| 1 | /* Tags file maker to go with GNU Emacs |
| 2 | Copyright (C) 1984,87,88,89,93,94 Free Software Foundation, Inc. and Ken Arnold |
| 3 | This file is not considered part of GNU Emacs. |
| 4 | |
| 5 | This program is free software; you can redistribute it and/or modify |
| 6 | it under the terms of the GNU General Public License as published by |
| 7 | the Free Software Foundation; either version 2 of the License, or |
| 8 | (at your option) any later version. |
| 9 | |
| 10 | This program is distributed in the hope that it will be useful, |
| 11 | but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 12 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 13 | GNU General Public License for more details. |
| 14 | |
| 15 | You should have received a copy of the GNU General Public License |
| 16 | along with this program; if not, write to the Free Software |
| 17 | Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */ |
| 18 | |
| 19 | /* |
| 20 | * Authors: |
| 21 | * Ctags originally by Ken Arnold. |
| 22 | * Fortran added by Jim Kleckner. |
| 23 | * Ed Pelegri-Llopart added C typedefs. |
| 24 | * Gnu Emacs TAGS format and modifications by RMS? |
| 25 | * Sam Kendall added C++. |
| 26 | * Francesco Potorti` reorganised C and C++ based on work by Joe Wells. |
| 27 | #ifdef ETAGS_REGEXPS |
| 28 | * Regexp tags by Tom Tromey. |
| 29 | #endif |
| 30 | * |
| 31 | * Francesco Potorti` (pot@cnuce.cnr.it) is the current maintainer. |
| 32 | */ |
| 33 | |
| 34 | char pot_etags_version[] = "@(#) pot revision number is 11.12"; |
| 35 | |
| 36 | #ifdef MSDOS |
| 37 | #include <fcntl.h> |
| 38 | #include <sys/param.h> |
| 39 | #endif /* MSDOS */ |
| 40 | |
| 41 | #ifdef WINDOWSNT |
| 42 | #include <stdlib.h> |
| 43 | #include <fcntl.h> |
| 44 | #include <string.h> |
| 45 | #define MAXPATHLEN _MAX_PATH |
| 46 | #endif |
| 47 | |
| 48 | #ifdef HAVE_CONFIG_H |
| 49 | #include <config.h> |
| 50 | /* On some systems, Emacs defines static as nothing for the sake |
| 51 | of unexec. We don't want that here since we don't use unexec. */ |
| 52 | #undef static |
| 53 | #endif |
| 54 | |
| 55 | #include <stdio.h> |
| 56 | #include <ctype.h> |
| 57 | #include <errno.h> |
| 58 | #ifndef errno |
| 59 | extern int errno; |
| 60 | #endif |
| 61 | #include <sys/types.h> |
| 62 | #include <sys/stat.h> |
| 63 | |
| 64 | #if !defined (S_ISREG) && defined (S_IFREG) |
| 65 | # define S_ISREG(m) (((m) & S_IFMT) == S_IFREG) |
| 66 | #endif |
| 67 | |
| 68 | #include <getopt.h> |
| 69 | |
| 70 | #ifdef ETAGS_REGEXPS |
| 71 | #include <regex.h> |
| 72 | #endif /* ETAGS_REGEXPS */ |
| 73 | |
| 74 | extern char *getenv (); |
| 75 | |
| 76 | /* Define CTAGS to make the program "ctags" compatible with the usual one. |
| 77 | Let it undefined to make the program "etags", which makes emacs-style |
| 78 | tag tables and tags typedefs, #defines and struct/union/enum by default. */ |
| 79 | #ifdef CTAGS |
| 80 | # undef CTAGS |
| 81 | # define CTAGS TRUE |
| 82 | #else |
| 83 | # define CTAGS FALSE |
| 84 | #endif |
| 85 | |
| 86 | /* Exit codes for success and failure. */ |
| 87 | #ifdef VMS |
| 88 | #define GOOD 1 |
| 89 | #define BAD 0 |
| 90 | #else |
| 91 | #define GOOD 0 |
