| 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 | This file is not considered part of GNU Emacs. |
| 5 | |
| 6 | This program is free software; you can redistribute it and/or modify |
| 7 | it under the terms of the GNU General Public License as published by |
| 8 | the Free Software Foundation; either version 2 of the License, or |
| 9 | (at your option) any later version. |
| 10 | |
| 11 | This program is distributed in the hope that it will be useful, |
| 12 | but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 13 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 14 | GNU General Public License for more details. |
| 15 | |
| 16 | You should have received a copy of the GNU General Public License |
| 17 | along with this program; if not, write to the Free Software |
| 18 | Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */ |
| 19 | |
| 20 | /* |
| 21 | * Authors: |
| 22 | * Ctags originally by Ken Arnold. |
| 23 | * Fortran added by Jim Kleckner. |
| 24 | * Ed Pelegri-Llopart added C typedefs. |
| 25 | * Gnu Emacs TAGS format and modifications by RMS? |
| 26 | * Sam Kendall added C++. |
| 27 | * Francesco Potorti` reorganised C and C++ based on work by Joe Wells. |
| 28 | #ifdef ETAGS_REGEXPS |
| 29 | * Regexp tags by Tom Tromey. |
| 30 | #endif |
| 31 | * |
| 32 | * Francesco Potorti` (pot@cnuce.cnr.it) is the current maintainer. |
| 33 | */ |
| 34 | |
| 35 | char pot_etags_version[] = "@(#) pot revision number is 11.28"; |
| 36 | |
| 37 | #define TRUE 1 |
| 38 | #define FALSE 0 |
| 39 | #ifndef DEBUG |
| 40 | # define DEBUG FALSE |
| 41 | #endif |
| 42 | |
| 43 | #ifdef MSDOS |
| 44 | #include <fcntl.h> |
| 45 | #include <sys/param.h> |
| 46 | #endif /* MSDOS */ |
| 47 | |
| 48 | #ifdef WINDOWSNT |
| 49 | #include <stdlib.h> |
| 50 | #include <fcntl.h> |
| 51 | #include <string.h> |
| 52 | #define MAXPATHLEN _MAX_PATH |
| 53 | #endif |
| 54 | |
| 55 | #ifdef HAVE_CONFIG_H |
| 56 | #include <config.h> |
| 57 | /* On some systems, Emacs defines static as nothing for the sake |
| 58 | of unexec. We don't want that here since we don't use unexec. */ |
| 59 | #undef static |
| 60 | #endif |
| 61 | |
| 62 | #include <stdio.h> |
| 63 | #include <ctype.h> |
| 64 | #include <errno.h> |
| 65 | #ifndef errno |
| 66 | extern int errno; |
| 67 | #endif |
| 68 | #include <sys/types.h> |
| 69 | #include <sys/stat.h> |
| 70 | |
| 71 | #if !defined (S_ISREG) && defined (S_IFREG) |
| 72 | # define S_ISREG(m) (((m) & S_IFMT) == S_IFREG) |
| 73 | #endif |
| 74 | |
| 75 | #include <getopt.h> |
| 76 | |
| 77 | #ifdef ETAGS_REGEXPS |
| 78 | #include <regex.h> |
| 79 | #endif /* ETAGS_REGEXPS */ |
| 80 | |
| 81 | /* Define CTAGS to make the program "ctags" compatible with the usual one. |
| 82 | Let it undefined to make the program "etags", which makes emacs-style |
| 83 | tag tables and tags typedefs, #defines and struct/union/enum by default. */ |
| 84 | #ifdef CTAGS |
| 85 | # undef CTAGS |
| 86 | # define CTAGS TRUE |
| 87 | #else |
| 88 | # define CTAGS FALSE |
| 89 | #endif |
| 90 | |
| 91 | /* Exit codes for success and failure. */ |
| 92 | #ifdef VMS |
| 93 | #define GOOD 1 |
| 94 | #define BAD 0 |
| 95 | #else |
| 96 | #define GOOD 0 |
| 97 | #define BAD 1 |
| 98 | #endif |
| 99 | |
| 100 | /* C extensions. */ |
| 101 | #define C_PLPL 0x00001 /* C++ */ |
| 102 | #define C_STAR 0x00003 /* C* */ |
| 103 | #define YACC 0x10000 /* yacc file */ |
| 104 | |
| 105 | #define streq(s,t) (strcmp (s, t) == 0) |
| 106 | #define strneq(s,t,n) (strncmp (s, t, n) == 0) |
| 107 | |
| 108 | #define iswhite(arg) (_wht[arg]) /* T if char is white */ |
| 109 | #define begtoken(arg) (_btk[arg]) /* T if char can start token */ |
| 110 | #define intoken(arg) (_itk[arg]) /* T if char can be in token */ |
| 111 | #define endtoken(arg) (_etk[arg]) /* T if char ends tokens */ |
| 112 | |
| 113 | /* |
| 114 | * xnew -- allocate storage |
| 115 | * |
| 116 | * SYNOPSIS: Type *xnew (int n, Type); |
| 117 | */ |
| 118 | #define xnew(n,Type) ((Type *) xmalloc ((n) * sizeof (Type))) |
| 119 | |
| 120 | typedef int logical; |
| 121 | |
| 122 | typedef struct nd_st |
| 123 | { /* sorting structure */ |
| 124 | char *name; /* function or type name */ |
| 125 | char *file; /* file name */ |
| 126 | logical is_func; /* use pattern or line no */ |
| 127 | logical named; /* list name separately */ |
| 128 | logical been_warned; /* set if noticed dup */ |
| 129 | int lno; /* line number tag is on */ |
| 130 | long cno; /* character number line starts on */ |
| 131 | char *pat; /* search pattern */ |
| 132 | struct nd_st *left, *right; /* left and right sons */ |
| 133 | } NODE; |
| 134 | |
| 135 | extern char *getenv (); |
| 136 | |
| 137 | char *concat (); |
| 138 | char *savenstr (), *savestr (); |
| 139 | char *etags_strchr (), *etags_strrchr (); |
| 140 | char *etags_getcwd (); |
| 141 | char *relative_filename (), *absolute_filename (), *absolute_dirname (); |
| 142 | long *xmalloc (), *xrealloc (); |
| 143 | |
| 144 | typedef void Lang_function (); |
| 145 | #if FALSE /* many compilers barf on this */ |
| 146 | Lang_function Asm_labels; |
| 147 | Lang_function default_C_entries; |
| 148 | Lang_function C_entries; |
| 149 | Lang_function Cplusplus_entries; |
| 150 | Lang_function Cstar_entries; |
| 151 | Lang_function Fortran_functions; |
| 152 | Lang_function Yacc_entries; |
| 153 | Lang_function Lisp_functions; |
| 154 | Lang_function Pascal_functions; |
| 155 | Lang_function Prolog_functions; |
| 156 | Lang_function Scheme_functions; |
| 157 | Lang_function TeX_functions; |
| 158 | Lang_function just_read_file; |
| 159 | #else /* so let's write it this way */ |
| 160 | void Asm_labels (); |
| 161 | void default_C_entries (); |
| 162 | void C_entries (); |
| 163 | void Cplusplus_entries (); |
| 164 | void Cstar_entries (); |
| 165 | void Fortran_functions (); |
| 166 | void Yacc_entries (); |
| 167 | void Lisp_functions (); |
| 168 | void Pascal_functions (); |
| 169 | void Prolog_functions (); |
| 170 | void Scheme_functions (); |
| 171 | void TeX_functions (); |
| 172 | void just_read_file (); |
| 173 | #endif |
| 174 | |
| 175 | logical get_language (); |
| 176 | int total_size_of_entries (); |
| 177 | long readline (); |
| 178 | long readline_internal (); |
| 179 | #ifdef ETAGS_REGEXPS |
| 180 | void add_regex (); |
| 181 | #endif |
| 182 | void add_node (); |
| 183 | void error (); |
| 184 | void fatal (), pfatal (); |
| 185 | void find_entries (); |
| 186 | void free_tree (); |
| 187 | void getit (); |
| 188 | void init (); |
| 189 | void initbuffer (); |
| 190 | void pfnote (); |
| 191 | void process_file (); |
| 192 | void put_entries (); |
| 193 | void takeprec (); |
| 194 | |
| 195 | \f |
| 196 | char searchar = '/'; /* use /.../ searches */ |
| 197 | |
| 198 | int lineno; /* line number of current line */ |
| 199 | long charno; /* current character number */ |
| 200 | |
| 201 | long linecharno; /* charno of start of line; not used by C, |
| 202 | but by every other language. */ |
| 203 | |
| 204 | char *curfile; /* current input file name */ |
| 205 | char *tagfile; /* output file */ |
| 206 | char *progname; /* name this program was invoked with */ |
| 207 | char *cwd; /* current working directory */ |
| 208 | char *tagfiledir; /* directory of tagfile */ |
| 209 | |
| 210 | FILE *tagf; /* ioptr for tags file */ |
| 211 | NODE *head; /* the head of the binary tree of tags */ |
| 212 | |
| 213 | /* |
| 214 | * A `struct linebuffer' is a structure which holds a line of text. |
| 215 | * `readline' reads a line from a stream into a linebuffer and works |
| 216 | * regardless of the length of the line. |
| 217 | */ |
| 218 | struct linebuffer |
| 219 | { |
| 220 | long size; |
| 221 | char *buffer; |
| 222 | }; |
| 223 | |
| 224 | struct linebuffer lb; /* the current line */ |
| 225 | struct linebuffer token_name; /* used by C_entries as temporary area */ |
| 226 | struct |
| 227 | { |
| 228 | long linepos; |
| 229 | struct linebuffer lb; /* used by C_entries instead of lb */ |
| 230 | } lbs[2]; |
| 231 | |
| 232 | /* boolean "functions" (see init) */ |
| 233 | logical _wht[0177], _etk[0177], _itk[0177], _btk[0177]; |
| 234 | char |
| 235 | *white = " \f\t\n\013", /* white chars */ |
| 236 | *endtk = " \t\n\013\"'#()[]{}=-+%*/&|^~!<>;,.:?", /* token ending chars */ |
| 237 | /* token starting chars */ |
| 238 | *begtk = "ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz$~", |
| 239 | /* valid in-token chars */ |
| 240 | *intk = "ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz$0123456789"; |
| 241 | |
| 242 | logical append_to_tagfile; /* -a: append to tags */ |
| 243 | /* The following three default to TRUE for etags, but to FALSE for ctags. */ |
| 244 | logical typedefs; /* -t: create tags for typedefs */ |
| 245 | logical typedefs_and_cplusplus; /* -T: create tags for typedefs, level */ |
| 246 | /* 0 struct/enum/union decls, and C++ */ |
| 247 | /* member functions. */ |
| 248 | logical constantypedefs; /* -d: create tags for C #define and enum */ |
| 249 | /* constants. Enum consts not implemented. */ |
| 250 | /* -D: opposite of -d. Default under ctags. */ |
| 251 | logical update; /* -u: update tags */ |
| 252 | logical vgrind_style; /* -v: create vgrind style index output */ |
| 253 | logical no_warnings; /* -w: suppress warnings */ |
| 254 | logical cxref_style; /* -x: create cxref style output */ |
| 255 | logical cplusplus; /* .[hc] means C++, not C */ |
| 256 | logical noindentypedefs; /* -I: ignore indentation in C */ |
| 257 | #define permit_duplicates TRUE /* allow duplicate tags */ |
| 258 | |
| 259 | struct option longopts[] = |
| 260 | { |
| 261 | { "append", no_argument, NULL, 'a' }, |
| 262 | { "backward-search", no_argument, NULL, 'B' }, |
| 263 | { "c++", no_argument, NULL, 'C' }, |
| 264 | { "cxref", no_argument, NULL, 'x' }, |
| 265 | { "defines", no_argument, NULL, 'd' }, |
| 266 | { "help", no_argument, NULL, 'h' }, |
| 267 | { "help", no_argument, NULL, 'H' }, |
| 268 | { "ignore-indentation", no_argument, NULL, 'I' }, |
| 269 | { "include", required_argument, NULL, 'i' }, |
| 270 | { "language", required_argument, NULL, 'l' }, |
| 271 | { "no-defines", no_argument, NULL, 'D' }, |
| 272 | { "no-regex", no_argument, NULL, 'R' }, |
| 273 | { "no-warn", no_argument, NULL, 'w' }, |
| 274 | { "output", required_argument, NULL, 'o' }, |
| 275 | { "regex", required_argument, NULL, 'r' }, |
| 276 | { "typedefs", no_argument, NULL, 't' }, |
| 277 | { "typedefs-and-c++", no_argument, NULL, 'T' }, |
| 278 | { "update", no_argument, NULL, 'u' }, |
| 279 | { "version", no_argument, NULL, 'V' }, |
| 280 | { "vgrind", no_argument, NULL, 'v' }, |
| 281 | { 0 } |
| 282 | }; |
| 283 | |
| 284 | #ifdef ETAGS_REGEXPS |
| 285 | /* Structure defining a regular expression. Elements are |
| 286 | the compiled pattern, and the name string. */ |
| 287 | struct pattern |
| 288 | { |
| 289 | struct re_pattern_buffer *pattern; |
| 290 | struct re_registers regs; |
| 291 | char *name_pattern; |
| 292 | logical error_signaled; |
| 293 | }; |
| 294 | |
| 295 | /* Number of regexps found. */ |
| 296 | int num_patterns = 0; |
| 297 | |
| 298 | /* Array of all regexps. */ |
| 299 | struct pattern *patterns = NULL; |
| 300 | #endif /* ETAGS_REGEXPS */ |
| 301 | |
| 302 | /* Language stuff. */ |
| 303 | struct lang_entry |
| 304 | { |
| 305 | char *extension; |
| 306 | Lang_function *function; |
| 307 | }; |
| 308 | |
| 309 | /* Table of language names and corresponding functions. */ |
| 310 | /* It is ok for a given function to be listed under more than one |
| 311 | name. I just didn't. */ |
| 312 | /* "auto" language reverts to default behavior. */ |
| 313 | struct lang_entry lang_names[] = |
| 314 | { |
| 315 | { "asm", Asm_labels }, |
| 316 | { "c", default_C_entries }, |
| 317 | { "c++", Cplusplus_entries }, |
| 318 | { "c*", Cstar_entries }, |
| 319 | { "fortran", Fortran_functions }, |
| 320 | { "lisp", Lisp_functions }, |
| 321 | { "none", just_read_file }, |
| 322 | { "pascal", Pascal_functions }, |
| 323 | { "scheme" , Scheme_functions }, |
| 324 | { "tex", TeX_functions }, |
| 325 | { "auto", NULL }, |
| 326 | { NULL, NULL } |
| 327 | }; |
| 328 | |
| 329 | /* Table of file extensions and corresponding language functions. */ |
| 330 | struct lang_entry lang_extensions[] = |
| 331 | { |
| 332 | /* Assume that ".s" or ".a" is assembly code. -wolfgang. |
| 333 | Or even ".sa". */ |
| 334 | { "a", Asm_labels }, /* Unix assembler */ |
| 335 | { "asm", Asm_labels }, /* Microcontroller assembly */ |
| 336 | { "def", Asm_labels }, /* BSO/Tasking definition includes */ |
| 337 | { "inc", Asm_labels }, /* Microcontroller include files */ |
| 338 | { "ins", Asm_labels }, /* Microcontroller include files */ |
| 339 | { "s", Asm_labels }, |
| 340 | { "sa", Asm_labels }, /* Unix assembler */ |
| 341 | { "src", Asm_labels }, /* BSO/Tasking C compiler output */ |
| 342 | |
| 343 | /* .aux, .bbl, .clo, .cls, .dtx or .tex implies LaTeX source code. */ |
| 344 | { "aux", TeX_functions }, |
| 345 | { "bbl", TeX_functions }, |
| 346 | { "clo", TeX_functions }, |
| 347 | { "cls", TeX_functions }, |
| 348 | { "dtx", TeX_functions }, |
| 349 | { "sty", TeX_functions }, |
| 350 | { "tex", TeX_functions }, |
| 351 | |
| 352 | /* .l or .el or .lisp (or .cl or .clisp or ...) implies lisp source code */ |
| 353 | { "cl", Lisp_functions }, |
| 354 | { "clisp", Lisp_functions }, |
| 355 | { "el", Lisp_functions }, |
| 356 | { "l", Lisp_functions }, |
| 357 | { "lisp", Lisp_functions }, |
| 358 | { "lsp", Lisp_functions }, |
| 359 | |
| 360 | /* .scm or .sm or .scheme implies scheme source code */ |
| 361 | { "SCM", Scheme_functions }, |
| 362 | { "SM", Scheme_functions }, |
| 363 | { "oak", Scheme_functions }, |
| 364 | { "sch", Scheme_functions }, |
| 365 | { "scheme", Scheme_functions }, |
| 366 | { "scm", Scheme_functions }, |
| 367 | { "sm", Scheme_functions }, |
| 368 | { "t", Scheme_functions }, |
| 369 | /* FIXME Can't do the `SCM' or `scm' prefix with a version number */ |
| 370 | |
| 371 | /* Note that ".c" and ".h" can be considered C++, if the --c++ |
| 372 | flag was given. That is why default_C_entries is called here. */ |
| 373 | { "c", default_C_entries }, |
| 374 | { "h", default_C_entries }, |
| 375 | |
| 376 | /* .C or .H or .c++ or .cc or .cpp or .cxx or .h++ or .hh or .hxx: |
| 377 | a C++ file */ |
| 378 | { "C", Cplusplus_entries }, |
| 379 | { "H", Cplusplus_entries }, |
| 380 | { "c++", Cplusplus_entries }, |
| 381 | { "cc", Cplusplus_entries }, |
| 382 | { "cpp", Cplusplus_entries }, |
| 383 | { "cxx", Cplusplus_entries }, |
| 384 | { "h++", Cplusplus_entries }, |
| 385 | { "hh", Cplusplus_entries }, |
| 386 | { "hxx", Cplusplus_entries }, |
| 387 | |
