| 1 | /* Tags file maker to go with GNU Emacs |
| 2 | Copyright (C) 1984, 87, 88, 89, 93, 94, 95 |
| 3 | Free Software Foundation, Inc. and Ken Arnold |
| 4 | |
| 5 | This file is not considered part of GNU Emacs. |
| 6 | |
| 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. |
| 11 | |
| 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. |
| 16 | |
| 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. */ |
| 20 | |
| 21 | /* |
| 22 | * Authors: |
| 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. |
| 30 | * |
| 31 | * Francesco Potorti` (F.Potorti@cnuce.cnr.it) is the current maintainer. |
| 32 | */ |
| 33 | |
| 34 | char pot_etags_version[] = "@(#) pot revision number is 11.71"; |
| 35 | |
| 36 | #define TRUE 1 |
| 37 | #define FALSE 0 |
| 38 | |
| 39 | #ifndef DEBUG |
| 40 | # define DEBUG FALSE |
| 41 | #endif |
| 42 | |
| 43 | #ifdef MSDOS |
| 44 | # include <string.h> |
| 45 | # include <fcntl.h> |
| 46 | # include <sys/param.h> |
| 47 | #endif /* MSDOS */ |
| 48 | |
| 49 | #ifdef WINDOWSNT |
| 50 | # include <stdlib.h> |
| 51 | # include <fcntl.h> |
| 52 | # include <string.h> |
| 53 | # include <io.h> |
| 54 | # define MAXPATHLEN _MAX_PATH |
| 55 | #endif |
| 56 | |
| 57 | #ifdef HAVE_CONFIG_H |
| 58 | # include <config.h> |
| 59 | /* On some systems, Emacs defines static as nothing for the sake |
| 60 | of unexec. We don't want that here since we don't use unexec. */ |
| 61 | # undef static |
| 62 | #endif |
| 63 | |
| 64 | #include <stdio.h> |
| 65 | #include <ctype.h> |
| 66 | #include <errno.h> |
| 67 | #ifndef errno |
| 68 | extern int errno; |
| 69 | #endif |
| 70 | #include <sys/types.h> |
| 71 | #include <sys/stat.h> |
| 72 | |
| 73 | #if !defined (S_ISREG) && defined (S_IFREG) |
| 74 | # define S_ISREG(m) (((m) & S_IFMT) == S_IFREG) |
| 75 | #endif |
| 76 | |
| 77 | #include <getopt.h> |
| 78 | |
| 79 | #ifdef ETAGS_REGEXPS |
| 80 | # include <regex.h> |
| 81 | #endif /* ETAGS_REGEXPS */ |
| 82 | |
| 83 | /* Define CTAGS to make the program "ctags" compatible with the usual one. |
| 84 | Let it undefined to make the program "etags", which makes emacs-style |
| 85 | tag tables and tags typedefs, #defines and struct/union/enum by default. */ |
| 86 | #ifdef CTAGS |
| 87 | # undef CTAGS |
| 88 | # define CTAGS TRUE |
| 89 | #else |
| 90 | # define CTAGS FALSE |
| 91 | #endif |
| 92 | |
| 93 | /* Exit codes for success and failure. */ |
| 94 | #ifdef VMS |
| 95 | # define GOOD 1 |
| 96 | # define BAD 0 |
| 97 | #else |
| 98 | # define GOOD 0 |
| 99 | # define BAD 1 |
| 100 | #endif |
| 101 | |
| 102 | /* C extensions. */ |
| 103 | #define C_PLPL 0x00001 /* C++ */ |
| 104 | #define C_STAR 0x00003 /* C* */ |
| 105 | #define YACC 0x10000 /* yacc file */ |
| 106 | |
| 107 | #define streq(s,t) ((DEBUG &&!(s)&&!(t)&&(abort(),1)) || !strcmp(s,t)) |
| 108 | #define strneq(s,t,n) ((DEBUG &&!(s)&&!(t)&&(abort(),1)) || !strncmp(s,t,n)) |
| 109 | |
| 110 | #define lowcase(c) tolower ((unsigned char)c) |
| 111 | |
| 112 | #define iswhite(arg) (_wht[arg]) /* T if char is white */ |
| 113 | #define begtoken(arg) (_btk[arg]) /* T if char can start token */ |
| 114 | #define intoken(arg) (_itk[arg]) /* T if char can be in token */ |
| 115 | #define endtoken(arg) (_etk[arg]) /* T if char ends tokens */ |
| 116 | |
| 117 | #ifdef DOS_NT |
| 118 | # define absolutefn(fn) (fn[0] == '/' \ |
| 119 | || (fn[1] == ':' && fn[2] == '/')) |
| 120 | #else |
| 121 | # define absolutefn(fn) (fn[0] == '/') |
| 122 | #endif |
| 123 | |
| 124 | |
| 125 | /* |
| 126 | * xnew -- allocate storage |
| 127 | * |
| 128 | * SYNOPSIS: Type *xnew (int n, Type); |
| 129 | */ |
| 130 | #define xnew(n,Type) ((Type *) xmalloc ((n) * sizeof (Type))) |
| 131 | |
| 132 | typedef int logical; |
| 133 | |
| 134 | typedef struct nd_st |
| 135 | { /* sorting structure */ |
| 136 | char *name; /* function or type name */ |
| 137 | char *file; /* file name */ |
| 138 | logical is_func; /* use pattern or line no */ |
| 139 | logical been_warned; /* set if noticed dup */ |
| 140 | int lno; /* line number tag is on */ |
| 141 | long cno; /* character number line starts on */ |
| 142 | char *pat; /* search pattern */ |
| 143 | struct nd_st *left, *right; /* left and right sons */ |
| 144 | } NODE; |
| 145 | |
| 146 | extern char *getenv (); |
| 147 | |
| 148 | char *concat (); |
| 149 | char *savenstr (), *savestr (); |
| 150 | char *etags_strchr (), *etags_strrchr (); |
| 151 | char *etags_getcwd (); |
| 152 | char *relative_filename (), *absolute_filename (), *absolute_dirname (); |
| 153 | long *xmalloc (), *xrealloc (); |
| 154 | |
| 155 | typedef void Lang_function (); |
| 156 | #if FALSE /* many compilers barf on this */ |
| 157 | Lang_function Asm_labels; |
| 158 | Lang_function default_C_entries; |
| 159 | Lang_function C_entries; |
| 160 | Lang_function Cplusplus_entries; |
| 161 | Lang_function Cstar_entries; |
| 162 | Lang_function Erlang_functions; |
| 163 | Lang_function Fortran_functions; |
| 164 | Lang_function Yacc_entries; |
| 165 | Lang_function Lisp_functions; |
| 166 | Lang_function Pascal_functions; |
| 167 | Lang_function Perl_functions; |
| 168 | Lang_function Prolog_functions; |
| 169 | Lang_function Scheme_functions; |
| 170 | Lang_function TeX_functions; |
| 171 | Lang_function just_read_file; |
| 172 | #else /* so let's write it this way */ |
| 173 | void Asm_labels (); |
| 174 | void C_entries (); |
| 175 | void default_C_entries (); |
| 176 | void plain_C_entries (); |
| 177 | void Cplusplus_entries (); |
| 178 | void Cstar_entries (); |
| 179 | void Erlang_functions (); |
| 180 | void Fortran_functions (); |
| 181 | void Yacc_entries (); |
| 182 | void Lisp_functions (); |
| 183 | void Pascal_functions (); |
| 184 | void Perl_functions (); |
| 185 | void Prolog_functions (); |
| 186 | void Scheme_functions (); |
| 187 | void TeX_functions (); |
| 188 | void just_read_file (); |
| 189 | #endif |
| 190 | |
| 191 | Lang_function *get_language_from_name (); |
| 192 | Lang_function *get_language_from_interpreter (); |
| 193 | Lang_function *get_language_from_suffix (); |
| 194 | int total_size_of_entries (); |
| 195 | long readline (); |
| 196 | long readline_internal (); |
| 197 | #ifdef ETAGS_REGEXPS |
| 198 | void add_regex (); |
| 199 | #endif |
| 200 | void add_node (); |
| 201 | void error (); |
| 202 | void suggest_asking_for_help (); |
| 203 | void fatal (), pfatal (); |
| 204 | void find_entries (); |
| 205 | void free_tree (); |
| 206 | void getit (); |
| 207 | void init (); |
| 208 | void initbuffer (); |
| 209 | void pfnote (); |
| 210 | void process_file (); |
| 211 | void put_entries (); |
| 212 | void takeprec (); |
| 213 | |
| 214 | \f |
| 215 | char searchar = '/'; /* use /.../ searches */ |
| 216 | |
| 217 | int lineno; /* line number of current line */ |
| 218 | long charno; /* current character number */ |
| 219 | long linecharno; /* charno of start of line */ |
| 220 | |
| 221 | char *curfile; /* current input file name */ |
| 222 | char *tagfile; /* output file */ |
| 223 | char *progname; /* name this program was invoked with */ |
| 224 | char *cwd; /* current working directory */ |
| 225 | char *tagfiledir; /* directory of tagfile */ |
| 226 | |
| 227 | FILE *tagf; /* ioptr for tags file */ |
| 228 | NODE *head; /* the head of the binary tree of tags */ |
| 229 | |
| 230 | /* |
| 231 | * A `struct linebuffer' is a structure which holds a line of text. |
| 232 | * `readline' reads a line from a stream into a linebuffer and works |
| 233 | * regardless of the length of the line. |
| 234 | */ |
| 235 | #define GROW_LINEBUFFER(buf,toksize) \ |
| 236 | while (buf.size < toksize) \ |
| 237 | buf.buffer = (char *) xrealloc (buf.buffer, buf.size *= 2) |
| 238 | struct linebuffer |
| 239 | { |
| 240 | long size; |
| 241 | char *buffer; |
| 242 | }; |
| 243 | |
| 244 | struct linebuffer lb; /* the current line */ |
| 245 | struct linebuffer token_name; /* used by C_entries as a temporary area */ |
| 246 | struct |
| 247 | { |
| 248 | long linepos; |
| 249 | struct linebuffer lb; /* used by C_entries instead of lb */ |
| 250 | } lbs[2]; |
| 251 | |
| 252 | /* boolean "functions" (see init) */ |
| 253 | logical _wht[0177], _etk[0177], _itk[0177], _btk[0177]; |
| 254 | char |
| 255 | /* white chars */ |
| 256 | *white = " \f\t\n\013", |
| 257 | /* token ending chars */ |
| 258 | *endtk = " \t\n\013\"'#()[]{}=-+%*/&|^~!<>;,.:?", |
| 259 | /* token starting chars */ |
| 260 | *begtk = "ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz$~@", |
| 261 | /* valid in-token chars */ |
| 262 | *intk = "ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz$0123456789"; |
| 263 | |
| 264 | logical append_to_tagfile; /* -a: append to tags */ |
| 265 | /* The following three default to TRUE for etags, but to FALSE for ctags. */ |
| 266 | logical typedefs; /* -t: create tags for typedefs */ |
| 267 | logical typedefs_and_cplusplus; /* -T: create tags for typedefs, level */ |
| 268 | /* 0 struct/enum/union decls, and C++ */ |
| 269 | /* member functions. */ |
| 270 | logical constantypedefs; /* -d: create tags for C #define and enum */ |
| 271 | /* constants. Enum consts not implemented. */ |
| 272 | /* -D: opposite of -d. Default under ctags. */ |
| 273 | logical update; /* -u: update tags */ |
| 274 | logical vgrind_style; /* -v: create vgrind style index output */ |
| 275 | logical no_warnings; /* -w: suppress warnings */ |
| 276 | logical cxref_style; /* -x: create cxref style output */ |
| 277 | logical cplusplus; /* .[hc] means C++, not C */ |
| 278 | logical noindentypedefs; /* -I: ignore indentation in C */ |
| 279 | |
| 280 | struct option longopts[] = |
| 281 | { |
| 282 | { "append", no_argument, NULL, 'a' }, |
| 283 | { "backward-search", no_argument, NULL, 'B' }, |
| 284 | { "c++", no_argument, NULL, 'C' }, |
| 285 | { "cxref", no_argument, NULL, 'x' }, |
| 286 | { "defines", no_argument, NULL, 'd' }, |
| 287 | { "help", no_argument, NULL, 'h' }, |
| 288 | { "help", no_argument, NULL, 'H' }, |
| 289 | { "ignore-indentation", no_argument, NULL, 'I' }, |
| 290 | { "include", required_argument, NULL, 'i' }, |
| 291 | { "language", required_argument, NULL, 'l' }, |
| 292 | { "no-defines", no_argument, NULL, 'D' }, |
| 293 | { "no-regex", no_argument, NULL, 'R' }, |
| 294 | { "no-warn", no_argument, NULL, 'w' }, |
| 295 | { "output", required_argument, NULL, 'o' }, |
| 296 | { "regex", required_argument, NULL, 'r' }, |
| 297 | { "typedefs", no_argument, NULL, 't' }, |
| 298 | { "typedefs-and-c++", no_argument, NULL, 'T' }, |
| 299 | { "update", no_argument, NULL, 'u' }, |
| 300 | { "version", no_argument, NULL, 'V' }, |
| 301 | { "vgrind", no_argument, NULL, 'v' }, |
| 302 | { 0 } |
| 303 | }; |
| 304 | |
| 305 | #ifdef ETAGS_REGEXPS |
| 306 | /* Structure defining a regular expression. Elements are |
| 307 | the compiled pattern, and the name string. */ |
| 308 | struct pattern |
| 309 | { |
| 310 | struct re_pattern_buffer *pattern; |
| 311 | struct re_registers regs; |
| 312 | char *name_pattern; |
| 313 | logical error_signaled; |
| 314 | }; |
| 315 | |
| 316 | /* Number of regexps found. */ |
| 317 | int num_patterns = 0; |
| 318 | |
| 319 | /* Array of all regexps. */ |
| 320 | struct pattern *patterns = NULL; |
| 321 | #endif /* ETAGS_REGEXPS */ |
| 322 | |
| 323 | /* |
| 324 | * Language stuff. |
| 325 | */ |
| 326 | |
| 327 | /* Non-NULL if language fixed. */ |
| 328 | Lang_function *lang_func = NULL; |
| 329 | |
| 330 | /* Assembly code */ |
| 331 | char *Asm_suffixes [] = { "a", /* Unix assembler */ |
| 332 | "asm", /* Microcontroller assembly */ |
| 333 | "def", /* BSO/Tasking definition includes */ |
| 334 | "inc", /* Microcontroller include files */ |
| 335 | "ins", /* Microcontroller include files */ |
| 336 | "s", "sa", /* Unix assembler */ |
| 337 | "src", /* BSO/Tasking C compiler output */ |
| 338 | NULL |
| 339 | }; |
| 340 | |
| 341 | /* Note that .c and .h can be considered C++, if the --c++ flag was |
| 342 | given. That is why default_C_entries is called here. */ |
| 343 | char *default_C_suffixes [] = |
| 344 | { "c", "h", NULL }; |
| 345 | |
| 346 | /* .M is for Objective C++ files. */ |
| 347 | char *Cplusplus_suffixes [] = |
| 348 | { "C", "H", "c++", "cc", "cpp", "cxx", "h++", "hh", "hpp", "hxx", "M", NULL}; |
| 349 | |
| 350 | char *Cstar_suffixes [] = |
| 351 | { "cs", "hs", NULL }; |
| 352 | |
| 353 | char *Erlang_suffixes [] = |
| 354 | { "erl", "hrl", NULL }; |
| 355 | |
| 356 | char *Fortran_suffixes [] = |
| 357 | { "F", "f", "f90", "for", NULL }; |
| 358 | |
| 359 | char *Lisp_suffixes [] = |
| 360 | { "cl", "clisp", "el", "l", "lisp", "lsp", "ml", NULL }; |
| 361 | |
| 362 | char *Pascal_suffixes [] = |
| 363 | { "p", "pas", NULL }; |
| 364 | |
| 365 | char *Perl_suffixes [] = |
| 366 | { "pl", "pm", NULL }; |
| 367 | char *Perl_interpreters [] = |
| 368 | { "perl", "@PERL@", NULL }; |
| 369 | |
| 370 | char *plain_C_suffixes [] = |
| 371 | { "pc", /* Pro*C file */ |
| 372 | "m", /* Objective C file */ |
| 373 | "lm", /* Objective lex file */ |
| 374 | NULL }; |
| 375 | |
| 376 | char *Prolog_suffixes [] = |
| 377 | { "prolog", NULL }; |
| 378 | |
| 379 | /* Can't do the `SCM' or `scm' prefix with a version number. */ |
| 380 | char *Scheme_suffixes [] = |
| 381 | { "SCM", "SM", "oak", "sch", "scheme", "scm", "sm", "t", NULL }; |
| 382 | |
| 383 | char *TeX_suffixes [] = |
| 384 | { "TeX", "bib", "clo", "cls", "ltx", "sty", "tex", NULL }; |
| 385 | |
| 386 | char *Yacc_suffixes [] = |
| 387 | { "y", "ym", NULL }; /* .ym is Objective yacc file */ |
| 388 | |
| 389 | /* Table of language names and corresponding functions, file suffixes |
| 390 | and interpreter names. |
| 391 | It is ok for a given function to be listed under more than one |
| 392 | name. I just didn't. */ |
| 393 | struct lang_entry |
| 394 | { |
| 395 | char *name; |
| 396 | Lang_function *function; |
| 397 | char **suffixes; |
| 398 | char **interpreters; |
| 399 | }; |
| 400 | |
| 401 | struct lang_entry lang_names [] = |
| 402 | { |
| 403 | { "asm", Asm_labels, Asm_suffixes, NULL }, |
| 404 | { "c", default_C_entries, default_C_suffixes, NULL }, |
| 405 | { "c++", Cplusplus_entries, Cplusplus_suffixes, NULL }, |
| 406 | { "c*", Cstar_entries, Cstar_suffixes, NULL }, |
| 407 | { "erlang", Erlang_functions, Erlang_suffixes, NULL }, |
| 408 | { "fortran", Fortran_functions, Fortran_suffixes, NULL }, |
| 409 | { "lisp", Lisp_functions, Lisp_suffixes, NULL }, |
| 410 | { "pascal", Pascal_functions, Pascal_suffixes, NULL }, |
| 411 | { "perl", Perl_functions, Perl_suffixes, Perl_interpreters }, |
| 412 | { "proc", plain_C_entries, plain_C_suffixes, NULL }, |
| 413 | { "prolog", Prolog_functions, Prolog_suffixes, NULL }, |
| 414 | { "scheme", Scheme_functions, Scheme_suffixes, NULL }, |
| 415 | { "tex", TeX_functions, TeX_suffixes, NULL }, |
| 416 | { "yacc", Yacc_entries, Yacc_suffixes, NULL }, |
| 417 | { "auto", NULL }, /* default guessing scheme */ |
| 418 | { "none", just_read_file }, /* regexp matching only */ |
| 419 | { NULL, NULL } /* end of list */ |
| 420 | }; |
| 421 | |
| 422 | \f |
| 423 | void |
| 424 | print_language_names () |
| 425 | { |
| 426 | struct lang_entry *lang; |
| 427 | char **ext; |
| 428 | |
| 429 | puts ("\nThese are the currently supported languages, along with the\n\ |
| 430 | default file name suffixes:"); |
| 431 | for (lang = lang_names; lang->name != NULL; lang++) |
| 432 | { |
| 433 | printf ("\t%s\t", lang->name); |
| 434 | if (lang->suffixes != NULL) |
| 435 | for (ext = lang->suffixes; *ext != NULL; ext++) |
| 436 | printf (" .%s", *ext); |
| 437 | puts (""); |
| 438 | } |
| 439 | puts ("Where `auto' means use default language for files based on file\n\ |
| 440 | name suffix, and `none' means only do regexp processing on files.\n\ |
| 441 | If no language is specified and no matching suffix is found,\n\ |
| 442 | the first line of the file is read for a sharp-bang (#!) sequence\n\ |
| 443 | followed by the name of an interpreter. If no such sequence is found,\n\ |
| 444 | Fortran is tried first; if no tags are found, C is tried next."); |
| 445 | } |
| 446 | |
| 447 | #ifndef VERSION |
| 448 | # define VERSION "19" |
| 449 | #endif |
| 450 | void |
| 451 | print_version () |
| 452 | { |
| 453 | printf ("%s (GNU Emacs %s)\n", (CTAGS) ? "ctags" : "etags", VERSION); |
| 454 | puts ("Copyright (C) 1996 Free Software Foundation, Inc. and Ken Arnold"); |
| 455 | puts ("This program is distributed under the same terms as Emacs"); |
| 456 | |
| 457 | exit (GOOD); |
| 458 | } |
| 459 | |
| 460 | void |
| 461 | print_help () |
| 462 | { |
| 463 | printf ("These are the options accepted by %s. You may use unambiguous\n\ |
| 464 | abbreviations for the long option names. A - as file name means read\n\ |
| 465 | names from stdin.", progname); |
| 466 | if (!CTAGS) |
| 467 | printf (" Absolute names are stored in the output file as they\n\ |
| 468 | are. Relative ones are stored relative to the output file's directory."); |
| 469 | puts ("\n"); |
| 470 | |
| 471 | puts ("-a, --append\n\ |
| 472 | Append tag entries to existing tags file."); |
| 473 | |
| 474 | if (CTAGS) |
| 475 | puts ("-B, --backward-search\n\ |
| 476 | Write the search commands for the tag entries using '?', the\n\ |
| 477 | backward-search command instead of '/', the forward-search command."); |
| 478 | |
| 479 | puts ("-C, --c++\n\ |