| 92 | #define BAD 1 |
| 93 | #endif |
| 94 | |
| 95 | /* |
| 96 | * The FILEPOS abstract type, which represents a position in a file, |
| 97 | * plus the following accessor functions: |
| 98 | * |
| 99 | * long GET_CHARNO (pos) |
| 100 | * returns absolute char number. |
| 101 | * void SET_FILEPOS (pos, fp, charno) |
| 102 | * FILE *fp; long charno; |
| 103 | * sets `pos' from the current file |
| 104 | * position of `fp' and from `charno', |
| 105 | * which must be the absolute character |
| 106 | * number corresponding to the current |
| 107 | * position of `fp'. |
| 108 | * |
| 109 | * The `pos' parameter is an lvalue expression of type FILEPOS. |
| 110 | * Parameters to the accessor functions are evaluated 0 or more times, |
| 111 | * and so must have no side effects. |
| 112 | * |
| 113 | * FILEPOS objects can also be assigned and passed to and from |
| 114 | * functions in the normal C manner. |
| 115 | * |
| 116 | * Implementation notes: the `+ 0' is to enforce rvalue-ness. |
| 117 | */ |
| 118 | |
| 119 | #ifndef DEBUG |
| 120 | #define DEBUG FALSE |
| 121 | |
| 122 | /* real implementation */ |
| 123 | typedef long FILEPOS; |
| 124 | #define GET_CHARNO(pos) ((pos) + 0) |
| 125 | #define SET_FILEPOS(pos,fp,cno) ((void) ((pos) = (cno))) |
| 126 | |
| 127 | #else |
| 128 | #undef DEBUG |
| 129 | #define DEBUG TRUE |
| 130 | |
| 131 | /* debugging implementation */ |
| 132 | typedef struct |
| 133 | { |
| 134 | long charno; |
| 135 | } FILEPOS; |
| 136 | |
| 137 | #define GET_CHARNO(pos) ((pos).charno + 0) |
| 138 | #define SET_FILEPOS(pos,fp,cno) \ |
| 139 | ((void) ((pos).charno = (cno), \ |
| 140 | (cno) != ftell (fp) ? (error ("SET_FILEPOS inconsistency"), 0) \ |
| 141 | : 0)) |
| 142 | #endif |
| 143 | |
| 144 | #define streq(s,t) (strcmp (s, t) == 0) |
| 145 | #define strneq(s,t,n) (strncmp (s, t, n) == 0) |
| 146 | |
| 147 | #define TRUE 1 |
| 148 | #define FALSE 0 |
| 149 | |
| 150 | #define iswhite(arg) (_wht[arg]) /* T if char is white */ |
| 151 | #define begtoken(arg) (_btk[arg]) /* T if char can start token */ |
| 152 | #define intoken(arg) (_itk[arg]) /* T if char can be in token */ |
| 153 | #define endtoken(arg) (_etk[arg]) /* T if char ends tokens */ |
| 154 | |
| 155 | #define max(I1,I2) ((I1) > (I2) ? (I1) : (I2)) |
| 156 | |
| 157 | typedef int logical; |
| 158 | |
| 159 | struct nd_st |
| 160 | { /* sorting structure */ |
| 161 | char *name; /* function or type name */ |
| 162 | char *file; /* file name */ |
| 163 | logical is_func; /* use pattern or line no */ |
| 164 | logical named; /* list name separately */ |
| 165 | logical been_warned; /* set if noticed dup */ |
| 166 | int lno; /* line number tag is on */ |
| 167 | long cno; /* character number line starts on */ |
| 168 | char *pat; /* search pattern */ |
| 169 | struct nd_st *left, *right; /* left and right sons */ |
| 170 | }; |
| 171 | |
| 172 | typedef struct nd_st NODE; |
| 173 | |
| 174 | /* boolean "functions" (see init) */ |
| 175 | logical _wht[0177], _etk[0177], _itk[0177], _btk[0177]; |
| 176 | |
| 177 | char *cwd; /* current working directory */ |
| 178 | char *tagfiledir; /* directory of tagfile */ |
| 179 | |
| 180 | char *concat (); |