| 388 | /* .y: a yacc file */ |
| 389 | { "y", Yacc_entries }, |
| 390 | |
| 391 | /* .cs or .hs: a C* file */ |
| 392 | { "cs", Cstar_entries }, |
| 393 | { "hs", Cstar_entries }, |
| 394 | |
| 395 | /* .f and .for are FORTRAN. */ |
| 396 | { "F", Fortran_functions }, |
| 397 | { "f", Fortran_functions }, |
| 398 | { "for", Fortran_functions }, |
| 399 | |
| 400 | /* .pl implies prolog source code */ |
| 401 | { "pl", Prolog_functions }, |
| 402 | |
| 403 | /* .p or .pas: a Pascal file */ |
| 404 | { "p", Pascal_functions }, |
| 405 | { "pas", Pascal_functions }, |
| 406 | |
| 407 | { NULL, NULL } |
| 408 | }; |
| 409 | |
| 410 | /* Non-NULL if language fixed. */ |
| 411 | Lang_function *lang_func = NULL; |
| 412 | |
| 413 | \f |
| 414 | void |
| 415 | print_language_names () |
| 416 | { |
| 417 | struct lang_entry *name, *ext; |
| 418 | |
| 419 | puts ("\nThese are the currently supported languages, along with the\n\ |
| 420 | default extensions for files:"); |
| 421 | for (name = lang_names; name->extension; ++name) |
| 422 | { |
| 423 | printf ("\t%s\t", name->extension); |
| 424 | for (ext = lang_extensions; ext->extension; ++ext) |
| 425 | if (name->function == ext->function) |
| 426 | printf (" .%s", ext->extension); |
| 427 | puts (""); |
| 428 | } |
| 429 | puts ("Where `auto' means use default language for files based on filename\n\ |
| 430 | extension, and `none' means only do regexp processing on files.\n\ |
| 431 | If no language is specified and no extension is found for some file,\n\ |
| 432 | Fortran is tried first; if no tags are found, C is tried next."); |
| 433 | } |
| 434 | |
| 435 | void |
| 436 | print_version () |
| 437 | { |
| 438 | #ifdef VERSION |
| 439 | printf ("%s for Emacs version %s.\n", (CTAGS) ? "CTAGS" : "ETAGS", VERSION); |
| 440 | #else |
| 441 | printf ("%s for Emacs version 19.\n", (CTAGS) ? "CTAGS" : "ETAGS"); |
| 442 | #endif |
| 443 | |
| 444 | exit (GOOD); |
| 445 | } |
| 446 | |
| 447 | void |
| 448 | print_help () |
| 449 | { |
| 450 | printf ("These are the options accepted by %s. You may use unambiguous\n\ |
| 451 | abbreviations for the long option names. A - as file name means read\n\ |
| 452 | names from stdin.\n\n", progname); |
| 453 | |
| 454 | puts ("-a, --append\n\ |
| 455 | Append tag entries to existing tags file."); |
| 456 | |
| 457 | if (CTAGS) |
| 458 | puts ("-B, --backward-search\n\ |
| 459 | Write the search commands for the tag entries using '?', the\n\ |
| 460 | backward-search command instead of '/', the forward-search command."); |
| 461 | |
| 462 | puts ("-C, --c++\n\ |
| 463 | Treat files whose extension defaults to C language as C++ files."); |
| 464 | |
| 465 | if (CTAGS) |
| 466 | puts ("-d, --defines\n\ |
| 467 | Create tag entries for constant C #defines, too."); |
| 468 | else |
| 469 | puts ("-D, --no-defines\n\ |
| 470 | Don't create tag entries for constant C #defines. This makes\n\ |
| 471 | the tags file smaller."); |
| 472 | |
| 473 | if (!CTAGS) |
| 474 | { |
| 475 | puts ("-i FILE, --include=FILE\n\ |
| 476 | Include a note in tag file indicating that, when searching for\n\ |
| 477 | a tag, one should also consult the tags file FILE after\n\ |
| 478 | checking the current file."); |
| 479 | puts ("-l LANG, --language=LANG\n\ |
| 480 | Force the following files to be considered as written in the\n\ |
| 481 | named language up to the next --language=LANG option."); |
| 482 | } |
| 483 | |
| 484 | #ifdef ETAGS_REGEXPS |
| 485 | puts ("-r /REGEXP/, --regex=/REGEXP/\n\ |
| 486 | Make a tag for each line matching pattern REGEXP in the\n\ |
| 487 | following files. REGEXP is anchored (as if preceded by ^).\n\ |
| 488 | The form /REGEXP/NAME/ creates a named tag. For example Tcl\n\ |
| 489 | named tags can be created with:\n\ |
| 490 | --regex=/proc[ \\t]+\\([^ \\t]+\\)/\\1/."); |
| 491 | puts ("-R, --no-regex\n\ |
| 492 | Don't create tags from regexps for the following files."); |
| 493 | #endif /* ETAGS_REGEXPS */ |
| 494 | puts ("-o FILE, --output=FILE\n\ |
| 495 | Write the tags to FILE."); |
| 496 | puts ("-I, --ignore-indentation\n\ |
| 497 | Don't rely on indentation quite as much as normal. Currently,\n\ |
| 498 | this means not to assume that a closing brace in the first\n\ |
| 499 | column is the final brace of a function or structure\n\ |
| 500 | definition in C and C++."); |
| 501 | |
| 502 | if (CTAGS) |
| 503 | { |
| 504 | puts ("-t, --typedefs\n\ |
| 505 | Generate tag entries for C typedefs."); |
| 506 | puts ("-T, --typedefs-and-c++\n\ |
| 507 | Generate tag entries for C typedefs, C struct/enum/union tags,\n\ |
| 508 | and C++ member functions."); |
| 509 | puts ("-u, --update\n\ |
| 510 | Update the tag entries for the given files, leaving tag\n\ |
| 511 | entries for other files in place. Currently, this is\n\ |
| 512 | implemented by deleting the existing entries for the given\n\ |
| 513 | files and then rewriting the new entries at the end of the\n\ |
| 514 | tags file. It is often faster to simply rebuild the entire\n\ |
| 515 | tag file than to use this."); |
| 516 | puts ("-v, --vgrind\n\ |
| 517 | Generates an index of items intended for human consumption,\n\ |
| 518 | similar to the output of vgrind. The index is sorted, and\n\ |
| 519 | gives the page number of each item."); |
| 520 | puts ("-w, --no-warn\n\ |
| 521 | Suppress warning messages about entries defined in multiple\n\ |
| 522 | files."); |
| 523 | puts ("-x, --cxref\n\ |
| 524 | Like --vgrind, but in the style of cxref, rather than vgrind.\n\ |
| 525 | The output uses line numbers instead of page numbers, but\n\ |
| 526 | beyond that the differences are cosmetic; try both to see\n\ |
| 527 | which you like."); |
| 528 | } |
| 529 | |
| 530 | puts ("-V, --version\n\ |
| 531 | Print the version of the program.\n\ |
| 532 | -h, --help\n\ |
| 533 | Print this help message."); |
| 534 | |
| 535 | print_language_names (); |
| 536 | |
| 537 | exit (GOOD); |
| 538 | } |
| 539 | |
| 540 | \f |
| 541 | enum argument_type |
| 542 | { |
| 543 | at_language, |
| 544 | at_regexp, |
| 545 | at_filename |
| 546 | }; |
| 547 | |
| 548 | /* This structure helps us allow mixing of --lang and filenames. */ |
| 549 | typedef struct |
| 550 | { |
| 551 | enum argument_type arg_type; |
| 552 | char *what; |
| 553 | Lang_function *function; |
| 554 | } ARGUMENT; |
| 555 | |
| 556 | #ifdef VMS /* VMS specific functions */ |
| 557 | |
| 558 | #define EOS '\0' |
| 559 | |
| 560 | /* This is a BUG! ANY arbitrary limit is a BUG! |
| 561 | Won't someone please fix this? */ |
| 562 | #define MAX_FILE_SPEC_LEN 255 |
| 563 | typedef struct { |
| 564 | short curlen; |
| 565 | char body[MAX_FILE_SPEC_LEN + 1]; |
| 566 | } vspec; |
| 567 | |
| 568 | /* |
| 569 | v1.05 nmm 26-Jun-86 fn_exp - expand specification of list of file names |
| 570 | returning in each successive call the next filename matching the input |
| 571 | spec. The function expects that each in_spec passed |
| 572 | to it will be processed to completion; in particular, up to and |
| 573 | including the call following that in which the last matching name |
| 574 | is returned, the function ignores the value of in_spec, and will |
| 575 | only start processing a new spec with the following call. |
| 576 | If an error occurs, on return out_spec contains the value |
| 577 | of in_spec when the error occurred. |
| 578 | |
| 579 | With each successive filename returned in out_spec, the |
| 580 | function's return value is one. When there are no more matching |
| 581 | names the function returns zero. If on the first call no file |
| 582 | matches in_spec, or there is any other error, -1 is returned. |
| 583 | */ |
| 584 | |
| 585 | #include <rmsdef.h> |
| 586 | #include <descrip.h> |
| 587 | #define OUTSIZE MAX_FILE_SPEC_LEN |
| 588 | short |
| 589 | fn_exp (out, in) |
| 590 | vspec *out; |
| 591 | char *in; |
| 592 | { |
| 593 | static long context = 0; |
| 594 | static struct dsc$descriptor_s o; |
| 595 | static struct dsc$descriptor_s i; |
| 596 | static logical pass1 = TRUE; |
| 597 | long status; |
| 598 | short retval; |
| 599 | |
| 600 | if (pass1) |
| 601 | { |
| 602 | pass1 = FALSE; |
| 603 | o.dsc$a_pointer = (char *) out; |
| 604 | o.dsc$w_length = (short)OUTSIZE; |
| 605 | i.dsc$a_pointer = in; |
| 606 | i.dsc$w_length = (short)strlen(in); |
| 607 | i.dsc$b_dtype = DSC$K_DTYPE_T; |
| 608 | i.dsc$b_class = DSC$K_CLASS_S; |
| 609 | o.dsc$b_dtype = DSC$K_DTYPE_VT; |
| 610 | o.dsc$b_class = DSC$K_CLASS_VS; |
| 611 | } |
| 612 | if ((status = lib$find_file(&i, &o, &context, 0, 0)) == RMS$_NORMAL) |
| 613 | { |
| 614 | out->body[out->curlen] = EOS; |
| 615 | return 1; |
| 616 | } |
| 617 | else if (status == RMS$_NMF) |
| 618 | retval = 0; |
| 619 | else |
| 620 | { |
| 621 | strcpy(out->body, in); |
| 622 | retval = -1; |
| 623 | } |
| 624 | lib$find_file_end(&context); |
| 625 | pass1 = TRUE; |
| 626 | return retval; |
| 627 | } |
| 628 | |
| 629 | /* |
| 630 | v1.01 nmm 19-Aug-85 gfnames - return in successive calls the |
| 631 | name of each file specified by the provided arg expanding wildcards. |
| 632 | */ |
| 633 | char * |
| 634 | gfnames (arg, p_error) |
| 635 | char *arg; |
| 636 | logical *p_error; |
| 637 | { |
| 638 | static vspec filename = {MAX_FILE_SPEC_LEN, "\0"}; |
| 639 | |
| 640 | switch (fn_exp (&filename, arg)) |
| 641 | { |
| 642 | case 1: |
| 643 | *p_error = FALSE; |
| 644 | return filename.body; |
| 645 | case 0: |
| 646 | *p_error = FALSE; |
| 647 | return NULL; |
| 648 | default: |
| 649 | *p_error = TRUE; |
| 650 | return filename.body; |
| 651 | } |
| 652 | } |
| 653 | |
| 654 | #ifndef OLD /* Newer versions of VMS do provide `system'. */ |
| 655 | system (cmd) |
| 656 | char *cmd; |
| 657 | { |
| 658 | fprintf (stderr, "system() function not implemented under VMS\n"); |
| 659 | } |
| 660 | #endif |
| 661 | |
| 662 | #define VERSION_DELIM ';' |
| 663 | char *massage_name (s) |
| 664 | char *s; |
| 665 | { |
| 666 | char *start = s; |
| 667 | |
| 668 | for ( ; *s; s++) |
| 669 | if (*s == VERSION_DELIM) |
| 670 | { |
| 671 | *s = EOS; |
| 672 | break; |
| 673 | } |
| 674 | else |
| 675 | *s = tolower(*s); |
| 676 | return start; |
| 677 | } |
| 678 | #endif /* VMS */ |
| 679 | |
| 680 | \f |
| 681 | void |
| 682 | main (argc, argv) |
| 683 | int argc; |
| 684 | char *argv[]; |
| 685 | { |
| 686 | int i; |
| 687 | unsigned int nincluded_files = 0; |
| 688 | char **included_files = xnew (argc, char *); |
| 689 | char *this_file; |
| 690 | ARGUMENT *argbuffer; |
| 691 | int current_arg = 0, file_count = 0; |
| 692 | struct linebuffer filename_lb; |
| 693 | #ifdef VMS |
| 694 | logical got_err; |
| 695 | #endif |
| 696 | |
| 697 | #ifdef DOS_NT |
| 698 | _fmode = O_BINARY; /* all of files are treated as binary files */ |
| 699 | #endif /* DOS_NT */ |
| 700 | |
| 701 | progname = argv[0]; |
| 702 | |
| 703 | /* Allocate enough no matter what happens. Overkill, but each one |
| 704 | is small. */ |
| 705 | argbuffer = xnew (argc, ARGUMENT); |
| 706 | |
| 707 | #ifdef ETAGS_REGEXPS |
| 708 | /* Set syntax for regular expression routines. */ |
| 709 | re_set_syntax (RE_SYNTAX_EMACS); |
| 710 | #endif /* ETAGS_REGEXPS */ |
| 711 | |
| 712 | /* |
| 713 | * If etags, always find typedefs and structure tags. Why not? |
| 714 | * Also default is to find macro constants. |
| 715 | */ |
| 716 | if (!CTAGS) |
| 717 | typedefs = typedefs_and_cplusplus = constantypedefs = TRUE; |
| 718 | |
| 719 | while (1) |
| 720 | { |
| 721 | int opt = getopt_long (argc, argv, |
| 722 | "-aCdDf:Il:o:r:RStTi:BuvxwVhH", longopts, 0); |
| 723 | |
| 724 | if (opt == EOF) |
| 725 | break; |
| 726 | |
| 727 | switch (opt) |
| 728 | { |
| 729 | case 0: |
| 730 | /* If getopt returns 0, then it has already processed a |
| 731 | long-named option. We should do nothing. */ |
| 732 | break; |
| 733 | |
| 734 | case 1: |
| 735 | /* This means that a filename has been seen. Record it. */ |
| 736 | argbuffer[current_arg].arg_type = at_filename; |
| 737 | argbuffer[current_arg].what = optarg; |
| 738 | ++current_arg; |
| 739 | ++file_count; |
| 740 | break; |
| 741 | |
| 742 | /* Common options. */ |
| 743 | case 'a': |
| 744 | append_to_tagfile = TRUE; |
| 745 | break; |
| 746 | case 'C': |
| 747 | cplusplus = TRUE; |
| 748 | break; |
| 749 | case 'd': |
| 750 | constantypedefs = TRUE; |
| 751 | break; |
| 752 | case 'D': |
| 753 | constantypedefs = FALSE; |
| 754 | break; |
| 755 | case 'f': /* for compatibility with old makefiles */ |
| 756 | case 'o': |
| 757 | if (tagfile) |
| 758 | { |
| 759 | fprintf (stderr, "%s: -%c option may only be given once.\n", |
| 760 | progname, opt); |
| 761 | goto usage; |
| 762 | } |
| 763 | tagfile = optarg; |
| 764 | break; |
| 765 | case 'I': |
| 766 | case 'S': /* for backward compatibility */ |
| 767 | noindentypedefs = TRUE; |
| 768 | break; |
| 769 | case 'l': |
| 770 | if (!get_language (optarg, &argbuffer[current_arg].function)) |
| 771 | { |
| 772 | fprintf (stderr, "%s: language \"%s\" not recognized.\n", |
| 773 | progname, optarg); |
| 774 | goto usage; |
| 775 | } |
| 776 | argbuffer[current_arg].arg_type = at_language; |
| 777 | ++current_arg; |
| 778 | break; |
| 779 | #ifdef ETAGS_REGEXPS |
| 780 | case 'r': |
| 781 | argbuffer[current_arg].arg_type = at_regexp; |
| 782 | argbuffer[current_arg].what = optarg; |
| 783 | ++current_arg; |
| 784 | break; |
| 785 | case 'R': |
| 786 | argbuffer[current_arg].arg_type = at_regexp; |
| 787 | argbuffer[current_arg].what = NULL; |
| 788 | ++current_arg; |
| 789 | break; |
| 790 | #endif /* ETAGS_REGEXPS */ |
| 791 | case 'V': |
| 792 | print_version (); |
| 793 | break; |
| 794 | case 'h': |
| 795 | case 'H': |
| 796 | print_help (); |
| 797 | break; |
| 798 | case 't': |
| 799 | typedefs = TRUE; |
| 800 | break; |
| 801 | case 'T': |
| 802 | typedefs = typedefs_and_cplusplus = TRUE; |
| 803 | break; |
| 804 | #if (!CTAGS) |
| 805 | /* Etags options */ |
| 806 | case 'i': |
| 807 | included_files[nincluded_files++] = optarg; |
| 808 | break; |
| 809 | #else /* CTAGS */ |
| 810 | /* Ctags options. */ |
| 811 | case 'B': |
| 812 | searchar = '?'; |
| 813 | break; |
| 814 | case 'u': |
| 815 | update = TRUE; |
| 816 | break; |
| 817 | case 'v': |
| 818 | vgrind_style = TRUE; |
| 819 | /*FALLTHRU*/ |
| 820 | case 'x': |
| 821 | cxref_style = TRUE; |
| 822 | break; |
| 823 | case 'w': |
| 824 | no_warnings = TRUE; |
| 825 | break; |
| 826 | #endif /* CTAGS */ |
| 827 | default: |
| 828 | goto usage; |
| 829 | } |
| 830 | } |
| 831 | |
| 832 | for (; optind < argc; ++optind) |
| 833 | { |
| 834 | argbuffer[current_arg].arg_type = at_filename; |
| 835 | argbuffer[current_arg].what = argv[optind]; |
| 836 | ++current_arg; |
| 837 | ++file_count; |
| 838 | } |
| 839 | |
| 840 | if (nincluded_files == 0 && file_count == 0) |