| 480 | Treat files whose name suffix defaults to C language as C++ files."); |
| 481 | |
| 482 | if (CTAGS) |
| 483 | puts ("-d, --defines\n\ |
| 484 | Create tag entries for constant C #defines, too."); |
| 485 | else |
| 486 | puts ("-D, --no-defines\n\ |
| 487 | Don't create tag entries for constant C #defines. This makes\n\ |
| 488 | the tags file smaller."); |
| 489 | |
| 490 | if (!CTAGS) |
| 491 | { |
| 492 | puts ("-i FILE, --include=FILE\n\ |
| 493 | Include a note in tag file indicating that, when searching for\n\ |
| 494 | a tag, one should also consult the tags file FILE after\n\ |
| 495 | checking the current file."); |
| 496 | puts ("-l LANG, --language=LANG\n\ |
| 497 | Force the following files to be considered as written in the\n\ |
| 498 | named language up to the next --language=LANG option."); |
| 499 | } |
| 500 | |
| 501 | #ifdef ETAGS_REGEXPS |
| 502 | puts ("-r /REGEXP/, --regex=/REGEXP/\n\ |
| 503 | Make a tag for each line matching pattern REGEXP in the\n\ |
| 504 | following files. REGEXP is anchored (as if preceded by ^).\n\ |
| 505 | The form /REGEXP/NAME/ creates a named tag. For example Tcl\n\ |
| 506 | named tags can be created with:\n\ |
| 507 | --regex=/proc[ \\t]+\\([^ \\t]+\\)/\\1/."); |
| 508 | puts ("-R, --no-regex\n\ |
| 509 | Don't create tags from regexps for the following files."); |
| 510 | #endif /* ETAGS_REGEXPS */ |
| 511 | puts ("-o FILE, --output=FILE\n\ |
| 512 | Write the tags to FILE."); |
| 513 | puts ("-I, --ignore-indentation\n\ |
| 514 | Don't rely on indentation quite as much as normal. Currently,\n\ |
| 515 | this means not to assume that a closing brace in the first\n\ |
| 516 | column is the final brace of a function or structure\n\ |
| 517 | definition in C and C++."); |
| 518 | |
| 519 | if (CTAGS) |
| 520 | { |
| 521 | puts ("-t, --typedefs\n\ |
| 522 | Generate tag entries for C typedefs."); |
| 523 | puts ("-T, --typedefs-and-c++\n\ |
| 524 | Generate tag entries for C typedefs, C struct/enum/union tags,\n\ |
| 525 | and C++ member functions."); |
| 526 | puts ("-u, --update\n\ |
| 527 | Update the tag entries for the given files, leaving tag\n\ |
| 528 | entries for other files in place. Currently, this is\n\ |
| 529 | implemented by deleting the existing entries for the given\n\ |
| 530 | files and then rewriting the new entries at the end of the\n\ |
| 531 | tags file. It is often faster to simply rebuild the entire\n\ |
| 532 | tag file than to use this."); |
| 533 | puts ("-v, --vgrind\n\ |
| 534 | Generates an index of items intended for human consumption,\n\ |
| 535 | similar to the output of vgrind. The index is sorted, and\n\ |
| 536 | gives the page number of each item."); |
| 537 | puts ("-w, --no-warn\n\ |
| 538 | Suppress warning messages about entries defined in multiple\n\ |
| 539 | files."); |
| 540 | puts ("-x, --cxref\n\ |
| 541 | Like --vgrind, but in the style of cxref, rather than vgrind.\n\ |
| 542 | The output uses line numbers instead of page numbers, but\n\ |
| 543 | beyond that the differences are cosmetic; try both to see\n\ |
| 544 | which you like."); |
| 545 | } |
| 546 | |
| 547 | puts ("-V, --version\n\ |
| 548 | Print the version of the program.\n\ |
| 549 | -h, --help\n\ |
| 550 | Print this help message."); |
| 551 | |
| 552 | print_language_names (); |
| 553 | |
| 554 | puts (""); |
| 555 | puts ("Report bugs to bug-gnu-emacs@prep.ai.mit.edu"); |
| 556 | |
| 557 | exit (GOOD); |
| 558 | } |
| 559 | |
| 560 | \f |
| 561 | enum argument_type |
| 562 | { |
| 563 | at_language, |
| 564 | at_regexp, |
| 565 | at_filename |
| 566 | }; |
| 567 | |
| 568 | /* This structure helps us allow mixing of --lang and filenames. */ |
| 569 | typedef struct |
| 570 | { |
| 571 | enum argument_type arg_type; |
| 572 | char *what; |
| 573 | Lang_function *function; |
| 574 | } argument; |
| 575 | |
| 576 | #ifdef VMS /* VMS specific functions */ |
| 577 | |
| 578 | #define EOS '\0' |
| 579 | |
| 580 | /* This is a BUG! ANY arbitrary limit is a BUG! |
| 581 | Won't someone please fix this? */ |
| 582 | #define MAX_FILE_SPEC_LEN 255 |
| 583 | typedef struct { |
| 584 | short curlen; |
| 585 | char body[MAX_FILE_SPEC_LEN + 1]; |
| 586 | } vspec; |
| 587 | |
| 588 | /* |
| 589 | v1.05 nmm 26-Jun-86 fn_exp - expand specification of list of file names |
| 590 | returning in each successive call the next filename matching the input |
| 591 | spec. The function expects that each in_spec passed |
| 592 | to it will be processed to completion; in particular, up to and |
| 593 | including the call following that in which the last matching name |
| 594 | is returned, the function ignores the value of in_spec, and will |
| 595 | only start processing a new spec with the following call. |
| 596 | If an error occurs, on return out_spec contains the value |
| 597 | of in_spec when the error occurred. |
| 598 | |
| 599 | With each successive filename returned in out_spec, the |
| 600 | function's return value is one. When there are no more matching |
| 601 | names the function returns zero. If on the first call no file |
| 602 | matches in_spec, or there is any other error, -1 is returned. |
| 603 | */ |
| 604 | |
| 605 | #include <rmsdef.h> |
| 606 | #include <descrip.h> |
| 607 | #define OUTSIZE MAX_FILE_SPEC_LEN |
| 608 | short |
| 609 | fn_exp (out, in) |
| 610 | vspec *out; |
| 611 | char *in; |
| 612 | { |
| 613 | static long context = 0; |
| 614 | static struct dsc$descriptor_s o; |
| 615 | static struct dsc$descriptor_s i; |
| 616 | static logical pass1 = TRUE; |
| 617 | long status; |
| 618 | short retval; |
| 619 | |
| 620 | if (pass1) |
| 621 | { |
| 622 | pass1 = FALSE; |
| 623 | o.dsc$a_pointer = (char *) out; |
| 624 | o.dsc$w_length = (short)OUTSIZE; |
| 625 | i.dsc$a_pointer = in; |
| 626 | i.dsc$w_length = (short)strlen(in); |
| 627 | i.dsc$b_dtype = DSC$K_DTYPE_T; |
| 628 | i.dsc$b_class = DSC$K_CLASS_S; |
| 629 | o.dsc$b_dtype = DSC$K_DTYPE_VT; |
| 630 | o.dsc$b_class = DSC$K_CLASS_VS; |
| 631 | } |
| 632 | if ((status = lib$find_file(&i, &o, &context, 0, 0)) == RMS$_NORMAL) |
| 633 | { |
| 634 | out->body[out->curlen] = EOS; |
| 635 | return 1; |
| 636 | } |
| 637 | else if (status == RMS$_NMF) |
| 638 | retval = 0; |
| 639 | else |
| 640 | { |
| 641 | strcpy(out->body, in); |
| 642 | retval = -1; |
| 643 | } |
| 644 | lib$find_file_end(&context); |
| 645 | pass1 = TRUE; |
| 646 | return retval; |
| 647 | } |
| 648 | |
| 649 | /* |
| 650 | v1.01 nmm 19-Aug-85 gfnames - return in successive calls the |
| 651 | name of each file specified by the provided arg expanding wildcards. |
| 652 | */ |
| 653 | char * |
| 654 | gfnames (arg, p_error) |
| 655 | char *arg; |
| 656 | logical *p_error; |
| 657 | { |
| 658 | static vspec filename = {MAX_FILE_SPEC_LEN, "\0"}; |
| 659 | |
| 660 | switch (fn_exp (&filename, arg)) |
| 661 | { |
| 662 | case 1: |
| 663 | *p_error = FALSE; |
| 664 | return filename.body; |
| 665 | case 0: |
| 666 | *p_error = FALSE; |
| 667 | return NULL; |
| 668 | default: |
| 669 | *p_error = TRUE; |
| 670 | return filename.body; |
| 671 | } |
| 672 | } |
| 673 | |
| 674 | #ifndef OLD /* Newer versions of VMS do provide `system'. */ |
| 675 | system (cmd) |
| 676 | char *cmd; |
| 677 | { |
| 678 | fprintf (stderr, "system() function not implemented under VMS\n"); |
| 679 | } |
| 680 | #endif |
| 681 | |
| 682 | #define VERSION_DELIM ';' |
| 683 | char *massage_name (s) |
| 684 | char *s; |
| 685 | { |
| 686 | char *start = s; |
| 687 | |
| 688 | for ( ; *s; s++) |
| 689 | if (*s == VERSION_DELIM) |
| 690 | { |
| 691 | *s = EOS; |
| 692 | break; |
| 693 | } |
| 694 | else |
| 695 | *s = lowcase (*s); |
| 696 | return start; |
| 697 | } |
| 698 | #endif /* VMS */ |
| 699 | |
| 700 | \f |
| 701 | int |
| 702 | main (argc, argv) |
| 703 | int argc; |
| 704 | char *argv[]; |
| 705 | { |
| 706 | int i; |
| 707 | unsigned int nincluded_files = 0; |
| 708 | char **included_files = xnew (argc, char *); |
| 709 | char *this_file; |
| 710 | argument *argbuffer; |
| 711 | int current_arg = 0, file_count = 0; |
| 712 | struct linebuffer filename_lb; |
| 713 | #ifdef VMS |
| 714 | logical got_err; |
| 715 | #endif |
| 716 | |
| 717 | #ifdef DOS_NT |
| 718 | _fmode = O_BINARY; /* all of files are treated as binary files */ |
| 719 | #endif /* DOS_NT */ |
| 720 | |
| 721 | progname = argv[0]; |
| 722 | |
| 723 | /* Allocate enough no matter what happens. Overkill, but each one |
| 724 | is small. */ |
| 725 | argbuffer = xnew (argc, argument); |
| 726 | |
| 727 | #ifdef ETAGS_REGEXPS |
| 728 | /* Set syntax for regular expression routines. */ |
| 729 | re_set_syntax (RE_SYNTAX_EMACS); |
| 730 | #endif /* ETAGS_REGEXPS */ |
| 731 | |
| 732 | /* |
| 733 | * If etags, always find typedefs and structure tags. Why not? |
| 734 | * Also default is to find macro constants. |
| 735 | */ |
| 736 | if (!CTAGS) |
| 737 | typedefs = typedefs_and_cplusplus = constantypedefs = TRUE; |
| 738 | |
| 739 | while (1) |
| 740 | { |
| 741 | int opt = getopt_long (argc, argv, |
| 742 | "-aCdDf:Il:o:r:RStTi:BuvxwVhH", longopts, 0); |
| 743 | |
| 744 | if (opt == EOF) |
| 745 | break; |
| 746 | |
| 747 | switch (opt) |
| 748 | { |
| 749 | case 0: |
| 750 | /* If getopt returns 0, then it has already processed a |
| 751 | long-named option. We should do nothing. */ |
| 752 | break; |
| 753 | |
| 754 | case 1: |
| 755 | /* This means that a filename has been seen. Record it. */ |
| 756 | argbuffer[current_arg].arg_type = at_filename; |
| 757 | argbuffer[current_arg].what = optarg; |
| 758 | ++current_arg; |
| 759 | ++file_count; |
| 760 | break; |
| 761 | |
| 762 | /* Common options. */ |
| 763 | case 'a': |
| 764 | append_to_tagfile = TRUE; |
| 765 | break; |
| 766 | case 'C': |
| 767 | cplusplus = TRUE; |
| 768 | break; |
| 769 | case 'd': |
| 770 | constantypedefs = TRUE; |
| 771 | break; |
| 772 | case 'D': |
| 773 | constantypedefs = FALSE; |
| 774 | break; |
| 775 | case 'f': /* for compatibility with old makefiles */ |
| 776 | case 'o': |
| 777 | if (tagfile) |
| 778 | { |
| 779 | fprintf (stderr, "%s: -%c option may only be given once.\n", |
| 780 | progname, opt); |
| 781 | suggest_asking_for_help (); |
| 782 | } |
| 783 | tagfile = optarg; |
| 784 | break; |
| 785 | case 'I': |
| 786 | case 'S': /* for backward compatibility */ |
| 787 | noindentypedefs = TRUE; |
| 788 | break; |
| 789 | case 'l': |
| 790 | argbuffer[current_arg].function = get_language_from_name (optarg); |
| 791 | argbuffer[current_arg].arg_type = at_language; |
| 792 | ++current_arg; |
| 793 | break; |
| 794 | #ifdef ETAGS_REGEXPS |
| 795 | case 'r': |
| 796 | argbuffer[current_arg].arg_type = at_regexp; |
| 797 | argbuffer[current_arg].what = optarg; |
| 798 | ++current_arg; |
| 799 | break; |
| 800 | case 'R': |
| 801 | argbuffer[current_arg].arg_type = at_regexp; |
| 802 | argbuffer[current_arg].what = NULL; |
| 803 | ++current_arg; |
| 804 | break; |
| 805 | #endif /* ETAGS_REGEXPS */ |
| 806 | case 'V': |
| 807 | print_version (); |
| 808 | break; |
| 809 | case 'h': |
| 810 | case 'H': |
| 811 | print_help (); |
| 812 | break; |
| 813 | case 't': |
| 814 | typedefs = TRUE; |
| 815 | break; |
| 816 | case 'T': |
| 817 | typedefs = typedefs_and_cplusplus = TRUE; |
| 818 | break; |
| 819 | #if (!CTAGS) |
| 820 | /* Etags options */ |
| 821 | case 'i': |
| 822 | included_files[nincluded_files++] = optarg; |
| 823 | break; |
| 824 | #else /* CTAGS */ |
| 825 | /* Ctags options. */ |
| 826 | case 'B': |
| 827 | searchar = '?'; |
| 828 | break; |
| 829 | case 'u': |
| 830 | update = TRUE; |
| 831 | break; |
| 832 | case 'v': |
| 833 | vgrind_style = TRUE; |
| 834 | /*FALLTHRU*/ |
| 835 | case 'x': |
| 836 | cxref_style = TRUE; |
| 837 | break; |
| 838 | case 'w': |
| 839 | no_warnings = TRUE; |
| 840 | break; |
| 841 | #endif /* CTAGS */ |
| 842 | default: |
| 843 | suggest_asking_for_help (); |
| 844 | } |
| 845 | } |
| 846 | |
| 847 | for (; optind < argc; ++optind) |
| 848 | { |
| 849 | argbuffer[current_arg].arg_type = at_filename; |
| 850 | argbuffer[current_arg].what = argv[optind]; |
| 851 | ++current_arg; |
| 852 | ++file_count; |
| 853 | } |
| 854 | |
| 855 | if (nincluded_files == 0 && file_count == 0) |
| 856 | { |
| 857 | fprintf (stderr, "%s: No input files specified.\n", progname); |
| 858 | suggest_asking_for_help (); |
| 859 | } |
| 860 | |
| 861 | if (tagfile == NULL) |
| 862 | tagfile = CTAGS ? "tags" : "TAGS"; |
| 863 | cwd = etags_getcwd (); /* the current working directory */ |
| 864 | if (cwd[strlen (cwd) - 1] != '/') |
| 865 | cwd = concat (cwd, "/", ""); |
| 866 | if (streq (tagfile, "-")) |
| 867 | tagfiledir = cwd; |
| 868 | else |
| 869 | tagfiledir = absolute_dirname (tagfile, cwd); |
| 870 | |
| 871 | init (); /* set up boolean "functions" */ |
| 872 | |
| 873 | initbuffer (&lb); |
| 874 | initbuffer (&token_name); |
| 875 | initbuffer (&lbs[0].lb); |
| 876 | initbuffer (&lbs[1].lb); |
| 877 | initbuffer (&filename_lb); |
| 878 | |
| 879 | if (!CTAGS) |
| 880 | { |
| 881 | if (streq (tagfile, "-")) |
| 882 | { |
| 883 | tagf = stdout; |
| 884 | #ifdef DOS_NT |
| 885 | /* Switch redirected `stdout' to binary mode (setting `_fmode' |
| 886 | doesn't take effect until after `stdout' is already open). */ |
| 887 | if (!isatty (fileno (stdout))) |
| 888 | setmode (fileno (stdout), O_BINARY); |
| 889 | #endif /* DOS_NT */ |
| 890 | } |
| 891 | else |
| 892 | tagf = fopen (tagfile, append_to_tagfile ? "a" : "w"); |
| 893 | if (tagf == NULL) |
| 894 | pfatal (tagfile); |
| 895 | } |
| 896 | |
| 897 | /* |
| 898 | * Loop through files finding functions. |
| 899 | */ |
| 900 | for (i = 0; i < current_arg; ++i) |
| 901 | { |
| 902 | switch (argbuffer[i].arg_type) |
| 903 | { |
| 904 | case at_language: |
| 905 | lang_func = argbuffer[i].function; |
| 906 | break; |
| 907 | #ifdef ETAGS_REGEXPS |
| 908 | case at_regexp: |
| 909 | add_regex (argbuffer[i].what); |
| 910 | break; |
| 911 | #endif |
| 912 | case at_filename: |
| 913 | #ifdef VMS |
| 914 | while ((this_file = gfnames (argbuffer[i].what, &got_err)) != NULL) |
| 915 | { |
| 916 | if (got_err) |
| 917 | { |
| 918 | error ("Can't find file %s\n", this_file); |
| 919 | argc--, argv++; |
| 920 | } |
| 921 | else |
| 922 | { |
| 923 | this_file = massage_name (this_file); |
| 924 | } |
| 925 | #else |
| 926 | this_file = argbuffer[i].what; |
| 927 | #endif |
| 928 | /* Input file named "-" means read file names from stdin |
| 929 | and use them. */ |
| 930 | if (streq (this_file, "-")) |
| 931 | while (readline_internal (&filename_lb, stdin) > 0) |
| 932 | process_file (filename_lb.buffer); |
| 933 | else |
| 934 | process_file (this_file); |
| 935 | #ifdef VMS |
| 936 | } |
| 937 | #endif |
| 938 | break; |
| 939 | } |
| 940 | } |
| 941 | |
| 942 | if (!CTAGS) |
| 943 | { |
| 944 | while (nincluded_files-- > 0) |
| 945 | fprintf (tagf, "\f\n%s,include\n", *included_files++); |
| 946 | |
| 947 | fclose (tagf); |
| 948 | exit (GOOD); |
| 949 | } |
| 950 | |
| 951 | /* If CTAGS, we are here. process_file did not write the tags yet, |
| 952 | because we want them ordered. Let's do it now. */ |
| 953 | if (cxref_style) |
| 954 | { |
| 955 | put_entries (head); |
| 956 | exit (GOOD); |
| 957 | } |
| 958 | |
| 959 | if (update) |
| 960 | { |
| 961 | char cmd[BUFSIZ]; |
| 962 | for (i = 0; i < current_arg; ++i) |
| 963 | { |
| 964 | if (argbuffer[i].arg_type != at_filename) |
| 965 | continue; |
| 966 | sprintf (cmd, |
| 967 | "mv %s OTAGS;fgrep -v '\t%s\t' OTAGS >%s;rm OTAGS", |
| 968 | tagfile, argbuffer[i].what, tagfile); |
| 969 | if (system (cmd) != GOOD) |
| 970 | fatal ("failed to execute shell command"); |
| 971 | } |
| 972 | append_to_tagfile = TRUE; |
| 973 | } |
| 974 | |
| 975 | tagf = fopen (tagfile, append_to_tagfile ? "a" : "w"); |
| 976 | if (tagf == NULL) |
| 977 | pfatal (tagfile); |
| 978 | put_entries (head); |
| 979 | fclose (tagf); |
| 980 | |
| 981 | if (update) |
| 982 | { |
| 983 | char cmd[BUFSIZ]; |
| 984 | sprintf (cmd, "sort %s -o %s", tagfile, tagfile); |
| 985 | exit (system (cmd)); |
| 986 | } |
| 987 | return GOOD; |
| 988 | } |
| 989 | |
| 990 | |
| 991 | /* |
| 992 | * Return a Lang_function given the name. |
| 993 | */ |
| 994 | Lang_function * |
| 995 | get_language_from_name (name) |
| 996 | char *name; |
| 997 | { |
| 998 | struct lang_entry *lang; |
| 999 | |
| 1000 | if (name != NULL) |
| 1001 | for (lang = lang_names; lang->name != NULL; lang++) |
| 1002 | { |
| 1003 | if (streq (name, lang->name)) |
| 1004 | return lang->function; |
| 1005 | } |
| 1006 | |
| 1007 | fprintf (stderr, "%s: language \"%s\" not recognized.\n", |
| 1008 | progname, optarg); |
| 1009 | suggest_asking_for_help (); |
| 1010 | |
| 1011 | /* This point should never be reached. The function should either |
| 1012 | return a function pointer or never return. Note that a NULL |
| 1013 | pointer cannot be considered as an error, as it means that the |
| 1014 | language has not been explicitely imposed by the user ("auto"). */ |
| 1015 | return NULL; /* avoid warnings from compiler */ |
| 1016 | } |
| 1017 | |
| 1018 | |
| 1019 | /* |
| 1020 | * Return a Lang_function given the interpreter name. |
| 1021 | */ |
| 1022 | Lang_function * |
| 1023 | get_language_from_interpreter (interpreter) |
| 1024 | char *interpreter; |
| 1025 | { |
| 1026 | struct lang_entry *lang; |
| 1027 | char **iname; |
| 1028 | |
| 1029 | if (interpreter == NULL) |
| 1030 | return NULL; |