| 181 | char *savenstr (), *savestr (); |
| 182 | char *etags_strchr (), *etags_strrchr (); |
| 183 | char *etags_getcwd (); |
| 184 | char *relative_filename (), *absolute_filename (), *absolute_dirname (); |
| 185 | char *xmalloc (), *xrealloc (); |
| 186 | |
| 187 | typedef void Lang_function (); |
| 188 | #if 0 /* many compilers barf on this */ |
| 189 | Lang_function Asm_labels; |
| 190 | Lang_function default_C_entries; |
| 191 | Lang_function C_entries; |
| 192 | Lang_function Cplusplus_entries; |
| 193 | Lang_function Cstar_entries; |
| 194 | Lang_function Fortran_functions; |
| 195 | Lang_function Yacc_entries; |
| 196 | Lang_function Lisp_functions; |
| 197 | Lang_function Pascal_functions; |
| 198 | Lang_function Prolog_functions; |
| 199 | Lang_function Scheme_functions; |
| 200 | Lang_function TeX_functions; |
| 201 | Lang_function just_read_file; |
| 202 | #else /* so let's write it this way */ |
| 203 | void Asm_labels (); |
| 204 | void default_C_entries (); |
| 205 | void C_entries (); |
| 206 | void Cplusplus_entries (); |
| 207 | void Cstar_entries (); |
| 208 | void Fortran_functions (); |
| 209 | void Yacc_entries (); |
| 210 | void Lisp_functions (); |
| 211 | void Pascal_functions (); |
| 212 | void Prolog_functions (); |
| 213 | void Scheme_functions (); |
| 214 | void TeX_functions (); |
| 215 | void just_read_file (); |
| 216 | #endif |
| 217 | |
| 218 | logical get_language (); |
| 219 | int total_size_of_entries (); |
| 220 | long readline (); |
| 221 | long readline_internal (); |
| 222 | #ifdef ETAGS_REGEXPS |
| 223 | void add_regex (); |
| 224 | #endif |
| 225 | void add_node (); |
| 226 | void error (); |
| 227 | void fatal (); |
| 228 | logical find_entries (); |
| 229 | void free_tree (); |
| 230 | void getit (); |
| 231 | void init (); |
| 232 | void initbuffer (); |
| 233 | void pfnote (); |
| 234 | void process_file (); |
| 235 | void put_entries (); |
| 236 | void takeprec (); |
| 237 | |
| 238 | /* |
| 239 | * MACRO |
| 240 | * xnew -- allocate storage |
| 241 | * |
| 242 | * SYNOPSIS |
| 243 | * Type *xnew (int n, Type); |
| 244 | */ |
| 245 | #define xnew(n,Type) ((Type *) xmalloc ((n) * sizeof (Type))) |
| 246 | |
| 247 | /* |
| 248 | * Symbol table types. |
| 249 | */ |
| 250 | enum sym_type |
| 251 | { |
| 252 | st_none, st_C_struct, st_C_enum, st_C_define, st_C_typedef, st_C_typespec |
| 253 | }; |
| 254 | \f |
| 255 | |
| 256 | |
| 257 | typedef int LINENO; |
| 258 | |
| 259 | typedef struct |
| 260 | { |
| 261 | char *p; |
| 262 | int len; |
| 263 | LINENO lineno; |
| 264 | logical named; |
| 265 | } TOKEN; |
| 266 | |
| 267 | /* C extensions. |
| 268 | */ |
| 269 | #define C_PLPL 0x00001 /* C++ */ |
| 270 | #define C_STAR 0x00003 /* C* */ |
| 271 | #define YACC 0x10000 /* yacc file */ |
| 272 | |
| 273 | char searchar = '/'; /* use /.../ searches */ |
| 274 | |
| 275 | LINENO lineno; /* line number of current line */ |
| 276 | long charno; /* current character number */ |
| 277 | |
| 278 | long linecharno; /* charno of start of line; not used by C, but |
| 279 | * by every other language. |
| 280 | */ |
| 281 | |
| 282 | char *curfile, /* current input file name */ |