| 841 | { |
| 842 | fprintf (stderr, "%s: No input files specified.\n", progname); |
| 843 | |
| 844 | usage: |
| 845 | fprintf (stderr, "\tTry `%s --help' for a complete list of options.\n", |
| 846 | progname); |
| 847 | exit (BAD); |
| 848 | } |
| 849 | |
| 850 | if (tagfile == NULL) |
| 851 | { |
| 852 | tagfile = CTAGS ? "tags" : "TAGS"; |
| 853 | } |
| 854 | cwd = etags_getcwd (); /* the current working directory */ |
| 855 | strcat (cwd, "/"); |
| 856 | if (streq (tagfile, "-")) |
| 857 | { |
| 858 | tagfiledir = cwd; |
| 859 | } |
| 860 | else |
| 861 | { |
| 862 | tagfiledir = absolute_dirname (tagfile, cwd); |
| 863 | } |
| 864 | |
| 865 | init (); /* set up boolean "functions" */ |
| 866 | |
| 867 | initbuffer (&lb); |
| 868 | initbuffer (&token_name); |
| 869 | initbuffer (&lbs[0].lb); |
| 870 | initbuffer (&lbs[1].lb); |
| 871 | initbuffer (&filename_lb); |
| 872 | |
| 873 | if (!CTAGS) |
| 874 | { |
| 875 | if (streq (tagfile, "-")) |
| 876 | tagf = stdout; |
| 877 | else |
| 878 | tagf = fopen (tagfile, append_to_tagfile ? "a" : "w"); |
| 879 | if (tagf == NULL) |
| 880 | pfatal (tagfile); |
| 881 | } |
| 882 | |
| 883 | /* |
| 884 | * Loop through files finding functions. |
| 885 | */ |
| 886 | for (i = 0; i < current_arg; ++i) |
| 887 | { |
| 888 | switch (argbuffer[i].arg_type) |
| 889 | { |
| 890 | case at_language: |
| 891 | lang_func = argbuffer[i].function; |
| 892 | break; |
| 893 | #ifdef ETAGS_REGEXPS |
| 894 | case at_regexp: |
| 895 | add_regex (argbuffer[i].what); |
| 896 | break; |
| 897 | #endif |
| 898 | case at_filename: |
| 899 | #ifdef VMS |
| 900 | while ((this_file = gfnames (argbuffer[i].what, &got_err)) != NULL) |
| 901 | { |
| 902 | if (got_err) |
| 903 | { |
| 904 | error ("Can't find file %s\n", this_file); |
| 905 | argc--, argv++; |
| 906 | } |
| 907 | else |
| 908 | { |
| 909 | this_file = massage_name (this_file); |
| 910 | } |
| 911 | #else |
| 912 | this_file = argbuffer[i].what; |
| 913 | #endif |
| 914 | /* Input file named "-" means read file names from stdin |
| 915 | and use them. */ |
| 916 | if (streq (this_file, "-")) |
| 917 | while (readline_internal (&filename_lb, stdin) > 0) |
| 918 | process_file (filename_lb.buffer); |
| 919 | else |
| 920 | process_file (this_file); |
| 921 | #ifdef VMS |
| 922 | } |
| 923 | #endif |
| 924 | break; |
| 925 | } |
| 926 | } |
| 927 | |
| 928 | if (!CTAGS) |
| 929 | { |
| 930 | while (nincluded_files-- > 0) |
| 931 | fprintf (tagf, "\f\n%s,include\n", *included_files++); |
| 932 | |
| 933 | fclose (tagf); |
| 934 | exit (GOOD); |
| 935 | } |
| 936 | |
| 937 | /* If CTAGS, we are here. process_file did not write the tags yet, |
| 938 | because we want them ordered. Let's do it now. */ |
| 939 | if (cxref_style) |
| 940 | { |
| 941 | tagf = fopen (tagfile, append_to_tagfile ? "a" : "w"); |
| 942 | if (tagf == NULL) |
| 943 | pfatal (tagfile); |
| 944 | put_entries (head); |
| 945 | exit (GOOD); |
| 946 | } |
| 947 | |
| 948 | if (update) |
| 949 | { |
| 950 | char cmd[BUFSIZ]; |
| 951 | for (i = 0; i < current_arg; ++i) |
| 952 | { |
| 953 | if (argbuffer[i].arg_type != at_filename) |
| 954 | continue; |
| 955 | sprintf (cmd, |
| 956 | "mv %s OTAGS;fgrep -v '\t%s\t' OTAGS >%s;rm OTAGS", |
| 957 | tagfile, argbuffer[i].what, tagfile); |
| 958 | if (system (cmd) != GOOD) |
| 959 | fatal ("failed to execute shell command"); |
| 960 | } |
| 961 | append_to_tagfile = TRUE; |
| 962 | } |
| 963 | |
| 964 | tagf = fopen (tagfile, append_to_tagfile ? "a" : "w"); |
| 965 | if (tagf == NULL) |
| 966 | pfatal (tagfile); |
| 967 | put_entries (head); |
| 968 | fclose (tagf); |
| 969 | |
| 970 | if (update) |
| 971 | { |
| 972 | char cmd[BUFSIZ]; |
| 973 | sprintf (cmd, "sort %s -o %s", tagfile, tagfile); |
| 974 | exit (system (cmd)); |
| 975 | } |
| 976 | exit (GOOD); |
| 977 | } |
| 978 | |
| 979 | |
| 980 | /* |
| 981 | * Set the language, given the name. |
| 982 | */ |
| 983 | logical |
| 984 | get_language (language, func) |
| 985 | char *language; |
| 986 | Lang_function **func; |
| 987 | { |
| 988 | struct lang_entry *lang; |
| 989 | |
| 990 | for (lang = lang_names; lang->extension; ++lang) |
| 991 | { |
| 992 | if (streq (language, lang->extension)) |
| 993 | { |
| 994 | *func = lang->function; |
| 995 | return TRUE; |
| 996 | } |
| 997 | } |
| 998 | |
| 999 | return FALSE; |
| 1000 | } |
| 1001 | |
| 1002 | |
| 1003 | /* |
| 1004 | * This routine is called on each file argument. |
| 1005 | */ |
| 1006 | void |
| 1007 | process_file (file) |
| 1008 | char *file; |
| 1009 | { |
| 1010 | struct stat stat_buf; |
| 1011 | FILE *inf; |
| 1012 | |
| 1013 | if (stat (file, &stat_buf) == 0 && !S_ISREG (stat_buf.st_mode)) |
| 1014 | { |
| 1015 | fprintf (stderr, "Skipping %s: it is not a regular file.\n", file); |
| 1016 | return; |
| 1017 | } |
| 1018 | if (streq (file, tagfile) && !streq (tagfile, "-")) |
| 1019 | { |
| 1020 | fprintf (stderr, "Skipping inclusion of %s in self.\n", file); |
| 1021 | return; |
| 1022 | } |
| 1023 | inf = fopen (file, "r"); |
| 1024 | if (inf == NULL) |
| 1025 | { |
| 1026 | perror (file); |
| 1027 | return; |
| 1028 | } |
| 1029 | |
| 1030 | find_entries (file, inf); |
| 1031 | |
| 1032 | if (!CTAGS) |
| 1033 | { |
| 1034 | char *filename; |
| 1035 | |
| 1036 | if (file[0] == '/') |
| 1037 | { |
| 1038 | /* file is an absolute filename. Canonicalise it. */ |
| 1039 | filename = absolute_filename (file, cwd); |
| 1040 | } |
| 1041 | else |
| 1042 | { |
| 1043 | /* file is a filename relative to cwd. Make it relative |
| 1044 | to the directory of the tags file. */ |
| 1045 | filename = relative_filename (file, tagfiledir); |
| 1046 | } |
| 1047 | fprintf (tagf, "\f\n%s,%d\n", filename, total_size_of_entries (head)); |
| 1048 | free (filename); |
| 1049 | put_entries (head); |
| 1050 | free_tree (head); |
| 1051 | head = NULL; |
| 1052 | } |
| 1053 | } |
| 1054 | |
| 1055 | /* |
| 1056 | * This routine sets up the boolean pseudo-functions which work |
| 1057 | * by setting boolean flags dependent upon the corresponding character |
| 1058 | * Every char which is NOT in that string is not a white char. Therefore, |
| 1059 | * all of the array "_wht" is set to FALSE, and then the elements |
| 1060 | * subscripted by the chars in "white" are set to TRUE. Thus "_wht" |
| 1061 | * of a char is TRUE if it is the string "white", else FALSE. |
| 1062 | */ |
| 1063 | void |
| 1064 | init () |
| 1065 | { |
| 1066 | register char *sp; |
| 1067 | register int i; |
| 1068 | |
| 1069 | for (i = 0; i < 0177; i++) |
| 1070 | _wht[i] = _etk[i] = _itk[i] = _btk[i] = FALSE; |
| 1071 | for (sp = white; *sp; sp++) |
| 1072 | _wht[*sp] = TRUE; |
| 1073 | for (sp = endtk; *sp; sp++) |
| 1074 | _etk[*sp] = TRUE; |
| 1075 | for (sp = intk; *sp; sp++) |
| 1076 | _itk[*sp] = TRUE; |
| 1077 | for (sp = begtk; *sp; sp++) |
| 1078 | _btk[*sp] = TRUE; |
| 1079 | _wht[0] = _wht['\n']; |
| 1080 | _etk[0] = _etk['\n']; |
| 1081 | _btk[0] = _btk['\n']; |
| 1082 | _itk[0] = _itk['\n']; |
| 1083 | } |
| 1084 | |
| 1085 | /* |
| 1086 | * This routine opens the specified file and calls the function |
| 1087 | * which finds the function and type definitions. |
| 1088 | */ |
| 1089 | void |
| 1090 | find_entries (file, inf) |
| 1091 | char *file; |
| 1092 | FILE *inf; |
| 1093 | { |
| 1094 | char *cp; |
| 1095 | struct lang_entry *lang; |
| 1096 | NODE *old_last_node; |
| 1097 | extern NODE *last_node; |
| 1098 | |
| 1099 | curfile = savestr (file); |
| 1100 | cp = etags_strrchr (file, '.'); |
| 1101 | |
| 1102 | /* If user specified a language, use it. */ |
| 1103 | if (lang_func != NULL) |
| 1104 | { |
| 1105 | lang_func (inf); |
| 1106 | fclose (inf); |
| 1107 | return; |
| 1108 | } |
| 1109 | |
| 1110 | if (cp) |
| 1111 | { |
| 1112 | ++cp; |
| 1113 | for (lang = lang_extensions; lang->extension; ++lang) |
| 1114 | { |
| 1115 | if (streq (cp, lang->extension)) |
| 1116 | { |
| 1117 | lang->function (inf); |
| 1118 | fclose (inf); |
| 1119 | return; |
| 1120 | } |
| 1121 | } |
| 1122 | } |
| 1123 | |
| 1124 | /* Try Fortran. */ |
| 1125 | old_last_node = last_node; |
| 1126 | Fortran_functions (inf); |
| 1127 | |
| 1128 | /* No Fortran entries found. Try C. */ |
| 1129 | if (old_last_node == last_node) |
| 1130 | default_C_entries (inf); |
| 1131 | fclose (inf); |
| 1132 | } |
| 1133 | \f |
| 1134 | /* Record a tag. */ |
| 1135 | void |
| 1136 | pfnote (name, is_func, named, linestart, linelen, lno, cno) |
| 1137 | char *name; /* tag name */ |
| 1138 | logical is_func; /* tag is a function */ |
| 1139 | logical named; /* tag different from text of definition */ |
| 1140 | char *linestart; /* start of the line where tag is */ |
| 1141 | int linelen; /* length of the line where tag is */ |
| 1142 | int lno; /* line number */ |
| 1143 | long cno; /* character number */ |
| 1144 | { |
| 1145 | register NODE *np = xnew (1, NODE); |
| 1146 | register char *fp; |
| 1147 | |
| 1148 | /* If ctags mode, change name "main" to M<thisfilename>. */ |
| 1149 | if (CTAGS && !cxref_style && streq (name, "main")) |
| 1150 | { |
| 1151 | fp = etags_strrchr (curfile, '/'); |
| 1152 | np->name = concat ("M", fp == 0 ? curfile : fp + 1, ""); |
| 1153 | fp = etags_strrchr (np->name, '.'); |
| 1154 | if (fp && fp[1] != '\0' && fp[2] == '\0') |
| 1155 | fp[0] = 0; |
| 1156 | np->named = TRUE; |
| 1157 | } |
| 1158 | else |
| 1159 | { |
| 1160 | np->name = name; |
| 1161 | np->named = named; |
| 1162 | } |
| 1163 | np->been_warned = FALSE; |
| 1164 | np->file = curfile; |
| 1165 | np->is_func = is_func; |
| 1166 | np->lno = lno; |
| 1167 | /* Our char numbers are 0-base, because of C language tradition? |
| 1168 | ctags compatibility? old versions compatibility? I don't know. |
| 1169 | Anyway, since emacs's are 1-base we espect etags.el to take care |
| 1170 | of the difference. If we wanted to have 1-based numbers, we would |
| 1171 | uncomment the +1 below. */ |
| 1172 | np->cno = cno /* + 1 */ ; |
| 1173 | np->left = np->right = NULL; |
| 1174 | np->pat = savenstr (linestart, ((CTAGS && !cxref_style) ? 50 : linelen)); |
| 1175 | |
| 1176 | add_node (np, &head); |
| 1177 | } |
| 1178 | |
| 1179 | /* |
| 1180 | * free_tree () |
| 1181 | * recurse on left children, iterate on right children. |
| 1182 | */ |
| 1183 | void |
| 1184 | free_tree (node) |
| 1185 | register NODE *node; |
| 1186 | { |
| 1187 | while (node) |
| 1188 | { |
| 1189 | register NODE *node_right = node->right; |
| 1190 | free_tree (node->left); |
| 1191 | if (node->named) |
| 1192 | free (node->name); |
| 1193 | free (node->pat); |
| 1194 | free ((char *) node); |
| 1195 | node = node_right; |
| 1196 | } |
| 1197 | } |
| 1198 | |
| 1199 | /* |
| 1200 | * add_node () |
| 1201 | * Adds a node to the tree of nodes. In etags mode, we don't keep |
| 1202 | * it sorted; we just keep a linear list. In ctags mode, maintain |
| 1203 | * an ordered tree, with no attempt at balancing. |
| 1204 | * |
| 1205 | * add_node is the only function allowed to add nodes, so it can |
| 1206 | * maintain state. |
| 1207 | */ |
| 1208 | NODE *last_node = NULL; |
| 1209 | void |
| 1210 | add_node (node, cur_node_p) |
| 1211 | NODE *node, **cur_node_p; |
| 1212 | { |
| 1213 | register int dif; |
| 1214 | register NODE *cur_node = *cur_node_p; |
| 1215 | |
| 1216 | if (cur_node == NULL) |
| 1217 | { |
| 1218 | *cur_node_p = node; |
| 1219 | last_node = node; |
| 1220 | return; |
| 1221 | } |
| 1222 | |
| 1223 | if (!CTAGS) |
| 1224 | { |
| 1225 | /* Etags Mode */ |
| 1226 | if (last_node == NULL) |
| 1227 | fatal ("internal error in add_node", 0); |
| 1228 | last_node->right = node; |
| 1229 | last_node = node; |
| 1230 | } |
| 1231 | else |
| 1232 | { |
| 1233 | /* Ctags Mode */ |
| 1234 | dif = strcmp (node->name, cur_node->name); |
| 1235 | |
| 1236 | /* |
| 1237 | * If this tag name matches an existing one, then |
| 1238 | * do not add the node, but maybe print a warning. |
| 1239 | */ |
| 1240 | if (!dif) |
| 1241 | { |
| 1242 | if (node->file == cur_node->file) |
| 1243 | { |
| 1244 | if (!no_warnings) |
| 1245 | { |
| 1246 | fprintf (stderr, "Duplicate entry in file %s, line %d: %s\n", |
| 1247 | node->file, lineno, node->name); |
| 1248 | fprintf (stderr, "Second entry ignored\n"); |
| 1249 | } |
| 1250 | return; |
| 1251 | } |
| 1252 | if (!cur_node->been_warned && !no_warnings) |
| 1253 | { |
| 1254 | fprintf (stderr, |
| 1255 | "Duplicate entry in files %s and %s: %s (Warning only)\n", |
| 1256 | node->file, cur_node->file, node->name); |
| 1257 | } |
| 1258 | cur_node->been_warned = TRUE; |
| 1259 | return; |
| 1260 | } |
| 1261 | |
| 1262 | /* Maybe refuse to add duplicate nodes. */ |
| 1263 | if (!permit_duplicates) |
| 1264 | { |
| 1265 | if (streq (node->name, cur_node->name) |
| 1266 | && streq (node->file, cur_node->file)) |
| 1267 | return; |
| 1268 | } |
| 1269 | |
| 1270 | /* Actually add the node */ |
| 1271 | add_node (node, dif < 0 ? &cur_node->left : &cur_node->right); |
| 1272 | } |
| 1273 | } |
| 1274 | \f |
| 1275 | void |
| 1276 | put_entries (node) |
| 1277 | register NODE *node; |
| 1278 | { |
| 1279 | register char *sp; |
| 1280 | |
| 1281 | if (node == NULL) |
| 1282 | return; |
| 1283 | |
| 1284 | /* Output subentries that precede this one */ |
| 1285 | put_entries (node->left); |
| 1286 | |
| 1287 | /* Output this entry */ |
| 1288 | |
| 1289 | if (!CTAGS) |
| 1290 | { |
| 1291 | if (node->named) |
| 1292 | { |
| 1293 | fprintf (tagf, "%s\177%s\001%d,%d\n", |
| 1294 | node->pat, node->name, |
| 1295 | node->lno, node->cno); |
| 1296 | } |
| 1297 | else |
| 1298 | { |
| 1299 | fprintf (tagf, "%s\177%d,%d\n", |
| 1300 | node->pat, |
| 1301 | node->lno, node->cno); |
| 1302 | } |
| 1303 | } |
| 1304 | else if (!cxref_style) |
| 1305 | { |
| 1306 | fprintf (tagf, "%s\t%s\t", |
| 1307 | node->name, node->file); |
| 1308 | |
| 1309 | if (node->is_func) |
| 1310 | { /* a function */ |
| 1311 | putc (searchar, tagf); |
| 1312 | putc ('^', tagf); |
| 1313 | |
| 1314 | for (sp = node->pat; *sp; sp++) |
| 1315 | { |
| 1316 | if (*sp == '\\' || *sp == searchar) |
| 1317 | putc ('\\', tagf); |
| 1318 | putc (*sp, tagf); |
| 1319 | } |
| 1320 | putc (searchar, tagf); |
| 1321 | } |
| 1322 | else |
| 1323 | { /* a typedef; text pattern inadequate */ |
| 1324 | fprintf (tagf, "%d", node->lno); |
| 1325 | } |
| 1326 | putc ('\n', tagf); |
| 1327 | } |
| 1328 | else if (vgrind_style) |
| 1329 | fprintf (stdout, "%s %s %d\n", |
| 1330 | node->name, node->file, (node->lno + 63) / 64); |
| 1331 | else |
| 1332 | fprintf (stdout, "%-16s %3d %-16s %s\n", |
| 1333 | node->name, node->lno, node->file, node->pat); |
| 1334 | |
| 1335 | /* Output subentries that follow this one */ |
| 1336 | put_entries (node->right); |
| 1337 | } |
