineno = charno = 0; /* loop initialization */ \
3307 !feof (file_pointer) /* loop test */ \
3308 && (lineno++, /* instructions at start of loop */ \
3309 linecharno = charno, \
3310 charno += readline (&line_buffer, file_pointer), \
3311 char_pointer = lb.buffer, \
3317 * Read a file, but do no processing. This is used to do regexp
3318 * matching on files that have no language defined.
3321 just_read_file (inf
)
3324 register char *dummy
;
3326 LOOP_ON_INPUT_LINES (inf
, lb
, dummy
)
3330 /* Fortran parsing */
3336 register int len
= 0;
3338 while (*cp
!= '\0' && lowcase (*cp
) == lowcase (dbp
[len
]))
3340 if (*cp
== '\0' && !intoken (dbp
[len
]))
3351 dbp
= skip_spaces (dbp
);
3355 dbp
= skip_spaces (dbp
);
3356 if (strneq (dbp
, "(*)", 3))
3361 if (!isdigit (*dbp
))
3363 --dbp
; /* force failure */
3368 while (isdigit (*dbp
));
3377 dbp
= skip_spaces (dbp
);
3381 linecharno
= charno
;
3382 charno
+= readline (&lb
, inf
);
3387 dbp
= skip_spaces (dbp
);
3389 if (!isalpha (*dbp
) && *dbp
!= '_' && *dbp
!= '$')
3391 for (cp
= dbp
+ 1; *cp
!= '\0' && intoken (*cp
); cp
++)
3393 pfnote (savenstr (dbp
, cp
-dbp
), TRUE
,
3394 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
3399 Fortran_functions (inf
)
3402 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
3405 dbp
++; /* Ratfor escape to fortran */
3406 dbp
= skip_spaces (dbp
);
3409 switch (lowcase (*dbp
))
3412 if (tail ("integer"))
3420 if (tail ("logical"))
3424 if (tail ("complex") || tail ("character"))
3428 if (tail ("double"))
3430 dbp
= skip_spaces (dbp
);
3433 if (tail ("precision"))
3439 dbp
= skip_spaces (dbp
);
3442 switch (lowcase (*dbp
))
3445 if (tail ("function"))
3449 if (tail ("subroutine"))
3457 if (tail ("blockdata") || tail ("block data"))
3459 dbp
= skip_spaces (dbp
);
3460 if (*dbp
== '\0') /* assume un-named */
3461 pfnote (savestr ("blockdata"), TRUE
,
3462 lb
.buffer
, dbp
- lb
.buffer
, lineno
, linecharno
);
3464 getit (inf
); /* look for name */
3472 * Philippe Waroquiers <philippe.waroquiers@eurocontrol.be>, 1998-04-24
3475 /* Once we are positioned after an "interesting" keyword, let's get
3476 the real tag value necessary. */
3478 adagetit (inf
, name_qualifier
)
3480 char *name_qualifier
;
3488 dbp
= skip_spaces (dbp
);
3490 || (dbp
[0] == '-' && dbp
[1] == '-'))
3493 linecharno
= charno
;
3494 charno
+= readline (&lb
, inf
);
3503 /* Skipping body of procedure body or package body or ....
3504 resetting qualifier to body instead of spec. */
3505 name_qualifier
= "/b";
3511 /* Skipping type of task type or protected type ... */
3519 for (cp
= dbp
; *cp
!= '\0' && *cp
!= '"'; cp
++)
3524 dbp
= skip_spaces (dbp
);
3527 && (isalpha (*cp
) || isdigit (*cp
) || *cp
== '_' || *cp
== '.'));
3535 name
= concat (dbp
, name_qualifier
, "");
3537 pfnote (name
, TRUE
, lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
3548 bool inquote
= FALSE
;
3550 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
3552 while (*dbp
!= '\0')
3554 /* Skip a string i.e. "abcd". */
3555 if (inquote
|| (*dbp
== '"'))
3557 dbp
= etags_strchr ((inquote
) ? dbp
: dbp
+1, '"');
3562 continue; /* advance char */
3567 break; /* advance line */
3571 /* Skip comments. */
3572 if (dbp
[0] == '-' && dbp
[1] == '-')
3573 break; /* advance line */
3575 /* Skip character enclosed in single quote i.e. 'a'
3576 and skip single quote starting an attribute i.e. 'Image. */
3585 /* Search for beginning of a token. */
3586 if (!begtoken (*dbp
))
3589 continue; /* advance char */
3592 /* We are at the beginning of a token. */
3597 if (!packages_only
&& tail ("function"))
3598 adagetit (inf
, "/f");
3600 break; /* from switch */
3601 continue; /* advance char */
3604 if (!packages_only
&& tail ("procedure"))
3605 adagetit (inf
, "/p");
3606 else if (tail ("package"))
3607 adagetit (inf
, "/s");
3608 else if (tail ("protected")) /* protected type */
3609 adagetit (inf
, "/t");
3611 break; /* from switch */
3612 continue; /* advance char */
3615 if (!packages_only