| 283 | *tagfile, /* output file */ |
| 284 | *white = " \f\t\n\013", /* white chars */ |
| 285 | *endtk = " \t\n\013\"'#()[]{}=-+%*/&|^~!<>;,.:?", /* token ending chars */ |
| 286 | /* token starting chars */ |
| 287 | *begtk = "ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz$~", |
| 288 | /* valid in-token chars */ |
| 289 | *intk = "ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz$0123456789"; |
| 290 | |
| 291 | int append_to_tagfile; /* -a: append to tags */ |
| 292 | /* The following three default to 1 for etags, but to 0 for ctags. */ |
| 293 | int typedefs; /* -t: create tags for typedefs */ |
| 294 | int typedefs_and_cplusplus; /* -T: create tags for typedefs, level */ |
| 295 | /* 0 struct/enum/union decls, and C++ */ |
| 296 | /* member functions. */ |
| 297 | int constantypedefs; /* -d: create tags for C #define and enum */ |
| 298 | /* constants. Enum consts not implemented. */ |
| 299 | /* -D: opposite of -d. Default under ctags. */ |
| 300 | int update; /* -u: update tags */ |
| 301 | int vgrind_style; /* -v: create vgrind style index output */ |
| 302 | int no_warnings; /* -w: suppress warnings */ |
| 303 | int cxref_style; /* -x: create cxref style output */ |
| 304 | int cplusplus; /* .[hc] means C++, not C */ |
| 305 | int noindentypedefs; /* -S: ignore indentation in C */ |
| 306 | |
| 307 | /* Name this program was invoked with. */ |
| 308 | char *progname; |
| 309 | |
| 310 | struct option longopts[] = |
| 311 | { |
| 312 | { "append", no_argument, NULL, 'a' }, |
| 313 | { "backward-search", no_argument, NULL, 'B' }, |
| 314 | { "c++", no_argument, NULL, 'C' }, |
| 315 | { "cxref", no_argument, NULL, 'x' }, |
| 316 | { "defines", no_argument, NULL, 'd' }, |
| 317 | { "help", no_argument, NULL, 'h' }, |
| 318 | { "help", no_argument, NULL, 'H' }, |
| 319 | { "ignore-indentation", no_argument, NULL, 'S' }, |
| 320 | { "include", required_argument, NULL, 'i' }, |
| 321 | { "language", required_argument, NULL, 'l' }, |
| 322 | { "no-defines", no_argument, NULL, 'D' }, |
| 323 | { "no-regex", no_argument, NULL, 'R' }, |
| 324 | { "no-warn", no_argument, NULL, 'w' }, |
| 325 | { "output", required_argument, NULL, 'o' }, |
| 326 | { "regex", required_argument, NULL, 'r' }, |
| 327 | { "typedefs", no_argument, NULL, 't' }, |
| 328 | { "typedefs-and-c++", no_argument, NULL, 'T' }, |
| 329 | { "update", no_argument, NULL, 'u' }, |
| 330 | { "version", no_argument, NULL, 'V' }, |
| 331 | { "vgrind", no_argument, NULL, 'v' }, |
| 332 | { 0 } |
| 333 | }; |
| 334 | |
| 335 | FILE *tagf; /* ioptr for tags file */ |
| 336 | NODE *head; /* the head of the binary tree of tags */ |
| 337 | logical permit_duplicates = TRUE; /* allow duplicate tags */ |
| 338 | |
| 339 | /* |
| 340 | * A `struct linebuffer' is a structure which holds a line of text. |
| 341 | * `readline' reads a line from a stream into a linebuffer and works |
| 342 | * regardless of the length of the line. |
| 343 | */ |
| 344 | struct linebuffer |
| 345 | { |
| 346 | long size; |
| 347 | char *buffer; |
| 348 | }; |
| 349 | |
| 350 | struct linebuffer lb; /* the current line */ |
| 351 | struct linebuffer filename_lb; /* used to read in filenames */ |
| 352 | struct |