| 1338 | |
| 1339 | /* Length of a number's decimal representation. */ |
| 1340 | int |
| 1341 | number_len (num) |
| 1342 | long num; |
| 1343 | { |
| 1344 | int len = 0; |
| 1345 | if (!num) |
| 1346 | return 1; |
| 1347 | for (; num; num /= 10) |
| 1348 | ++len; |
| 1349 | return len; |
| 1350 | } |
| 1351 | |
| 1352 | /* |
| 1353 | * Return total number of characters that put_entries will output for |
| 1354 | * the nodes in the subtree of the specified node. Works only if |
| 1355 | * we are not ctags, but called only in that case. This count |
| 1356 | * is irrelevant with the new tags.el, but is still supplied for |
| 1357 | * backward compatibility. |
| 1358 | */ |
| 1359 | int |
| 1360 | total_size_of_entries (node) |
| 1361 | register NODE *node; |
| 1362 | { |
| 1363 | register int total; |
| 1364 | |
| 1365 | if (node == NULL) |
| 1366 | return 0; |
| 1367 | |
| 1368 | total = 0; |
| 1369 | for (; node; node = node->right) |
| 1370 | { |
| 1371 | /* Count left subentries. */ |
| 1372 | total += total_size_of_entries (node->left); |
| 1373 | |
| 1374 | /* Count this entry */ |
| 1375 | total += strlen (node->pat) + 1; |
| 1376 | total += number_len ((long) node->lno) + 1 + number_len (node->cno) + 1; |
| 1377 | if (node->named) |
| 1378 | total += 1 + strlen (node->name); /* \001name */ |
| 1379 | } |
| 1380 | |
| 1381 | return total; |
| 1382 | } |
| 1383 | \f |
| 1384 | /* |
| 1385 | * The C symbol tables. |
| 1386 | */ |
| 1387 | enum sym_type |
| 1388 | { |
| 1389 | st_none, st_C_struct, st_C_enum, st_C_define, st_C_typedef, st_C_typespec |
| 1390 | }; |
| 1391 | |
| 1392 | /* Feed stuff between (but not including) %[ and %] lines to: |
| 1393 | gperf -c -k1,3 -o -p -r -t |
| 1394 | %[ |
| 1395 | struct C_stab_entry { char *name; int c_ext; enum sym_type type; } |
| 1396 | %% |
| 1397 | class, C_PLPL, st_C_struct |
| 1398 | domain, C_STAR, st_C_struct |
| 1399 | union, 0, st_C_struct |
| 1400 | struct, 0, st_C_struct |
| 1401 | enum, 0, st_C_enum |
| 1402 | typedef, 0, st_C_typedef |
| 1403 | define, 0, st_C_define |
| 1404 | long, 0, st_C_typespec |
| 1405 | short, 0, st_C_typespec |
| 1406 | int, 0, st_C_typespec |
| 1407 | char, 0, st_C_typespec |
| 1408 | float, 0, st_C_typespec |
| 1409 | double, 0, st_C_typespec |
| 1410 | signed, 0, st_C_typespec |
| 1411 | unsigned, 0, st_C_typespec |
| 1412 | auto, 0, st_C_typespec |
| 1413 | void, 0, st_C_typespec |
| 1414 | extern, 0, st_C_typespec |
| 1415 | static, 0, st_C_typespec |
| 1416 | const, 0, st_C_typespec |
| 1417 | volatile, 0, st_C_typespec |
| 1418 | %] |
| 1419 | and replace lines between %< and %> with its output. */ |
| 1420 | /*%<*/ |
| 1421 | /* C code produced by gperf version 1.8.1 (K&R C version) */ |
| 1422 | /* Command-line: gperf -c -k1,3 -o -p -r -t */ |
| 1423 | |
| 1424 | |
| 1425 | struct C_stab_entry { char *name; int c_ext; enum sym_type type; }; |
| 1426 | |
| 1427 | #define MIN_WORD_LENGTH 3 |
| 1428 | #define MAX_WORD_LENGTH 8 |
| 1429 | #define MIN_HASH_VALUE 10 |
| 1430 | #define MAX_HASH_VALUE 62 |
| 1431 | /* |
| 1432 | 21 keywords |
| 1433 | 53 is the maximum key range |
| 1434 | */ |
| 1435 | |
| 1436 | static int |
| 1437 | hash (str, len) |
| 1438 | register char *str; |
| 1439 | register int len; |
| 1440 | { |
| 1441 | static unsigned char hash_table[] = |
| 1442 | { |
| 1443 | 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, |
| 1444 | 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, |
| 1445 | 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, |
| 1446 | 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, |
| 1447 | 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, |
| 1448 | 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, |
| 1449 | 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, |
| 1450 | 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, |
| 1451 | 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, |
| 1452 | 62, 62, 62, 62, 62, 62, 62, 2, 62, 7, |
| 1453 | 6, 9, 15, 30, 62, 24, 62, 62, 1, 24, |
| 1454 | 7, 27, 13, 62, 19, 26, 18, 27, 1, 62, |
| 1455 | 62, 62, 62, 62, 62, 62, 62, 62, |
| 1456 | }; |
| 1457 | return len + hash_table[str[2]] + hash_table[str[0]]; |
| 1458 | } |
| 1459 | |
| 1460 | struct C_stab_entry * |
| 1461 | in_word_set (str, len) |
| 1462 | register char *str; |
| 1463 | register int len; |
| 1464 | { |
| 1465 | |
| 1466 | static struct C_stab_entry wordlist[] = |
| 1467 | { |
| 1468 | {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, |
| 1469 | {"",}, |
| 1470 | {"volatile", 0, st_C_typespec}, |
| 1471 | {"",}, |
| 1472 | {"long", 0, st_C_typespec}, |
| 1473 | {"char", 0, st_C_typespec}, |
| 1474 | {"class", C_PLPL, st_C_struct}, |
| 1475 | {"",}, {"",}, {"",}, {"",}, |
| 1476 | {"const", 0, st_C_typespec}, |
| 1477 | {"",}, {"",}, {"",}, {"",}, |
| 1478 | {"auto", 0, st_C_typespec}, |
| 1479 | {"",}, {"",}, |
| 1480 | {"define", 0, st_C_define}, |
| 1481 | {"",}, |
| 1482 | {"void", 0, st_C_typespec}, |
| 1483 | {"",}, {"",}, {"",}, |
| 1484 | {"extern", 0, st_C_typespec}, |
| 1485 | {"static", 0, st_C_typespec}, |
| 1486 | {"",}, |
| 1487 | {"domain", C_STAR, st_C_struct}, |
| 1488 | {"",}, |
| 1489 | {"typedef", 0, st_C_typedef}, |
| 1490 | {"double", 0, st_C_typespec}, |
| 1491 | {"enum", 0, st_C_enum}, |
| 1492 | {"",}, {"",}, {"",}, {"",}, |
| 1493 | {"int", 0, st_C_typespec}, |
| 1494 | {"",}, |
| 1495 | {"float", 0, st_C_typespec}, |
| 1496 | {"",}, {"",}, {"",}, |
| 1497 | {"struct", 0, st_C_struct}, |
| 1498 | {"",}, {"",}, {"",}, {"",}, |
| 1499 | {"union", 0, st_C_struct}, |
| 1500 | {"",}, |
| 1501 | {"short", 0, st_C_typespec}, |
| 1502 | {"",}, {"",}, |
| 1503 | {"unsigned", 0, st_C_typespec}, |
| 1504 | {"signed", 0, st_C_typespec}, |
| 1505 | }; |
| 1506 | |
| 1507 | if (len <= MAX_WORD_LENGTH && len >= MIN_WORD_LENGTH) |
| 1508 | { |
| 1509 | register int key = hash (str, len); |
| 1510 | |
| 1511 | if (key <= MAX_HASH_VALUE && key >= MIN_HASH_VALUE) |
| 1512 | { |
| 1513 | register char *s = wordlist[key].name; |
| 1514 | |
| 1515 | if (*s == *str && strneq (str + 1, s + 1, len - 1)) |
| 1516 | return &wordlist[key]; |
| 1517 | } |
| 1518 | } |
| 1519 | return 0; |
| 1520 | } |
| 1521 | /*%>*/ |
| 1522 | |
| 1523 | enum sym_type |
| 1524 | C_symtype(str, len, c_ext) |
| 1525 | char *str; |
| 1526 | int len; |
| 1527 | int c_ext; |
| 1528 | { |
| 1529 | register struct C_stab_entry *se = in_word_set(str, len); |
| 1530 | |
| 1531 | if (se == NULL || (se->c_ext && !(c_ext & se->c_ext))) |
| 1532 | return st_none; |
| 1533 | return se->type; |
| 1534 | } |
| 1535 | \f |
| 1536 | /* |
| 1537 | * C functions are recognized using a simple finite automaton. |
| 1538 | * funcdef is its state variable. |
| 1539 | */ |
| 1540 | typedef enum |
| 1541 | { |
| 1542 | fnone, /* nothing seen */ |
| 1543 | ftagseen, /* function-like tag seen */ |
| 1544 | fstartlist, /* just after open parenthesis */ |
| 1545 | finlist, /* in parameter list */ |
| 1546 | flistseen, /* after parameter list */ |
| 1547 | fignore /* before open brace */ |
| 1548 | } FUNCST; |
| 1549 | FUNCST funcdef; |
| 1550 | |
| 1551 | |
| 1552 | /* |
| 1553 | * typedefs are recognized using a simple finite automaton. |
| 1554 | * typeddef is its state variable. |
| 1555 | */ |
| 1556 | typedef enum |
| 1557 | { |
| 1558 | tnone, /* nothing seen */ |
| 1559 | ttypedseen, /* typedef keyword seen */ |
| 1560 | tinbody, /* inside typedef body */ |
| 1561 | tend, /* just before typedef tag */ |
| 1562 | tignore /* junk after typedef tag */ |
| 1563 | } TYPEDST; |
| 1564 | TYPEDST typdef; |
| 1565 | |
| 1566 | |
| 1567 | /* |
| 1568 | * struct-like structures (enum, struct and union) are recognized |
| 1569 | * using another simple finite automaton. `structdef' is its state |
| 1570 | * variable. |
| 1571 | */ |
| 1572 | typedef enum |
| 1573 | { |
| 1574 | snone, /* nothing seen yet */ |
| 1575 | skeyseen, /* struct-like keyword seen */ |
| 1576 | stagseen, /* struct-like tag seen */ |
| 1577 | scolonseen, /* colon seen after struct-like tag */ |
| 1578 | sinbody /* in struct body: recognize member func defs*/ |
| 1579 | } STRUCTST; |
| 1580 | STRUCTST structdef; |
| 1581 | |
| 1582 | /* |
| 1583 | * When structdef is stagseen, scolonseen, or sinbody, structtag is the |
| 1584 | * struct tag, and structtype is the type of the preceding struct-like |
| 1585 | * keyword. |
| 1586 | */ |
| 1587 | char *structtag = "<uninited>"; |
| 1588 | enum sym_type structtype; |
| 1589 | |
| 1590 | /* |
| 1591 | * Yet another little state machine to deal with preprocessor lines. |
| 1592 | */ |
| 1593 | typedef enum |
| 1594 | { |
| 1595 | dnone, /* nothing seen */ |
| 1596 | dsharpseen, /* '#' seen as first char on line */ |
| 1597 | ddefineseen, /* '#' and 'define' seen */ |
| 1598 | dignorerest /* ignore rest of line */ |
| 1599 | } DEFINEST; |
| 1600 | DEFINEST definedef; |
| 1601 | |
| 1602 | /* |
| 1603 | * Set this to TRUE, and the next token considered is called a function. |
| 1604 | * Used only for GNU emacs's function-defining macros. |
| 1605 | */ |
| 1606 | logical next_token_is_func; |
| 1607 | |
| 1608 | /* |
| 1609 | * TRUE in the rules part of a yacc file, FALSE outside (parse as C). |
| 1610 | */ |
| 1611 | logical yacc_rules; |
| 1612 | |
| 1613 | /* |
| 1614 | * consider_token () |
| 1615 | * checks to see if the current token is at the start of a |
| 1616 | * function, or corresponds to a typedef, or is a struct/union/enum |
| 1617 | * tag. |
| 1618 | * |
| 1619 | * *IS_FUNC gets TRUE iff the token is a function or macro with args. |
| 1620 | * C_EXT is which language we are looking at. |
| 1621 | * |
| 1622 | * In the future we will need some way to adjust where the end of |
| 1623 | * the token is; for instance, implementing the C++ keyword |
| 1624 | * `operator' properly will adjust the end of the token to be after |
| 1625 | * whatever follows `operator'. |
| 1626 | * |
| 1627 | * Globals |
| 1628 | * funcdef IN OUT |
| 1629 | * structdef IN OUT |
| 1630 | * definedef IN OUT |
| 1631 | * typdef IN OUT |
| 1632 | * next_token_is_func IN OUT |
| 1633 | */ |
| 1634 | |
| 1635 | logical |
| 1636 | consider_token (str, len, c, c_ext, cblev, is_func) |
| 1637 | register char *str; /* IN: token pointer */ |
| 1638 | register int len; /* IN: token length */ |
| 1639 | register char c; /* IN: first char after the token */ |
| 1640 | int c_ext; /* IN: C extensions mask */ |
| 1641 | int cblev; /* IN: curly brace level */ |
| 1642 | logical *is_func; /* OUT: function found */ |
| 1643 | { |
| 1644 | enum sym_type toktype = C_symtype (str, len, c_ext); |
| 1645 | |
| 1646 | /* |
| 1647 | * Advance the definedef state machine. |
| 1648 | */ |
| 1649 | switch (definedef) |
| 1650 | { |
| 1651 | case dnone: |
| 1652 | /* We're not on a preprocessor line. */ |
| 1653 | break; |
| 1654 | case dsharpseen: |
| 1655 | if (toktype == st_C_define) |
| 1656 | { |
| 1657 | definedef = ddefineseen; |
| 1658 | } |
| 1659 | else |
| 1660 | { |
| 1661 | definedef = dignorerest; |
| 1662 | } |
| 1663 | return FALSE; |
| 1664 | case ddefineseen: |
| 1665 | /* |
| 1666 | * Make a tag for any macro, unless it is a constant |
| 1667 | * and constantypedefs is FALSE. |
| 1668 | */ |
| 1669 | definedef = dignorerest; |
| 1670 | *is_func = (c == '('); |
| 1671 | if (!*is_func && !constantypedefs) |
| 1672 | return FALSE; |
| 1673 | else |
| 1674 | return TRUE; |
| 1675 | case dignorerest: |
| 1676 | return FALSE; |
| 1677 | default: |
| 1678 | error ("internal error: definedef value.", 0); |
| 1679 | } |
| 1680 | |
| 1681 | /* |
| 1682 | * Now typedefs |
| 1683 | */ |
| 1684 | switch (typdef) |
| 1685 | { |
| 1686 | case tnone: |
| 1687 | if (toktype == st_C_typedef) |
| 1688 | { |
| 1689 | if (typedefs) |
| 1690 | typdef = ttypedseen; |
| 1691 | funcdef = fnone; |
| 1692 | return FALSE; |
| 1693 | } |
| 1694 | break; |
| 1695 | case ttypedseen: |
| 1696 | switch (toktype) |
| 1697 | { |
| 1698 | case st_none: |
| 1699 | case st_C_typespec: |
| 1700 | typdef = tend; |
| 1701 | break; |
| 1702 | case st_C_struct: |
| 1703 | case st_C_enum: |
| 1704 | break; |
| 1705 | } |
| 1706 | /* Do not return here, so the structdef stuff has a chance. */ |
| 1707 | break; |
| 1708 | case tend: |
| 1709 | switch (toktype) |
| 1710 | { |
| 1711 | case st_C_typespec: |
| 1712 | case st_C_struct: |
| 1713 | case st_C_enum: |
| 1714 | return FALSE; |
| 1715 | } |
| 1716 | return TRUE; |
| 1717 | } |
| 1718 | |
| 1719 | /* |
| 1720 | * This structdef business is currently only invoked when cblev==0. |
| 1721 | * It should be recursively invoked whatever the curly brace level, |
| 1722 | * and a stack of states kept, to allow for definitions of structs |
| 1723 | * within structs. |
| 1724 | * |
| 1725 | * This structdef business is NOT invoked when we are ctags and the |
| 1726 | * file is plain C. This is because a struct tag may have the same |
| 1727 | * name as another tag, and this loses with ctags. |
| 1728 | * |
| 1729 | * This if statement deals with the typdef state machine as |
| 1730 | * follows: if typdef==ttypedseen and token is struct/union/class/enum, |
| 1731 | * return FALSE. All the other code here is for the structdef |
| 1732 | * state machine. |
| 1733 | */ |
| 1734 | switch (toktype) |
| 1735 | { |
| 1736 | case st_C_struct: |
| 1737 | case st_C_enum: |
| 1738 | if (typdef == ttypedseen |
| 1739 | || (typedefs_and_cplusplus && cblev == 0 && structdef == snone)) |
| 1740 | { |
| 1741 | structdef = skeyseen; |
| 1742 | structtype = toktype; |
| 1743 | } |
| 1744 | return FALSE; |
| 1745 | } |
| 1746 | if (structdef == skeyseen) |
| 1747 | { |
| 1748 | /* Save the tag for struct/union/class, for functions that may be |
| 1749 | defined inside. */ |
| 1750 | if (structtype == st_C_struct) |
| 1751 | structtag = savenstr (str, len); |
| 1752 | else |
| 1753 | structtag = "<enum>"; |
| 1754 | structdef = stagseen; |
| 1755 | return TRUE; |
| 1756 | } |
| 1757 | |
| 1758 | /* Avoid entering funcdef stuff if typdef is going on. */ |
| 1759 | if (typdef != tnone) |
| 1760 | { |
| 1761 | definedef = dnone; |
| 1762 | return FALSE; |
| 1763 | } |
| 1764 | |
| 1765 | /* Detect GNU macros. */ |
| 1766 | if (definedef == dnone) |
| 1767 | if (strneq (str, "DEFUN", 5) /* Used in emacs */ |
| 1768 | #if FALSE |
| 1769 | These are defined inside C functions, so currently they |
| 1770 | are not met anyway. |
| 1771 | || strneq (str, "EXFUN", 5) /* Used in glibc */ |
| 1772 | || strneq (str, "DEFVAR_", 7) /* Used in emacs */ |
| 1773 | #endif |
| 1774 | || strneq (str, "SYSCALL", 7) /* Used in glibc (mach) */ |
| 1775 | || strneq (str, "ENTRY", 5) /* Used in glibc */ |
| 1776 | || strneq (str, "PSEUDO", 6)) /* Used in glibc */ |
| 1777 | |