&& tail ("task"))
3616 adagetit (inf
, "/k");
3617 else if (typedefs
&& !packages_only
&& tail ("type"))
3619 adagetit (inf
, "/t");
3620 while (*dbp
!= '\0')
3624 break; /* from switch */
3625 continue; /* advance char */
3628 /* Look for the end of the token. */
3629 while (!endtoken (*dbp
))
3632 } /* advance char */
3633 } /* advance line */
3637 * Bob Weiner, Motorola Inc., 4/3/94
3638 * Unix and microcontroller assembly tag handling
3639 * look for '^[a-zA-Z_.$][a-zA_Z0-9_.$]*[: ^I^J]'
3647 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
3649 /* If first char is alphabetic or one of [_.$], test for colon
3650 following identifier. */
3651 if (isalpha (*cp
) || *cp
== '_' || *cp
== '.' || *cp
== '$')
3653 /* Read past label. */
3655 while (isalnum (*cp
) || *cp
== '_' || *cp
== '.' || *cp
== '$')
3657 if (*cp
== ':' || isspace (*cp
))
3659 /* Found end of label, so copy it and add it to the table. */
3660 pfnote (savenstr(lb
.buffer
, cp
-lb
.buffer
), TRUE
,
3661 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
3668 * Perl support by Bart Robinson <lomew@cs.utah.edu>
3669 * enhanced by Michael Ernst <mernst@alum.mit.edu>
3670 * Perl sub names: look for /^sub[ \t\n]+[^ \t\n{]+/
3671 * Perl variable names: /^(my|local).../
3674 Perl_functions (inf
)
3679 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
3683 && *cp
++ == 'b' && isspace (*cp
++))
3685 cp
= skip_spaces (cp
);
3690 && !isspace (*cp
) && *cp
!= '{' && *cp
!= '(')
3692 pfnote (savenstr (sp
, cp
-sp
), TRUE
,
3693 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
3696 else if (globals
/* only if tagging global vars is enabled */
3697 && ((cp
= lb
.buffer
,
3706 && (*cp
== '(' || isspace (*cp
)))
3708 /* After "my" or "local", but before any following paren or space. */
3709 char *varname
= NULL
;
3711 cp
= skip_spaces (cp
);
3712 if (*cp
== '$' || *cp
== '@' || *cp
== '%')
3714 char* varstart
= ++cp
;
3715 while (isalnum (*cp
) || *cp
== '_')
3717 varname
= savenstr (varstart
, cp
-varstart
);
3721 /* Should be examining a variable list at this point;
3722 could insist on seeing an open parenthesis. */
3723 while (*cp
!= '\0' && *cp
!= ';' && *cp
!= '=' && *cp
!= ')')
3727 /* Perhaps I should back cp up one character, so the TAGS table
3728 doesn't mention (and so depend upon) the following char. */
3729 pfnote ((CTAGS
) ? savenstr (lb
.buffer
, cp
-lb
.buffer
) : varname
,
3730 FALSE
, lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
3736 * Python support by Eric S. Raymond <esr@thyrsus.com>
3737 * Look for /^def[ \t\n]+[^ \t\n(:]+/ or /^class[ \t\n]+[^ \t\n(:]+/
3740 Python_functions (inf
)
3745 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
3749 && *cp
++ == 'f' && isspace (*cp
++))
3751 cp
= skip_spaces (cp
);
3752 while (*cp
!= '\0' && !isspace (*cp
) && *cp
!= '(' && *cp
!= ':')
3755 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
3763 && *cp
++ == 's' && isspace (*cp
++))
3765 cp
= skip_spaces (cp
);
3766 while (*cp
!= '\0' && !isspace (*cp
) && *cp
!= '(' && *cp
!= ':')
3769 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
3774 /* Idea by Corny de Souza
3775 * Cobol tag functions
3776 * We could look for anything that could be a paragraph name.
3777 * i.e. anything that starts in column 8 is one word and ends in a full stop.
3780 Cobol_paragraphs (inf
)
3783 register char *bp
, *ep
;
3785 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
3791 /* If eoln, compiler option or comment ignore whole line. */
3792 if (bp
[-1] != ' ' || !isalnum (bp
[0]))
3795 for (ep
= bp
; isalnum (*ep
) || *ep
== '-'; ep
++)
3798 pfnote (savenstr (bp
, ep
-bp
), TRUE
,
3799 lb
.buffer
, ep
- lb
.buffer
+ 1, lineno
, linecharno
);
3803 /* Added by Mosur Mohan, 4/22/88 */
3804 /* Pascal parsing */
3807 * Locates tags for procedures & functions. Doesn't do any type- or
3808 * var-definitions. It does look for the keyword "extern" or
3809 * "forward" immediately following the procedure statement; if found,
3810 * the tag is skipped.