| 1031 | for (lang = lang_names; lang->name != NULL; lang++) |
| 1032 | if (lang->interpreters != NULL) |
| 1033 | for (iname = lang->interpreters; *iname != NULL; iname++) |
| 1034 | if (streq (*iname, interpreter)) |
| 1035 | return lang->function; |
| 1036 | |
| 1037 | return NULL; |
| 1038 | } |
| 1039 | |
| 1040 | |
| 1041 | |
| 1042 | /* |
| 1043 | * Return a Lang_function given the file suffix. |
| 1044 | */ |
| 1045 | Lang_function * |
| 1046 | get_language_from_suffix (suffix) |
| 1047 | char *suffix; |
| 1048 | { |
| 1049 | struct lang_entry *lang; |
| 1050 | char **ext; |
| 1051 | |
| 1052 | if (suffix == NULL) |
| 1053 | return NULL; |
| 1054 | for (lang = lang_names; lang->name != NULL; lang++) |
| 1055 | if (lang->suffixes != NULL) |
| 1056 | for (ext = lang->suffixes; *ext != NULL; ext++) |
| 1057 | if (streq (*ext, suffix)) |
| 1058 | return lang->function; |
| 1059 | |
| 1060 | return NULL; |
| 1061 | } |
| 1062 | |
| 1063 | |
| 1064 | /* |
| 1065 | * This routine is called on each file argument. |
| 1066 | */ |
| 1067 | void |
| 1068 | process_file (file) |
| 1069 | char *file; |
| 1070 | { |
| 1071 | struct stat stat_buf; |
| 1072 | FILE *inf; |
| 1073 | #ifdef DOS_NT |
| 1074 | char *p; |
| 1075 | |
| 1076 | for (p = file; *p != '\0'; p++) |
| 1077 | if (*p == '\\') |
| 1078 | *p = '/'; |
| 1079 | #endif |
| 1080 | |
| 1081 | if (stat (file, &stat_buf) == 0 && !S_ISREG (stat_buf.st_mode)) |
| 1082 | { |
| 1083 | fprintf (stderr, "Skipping %s: it is not a regular file.\n", file); |
| 1084 | return; |
| 1085 | } |
| 1086 | if (streq (file, tagfile) && !streq (tagfile, "-")) |
| 1087 | { |
| 1088 | fprintf (stderr, "Skipping inclusion of %s in self.\n", file); |
| 1089 | return; |
| 1090 | } |
| 1091 | inf = fopen (file, "r"); |
| 1092 | if (inf == NULL) |
| 1093 | { |
| 1094 | perror (file); |
| 1095 | return; |
| 1096 | } |
| 1097 | |
| 1098 | find_entries (file, inf); |
| 1099 | |
| 1100 | if (!CTAGS) |
| 1101 | { |
| 1102 | char *filename; |
| 1103 | |
| 1104 | if (absolutefn (file)) |
| 1105 | { |
| 1106 | /* file is an absolute filename. Canonicalise it. */ |
| 1107 | filename = absolute_filename (file, cwd); |
| 1108 | } |
| 1109 | else |
| 1110 | { |
| 1111 | /* file is a filename relative to cwd. Make it relative |
| 1112 | to the directory of the tags file. */ |
| 1113 | filename = relative_filename (file, tagfiledir); |
| 1114 | } |
| 1115 | fprintf (tagf, "\f\n%s,%d\n", filename, total_size_of_entries (head)); |
| 1116 | free (filename); |
| 1117 | put_entries (head); |
| 1118 | free_tree (head); |
| 1119 | head = NULL; |
| 1120 | } |
| 1121 | } |
| 1122 | |
| 1123 | /* |
| 1124 | * This routine sets up the boolean pseudo-functions which work |
| 1125 | * by setting boolean flags dependent upon the corresponding character |
| 1126 | * Every char which is NOT in that string is not a white char. Therefore, |
| 1127 | * all of the array "_wht" is set to FALSE, and then the elements |
| 1128 | * subscripted by the chars in "white" are set to TRUE. Thus "_wht" |
| 1129 | * of a char is TRUE if it is the string "white", else FALSE. |
| 1130 | */ |
| 1131 | void |
| 1132 | init () |
| 1133 | { |
| 1134 | register char *sp; |
| 1135 | register int i; |
| 1136 | |
| 1137 | for (i = 0; i < 0177; i++) |
| 1138 | _wht[i] = _etk[i] = _itk[i] = _btk[i] = FALSE; |
| 1139 | for (sp = white; *sp; sp++) |
| 1140 | _wht[*sp] = TRUE; |
| 1141 | for (sp = endtk; *sp; sp++) |
| 1142 | _etk[*sp] = TRUE; |
| 1143 | for (sp = intk; *sp; sp++) |
| 1144 | _itk[*sp] = TRUE; |
| 1145 | for (sp = begtk; *sp; sp++) |
| 1146 | _btk[*sp] = TRUE; |
| 1147 | _wht[0] = _wht['\n']; |
| 1148 | _etk[0] = _etk['\n']; |
| 1149 | _btk[0] = _btk['\n']; |
| 1150 | _itk[0] = _itk['\n']; |
| 1151 | } |
| 1152 | |
| 1153 | /* |
| 1154 | * This routine opens the specified file and calls the function |
| 1155 | * which finds the function and type definitions. |
| 1156 | */ |
| 1157 | void |
| 1158 | find_entries (file, inf) |
| 1159 | char *file; |
| 1160 | FILE *inf; |
| 1161 | { |
| 1162 | char *cp; |
| 1163 | Lang_function *function; |
| 1164 | NODE *old_last_node; |
| 1165 | extern NODE *last_node; |
| 1166 | |
| 1167 | |
| 1168 | /* Memory leakage here: the memory block pointed by curfile is never |
| 1169 | released. The amount of memory leaked here is the sum of the |
| 1170 | lengths of the input file names. */ |
| 1171 | curfile = savestr (file); |
| 1172 | |
| 1173 | /* If user specified a language, use it. */ |
| 1174 | function = lang_func; |
| 1175 | if (function != NULL) |
| 1176 | { |
| 1177 | function (inf); |
| 1178 | fclose (inf); |
| 1179 | return; |
| 1180 | } |
| 1181 | |
| 1182 | cp = etags_strrchr (file, '.'); |
| 1183 | if (cp != NULL) |
| 1184 | { |
| 1185 | cp += 1; |
| 1186 | function = get_language_from_suffix (cp); |
| 1187 | if (function != NULL) |
| 1188 | { |
| 1189 | function (inf); |
| 1190 | fclose (inf); |
| 1191 | return; |
| 1192 | } |
| 1193 | } |
| 1194 | |
| 1195 | /* Look for sharp-bang as the first two characters. */ |
| 1196 | if (readline_internal (&lb, inf) > 2 |
| 1197 | && lb.buffer[0] == '#' |
| 1198 | && lb.buffer[1] == '!') |
| 1199 | { |
| 1200 | char *lp; |
| 1201 | |
| 1202 | /* Set lp to point at the first char after the last slash in the |
| 1203 | line or, if no slashes, at the first nonblank. Then set cp to |
| 1204 | the first successive blank and terminate the string. */ |
| 1205 | lp = etags_strrchr (lb.buffer+2, '/'); |
| 1206 | if (lp != NULL) |
| 1207 | lp += 1; |
| 1208 | else |
| 1209 | for (lp = lb.buffer+2; *lp != '\0' && isspace (*lp); lp++) |
| 1210 | continue; |
| 1211 | for (cp = lp; *cp != '\0' && !isspace (*cp); cp++) |
| 1212 | continue; |
| 1213 | *cp = '\0'; |
| 1214 | |
| 1215 | if (strlen (lp) > 0) |
| 1216 | { |
| 1217 | function = get_language_from_interpreter (lp); |
| 1218 | if (function != NULL) |
| 1219 | { |
| 1220 | function (inf); |
| 1221 | fclose (inf); |
| 1222 | return; |
| 1223 | } |
| 1224 | } |
| 1225 | } |
| 1226 | rewind (inf); |
| 1227 | |
| 1228 | /* Try Fortran. */ |
| 1229 | old_last_node = last_node; |
| 1230 | Fortran_functions (inf); |
| 1231 | |
| 1232 | /* No Fortran entries found. Try C. */ |
| 1233 | if (old_last_node == last_node) |
| 1234 | { |
| 1235 | rewind (inf); |
| 1236 | default_C_entries (inf); |
| 1237 | } |
| 1238 | fclose (inf); |
| 1239 | return; |
| 1240 | } |
| 1241 | \f |
| 1242 | /* Record a tag. */ |
| 1243 | void |
| 1244 | pfnote (name, is_func, linestart, linelen, lno, cno) |
| 1245 | char *name; /* tag name, or NULL if unnamed */ |
| 1246 | logical is_func; /* tag is a function */ |
| 1247 | char *linestart; /* start of the line where tag is */ |
| 1248 | int linelen; /* length of the line where tag is */ |
| 1249 | int lno; /* line number */ |
| 1250 | long cno; /* character number */ |
| 1251 | { |
| 1252 | register NODE *np; |
| 1253 | |
| 1254 | if (CTAGS && name == NULL) |
| 1255 | return; |
| 1256 | |
| 1257 | np = xnew (1, NODE); |
| 1258 | |
| 1259 | /* If ctags mode, change name "main" to M<thisfilename>. */ |
| 1260 | if (CTAGS && !cxref_style && streq (name, "main")) |
| 1261 | { |
| 1262 | register char *fp = etags_strrchr (curfile, '/'); |
| 1263 | np->name = concat ("M", fp == 0 ? curfile : fp + 1, ""); |
| 1264 | fp = etags_strrchr (np->name, '.'); |
| 1265 | if (fp && fp[1] != '\0' && fp[2] == '\0') |
| 1266 | fp[0] = 0; |
| 1267 | } |
| 1268 | else |
| 1269 | np->name = name; |
| 1270 | np->been_warned = FALSE; |
| 1271 | np->file = curfile; |
| 1272 | np->is_func = is_func; |
| 1273 | np->lno = lno; |
| 1274 | /* Our char numbers are 0-base, because of C language tradition? |
| 1275 | ctags compatibility? old versions compatibility? I don't know. |
| 1276 | Anyway, since emacs's are 1-base we expect etags.el to take care |
| 1277 | of the difference. If we wanted to have 1-based numbers, we would |
| 1278 | uncomment the +1 below. */ |
| 1279 | np->cno = cno /* + 1 */ ; |
| 1280 | np->left = np->right = NULL; |
| 1281 | if (CTAGS && !cxref_style) |
| 1282 | { |
| 1283 | if (strlen (linestart) < 50) |
| 1284 | np->pat = concat (linestart, "$", ""); |
| 1285 | else |
| 1286 | np->pat = savenstr (linestart, 50); |
| 1287 | } |
| 1288 | else |
| 1289 | np->pat = savenstr (linestart, linelen); |
| 1290 | |
| 1291 | add_node (np, &head); |
| 1292 | } |
| 1293 | |
| 1294 | /* |
| 1295 | * free_tree () |
| 1296 | * recurse on left children, iterate on right children. |
| 1297 | */ |
| 1298 | void |
| 1299 | free_tree (node) |
| 1300 | register NODE *node; |
| 1301 | { |
| 1302 | while (node) |
| 1303 | { |
| 1304 | register NODE *node_right = node->right; |
| 1305 | free_tree (node->left); |
| 1306 | if (node->name != NULL) |
| 1307 | free (node->name); |
| 1308 | free (node->pat); |
| 1309 | free ((char *) node); |
| 1310 | node = node_right; |
| 1311 | } |
| 1312 | } |
| 1313 | |
| 1314 | /* |
| 1315 | * add_node () |
| 1316 | * Adds a node to the tree of nodes. In etags mode, we don't keep |
| 1317 | * it sorted; we just keep a linear list. In ctags mode, maintain |
| 1318 | * an ordered tree, with no attempt at balancing. |
| 1319 | * |
| 1320 | * add_node is the only function allowed to add nodes, so it can |
| 1321 | * maintain state. |
| 1322 | */ |
| 1323 | NODE *last_node = NULL; |
| 1324 | void |
| 1325 | add_node (node, cur_node_p) |
| 1326 | NODE *node, **cur_node_p; |
| 1327 | { |
| 1328 | register int dif; |
| 1329 | register NODE *cur_node = *cur_node_p; |
| 1330 | |
| 1331 | if (cur_node == NULL) |
| 1332 | { |
| 1333 | *cur_node_p = node; |
| 1334 | last_node = node; |
| 1335 | return; |
| 1336 | } |
| 1337 | |
| 1338 | if (!CTAGS) |
| 1339 | { |
| 1340 | /* Etags Mode */ |
| 1341 | if (last_node == NULL) |
| 1342 | fatal ("internal error in add_node", 0); |
| 1343 | last_node->right = node; |
| 1344 | last_node = node; |
| 1345 | } |
| 1346 | else |
| 1347 | { |
| 1348 | /* Ctags Mode */ |
| 1349 | dif = strcmp (node->name, cur_node->name); |
| 1350 | |
| 1351 | /* |
| 1352 | * If this tag name matches an existing one, then |
| 1353 | * do not add the node, but maybe print a warning. |
| 1354 | */ |
| 1355 | if (!dif) |
| 1356 | { |
| 1357 | if (streq (node->file, cur_node->file)) |
| 1358 | { |
| 1359 | if (!no_warnings) |
| 1360 | { |
| 1361 | fprintf (stderr, "Duplicate entry in file %s, line %d: %s\n", |
| 1362 | node->file, lineno, node->name); |
| 1363 | fprintf (stderr, "Second entry ignored\n"); |
| 1364 | } |
| 1365 | } |
| 1366 | else if (!cur_node->been_warned && !no_warnings) |
| 1367 | { |
| 1368 | fprintf |
| 1369 | (stderr, |
| 1370 | "Duplicate entry in files %s and %s: %s (Warning only)\n", |
| 1371 | node->file, cur_node->file, node->name); |
| 1372 | cur_node->been_warned = TRUE; |
| 1373 | } |
| 1374 | return; |
| 1375 | } |
| 1376 | |
| 1377 | /* Actually add the node */ |
| 1378 | add_node (node, dif < 0 ? &cur_node->left : &cur_node->right); |
| 1379 | } |
| 1380 | } |
| 1381 | \f |
| 1382 | void |
| 1383 | put_entries (node) |
| 1384 | register NODE *node; |
| 1385 | { |
| 1386 | register char *sp; |
| 1387 | |
| 1388 | if (node == NULL) |
| 1389 | return; |
| 1390 | |
| 1391 | /* Output subentries that precede this one */ |
| 1392 | put_entries (node->left); |
| 1393 | |
| 1394 | /* Output this entry */ |
| 1395 | |
| 1396 | if (!CTAGS) |
| 1397 | { |
| 1398 | if (node->name != NULL) |
| 1399 | fprintf (tagf, "%s\177%s\001%d,%d\n", |
| 1400 | node->pat, node->name, node->lno, node->cno); |
| 1401 | else |
| 1402 | fprintf (tagf, "%s\177%d,%d\n", |
| 1403 | node->pat, node->lno, node->cno); |
| 1404 | } |
| 1405 | else |
| 1406 | { |
| 1407 | if (node->name == NULL) |
| 1408 | error ("internal error: NULL name in ctags mode.", 0); |
| 1409 | |
| 1410 | if (cxref_style) |
| 1411 | { |
| 1412 | if (vgrind_style) |
| 1413 | fprintf (stdout, "%s %s %d\n", |
| 1414 | node->name, node->file, (node->lno + 63) / 64); |
| 1415 | else |
| 1416 | fprintf (stdout, "%-16s %3d %-16s %s\n", |
| 1417 | node->name, node->lno, node->file, node->pat); |
| 1418 | } |
| 1419 | else |
| 1420 | { |
| 1421 | fprintf (tagf, "%s\t%s\t", node->name, node->file); |
| 1422 | |
| 1423 | if (node->is_func) |
| 1424 | { /* a function */ |
| 1425 | putc (searchar, tagf); |
| 1426 | putc ('^', tagf); |
| 1427 | |
| 1428 | for (sp = node->pat; *sp; sp++) |
| 1429 | { |
| 1430 | if (*sp == '\\' || *sp == searchar) |
| 1431 | putc ('\\', tagf); |
| 1432 | putc (*sp, tagf); |
| 1433 | } |
| 1434 | putc (searchar, tagf); |
| 1435 | } |
| 1436 | else |
| 1437 | { /* a typedef; text pattern inadequate */ |
| 1438 | fprintf (tagf, "%d", node->lno); |
| 1439 | } |
| 1440 | putc ('\n', tagf); |
| 1441 | } |
| 1442 | } |
| 1443 | |
| 1444 | /* Output subentries that follow this one */ |
| 1445 | put_entries (node->right); |
| 1446 | } |
| 1447 | |
| 1448 | /* Length of a number's decimal representation. */ |
| 1449 | int |
| 1450 | number_len (num) |
| 1451 | long num; |
| 1452 | { |
| 1453 | int len = 0; |
| 1454 | if (!num) |
| 1455 | return 1; |
| 1456 | for (; num; num /= 10) |
| 1457 | ++len; |
| 1458 | return len; |
| 1459 | } |
| 1460 | |
| 1461 | /* |
| 1462 | * Return total number of characters that put_entries will output for |
| 1463 | * the nodes in the subtree of the specified node. Works only if |
| 1464 | * we are not ctags, but called only in that case. This count |
| 1465 | * is irrelevant with the new tags.el, but is still supplied for |
| 1466 | * backward compatibility. |
| 1467 | */ |
| 1468 | int |
| 1469 | total_size_of_entries (node) |
| 1470 | register NODE *node; |
| 1471 | { |
| 1472 | register int total; |
| 1473 | |
| 1474 | if (node == NULL) |
| 1475 | return 0; |
| 1476 | |
| 1477 | total = 0; |
| 1478 | for (; node; node = node->right) |
| 1479 | { |
| 1480 | /* Count left subentries. */ |
| 1481 | total += total_size_of_entries (node->left); |
| 1482 | |
| 1483 | /* Count this entry */ |
| 1484 | total += strlen (node->pat) + 1; |
| 1485 | total += number_len ((long) node->lno) + 1 + number_len (node->cno) + 1; |
| 1486 | if (node->name != NULL) |
| 1487 | total += 1 + strlen (node->name); /* \001name */ |
| 1488 | } |
| 1489 | |
| 1490 | return total; |
| 1491 | } |
| 1492 | \f |
| 1493 | /* |
| 1494 | * The C symbol tables. |
| 1495 | */ |
| 1496 | enum sym_type |
| 1497 | { |
| 1498 | st_none, st_C_objprot, st_C_objimpl, st_C_objend, st_C_gnumacro, |
| 1499 | st_C_struct, st_C_enum, st_C_define, st_C_typedef, st_C_typespec |
| 1500 | }; |
| 1501 | |
| 1502 | /* Feed stuff between (but not including) %[ and %] lines to: |
| 1503 | gperf -c -k 1,3 -o -p -r -t |
| 1504 | %[ |
| 1505 | struct C_stab_entry { char *name; int c_ext; enum sym_type type; } |
| 1506 | %% |
| 1507 | @interface, 0, st_C_objprot |
| 1508 | @protocol, 0, st_C_objprot |
| 1509 | @implementation,0, st_C_objimpl |
| 1510 | @end, 0, st_C_objend |
| 1511 | class, C_PLPL, st_C_struct |
| 1512 | namespace, C_PLPL, st_C_struct |
| 1513 | domain, C_STAR, st_C_struct |
| 1514 | union, 0, st_C_struct |
| 1515 | struct, 0, st_C_struct |
| 1516 | enum, 0, st_C_enum |
| 1517 | typedef, 0, st_C_typedef |
| 1518 | define, 0, st_C_define |
| 1519 | bool, C_PLPL, st_C_typespec |
| 1520 | long, 0, st_C_typespec |
| 1521 | short, 0, st_C_typespec |
| 1522 | int, 0, st_C_typespec |
| 1523 | char, 0, st_C_typespec |
| 1524 | float, 0, st_C_typespec |
| 1525 | double, 0, st_C_typespec |
| 1526 | signed, 0, st_C_typespec |
| 1527 | unsigned, 0, st_C_typespec |
| 1528 | auto, 0, st_C_typespec |
| 1529 | void, 0, st_C_typespec |
| 1530 | extern, 0, st_C_typespec |
| 1531 | static, 0, st_C_typespec |
| 1532 | const, 0, st_C_typespec |
| 1533 | volatile, 0, st_C_typespec |
| 1534 | explicit, C_PLPL, st_C_typespec |
| 1535 | mutable, C_PLPL, st_C_typespec |
| 1536 | typename, C_PLPL, st_C_typespec |
| 1537 | # DEFUN used in emacs, the next three used in glibc (SYSCALL only for mach). |
| 1538 | DEFUN, 0, st_C_gnumacro |
| 1539 | SYSCALL, 0, st_C_gnumacro |
| 1540 | ENTRY, 0, st_C_gnumacro |
| 1541 | PSEUDO, 0, st_C_gnumacro |
| 1542 | # These are defined inside C functions, so currently they are not met. |
| 1543 | # EXFUN used in glibc, DEFVAR_* in emacs. |
| 1544 | #EXFUN, 0, st_C_gnumacro |
| 1545 | #DEFVAR_, 0, st_C_gnumacro |
| 1546 | %] |
| 1547 | and replace lines between %< and %> with its output. */ |
| 1548 | /*%<*/ |
| 1549 | /* C code produced by gperf version 2.1 (K&R C version) */ |
| 1550 | /* Command-line: gperf -c -k 1,3 -o -p -r -t */ |
| 1551 | |
| 1552 | |
| 1553 | struct C_stab_entry { char *name; int c_ext; enum sym_type type; }; |
| 1554 | |
| 1555 | #define MIN_WORD_LENGTH 3 |
| 1556 | #define MAX_WORD_LENGTH 15 |
| 1557 | #define MIN_HASH_VALUE 34 |
| 1558 | #define MAX_HASH_VALUE 121 |
| 1559 | /* |
| 1560 | 34 keywords |
| 1561 | 88 is the maximum key range |
| 1562 | */ |
| 1563 | |
| 1564 | static int |
| 1565 | hash (str, len) |
| 1566 | register char *str; |
| 1567 | register unsigned int len; |
| 1568 | { |
| 1569 | static unsigned char hash_table[] = |
| 1570 | { |
| 1571 | 121, 121, 121, 121, 121, 121, 121, 121, 121, 121, |
| 1572 | 121, 121, 121, 121, 121, 121, 121, 121, 121, 121, |
| 1573 | 121, 121, 121, 121, 121, 121, 121, 121, 121, 121, |
| 1574 | 121, 121, 121, 121, 121, 121, 121, 121, 121, 121, |
| 1575 | 121, 121, 121, 121, 121, 121, 121, 121, 121, 121, |
| 1576 | 121, 121, 121, 121, 121, 121, 121, 121, 121, 121, |