| 1778 | { |
| 1779 | next_token_is_func = TRUE; |
| 1780 | return FALSE; |
| 1781 | } |
| 1782 | if (next_token_is_func) |
| 1783 | { |
| 1784 | next_token_is_func = FALSE; |
| 1785 | funcdef = fignore; |
| 1786 | *is_func = TRUE; |
| 1787 | return TRUE; |
| 1788 | } |
| 1789 | |
| 1790 | /* A function? */ |
| 1791 | switch (toktype) |
| 1792 | { |
| 1793 | case st_C_typespec: |
| 1794 | if (funcdef != finlist && funcdef != fignore) |
| 1795 | funcdef = fnone; /* should be useless */ |
| 1796 | return FALSE; |
| 1797 | default: |
| 1798 | if (funcdef == fnone) |
| 1799 | { |
| 1800 | funcdef = ftagseen; |
| 1801 | *is_func = TRUE; |
| 1802 | return TRUE; |
| 1803 | } |
| 1804 | } |
| 1805 | |
| 1806 | return FALSE; |
| 1807 | } |
| 1808 | |
| 1809 | /* |
| 1810 | * C_entries () |
| 1811 | * This routine finds functions, typedefs, #define's and |
| 1812 | * struct/union/enum definitions in C syntax and adds them |
| 1813 | * to the list. |
| 1814 | */ |
| 1815 | typedef struct |
| 1816 | { |
| 1817 | logical valid; |
| 1818 | char *str; |
| 1819 | logical named; |
| 1820 | int linelen; |
| 1821 | int lineno; |
| 1822 | long linepos; |
| 1823 | char *buffer; |
| 1824 | } TOKEN; |
| 1825 | |
| 1826 | #define current_lb_is_new (newndx == curndx) |
| 1827 | #define switch_line_buffers() (curndx = 1 - curndx) |
| 1828 | |
| 1829 | #define curlb (lbs[curndx].lb) |
| 1830 | #define othlb (lbs[1-curndx].lb) |
| 1831 | #define newlb (lbs[newndx].lb) |
| 1832 | #define curlinepos (lbs[curndx].linepos) |
| 1833 | #define othlinepos (lbs[1-curndx].linepos) |
| 1834 | #define newlinepos (lbs[newndx].linepos) |
| 1835 | |
| 1836 | #define CNL_SAVE_DEFINEDEF \ |
| 1837 | do { \ |
| 1838 | curlinepos = charno; \ |
| 1839 | lineno++; \ |
| 1840 | charno += readline (&curlb, inf); \ |
| 1841 | lp = curlb.buffer; \ |
| 1842 | quotednl = FALSE; \ |
| 1843 | newndx = curndx; \ |
| 1844 | } while (0) |
| 1845 | |
| 1846 | #define CNL \ |
| 1847 | do { \ |
| 1848 | CNL_SAVE_DEFINEDEF; \ |
| 1849 | if (savetok.valid) \ |
| 1850 | { \ |
| 1851 | tok = savetok; \ |
| 1852 | savetok.valid = FALSE; \ |
| 1853 | } \ |
| 1854 | definedef = dnone; \ |
| 1855 | } while (0) |
| 1856 | |
| 1857 | #define make_tag(isfun) do \ |
| 1858 | { \ |
| 1859 | if (tok.valid) \ |
| 1860 | pfnote (savestr (token_name.buffer), isfun, tok.named, \ |
| 1861 | tok.buffer, tok.linelen, tok.lineno, tok.linepos); \ |
| 1862 | else if (DEBUG) abort (); \ |
| 1863 | tok.valid = FALSE; \ |
| 1864 | } while (0) |
| 1865 | |
| 1866 | void |
| 1867 | C_entries (c_ext, inf) |
| 1868 | int c_ext; /* extension of C */ |
| 1869 | FILE *inf; /* input file */ |
| 1870 | { |
| 1871 | register char c; /* latest char read; '\0' for end of line */ |
| 1872 | register char *lp; /* pointer one beyond the character `c' */ |
| 1873 | int curndx, newndx; /* indices for current and new lb */ |
| 1874 | TOKEN tok; /* latest token read */ |
| 1875 | register int tokoff; /* offset in line of start of current token */ |
| 1876 | register int toklen; /* length of current token */ |
| 1877 | int cblev; /* current curly brace level */ |
| 1878 | int parlev; /* current parenthesis level */ |
| 1879 | logical incomm, inquote, inchar, quotednl, midtoken; |
| 1880 | logical cplpl; |
| 1881 | TOKEN savetok; /* token saved during preprocessor handling */ |
| 1882 | |
| 1883 | |
| 1884 | curndx = newndx = 0; |
| 1885 | lineno = 0; |
| 1886 | charno = 0; |
| 1887 | lp = curlb.buffer; |
| 1888 | *lp = 0; |
| 1889 | |
| 1890 | definedef = dnone; funcdef = fnone; typdef = tnone; structdef = snone; |
| 1891 | next_token_is_func = yacc_rules = FALSE; |
| 1892 | midtoken = inquote = inchar = incomm = quotednl = FALSE; |
| 1893 | tok.valid = savetok.valid = FALSE; |
| 1894 | cblev = 0; |
| 1895 | parlev = 0; |
| 1896 | cplpl = c_ext & C_PLPL; |
| 1897 | |
| 1898 | while (!feof (inf)) |
| 1899 | { |
| 1900 | c = *lp++; |
| 1901 | if (c == '\\') |
| 1902 | { |
| 1903 | /* If we're at the end of the line, the next character is a |
| 1904 | '\0'; don't skip it, because it's the thing that tells us |
| 1905 | to read the next line. */ |
| 1906 | if (*lp == '\0') |
| 1907 | { |
| 1908 | quotednl = TRUE; |
| 1909 | continue; |
| 1910 | } |
| 1911 | lp++; |
| 1912 | c = ' '; |
| 1913 | } |
| 1914 | else if (incomm) |
| 1915 | { |
| 1916 | switch (c) |
| 1917 | { |
| 1918 | case '*': |
| 1919 | if (*lp == '/') |
| 1920 | { |
| 1921 | c = *lp++; |
| 1922 | incomm = FALSE; |
| 1923 | } |
| 1924 | break; |
| 1925 | case '\0': |
| 1926 | /* Newlines inside comments do not end macro definitions in |
| 1927 | traditional cpp. */ |
| 1928 | CNL_SAVE_DEFINEDEF; |
| 1929 | break; |
| 1930 | } |
| 1931 | continue; |
| 1932 | } |
| 1933 | else if (inquote) |
| 1934 | { |
| 1935 | switch (c) |
| 1936 | { |
| 1937 | case '"': |
| 1938 | inquote = FALSE; |
| 1939 | break; |
| 1940 | case '\0': |
| 1941 | /* Newlines inside strings do not end macro definitions |
| 1942 | in traditional cpp, even though compilers don't |
| 1943 | usually accept them. */ |
| 1944 | CNL_SAVE_DEFINEDEF; |
| 1945 | break; |
| 1946 | } |
| 1947 | continue; |
| 1948 | } |
| 1949 | else if (inchar) |
| 1950 | { |
| 1951 | switch (c) |
| 1952 | { |
| 1953 | case '\0': |
| 1954 | /* Hmmm, something went wrong. */ |
| 1955 | CNL; |
| 1956 | /* FALLTHRU */ |
| 1957 | case '\'': |
| 1958 | inchar = FALSE; |
| 1959 | break; |
| 1960 | } |
| 1961 | continue; |
| 1962 | } |
| 1963 | else |
| 1964 | switch (c) |
| 1965 | { |
| 1966 | case '"': |
| 1967 | inquote = TRUE; |
| 1968 | if (funcdef != finlist && funcdef != fignore) |
| 1969 | funcdef = fnone; |
| 1970 | continue; |
| 1971 | case '\'': |
| 1972 | inchar = TRUE; |
| 1973 | if (funcdef != finlist && funcdef != fignore) |
| 1974 | funcdef = fnone; |
| 1975 | continue; |
| 1976 | case '/': |
| 1977 | if (*lp == '*') |
| 1978 | { |
| 1979 | lp++; |
| 1980 | incomm = TRUE; |
| 1981 | continue; |
| 1982 | } |
| 1983 | else if (cplpl && *lp == '/') |
| 1984 | { |
| 1985 | c = 0; |
| 1986 | break; |
| 1987 | } |
| 1988 | else |
| 1989 | break; |
| 1990 | case '%': |
| 1991 | if ((c_ext & YACC) && *lp == '%') |
| 1992 | { |
| 1993 | /* entering or exiting rules section in yacc file */ |
| 1994 | lp++; |
| 1995 | definedef = dnone; funcdef = fnone; |
| 1996 | typdef = tnone; structdef = snone; |
| 1997 | next_token_is_func = FALSE; |
| 1998 | midtoken = inquote = inchar = incomm = quotednl = FALSE; |
| 1999 | cblev = 0; |
| 2000 | yacc_rules = !yacc_rules; |
| 2001 | continue; |
| 2002 | } |
| 2003 | else |
| 2004 | break; |
| 2005 | case '#': |
| 2006 | if (definedef == dnone) |
| 2007 | { |
| 2008 | char *cp; |
| 2009 | logical cpptoken = TRUE; |
| 2010 | |
| 2011 | /* Look back on this line. If all blanks, or nonblanks |
| 2012 | followed by an end of comment, this is a preprocessor |
| 2013 | token. */ |
| 2014 | for (cp = newlb.buffer; cp < lp-1; cp++) |
| 2015 | if (!iswhite (*cp)) |
| 2016 | { |
| 2017 | if (*cp == '*' && *(cp+1) == '/') |
| 2018 | { |
| 2019 | cp++; |
| 2020 | cpptoken = TRUE; |
| 2021 | } |
| 2022 | else |
| 2023 | cpptoken = FALSE; |
| 2024 | } |
| 2025 | if (cpptoken) |
| 2026 | definedef = dsharpseen; |
| 2027 | } /* if (definedef == dnone) */ |
| 2028 | |
| 2029 | continue; |
| 2030 | } /* switch (c) */ |
| 2031 | |
| 2032 | |
| 2033 | /* Consider token only if some complicated conditions are satisfied. */ |
| 2034 | if ((definedef != dnone |
| 2035 | || (cblev == 0 && structdef != scolonseen) |
| 2036 | || (cblev == 1 && cplpl && structdef == sinbody)) |
| 2037 | && typdef != tignore |
| 2038 | && definedef != dignorerest |
| 2039 | && funcdef != finlist) |
| 2040 | { |
| 2041 | if (midtoken) |
| 2042 | { |
| 2043 | if (endtoken (c)) |
| 2044 | { |
| 2045 | if (cplpl && c == ':' && *lp == ':' && begtoken(*(lp + 1))) |
| 2046 | { |
| 2047 | /* |
| 2048 | * This handles :: in the middle, but not at the |
| 2049 | * beginning of an identifier. |
| 2050 | */ |
| 2051 | lp += 2; |
| 2052 | toklen += 3; |
| 2053 | } |
| 2054 | else |
| 2055 | { |
| 2056 | logical is_func = FALSE; |
| 2057 | |
| 2058 | if (yacc_rules |
| 2059 | || consider_token (newlb.buffer + tokoff, toklen, |
| 2060 | c, c_ext, cblev, &is_func)) |
| 2061 | { |
| 2062 | if (structdef == sinbody |
| 2063 | && definedef == dnone |
| 2064 | && is_func) |
| 2065 | /* function defined in C++ class body */ |
| 2066 | { |
| 2067 | int strsize = strlen(structtag) + 2 + toklen + 1; |
| 2068 | while (token_name.size < strsize) |
| 2069 | { |
| 2070 | token_name.size *= 2; |
| 2071 | token_name.buffer |
| 2072 | = (char *) xrealloc (token_name.buffer, |
| 2073 | token_name.size); |
| 2074 | } |
| 2075 | strcpy (token_name.buffer, structtag); |
| 2076 | strcat (token_name.buffer, "::"); |
| 2077 | strncat (token_name.buffer, |
| 2078 | newlb.buffer+tokoff, toklen); |
| 2079 | tok.named = TRUE; |
| 2080 | } |
| 2081 | else |
| 2082 | { |
| 2083 | while (token_name.size < toklen + 1) |
| 2084 | { |
| 2085 | token_name.size *= 2; |
| 2086 | token_name.buffer |
| 2087 | = (char *) xrealloc (token_name.buffer, |
| 2088 | token_name.size); |
| 2089 | } |
| 2090 | strncpy (token_name.buffer, |
| 2091 | newlb.buffer+tokoff, toklen); |
| 2092 | token_name.buffer[toklen] = '\0'; |
| 2093 | if (structdef == stagseen |
| 2094 | || typdef == tend |
| 2095 | || (is_func |
| 2096 | && definedef == dignorerest)) /* macro */ |
| 2097 | tok.named = TRUE; |
| 2098 | else |
| 2099 | tok.named = FALSE; |
| 2100 | } |
| 2101 | tok.lineno = lineno; |
| 2102 | tok.linelen = tokoff + toklen + 1; |
| 2103 | tok.buffer = newlb.buffer; |
| 2104 | tok.linepos = newlinepos; |
| 2105 | tok.valid = TRUE; |
| 2106 | |
| 2107 | if (definedef == dnone |
| 2108 | && (funcdef == ftagseen |
| 2109 | || structdef == stagseen |
| 2110 | || typdef == tend)) |
| 2111 | { |
| 2112 | if (current_lb_is_new) |
| 2113 | switch_line_buffers (); |
| 2114 | } |
| 2115 | else |
| 2116 | make_tag (is_func); |
| 2117 | } |
| 2118 | midtoken = FALSE; |
| 2119 | } |
| 2120 | } /* if (endtoken (c)) */ |
| 2121 | else if (intoken (c)) |
| 2122 | { |
| 2123 | toklen++; |
| 2124 | continue; |
| 2125 | } |
| 2126 | } /* if (midtoken) */ |
| 2127 | else if (begtoken (c)) |
| 2128 | { |
| 2129 | switch (definedef) |
| 2130 | { |
| 2131 | case dnone: |
| 2132 | switch (funcdef) |
| 2133 | { |
| 2134 | case fstartlist: |
| 2135 | funcdef = finlist; |
| 2136 | continue; |
| 2137 | case flistseen: |
| 2138 | make_tag (TRUE); |
| 2139 | funcdef = fignore; |
| 2140 | break; |
| 2141 | case ftagseen: |
| 2142 | funcdef = fnone; |
| 2143 | break; |
| 2144 | } |
| 2145 | if (structdef == stagseen) |
| 2146 | structdef = snone; |
| 2147 | break; |
| 2148 | case dsharpseen: |
| 2149 | savetok = tok; |
| 2150 | } |
| 2151 | if (!yacc_rules || lp == newlb.buffer + 1) |
| 2152 | { |
| 2153 | tokoff = lp - 1 - newlb.buffer; |
| 2154 | toklen = 1; |
| 2155 | midtoken = TRUE; |
| 2156 | } |
| 2157 | continue; |
| 2158 | } /* if (begtoken) */ |
| 2159 | } /* if must look at token */ |
| 2160 | |
| 2161 | |
| 2162 | /* Detect end of line, colon, comma, semicolon and various braces |
| 2163 | after having handled a token.*/ |
| 2164 | switch (c) |
| 2165 | { |
| 2166 | case ':': |
| 2167 | if (definedef != dnone) |
| 2168 | break; |
| 2169 | if (structdef == stagseen) |
| 2170 | structdef = scolonseen; |
| 2171 | else |
| 2172 | switch (funcdef) |
| 2173 | { |
| 2174 | case ftagseen: |
| 2175 | if (yacc_rules) |
| 2176 | { |
| 2177 | make_tag (FALSE); |
| 2178 | funcdef = fignore; |
| 2179 | } |
| 2180 | break; |
| 2181 | case fstartlist: |
| 2182 | funcdef = fnone; |
| 2183 | break; |
| 2184 | } |
| 2185 | break; |
| 2186 | case ';': |
| 2187 | if (definedef != dnone) |
| 2188 | break; |
| 2189 | if (cblev == 0) |
| 2190 | switch (typdef) |
| 2191 | { |
| 2192 | case tend: |
| 2193 | make_tag (FALSE); |
| 2194 | /* FALLTHRU */ |
| 2195 | default: |
| 2196 | typdef = tnone; |
| 2197 | } |
| 2198 | if (funcdef != fignore) |
| 2199 | funcdef = fnone; |
| 2200 | if (structdef == stagseen) |
| 2201 | structdef = snone; |
| 2202 | break; |
| 2203 | case ',': |
| 2204 | if (definedef != dnone) |
| 2205 | break; |
| 2206 | if (funcdef != finlist && funcdef != fignore) |
| 2207 | funcdef = fnone; |
| 2208 | if (structdef == stagseen) |
| 2209 | structdef = snone; |
| 2210 | break; |
| 2211 | case '[': |
| 2212 | if (definedef != dnone) |
| 2213 | break; |
| 2214 | if (cblev == 0 && typdef == tend) |
| 2215 | { |
| 2216 | typdef = tignore; |
| 2217 | make_tag (FALSE); |
| 2218 | break; |
| 2219 | } |
| 2220 | if (funcdef != finlist && funcdef != fignore) |
| 2221 | funcdef = fnone; |
| 2222 | if (structdef == stagseen) |
| 2223 | structdef = snone; |
| 2224 | break; |
| 2225 | case '(': |
| 2226 | if (definedef != dnone) |
| 2227 | break; |
| 2228 | switch (funcdef) |
| 2229 | { |
| 2230 | case fnone: |
| 2231 | switch (typdef) |
| 2232 | { |
| 2233 | case ttypedseen: |
| 2234 | case tend: |
| 2235 | /* Make sure that the next char is not a '*'. |
| 2236 | This handles constructs like: |
| 2237 | typedef void OperatorFun (int fun); */ |
| 2238 | if (*lp != '*') |
| 2239 | { |
| 2240 | typdef = tignore; |
| 2241 | make_tag (FALSE); |
| 2242 | } |
| 2243 | break; |
| 2244 | } /* switch (typdef) */ |
| 2245 | break; |
| 2246 | case ftagseen: |
| 2247 | funcdef = fstartlist; |
| 2248 | break; |
| 2249 | case flistseen: |
| 2250 | funcdef = finlist; |
| 2251 | break; |
| 2252 | } |
| 2253 | parlev++; |
| 2254 | break; |
| 2255 | case ')': |
| 2256 | if (definedef != dnone) |
| 2257 | break; |
| 2258 | if (--parlev == 0) |
| 2259 | { |
| 2260 | switch (funcdef) |
| 2261 | { |
| 2262 | case fstartlist: |
| 2263 | case finlist: |
| 2264 | funcdef = flistseen; |
| 2265 | break; |
| 2266 | } |
| 2267 | if (cblev == 0 && typdef == tend) |
| 2268 | { |
| 2269 | typdef = tignore; |
| 2270 | make_tag (FALSE); |
| 2271 | } |
| 2272 | } |
| 2273 | else if (parlev < 0) /* can happen due to ill-conceived #if's. */ |
| 2274 | parlev = 0; |
| 2275 | break; |
| 2276 | case '{': |
| 2277 | if (definedef != dnone) |
| 2278 | break; |
| 2279 | if (typdef == ttypedseen) |
| 2280 | typdef = tinbody; |
| 2281 | switch (structdef) |
| 2282 | { |
| 2283 | case skeyseen: /* unnamed struct */ |
| 2284 | structtag = "_anonymous_"; |
| 2285 | structdef = sinbody; |
| 2286 | break; |
| 2287 | case stagseen: |
| 2288 | case scolonseen: /* named struct */ |
| 2289 | structdef = sinbody; |
| 2290 | make_tag (FALSE); |
| 2291 | break; |
| 2292 | } |
| 2293 | switch (funcdef) |
| 2294 | { |
| 2295 | case flistseen: |
| 2296 | make_tag (TRUE); |
| 2297 | /* FALLTHRU */ |
| 2298 | case fignore: |
| 2299 | funcdef = fnone; |
| 2300 | break; |
| 2301 | case fnone: |
| 2302 | /* Neutralize `extern "C" {' grot and look inside structs. */ |