3813 Pascal_functions (inf
)
3816 linebuffer tline
; /* mostly copied from C_entries */
3818 int save_lineno
, save_len
;
3819 char c
, *cp
, *namebuf
;
3821 bool /* each of these flags is TRUE iff: */
3822 incomment
, /* point is inside a comment */
3823 inquote
, /* point is inside '..' string */
3824 get_tagname
, /* point is after PROCEDURE/FUNCTION
3825 keyword, so next item = potential tag */
3826 found_tag
, /* point is after a potential tag */
3827 inparms
, /* point is within parameter-list */
3828 verify_tag
; /* point has passed the parm-list, so the
3829 next token will determine whether this
3830 is a FORWARD/EXTERN to be ignored, or
3831 whether it is a real tag */
3833 save_lcno
= save_lineno
= save_len
= 0; /* keep compiler quiet */
3834 namebuf
= NULL
; /* keep compiler quiet */
3839 initbuffer (&tline
);
3841 incomment
= inquote
= FALSE
;
3842 found_tag
= FALSE
; /* have a proc name; check if extern */
3843 get_tagname
= FALSE
; /* have found "procedure" keyword */
3844 inparms
= FALSE
; /* found '(' after "proc" */
3845 verify_tag
= FALSE
; /* check if "extern" is ahead */
3848 while (!feof (inf
)) /* long main loop to get next char */
3851 if (c
== '\0') /* if end of line */
3854 linecharno
= charno
;
3855 charno
+= readline (&lb
, inf
);
3859 if (!((found_tag
&& verify_tag
)
3861 c
= *dbp
++; /* only if don't need *dbp pointing
3862 to the beginning of the name of
3863 the procedure or function */
3867 if (c
== '}') /* within { } comments */
3869 else if (c
== '*' && *dbp
== ')') /* within (* *) comments */
3886 inquote
= TRUE
; /* found first quote */
3888 case '{': /* found open { comment */
3892 if (*dbp
== '*') /* found open (* comment */
3897 else if (found_tag
) /* found '(' after tag, i.e., parm-list */
3900 case ')': /* end of parms list */
3905 if (found_tag
&& !inparms
) /* end of proc or fn stmt */
3912 if (found_tag
&& verify_tag
&& (*dbp
!= ' '))
3914 /* check if this is an "extern" declaration */
3917 if (lowcase (*dbp
== 'e'))
3919 if (tail ("extern")) /* superfluous, really! */
3925 else if (lowcase (*dbp
) == 'f')
3927 if (tail ("forward")) /* check for forward reference */
3933 if (found_tag
&& verify_tag
) /* not external proc, so make tag */
3937 pfnote (namebuf
, TRUE
,
3938 tline
.buffer
, save_len
, save_lineno
, save_lcno
);
3942 if (get_tagname
) /* grab name of proc or fn */
3947 /* save all values for later tagging */
3948 grow_linebuffer (&tline
, lb
.len
+ 1);
3949 strcpy (tline
.buffer
, lb
.buffer
);
3950 save_lineno
= lineno
;
3951 save_lcno
= linecharno
;
3953 /* grab block name */
3954 for (cp
= dbp
+ 1; *cp
!= '\0' && !endtoken (*cp
); cp
++)
3956 namebuf
= savenstr (dbp
, cp
-dbp
);
3957 dbp
= cp
; /* set dbp to e-o-token */
3958 save_len
= dbp
- lb
.buffer
+ 1;
3959 get_tagname
= FALSE
;
3963 /* and proceed to check for "extern" */
3965 else if (!incomment
&& !inquote
&& !found_tag
)
3967 /* check for proc/fn keywords */
3968 switch (lowcase (c
))
3971 if (tail ("rocedure")) /* c = 'p', dbp has advanced */
3975 if (tail ("unction"))
3980 } /* while not eof */
3982 free (tline
.buffer
);
3986 * lisp tag functions
3987 * look for (def or (DEF, quote or QUOTE
3991 register char *strp
;
3993 return ((strp
[1] == 'd' || strp
[1] == 'D')
3994 && (strp
[2] == 'e' || strp
[2] == 'E')
3995 && (strp
[3] == 'f' || strp
[3] == 'F'));
4000 register char *strp
;
4002 return ((*++strp
== 'q' || *strp
== 'Q')
4003 && (*++strp
== 'u' || *strp
== 'U')
4004 && (*++strp
== 'o' || *strp
== 'O')
4005 && (*++strp
== 't' || *strp
== 'T')
4006 && (*++strp
== 'e' || *strp
== 'E')
4007 && isspace (*++strp
));
4015 if (*dbp
== '\'') /* Skip prefix quote */
4017 else if (*dbp
== '(')
4019 if (L_isquote (dbp
))
4020 dbp
+= 7; /* Skip "(quote " */
4022 dbp
+= 1; /* Skip "(" before name in (defstruct (foo)) */
4023 dbp
= skip_spaces (dbp
);
4026 for (cp
= dbp
/*+1*/;
4027 *cp
!= '\0' && *cp
!= '(' && !isspace(*cp
) && *cp
!= ')';
4033 pfnote (savenstr (dbp
, cp
-dbp
), TRUE
,
4034 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4038 Lisp_functions (inf
)
4041 LOOP_ON_INPUT_LINES (inf
, lb
, dbp
)
4047 dbp
= skip_non_spaces (dbp
);
4048 dbp
= skip_spaces (dbp
);
4053 /* Check for (foo::defmumble name-defined ... */
4056 while (*dbp
!= '\0' && !isspace (*dbp
)
4057 && *dbp
!= ':' && *dbp
!= '(' && *dbp
!= ')');
4062 while (*dbp
== ':');
4064 if (L_isdef (dbp
- 1))
4066 dbp
= skip_non_spaces (dbp
);
4067 dbp
= skip_spaces (dbp
);
4077 * Postscript tag functions
4078 * Just look for lines where the first character is '/'
4079 * Richard Mlynarik <mly@adoc.xerox.com>
4080 * Also look at "defineps" for PSWrap
4081 * suggested by Masatake YAMATO <masata-y@is.aist-nara.ac.jp>
4084 Postscript_functions (inf
)
4087 register char *bp
, *ep
;
4089 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4094 *ep
!= '\0' && *ep
!= ' ' && *ep
!= '{';
4097 pfnote (savenstr (bp
, ep
-bp
), TRUE
,
4098 lb
.buffer
, ep
- lb
.buffer
+ 1, lineno
, linecharno
);
4100 else if (strneq (bp
, "defineps", 8))
4102 bp
= skip_non_spaces (bp
);
4103 bp
= skip_spaces (bp
);
4111 * Scheme tag functions
4112 * look for (def... xyzzy
4113 * look for (def... (xyzzy
4114 * look for (def ... ((...(xyzzy ....