| 1577 | 121, 121, 121, 121, 45, 121, 121, 121, 16, 19, |
| 1578 | 61, 121, 121, 121, 121, 121, 121, 121, 121, 121, |
| 1579 | 10, 121, 121, 20, 53, 121, 121, 121, 121, 121, |
| 1580 | 121, 121, 121, 121, 121, 121, 121, 41, 45, 22, |
| 1581 | 60, 47, 37, 28, 121, 55, 121, 121, 20, 14, |
| 1582 | 29, 30, 5, 121, 50, 59, 30, 54, 6, 121, |
| 1583 | 121, 121, 121, 121, 121, 121, 121, 121, |
| 1584 | }; |
| 1585 | return len + hash_table[str[2]] + hash_table[str[0]]; |
| 1586 | } |
| 1587 | |
| 1588 | struct C_stab_entry * |
| 1589 | in_word_set (str, len) |
| 1590 | register char *str; |
| 1591 | register unsigned int len; |
| 1592 | { |
| 1593 | |
| 1594 | static struct C_stab_entry wordlist[] = |
| 1595 | { |
| 1596 | {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, |
| 1597 | {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, |
| 1598 | {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, |
| 1599 | {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, |
| 1600 | {"volatile", 0, st_C_typespec}, |
| 1601 | {"PSEUDO", 0, st_C_gnumacro}, |
| 1602 | {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, |
| 1603 | {"typedef", 0, st_C_typedef}, |
| 1604 | {"typename", C_PLPL, st_C_typespec}, |
| 1605 | {"",}, {"",}, {"",}, |
| 1606 | {"SYSCALL", 0, st_C_gnumacro}, |
| 1607 | {"",}, {"",}, {"",}, |
| 1608 | {"mutable", C_PLPL, st_C_typespec}, |
| 1609 | {"namespace", C_PLPL, st_C_struct}, |
| 1610 | {"long", 0, st_C_typespec}, |
| 1611 | {"",}, {"",}, |
| 1612 | {"const", 0, st_C_typespec}, |
| 1613 | {"",}, {"",}, {"",}, |
| 1614 | {"explicit", C_PLPL, st_C_typespec}, |
| 1615 | {"",}, {"",}, {"",}, {"",}, |
| 1616 | {"void", 0, st_C_typespec}, |
| 1617 | {"",}, |
| 1618 | {"char", 0, st_C_typespec}, |
| 1619 | {"class", C_PLPL, st_C_struct}, |
| 1620 | {"",}, {"",}, {"",}, |
| 1621 | {"float", 0, st_C_typespec}, |
| 1622 | {"",}, |
| 1623 | {"@implementation", 0, st_C_objimpl}, |
| 1624 | {"auto", 0, st_C_typespec}, |
| 1625 | {"",}, |
| 1626 | {"ENTRY", 0, st_C_gnumacro}, |
| 1627 | {"@end", 0, st_C_objend}, |
| 1628 | {"bool", C_PLPL, st_C_typespec}, |
| 1629 | {"domain", C_STAR, st_C_struct}, |
| 1630 | {"",}, |
| 1631 | {"DEFUN", 0, st_C_gnumacro}, |
| 1632 | {"extern", 0, st_C_typespec}, |
| 1633 | {"@interface", 0, st_C_objprot}, |
| 1634 | {"",}, {"",}, {"",}, |
| 1635 | {"int", 0, st_C_typespec}, |
| 1636 | {"",}, {"",}, {"",}, {"",}, |
| 1637 | {"signed", 0, st_C_typespec}, |
| 1638 | {"short", 0, st_C_typespec}, |
| 1639 | {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, |
| 1640 | {"define", 0, st_C_define}, |
| 1641 | {"@protocol", 0, st_C_objprot}, |
| 1642 | {"enum", 0, st_C_enum}, |
| 1643 | {"static", 0, st_C_typespec}, |
| 1644 | {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, |
| 1645 | {"union", 0, st_C_struct}, |
| 1646 | {"struct", 0, st_C_struct}, |
| 1647 | {"",}, {"",}, {"",}, {"",}, |
| 1648 | {"double", 0, st_C_typespec}, |
| 1649 | {"unsigned", 0, st_C_typespec}, |
| 1650 | }; |
| 1651 | |
| 1652 | if (len <= MAX_WORD_LENGTH && len >= MIN_WORD_LENGTH) |
| 1653 | { |
| 1654 | register int key = hash (str, len); |
| 1655 | |
| 1656 | if (key <= MAX_HASH_VALUE && key >= MIN_HASH_VALUE) |
| 1657 | { |
| 1658 | register char *s = wordlist[key].name; |
| 1659 | |
| 1660 | if (*s == *str && !strncmp (str + 1, s + 1, len - 1)) |
| 1661 | return &wordlist[key]; |
| 1662 | } |
| 1663 | } |
| 1664 | return 0; |
| 1665 | } |
| 1666 | /*%>*/ |
| 1667 | |
| 1668 | enum sym_type |
| 1669 | C_symtype(str, len, c_ext) |
| 1670 | char *str; |
| 1671 | int len; |
| 1672 | int c_ext; |
| 1673 | { |
| 1674 | register struct C_stab_entry *se = in_word_set(str, len); |
| 1675 | |
| 1676 | if (se == NULL || (se->c_ext && !(c_ext & se->c_ext))) |
| 1677 | return st_none; |
| 1678 | return se->type; |
| 1679 | } |
| 1680 | \f |
| 1681 | /* |
| 1682 | * C functions are recognized using a simple finite automaton. |
| 1683 | * funcdef is its state variable. |
| 1684 | */ |
| 1685 | enum |
| 1686 | { |
| 1687 | fnone, /* nothing seen */ |
| 1688 | ftagseen, /* function-like tag seen */ |
| 1689 | fstartlist, /* just after open parenthesis */ |
| 1690 | finlist, /* in parameter list */ |
| 1691 | flistseen, /* after parameter list */ |
| 1692 | fignore /* before open brace */ |
| 1693 | } funcdef; |
| 1694 | |
| 1695 | |
| 1696 | /* |
| 1697 | * typedefs are recognized using a simple finite automaton. |
| 1698 | * typdef is its state variable. |
| 1699 | */ |
| 1700 | enum |
| 1701 | { |
| 1702 | tnone, /* nothing seen */ |
| 1703 | ttypedseen, /* typedef keyword seen */ |
| 1704 | tinbody, /* inside typedef body */ |
| 1705 | tend, /* just before typedef tag */ |
| 1706 | tignore /* junk after typedef tag */ |
| 1707 | } typdef; |
| 1708 | |
| 1709 | |
| 1710 | /* |
| 1711 | * struct-like structures (enum, struct and union) are recognized |
| 1712 | * using another simple finite automaton. `structdef' is its state |
| 1713 | * variable. |
| 1714 | */ |
| 1715 | enum |
| 1716 | { |
| 1717 | snone, /* nothing seen yet */ |
| 1718 | skeyseen, /* struct-like keyword seen */ |
| 1719 | stagseen, /* struct-like tag seen */ |
| 1720 | scolonseen, /* colon seen after struct-like tag */ |
| 1721 | sinbody /* in struct body: recognize member func defs*/ |
| 1722 | } structdef; |
| 1723 | |
| 1724 | /* |
| 1725 | * When structdef is stagseen, scolonseen, or sinbody, structtag is the |
| 1726 | * struct tag, and structtype is the type of the preceding struct-like |
| 1727 | * keyword. |
| 1728 | */ |
| 1729 | char *structtag = "<uninited>"; |
| 1730 | enum sym_type structtype; |
| 1731 | |
| 1732 | /* |
| 1733 | * When objdef is different from onone, objtag is the name of the class. |
| 1734 | */ |
| 1735 | char *objtag = "<uninited>"; |
| 1736 | |
| 1737 | /* |
| 1738 | * Yet another little state machine to deal with preprocessor lines. |
| 1739 | */ |
| 1740 | enum |
| 1741 | { |
| 1742 | dnone, /* nothing seen */ |
| 1743 | dsharpseen, /* '#' seen as first char on line */ |
| 1744 | ddefineseen, /* '#' and 'define' seen */ |
| 1745 | dignorerest /* ignore rest of line */ |
| 1746 | } definedef; |
| 1747 | |
| 1748 | /* |
| 1749 | * State machine for Objective C protocols and implementations. |
| 1750 | */ |
| 1751 | enum |
| 1752 | { |
| 1753 | onone, /* nothing seen */ |
| 1754 | oprotocol, /* @interface or @protocol seen */ |
| 1755 | oimplementation, /* @implementations seen */ |
| 1756 | otagseen, /* class name seen */ |
| 1757 | oparenseen, /* parenthesis before category seen */ |
| 1758 | ocatseen, /* category name seen */ |
| 1759 | oinbody, /* in @implementation body */ |
| 1760 | omethodsign, /* in @implementation body, after +/- */ |
| 1761 | omethodtag, /* after method name */ |
| 1762 | omethodcolon, /* after method colon */ |
| 1763 | omethodparm, /* after method parameter */ |
| 1764 | oignore /* wait for @end */ |
| 1765 | } objdef; |
| 1766 | |
| 1767 | /* |
| 1768 | * Set this to TRUE, and the next token considered is called a function. |
| 1769 | * Used only for GNU emacs's function-defining macros. |
| 1770 | */ |
| 1771 | logical next_token_is_func; |
| 1772 | |
| 1773 | /* |
| 1774 | * TRUE in the rules part of a yacc file, FALSE outside (parse as C). |
| 1775 | */ |
| 1776 | logical yacc_rules; |
| 1777 | |
| 1778 | /* |
| 1779 | * methodlen is the length of the method name stored in token_name. |
| 1780 | */ |
| 1781 | int methodlen; |
| 1782 | |
| 1783 | /* |
| 1784 | * consider_token () |
| 1785 | * checks to see if the current token is at the start of a |
| 1786 | * function, or corresponds to a typedef, or is a struct/union/enum |
| 1787 | * tag. |
| 1788 | * |
| 1789 | * *IS_FUNC gets TRUE iff the token is a function or macro with args. |
| 1790 | * C_EXT is which language we are looking at. |
| 1791 | * |
| 1792 | * In the future we will need some way to adjust where the end of |
| 1793 | * the token is; for instance, implementing the C++ keyword |
| 1794 | * `operator' properly will adjust the end of the token to be after |
| 1795 | * whatever follows `operator'. |
| 1796 | * |
| 1797 | * Globals |
| 1798 | * funcdef IN OUT |
| 1799 | * structdef IN OUT |
| 1800 | * definedef IN OUT |
| 1801 | * typdef IN OUT |
| 1802 | * objdef IN OUT |
| 1803 | * next_token_is_func IN OUT |
| 1804 | */ |
| 1805 | |
| 1806 | logical |
| 1807 | consider_token (str, len, c, c_ext, cblev, parlev, is_func) |
| 1808 | register char *str; /* IN: token pointer */ |
| 1809 | register int len; /* IN: token length */ |
| 1810 | register char c; /* IN: first char after the token */ |
| 1811 | int c_ext; /* IN: C extensions mask */ |
| 1812 | int cblev; /* IN: curly brace level */ |
| 1813 | int parlev; /* IN: parenthesis level */ |
| 1814 | logical *is_func; /* OUT: function found */ |
| 1815 | { |
| 1816 | enum sym_type toktype = C_symtype (str, len, c_ext); |
| 1817 | |
| 1818 | /* |
| 1819 | * Advance the definedef state machine. |
| 1820 | */ |
| 1821 | switch (definedef) |
| 1822 | { |
| 1823 | case dnone: |
| 1824 | /* We're not on a preprocessor line. */ |
| 1825 | break; |
| 1826 | case dsharpseen: |
| 1827 | if (toktype == st_C_define) |
| 1828 | { |
| 1829 | definedef = ddefineseen; |
| 1830 | } |
| 1831 | else |
| 1832 | { |
| 1833 | definedef = dignorerest; |
| 1834 | } |
| 1835 | return FALSE; |
| 1836 | case ddefineseen: |
| 1837 | /* |
| 1838 | * Make a tag for any macro, unless it is a constant |
| 1839 | * and constantypedefs is FALSE. |
| 1840 | */ |
| 1841 | definedef = dignorerest; |
| 1842 | *is_func = (c == '('); |
| 1843 | if (!*is_func && !constantypedefs) |
| 1844 | return FALSE; |
| 1845 | else |
| 1846 | return TRUE; |
| 1847 | case dignorerest: |
| 1848 | return FALSE; |
| 1849 | default: |
| 1850 | error ("internal error: definedef value.", 0); |
| 1851 | } |
| 1852 | |
| 1853 | /* |
| 1854 | * Now typedefs |
| 1855 | */ |
| 1856 | switch (typdef) |
| 1857 | { |
| 1858 | case tnone: |
| 1859 | if (toktype == st_C_typedef) |
| 1860 | { |
| 1861 | if (typedefs) |
| 1862 | typdef = ttypedseen; |
| 1863 | funcdef = fnone; |
| 1864 | return FALSE; |
| 1865 | } |
| 1866 | break; |
| 1867 | case ttypedseen: |
| 1868 | switch (toktype) |
| 1869 | { |
| 1870 | case st_none: |
| 1871 | case st_C_typespec: |
| 1872 | typdef = tend; |
| 1873 | break; |
| 1874 | case st_C_struct: |
| 1875 | case st_C_enum: |
| 1876 | break; |
| 1877 | } |
| 1878 | /* Do not return here, so the structdef stuff has a chance. */ |
| 1879 | break; |
| 1880 | case tend: |
| 1881 | switch (toktype) |
| 1882 | { |
| 1883 | case st_C_typespec: |
| 1884 | case st_C_struct: |
| 1885 | case st_C_enum: |
| 1886 | return FALSE; |
| 1887 | } |
| 1888 | return TRUE; |
| 1889 | } |
| 1890 | |
| 1891 | /* |
| 1892 | * This structdef business is currently only invoked when cblev==0. |
| 1893 | * It should be recursively invoked whatever the curly brace level, |
| 1894 | * and a stack of states kept, to allow for definitions of structs |
| 1895 | * within structs. |
| 1896 | * |
| 1897 | * This structdef business is NOT invoked when we are ctags and the |
| 1898 | * file is plain C. This is because a struct tag may have the same |
| 1899 | * name as another tag, and this loses with ctags. |
| 1900 | * |
| 1901 | * This if statement deals with the typdef state machine as |
| 1902 | * follows: if typdef==ttypedseen and token is struct/union/class/enum, |
| 1903 | * return FALSE. All the other code here is for the structdef |
| 1904 | * state machine. |
| 1905 | */ |
| 1906 | switch (toktype) |
| 1907 | { |
| 1908 | case st_C_struct: |
| 1909 | case st_C_enum: |
| 1910 | if (typdef == ttypedseen |
| 1911 | || (typedefs_and_cplusplus && cblev == 0 && structdef == snone)) |
| 1912 | { |
| 1913 | structdef = skeyseen; |
| 1914 | structtype = toktype; |
| 1915 | } |
| 1916 | return FALSE; |
| 1917 | } |
| 1918 | if (structdef == skeyseen) |
| 1919 | { |
| 1920 | /* Save the tag for struct/union/class, for functions that may be |
| 1921 | defined inside. */ |
| 1922 | if (structtype == st_C_struct) |
| 1923 | structtag = savenstr (str, len); |
| 1924 | else |
| 1925 | structtag = "<enum>"; |
| 1926 | structdef = stagseen; |
| 1927 | return TRUE; |
| 1928 | } |
| 1929 | |
| 1930 | /* Avoid entering funcdef stuff if typdef is going on. */ |
| 1931 | if (typdef != tnone) |
| 1932 | { |
| 1933 | definedef = dnone; |
| 1934 | return FALSE; |
| 1935 | } |
| 1936 | |
| 1937 | /* Detect GNU macros. */ |
| 1938 | if (definedef == dnone && toktype == st_C_gnumacro) |
| 1939 | { |
| 1940 | next_token_is_func = TRUE; |
| 1941 | return FALSE; |
| 1942 | } |
| 1943 | if (next_token_is_func) |
| 1944 | { |
| 1945 | next_token_is_func = FALSE; |
| 1946 | funcdef = fignore; |
| 1947 | *is_func = TRUE; |
| 1948 | return TRUE; |
| 1949 | } |
| 1950 | |
| 1951 | /* |
| 1952 | * Detecting Objective C constructs. |
| 1953 | */ |
| 1954 | switch (objdef) |
| 1955 | { |
| 1956 | case onone: |
| 1957 | switch (toktype) |
| 1958 | { |
| 1959 | case st_C_objprot: |
| 1960 | objdef = oprotocol; |
| 1961 | return FALSE; |
| 1962 | case st_C_objimpl: |
| 1963 | objdef = oimplementation; |
| 1964 | return FALSE; |
| 1965 | } |
| 1966 | break; |
| 1967 | case oimplementation: |
| 1968 | /* Save the class tag for functions that may be defined inside. */ |
| 1969 | objtag = savenstr (str, len); |
| 1970 | objdef = oinbody; |
| 1971 | return FALSE; |
| 1972 | case oprotocol: |
| 1973 | /* Save the class tag for categories. */ |
| 1974 | objtag = savenstr (str, len); |
| 1975 | objdef = otagseen; |
| 1976 | *is_func = TRUE; |
| 1977 | return TRUE; |
| 1978 | case oparenseen: |
| 1979 | objdef = ocatseen; |
| 1980 | *is_func = TRUE; |
| 1981 | return TRUE; |
| 1982 | case oinbody: |
| 1983 | break; |
| 1984 | case omethodsign: |
| 1985 | if (parlev == 0) |
| 1986 | { |
| 1987 | objdef = omethodtag; |
| 1988 | methodlen = len; |
| 1989 | GROW_LINEBUFFER (token_name, methodlen+1); |
| 1990 | strncpy (token_name.buffer, str, len); |
| 1991 | token_name.buffer[methodlen] = '\0'; |
| 1992 | return TRUE; |
| 1993 | } |
| 1994 | return FALSE; |
| 1995 | case omethodcolon: |
| 1996 | if (parlev == 0) |
| 1997 | objdef = omethodparm; |
| 1998 | return FALSE; |
| 1999 | case omethodparm: |
| 2000 | if (parlev == 0) |
| 2001 | { |
| 2002 | objdef = omethodtag; |
| 2003 | methodlen += len; |
| 2004 | GROW_LINEBUFFER (token_name, methodlen+1); |
| 2005 | strncat (token_name.buffer, str, len); |
| 2006 | return TRUE; |
| 2007 | } |
| 2008 | return FALSE; |
| 2009 | case oignore: |
| 2010 | if (toktype == st_C_objend) |
| 2011 | { |
| 2012 | /* Memory leakage here: the string pointed by objtag is |
| 2013 | never released, because many tests would be needed to |
| 2014 | avoid breaking on incorrect input code. The amount of |
| 2015 | memory leaked here is the sum of the lengths of the |
| 2016 | class tags. |
| 2017 | free (objtag); */ |
| 2018 | objdef = onone; |
| 2019 | } |
| 2020 | return FALSE; |
| 2021 | } |
| 2022 | |
| 2023 | /* A function? */ |
| 2024 | switch (toktype) |
| 2025 | { |
| 2026 | case st_C_typespec: |
| 2027 | if (funcdef != finlist && funcdef != fignore) |
| 2028 | funcdef = fnone; /* should be useless */ |
| 2029 | return FALSE; |
| 2030 | default: |
| 2031 | if (funcdef == fnone) |
| 2032 | { |
| 2033 | funcdef = ftagseen; |
| 2034 | *is_func = TRUE; |
| 2035 | return TRUE; |
| 2036 | } |
| 2037 | } |
| 2038 | |
| 2039 | return FALSE; |
| 2040 | } |
| 2041 | |
| 2042 | /* |
| 2043 | * C_entries () |
| 2044 | * This routine finds functions, typedefs, #define's and |
| 2045 | * struct/union/enum definitions in C syntax and adds them |
| 2046 | * to the list. |
| 2047 | */ |
| 2048 | typedef struct |
| 2049 | { |
| 2050 | logical valid; |
| 2051 | char *str; |
| 2052 | logical named; |
| 2053 | int linelen; |
| 2054 | int lineno; |
| 2055 | long linepos; |
| 2056 | char *buffer; |
| 2057 | } TOKEN; |
| 2058 | |
| 2059 | #define current_lb_is_new (newndx == curndx) |
| 2060 | #define switch_line_buffers() (curndx = 1 - curndx) |
| 2061 | |
| 2062 | #define curlb (lbs[curndx].lb) |
| 2063 | #define othlb (lbs[1-curndx].lb) |
| 2064 | #define newlb (lbs[newndx].lb) |
| 2065 | #define curlinepos (lbs[curndx].linepos) |
| 2066 | #define othlinepos (lbs[1-curndx].linepos) |
| 2067 | #define newlinepos (lbs[newndx].linepos) |
| 2068 | |
| 2069 | #define CNL_SAVE_DEFINEDEF \ |
| 2070 | do { \ |
| 2071 | curlinepos = charno; \ |
| 2072 | lineno++; \ |
| 2073 | linecharno = charno; \ |
| 2074 | charno += readline (&curlb, inf); \ |
| 2075 | lp = curlb.buffer; \ |
| 2076 | quotednl = FALSE; \ |
| 2077 | newndx = curndx; \ |
| 2078 | } while (0) |
| 2079 | |
| 2080 | #define CNL \ |
| 2081 | do { \ |
| 2082 | CNL_SAVE_DEFINEDEF; \ |
| 2083 | if (savetok.valid) \ |
| 2084 | { \ |
| 2085 | tok = savetok; \ |
| 2086 | savetok.valid = FALSE; \ |
| 2087 | } \ |
| 2088 | definedef = dnone; \ |
| 2089 | } while (0) |
| 2090 | |
| 2091 | /* Ideally this macro should never be called wihen tok.valid is FALSE, |
| 2092 | but this would mean that the state machines always guess right. */ |
| 2093 | #define make_tag(isfun) do \ |
| 2094 | if (tok.valid) { \ |
| 2095 | char *name = NULL; \ |
| 2096 | if (CTAGS || tok.named) \ |
| 2097 | name = savestr (token_name.buffer); \ |
| 2098 | pfnote (name, isfun, tok.buffer, tok.linelen, tok.lineno, tok.linepos); \ |
| 2099 | tok.valid = FALSE; \ |
| 2100 | } while (0) |
| 2101 | |
| 2102 | void |