| 2303 | if (cblev == 0 && structdef == snone && typdef == tnone) |
| 2304 | cblev = -1; |
| 2305 | } |
| 2306 | cblev++; |
| 2307 | break; |
| 2308 | case '*': |
| 2309 | if (definedef != dnone) |
| 2310 | break; |
| 2311 | if (funcdef == fstartlist) |
| 2312 | funcdef = fnone; /* avoid tagging `foo' in `foo (*bar()) ()' */ |
| 2313 | break; |
| 2314 | case '}': |
| 2315 | if (definedef != dnone) |
| 2316 | break; |
| 2317 | if (!noindentypedefs && lp == newlb.buffer + 1) |
| 2318 | { |
| 2319 | cblev = 0; /* reset curly brace level if first column */ |
| 2320 | parlev = 0; /* also reset paren level, just in case... */ |
| 2321 | } |
| 2322 | else if (cblev > 0) |
| 2323 | cblev--; |
| 2324 | if (cblev == 0) |
| 2325 | { |
| 2326 | if (typdef == tinbody) |
| 2327 | typdef = tend; |
| 2328 | #if FALSE /* too risky */ |
| 2329 | if (structdef == sinbody) |
| 2330 | free (structtag); |
| 2331 | #endif |
| 2332 | |
| 2333 | structdef = snone; |
| 2334 | structtag = "<error>"; |
| 2335 | } |
| 2336 | break; |
| 2337 | case '=': |
| 2338 | case '#': case '+': case '-': case '~': case '&': case '%': case '/': |
| 2339 | case '|': case '^': case '!': case '<': case '>': case '.': case '?': |
| 2340 | if (definedef != dnone) |
| 2341 | break; |
| 2342 | /* These surely cannot follow a function tag. */ |
| 2343 | if (funcdef != finlist && funcdef != fignore) |
| 2344 | funcdef = fnone; |
| 2345 | break; |
| 2346 | case '\0': |
| 2347 | /* If a macro spans multiple lines don't reset its state. */ |
| 2348 | if (quotednl) |
| 2349 | CNL_SAVE_DEFINEDEF; |
| 2350 | else |
| 2351 | CNL; |
| 2352 | break; |
| 2353 | } /* switch (c) */ |
| 2354 | |
| 2355 | } /* while not eof */ |
| 2356 | } |
| 2357 | |
| 2358 | /* |
| 2359 | * Process either a C++ file or a C file depending on the setting |
| 2360 | * of a global flag. |
| 2361 | */ |
| 2362 | void |
| 2363 | default_C_entries (inf) |
| 2364 | FILE *inf; |
| 2365 | { |
| 2366 | C_entries (cplusplus ? C_PLPL : 0, inf); |
| 2367 | } |
| 2368 | |
| 2369 | /* Always do C++. */ |
| 2370 | void |
| 2371 | Cplusplus_entries (inf) |
| 2372 | FILE *inf; |
| 2373 | { |
| 2374 | C_entries (C_PLPL, inf); |
| 2375 | } |
| 2376 | |
| 2377 | /* Always do C*. */ |
| 2378 | void |
| 2379 | Cstar_entries (inf) |
| 2380 | FILE *inf; |
| 2381 | { |
| 2382 | C_entries (C_STAR, inf); |
| 2383 | } |
| 2384 | |
| 2385 | /* Always do Yacc. */ |
| 2386 | void |
| 2387 | Yacc_entries (inf) |
| 2388 | FILE *inf; |
| 2389 | { |
| 2390 | C_entries (YACC, inf); |
| 2391 | } |
| 2392 | \f |
| 2393 | /* Fortran parsing */ |
| 2394 | |
| 2395 | char *dbp; |
| 2396 | |
| 2397 | logical |
| 2398 | tail (cp) |
| 2399 | char *cp; |
| 2400 | { |
| 2401 | register int len = 0; |
| 2402 | |
| 2403 | while (*cp && (*cp | ' ') == (dbp[len] | ' ')) |
| 2404 | cp++, len++; |
| 2405 | if (*cp == 0) |
| 2406 | { |
| 2407 | dbp += len; |
| 2408 | return TRUE; |
| 2409 | } |
| 2410 | return FALSE; |
| 2411 | } |
| 2412 | |
| 2413 | void |
| 2414 | takeprec () |
| 2415 | { |
| 2416 | while (isspace (*dbp)) |
| 2417 | dbp++; |
| 2418 | if (*dbp != '*') |
| 2419 | return; |
| 2420 | dbp++; |
| 2421 | while (isspace (*dbp)) |
| 2422 | dbp++; |
| 2423 | if (tail ("(*)")) |
| 2424 | return; |
| 2425 | if (!isdigit (*dbp)) |
| 2426 | { |
| 2427 | --dbp; /* force failure */ |
| 2428 | return; |
| 2429 | } |
| 2430 | do |
| 2431 | dbp++; |
| 2432 | while (isdigit (*dbp)); |
| 2433 | } |
| 2434 | |
| 2435 | void |
| 2436 | getit (inf) |
| 2437 | FILE *inf; |
| 2438 | { |
| 2439 | register char *cp; |
| 2440 | |
| 2441 | while (isspace (*dbp)) |
| 2442 | dbp++; |
| 2443 | if (*dbp == '\0') |
| 2444 | { |
| 2445 | lineno++; |
| 2446 | linecharno = charno; |
| 2447 | charno += readline (&lb, inf); |
| 2448 | dbp = lb.buffer; |
| 2449 | if (dbp[5] != '&') |
| 2450 | return; |
| 2451 | dbp += 6; |
| 2452 | while (isspace (*dbp)) |
| 2453 | dbp++; |
| 2454 | } |
| 2455 | if (!isalpha (*dbp) |
| 2456 | && *dbp != '_' |
| 2457 | && *dbp != '$') |
| 2458 | return; |
| 2459 | for (cp = dbp + 1; |
| 2460 | (*cp |
| 2461 | && (isalpha (*cp) || isdigit (*cp) || (*cp == '_') || (*cp == '$'))); |
| 2462 | cp++) |
| 2463 | continue; |
| 2464 | pfnote (savenstr (dbp, cp-dbp), TRUE, FALSE, lb.buffer, |
| 2465 | cp - lb.buffer + 1, lineno, linecharno); |
| 2466 | } |
| 2467 | |
| 2468 | void |
| 2469 | Fortran_functions (inf) |
| 2470 | FILE *inf; |
| 2471 | { |
| 2472 | lineno = 0; |
| 2473 | charno = 0; |
| 2474 | |
| 2475 | while (!feof (inf)) |
| 2476 | { |
| 2477 | lineno++; |
| 2478 | linecharno = charno; |
| 2479 | charno += readline (&lb, inf); |
| 2480 | dbp = lb.buffer; |
| 2481 | if (*dbp == '%') |
| 2482 | dbp++; /* Ratfor escape to fortran */ |
| 2483 | while (isspace (*dbp)) |
| 2484 | dbp++; |
| 2485 | if (*dbp == 0) |
| 2486 | continue; |
| 2487 | switch (*dbp | ' ') |
| 2488 | { |
| 2489 | case 'i': |
| 2490 | if (tail ("integer")) |
| 2491 | takeprec (); |
| 2492 | break; |
| 2493 | case 'r': |
| 2494 | if (tail ("real")) |
| 2495 | takeprec (); |
| 2496 | break; |
| 2497 | case 'l': |
| 2498 | if (tail ("logical")) |
| 2499 | takeprec (); |
| 2500 | break; |
| 2501 | case 'c': |
| 2502 | if (tail ("complex") || tail ("character")) |
| 2503 | takeprec (); |
| 2504 | break; |
| 2505 | case 'd': |
| 2506 | if (tail ("double")) |
| 2507 | { |
| 2508 | while (isspace (*dbp)) |
| 2509 | dbp++; |
| 2510 | if (*dbp == 0) |
| 2511 | continue; |
| 2512 | if (tail ("precision")) |
| 2513 | break; |
| 2514 | continue; |
| 2515 | } |
| 2516 | break; |
| 2517 | } |
| 2518 | while (isspace (*dbp)) |
| 2519 | dbp++; |
| 2520 | if (*dbp == 0) |
| 2521 | continue; |
| 2522 | switch (*dbp | ' ') |
| 2523 | { |
| 2524 | case 'f': |
| 2525 | if (tail ("function")) |
| 2526 | getit (inf); |
| 2527 | continue; |
| 2528 | case 's': |
| 2529 | if (tail ("subroutine")) |
| 2530 | getit (inf); |
| 2531 | continue; |
| 2532 | case 'e': |
| 2533 | if (tail ("entry")) |
| 2534 | getit (inf); |
| 2535 | continue; |
| 2536 | case 'p': |
| 2537 | if (tail ("program")) |
| 2538 | { |
| 2539 | getit (inf); |
| 2540 | continue; |
| 2541 | } |
| 2542 | if (tail ("procedure")) |
| 2543 | getit (inf); |
| 2544 | continue; |
| 2545 | } |
| 2546 | } |
| 2547 | } |
| 2548 | \f |
| 2549 | /* |
| 2550 | * Bob Weiner, Motorola Inc., 4/3/94 |
| 2551 | * Unix and microcontroller assembly tag handling |
| 2552 | * look for '^[a-zA-Z_.$][a-zA_Z0-9_.$]*[: ^I^J]' |
| 2553 | */ |
| 2554 | void |
| 2555 | Asm_labels (inf) |
| 2556 | FILE *inf; |
| 2557 | { |
| 2558 | register char *cp; |
| 2559 | |
| 2560 | lineno = 0; |
| 2561 | charno = 0; |
| 2562 | |
| 2563 | while (!feof (inf)) |
| 2564 | { |
| 2565 | lineno++; |
| 2566 | linecharno = charno; |
| 2567 | charno += readline (&lb, inf); |
| 2568 | cp = lb.buffer; |
| 2569 | |
| 2570 | /* If first char is alphabetic or one of [_.$], test for colon |
| 2571 | following identifier. */ |
| 2572 | if (isalpha (*cp) || *cp == '_' || *cp == '.' || *cp == '$') |
| 2573 | { |
| 2574 | /* Read past label. */ |
| 2575 | cp++; |
| 2576 | while (isalnum (*cp) || *cp == '_' || *cp == '.' || *cp == '$') |
| 2577 | cp++; |
| 2578 | if (*cp == ':' || isspace (*cp)) |
| 2579 | { |
| 2580 | /* Found end of label, so copy it and add it to the table. */ |
| 2581 | pfnote (savenstr (lb.buffer, cp-lb.buffer), TRUE, FALSE, |
| 2582 | lb.buffer, cp - lb.buffer + 1, lineno, linecharno); |
| 2583 | } |
| 2584 | } |
| 2585 | } |
| 2586 | } |
| 2587 | \f |
| 2588 | /* Added by Mosur Mohan, 4/22/88 */ |
| 2589 | /* Pascal parsing */ |
| 2590 | |
| 2591 | #define GET_NEW_LINE \ |
| 2592 | { \ |
| 2593 | linecharno = charno; lineno++; \ |
| 2594 | charno += 1 + readline (&lb, inf); \ |
| 2595 | dbp = lb.buffer; \ |
| 2596 | } |
| 2597 | |
| 2598 | /* |
| 2599 | * Locates tags for procedures & functions. Doesn't do any type- or |
| 2600 | * var-definitions. It does look for the keyword "extern" or |
| 2601 | * "forward" immediately following the procedure statement; if found, |
| 2602 | * the tag is skipped. |
| 2603 | */ |
| 2604 | void |
| 2605 | Pascal_functions (inf) |
| 2606 | FILE *inf; |
| 2607 | { |
| 2608 | struct linebuffer tline; /* mostly copied from C_entries */ |
| 2609 | long save_lcno; |
| 2610 | int save_lineno; |
| 2611 | char c, *cp; |
| 2612 | char *nambuf; |
| 2613 | |
| 2614 | logical /* each of these flags is TRUE iff: */ |
| 2615 | incomment, /* point is inside a comment */ |
| 2616 | inquote, /* point is inside '..' string */ |
| 2617 | get_tagname, /* point is after PROCEDURE/FUNCTION */ |
| 2618 | /* keyword, so next item = potential tag */ |
| 2619 | found_tag, /* point is after a potential tag */ |
| 2620 | inparms, /* point is within parameter-list */ |
| 2621 | verify_tag; /* point has passed the parm-list, so the */ |
| 2622 | /* next token will determine whether */ |
| 2623 | /* this is a FORWARD/EXTERN to be */ |
| 2624 | /* ignored, or whether it is a real tag */ |
| 2625 | |
| 2626 | lineno = 0; |
| 2627 | charno = 0; |
| 2628 | dbp = lb.buffer; |
| 2629 | *dbp = 0; |
| 2630 | initbuffer (&tline); |
| 2631 | |
| 2632 | incomment = inquote = FALSE; |
| 2633 | found_tag = FALSE; /* have a proc name; check if extern */ |
| 2634 | get_tagname = FALSE; /* have found "procedure" keyword */ |
| 2635 | inparms = FALSE; /* found '(' after "proc" */ |
| 2636 | verify_tag = FALSE; /* check if "extern" is ahead */ |
| 2637 | |
| 2638 | /* long main loop to get next char */ |
| 2639 | while (!feof (inf)) |
| 2640 | { |
| 2641 | c = *dbp++; |
| 2642 | if (c == '\0') /* if end of line */ |
| 2643 | { |
| 2644 | GET_NEW_LINE; |
| 2645 | if (*dbp == '\0') |
| 2646 | continue; |
| 2647 | if (!((found_tag && verify_tag) || |
| 2648 | get_tagname)) |
| 2649 | c = *dbp++; /* only if don't need *dbp pointing */ |
| 2650 | /* to the beginning of the name of */ |
| 2651 | /* the procedure or function */ |
| 2652 | } |
| 2653 | if (incomment) |
| 2654 | { |
| 2655 | if (c == '}') /* within { - } comments */ |
| 2656 | incomment = FALSE; |
| 2657 | else if (c == '*' && dbp[1] == ')') /* within (* - *) comments */ |
| 2658 | { |
| 2659 | dbp++; |
| 2660 | incomment = FALSE; |
| 2661 | } |
| 2662 | continue; |
| 2663 | } |
| 2664 | else if (inquote) |
| 2665 | { |
| 2666 | if (c == '\'') |
| 2667 | inquote = FALSE; |
| 2668 | continue; |
| 2669 | } |
| 2670 | else |
| 2671 | switch (c) |
| 2672 | { |
| 2673 | case '\'': |
| 2674 | inquote = TRUE; /* found first quote */ |
| 2675 | continue; |
| 2676 | case '{': /* found open-{-comment */ |
| 2677 | incomment = TRUE; |
| 2678 | continue; |
| 2679 | case '(': |
| 2680 | if (*dbp == '*') /* found open-(*-comment */ |
| 2681 | { |
| 2682 | incomment = TRUE; |
| 2683 | dbp++; |
| 2684 | } |
| 2685 | else if (found_tag) /* found '(' after tag, i.e., parm-list */ |
| 2686 | inparms = TRUE; |
| 2687 | continue; |
| 2688 | case ')': /* end of parms list */ |
| 2689 | if (inparms) |
| 2690 | inparms = FALSE; |
| 2691 | continue; |
| 2692 | case ';': |
| 2693 | if ((found_tag) && (!inparms)) /* end of proc or fn stmt */ |
| 2694 | { |
| 2695 | verify_tag = TRUE; |
| 2696 | break; |
| 2697 | } |
| 2698 | continue; |
| 2699 | } |
| 2700 | if ((found_tag) && (verify_tag) && (*dbp != ' ')) |
| 2701 | { |
| 2702 | /* check if this is an "extern" declaration */ |
| 2703 | if (*dbp == 0) |
| 2704 | continue; |
| 2705 | if ((*dbp == 'e') || (*dbp == 'E')) |
| 2706 | { |
| 2707 | if (tail ("extern")) /* superfluous, really! */ |
| 2708 | { |
| 2709 | found_tag = FALSE; |
| 2710 | verify_tag = FALSE; |
| 2711 | } |
| 2712 | } |
| 2713 | else if ((*dbp == 'f') || (*dbp == 'F')) |
| 2714 | { |
| 2715 | if (tail ("forward")) /* check for forward reference */ |
| 2716 | { |
| 2717 | found_tag = FALSE; |
| 2718 | verify_tag = FALSE; |
| 2719 | } |
| 2720 | } |
| 2721 | if ((found_tag) && (verify_tag)) /* not external proc, so make tag */ |
| 2722 | { |
| 2723 | found_tag = FALSE; |
| 2724 | verify_tag = FALSE; |
| 2725 | pfnote (nambuf, TRUE, FALSE, tline.buffer, |
| 2726 | cp - tline.buffer + 1, save_lineno, save_lcno); |
| 2727 | continue; |
| 2728 | } |
| 2729 | } |
| 2730 | if (get_tagname) /* grab name of proc or fn */ |
| 2731 | { |
| 2732 | if (*dbp == 0) |
| 2733 | continue; |
| 2734 | |
| 2735 | /* save all values for later tagging */ |
| 2736 | tline.size = lb.size; |
| 2737 | strcpy (tline.buffer, lb.buffer); |
| 2738 | save_lineno = lineno; |
| 2739 | save_lcno = linecharno; |
| 2740 | |
| 2741 | /* grab block name */ |
| 2742 | for (cp = dbp + 1; *cp && (!endtoken (*cp)); cp++) |
| 2743 | continue; |
| 2744 | nambuf = savenstr (dbp, cp-dbp); |
| 2745 | dbp = cp; /* restore dbp to e-o-token */ |
| 2746 | get_tagname = FALSE; |
| 2747 | found_tag = TRUE; |
| 2748 | continue; |
| 2749 | |
| 2750 | /* and proceed to check for "extern" */ |
| 2751 | } |
| 2752 | else if (!incomment && !inquote && !found_tag) |
| 2753 | { |
| 2754 | /* check for proc/fn keywords */ |
| 2755 | switch (c | ' ') |
| 2756 | { |
| 2757 | case 'p': |
| 2758 | if (tail ("rocedure")) /* c = 'p', dbp has advanced */ |
| 2759 | get_tagname = TRUE; |
| 2760 | continue; |
| 2761 | case 'f': |
| 2762 | if (tail ("unction")) |
| 2763 | get_tagname = TRUE; |
| 2764 | continue; |
| 2765 | } |
| 2766 | } |
| 2767 | } /* while not eof */ |
| 2768 | } |
| 2769 | \f |
| 2770 | /* |
| 2771 | * lisp tag functions |
| 2772 | * look for (def or (DEF, quote or QUOTE |
| 2773 | */ |
| 2774 | int |
| 2775 | L_isdef (strp) |
| 2776 | register char *strp; |
| 2777 | { |
| 2778 | return ((strp[1] == 'd' || strp[1] == 'D') |
| 2779 | && (strp[2] == 'e' || strp[2] == 'E') |
| 2780 | && (strp[3] == 'f' || strp[3] == 'F')); |
| 2781 | } |
| 2782 | |
| 2783 | int |
| 2784 | L_isquote (strp) |
| 2785 | register char *strp; |
| 2786 | { |
| 2787 | return ((*(++strp) == 'q' || *strp == 'Q') |
| 2788 | && (*(++strp) == 'u' || *strp == 'U') |
| 2789 | && (*(++strp) == 'o' || *strp == 'O') |
| 2790 | && (*(++strp) == 't' || *strp == 'T') |
| 2791 | && (*(++strp) == 'e' || *strp == 'E') |
| 2792 | && isspace(*(++strp))); |
| 2793 | } |
| 2794 | |
| 2795 | void |
| 2796 | L_getit () |
| 2797 | { |
| 2798 | register char *cp; |
| 2799 | |
| 2800 | if (*dbp == '\'') /* Skip prefix quote */ |
| 2801 | dbp++; |
| 2802 | else if (*dbp == '(' && L_isquote (dbp)) /* Skip "(quote " */ |
| 2803 | { |
| 2804 | dbp += 7; |
| 2805 | while (isspace(*dbp)) |
| 2806 | dbp++; |
| 2807 | } |
| 2808 | for (cp = dbp /*+1*/; |
| 2809 | *cp && *cp != '(' && *cp != ' ' && *cp != ')'; |
| 2810 | cp++) |
| 2811 | continue; |
| 2812 | if (cp == dbp) |
| 2813 | return; |
| 2814 | |
| 2815 | pfnote (savenstr (dbp, cp-dbp), TRUE, FALSE, lb.buffer, |
| 2816 | cp - lb.buffer + 1, lineno, linecharno); |
| 2817 | } |
| 2818 | |