4115 * look for (set! xyzzy
4119 Scheme_functions (inf
)
4124 LOOP_ON_INPUT_LINES (inf
, lb
, bp
)
4127 && (bp
[1] == 'D' || bp
[1] == 'd')
4128 && (bp
[2] == 'E' || bp
[2] == 'e')
4129 && (bp
[3] == 'F' || bp
[3] == 'f'))
4131 bp
= skip_non_spaces (bp
);
4132 /* Skip over open parens and white space */
4133 while (isspace (*bp
) || *bp
== '(')
4138 && (bp
[1] == 'S' || bp
[1] == 's')
4139 && (bp
[2] == 'E' || bp
[2] == 'e')
4140 && (bp
[3] == 'T' || bp
[3] == 't')
4141 && (bp
[4] == '!' || bp
[4] == '!')
4142 && (isspace (bp
[5])))
4144 bp
= skip_non_spaces (bp
);
4145 bp
= skip_spaces (bp
);
4151 /* Find tags in TeX and LaTeX input files. */
4153 /* TEX_toktab is a table of TeX control sequences that define tags.
4154 Each TEX_tabent records one such control sequence.
4155 CONVERT THIS TO USE THE Stab TYPE!! */
4162 struct TEX_tabent
*TEX_toktab
= NULL
; /* Table with tag tokens */
4164 /* Default set of control sequences to put into TEX_toktab.
4165 The value of environment var TEXTAGS is prepended to this. */
4167 char *TEX_defenv
= "\
4168 :chapter:section:subsection:subsubsection:eqno:label:ref:cite:bibitem\
4169 :part:appendix:entry:index";
4172 struct TEX_tabent
*TEX_decode_env ();
4175 char TEX_esc
= '\\';
4176 char TEX_opgrp
= '{';
4177 char TEX_clgrp
= '}';
4180 * TeX/LaTeX scanning loop.
4189 /* Select either \ or ! as escape character. */
4192 /* Initialize token table once from environment. */
4194 TEX_toktab
= TEX_decode_env ("TEXTAGS", TEX_defenv
);
4196 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4199 /* Look at each esc in line. */
4200 while ((cp
= etags_strchr (cp
, TEX_esc
)) != NULL
)
4204 linecharno
+= cp
- lasthit
;
4206 i
= TEX_Token (lasthit
);
4209 /* We seem to include the TeX command in the tag name.
4211 for (p = lasthit + TEX_toktab[i].len;
4212 *p != '\0' && *p != TEX_clgrp;
4215 pfnote (/*savenstr (lasthit, p-lasthit)*/ (char *)NULL
, TRUE
,
4216 lb
.buffer
, lb
.len
, lineno
, linecharno
);
4217 break; /* We only tag a line once */
4223 #define TEX_LESC '\\'
4224 #define TEX_SESC '!'
4227 /* Figure out whether TeX's escapechar is '\\' or '!' and set grouping
4228 chars accordingly. */
4235 while ((c
= getc (inf
)) != EOF
)
4237 /* Skip to next line if we hit the TeX comment char. */
4241 else if (c
== TEX_LESC
|| c
== TEX_SESC
)
4257 /* If the input file is compressed, inf is a pipe, and rewind may fail.
4258 No attempt is made to correct the situation. */
4262 /* Read environment and prepend it to the default string.
4263 Build token table. */
4265 TEX_decode_env (evarname
, defenv
)
4269 register char *env
, *p
;
4271 struct TEX_tabent
*tab
;
4274 /* Append default string to environment. */
4275 env
= getenv (evarname
);
4281 env
= concat (oldenv
, defenv
, "");
4284 /* Allocate a token table */
4285 for (size
= 1, p
= env
; p
;)
4286 if ((p
= etags_strchr (p
, ':')) && *++p
!= '\0')
4288 /* Add 1 to leave room for null terminator. */
4289 tab
= xnew (size
+ 1, struct TEX_tabent
);
4291 /* Unpack environment string into token table. Be careful about */
4292 /* zero-length strings (leading ':', "::" and trailing ':') */
4295 p
= etags_strchr (env
, ':');
4296 if (!p
) /* End of environment string. */
4297 p
= env
+ strlen (env
);
4299 { /* Only non-zero strings. */
4300 tab
[i
].name
= savenstr (env
, p
- env
);
4301 tab
[i
].len
= strlen (tab
[i
].name
);
4308 tab
[i
].name
= NULL
; /* Mark end of table. */
4316 /* If the text at CP matches one of the tag-defining TeX command names,
4317 return the pointer to the first occurrence of that command in TEX_toktab.
4318 Otherwise return -1.
4319 Keep the capital `T' in `token' for dumb truncating compilers
4320 (this distinguishes it from `TEX_toktab' */
4327 for (i
= 0; TEX_toktab
[i
].len
> 0; i
++)
4328 if (strneq (TEX_toktab
[i
].name
, cp
, TEX_toktab
[i
].len
))
4334 * Prolog support (rewritten) by Anders Lindgren, Mar. 96
4336 * Assumes that the predicate starts at column 0.
4337 * Only the first clause of a predicate is added.