| 2103 | C_entries (c_ext, inf) |
| 2104 | int c_ext; /* extension of C */ |
| 2105 | FILE *inf; /* input file */ |
| 2106 | { |
| 2107 | register char c; /* latest char read; '\0' for end of line */ |
| 2108 | register char *lp; /* pointer one beyond the character `c' */ |
| 2109 | int curndx, newndx; /* indices for current and new lb */ |
| 2110 | TOKEN tok; /* latest token read */ |
| 2111 | register int tokoff; /* offset in line of start of current token */ |
| 2112 | register int toklen; /* length of current token */ |
| 2113 | int cblev; /* current curly brace level */ |
| 2114 | int parlev; /* current parenthesis level */ |
| 2115 | logical incomm, inquote, inchar, quotednl, midtoken; |
| 2116 | logical cplpl; |
| 2117 | TOKEN savetok; /* token saved during preprocessor handling */ |
| 2118 | |
| 2119 | |
| 2120 | curndx = newndx = 0; |
| 2121 | lineno = 0; |
| 2122 | charno = 0; |
| 2123 | lp = curlb.buffer; |
| 2124 | *lp = 0; |
| 2125 | |
| 2126 | funcdef = fnone; typdef = tnone; structdef = snone; |
| 2127 | definedef = dnone; objdef = onone; |
| 2128 | next_token_is_func = yacc_rules = FALSE; |
| 2129 | midtoken = inquote = inchar = incomm = quotednl = FALSE; |
| 2130 | tok.valid = savetok.valid = FALSE; |
| 2131 | cblev = 0; |
| 2132 | parlev = 0; |
| 2133 | cplpl = c_ext & C_PLPL; |
| 2134 | |
| 2135 | while (!feof (inf)) |
| 2136 | { |
| 2137 | c = *lp++; |
| 2138 | if (c == '\\') |
| 2139 | { |
| 2140 | /* If we're at the end of the line, the next character is a |
| 2141 | '\0'; don't skip it, because it's the thing that tells us |
| 2142 | to read the next line. */ |
| 2143 | if (*lp == '\0') |
| 2144 | { |
| 2145 | quotednl = TRUE; |
| 2146 | continue; |
| 2147 | } |
| 2148 | lp++; |
| 2149 | c = ' '; |
| 2150 | } |
| 2151 | else if (incomm) |
| 2152 | { |
| 2153 | switch (c) |
| 2154 | { |
| 2155 | case '*': |
| 2156 | if (*lp == '/') |
| 2157 | { |
| 2158 | c = *lp++; |
| 2159 | incomm = FALSE; |
| 2160 | } |
| 2161 | break; |
| 2162 | case '\0': |
| 2163 | /* Newlines inside comments do not end macro definitions in |
| 2164 | traditional cpp. */ |
| 2165 | CNL_SAVE_DEFINEDEF; |
| 2166 | break; |
| 2167 | } |
| 2168 | continue; |
| 2169 | } |
| 2170 | else if (inquote) |
| 2171 | { |
| 2172 | switch (c) |
| 2173 | { |
| 2174 | case '"': |
| 2175 | inquote = FALSE; |
| 2176 | break; |
| 2177 | case '\0': |
| 2178 | /* Newlines inside strings do not end macro definitions |
| 2179 | in traditional cpp, even though compilers don't |
| 2180 | usually accept them. */ |
| 2181 | CNL_SAVE_DEFINEDEF; |
| 2182 | break; |
| 2183 | } |
| 2184 | continue; |
| 2185 | } |
| 2186 | else if (inchar) |
| 2187 | { |
| 2188 | switch (c) |
| 2189 | { |
| 2190 | case '\0': |
| 2191 | /* Hmmm, something went wrong. */ |
| 2192 | CNL; |
| 2193 | /* FALLTHRU */ |
| 2194 | case '\'': |
| 2195 | inchar = FALSE; |
| 2196 | break; |
| 2197 | } |
| 2198 | continue; |
| 2199 | } |
| 2200 | else |
| 2201 | switch (c) |
| 2202 | { |
| 2203 | case '"': |
| 2204 | inquote = TRUE; |
| 2205 | if (funcdef != finlist && funcdef != fignore) |
| 2206 | funcdef = fnone; |
| 2207 | continue; |
| 2208 | case '\'': |
| 2209 | inchar = TRUE; |
| 2210 | if (funcdef != finlist && funcdef != fignore) |
| 2211 | funcdef = fnone; |
| 2212 | continue; |
| 2213 | case '/': |
| 2214 | if (*lp == '*') |
| 2215 | { |
| 2216 | lp++; |
| 2217 | incomm = TRUE; |
| 2218 | continue; |
| 2219 | } |
| 2220 | else if (/* cplpl && */ *lp == '/') |
| 2221 | { |
| 2222 | c = '\0'; |
| 2223 | break; |
| 2224 | } |
| 2225 | else |
| 2226 | break; |
| 2227 | case '%': |
| 2228 | if ((c_ext & YACC) && *lp == '%') |
| 2229 | { |
| 2230 | /* entering or exiting rules section in yacc file */ |
| 2231 | lp++; |
| 2232 | definedef = dnone; funcdef = fnone; |
| 2233 | typdef = tnone; structdef = snone; |
| 2234 | next_token_is_func = FALSE; |
| 2235 | midtoken = inquote = inchar = incomm = quotednl = FALSE; |
| 2236 | cblev = 0; |
| 2237 | yacc_rules = !yacc_rules; |
| 2238 | continue; |
| 2239 | } |
| 2240 | else |
| 2241 | break; |
| 2242 | case '#': |
| 2243 | if (definedef == dnone) |
| 2244 | { |
| 2245 | char *cp; |
| 2246 | logical cpptoken = TRUE; |
| 2247 | |
| 2248 | /* Look back on this line. If all blanks, or nonblanks |
| 2249 | followed by an end of comment, this is a preprocessor |
| 2250 | token. */ |
| 2251 | for (cp = newlb.buffer; cp < lp-1; cp++) |
| 2252 | if (!iswhite (*cp)) |
| 2253 | { |
| 2254 | if (*cp == '*' && *(cp+1) == '/') |
| 2255 | { |
| 2256 | cp++; |
| 2257 | cpptoken = TRUE; |
| 2258 | } |
| 2259 | else |
| 2260 | cpptoken = FALSE; |
| 2261 | } |
| 2262 | if (cpptoken) |
| 2263 | definedef = dsharpseen; |
| 2264 | } /* if (definedef == dnone) */ |
| 2265 | |
| 2266 | continue; |
| 2267 | } /* switch (c) */ |
| 2268 | |
| 2269 | |
| 2270 | /* Consider token only if some complicated conditions are satisfied. */ |
| 2271 | if ((definedef != dnone |
| 2272 | || (cblev == 0 && structdef != scolonseen) |
| 2273 | || (cblev == 1 && cplpl && structdef == sinbody)) |
| 2274 | && typdef != tignore |
| 2275 | && definedef != dignorerest |
| 2276 | && funcdef != finlist) |
| 2277 | { |
| 2278 | if (midtoken) |
| 2279 | { |
| 2280 | if (endtoken (c)) |
| 2281 | { |
| 2282 | if (c == ':' && cplpl && *lp == ':' && begtoken(*(lp + 1))) |
| 2283 | { |
| 2284 | /* |
| 2285 | * This handles :: in the middle, but not at the |
| 2286 | * beginning of an identifier. |
| 2287 | */ |
| 2288 | lp += 2; |
| 2289 | toklen += 3; |
| 2290 | } |
| 2291 | else |
| 2292 | { |
| 2293 | logical is_func = FALSE; |
| 2294 | |
| 2295 | if (yacc_rules |
| 2296 | || consider_token (newlb.buffer + tokoff, toklen, c, |
| 2297 | c_ext, cblev, parlev, &is_func)) |
| 2298 | { |
| 2299 | if (structdef == sinbody |
| 2300 | && definedef == dnone |
| 2301 | && is_func) |
| 2302 | /* function defined in C++ class body */ |
| 2303 | { |
| 2304 | GROW_LINEBUFFER (token_name, |
| 2305 | strlen(structtag)+2+toklen+1); |
| 2306 | strcpy (token_name.buffer, structtag); |
| 2307 | strcat (token_name.buffer, "::"); |
| 2308 | strncat (token_name.buffer, |
| 2309 | newlb.buffer+tokoff, toklen); |
| 2310 | tok.named = TRUE; |
| 2311 | } |
| 2312 | else if (objdef == ocatseen) |
| 2313 | /* Objective C category */ |
| 2314 | { |
| 2315 | GROW_LINEBUFFER (token_name, |
| 2316 | strlen(objtag)+2+toklen+1); |
| 2317 | strcpy (token_name.buffer, objtag); |
| 2318 | strcat (token_name.buffer, "("); |
| 2319 | strncat (token_name.buffer, |
| 2320 | newlb.buffer+tokoff, toklen); |
| 2321 | strcat (token_name.buffer, ")"); |
| 2322 | tok.named = TRUE; |
| 2323 | } |
| 2324 | else if (objdef == omethodtag |
| 2325 | || objdef == omethodparm) |
| 2326 | /* Objective C method */ |
| 2327 | { |
| 2328 | tok.named = TRUE; |
| 2329 | } |
| 2330 | else |
| 2331 | { |
| 2332 | GROW_LINEBUFFER (token_name, toklen+1); |
| 2333 | strncpy (token_name.buffer, |
| 2334 | newlb.buffer+tokoff, toklen); |
| 2335 | token_name.buffer[toklen] = '\0'; |
| 2336 | if (structdef == stagseen |
| 2337 | || typdef == tend |
| 2338 | || (is_func |
| 2339 | && definedef == dignorerest)) /* macro */ |
| 2340 | tok.named = TRUE; |
| 2341 | else |
| 2342 | tok.named = FALSE; |
| 2343 | } |
| 2344 | tok.lineno = lineno; |
| 2345 | tok.linelen = tokoff + toklen + 1; |
| 2346 | tok.buffer = newlb.buffer; |
| 2347 | tok.linepos = newlinepos; |
| 2348 | tok.valid = TRUE; |
| 2349 | |
| 2350 | if (definedef == dnone |
| 2351 | && (funcdef == ftagseen |
| 2352 | || structdef == stagseen |
| 2353 | || typdef == tend |
| 2354 | || objdef != onone)) |
| 2355 | { |
| 2356 | if (current_lb_is_new) |
| 2357 | switch_line_buffers (); |
| 2358 | } |
| 2359 | else |
| 2360 | make_tag (is_func); |
| 2361 | } |
| 2362 | midtoken = FALSE; |
| 2363 | } |
| 2364 | } /* if (endtoken (c)) */ |
| 2365 | else if (intoken (c)) |
| 2366 | { |
| 2367 | toklen++; |
| 2368 | continue; |
| 2369 | } |
| 2370 | } /* if (midtoken) */ |
| 2371 | else if (begtoken (c)) |
| 2372 | { |
| 2373 | switch (definedef) |
| 2374 | { |
| 2375 | case dnone: |
| 2376 | switch (funcdef) |
| 2377 | { |
| 2378 | case fstartlist: |
| 2379 | funcdef = finlist; |
| 2380 | continue; |
| 2381 | case flistseen: |
| 2382 | make_tag (TRUE); |
| 2383 | funcdef = fignore; |
| 2384 | break; |
| 2385 | case ftagseen: |
| 2386 | funcdef = fnone; |
| 2387 | break; |
| 2388 | } |
| 2389 | if (structdef == stagseen) |
| 2390 | structdef = snone; |
| 2391 | break; |
| 2392 | case dsharpseen: |
| 2393 | savetok = tok; |
| 2394 | } |
| 2395 | if (!yacc_rules || lp == newlb.buffer + 1) |
| 2396 | { |
| 2397 | tokoff = lp - 1 - newlb.buffer; |
| 2398 | toklen = 1; |
| 2399 | midtoken = TRUE; |
| 2400 | } |
| 2401 | continue; |
| 2402 | } /* if (begtoken) */ |
| 2403 | } /* if must look at token */ |
| 2404 | |
| 2405 | |
| 2406 | /* Detect end of line, colon, comma, semicolon and various braces |
| 2407 | after having handled a token.*/ |
| 2408 | switch (c) |
| 2409 | { |
| 2410 | case ':': |
| 2411 | if (definedef != dnone) |
| 2412 | break; |
| 2413 | switch (objdef) |
| 2414 | { |
| 2415 | case otagseen: |
| 2416 | objdef = oignore; |
| 2417 | make_tag (TRUE); |
| 2418 | break; |
| 2419 | case omethodtag: |
| 2420 | case omethodparm: |
| 2421 | objdef = omethodcolon; |
| 2422 | methodlen += 1; |
| 2423 | GROW_LINEBUFFER (token_name, methodlen+1); |
| 2424 | strcat (token_name.buffer, ":"); |
| 2425 | break; |
| 2426 | } |
| 2427 | if (structdef == stagseen) |
| 2428 | structdef = scolonseen; |
| 2429 | else |
| 2430 | switch (funcdef) |
| 2431 | { |
| 2432 | case ftagseen: |
| 2433 | if (yacc_rules) |
| 2434 | { |
| 2435 | make_tag (FALSE); |
| 2436 | funcdef = fignore; |
| 2437 | } |
| 2438 | break; |
| 2439 | case fstartlist: |
| 2440 | funcdef = fnone; |
| 2441 | break; |
| 2442 | } |
| 2443 | break; |
| 2444 | case ';': |
| 2445 | if (definedef != dnone) |
| 2446 | break; |
| 2447 | if (cblev == 0) |
| 2448 | switch (typdef) |
| 2449 | { |
| 2450 | case tend: |
| 2451 | make_tag (FALSE); |
| 2452 | /* FALLTHRU */ |
| 2453 | default: |
| 2454 | typdef = tnone; |
| 2455 | } |
| 2456 | if (funcdef != fignore) |
| 2457 | { |
| 2458 | funcdef = fnone; |
| 2459 | /* The following instruction invalidates the token. |
| 2460 | Probably the token should be invalidated in all |
| 2461 | other cases where some state machine is reset. */ |
| 2462 | tok.valid = FALSE; |
| 2463 | } |
| 2464 | if (structdef == stagseen) |
| 2465 | structdef = snone; |
| 2466 | break; |
| 2467 | case ',': |
| 2468 | if (definedef != dnone) |
| 2469 | break; |
| 2470 | switch (objdef) |
| 2471 | { |
| 2472 | case omethodtag: |
| 2473 | case omethodparm: |
| 2474 | make_tag (TRUE); |
| 2475 | objdef = oinbody; |
| 2476 | break; |
| 2477 | } |
| 2478 | if (funcdef != finlist && funcdef != fignore) |
| 2479 | funcdef = fnone; |
| 2480 | if (structdef == stagseen) |
| 2481 | structdef = snone; |
| 2482 | break; |
| 2483 | case '[': |
| 2484 | if (definedef != dnone) |
| 2485 | break; |
| 2486 | if (cblev == 0 && typdef == tend) |
| 2487 | { |
| 2488 | typdef = tignore; |
| 2489 | make_tag (FALSE); |
| 2490 | break; |
| 2491 | } |
| 2492 | if (funcdef != finlist && funcdef != fignore) |
| 2493 | funcdef = fnone; |
| 2494 | if (structdef == stagseen) |
| 2495 | structdef = snone; |
| 2496 | break; |
| 2497 | case '(': |
| 2498 | if (definedef != dnone) |
| 2499 | break; |
| 2500 | if (objdef == otagseen && parlev == 0) |
| 2501 | objdef = oparenseen; |
| 2502 | switch (funcdef) |
| 2503 | { |
| 2504 | case fnone: |
| 2505 | switch (typdef) |
| 2506 | { |
| 2507 | case ttypedseen: |
| 2508 | case tend: |
| 2509 | /* Make sure that the next char is not a '*'. |
| 2510 | This handles constructs like: |
| 2511 | typedef void OperatorFun (int fun); */ |
| 2512 | if (*lp != '*') |
| 2513 | { |
| 2514 | typdef = tignore; |
| 2515 | make_tag (FALSE); |
| 2516 | } |
| 2517 | break; |
| 2518 | } /* switch (typdef) */ |
| 2519 | break; |
| 2520 | case ftagseen: |
| 2521 | funcdef = fstartlist; |
| 2522 | break; |
| 2523 | case flistseen: |
| 2524 | funcdef = finlist; |
| 2525 | break; |
| 2526 | } |
| 2527 | parlev++; |
| 2528 | break; |
| 2529 | case ')': |
| 2530 | if (definedef != dnone) |
| 2531 | break; |
| 2532 | if (objdef == ocatseen && parlev == 1) |
| 2533 | { |
| 2534 | make_tag (TRUE); |
| 2535 | objdef = oignore; |
| 2536 | } |
| 2537 | if (--parlev == 0) |
| 2538 | { |
| 2539 | switch (funcdef) |
| 2540 | { |
| 2541 | case fstartlist: |
| 2542 | case finlist: |
| 2543 | funcdef = flistseen; |
| 2544 | break; |
| 2545 | } |
| 2546 | if (cblev == 0 && typdef == tend) |
| 2547 | { |
| 2548 | typdef = tignore; |
| 2549 | make_tag (FALSE); |
| 2550 | } |
| 2551 | } |
| 2552 | else if (parlev < 0) /* can happen due to ill-conceived #if's. */ |
| 2553 | parlev = 0; |
| 2554 | break; |
| 2555 | case '{': |
| 2556 | if (definedef != dnone) |
| 2557 | break; |
| 2558 | if (typdef == ttypedseen) |
| 2559 | typdef = tinbody; |
| 2560 | switch (structdef) |
| 2561 | { |
| 2562 | case skeyseen: /* unnamed struct */ |
| 2563 | structtag = "_anonymous_"; |
| 2564 | structdef = sinbody; |
| 2565 | break; |
| 2566 | case stagseen: |
| 2567 | case scolonseen: /* named struct */ |
| 2568 | structdef = sinbody; |
| 2569 | make_tag (FALSE); |
| 2570 | break; |
| 2571 | } |
| 2572 | switch (funcdef) |
| 2573 | { |
| 2574 | case flistseen: |
| 2575 | make_tag (TRUE); |
| 2576 | /* FALLTHRU */ |
| 2577 | case fignore: |
| 2578 | funcdef = fnone; |
| 2579 | break; |
| 2580 | case fnone: |
| 2581 | switch (objdef) |
| 2582 | { |
| 2583 | case otagseen: |
| 2584 | make_tag (TRUE); |
| 2585 | objdef = oignore; |
| 2586 | break; |
| 2587 | case omethodtag: |
| 2588 | case omethodparm: |
| 2589 | make_tag (TRUE); |
| 2590 | objdef = oinbody; |
| 2591 | break; |
| 2592 | default: |
| 2593 | /* Neutralize `extern "C" {' grot and look inside structs. */ |
| 2594 | if (cblev == 0 && structdef == snone && typdef == tnone) |
| 2595 | cblev = -1; |
| 2596 | } |
| 2597 | } |
| 2598 | cblev++; |
| 2599 | break; |
| 2600 | case '*': |
| 2601 | if (definedef != dnone) |
| 2602 | break; |
| 2603 | if (funcdef == fstartlist) |
| 2604 | funcdef = fnone; /* avoid tagging `foo' in `foo (*bar()) ()' */ |
| 2605 | break; |
| 2606 | case '}': |
| 2607 | if (definedef != dnone) |
| 2608 | break; |
| 2609 | if (!noindentypedefs && lp == newlb.buffer + 1) |
| 2610 | { |
| 2611 | cblev = 0; /* reset curly brace level if first column */ |
| 2612 | parlev = 0; /* also reset paren level, just in case... */ |
| 2613 | } |
| 2614 | else if (cblev > 0) |
| 2615 | cblev--; |
| 2616 | if (cblev == 0) |
| 2617 | { |
| 2618 | if (typdef == tinbody) |
| 2619 | typdef = tend; |
| 2620 | /* Memory leakage here: the string pointed by structtag is |
| 2621 | never released, because I fear to miss something and |
| 2622 | break things while freeing the area. The amount of |
| 2623 | memory leaked here is the sum of the lengths of the |
| 2624 | struct tags. |
| 2625 | if (structdef == sinbody) |
| 2626 | free (structtag); */ |
| 2627 | |
| 2628 | structdef = snone; |
| 2629 | structtag = "<error>"; |
| 2630 | } |
| 2631 | break; |
| 2632 | case '+': |
| 2633 | case '-': |
| 2634 | if (objdef == oinbody && cblev == 0) |
| 2635 | { |
| 2636 | objdef = omethodsign; |
| 2637 | break; |
| 2638 | } |
| 2639 | /* FALLTHRU */ |
| 2640 | case '=': case '#': case '~': case '&': case '%': case '/': |
| 2641 | case '|': case '^': case '!': case '<': case '>': case '.': case '?': |
| 2642 | if (definedef != dnone) |
| 2643 | break; |
| 2644 | /* These surely cannot follow a function tag. */ |
| 2645 | if (funcdef != finlist && funcdef != fignore) |
| 2646 | funcdef = fnone; |
| 2647 | break; |
| 2648 | case '\0': |
| 2649 | if (objdef == otagseen) |
| 2650 | { |
| 2651 | make_tag (TRUE); |
| 2652 | objdef = oignore; |
| 2653 | } |
| 2654 | /* If a macro spans multiple lines don't reset its state. */ |
| 2655 | if (quotednl) |
| 2656 | CNL_SAVE_DEFINEDEF; |
| 2657 | else |
| 2658 | CNL; |
| 2659 | break; |
| 2660 | } /* switch (c) */ |
| 2661 | |
| 2662 | } /* while not eof */ |
| 2663 | } |
| 2664 | |
| 2665 | /* |
| 2666 | * Process either a C++ file or a C file depending on the setting |
| 2667 | * of a global flag. |
| 2668 | */ |
| 2669 | void |
| 2670 | default_C_entries (inf) |
| 2671 | FILE *inf; |
| 2672 | { |
| 2673 | C_entries (cplusplus ? C_PLPL : 0, inf); |
| 2674 | } |
| 2675 | |
| 2676 | /* Always do plain ANSI C. */ |
| 2677 | void |
| 2678 | plain_C_entries (inf) |
| 2679 | FILE *inf; |
| 2680 | { |
| 2681 | C_entries (0, inf); |
| 2682 | } |
| 2683 | |
| 2684 | /* Always do C++. */ |
| 2685 | void |
| 2686 | Cplusplus_entries (inf) |
| 2687 | FILE *inf; |
| 2688 | { |
| 2689 | C_entries (C_PLPL, inf); |
| 2690 | } |
| 2691 | |
| 2692 | /* Always do C*. */ |
| 2693 | void |
| 2694 | Cstar_entries (inf) |
| 2695 | FILE *inf; |
| 2696 | { |
| 2697 | C_entries (C_STAR, inf); |
| 2698 | } |
| 2699 | |
| 2700 | /* Always do Yacc. */ |
| 2701 | void |
| 2702 | Yacc_entries (inf) |
| 2703 | FILE *inf; |
| 2704 | { |
| 2705 | C_entries (YACC, inf); |
| 2706 | } |
| 2707 | \f |
| 2708 | /* Fortran parsing */ |
| 2709 | |
| 2710 | char *dbp; |
| 2711 | |
| 2712 | logical |
| 2713 | tail (cp) |
| 2714 | char *cp; |
| 2715 | { |
| 2716 | register int len = 0; |
| 2717 | |
| 2718 | while (*cp && lowcase(*cp) == lowcase(dbp[len])) |
| 2719 | cp++, len++; |