| 2819 | void |
| 2820 | Lisp_functions (inf) |
| 2821 | FILE *inf; |
| 2822 | { |
| 2823 | lineno = 0; |
| 2824 | charno = 0; |
| 2825 | |
| 2826 | while (!feof (inf)) |
| 2827 | { |
| 2828 | lineno++; |
| 2829 | linecharno = charno; |
| 2830 | charno += readline (&lb, inf); |
| 2831 | dbp = lb.buffer; |
| 2832 | if (dbp[0] == '(') |
| 2833 | { |
| 2834 | if (L_isdef (dbp)) |
| 2835 | { |
| 2836 | while (!isspace (*dbp)) |
| 2837 | dbp++; |
| 2838 | while (isspace (*dbp)) |
| 2839 | dbp++; |
| 2840 | L_getit (); |
| 2841 | } |
| 2842 | else |
| 2843 | { |
| 2844 | /* Check for (foo::defmumble name-defined ... */ |
| 2845 | do |
| 2846 | dbp++; |
| 2847 | while (*dbp && !isspace (*dbp) |
| 2848 | && *dbp != ':' && *dbp != '(' && *dbp != ')'); |
| 2849 | if (*dbp == ':') |
| 2850 | { |
| 2851 | do |
| 2852 | dbp++; |
| 2853 | while (*dbp == ':'); |
| 2854 | |
| 2855 | if (L_isdef (dbp - 1)) |
| 2856 | { |
| 2857 | while (!isspace (*dbp)) |
| 2858 | dbp++; |
| 2859 | while (isspace (*dbp)) |
| 2860 | dbp++; |
| 2861 | L_getit (); |
| 2862 | } |
| 2863 | } |
| 2864 | } |
| 2865 | } |
| 2866 | } |
| 2867 | } |
| 2868 | \f |
| 2869 | /* |
| 2870 | * Scheme tag functions |
| 2871 | * look for (def... xyzzy |
| 2872 | * look for (def... (xyzzy |
| 2873 | * look for (def ... ((...(xyzzy .... |
| 2874 | * look for (set! xyzzy |
| 2875 | */ |
| 2876 | |
| 2877 | void get_scheme (); |
| 2878 | |
| 2879 | void |
| 2880 | Scheme_functions (inf) |
| 2881 | FILE *inf; |
| 2882 | { |
| 2883 | lineno = 0; |
| 2884 | charno = 0; |
| 2885 | |
| 2886 | while (!feof (inf)) |
| 2887 | { |
| 2888 | lineno++; |
| 2889 | linecharno = charno; |
| 2890 | charno += readline (&lb, inf); |
| 2891 | dbp = lb.buffer; |
| 2892 | if (dbp[0] == '(' && |
| 2893 | (dbp[1] == 'D' || dbp[1] == 'd') && |
| 2894 | (dbp[2] == 'E' || dbp[2] == 'e') && |
| 2895 | (dbp[3] == 'F' || dbp[3] == 'f')) |
| 2896 | { |
| 2897 | while (!isspace (*dbp)) |
| 2898 | dbp++; |
| 2899 | /* Skip over open parens and white space */ |
| 2900 | while (*dbp && (isspace (*dbp) || *dbp == '(')) |
| 2901 | dbp++; |
| 2902 | get_scheme (); |
| 2903 | } |
| 2904 | if (dbp[0] == '(' && |
| 2905 | (dbp[1] == 'S' || dbp[1] == 's') && |
| 2906 | (dbp[2] == 'E' || dbp[2] == 'e') && |
| 2907 | (dbp[3] == 'T' || dbp[3] == 't') && |
| 2908 | (dbp[4] == '!' || dbp[4] == '!') && |
| 2909 | (isspace (dbp[5]))) |
| 2910 | { |
| 2911 | while (!isspace (*dbp)) |
| 2912 | dbp++; |
| 2913 | /* Skip over white space */ |
| 2914 | while (isspace (*dbp)) |
| 2915 | dbp++; |
| 2916 | get_scheme (); |
| 2917 | } |
| 2918 | } |
| 2919 | } |
| 2920 | |
| 2921 | void |
| 2922 | get_scheme () |
| 2923 | { |
| 2924 | register char *cp; |
| 2925 | |
| 2926 | if (*dbp == 0) |
| 2927 | return; |
| 2928 | /* Go till you get to white space or a syntactic break */ |
| 2929 | for (cp = dbp + 1; |
| 2930 | *cp && *cp != '(' && *cp != ')' && !isspace (*cp); |
| 2931 | cp++) |
| 2932 | continue; |
| 2933 | pfnote (savenstr (dbp, cp-dbp), TRUE, FALSE, |
| 2934 | lb.buffer, cp - lb.buffer + 1, lineno, linecharno); |
| 2935 | } |
| 2936 | \f |
| 2937 | /* Find tags in TeX and LaTeX input files. */ |
| 2938 | |
| 2939 | /* TEX_toktab is a table of TeX control sequences that define tags. |
| 2940 | Each TEX_tabent records one such control sequence. |
| 2941 | CONVERT THIS TO USE THE Stab TYPE!! */ |
| 2942 | struct TEX_tabent |
| 2943 | { |
| 2944 | char *name; |
| 2945 | int len; |
| 2946 | }; |
| 2947 | |
| 2948 | struct TEX_tabent *TEX_toktab = NULL; /* Table with tag tokens */ |
| 2949 | |
| 2950 | /* Default set of control sequences to put into TEX_toktab. |
| 2951 | The value of environment var TEXTAGS is prepended to this. */ |
| 2952 | |
| 2953 | char *TEX_defenv = "\ |
| 2954 | :chapter:section:subsection:subsubsection:eqno:label:ref:cite:bibitem:typeout"; |
| 2955 | |
| 2956 | void TEX_mode (); |
| 2957 | struct TEX_tabent *TEX_decode_env (); |
| 2958 | void TEX_getit (); |
| 2959 | int TEX_Token (); |
| 2960 | |
| 2961 | char TEX_esc = '\\'; |
| 2962 | char TEX_opgrp = '{'; |
| 2963 | char TEX_clgrp = '}'; |
| 2964 | |
| 2965 | /* |
| 2966 | * TeX/LaTeX scanning loop. |
| 2967 | */ |
| 2968 | void |
| 2969 | TeX_functions (inf) |
| 2970 | FILE *inf; |
| 2971 | { |
| 2972 | char *lasthit; |
| 2973 | |
| 2974 | lineno = 0; |
| 2975 | charno = 0; |
| 2976 | |
| 2977 | /* Select either \ or ! as escape character. */ |
| 2978 | TEX_mode (inf); |
| 2979 | |
| 2980 | /* Initialize token table once from environment. */ |
| 2981 | if (!TEX_toktab) |
| 2982 | TEX_toktab = TEX_decode_env ("TEXTAGS", TEX_defenv); |
| 2983 | |
| 2984 | while (!feof (inf)) |
| 2985 | { /* Scan each line in file */ |
| 2986 | lineno++; |
| 2987 | linecharno = charno; |
| 2988 | charno += readline (&lb, inf); |
| 2989 | dbp = lb.buffer; |
| 2990 | lasthit = dbp; |
| 2991 | while (dbp = etags_strchr (dbp, TEX_esc)) /* Look at each esc in line */ |
| 2992 | { |
| 2993 | register int i; |
| 2994 | |
| 2995 | if (!*(++dbp)) |
| 2996 | break; |
| 2997 | linecharno += dbp - lasthit; |
| 2998 | lasthit = dbp; |
| 2999 | i = TEX_Token (lasthit); |
| 3000 | if (0 <= i) |
| 3001 | { |
| 3002 | TEX_getit (lasthit, TEX_toktab[i].len); |
| 3003 | break; /* We only save a line once */ |
| 3004 | } |
| 3005 | } |
| 3006 | } |
| 3007 | } |
| 3008 | |
| 3009 | #define TEX_LESC '\\' |
| 3010 | #define TEX_SESC '!' |
| 3011 | #define TEX_cmt '%' |
| 3012 | |
| 3013 | /* Figure out whether TeX's escapechar is '\\' or '!' and set grouping |
| 3014 | chars accordingly. */ |
| 3015 | void |
| 3016 | TEX_mode (inf) |
| 3017 | FILE *inf; |
| 3018 | { |
| 3019 | int c; |
| 3020 | |
| 3021 | while ((c = getc (inf)) != EOF) |
| 3022 | { |
| 3023 | /* Skip to next line if we hit the TeX comment char. */ |
| 3024 | if (c == TEX_cmt) |
| 3025 | while (c != '\n') |
| 3026 | c = getc (inf); |
| 3027 | else if (c == TEX_LESC || c == TEX_SESC ) |
| 3028 | break; |
| 3029 | } |
| 3030 | |
| 3031 | if (c == TEX_LESC) |
| 3032 | { |
| 3033 | TEX_esc = TEX_LESC; |
| 3034 | TEX_opgrp = '{'; |
| 3035 | TEX_clgrp = '}'; |
| 3036 | } |
| 3037 | else |
| 3038 | { |
| 3039 | TEX_esc = TEX_SESC; |
| 3040 | TEX_opgrp = '<'; |
| 3041 | TEX_clgrp = '>'; |
| 3042 | } |
| 3043 | rewind (inf); |
| 3044 | } |
| 3045 | |
| 3046 | /* Read environment and prepend it to the default string. |
| 3047 | Build token table. */ |
| 3048 | struct TEX_tabent * |
| 3049 | TEX_decode_env (evarname, defenv) |
| 3050 | char *evarname; |
| 3051 | char *defenv; |
| 3052 | { |
| 3053 | register char *env, *p; |
| 3054 | |
| 3055 | struct TEX_tabent *tab; |
| 3056 | int size, i; |
| 3057 | |
| 3058 | /* Append default string to environment. */ |
| 3059 | env = getenv (evarname); |
| 3060 | if (!env) |
| 3061 | env = defenv; |
| 3062 | else |
| 3063 | env = concat (env, defenv, ""); |
| 3064 | |
| 3065 | /* Allocate a token table */ |
| 3066 | for (size = 1, p = env; p;) |
| 3067 | if ((p = etags_strchr (p, ':')) && *(++p)) |
| 3068 | size++; |
| 3069 | /* Add 1 to leave room for null terminator. */ |
| 3070 | tab = xnew (size + 1, struct TEX_tabent); |
| 3071 | |
| 3072 | /* Unpack environment string into token table. Be careful about */ |
| 3073 | /* zero-length strings (leading ':', "::" and trailing ':') */ |
| 3074 | for (i = 0; *env;) |
| 3075 | { |
| 3076 | p = etags_strchr (env, ':'); |
| 3077 | if (!p) /* End of environment string. */ |
| 3078 | p = env + strlen (env); |
| 3079 | if (p - env > 0) |
| 3080 | { /* Only non-zero strings. */ |
| 3081 | tab[i].name = savenstr (env, p - env); |
| 3082 | tab[i].len = strlen (tab[i].name); |
| 3083 | i++; |
| 3084 | } |
| 3085 | if (*p) |
| 3086 | env = p + 1; |
| 3087 | else |
| 3088 | { |
| 3089 | tab[i].name = NULL; /* Mark end of table. */ |
| 3090 | tab[i].len = 0; |
| 3091 | break; |
| 3092 | } |
| 3093 | } |
| 3094 | return tab; |
| 3095 | } |
| 3096 | |
| 3097 | /* Record a tag defined by a TeX command of length LEN and starting at NAME. |
| 3098 | The name being defined actually starts at (NAME + LEN + 1). |
| 3099 | But we seem to include the TeX command in the tag name. */ |
| 3100 | void |
| 3101 | TEX_getit (name, len) |
| 3102 | char *name; |
| 3103 | int len; |
| 3104 | { |
| 3105 | char *p = name + len; |
| 3106 | |
| 3107 | if (*name == 0) |
| 3108 | return; |
| 3109 | |
| 3110 | /* Let tag name extend to next group close (or end of line) */ |
| 3111 | while (*p && *p != TEX_clgrp) |
| 3112 | p++; |
| 3113 | pfnote (savenstr (name, p-name), TRUE, FALSE, lb.buffer, |
| 3114 | strlen (lb.buffer), lineno, linecharno); |
| 3115 | } |
| 3116 | |
| 3117 | /* If the text at CP matches one of the tag-defining TeX command names, |
| 3118 | return the pointer to the first occurrence of that command in TEX_toktab. |
| 3119 | Otherwise return -1. |
| 3120 | Keep the capital `T' in `Token' for dumb truncating compilers |
| 3121 | (this distinguishes it from `TEX_toktab' */ |
| 3122 | int |
| 3123 | TEX_Token (cp) |
| 3124 | char *cp; |
| 3125 | { |
| 3126 | int i; |
| 3127 | |
| 3128 | for (i = 0; TEX_toktab[i].len > 0; i++) |
| 3129 | if (strneq (TEX_toktab[i].name, cp, TEX_toktab[i].len)) |
| 3130 | return i; |
| 3131 | return -1; |
| 3132 | } |
| 3133 | \f |
| 3134 | /* Support for Prolog. */ |
| 3135 | |
| 3136 | /* Whole head (not only functor, but also arguments) |
| 3137 | is gotten in compound term. */ |
| 3138 | void |
| 3139 | prolog_getit (s) |
| 3140 | char *s; |
| 3141 | { |
| 3142 | char *save_s; |
| 3143 | int insquote, npar; |
| 3144 | |
| 3145 | save_s = s; |
| 3146 | insquote = FALSE; |
| 3147 | npar = 0; |
| 3148 | while (1) |
| 3149 | { |
| 3150 | if (s[0] == '\0') /* syntax error. */ |
| 3151 | return; |
| 3152 | else if (insquote && s[0] == '\'' && s[1] == '\'') |
| 3153 | s += 2; |
| 3154 | else if (s[0] == '\'') |
| 3155 | { |
| 3156 | insquote = !insquote; |
| 3157 | s++; |
| 3158 | } |
| 3159 | else if (!insquote && s[0] == '(') |
| 3160 | { |
| 3161 | npar++; |
| 3162 | s++; |
| 3163 | } |
| 3164 | else if (!insquote && s[0] == ')') |
| 3165 | { |
| 3166 | npar--; |
| 3167 | s++; |
| 3168 | if (npar == 0) |
| 3169 | break; |
| 3170 | else if (npar < 0) /* syntax error. */ |
| 3171 | return; |
| 3172 | } |
| 3173 | else if (!insquote && s[0] == '.' |
| 3174 | && (isspace (s[1]) || s[1] == '\0')) |
| 3175 | { /* fullstop. */ |
| 3176 | if (npar != 0) /* syntax error. */ |
| 3177 | return; |
| 3178 | s++; |
| 3179 | break; |
| 3180 | } |
| 3181 | else |
| 3182 | s++; |
| 3183 | } |
| 3184 | pfnote (savenstr (save_s, s-save_s), TRUE, FALSE, |
| 3185 | save_s, s-save_s, lineno, linecharno); |
| 3186 | } |
| 3187 | |
| 3188 | /* It is assumed that prolog predicate starts from column 0. */ |
| 3189 | void |
| 3190 | Prolog_functions (inf) |
| 3191 | FILE *inf; |
| 3192 | { |
| 3193 | void skip_comment (), prolog_getit (); |
| 3194 | |
| 3195 | lineno = linecharno = charno = 0; |
| 3196 | while (!feof (inf)) |
| 3197 | { |
| 3198 | lineno++; |
| 3199 | linecharno += charno; |
| 3200 | charno = readline (&lb, inf) + 1; /* 1 for newline. */ |
| 3201 | dbp = lb.buffer; |
| 3202 | if (isspace (dbp[0])) /* not predicate header. */ |
| 3203 | continue; |
| 3204 | else if (dbp[0] == '%') /* comment. */ |
| 3205 | continue; |
| 3206 | else if (dbp[0] == '/' && dbp[1] == '*') /* comment. */ |
| 3207 | skip_comment (&lb, inf, &lineno, &linecharno); |
| 3208 | else /* found. */ |
| 3209 | prolog_getit (dbp); |
| 3210 | } |
| 3211 | } |
| 3212 | |
| 3213 | void |
| 3214 | skip_comment (plb, inf, plineno, plinecharno) |
| 3215 | struct linebuffer *plb; |
| 3216 | FILE *inf; |
| 3217 | int *plineno; /* result */ |
| 3218 | long *plinecharno; /* result */ |
| 3219 | { |
| 3220 | char *cp; |
| 3221 | |
| 3222 | do |
| 3223 | { |
| 3224 | for (cp = plb->buffer; *cp != '\0'; cp++) |
| 3225 | if (cp[0] == '*' && cp[1] == '/') |
| 3226 | return; |
| 3227 | (*plineno)++; |
| 3228 | *plinecharno += readline (plb, inf) + 1; /* 1 for newline. */ |
| 3229 | } |
| 3230 | while (!feof(inf)); |
| 3231 | } |
| 3232 | \f |
| 3233 | #ifdef ETAGS_REGEXPS |
| 3234 | /* Take a string like "/blah/" and turn it into "blah", making sure |
| 3235 | that the first and last characters are the same, and handling |
| 3236 | quoted separator characters. Actually, stops on the occurence of |
| 3237 | an unquoted separator. Also turns "\t" into a Tab character. |
| 3238 | Returns pointer to terminating separator. Works in place. Null |
| 3239 | terminates name string. */ |
| 3240 | char * |
| 3241 | scan_separators (name) |
| 3242 | char *name; |
| 3243 | { |
| 3244 | char sep = name[0]; |
| 3245 | char *copyto = name; |
| 3246 | logical quoted = FALSE; |
| 3247 | |
| 3248 | for (++name; *name != '\0'; ++name) |
| 3249 | { |
| 3250 | if (quoted) |
| 3251 | { |
| 3252 | if (*name == 't') |
| 3253 | *copyto++ = '\t'; |
| 3254 | else if (*name == sep) |
| 3255 | *copyto++ = sep; |
| 3256 | else |
| 3257 | { |
| 3258 | /* Something else is quoted, so preserve the quote. */ |
| 3259 | *copyto++ = '\\'; |
| 3260 | *copyto++ = *name; |
| 3261 | } |
| 3262 | quoted = FALSE; |
| 3263 | } |
| 3264 | else if (*name == '\\') |
| 3265 | quoted = TRUE; |
| 3266 | else if (*name == sep) |
| 3267 | break; |
| 3268 | else |
| 3269 | *copyto++ = *name; |
| 3270 | } |
| 3271 | |
| 3272 | /* Terminate copied string. */ |
| 3273 | *copyto = '\0'; |
| 3274 | return name; |
| 3275 | } |
| 3276 | |
| 3277 | /* Turn a name, which is an ed-style (but Emacs syntax) regular |
| 3278 | expression, into a real regular expression by compiling it. */ |
| 3279 | void |
| 3280 | add_regex (regexp_pattern) |
| 3281 | char *regexp_pattern; |
| 3282 | { |
| 3283 | char *name; |
| 3284 | const char *err; |
| 3285 | struct re_pattern_buffer *patbuf; |
| 3286 | |
| 3287 | if (regexp_pattern == NULL) |
| 3288 | { |
| 3289 | /* Remove existing regexps. */ |
| 3290 | num_patterns = 0; |
| 3291 | patterns = NULL; |
| 3292 | return; |
| 3293 | } |
| 3294 | |
| 3295 | if (regexp_pattern[0] == '\0') |
| 3296 | { |
| 3297 | error ("missing regexp", 0); |
| 3298 | return; |
| 3299 | } |
| 3300 | if (regexp_pattern[strlen(regexp_pattern)-1] != regexp_pattern[0]) |
| 3301 | { |
| 3302 | error ("%s: unterminated regexp", regexp_pattern); |
| 3303 | return; |
| 3304 | } |
| 3305 | name = scan_separators (regexp_pattern); |
| 3306 | if (regexp_pattern[0] == '\0') |
| 3307 | { |
| 3308 | error ("null regexp", 0); |
| 3309 | return; |
| 3310 | } |
| 3311 | (void) scan_separators (name); |
| 3312 | |
| 3313 | patbuf = xnew (1, struct re_pattern_buffer); |
| 3314 | patbuf->translate = NULL; |
| 3315 | patbuf->fastmap = NULL; |
| 3316 | patbuf->buffer = NULL; |
| 3317 | patbuf->allocated = 0; |
| 3318 | |
| 3319 | err = re_compile_pattern (regexp_pattern, strlen (regexp_pattern), patbuf); |