4340 void prolog_skip_comment ();
4344 Prolog_functions (inf
)
4355 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4357 if (cp
[0] == '\0') /* Empty line */
4359 else if (isspace (cp
[0])) /* Not a predicate */
4361 else if (cp
[0] == '/' && cp
[1] == '*') /* comment. */
4362 prolog_skip_comment (&lb
, inf
);
4363 else if ((len
= prolog_pred (cp
, last
)) > 0)
4365 /* Predicate. Store the function name so that we only
4366 generate a tag for the first clause. */
4368 last
= xnew(len
+ 1, char);
4369 else if (len
+ 1 > allocated
)
4370 last
= xrnew (last
, len
+ 1, char);
4371 allocated
= len
+ 1;
4372 strncpy (last
, cp
, len
);
4380 prolog_skip_comment (plb
, inf
)
4388 for (cp
= plb
->buffer
; *cp
!= '\0'; cp
++)
4389 if (cp
[0] == '*' && cp
[1] == '/')
4392 linecharno
+= readline (plb
, inf
);
4398 * A predicate definition is added if it matches:
4399 * <beginning of line><Prolog Atom><whitespace>(
4401 * It is added to the tags database if it doesn't match the
4402 * name of the previous clause header.
4404 * Return the size of the name of the predicate, or 0 if no header
4408 prolog_pred (s
, last
)
4410 char *last
; /* Name of last clause. */
4415 pos
= prolog_atom (s
, 0);
4420 pos
= skip_spaces (s
+ pos
) - s
;
4422 if ((s
[pos
] == '(') || (s
[pos
] == '.'))
4427 /* Save only the first clause. */
4429 || len
!= (int)strlen (last
)
4430 || !strneq (s
, last
, len
))
4432 pfnote (savenstr (s
, len
), TRUE
, s
, pos
, lineno
, linecharno
);
4440 * Consume a Prolog atom.
4441 * Return the number of bytes consumed, or -1 if there was an error.
4443 * A prolog atom, in this context, could be one of:
4444 * - An alphanumeric sequence, starting with a lower case letter.
4445 * - A quoted arbitrary string. Single quotes can escape themselves.
4446 * Backslash quotes everything.
4449 prolog_atom (s
, pos
)
4457 if (islower(s
[pos
]) || (s
[pos
] == '_'))
4459 /* The atom is unquoted. */
4461 while (isalnum(s
[pos
]) || (s
[pos
] == '_'))
4465 return pos
- origpos
;
4467 else if (s
[pos
] == '\'')
4478 pos
++; /* A double quote */
4480 else if (s
[pos
] == '\0')
4481 /* Multiline quoted atoms are ignored. */
4483 else if (s
[pos
] == '\\')
4485 if (s
[pos
+1] == '\0')
4492 return pos
- origpos
;
4499 * Support for Erlang -- Anders Lindgren, Feb 1996.
4501 * Generates tags for functions, defines, and records.
4503 * Assumes that Erlang functions start at column 0.
4506 void erlang_attribute ();
4510 Erlang_functions (inf
)
4521 LOOP_ON_INPUT_LINES (inf
, lb
, cp
)
4523 if (cp
[0] == '\0') /* Empty line */
4525 else if (isspace (cp
[0])) /* Not function nor attribute */
4527 else if (cp
[0] == '%') /* comment */
4529 else if (cp
[0] == '"') /* Sometimes, strings start in column one */
4531 else if (cp
[0] == '-') /* attribute, e.g. "-define" */
4533 erlang_attribute (cp
);
4536 else if ((len
= erlang_func (cp
, last
)) > 0)
4539 * Function. Store the function name so that we only
4540 * generates a tag for the first clause.
4543 last
= xnew (len
+ 1, char);
4544 else if (len
+ 1 > allocated
)
4545 last
= xrnew (last
, len
+ 1, char);
4546 allocated
= len
+ 1;
4547 strncpy (last
, cp
, len
);
4555 * A function definition is added if it matches:
4556 * <beginning of line><Erlang Atom><whitespace>(
4558 * It is added to the tags database if it doesn't match the
4559 * name of the previous clause header.
4561 * Return the size of the name of the function, or 0 if no function
4565 erlang_func (s
, last
)
4567 char *last
; /* Name of last clause. */
4572 pos
= erlang_atom (s
, 0);
4577 pos
= skip_spaces (s
+ pos
) - s
;
4579 /* Save only the first clause. */
4582 || len
!= (int)strlen (last
)
4583 || !strneq (s
, last
, len
)))
4585 pfnote (savenstr (s
, len
), TRUE
, s
, pos
, lineno
, linecharno
);
4594 * Handle attributes. Currently, tags are generated for defines
4597 * They are on the form:
4598 * -define(foo, bar).
4599 * -define(Foo(M, N), M+N).
4600 * -record(graph, {vtab = notable, cyclic = true}).
4603 erlang_attribute (s
)
4609 if (strneq (s
, "-define", 7) || strneq (s
, "-record", 7))
4611 pos
= skip_spaces (s
+ 7) - s
;
4612 if (s
[pos
++] == '(')
4614 pos
= skip_spaces (s
+ pos
) - s
;
4615 len
= erlang_atom (s
, pos
);
4617 pfnote (savenstr (& s
[pos
], len
), TRUE
,
4618 s
, pos
+ len
, lineno
, linecharno
);
4626 * Consume an Erlang atom (or variable).
4627 * Return the number of bytes consumed, or -1 if there was an error.