| 2720 | if (*cp == '\0' && !intoken(dbp[len])) |
| 2721 | { |
| 2722 | dbp += len; |
| 2723 | return TRUE; |
| 2724 | } |
| 2725 | return FALSE; |
| 2726 | } |
| 2727 | |
| 2728 | void |
| 2729 | takeprec () |
| 2730 | { |
| 2731 | while (isspace (*dbp)) |
| 2732 | dbp++; |
| 2733 | if (*dbp != '*') |
| 2734 | return; |
| 2735 | dbp++; |
| 2736 | while (isspace (*dbp)) |
| 2737 | dbp++; |
| 2738 | if (strneq (dbp, "(*)", 3)) |
| 2739 | { |
| 2740 | dbp += 3; |
| 2741 | return; |
| 2742 | } |
| 2743 | if (!isdigit (*dbp)) |
| 2744 | { |
| 2745 | --dbp; /* force failure */ |
| 2746 | return; |
| 2747 | } |
| 2748 | do |
| 2749 | dbp++; |
| 2750 | while (isdigit (*dbp)); |
| 2751 | } |
| 2752 | |
| 2753 | void |
| 2754 | getit (inf) |
| 2755 | FILE *inf; |
| 2756 | { |
| 2757 | register char *cp; |
| 2758 | |
| 2759 | while (isspace (*dbp)) |
| 2760 | dbp++; |
| 2761 | if (*dbp == '\0') |
| 2762 | { |
| 2763 | lineno++; |
| 2764 | linecharno = charno; |
| 2765 | charno += readline (&lb, inf); |
| 2766 | dbp = lb.buffer; |
| 2767 | if (dbp[5] != '&') |
| 2768 | return; |
| 2769 | dbp += 6; |
| 2770 | while (isspace (*dbp)) |
| 2771 | dbp++; |
| 2772 | } |
| 2773 | if (!isalpha (*dbp) |
| 2774 | && *dbp != '_' |
| 2775 | && *dbp != '$') |
| 2776 | return; |
| 2777 | for (cp = dbp + 1; |
| 2778 | (*cp |
| 2779 | && (isalpha (*cp) || isdigit (*cp) || (*cp == '_') || (*cp == '$'))); |
| 2780 | cp++) |
| 2781 | continue; |
| 2782 | pfnote ((CTAGS) ? savenstr (dbp, cp-dbp) : NULL, TRUE, |
| 2783 | lb.buffer, cp - lb.buffer + 1, lineno, linecharno); |
| 2784 | } |
| 2785 | |
| 2786 | void |
| 2787 | Fortran_functions (inf) |
| 2788 | FILE *inf; |
| 2789 | { |
| 2790 | lineno = 0; |
| 2791 | charno = 0; |
| 2792 | |
| 2793 | while (!feof (inf)) |
| 2794 | { |
| 2795 | lineno++; |
| 2796 | linecharno = charno; |
| 2797 | charno += readline (&lb, inf); |
| 2798 | dbp = lb.buffer; |
| 2799 | if (*dbp == '%') |
| 2800 | dbp++; /* Ratfor escape to fortran */ |
| 2801 | while (isspace (*dbp)) |
| 2802 | dbp++; |
| 2803 | if (*dbp == '\0') |
| 2804 | continue; |
| 2805 | switch (lowcase (*dbp)) |
| 2806 | { |
| 2807 | case 'i': |
| 2808 | if (tail ("integer")) |
| 2809 | takeprec (); |
| 2810 | break; |
| 2811 | case 'r': |
| 2812 | if (tail ("real")) |
| 2813 | takeprec (); |
| 2814 | break; |
| 2815 | case 'l': |
| 2816 | if (tail ("logical")) |
| 2817 | takeprec (); |
| 2818 | break; |
| 2819 | case 'c': |
| 2820 | if (tail ("complex") || tail ("character")) |
| 2821 | takeprec (); |
| 2822 | break; |
| 2823 | case 'd': |
| 2824 | if (tail ("double")) |
| 2825 | { |
| 2826 | while (isspace (*dbp)) |
| 2827 | dbp++; |
| 2828 | if (*dbp == '\0') |
| 2829 | continue; |
| 2830 | if (tail ("precision")) |
| 2831 | break; |
| 2832 | continue; |
| 2833 | } |
| 2834 | break; |
| 2835 | } |
| 2836 | while (isspace (*dbp)) |
| 2837 | dbp++; |
| 2838 | if (*dbp == '\0') |
| 2839 | continue; |
| 2840 | switch (lowcase (*dbp)) |
| 2841 | { |
| 2842 | case 'f': |
| 2843 | if (tail ("function")) |
| 2844 | getit (inf); |
| 2845 | continue; |
| 2846 | case 's': |
| 2847 | if (tail ("subroutine")) |
| 2848 | getit (inf); |
| 2849 | continue; |
| 2850 | case 'e': |
| 2851 | if (tail ("entry")) |
| 2852 | getit (inf); |
| 2853 | continue; |
| 2854 | case 'p': |
| 2855 | if (tail ("program")) |
| 2856 | { |
| 2857 | getit (inf); |
| 2858 | continue; |
| 2859 | } |
| 2860 | if (tail ("procedure")) |
| 2861 | getit (inf); |
| 2862 | continue; |
| 2863 | } |
| 2864 | } |
| 2865 | } |
| 2866 | \f |
| 2867 | /* |
| 2868 | * Bob Weiner, Motorola Inc., 4/3/94 |
| 2869 | * Unix and microcontroller assembly tag handling |
| 2870 | * look for '^[a-zA-Z_.$][a-zA_Z0-9_.$]*[: ^I^J]' |
| 2871 | */ |
| 2872 | void |
| 2873 | Asm_labels (inf) |
| 2874 | FILE *inf; |
| 2875 | { |
| 2876 | register char *cp; |
| 2877 | |
| 2878 | lineno = 0; |
| 2879 | charno = 0; |
| 2880 | |
| 2881 | while (!feof (inf)) |
| 2882 | { |
| 2883 | lineno++; |
| 2884 | linecharno = charno; |
| 2885 | charno += readline (&lb, inf); |
| 2886 | cp = lb.buffer; |
| 2887 | |
| 2888 | /* If first char is alphabetic or one of [_.$], test for colon |
| 2889 | following identifier. */ |
| 2890 | if (isalpha (*cp) || *cp == '_' || *cp == '.' || *cp == '$') |
| 2891 | { |
| 2892 | /* Read past label. */ |
| 2893 | cp++; |
| 2894 | while (isalnum (*cp) || *cp == '_' || *cp == '.' || *cp == '$') |
| 2895 | cp++; |
| 2896 | if (*cp == ':' || isspace (*cp)) |
| 2897 | { |
| 2898 | /* Found end of label, so copy it and add it to the table. */ |
| 2899 | pfnote ((CTAGS) ? savenstr(lb.buffer, cp-lb.buffer) : NULL, TRUE, |
| 2900 | lb.buffer, cp - lb.buffer + 1, lineno, linecharno); |
| 2901 | } |
| 2902 | } |
| 2903 | } |
| 2904 | } |
| 2905 | \f |
| 2906 | /* |
| 2907 | * Perl support by Bart Robinson <lomew@cs.utah.edu> |
| 2908 | * Perl sub names: look for /^sub[ \t\n]+[^ \t\n{]+/ |
| 2909 | */ |
| 2910 | void |
| 2911 | Perl_functions (inf) |
| 2912 | FILE *inf; |
| 2913 | { |
| 2914 | register char *cp; |
| 2915 | |
| 2916 | lineno = 0; |
| 2917 | charno = 0; |
| 2918 | |
| 2919 | while (!feof (inf)) |
| 2920 | { |
| 2921 | lineno++; |
| 2922 | linecharno = charno; |
| 2923 | charno += readline (&lb, inf); |
| 2924 | cp = lb.buffer; |
| 2925 | |
| 2926 | if (*cp++ == 's' && *cp++ == 'u' && *cp++ == 'b' && isspace(*cp++)) |
| 2927 | { |
| 2928 | while (*cp && isspace(*cp)) |
| 2929 | cp++; |
| 2930 | while (*cp && ! isspace(*cp) && *cp != '{') |
| 2931 | cp++; |
| 2932 | pfnote ((CTAGS) ? savenstr (lb.buffer, cp-lb.buffer) : NULL, TRUE, |
| 2933 | lb.buffer, cp - lb.buffer + 1, lineno, linecharno); |
| 2934 | } |
| 2935 | } |
| 2936 | } |
| 2937 | \f |
| 2938 | /* Added by Mosur Mohan, 4/22/88 */ |
| 2939 | /* Pascal parsing */ |
| 2940 | |
| 2941 | /* |
| 2942 | * Locates tags for procedures & functions. Doesn't do any type- or |
| 2943 | * var-definitions. It does look for the keyword "extern" or |
| 2944 | * "forward" immediately following the procedure statement; if found, |
| 2945 | * the tag is skipped. |
| 2946 | */ |
| 2947 | void |
| 2948 | Pascal_functions (inf) |
| 2949 | FILE *inf; |
| 2950 | { |
| 2951 | struct linebuffer tline; /* mostly copied from C_entries */ |
| 2952 | long save_lcno; |
| 2953 | int save_lineno, save_len; |
| 2954 | char c, *cp, *namebuf; |
| 2955 | |
| 2956 | logical /* each of these flags is TRUE iff: */ |
| 2957 | incomment, /* point is inside a comment */ |
| 2958 | inquote, /* point is inside '..' string */ |
| 2959 | get_tagname, /* point is after PROCEDURE/FUNCTION |
| 2960 | keyword, so next item = potential tag */ |
| 2961 | found_tag, /* point is after a potential tag */ |
| 2962 | inparms, /* point is within parameter-list */ |
| 2963 | verify_tag; /* point has passed the parm-list, so the |
| 2964 | next token will determine whether this |
| 2965 | is a FORWARD/EXTERN to be ignored, or |
| 2966 | whether it is a real tag */ |
| 2967 | |
| 2968 | lineno = 0; |
| 2969 | charno = 0; |
| 2970 | dbp = lb.buffer; |
| 2971 | *dbp = '\0'; |
| 2972 | save_len = 0; |
| 2973 | initbuffer (&tline); |
| 2974 | |
| 2975 | incomment = inquote = FALSE; |
| 2976 | found_tag = FALSE; /* have a proc name; check if extern */ |
| 2977 | get_tagname = FALSE; /* have found "procedure" keyword */ |
| 2978 | inparms = FALSE; /* found '(' after "proc" */ |
| 2979 | verify_tag = FALSE; /* check if "extern" is ahead */ |
| 2980 | |
| 2981 | /* long main loop to get next char */ |
| 2982 | while (!feof (inf)) |
| 2983 | { |
| 2984 | c = *dbp++; |
| 2985 | if (c == '\0') /* if end of line */ |
| 2986 | { |
| 2987 | lineno++; |
| 2988 | linecharno = charno; |
| 2989 | charno += readline (&lb, inf); |
| 2990 | dbp = lb.buffer; |
| 2991 | if (*dbp == '\0') |
| 2992 | continue; |
| 2993 | if (!((found_tag && verify_tag) || |
| 2994 | get_tagname)) |
| 2995 | c = *dbp++; /* only if don't need *dbp pointing |
| 2996 | to the beginning of the name of |
| 2997 | the procedure or function */ |
| 2998 | } |
| 2999 | if (incomment) |
| 3000 | { |
| 3001 | if (c == '}') /* within { } comments */ |
| 3002 | incomment = FALSE; |
| 3003 | else if (c == '*' && *dbp == ')') /* within (* *) comments */ |
| 3004 | { |
| 3005 | dbp++; |
| 3006 | incomment = FALSE; |
| 3007 | } |
| 3008 | continue; |
| 3009 | } |
| 3010 | else if (inquote) |
| 3011 | { |
| 3012 | if (c == '\'') |
| 3013 | inquote = FALSE; |
| 3014 | continue; |
| 3015 | } |
| 3016 | else |
| 3017 | switch (c) |
| 3018 | { |
| 3019 | case '\'': |
| 3020 | inquote = TRUE; /* found first quote */ |
| 3021 | continue; |
| 3022 | case '{': /* found open { comment */ |
| 3023 | incomment = TRUE; |
| 3024 | continue; |
| 3025 | case '(': |
| 3026 | if (*dbp == '*') /* found open (* comment */ |
| 3027 | { |
| 3028 | incomment = TRUE; |
| 3029 | dbp++; |
| 3030 | } |
| 3031 | else if (found_tag) /* found '(' after tag, i.e., parm-list */ |
| 3032 | inparms = TRUE; |
| 3033 | continue; |
| 3034 | case ')': /* end of parms list */ |
| 3035 | if (inparms) |
| 3036 | inparms = FALSE; |
| 3037 | continue; |
| 3038 | case ';': |
| 3039 | if (found_tag && !inparms) /* end of proc or fn stmt */ |
| 3040 | { |
| 3041 | verify_tag = TRUE; |
| 3042 | break; |
| 3043 | } |
| 3044 | continue; |
| 3045 | } |
| 3046 | if (found_tag && verify_tag && (*dbp != ' ')) |
| 3047 | { |
| 3048 | /* check if this is an "extern" declaration */ |
| 3049 | if (*dbp == '\0') |
| 3050 | continue; |
| 3051 | if (lowcase (*dbp == 'e')) |
| 3052 | { |
| 3053 | if (tail ("extern")) /* superfluous, really! */ |
| 3054 | { |
| 3055 | found_tag = FALSE; |
| 3056 | verify_tag = FALSE; |
| 3057 | } |
| 3058 | } |
| 3059 | else if (lowcase (*dbp) == 'f') |
| 3060 | { |
| 3061 | if (tail ("forward")) /* check for forward reference */ |
| 3062 | { |
| 3063 | found_tag = FALSE; |
| 3064 | verify_tag = FALSE; |
| 3065 | } |
| 3066 | } |
| 3067 | if (found_tag && verify_tag) /* not external proc, so make tag */ |
| 3068 | { |
| 3069 | found_tag = FALSE; |
| 3070 | verify_tag = FALSE; |
| 3071 | pfnote (namebuf, TRUE, |
| 3072 | tline.buffer, save_len, save_lineno, save_lcno); |
| 3073 | continue; |
| 3074 | } |
| 3075 | } |
| 3076 | if (get_tagname) /* grab name of proc or fn */ |
| 3077 | { |
| 3078 | if (*dbp == '\0') |
| 3079 | continue; |
| 3080 | |
| 3081 | /* save all values for later tagging */ |
| 3082 | GROW_LINEBUFFER (tline, strlen (lb.buffer) + 1); |
| 3083 | strcpy (tline.buffer, lb.buffer); |
| 3084 | save_lineno = lineno; |
| 3085 | save_lcno = linecharno; |
| 3086 | |
| 3087 | /* grab block name */ |
| 3088 | for (cp = dbp + 1; *cp && (!endtoken (*cp)); cp++) |
| 3089 | continue; |
| 3090 | namebuf = (CTAGS) ? savenstr (dbp, cp-dbp) : NULL; |
| 3091 | dbp = cp; /* set dbp to e-o-token */ |
| 3092 | save_len = dbp - lb.buffer + 1; |
| 3093 | get_tagname = FALSE; |
| 3094 | found_tag = TRUE; |
| 3095 | continue; |
| 3096 | |
| 3097 | /* and proceed to check for "extern" */ |
| 3098 | } |
| 3099 | else if (!incomment && !inquote && !found_tag) |
| 3100 | { |
| 3101 | /* check for proc/fn keywords */ |
| 3102 | switch (lowcase (c)) |
| 3103 | { |
| 3104 | case 'p': |
| 3105 | if (tail ("rocedure")) /* c = 'p', dbp has advanced */ |
| 3106 | get_tagname = TRUE; |
| 3107 | continue; |
| 3108 | case 'f': |
| 3109 | if (tail ("unction")) |
| 3110 | get_tagname = TRUE; |
| 3111 | continue; |
| 3112 | } |
| 3113 | } |
| 3114 | } /* while not eof */ |
| 3115 | |
| 3116 | free (tline.buffer); |
| 3117 | } |
| 3118 | \f |
| 3119 | /* |
| 3120 | * lisp tag functions |
| 3121 | * look for (def or (DEF, quote or QUOTE |
| 3122 | */ |
| 3123 | int |
| 3124 | L_isdef (strp) |
| 3125 | register char *strp; |
| 3126 | { |
| 3127 | return ((strp[1] == 'd' || strp[1] == 'D') |
| 3128 | && (strp[2] == 'e' || strp[2] == 'E') |
| 3129 | && (strp[3] == 'f' || strp[3] == 'F')); |
| 3130 | } |
| 3131 | |
| 3132 | int |
| 3133 | L_isquote (strp) |
| 3134 | register char *strp; |
| 3135 | { |
| 3136 | return ((*(++strp) == 'q' || *strp == 'Q') |
| 3137 | && (*(++strp) == 'u' || *strp == 'U') |
| 3138 | && (*(++strp) == 'o' || *strp == 'O') |
| 3139 | && (*(++strp) == 't' || *strp == 'T') |
| 3140 | && (*(++strp) == 'e' || *strp == 'E') |
| 3141 | && isspace(*(++strp))); |
| 3142 | } |
| 3143 | |
| 3144 | void |
| 3145 | L_getit () |
| 3146 | { |
| 3147 | register char *cp; |
| 3148 | |
| 3149 | if (*dbp == '\'') /* Skip prefix quote */ |
| 3150 | dbp++; |
| 3151 | else if (*dbp == '(' && L_isquote (dbp)) /* Skip "(quote " */ |
| 3152 | { |
| 3153 | dbp += 7; |
| 3154 | while (isspace(*dbp)) |
| 3155 | dbp++; |
| 3156 | } |
| 3157 | for (cp = dbp /*+1*/; |
| 3158 | *cp && *cp != '(' && *cp != ' ' && *cp != ')'; |
| 3159 | cp++) |
| 3160 | continue; |
| 3161 | if (cp == dbp) |
| 3162 | return; |
| 3163 | |
| 3164 | pfnote ((CTAGS) ? savenstr (dbp, cp-dbp) : NULL, TRUE, |
| 3165 | lb.buffer, cp - lb.buffer + 1, lineno, linecharno); |
| 3166 | } |
| 3167 | |
| 3168 | void |
| 3169 | Lisp_functions (inf) |
| 3170 | FILE *inf; |
| 3171 | { |
| 3172 | lineno = 0; |
| 3173 | charno = 0; |
| 3174 | |
| 3175 | while (!feof (inf)) |
| 3176 | { |
| 3177 | lineno++; |
| 3178 | linecharno = charno; |
| 3179 | charno += readline (&lb, inf); |
| 3180 | dbp = lb.buffer; |
| 3181 | if (dbp[0] == '(') |
| 3182 | { |
| 3183 | if (L_isdef (dbp)) |
| 3184 | { |
| 3185 | while (!isspace (*dbp)) |
| 3186 | dbp++; |
| 3187 | while (isspace (*dbp)) |
| 3188 | dbp++; |
| 3189 | L_getit (); |
| 3190 | } |
| 3191 | else |
| 3192 | { |
| 3193 | /* Check for (foo::defmumble name-defined ... */ |
| 3194 | do |
| 3195 | dbp++; |
| 3196 | while (*dbp && !isspace (*dbp) |
| 3197 | && *dbp != ':' && *dbp != '(' && *dbp != ')'); |
| 3198 | if (*dbp == ':') |
| 3199 | { |
| 3200 | do |
| 3201 | dbp++; |
| 3202 | while (*dbp == ':'); |
| 3203 | |
| 3204 | if (L_isdef (dbp - 1)) |
| 3205 | { |
| 3206 | while (!isspace (*dbp)) |
| 3207 | dbp++; |
| 3208 | while (isspace (*dbp)) |
| 3209 | dbp++; |
| 3210 | L_getit (); |
| 3211 | } |
| 3212 | } |
| 3213 | } |
| 3214 | } |
| 3215 | } |
| 3216 | } |
| 3217 | \f |
| 3218 | /* |
| 3219 | * Scheme tag functions |
| 3220 | * look for (def... xyzzy |
| 3221 | * look for (def... (xyzzy |
| 3222 | * look for (def ... ((...(xyzzy .... |
| 3223 | * look for (set! xyzzy |
| 3224 | */ |
| 3225 | |
| 3226 | void get_scheme (); |
| 3227 | |
| 3228 | void |
| 3229 | Scheme_functions (inf) |
| 3230 | FILE *inf; |
| 3231 | { |
| 3232 | lineno = 0; |
| 3233 | charno = 0; |
| 3234 | |
| 3235 | while (!feof (inf)) |
| 3236 | { |
| 3237 | lineno++; |
| 3238 | linecharno = charno; |
| 3239 | charno += readline (&lb, inf); |
| 3240 | dbp = lb.buffer; |
| 3241 | if (dbp[0] == '(' && |
| 3242 | (dbp[1] == 'D' || dbp[1] == 'd') && |
| 3243 | (dbp[2] == 'E' || dbp[2] == 'e') && |
| 3244 | (dbp[3] == 'F' || dbp[3] == 'f')) |
| 3245 | { |
| 3246 | while (!isspace (*dbp)) |
| 3247 | dbp++; |
| 3248 | /* Skip over open parens and white space */ |
| 3249 | while (*dbp && (isspace (*dbp) || *dbp == '(')) |
| 3250 | dbp++; |
| 3251 | get_scheme (); |
| 3252 | } |
| 3253 | if (dbp[0] == '(' && |
| 3254 | (dbp[1] == 'S' || dbp[1] == 's') && |
| 3255 | (dbp[2] == 'E' || dbp[2] == 'e') && |
| 3256 | (dbp[3] == 'T' || dbp[3] == 't') && |
| 3257 | (dbp[4] == '!' || dbp[4] == '!') && |
| 3258 | (isspace (dbp[5]))) |
| 3259 | { |
| 3260 | while (!isspace (*dbp)) |
| 3261 | dbp++; |
| 3262 | /* Skip over white space */ |
| 3263 | while (isspace (*dbp)) |
| 3264 | dbp++; |
| 3265 | get_scheme (); |
| 3266 | } |
| 3267 | } |
| 3268 | } |
| 3269 | |
| 3270 | void |
| 3271 | get_scheme () |
| 3272 | { |
| 3273 | register char *cp; |
| 3274 | |
| 3275 | if (*dbp == '\0') |
| 3276 | return; |
| 3277 | /* Go till you get to white space or a syntactic break */ |
| 3278 | for (cp = dbp + 1; |
| 3279 | *cp && *cp != '(' && *cp != ')' && !isspace (*cp); |
| 3280 | cp++) |
| 3281 | continue; |
| 3282 | pfnote ((CTAGS) ? savenstr (dbp, cp-dbp) : NULL, TRUE, |
| 3283 | lb.buffer, cp - lb.buffer + 1, lineno, linecharno); |
| 3284 | } |
| 3285 | \f |
| 3286 | /* Find tags in TeX and LaTeX input files. */ |
| 3287 | |
| 3288 | /* TEX_toktab is a table of TeX control sequences that define tags. |
| 3289 | Each TEX_tabent records one such control sequence. |
| 3290 | CONVERT THIS TO USE THE Stab TYPE!! */ |
| 3291 | struct TEX_tabent |
| 3292 | { |
| 3293 | char *name; |
| 3294 | int len; |
| 3295 | }; |
| 3296 | |
| 3297 | struct TEX_tabent *TEX_toktab = NULL; /* Table with tag tokens */ |
| 3298 | |
| 3299 | /* Default set of control sequences to put into TEX_toktab. |
| 3300 | The value of environment var TEXTAGS is prepended to this. */ |
| 3301 | |
| 3302 | char *TEX_defenv = "\ |
| 3303 | :chapter:section:subsection:subsubsection:eqno:label:ref:cite:bibitem\ |
| 3304 | :part:appendix:entry:index"; |
| 3305 | |
| 3306 | void TEX_mode (); |
| 3307 | struct TEX_tabent *TEX_decode_env (); |
| 3308 | int TEX_Token (); |
| 3309 | #if TeX_named_tokens |
| 3310 | void TEX_getit (); |
| 3311 | #endif |
| 3312 | |
| 3313 | char TEX_esc = '\\'; |
| 3314 | char TEX_opgrp = '{'; |
| 3315 | char TEX_clgrp = '}'; |
| 3316 | |
| 3317 | /* |
| 3318 | * TeX/LaTeX scanning loop. |
| 3319 | */ |
| 3320 | void |
| 3321 | TeX_functions (inf) |
| 3322 | FILE *inf; |
| 3323 | { |
| 3324 | char *lasthit; |
| 3325 | |
| 3326 | lineno = 0; |
| 3327 | charno = 0; |
| 3328 | |
| 3329 | /* Select either \ or ! as escape character. */ |