| 353 | { |
| 354 | FILEPOS linepos; |
| 355 | struct linebuffer lb; /* used by C_entries instead of lb */ |
| 356 | } lbs[2]; |
| 357 | |
| 358 | #ifdef ETAGS_REGEXPS |
| 359 | /* Structure defining a regular expression. Elements are |
| 360 | the compiled pattern, and the name string. */ |
| 361 | struct pattern |
| 362 | { |
| 363 | struct re_pattern_buffer *pattern; |
| 364 | struct re_registers regs; |
| 365 | char *name_pattern; |
| 366 | logical error_signaled; |
| 367 | }; |
| 368 | |
| 369 | /* Number of regexps found. */ |
| 370 | int num_patterns = 0; |
| 371 | |
| 372 | /* Array of all regexps. */ |
| 373 | struct pattern *patterns = NULL; |
| 374 | #endif /* ETAGS_REGEXPS */ |
| 375 | |
| 376 | /* Language stuff. */ |
| 377 | struct lang_entry |
| 378 | { |
| 379 | char *extension; |
| 380 | Lang_function *function; |
| 381 | }; |
| 382 | |
| 383 | /* Table of language names and corresponding functions. */ |
| 384 | /* It is ok for a given function to be listed under more than one |
| 385 | name. I just didn't. */ |
| 386 | /* "auto" language reverts to default behavior. */ |
| 387 | struct lang_entry lang_names[] = |
| 388 | { |
| 389 | { "asm", Asm_labels }, |
| 390 | { "c", default_C_entries }, |
| 391 | { "c++", Cplusplus_entries }, |
| 392 | { "c*", Cstar_entries }, |
| 393 | { "fortran", Fortran_functions }, |
| 394 | { "lisp", Lisp_functions }, |
| 395 | { "none", just_read_file }, |
| 396 | { "pascal", Pascal_functions }, |
| 397 | { "scheme" , Scheme_functions }, |
| 398 | { "tex", TeX_functions }, |
| 399 | { "auto", NULL }, |
| 400 | { NULL, NULL } |
| 401 | }; |
| 402 | |
| 403 | /* Table of file extensions and corresponding language functions. */ |
| 404 | struct lang_entry lang_extensions[] = |
| 405 | { |
| 406 | /* Assume that ".s" or ".a" is assembly code. -wolfgang. |
| 407 | Or even ".sa". */ |
| 408 | { "a", Asm_labels }, /* Unix assembler */ |
| 409 | { "asm", Asm_labels }, /* Microcontroller assembly */ |
| 410 | { "def", Asm_labels }, /* BSO/Tasking definition includes */ |
| 411 | { "inc", Asm_labels }, /* Microcontroller include files */ |
| 412 | { "ins", Asm_labels }, /* Microcontroller include files */ |
| 413 | { "s", Asm_labels }, |
| 414 | { "sa", Asm_labels }, /* Unix assembler */ |
| 415 | { "src", Asm_labels }, /* BSO/Tasking C compiler output */ |
| 416 | |
| 417 | /* .aux, .bbl, .clo, .cls, .dtx or .tex implies LaTeX source code. */ |
| 418 | { "aux", TeX_functions }, |
| 419 | { "bbl", TeX_functions }, |
| 420 | { "clo", TeX_functions }, |
| 421 | { "cls", TeX_functions }, |
| 422 | { "dtx", TeX_functions }, |
| 423 | { "sty", TeX_functions }, |
| 424 | { "tex", TeX_functions }, |
| 425 | |
| 426 | /* .l or .el or .lisp (or .cl or .clisp or ...) implies lisp source code */ |
| 427 | { "cl", Lisp_functions }, |
| 428 | { "clisp", Lisp_functions }, |
| 429 | { "el", Lisp_functions }, |
| 430 | { "l", Lisp_functions }, |
| 431 | { "lisp", Lisp_functions }, |
| 432 | { "lsp", Lisp_functions }, |
| 433 | |
| 434 | /* .scm or .sm or .scheme implies scheme source code */ |
| 435 | { "SCM", Scheme_functions }, |
| 436 | { "SM", Scheme_functions }, |
| 437 | { "oak", Scheme_functions }, |