4630 erlang_atom (s
, pos
)
4638 if (isalpha (s
[pos
]) || s
[pos
] == '_')
4640 /* The atom is unquoted. */
4642 while (isalnum (s
[pos
]) || s
[pos
] == '_')
4644 return pos
- origpos
;
4646 else if (s
[pos
] == '\'')
4657 else if (s
[pos
] == '\0')
4658 /* Multiline quoted atoms are ignored. */
4660 else if (s
[pos
] == '\\')
4662 if (s
[pos
+1] == '\0')
4669 return pos
- origpos
;
4675 #ifdef ETAGS_REGEXPS
4677 /* Take a string like "/blah/" and turn it into "blah", making sure
4678 that the first and last characters are the same, and handling
4679 quoted separator characters. Actually, stops on the occurrence of
4680 an unquoted separator. Also turns "\t" into a Tab character.
4681 Returns pointer to terminating separator. Works in place. Null
4682 terminates name string. */
4684 scan_separators (name
)
4688 char *copyto
= name
;
4689 bool quoted
= FALSE
;
4691 for (++name
; *name
!= '\0'; ++name
)
4697 else if (*name
== sep
)
4701 /* Something else is quoted, so preserve the quote. */
4707 else if (*name
== '\\')
4709 else if (*name
== sep
)
4715 /* Terminate copied string. */
4720 /* Look at the argument of --regex or --no-regex and do the right
4721 thing. Same for each line of a regexp file. */
4723 analyse_regex (regex_arg
, ignore_case
)
4727 if (regex_arg
== NULL
)
4728 free_patterns (); /* --no-regex: remove existing regexps */
4730 /* A real --regexp option or a line in a regexp file. */
4731 switch (regex_arg
[0])
4733 /* Comments in regexp file or null arg to --regex. */
4739 /* Read a regex file. This is recursive and may result in a
4740 loop, which will stop when the file descriptors are exhausted. */
4744 linebuffer regexbuf
;
4745 char *regexfile
= regex_arg
+ 1;
4747 /* regexfile is a file containing regexps, one per line. */
4748 regexfp
= fopen (regexfile
, "r");
4749 if (regexfp
== NULL
)
4754 initbuffer (®exbuf
);
4755 while (readline_internal (®exbuf
, regexfp
) > 0)
4756 analyse_regex (regexbuf
.buffer
, ignore_case
);
4757 free (regexbuf
.buffer
);
4762 /* Regexp to be used for a specific language only. */
4766 char *lang_name
= regex_arg
+ 1;
4769 for (cp
= lang_name
; *cp
!= '}'; cp
++)
4772 error ("unterminated language name in regex: %s", regex_arg
);
4776 lang
= get_language_from_name (lang_name
);
4779 add_regex (cp
+ 1, ignore_case
, lang
);
4783 /* Regexp to be used for any language. */
4785 add_regex (regex_arg
, ignore_case
, NULL
);
4790 /* Turn a name, which is an ed-style (but Emacs syntax) regular
4791 expression, into a real regular expression by compiling it. */
4793 add_regex (regexp_pattern
, ignore_case
, lang
)
4794 char *regexp_pattern
;
4800 struct re_pattern_buffer
*patbuf
;
4804 if (regexp_pattern
[strlen(regexp_pattern
)-1] != regexp_pattern
[0])
4806 error ("%s: unterminated regexp", regexp_pattern
);
4809 name
= scan_separators (regexp_pattern
);
4810 if (regexp_pattern
[0] == '\0')
4812 error ("null regexp", (char *)NULL
);
4815 (void) scan_separators (name
);
4817 patbuf
= xnew (1, struct re_pattern_buffer
);
4818 /* Translation table to fold case if appropriate. */
4819 patbuf
->translate
= (ignore_case
) ? lc_trans
: NULL
;
4820 patbuf
->fastmap
= NULL
;
4821 patbuf
->buffer
= NULL
;
4822 patbuf
->allocated
= 0;
4824 err
= re_compile_pattern (regexp_pattern
, strlen (regexp_pattern
), patbuf
);
4827 error ("%s while compiling pattern", err
);
4832 p_head
= xnew (1, pattern
);
4833 p_head
->regex
= savestr (regexp_pattern
);
4834 p_head
->p_next
= pp
;
4835 p_head
->language
= lang
;
4836 p_head
->pattern
= patbuf
;
4837 p_head
->name_pattern
= savestr (name
);
4838 p_head
->error_signaled
= FALSE
;
4842 * Do the substitutions indicated by the regular expression and
4846 substitute (in
, out
, regs
)
4848 struct re_registers
*regs
;
4851 int size
, dig
, diglen
;
4854 size
= strlen (out
);
4856 /* Pass 1: figure out how much to allocate by finding all \N strings. */
4857 if (out
[size
- 1] == '\\')
4858 fatal ("pattern error in \"%s\"", out
);
4859 for (t
= etags_strchr (out
, '\\');
4861 t
= etags_strchr (t
+ 2, '\\'))
4865 diglen
= regs
->end
[dig
] - regs
->start
[dig
];
4871 /* Allocate space and do the substitutions. */
4872 result
= xnew (size
+ 1, char);
4874 for (t
= result
; *out
!= '\0'; out
++)
4875 if (*out
== '\\' && isdigit (*++out
))
4877 /* Using "dig2" satisfies my debugger. Bleah. */
4879 diglen
= regs
->end
[dig
] - regs
->start
[dig
];
4880 strncpy (t
, in
+ regs
->start
[dig
], diglen
);
4887 if (DEBUG
&& (t
> result
+ size
|| t
- result
!= (int)strlen (result
)))
4893 /* Deallocate all patterns. */
4898 while (p_head
!= NULL
)
4900 pp
= p_head
->p_next
;
4901 free (p_head
->regex
);
4902 free (p_head
->name_pattern
);
4917 /* Go till you get to white space or a syntactic break */
4919 *cp
!= '\0' && *cp
!= '(' && *cp
!= ')' && !isspace (*cp
);
4922 pfnote (savenstr (bp
, cp
-bp
), TRUE
,
4923 lb
.buffer
, cp
- lb
.buffer
+ 1, lineno
, linecharno
);
4926 #endif /* ETAGS_REGEXPS */
4927 /* Initialize a linebuffer for use */
4933 lbp
->buffer
= xnew (200, char);
4937 * Read a line of text from `stream' into `lbp', excluding the
4938 * newline or CR-NL, if any. Return the number of characters read from
4939 * `stream', which is the length of the line including the newline.