| 3320 | if (err != NULL) |
| 3321 | { |
| 3322 | error ("%s while compiling pattern", err); |
| 3323 | return; |
| 3324 | } |
| 3325 | |
| 3326 | num_patterns += 1; |
| 3327 | if (num_patterns == 1) |
| 3328 | patterns = xnew (1, struct pattern); |
| 3329 | else |
| 3330 | patterns = ((struct pattern *) |
| 3331 | xrealloc (patterns, |
| 3332 | (num_patterns * sizeof (struct pattern)))); |
| 3333 | patterns[num_patterns - 1].pattern = patbuf; |
| 3334 | patterns[num_patterns - 1].name_pattern = savestr (name); |
| 3335 | patterns[num_patterns - 1].error_signaled = FALSE; |
| 3336 | } |
| 3337 | |
| 3338 | /* |
| 3339 | * Do the subtitutions indicated by the regular expression and |
| 3340 | * arguments. |
| 3341 | */ |
| 3342 | char * |
| 3343 | substitute (in, out, regs) |
| 3344 | char *in, *out; |
| 3345 | struct re_registers *regs; |
| 3346 | { |
| 3347 | char *result = NULL, *t; |
| 3348 | int size = 0; |
| 3349 | |
| 3350 | /* Pass 1: figure out how much size to allocate. */ |
| 3351 | for (t = out; *t; ++t) |
| 3352 | { |
| 3353 | if (*t == '\\') |
| 3354 | { |
| 3355 | ++t; |
| 3356 | if (!*t) |
| 3357 | { |
| 3358 | fprintf (stderr, "%s: pattern subtitution ends prematurely\n", |
| 3359 | progname); |
| 3360 | return NULL; |
| 3361 | } |
| 3362 | if (isdigit (*t)) |
| 3363 | { |
| 3364 | int dig = *t - '0'; |
| 3365 | size += regs->end[dig] - regs->start[dig]; |
| 3366 | } |
| 3367 | } |
| 3368 | } |
| 3369 | |
| 3370 | /* Allocate space and do the substitutions. */ |
| 3371 | result = xnew (size + 1, char); |
| 3372 | size = 0; |
| 3373 | for (; *out; ++out) |
| 3374 | { |
| 3375 | if (*out == '\\') |
| 3376 | { |
| 3377 | ++out; |
| 3378 | if (isdigit (*out)) |
| 3379 | { |
| 3380 | /* Using "dig2" satisfies my debugger. Bleah. */ |
| 3381 | int dig2 = *out - '0'; |
| 3382 | strncpy (result + size, in + regs->start[dig2], |
| 3383 | regs->end[dig2] - regs->start[dig2]); |
| 3384 | size += regs->end[dig2] - regs->start[dig2]; |
| 3385 | } |
| 3386 | else |
| 3387 | { |
| 3388 | switch (*out) |
| 3389 | { |
| 3390 | case '\t': |
| 3391 | result[size++] = '\t'; |
| 3392 | break; |
| 3393 | case '\\': |
| 3394 | *out = '\\'; |
| 3395 | break; |
| 3396 | default: |
| 3397 | result[size++] = *out; |
| 3398 | break; |
| 3399 | } |
| 3400 | } |
| 3401 | } |
| 3402 | else |
| 3403 | result[size++] = *out; |
| 3404 | } |
| 3405 | result[size] = '\0'; |
| 3406 | |
| 3407 | return result; |
| 3408 | } |
| 3409 | \f |
| 3410 | #endif /* ETAGS_REGEXPS */ |
| 3411 | /* Initialize a linebuffer for use */ |
| 3412 | void |
| 3413 | initbuffer (linebuffer) |
| 3414 | struct linebuffer *linebuffer; |
| 3415 | { |
| 3416 | linebuffer->size = 200; |
| 3417 | linebuffer->buffer = xnew (200, char); |
| 3418 | } |
| 3419 | |
| 3420 | /* |
| 3421 | * Read a line of text from `stream' into `linebuffer'. |
| 3422 | * Return the number of characters read from `stream', |
| 3423 | * which is the length of the line including the newline, if any. |
| 3424 | */ |
| 3425 | long |
| 3426 | readline_internal (linebuffer, stream) |
| 3427 | struct linebuffer *linebuffer; |
| 3428 | register FILE *stream; |
| 3429 | { |
| 3430 | char *buffer = linebuffer->buffer; |
| 3431 | register char *p = linebuffer->buffer; |
| 3432 | register char *pend; |
| 3433 | int chars_deleted; |
| 3434 | |
| 3435 | pend = p + linebuffer->size; /* Separate to avoid 386/IX compiler bug. */ |
| 3436 | |
| 3437 | while (1) |
| 3438 | { |
| 3439 | register int c = getc (stream); |
| 3440 | if (p == pend) |
| 3441 | { |
| 3442 | linebuffer->size *= 2; |
| 3443 | buffer = (char *) xrealloc (buffer, linebuffer->size); |
| 3444 | p += buffer - linebuffer->buffer; |
| 3445 | pend = buffer + linebuffer->size; |
| 3446 | linebuffer->buffer = buffer; |
| 3447 | } |
| 3448 | if (c == EOF) |
| 3449 | { |
| 3450 | chars_deleted = 0; |
| 3451 | break; |
| 3452 | } |
| 3453 | if (c == '\n') |
| 3454 | { |
| 3455 | if (p > buffer && p[-1] == '\r') |
| 3456 | { |
| 3457 | *--p = '\0'; |
| 3458 | chars_deleted = 2; |
| 3459 | } |
| 3460 | else |
| 3461 | { |
| 3462 | *p = '\0'; |
| 3463 | chars_deleted = 1; |
| 3464 | } |
| 3465 | break; |
| 3466 | } |
| 3467 | *p++ = c; |
| 3468 | } |
| 3469 | |
| 3470 | return p - buffer + chars_deleted; |
| 3471 | } |
| 3472 | |
| 3473 | /* |
| 3474 | * Like readline_internal, above, but try to match the input |
| 3475 | * line against any existing regular expressions. |
| 3476 | */ |
| 3477 | long |
| 3478 | readline (linebuffer, stream) |
| 3479 | struct linebuffer *linebuffer; |
| 3480 | FILE *stream; |
| 3481 | { |
| 3482 | /* Read new line. */ |
| 3483 | int i; |
| 3484 | long result = readline_internal (linebuffer, stream); |
| 3485 | |
| 3486 | #ifdef ETAGS_REGEXPS |
| 3487 | /* Match against all listed patterns. */ |
| 3488 | for (i = 0; i < num_patterns; ++i) |
| 3489 | { |
| 3490 | int match = re_match (patterns[i].pattern, linebuffer->buffer, |
| 3491 | (int)result, 0, &patterns[i].regs); |
| 3492 | switch (match) |
| 3493 | { |
| 3494 | case -2: |
| 3495 | /* Some error. */ |
| 3496 | if (!patterns[i].error_signaled) |
| 3497 | { |
| 3498 | error ("error while matching pattern %d", i); |
| 3499 | patterns[i].error_signaled = TRUE; |
| 3500 | } |
| 3501 | break; |
| 3502 | case -1: |
| 3503 | /* No match. */ |
| 3504 | break; |
| 3505 | default: |
| 3506 | /* Match occurred. Construct a tag. */ |
| 3507 | if (patterns[i].name_pattern[0] != '\0') |
| 3508 | { |
| 3509 | /* Make a named tag. */ |
| 3510 | char *name = substitute (linebuffer->buffer, |
| 3511 | patterns[i].name_pattern, |
| 3512 | &patterns[i].regs); |
| 3513 | if (name != NULL) |
| 3514 | pfnote (name, TRUE, TRUE, linebuffer->buffer, |
| 3515 | match, lineno, linecharno); |
| 3516 | } |
| 3517 | else |
| 3518 | { |
| 3519 | /* Make an unnamed tag. */ |
| 3520 | pfnote (NULL, TRUE, FALSE, linebuffer->buffer, |
| 3521 | match, lineno, linecharno); |
| 3522 | } |
| 3523 | break; |
| 3524 | } |
| 3525 | } |
| 3526 | #endif /* ETAGS_REGEXPS */ |
| 3527 | |
| 3528 | return result; |
| 3529 | } |
| 3530 | |
| 3531 | /* |
| 3532 | * Read a file, but do no processing. This is used to do regexp |
| 3533 | * matching on files that have no language defined. |
| 3534 | */ |
| 3535 | void |
| 3536 | just_read_file (inf) |
| 3537 | FILE *inf; |
| 3538 | { |
| 3539 | while (!feof (inf)) |
| 3540 | { |
| 3541 | ++lineno; |
| 3542 | linecharno = charno; |
| 3543 | charno += readline (&lb, inf) + 1; |
| 3544 | } |
| 3545 | } |
| 3546 | |
| 3547 | \f |
| 3548 | /* |
| 3549 | * Return a pointer to a space of size strlen(cp)+1 allocated |
| 3550 | * with xnew where the string CP has been copied. |
| 3551 | */ |
| 3552 | char * |
| 3553 | savestr (cp) |
| 3554 | char *cp; |
| 3555 | { |
| 3556 | return savenstr (cp, strlen (cp)); |
| 3557 | } |
| 3558 | |
| 3559 | /* |
| 3560 | * Return a pointer to a space of size LEN+1 allocated with xnew where |
| 3561 | * the string CP has been copied for at most the first LEN characters. |
| 3562 | */ |
| 3563 | char * |
| 3564 | savenstr (cp, len) |
| 3565 | char *cp; |
| 3566 | int len; |
| 3567 | { |
| 3568 | register char *dp; |
| 3569 | |
| 3570 | dp = xnew (len + 1, char); |
| 3571 | strncpy (dp, cp, len); |
| 3572 | dp[len] = '\0'; |
| 3573 | return dp; |
| 3574 | } |
| 3575 | |
| 3576 | /* |
| 3577 | * Return the ptr in sp at which the character c last |
| 3578 | * appears; NULL if not found |
| 3579 | * |
| 3580 | * Identical to System V strrchr, included for portability. |
| 3581 | */ |
| 3582 | char * |
| 3583 | etags_strrchr (sp, c) |
| 3584 | register char *sp, c; |
| 3585 | { |
| 3586 | register char *r; |
| 3587 | |
| 3588 | r = NULL; |
| 3589 | do |
| 3590 | { |
| 3591 | if (*sp == c) |
| 3592 | r = sp; |
| 3593 | } while (*sp++); |
| 3594 | return r; |
| 3595 | } |
| 3596 | |
| 3597 | |
| 3598 | /* |
| 3599 | * Return the ptr in sp at which the character c first |
| 3600 | * appears; NULL if not found |
| 3601 | * |
| 3602 | * Identical to System V strchr, included for portability. |
| 3603 | */ |
| 3604 | char * |
| 3605 | etags_strchr (sp, c) |
| 3606 | register char *sp, c; |
| 3607 | { |
| 3608 | do |
| 3609 | { |
| 3610 | if (*sp == c) |
| 3611 | return sp; |
| 3612 | } while (*sp++); |
| 3613 | return NULL; |
| 3614 | } |
| 3615 | |
| 3616 | /* Print error message and exit. */ |
| 3617 | void |
| 3618 | fatal (s1, s2) |
| 3619 | char *s1, *s2; |
| 3620 | { |
| 3621 | error (s1, s2); |
| 3622 | exit (BAD); |
| 3623 | } |
| 3624 | |
| 3625 | void |
| 3626 | pfatal (s1) |
| 3627 | char *s1; |
| 3628 | { |
| 3629 | perror (s1); |
| 3630 | exit (BAD); |
| 3631 | } |
| 3632 | |
| 3633 | /* Print error message. `s1' is printf control string, `s2' is arg for it. */ |
| 3634 | void |
| 3635 | error (s1, s2) |
| 3636 | char *s1, *s2; |
| 3637 | { |
| 3638 | fprintf (stderr, "%s: ", progname); |
| 3639 | fprintf (stderr, s1, s2); |
| 3640 | fprintf (stderr, "\n"); |
| 3641 | } |
| 3642 | |
| 3643 | /* Return a newly-allocated string whose contents |
| 3644 | concatenate those of s1, s2, s3. */ |
| 3645 | char * |
| 3646 | concat (s1, s2, s3) |
| 3647 | char *s1, *s2, *s3; |
| 3648 | { |
| 3649 | int len1 = strlen (s1), len2 = strlen (s2), len3 = strlen (s3); |
| 3650 | char *result = xnew (len1 + len2 + len3 + 1, char); |
| 3651 | |
| 3652 | strcpy (result, s1); |
| 3653 | strcpy (result + len1, s2); |
| 3654 | strcpy (result + len1 + len2, s3); |
| 3655 | result[len1 + len2 + len3] = '\0'; |
| 3656 | |
| 3657 | return result; |
| 3658 | } |
| 3659 | \f |
| 3660 | /* Does the same work as the system V getcwd, but does not need to |
| 3661 | guess buffer size in advance. */ |
| 3662 | char * |
| 3663 | etags_getcwd () |
| 3664 | { |
| 3665 | #ifdef DOS_NT |
| 3666 | char *p, path[MAXPATHLEN + 1]; /* Fixed size is safe on MSDOS. */ |
| 3667 | |
| 3668 | getwd (path); |
| 3669 | p = path; |
| 3670 | while (*p) |
| 3671 | if (*p == '\\') |
| 3672 | *p++ = '/'; |
| 3673 | else |
| 3674 | *p++ = tolower (*p); |
| 3675 | |
| 3676 | return strdup (path); |
| 3677 | #else /* not DOS_NT */ |
| 3678 | #if HAVE_GETCWD |
| 3679 | int bufsize = 200; |
| 3680 | char *path = xnew (bufsize, char); |
| 3681 | |
| 3682 | while (getcwd (path, bufsize) == NULL) |
| 3683 | { |
| 3684 | if (errno != ERANGE) |
| 3685 | pfatal ("getcwd"); |
| 3686 | bufsize *= 2; |
| 3687 | path = xnew (bufsize, char); |
| 3688 | } |
| 3689 | |
| 3690 | return path; |
| 3691 | #else /* not DOS_NT and not HAVE_GETCWD */ |
| 3692 | struct linebuffer path; |
| 3693 | FILE *pipe; |
| 3694 | |
| 3695 | initbuffer (&path); |
| 3696 | pipe = (FILE *) popen ("/bin/pwd 2>/dev/null", "r"); |
| 3697 | if (pipe == NULL || readline_internal (&path, pipe) == 0) |
| 3698 | pfatal ("/bin/pwd"); |
| 3699 | pclose (pipe); |
| 3700 | |
| 3701 | return path.buffer; |
| 3702 | #endif /* not HAVE_GETCWD */ |
| 3703 | #endif /* not DOS_NT */ |
| 3704 | } |
| 3705 | |
| 3706 | /* Return a newly allocated string containing the filename |
| 3707 | of FILE relative to the absolute directory DIR (which |
| 3708 | should end with a slash). */ |
| 3709 | char * |
| 3710 | relative_filename (file, dir) |
| 3711 | char *file, *dir; |
| 3712 | { |
| 3713 | char *fp, *dp, *res; |
| 3714 | |
| 3715 | /* Find the common root of file and dir. */ |
| 3716 | fp = absolute_filename (file, cwd); |
| 3717 | dp = dir; |
| 3718 | while (*fp++ == *dp++) |
| 3719 | continue; |
| 3720 | do |
| 3721 | { |
| 3722 | fp--; |
| 3723 | dp--; |
| 3724 | } |
| 3725 | while (*fp != '/'); |
| 3726 | |
| 3727 | /* Build a sequence of "../" strings for the resulting relative filename. */ |
| 3728 | for (dp = etags_strchr (dp + 1, '/'), res = ""; |
| 3729 | dp != NULL; |
| 3730 | dp = etags_strchr (dp + 1, '/')) |
| 3731 | { |
| 3732 | res = concat (res, "../", ""); |
| 3733 | } |
| 3734 | |
| 3735 | /* Add the filename relative to the common root of file and dir. */ |
| 3736 | res = concat (res, fp + 1, ""); |
| 3737 | |
| 3738 | return res; /* temporary stub */ |
| 3739 | } |
| 3740 | |
| 3741 | /* Return a newly allocated string containing the |
| 3742 | absolute filename of FILE given CWD (which should |
| 3743 | end with a slash). */ |
| 3744 | char * |
| 3745 | absolute_filename (file, cwd) |
| 3746 | char *file, *cwd; |
| 3747 | { |
| 3748 | char *slashp, *cp, *res; |
| 3749 | |
| 3750 | if (file[0] == '/') |
| 3751 | res = concat (file, "", ""); |
| 3752 | else |
| 3753 | res = concat (cwd, file, ""); |
| 3754 | |
| 3755 | /* Delete the "/dirname/.." and "/." substrings. */ |
| 3756 | slashp = etags_strchr (res, '/'); |
| 3757 | while (slashp != NULL && slashp[0] != '\0') |
| 3758 | { |
| 3759 | if (slashp[1] == '.') |
| 3760 | { |
| 3761 | if (slashp[2] == '.' |
| 3762 | && (slashp[3] == '/' || slashp[3] == '\0')) |
| 3763 | { |
| 3764 | cp = slashp; |
| 3765 | do |
| 3766 | cp--; |
| 3767 | while (cp >= res && *cp != '/'); |
| 3768 | if (*cp == '/') |
| 3769 | { |
| 3770 | strcpy (cp, slashp + 3); |
| 3771 | } |
| 3772 | else /* else (cp == res) */ |
| 3773 | { |
| 3774 | if (slashp[3] != '\0') |
| 3775 | strcpy (cp, slashp + 4); |
| 3776 | else |
| 3777 | return "."; |
| 3778 | } |
| 3779 | slashp = cp; |
| 3780 | continue; |
| 3781 | } |
| 3782 | else if (slashp[2] == '/' || slashp[2] == '\0') |
| 3783 | { |
| 3784 | strcpy (slashp, slashp + 2); |
| 3785 | continue; |
| 3786 | } |
| 3787 | } |
| 3788 | |
| 3789 | slashp = etags_strchr (slashp + 1, '/'); |
| 3790 | } |
| 3791 | |
| 3792 | return res; |
| 3793 | } |
| 3794 | |
| 3795 | /* Return a newly allocated string containing the absolute |
| 3796 | filename of dir where FILE resides given CWD (which should |
| 3797 | end with a slash). */ |
| 3798 | char * |
| 3799 | absolute_dirname (file, cwd) |
| 3800 | char *file, *cwd; |
| 3801 | { |
| 3802 | char *slashp, *res; |
| 3803 | char save; |
| 3804 | |
| 3805 | slashp = etags_strrchr (file, '/'); |
| 3806 | if (slashp == NULL) |
| 3807 | return cwd; |
| 3808 | save = slashp[1]; |
| 3809 | slashp[1] = '\0'; |
| 3810 | res = absolute_filename (file, cwd); |
| 3811 | slashp[1] = save; |
| 3812 | |
| 3813 | return res; |
| 3814 | } |
| 3815 | |
| 3816 | /* Like malloc but get fatal error if memory is exhausted. */ |
| 3817 | long * |
| 3818 | xmalloc (size) |
| 3819 | unsigned int size; |
| 3820 | { |
| 3821 | long *result = (long *) malloc (size); |
| 3822 | if (result == NULL) |
| 3823 | fatal ("virtual memory exhausted", 0); |
| 3824 | return result; |
| 3825 | } |
| 3826 | |
| 3827 | long * |
| 3828 | xrealloc (ptr, size) |
| 3829 | char *ptr; |
| 3830 | unsigned int size; |
| 3831 | { |
| 3832 | long *result = (long *) realloc (ptr, size); |
| 3833 | if (result == NULL) |
| 3834 | fatal ("virtual memory exhausted"); |
| 3835 | return result; |
| 3836 | } |