| 3330 | TEX_mode (inf); |
| 3331 | |
| 3332 | /* Initialize token table once from environment. */ |
| 3333 | if (!TEX_toktab) |
| 3334 | TEX_toktab = TEX_decode_env ("TEXTAGS", TEX_defenv); |
| 3335 | |
| 3336 | while (!feof (inf)) |
| 3337 | { /* Scan each line in file */ |
| 3338 | lineno++; |
| 3339 | linecharno = charno; |
| 3340 | charno += readline (&lb, inf); |
| 3341 | dbp = lb.buffer; |
| 3342 | lasthit = dbp; |
| 3343 | while (dbp = etags_strchr (dbp, TEX_esc)) /* Look at each esc in line */ |
| 3344 | { |
| 3345 | register int i; |
| 3346 | |
| 3347 | if (!*(++dbp)) |
| 3348 | break; |
| 3349 | linecharno += dbp - lasthit; |
| 3350 | lasthit = dbp; |
| 3351 | i = TEX_Token (lasthit); |
| 3352 | if (0 <= i) |
| 3353 | { |
| 3354 | pfnote (NULL, TRUE, |
| 3355 | lb.buffer, strlen (lb.buffer), lineno, linecharno); |
| 3356 | #if TeX_named_tokens |
| 3357 | TEX_getit (lasthit, TEX_toktab[i].len); |
| 3358 | #endif |
| 3359 | break; /* We only save a line once */ |
| 3360 | } |
| 3361 | } |
| 3362 | } |
| 3363 | } |
| 3364 | |
| 3365 | #define TEX_LESC '\\' |
| 3366 | #define TEX_SESC '!' |
| 3367 | #define TEX_cmt '%' |
| 3368 | |
| 3369 | /* Figure out whether TeX's escapechar is '\\' or '!' and set grouping |
| 3370 | chars accordingly. */ |
| 3371 | void |
| 3372 | TEX_mode (inf) |
| 3373 | FILE *inf; |
| 3374 | { |
| 3375 | int c; |
| 3376 | |
| 3377 | while ((c = getc (inf)) != EOF) |
| 3378 | { |
| 3379 | /* Skip to next line if we hit the TeX comment char. */ |
| 3380 | if (c == TEX_cmt) |
| 3381 | while (c != '\n') |
| 3382 | c = getc (inf); |
| 3383 | else if (c == TEX_LESC || c == TEX_SESC ) |
| 3384 | break; |
| 3385 | } |
| 3386 | |
| 3387 | if (c == TEX_LESC) |
| 3388 | { |
| 3389 | TEX_esc = TEX_LESC; |
| 3390 | TEX_opgrp = '{'; |
| 3391 | TEX_clgrp = '}'; |
| 3392 | } |
| 3393 | else |
| 3394 | { |
| 3395 | TEX_esc = TEX_SESC; |
| 3396 | TEX_opgrp = '<'; |
| 3397 | TEX_clgrp = '>'; |
| 3398 | } |
| 3399 | rewind (inf); |
| 3400 | } |
| 3401 | |
| 3402 | /* Read environment and prepend it to the default string. |
| 3403 | Build token table. */ |
| 3404 | struct TEX_tabent * |
| 3405 | TEX_decode_env (evarname, defenv) |
| 3406 | char *evarname; |
| 3407 | char *defenv; |
| 3408 | { |
| 3409 | register char *env, *p; |
| 3410 | |
| 3411 | struct TEX_tabent *tab; |
| 3412 | int size, i; |
| 3413 | |
| 3414 | /* Append default string to environment. */ |
| 3415 | env = getenv (evarname); |
| 3416 | if (!env) |
| 3417 | env = defenv; |
| 3418 | else |
| 3419 | env = concat (env, defenv, ""); |
| 3420 | |
| 3421 | /* Allocate a token table */ |
| 3422 | for (size = 1, p = env; p;) |
| 3423 | if ((p = etags_strchr (p, ':')) && *(++p)) |
| 3424 | size++; |
| 3425 | /* Add 1 to leave room for null terminator. */ |
| 3426 | tab = xnew (size + 1, struct TEX_tabent); |
| 3427 | |
| 3428 | /* Unpack environment string into token table. Be careful about */ |
| 3429 | /* zero-length strings (leading ':', "::" and trailing ':') */ |
| 3430 | for (i = 0; *env;) |
| 3431 | { |
| 3432 | p = etags_strchr (env, ':'); |
| 3433 | if (!p) /* End of environment string. */ |
| 3434 | p = env + strlen (env); |
| 3435 | if (p - env > 0) |
| 3436 | { /* Only non-zero strings. */ |
| 3437 | tab[i].name = savenstr (env, p - env); |
| 3438 | tab[i].len = strlen (tab[i].name); |
| 3439 | i++; |
| 3440 | } |
| 3441 | if (*p) |
| 3442 | env = p + 1; |
| 3443 | else |
| 3444 | { |
| 3445 | tab[i].name = NULL; /* Mark end of table. */ |
| 3446 | tab[i].len = 0; |
| 3447 | break; |
| 3448 | } |
| 3449 | } |
| 3450 | return tab; |
| 3451 | } |
| 3452 | |
| 3453 | #if TeX_named_tokens |
| 3454 | /* Record a tag defined by a TeX command of length LEN and starting at NAME. |
| 3455 | The name being defined actually starts at (NAME + LEN + 1). |
| 3456 | But we seem to include the TeX command in the tag name. */ |
| 3457 | void |
| 3458 | TEX_getit (name, len) |
| 3459 | char *name; |
| 3460 | int len; |
| 3461 | { |
| 3462 | char *p = name + len; |
| 3463 | |
| 3464 | if (*name == '\0') |
| 3465 | return; |
| 3466 | |
| 3467 | /* Let tag name extend to next group close (or end of line) */ |
| 3468 | while (*p && *p != TEX_clgrp) |
| 3469 | p++; |
| 3470 | pfnote (savenstr (name, p-name), TRUE, |
| 3471 | lb.buffer, strlen (lb.buffer), lineno, linecharno); |
| 3472 | } |
| 3473 | #endif |
| 3474 | |
| 3475 | /* If the text at CP matches one of the tag-defining TeX command names, |
| 3476 | return the pointer to the first occurrence of that command in TEX_toktab. |
| 3477 | Otherwise return -1. |
| 3478 | Keep the capital `T' in `Token' for dumb truncating compilers |
| 3479 | (this distinguishes it from `TEX_toktab' */ |
| 3480 | int |
| 3481 | TEX_Token (cp) |
| 3482 | char *cp; |
| 3483 | { |
| 3484 | int i; |
| 3485 | |
| 3486 | for (i = 0; TEX_toktab[i].len > 0; i++) |
| 3487 | if (strneq (TEX_toktab[i].name, cp, TEX_toktab[i].len)) |
| 3488 | return i; |
| 3489 | return -1; |
| 3490 | } |
| 3491 | \f |
| 3492 | /* |
| 3493 | * Prolog support (rewritten) by Anders Lindgren, Mar. 96 |
| 3494 | * |
| 3495 | * Assumes that the predicate starts at column 0. |
| 3496 | * Only the first clause of a predicate is added. |
| 3497 | */ |
| 3498 | void |
| 3499 | Prolog_functions (inf) |
| 3500 | FILE *inf; |
| 3501 | { |
| 3502 | int prolog_pred (); |
| 3503 | void prolog_skip_comment (); |
| 3504 | |
| 3505 | char * last; |
| 3506 | int len; |
| 3507 | int allocated; |
| 3508 | |
| 3509 | allocated = 0; |
| 3510 | len = 0; |
| 3511 | last = NULL; |
| 3512 | |
| 3513 | lineno = 0; |
| 3514 | linecharno = 0; |
| 3515 | charno = 0; |
| 3516 | |
| 3517 | while (!feof (inf)) |
| 3518 | { |
| 3519 | lineno++; |
| 3520 | linecharno += charno; |
| 3521 | charno = readline (&lb, inf); |
| 3522 | dbp = lb.buffer; |
| 3523 | if (dbp[0] == '\0') /* Empty line */ |
| 3524 | continue; |
| 3525 | else if (isspace (dbp[0])) /* Not a predicate */ |
| 3526 | continue; |
| 3527 | else if (dbp[0] == '/' && dbp[1] == '*') /* comment. */ |
| 3528 | prolog_skip_comment (&lb, inf, &lineno, &linecharno); |
| 3529 | else if (len = prolog_pred (dbp, last)) |
| 3530 | { |
| 3531 | /* Predicate. Store the function name so that we only |
| 3532 | * generates a tag for the first clause. */ |
| 3533 | if (last == NULL) |
| 3534 | last = xnew(len + 1, char); |
| 3535 | else if (len + 1 > allocated) |
| 3536 | last = (char *) xrealloc(last, len + 1); |
| 3537 | allocated = len + 1; |
| 3538 | strncpy (last, dbp, len); |
| 3539 | last[len] = '\0'; |
| 3540 | } |
| 3541 | } |
| 3542 | } |
| 3543 | |
| 3544 | |
| 3545 | void |
| 3546 | prolog_skip_comment (plb, inf) |
| 3547 | struct linebuffer *plb; |
| 3548 | FILE *inf; |
| 3549 | { |
| 3550 | char *cp; |
| 3551 | |
| 3552 | do |
| 3553 | { |
| 3554 | for (cp = plb->buffer; *cp != '\0'; cp++) |
| 3555 | if (cp[0] == '*' && cp[1] == '/') |
| 3556 | return; |
| 3557 | lineno++; |
| 3558 | linecharno += readline (plb, inf); |
| 3559 | } |
| 3560 | while (!feof(inf)); |
| 3561 | } |
| 3562 | |
| 3563 | /* |
| 3564 | * A predicate definition is added if it matches: |
| 3565 | * <beginning of line><Prolog Atom><whitespace>( |
| 3566 | * |
| 3567 | * It is added to the tags database if it doesn't match the |
| 3568 | * name of the previous clause header. |
| 3569 | * |
| 3570 | * Return the size of the name of the predicate, or 0 if no header |
| 3571 | * was found. |
| 3572 | */ |
| 3573 | int |
| 3574 | prolog_pred (s, last) |
| 3575 | char *s; |
| 3576 | char *last; /* Name of last clause. */ |
| 3577 | { |
| 3578 | int prolog_atom(); |
| 3579 | int prolog_white(); |
| 3580 | |
| 3581 | int pos; |
| 3582 | int len; |
| 3583 | |
| 3584 | pos = prolog_atom(s, 0); |
| 3585 | if (pos < 1) |
| 3586 | return 0; |
| 3587 | |
| 3588 | len = pos; |
| 3589 | pos += prolog_white(s, pos); |
| 3590 | |
| 3591 | if ((s[pos] == '(') || (s[pos] == '.')) |
| 3592 | { |
| 3593 | if (s[pos] == '(') |
| 3594 | pos++; |
| 3595 | |
| 3596 | /* Save only the first clause. */ |
| 3597 | if ((last == NULL) || |
| 3598 | (len != strlen(last)) || |
| 3599 | (strncmp(s, last, len) != 0)) |
| 3600 | { |
| 3601 | pfnote ((CTAGS) ? savenstr (s, len) : NULL, TRUE, |
| 3602 | s, pos, lineno, linecharno); |
| 3603 | return len; |
| 3604 | } |
| 3605 | } |
| 3606 | return 0; |
| 3607 | } |
| 3608 | |
| 3609 | /* |
| 3610 | * Consume a Prolog atom. |
| 3611 | * Return the number of bytes consumed, or -1 if there was an error. |
| 3612 | * |
| 3613 | * A prolog atom, in this context, could be one of: |
| 3614 | * - An alphanumeric sequence, starting with a lower case letter. |
| 3615 | * - A quoted arbitrary string. Single quotes can escape themselves. |
| 3616 | * Backslash quotes everything. |
| 3617 | */ |
| 3618 | int |
| 3619 | prolog_atom (s, pos) |
| 3620 | char *s; |
| 3621 | int pos; |
| 3622 | { |
| 3623 | int origpos; |
| 3624 | |
| 3625 | origpos = pos; |
| 3626 | |
| 3627 | if (islower(s[pos]) || (s[pos] == '_')) |
| 3628 | { |
| 3629 | /* The atom is unquoted. */ |
| 3630 | pos++; |
| 3631 | while (isalnum(s[pos]) || (s[pos] == '_')) |
| 3632 | { |
| 3633 | pos++; |
| 3634 | } |
| 3635 | return pos - origpos; |
| 3636 | } |
| 3637 | else if (s[pos] == '\'') |
| 3638 | { |
| 3639 | pos++; |
| 3640 | |
| 3641 | while (1) |
| 3642 | { |
| 3643 | if (s[pos] == '\'') |
| 3644 | { |
| 3645 | pos++; |
| 3646 | if (s[pos] != '\'') |
| 3647 | break; |
| 3648 | pos++; /* A double quote */ |
| 3649 | } |
| 3650 | else if (s[pos] == '\0') |
| 3651 | /* Multiline quoted atoms are ignored. */ |
| 3652 | return -1; |
| 3653 | else if (s[pos] == '\\') |
| 3654 | { |
| 3655 | if (s[pos+1] == '\0') |
| 3656 | return -1; |
| 3657 | pos += 2; |
| 3658 | } |
| 3659 | else |
| 3660 | pos++; |
| 3661 | } |
| 3662 | return pos - origpos; |
| 3663 | } |
| 3664 | else |
| 3665 | return -1; |
| 3666 | } |
| 3667 | |
| 3668 | /* Consume whitespace. Return the number of bytes eaten. */ |
| 3669 | int |
| 3670 | prolog_white (s, pos) |
| 3671 | char *s; |
| 3672 | int pos; |
| 3673 | { |
| 3674 | int origpos; |
| 3675 | |
| 3676 | origpos = pos; |
| 3677 | |
| 3678 | while (isspace(s[pos])) |
| 3679 | pos++; |
| 3680 | |
| 3681 | return pos - origpos; |
| 3682 | } |
| 3683 | \f |
| 3684 | /* |
| 3685 | * Support for Erlang -- Anders Lindgren, Feb 1996. |
| 3686 | * |
| 3687 | * Generates tags for functions, defines, and records. |
| 3688 | * |
| 3689 | * Assumes that Erlang functions start at column 0. |
| 3690 | */ |
| 3691 | void |
| 3692 | Erlang_functions (inf) |
| 3693 | FILE *inf; |
| 3694 | { |
| 3695 | int erlang_func (); |
| 3696 | void erlang_attribute (); |
| 3697 | |
| 3698 | char * last; |
| 3699 | int len; |
| 3700 | int allocated; |
| 3701 | |
| 3702 | allocated = 0; |
| 3703 | len = 0; |
| 3704 | last = NULL; |
| 3705 | |
| 3706 | lineno = 0; |
| 3707 | linecharno = 0; |
| 3708 | charno = 0; |
| 3709 | |
| 3710 | while (!feof (inf)) |
| 3711 | { |
| 3712 | lineno++; |
| 3713 | linecharno += charno; |
| 3714 | charno = readline (&lb, inf); |
| 3715 | dbp = lb.buffer; |
| 3716 | if (dbp[0] == '\0') /* Empty line */ |
| 3717 | continue; |
| 3718 | else if (isspace (dbp[0])) /* Not function nor attribute */ |
| 3719 | continue; |
| 3720 | else if (dbp[0] == '%') /* comment */ |
| 3721 | continue; |
| 3722 | else if (dbp[0] == '"') /* Sometimes, strings start in column one */ |
| 3723 | continue; |
| 3724 | else if (dbp[0] == '-') /* attribute, e.g. "-define" */ |
| 3725 | { |
| 3726 | erlang_attribute(dbp); |
| 3727 | last = NULL; |
| 3728 | } |
| 3729 | else if (len = erlang_func (dbp, last)) |
| 3730 | { |
| 3731 | /* |
| 3732 | * Function. Store the function name so that we only |
| 3733 | * generates a tag for the first clause. |
| 3734 | */ |
| 3735 | if (last == NULL) |
| 3736 | last = xnew(len + 1, char); |
| 3737 | else if (len + 1 > allocated) |
| 3738 | last = (char *) xrealloc(last, len + 1); |
| 3739 | allocated = len + 1; |
| 3740 | strncpy (last, dbp, len); |
| 3741 | last[len] = '\0'; |
| 3742 | } |
| 3743 | } |
| 3744 | } |
| 3745 | |
| 3746 | |
| 3747 | /* |
| 3748 | * A function definition is added if it matches: |
| 3749 | * <beginning of line><Erlang Atom><whitespace>( |
| 3750 | * |
| 3751 | * It is added to the tags database if it doesn't match the |
| 3752 | * name of the previous clause header. |
| 3753 | * |
| 3754 | * Return the size of the name of the function, or 0 if no function |
| 3755 | * was found. |
| 3756 | */ |
| 3757 | int |
| 3758 | erlang_func (s, last) |
| 3759 | char *s; |
| 3760 | char *last; /* Name of last clause. */ |
| 3761 | { |
| 3762 | int erlang_atom (); |
| 3763 | int erlang_white (); |
| 3764 | |
| 3765 | int pos; |
| 3766 | int len; |
| 3767 | |
| 3768 | pos = erlang_atom(s, 0); |
| 3769 | if (pos < 1) |
| 3770 | return 0; |
| 3771 | |
| 3772 | len = pos; |
| 3773 | pos += erlang_white(s, pos); |
| 3774 | |
| 3775 | if (s[pos++] == '(') |
| 3776 | { |
| 3777 | /* Save only the first clause. */ |
| 3778 | if ((last == NULL) || |
| 3779 | (len != strlen(last)) || |
| 3780 | (strncmp(s, last, len) != 0)) |
| 3781 | { |
| 3782 | pfnote ((CTAGS) ? savenstr (s, len) : NULL, TRUE, |
| 3783 | s, pos, lineno, linecharno); |
| 3784 | return len; |
| 3785 | } |
| 3786 | } |
| 3787 | return 0; |
| 3788 | } |
| 3789 | |
| 3790 | |
| 3791 | /* |
| 3792 | * Handle attributes. Currently, tags are generated for defines |
| 3793 | * and records. |
| 3794 | * |
| 3795 | * They are on the form: |
| 3796 | * -define(foo, bar). |
| 3797 | * -define(Foo(M, N), M+N). |
| 3798 | * -record(graph, {vtab = notable, cyclic = true}). |
| 3799 | */ |
| 3800 | void |
| 3801 | erlang_attribute (s) |
| 3802 | char *s; |
| 3803 | { |
| 3804 | int erlang_atom (); |
| 3805 | int erlang_white (); |
| 3806 | |
| 3807 | int pos; |
| 3808 | int len; |
| 3809 | |
| 3810 | if ((strncmp(s, "-define", 7) == 0) || |
| 3811 | (strncmp(s, "-record", 7) == 0)) |
| 3812 | { |
| 3813 | pos = 7; |
| 3814 | pos += erlang_white(s, pos); |
| 3815 | |
| 3816 | if (s[pos++] == '(') |
| 3817 | { |
| 3818 | pos += erlang_white(s, pos); |
| 3819 | |
| 3820 | if (len = erlang_atom(s, pos)) |
| 3821 | { |
| 3822 | pfnote ((CTAGS) ? savenstr (& s[pos], len) : NULL, TRUE, |
| 3823 | s, pos + len, lineno, linecharno); |
| 3824 | } |
| 3825 | } |
| 3826 | } |
| 3827 | return; |
| 3828 | } |
| 3829 | |
| 3830 | |
| 3831 | /* |
| 3832 | * Consume an Erlang atom (or variable). |
| 3833 | * Return the number of bytes consumed, or -1 if there was an error. |
| 3834 | */ |
| 3835 | int |
| 3836 | erlang_atom (s, pos) |
| 3837 | char *s; |
| 3838 | int pos; |
| 3839 | { |
| 3840 | int origpos; |
| 3841 | |
| 3842 | origpos = pos; |
| 3843 | |
| 3844 | if (isalpha (s[pos]) || s[pos] == '_') |
| 3845 | { |
| 3846 | /* The atom is unquoted. */ |
| 3847 | pos++; |
| 3848 | while (isalnum (s[pos]) || s[pos] == '_') |
| 3849 | pos++; |
| 3850 | return pos - origpos; |
| 3851 | } |
| 3852 | else if (s[pos] == '\'') |
| 3853 | { |
| 3854 | pos++; |
| 3855 | |
| 3856 | while (1) |
| 3857 | { |
| 3858 | if (s[pos] == '\'') |
| 3859 | { |
| 3860 | pos++; |
| 3861 | break; |
| 3862 | } |
| 3863 | else if (s[pos] == '\0') |
| 3864 | /* Multiline quoted atoms are ignored. */ |
| 3865 | return -1; |
| 3866 | else if (s[pos] == '\\') |
| 3867 | { |
| 3868 | if (s[pos+1] == '\0') |
| 3869 | return -1; |
| 3870 | pos += 2; |
| 3871 | } |
| 3872 | else |
| 3873 | pos++; |
| 3874 | } |
| 3875 | return pos - origpos; |
| 3876 | } |
| 3877 | else |
| 3878 | return -1; |
| 3879 | } |
| 3880 | |
| 3881 | /* Consume whitespace. Return the number of bytes eaten */ |
| 3882 | int |
| 3883 | erlang_white (s, pos) |
| 3884 | char *s; |
| 3885 | int pos; |
| 3886 | { |
| 3887 | int origpos; |
| 3888 | |
| 3889 | origpos = pos; |
| 3890 | |
| 3891 | while (isspace (s[pos])) |
| 3892 | pos++; |
| 3893 | |
| 3894 | return pos - origpos; |
| 3895 | } |
| 3896 | \f |
| 3897 | #ifdef ETAGS_REGEXPS |
| 3898 | /* Take a string like "/blah/" and turn it into "blah", making sure |
| 3899 | that the first and last characters are the same, and handling |
| 3900 | quoted separator characters. Actually, stops on the occurrence of |
| 3901 | an unquoted separator. Also turns "\t" into a Tab character. |
| 3902 | Returns pointer to terminating separator. Works in place. Null |
| 3903 | terminates name string. */ |
| 3904 | char * |
| 3905 | scan_separators (name) |
| 3906 | char *name; |
| 3907 | { |
| 3908 | char sep = name[0]; |
| 3909 | char *copyto = name; |
| 3910 | logical quoted = FALSE; |
| 3911 | |
| 3912 | for (++name; *name != '\0'; ++name) |
| 3913 | { |
| 3914 | if (quoted) |
| 3915 | { |
| 3916 | if (*name == 't') |
| 3917 | *copyto++ = '\t'; |
| 3918 | else if (*name == sep) |
| 3919 | *copyto++ = sep; |
| 3920 | else |
| 3921 | { |
| 3922 | /* Something else is quoted, so preserve the quote. */ |
| 3923 | *copyto++ = '\\'; |
| 3924 | *copyto++ = *name; |
| 3925 | } |
| 3926 | quoted = FALSE; |
| 3927 | } |
| 3928 | else if (*name == '\\') |
| 3929 | quoted = TRUE; |
| 3930 | else if (*name == sep) |
| 3931 | break; |
| 3932 | else |
| 3933 | *copyto++ = *name; |
| 3934 | } |
| 3935 | |
| 3936 | /* Terminate copied string. */ |
| 3937 | *copyto = '\0'; |
| 3938 | return name; |
| 3939 | } |
| 3940 | |
| 3941 | /* Turn a name, which is an ed-style (but Emacs syntax) regular |
| 3942 | expression, into a real regular expression by compiling it. */ |