4941 * On DOS or Windows we do not count the CR character, if any, before the
4942 * NL, in the returned length; this mirrors the behavior of emacs on those
4943 * platforms (for text files, it translates CR-NL to NL as it reads in the
4947 readline_internal (lbp
, stream
)
4949 register FILE *stream
;
4951 char *buffer
= lbp
->buffer
;
4952 register char *p
= lbp
->buffer
;
4953 register char *pend
;
4956 pend
= p
+ lbp
->size
; /* Separate to avoid 386/IX compiler bug. */
4960 register int c
= getc (stream
);
4963 /* We're at the end of linebuffer: expand it. */
4965 buffer
= xrnew (buffer
, lbp
->size
, char);
4966 p
+= buffer
- lbp
->buffer
;
4967 pend
= buffer
+ lbp
->size
;
4968 lbp
->buffer
= buffer
;
4978 if (p
> buffer
&& p
[-1] == '\r')
4982 /* Assume CRLF->LF translation will be performed by Emacs
4983 when loading this file, so CRs won't appear in the buffer.
4984 It would be cleaner to compensate within Emacs;
4985 however, Emacs does not know how many CRs were deleted
4986 before any given point in the file. */
5001 lbp
->len
= p
- buffer
;
5003 return lbp
->len
+ chars_deleted
;
5007 * Like readline_internal, above, but in addition try to match the
5008 * input line against relevant regular expressions.
5011 readline (lbp
, stream
)
5015 /* Read new line. */
5016 long result
= readline_internal (lbp
, stream
);
5017 #ifdef ETAGS_REGEXPS
5021 /* Match against relevant patterns. */
5023 for (pp
= p_head
; pp
!= NULL
; pp
= pp
->p_next
)
5025 /* Only use generic regexps or those for the current language. */
5026 if (pp
->language
!= NULL
&& pp
->language
!= curlang
)
5029 match
= re_match (pp
->pattern
, lbp
->buffer
, lbp
->len
, 0, &pp
->regs
);
5034 if (!pp
->error_signaled
)
5036 error ("error while matching \"%s\"", pp
->regex
);
5037 pp
->error_signaled
= TRUE
;
5044 /* Match occurred. Construct a tag. */
5045 if (pp
->name_pattern
[0] != '\0')
5047 /* Make a named tag. */
5048 char *name
= substitute (lbp
->buffer
,
5049 pp
->name_pattern
, &pp
->regs
);
5051 pfnote (name
, TRUE
, lbp
->buffer
, match
, lineno
, linecharno
);
5055 /* Make an unnamed tag. */
5056 pfnote ((char *)NULL
, TRUE
,
5057 lbp
->buffer
, match
, lineno
, linecharno
);
5062 #endif /* ETAGS_REGEXPS */
5068 * Return a pointer to a space of size strlen(cp)+1 allocated
5069 * with xnew where the string CP has been copied.
5075 return savenstr (cp
, strlen (cp
));
5079 * Return a pointer to a space of size LEN+1 allocated with xnew where
5080 * the string CP has been copied for at most the first LEN characters.
5089 dp
= xnew (len
+ 1, char);
5090 strncpy (dp
, cp
, len
);
5096 * Return the ptr in sp at which the character c last
5097 * appears; NULL if not found
5099 * Identical to System V strrchr, included for portability.
5102 etags_strrchr (sp
, c
)
5103 register char *sp
, c
;
5118 * Return the ptr in sp at which the character c first
5119 * appears; NULL if not found
5121 * Identical to System V strchr, included for portability.