| 438 | { "sch", Scheme_functions }, |
| 439 | { "scheme", Scheme_functions }, |
| 440 | { "scm", Scheme_functions }, |
| 441 | { "sm", Scheme_functions }, |
| 442 | { "t", Scheme_functions }, |
| 443 | /* FIXME Can't do the `SCM' or `scm' prefix with a version number */ |
| 444 | |
| 445 | /* Note that ".c" and ".h" can be considered C++, if the --c++ |
| 446 | flag was given. That is why default_C_entries is called here. */ |
| 447 | { "c", default_C_entries }, |
| 448 | { "h", default_C_entries }, |
| 449 | |
| 450 | /* .C or .H or .cpp or .cxx or .hxx or .hh or .cc or .cpp: a C++ file */ |
| 451 | { "C", Cplusplus_entries }, |
| 452 | { "H", Cplusplus_entries }, |
| 453 | { "cc", Cplusplus_entries }, |
| 454 | { "cpp", Cplusplus_entries }, |
| 455 | { "cxx", Cplusplus_entries }, |
| 456 | { "hh", Cplusplus_entries }, |
| 457 | { "hxx", Cplusplus_entries }, |
| 458 | |
| 459 | /* .y: a yacc file */ |
| 460 | { "y", Yacc_entries }, |
| 461 | |
| 462 | /* .cs or .hs: a C* file */ |
| 463 | { "cs", Cstar_entries }, |
| 464 | { "hs", Cstar_entries }, |
| 465 | |
| 466 | /* .f and .for are FORTRAN. */ |
| 467 | { "F", Fortran_functions }, |
| 468 | { "f", Fortran_functions }, |
| 469 | { "for", Fortran_functions }, |
| 470 | |
| 471 | /* .pl implies prolog source code */ |
| 472 | { "pl", Prolog_functions }, |
| 473 | |
| 474 | /* .p or .pas: a Pascal file */ |
| 475 | { "p", Pascal_functions }, |
| 476 | { "pas", Pascal_functions }, |
| 477 | |
| 478 | { NULL, NULL } |
| 479 | }; |
| 480 | |
| 481 | /* Non-NULL if language fixed. */ |
| 482 | Lang_function *lang_func = NULL; |
| 483 | |
| 484 | \f |
| 485 | void |
| 486 | print_language_names () |
| 487 | { |
| 488 | struct lang_entry *name, *ext; |
| 489 | |
| 490 | puts ("\nThese are the currently supported languages, along with the\n\ |
| 491 | default extensions for files:"); |
| 492 | for (name = lang_names; name->extension; ++name) |
| 493 | { |
| 494 | printf ("\t%s\t", name->extension); |
| 495 | for (ext = lang_extensions; ext->extension; ++ext) |
| 496 | if (name->function == ext->function) |
| 497 | printf (" .%s", ext->extension); |
| 498 | puts (""); |
| 499 | } |
| 500 | puts ("Where `auto' means use default language for files based on filename\n\ |
| 501 | extension, and `none' means only do regexp processing on files.\n\ |
| 502 | If no language is specified and no extension is found for some file,\n\ |
| 503 | Fortran is tried first; if no tags are found, C is tried next."); |
| 504 | } |
| 505 | |
| 506 | void |
| 507 | print_version () |
| 508 | { |
| 509 | #ifdef VERSION |
| 510 | printf ("%s for Emacs version %s.\n", (CTAGS) ? "CTAGS" : "ETAGS", VERSION); |
| 511 | #else |
| 512 | printf ("%s for Emacs version 19.\n", (CTAGS) ? "CTAGS" : "ETAGS"); |
| 513 | #endif |
| 514 | |
| 515 | exit (GOOD); |
| 516 | } |
| 517 | |
| 518 | void |
| 519 | print_help () |
| 520 | { |
| 521 | printf ("These are the options accepted by %s. You may use unambiguous\n\ |
| 522 | abbreviations for the long option names. A - as file name means read\n\ |
| 523 | names from stdin.\n\n", progname); |
| 524 | |
| 525 | puts ("-a, --append\n\ |
| 526 | Append tag entries to existing tags file."); |