| 3943 | void |
| 3944 | add_regex (regexp_pattern) |
| 3945 | char *regexp_pattern; |
| 3946 | { |
| 3947 | char *name; |
| 3948 | const char *err; |
| 3949 | struct re_pattern_buffer *patbuf; |
| 3950 | |
| 3951 | if (regexp_pattern == NULL) |
| 3952 | { |
| 3953 | /* Remove existing regexps. */ |
| 3954 | num_patterns = 0; |
| 3955 | patterns = NULL; |
| 3956 | return; |
| 3957 | } |
| 3958 | |
| 3959 | if (regexp_pattern[0] == '\0') |
| 3960 | { |
| 3961 | error ("missing regexp", 0); |
| 3962 | return; |
| 3963 | } |
| 3964 | if (regexp_pattern[strlen(regexp_pattern)-1] != regexp_pattern[0]) |
| 3965 | { |
| 3966 | error ("%s: unterminated regexp", regexp_pattern); |
| 3967 | return; |
| 3968 | } |
| 3969 | name = scan_separators (regexp_pattern); |
| 3970 | if (regexp_pattern[0] == '\0') |
| 3971 | { |
| 3972 | error ("null regexp", 0); |
| 3973 | return; |
| 3974 | } |
| 3975 | (void) scan_separators (name); |
| 3976 | |
| 3977 | patbuf = xnew (1, struct re_pattern_buffer); |
| 3978 | patbuf->translate = NULL; |
| 3979 | patbuf->fastmap = NULL; |
| 3980 | patbuf->buffer = NULL; |
| 3981 | patbuf->allocated = 0; |
| 3982 | |
| 3983 | err = re_compile_pattern (regexp_pattern, strlen (regexp_pattern), patbuf); |
| 3984 | if (err != NULL) |
| 3985 | { |
| 3986 | error ("%s while compiling pattern", err); |
| 3987 | return; |
| 3988 | } |
| 3989 | |
| 3990 | num_patterns += 1; |
| 3991 | if (num_patterns == 1) |
| 3992 | patterns = xnew (1, struct pattern); |
| 3993 | else |
| 3994 | patterns = ((struct pattern *) |
| 3995 | xrealloc (patterns, |
| 3996 | (num_patterns * sizeof (struct pattern)))); |
| 3997 | patterns[num_patterns - 1].pattern = patbuf; |
| 3998 | patterns[num_patterns - 1].name_pattern = savestr (name); |
| 3999 | patterns[num_patterns - 1].error_signaled = FALSE; |
| 4000 | } |
| 4001 | |
| 4002 | /* |
| 4003 | * Do the substitutions indicated by the regular expression and |
| 4004 | * arguments. |
| 4005 | */ |
| 4006 | char * |
| 4007 | substitute (in, out, regs) |
| 4008 | char *in, *out; |
| 4009 | struct re_registers *regs; |
| 4010 | { |
| 4011 | char *result = NULL, *t; |
| 4012 | int size = 0; |
| 4013 | |
| 4014 | /* Pass 1: figure out how much size to allocate. */ |
| 4015 | for (t = out; *t; ++t) |
| 4016 | { |
| 4017 | if (*t == '\\') |
| 4018 | { |
| 4019 | ++t; |
| 4020 | if (!*t) |
| 4021 | { |
| 4022 | fprintf (stderr, "%s: pattern substitution ends prematurely\n", |
| 4023 | progname); |
| 4024 | return NULL; |
| 4025 | } |
| 4026 | if (isdigit (*t)) |
| 4027 | { |
| 4028 | int dig = *t - '0'; |
| 4029 | size += regs->end[dig] - regs->start[dig]; |
| 4030 | } |
| 4031 | } |
| 4032 | } |
| 4033 | |
| 4034 | /* Allocate space and do the substitutions. */ |
| 4035 | result = xnew (size + 1, char); |
| 4036 | size = 0; |
| 4037 | for (; *out; ++out) |
| 4038 | { |
| 4039 | if (*out == '\\') |
| 4040 | { |
| 4041 | ++out; |
| 4042 | if (isdigit (*out)) |
| 4043 | { |
| 4044 | /* Using "dig2" satisfies my debugger. Bleah. */ |
| 4045 | int dig2 = *out - '0'; |
| 4046 | strncpy (result + size, in + regs->start[dig2], |
| 4047 | regs->end[dig2] - regs->start[dig2]); |
| 4048 | size += regs->end[dig2] - regs->start[dig2]; |
| 4049 | } |
| 4050 | else |
| 4051 | result[size++] = *out; |
| 4052 | } |
| 4053 | else |
| 4054 | result[size++] = *out; |
| 4055 | } |
| 4056 | result[size] = '\0'; |
| 4057 | |
| 4058 | return result; |
| 4059 | } |
| 4060 | \f |
| 4061 | #endif /* ETAGS_REGEXPS */ |
| 4062 | /* Initialize a linebuffer for use */ |
| 4063 | void |
| 4064 | initbuffer (linebuffer) |
| 4065 | struct linebuffer *linebuffer; |
| 4066 | { |
| 4067 | linebuffer->size = 200; |
| 4068 | linebuffer->buffer = xnew (200, char); |
| 4069 | } |
| 4070 | |
| 4071 | /* |
| 4072 | * Read a line of text from `stream' into `linebuffer'. |
| 4073 | * Return the number of characters read from `stream', |
| 4074 | * which is the length of the line including the newline, if any. |
| 4075 | */ |
| 4076 | long |
| 4077 | readline_internal (linebuffer, stream) |
| 4078 | struct linebuffer *linebuffer; |
| 4079 | register FILE *stream; |
| 4080 | { |
| 4081 | char *buffer = linebuffer->buffer; |
| 4082 | register char *p = linebuffer->buffer; |
| 4083 | register char *pend; |
| 4084 | int chars_deleted; |
| 4085 | |
| 4086 | pend = p + linebuffer->size; /* Separate to avoid 386/IX compiler bug. */ |
| 4087 | |
| 4088 | while (1) |
| 4089 | { |
| 4090 | register int c = getc (stream); |
| 4091 | if (p == pend) |
| 4092 | { |
| 4093 | linebuffer->size *= 2; |
| 4094 | buffer = (char *) xrealloc (buffer, linebuffer->size); |
| 4095 | p += buffer - linebuffer->buffer; |
| 4096 | pend = buffer + linebuffer->size; |
| 4097 | linebuffer->buffer = buffer; |
| 4098 | } |
| 4099 | if (c == EOF) |
| 4100 | { |
| 4101 | *p = '\0'; |
| 4102 | chars_deleted = 0; |
| 4103 | break; |
| 4104 | } |
| 4105 | if (c == '\n') |
| 4106 | { |
| 4107 | if (p > buffer && p[-1] == '\r') |
| 4108 | { |
| 4109 | *--p = '\0'; |
| 4110 | #ifdef DOS_NT |
| 4111 | /* Assume CRLF->LF translation will be performed by Emacs |
| 4112 | when loading this file, so CRs won't appear in the buffer. |
| 4113 | It would be cleaner to compensate within Emacs; |
| 4114 | however, Emacs does not know how many CRs were deleted |
| 4115 | before any given point in the file. */ |
| 4116 | chars_deleted = 1; |
| 4117 | #else |
| 4118 | chars_deleted = 2; |
| 4119 | #endif |
| 4120 | } |
| 4121 | else |
| 4122 | { |
| 4123 | *p = '\0'; |
| 4124 | chars_deleted = 1; |
| 4125 | } |
| 4126 | break; |
| 4127 | } |
| 4128 | *p++ = c; |
| 4129 | } |
| 4130 | |
| 4131 | return p - buffer + chars_deleted; |
| 4132 | } |
| 4133 | |
| 4134 | /* |
| 4135 | * Like readline_internal, above, but try to match the input |
| 4136 | * line against any existing regular expressions. |
| 4137 | */ |
| 4138 | long |
| 4139 | readline (linebuffer, stream) |
| 4140 | struct linebuffer *linebuffer; |
| 4141 | FILE *stream; |
| 4142 | { |
| 4143 | /* Read new line. */ |
| 4144 | long result = readline_internal (linebuffer, stream); |
| 4145 | #ifdef ETAGS_REGEXPS |
| 4146 | int i; |
| 4147 | |
| 4148 | /* Match against all listed patterns. */ |
| 4149 | for (i = 0; i < num_patterns; ++i) |
| 4150 | { |
| 4151 | int match = re_match (patterns[i].pattern, linebuffer->buffer, |
| 4152 | (int)result, 0, &patterns[i].regs); |
| 4153 | switch (match) |
| 4154 | { |
| 4155 | case -2: |
| 4156 | /* Some error. */ |
| 4157 | if (!patterns[i].error_signaled) |
| 4158 | { |
| 4159 | error ("error while matching pattern %d", i); |
| 4160 | patterns[i].error_signaled = TRUE; |
| 4161 | } |
| 4162 | break; |
| 4163 | case -1: |
| 4164 | /* No match. */ |
| 4165 | break; |
| 4166 | default: |
| 4167 | /* Match occurred. Construct a tag. */ |
| 4168 | if (patterns[i].name_pattern[0] != '\0') |
| 4169 | { |
| 4170 | /* Make a named tag. */ |
| 4171 | char *name = substitute (linebuffer->buffer, |
| 4172 | patterns[i].name_pattern, |
| 4173 | &patterns[i].regs); |
| 4174 | if (name != NULL) |
| 4175 | pfnote (name, TRUE, |
| 4176 | linebuffer->buffer, match, lineno, linecharno); |
| 4177 | } |
| 4178 | else |
| 4179 | { |
| 4180 | /* Make an unnamed tag. */ |
| 4181 | pfnote (NULL, TRUE, |
| 4182 | linebuffer->buffer, match, lineno, linecharno); |
| 4183 | } |
| 4184 | break; |
| 4185 | } |
| 4186 | } |
| 4187 | #endif /* ETAGS_REGEXPS */ |
| 4188 | |
| 4189 | return result; |
| 4190 | } |
| 4191 | |
| 4192 | /* |
| 4193 | * Read a file, but do no processing. This is used to do regexp |
| 4194 | * matching on files that have no language defined. |
| 4195 | */ |
| 4196 | void |
| 4197 | just_read_file (inf) |
| 4198 | FILE *inf; |
| 4199 | { |
| 4200 | lineno = 0; |
| 4201 | charno = 0; |
| 4202 | |
| 4203 | while (!feof (inf)) |
| 4204 | { |
| 4205 | ++lineno; |
| 4206 | linecharno = charno; |
| 4207 | charno += readline (&lb, inf) + 1; |
| 4208 | } |
| 4209 | } |
| 4210 | |
| 4211 | \f |
| 4212 | /* |
| 4213 | * Return a pointer to a space of size strlen(cp)+1 allocated |
| 4214 | * with xnew where the string CP has been copied. |
| 4215 | */ |
| 4216 | char * |
| 4217 | savestr (cp) |
| 4218 | char *cp; |
| 4219 | { |
| 4220 | return savenstr (cp, strlen (cp)); |
| 4221 | } |
| 4222 | |
| 4223 | /* |
| 4224 | * Return a pointer to a space of size LEN+1 allocated with xnew where |
| 4225 | * the string CP has been copied for at most the first LEN characters. |
| 4226 | */ |
| 4227 | char * |
| 4228 | savenstr (cp, len) |
| 4229 | char *cp; |
| 4230 | int len; |
| 4231 | { |
| 4232 | register char *dp; |
| 4233 | |
| 4234 | dp = xnew (len + 1, char); |
| 4235 | strncpy (dp, cp, len); |
| 4236 | dp[len] = '\0'; |
| 4237 | return dp; |
| 4238 | } |
| 4239 | |
| 4240 | /* |
| 4241 | * Return the ptr in sp at which the character c last |
| 4242 | * appears; NULL if not found |
| 4243 | * |
| 4244 | * Identical to System V strrchr, included for portability. |
| 4245 | */ |
| 4246 | char * |
| 4247 | etags_strrchr (sp, c) |
| 4248 | register char *sp, c; |
| 4249 | { |
| 4250 | register char *r; |
| 4251 | |
| 4252 | r = NULL; |
| 4253 | do |
| 4254 | { |
| 4255 | if (*sp == c) |
| 4256 | r = sp; |
| 4257 | } while (*sp++); |
| 4258 | return r; |
| 4259 | } |
| 4260 | |
| 4261 | |
| 4262 | /* |
| 4263 | * Return the ptr in sp at which the character c first |
| 4264 | * appears; NULL if not found |
| 4265 | * |
| 4266 | * Identical to System V strchr, included for portability. |
| 4267 | */ |
| 4268 | char * |
| 4269 | etags_strchr (sp, c) |
| 4270 | register char *sp, c; |
| 4271 | { |
| 4272 | do |
| 4273 | { |
| 4274 | if (*sp == c) |
| 4275 | return sp; |
| 4276 | } while (*sp++); |
| 4277 | return NULL; |
| 4278 | } |
| 4279 | |
| 4280 | /* Print error message and exit. */ |
| 4281 | void |
| 4282 | fatal (s1, s2) |
| 4283 | char *s1, *s2; |
| 4284 | { |
| 4285 | error (s1, s2); |
| 4286 | exit (BAD); |
| 4287 | } |
| 4288 | |
| 4289 | void |
| 4290 | pfatal (s1) |
| 4291 | char *s1; |
| 4292 | { |
| 4293 | perror (s1); |
| 4294 | exit (BAD); |
| 4295 | } |
| 4296 | |
| 4297 | void |
| 4298 | suggest_asking_for_help () |
| 4299 | { |
| 4300 | fprintf (stderr, "\tTry `%s --help' for a complete list of options.\n", |
| 4301 | progname); |
| 4302 | exit (BAD); |
| 4303 | } |
| 4304 | |
| 4305 | /* Print error message. `s1' is printf control string, `s2' is arg for it. */ |
| 4306 | void |
| 4307 | error (s1, s2) |
| 4308 | char *s1, *s2; |
| 4309 | { |
| 4310 | fprintf (stderr, "%s: ", progname); |
| 4311 | fprintf (stderr, s1, s2); |
| 4312 | fprintf (stderr, "\n"); |
| 4313 | } |
| 4314 | |
| 4315 | /* Return a newly-allocated string whose contents |
| 4316 | concatenate those of s1, s2, s3. */ |
| 4317 | char * |
| 4318 | concat (s1, s2, s3) |
| 4319 | char *s1, *s2, *s3; |
| 4320 | { |
| 4321 | int len1 = strlen (s1), len2 = strlen (s2), len3 = strlen (s3); |
| 4322 | char *result = xnew (len1 + len2 + len3 + 1, char); |
| 4323 | |
| 4324 | strcpy (result, s1); |
| 4325 | strcpy (result + len1, s2); |
| 4326 | strcpy (result + len1 + len2, s3); |
| 4327 | result[len1 + len2 + len3] = '\0'; |
| 4328 | |
| 4329 | return result; |
| 4330 | } |
| 4331 | \f |
| 4332 | /* Does the same work as the system V getcwd, but does not need to |
| 4333 | guess the buffer size in advance. */ |
| 4334 | char * |
| 4335 | etags_getcwd () |
| 4336 | { |
| 4337 | #ifdef HAVE_GETCWD |
| 4338 | int bufsize = 200; |
| 4339 | char *path = xnew (bufsize, char); |
| 4340 | char *p; |
| 4341 | |
| 4342 | while (getcwd (path, bufsize) == NULL) |
| 4343 | { |
| 4344 | if (errno != ERANGE) |
| 4345 | pfatal ("getcwd"); |
| 4346 | bufsize *= 2; |
| 4347 | path = xnew (bufsize, char); |
| 4348 | } |
| 4349 | |
| 4350 | /* Convert backslashes to slashes. */ |
| 4351 | #if WINDOWSNT |
| 4352 | for (p = path; *p != '\0'; p++) |
| 4353 | if (*p == '\\') |
| 4354 | *p = '/'; |
| 4355 | #endif |
| 4356 | |
| 4357 | return path; |
| 4358 | |
| 4359 | #else /* not HAVE_GETCWD */ |
| 4360 | #ifdef MSDOS |
| 4361 | char *p, path[MAXPATHLEN + 1]; /* Fixed size is safe on MSDOS. */ |
| 4362 | |
| 4363 | getwd (path); |
| 4364 | |
| 4365 | for (p = path; *p != '\0'; p++) |
| 4366 | if (*p == '\\') |
| 4367 | *p = '/'; |
| 4368 | else |
| 4369 | *p = lowcase (*p); |
| 4370 | |
| 4371 | return strdup (path); |
| 4372 | #else /* not MSDOS */ |
| 4373 | struct linebuffer path; |
| 4374 | FILE *pipe; |
| 4375 | |
| 4376 | initbuffer (&path); |
| 4377 | pipe = (FILE *) popen ("pwd 2>/dev/null", "r"); |
| 4378 | if (pipe == NULL || readline_internal (&path, pipe) == 0) |
| 4379 | pfatal ("pwd"); |
| 4380 | pclose (pipe); |
| 4381 | |
| 4382 | return path.buffer; |
| 4383 | #endif /* not MSDOS */ |
| 4384 | #endif /* not HAVE_GETCWD */ |
| 4385 | } |
| 4386 | |
| 4387 | /* Return a newly allocated string containing the filename |
| 4388 | of FILE relative to the absolute directory DIR (which |
| 4389 | should end with a slash). */ |
| 4390 | char * |
| 4391 | relative_filename (file, dir) |
| 4392 | char *file, *dir; |
| 4393 | { |
| 4394 | char *fp, *dp, *abs, *res; |
| 4395 | |
| 4396 | /* Find the common root of file and dir. */ |
| 4397 | abs = absolute_filename (file, cwd); |
| 4398 | fp = abs; |
| 4399 | dp = dir; |
| 4400 | while (*fp++ == *dp++) |
| 4401 | continue; |
| 4402 | do |
| 4403 | { |
| 4404 | fp--; |
| 4405 | dp--; |
| 4406 | } |
| 4407 | while (*fp != '/'); |
| 4408 | |
| 4409 | /* Build a sequence of "../" strings for the resulting relative filename. */ |
| 4410 | for (dp = etags_strchr (dp + 1, '/'), res = ""; |
| 4411 | dp != NULL; |
| 4412 | dp = etags_strchr (dp + 1, '/')) |
| 4413 | { |
| 4414 | res = concat (res, "../", ""); |
| 4415 | } |
| 4416 | |
| 4417 | /* Add the filename relative to the common root of file and dir. */ |
| 4418 | res = concat (res, fp + 1, ""); |
| 4419 | free (abs); |
| 4420 | |
| 4421 | return res; |
| 4422 | } |
| 4423 | |
| 4424 | /* Return a newly allocated string containing the |
| 4425 | absolute filename of FILE given CWD (which should |
| 4426 | end with a slash). */ |
| 4427 | char * |
| 4428 | absolute_filename (file, cwd) |
| 4429 | char *file, *cwd; |
| 4430 | { |
| 4431 | char *slashp, *cp, *res; |
| 4432 | |
| 4433 | if (absolutefn (file)) |
| 4434 | res = concat (file, "", ""); |
| 4435 | #ifdef DOS_NT |
| 4436 | /* We don't support non-absolute filenames with a drive |
| 4437 | letter, like `d:NAME' (it's too much hassle). */ |
| 4438 | else if (file[1] == ':') |
| 4439 | fatal ("%s: relative filenames with drive letters not supported", file); |
| 4440 | #endif |
| 4441 | else |
| 4442 | res = concat (cwd, file, ""); |
| 4443 | |
| 4444 | /* Delete the "/dirname/.." and "/." substrings. */ |
| 4445 | slashp = etags_strchr (res, '/'); |
| 4446 | while (slashp != NULL && slashp[0] != '\0') |
| 4447 | { |
| 4448 | if (slashp[1] == '.') |
| 4449 | { |
| 4450 | if (slashp[2] == '.' |
| 4451 | && (slashp[3] == '/' || slashp[3] == '\0')) |
| 4452 | { |
| 4453 | cp = slashp; |
| 4454 | do |
| 4455 | cp--; |
| 4456 | while (cp >= res && !absolutefn (cp)); |
| 4457 | if (*cp == '/') |
| 4458 | { |
| 4459 | strcpy (cp, slashp + 3); |
| 4460 | } |
| 4461 | #ifdef DOS_NT |
| 4462 | /* Under MSDOS and NT we get `d:/NAME' as absolute |
| 4463 | filename, so the luser could say `d:/../NAME'. |
| 4464 | We silently treat this as `d:/NAME'. */ |
| 4465 | else if (cp[1] == ':') |
| 4466 | strcpy (cp + 3, slashp + 4); |
| 4467 | #endif |
| 4468 | else /* else (cp == res) */ |
| 4469 | { |
| 4470 | if (slashp[3] != '\0') |
| 4471 | strcpy (cp, slashp + 4); |
| 4472 | else |
| 4473 | return "."; |
| 4474 | } |
| 4475 | slashp = cp; |
| 4476 | continue; |
| 4477 | } |
| 4478 | else if (slashp[2] == '/' || slashp[2] == '\0') |
| 4479 | { |
| 4480 | strcpy (slashp, slashp + 2); |
| 4481 | continue; |
| 4482 | } |
| 4483 | } |
| 4484 | |
| 4485 | slashp = etags_strchr (slashp + 1, '/'); |
| 4486 | } |
| 4487 | |
| 4488 | return res; |
| 4489 | } |
| 4490 | |
| 4491 | /* Return a newly allocated string containing the absolute |
| 4492 | filename of dir where FILE resides given CWD (which should |
| 4493 | end with a slash). */ |
| 4494 | char * |
| 4495 | absolute_dirname (file, cwd) |
| 4496 | char *file, *cwd; |
| 4497 | { |
| 4498 | char *slashp, *res; |
| 4499 | char save; |
| 4500 | #ifdef DOS_NT |
| 4501 | char *p; |
| 4502 | |
| 4503 | for (p = file; *p != '\0'; p++) |
| 4504 | if (*p == '\\') |
| 4505 | *p = '/'; |
| 4506 | #endif |
| 4507 | |
| 4508 | slashp = etags_strrchr (file, '/'); |
| 4509 | if (slashp == NULL) |
| 4510 | return cwd; |
| 4511 | save = slashp[1]; |
| 4512 | slashp[1] = '\0'; |
| 4513 | res = absolute_filename (file, cwd); |
| 4514 | slashp[1] = save; |
| 4515 | |
| 4516 | return res; |
| 4517 | } |
| 4518 | |
| 4519 | /* Like malloc but get fatal error if memory is exhausted. */ |
| 4520 | long * |
| 4521 | xmalloc (size) |
| 4522 | unsigned int size; |
| 4523 | { |
| 4524 | long *result = (long *) malloc (size); |
| 4525 | if (result == NULL) |
| 4526 | fatal ("virtual memory exhausted", 0); |
| 4527 | return result; |
| 4528 | } |
| 4529 | |
| 4530 | long * |
| 4531 | xrealloc (ptr, size) |
| 4532 | char *ptr; |
| 4533 | unsigned int size; |
| 4534 | { |
| 4535 | long *result = (long *) realloc (ptr, size); |
| 4536 | if (result == NULL) |
| 4537 | fatal ("virtual memory exhausted"); |
| 4538 | return result; |
| 4539 | } |