5124 etags_strchr (sp
, c
)
5125 register char *sp
, c
;
5135 /* Skip spaces, return new pointer. */
5140 while (isspace (*cp
)) /* isspace('\0')==FALSE */
5145 /* Skip non spaces, return new pointer. */
5147 skip_non_spaces (cp
)
5150 while (!iswhite (*cp
)) /* iswhite('\0')==TRUE */
5155 /* Print error message and exit. */
5173 suggest_asking_for_help ()
5175 fprintf (stderr
, "\tTry `%s %s' for a complete list of options.\n",
5186 /* Print error message. `s1' is printf control string, `s2' is arg for it. */
5191 fprintf (stderr
, "%s: ", progname
);
5192 fprintf (stderr
, s1
, s2
);
5193 fprintf (stderr
, "\n");
5196 /* Return a newly-allocated string whose contents
5197 concatenate those of s1, s2, s3. */
5202 int len1
= strlen (s1
), len2
= strlen (s2
), len3
= strlen (s3
);
5203 char *result
= xnew (len1
+ len2
+ len3
+ 1, char);
5205 strcpy (result
, s1
);
5206 strcpy (result
+ len1
, s2
);
5207 strcpy (result
+ len1
+ len2
, s3
);
5208 result
[len1
+ len2
+ len3
] = '\0';
5213 /* Does the same work as the system V getcwd, but does not need to
5214 guess the buffer size in advance. */
5220 char *path
= xnew (bufsize
, char);
5222 while (getcwd (path
, bufsize
) == NULL
)
5224 if (errno
!= ERANGE
)
5228 path
= xnew (bufsize
, char);
5231 canonicalize_filename (path
);
5234 #else /* not HAVE_GETCWD */
5236 char *p
, path
[MAXPATHLEN
+ 1]; /* Fixed size is safe on MSDOS. */
5240 for (p
= path
; *p
!= '\0'; p
++)
5246 return strdup (path
);
5247 #else /* not MSDOS */
5252 pipe
= (FILE *) popen ("pwd 2>/dev/null", "r");
5253 if (pipe
== NULL
|| readline_internal (&path
, pipe
) == 0)
5258 #endif /* not MSDOS */
5259 #endif /* not HAVE_GETCWD */
5262 /* Return a newly allocated string containing the file name of FILE
5263 relative to the absolute directory DIR (which should end with a slash). */
5265 relative_filename (file
, dir
)
5268 char *fp
, *dp
, *afn
, *res
;
5271 /* Find the common root of file and dir (with a trailing slash). */
5272 afn
= absolute_filename (file
, cwd
);
5275 while (*fp
++ == *dp
++)
5277 fp
--, dp
--; /* back to the first differing char */
5279 if (fp
== afn
&& afn
[0] != '/') /* cannot build a relative name */
5282 do /* look at the equal chars until '/' */
5286 /* Build a sequence of "../" strings for the resulting relative file name. */
5288 while ((dp
= etags_strchr (dp
+ 1, '/')) != NULL
)
5290 res
= xnew (3*i
+ strlen (fp
+ 1) + 1, char);
5293 strcat (res
, "../");
5295 /* Add the file name relative to the common root of file and dir. */
5296 strcat (res
, fp
+ 1);
5302 /* Return a newly allocated string containing the absolute file name
5303 of FILE given DIR (which should end with a slash). */
5305 absolute_filename (file
, dir
)
5308 char *slashp
, *cp
, *res
;
5310 if (filename_is_absolute (file
))
5311 res
= savestr (file
);
5313 /* We don't support non-absolute file names with a drive
5314 letter, like `d:NAME' (it's too much hassle). */
5315 else if (file
[1] == ':')
5316 fatal ("%s: relative file names with drive letters not supported", file
);
5319 res
= concat (dir
, file
, "");
5321 /* Delete the "/dirname/.." and "/." substrings. */
5322 slashp
= etags_strchr (res
, '/');
5323 while (slashp
!= NULL
&& slashp
[0] != '\0')
5325 if (slashp
[1] == '.')
5327 if (slashp
[2] == '.'
5328 && (slashp
[3] == '/' || slashp
[3] == '\0'))
5333 while (cp
>= res
&& !filename_is_absolute (cp
));
5335 cp
= slashp
; /* the absolute name begins with "/.." */
5337 /* Under MSDOS and NT we get `d:/NAME' as absolute
5338 file name, so the luser could say `d:/../NAME'.
5339 We silently treat this as `d:/NAME'. */
5340 else if (cp
[0] != '/')
5343 strcpy (cp
, slashp
+ 3);
5347 else if (slashp
[2] == '/' || slashp
[2] == '\0')
5349 strcpy (slashp
, slashp
+ 2);
5354 slashp
= etags_strchr (slashp
+ 1, '/');
5358 return savestr ("/");
5363 /* Return a newly allocated string containing the absolute
5364 file name of dir where FILE resides given DIR (which should
5365 end with a slash). */
5367 absolute_dirname (file
, dir
)
5373 canonicalize_filename (file
);
5374 slashp
= etags_strrchr (file
, '/');
5376 return savestr (dir
);
5379 res
= absolute_filename (file
, dir
);
5385 /* Whether the argument string is an absolute file name. The argument
5386 string must have been canonicalized with canonicalize_filename. */
5388 filename_is_absolute (fn
)
5391 return (fn
[0] == '/'
5393 || (isalpha(fn
[0]) && fn
[1] == ':' && fn
[2] == '/')
5398 /* Translate backslashes into slashes. Works in place. */
5400 canonicalize_filename (fn
)
5404 /* Canonicalize drive letter case. */
5405 if (islower (fn
[0]))
5406 fn
[0] = toupper (fn
[0]);
5407 /* Convert backslashes to slashes. */
5408 for (; *fn
!= '\0'; fn
++)
5413 fn
= NULL
; /* shut up the compiler */
5417 /* Increase the size of a linebuffer. */
5419 grow_linebuffer (lbp
, toksize
)
5423 while (lbp
->size
< toksize
)
5425 lbp
->buffer
= xrnew (lbp
->buffer
, lbp
->size
, char);
5428 /* Like malloc but get fatal error if memory is exhausted. */
5433 long *result
= (long *) malloc (size
);
5435 fatal ("virtual memory exhausted", (char *)NULL
);
5440 xrealloc (ptr
, size
)
5444 long *result
= (long *) realloc (ptr
, size
);
5446 fatal ("virtual memory exhausted", (char *)NULL
);