| 527 | |
| 528 | if (CTAGS) |
| 529 | puts ("-B, --backward-search\n\ |
| 530 | Write the search commands for the tag entries using '?', the\n\ |
| 531 | backward-search command instead of '/', the forward-search command."); |
| 532 | |
| 533 | puts ("-C, --c++\n\ |
| 534 | Treat files whose extension defaults to C language as C++ files."); |
| 535 | |
| 536 | if (CTAGS) |
| 537 | puts ("-d, --defines\n\ |
| 538 | Create tag entries for constant C #defines, too."); |
| 539 | else |
| 540 | puts ("-D, --no-defines\n\ |
| 541 | Don't create tag entries for constant C #defines. This makes\n\ |
| 542 | the tags file smaller."); |
| 543 | |
| 544 | if (!CTAGS) |
| 545 | { |
| 546 | puts ("-i FILE, --include=FILE\n\ |
| 547 | Include a note in tag file indicating that, when searching for\n\ |
| 548 | a tag, one should also consult the tags file FILE after\n\ |
| 549 | checking the current file."); |
| 550 | puts ("-l LANG, --language=LANG\n\ |
| 551 | Force the following files to be considered as written in the\n\ |
| 552 | named language up to the next --language=LANG option."); |
| 553 | } |
| 554 | |
| 555 | #ifdef ETAGS_REGEXPS |
| 556 | puts ("-r /REGEXP/, --regex=/REGEXP/\n\ |
| 557 | Make a tag for each line matching pattern REGEXP in the\n\ |
| 558 | following files. REGEXP is anchored (as if preceded by ^).\n\ |
| 559 | The form /REGEXP/NAME/ creates a named tag. For example Tcl\n\ |
| 560 | named tags can be created with:\n\ |
| 561 | --regex=/proc[ \\t]+\\([^ \\t]+\\)/\\1/."); |
| 562 | puts ("-R, --no-regex\n\ |
| 563 | Don't create tags from regexps for the following files."); |
| 564 | #endif /* ETAGS_REGEXPS */ |
| 565 | puts ("-o FILE, --output=FILE\n\ |
| 566 | Write the tags to FILE."); |
| 567 | puts ("-S, --ignore-indentation\n\ |
| 568 | Don't rely on indentation quite as much as normal. Currently,\n\ |
| 569 | this means not to assume that a closing brace in the first\n\ |
| 570 | column is the final brace of a function or structure\n\ |
| 571 | definition in C and C++."); |
| 572 | |
| 573 | if (CTAGS) |
| 574 | { |
| 575 | puts ("-t, --typedefs\n\ |
| 576 | Generate tag entries for C typedefs."); |
| 577 | puts ("-T, --typedefs-and-c++\n\ |
| 578 | Generate tag entries for C typedefs, C struct/enum/union tags,\n\ |
| 579 | and C++ member functions."); |
| 580 | puts ("-u, --update\n\ |
| 581 | Update the tag entries for the given files, leaving tag\n\ |
| 582 | entries for other files in place. Currently, this is\n\ |
| 583 | implemented by deleting the existing entries for the given\n\ |
| 584 | files and then rewriting the new entries at the end of the\n\ |
| 585 | tags file. It is often faster to simply rebuild the entire\n\ |
| 586 | tag file than to use this."); |
| 587 | puts ("-v, --vgrind\n\ |
| 588 | Generates an index of items intended for human consumption,\n\ |
| 589 | similar to the output of vgrind. The index is sorted, and\n\ |
| 590 | gives the page number of each item."); |
| 591 | puts ("-w, --no-warn\n\ |
| 592 | Suppress warning messages about entries defined in multiple\n\ |
| 593 | files."); |
| 594 | puts ("-x, --cxref\n\ |
| 595 | Like --vgrind, but in the style of cxref, rather than vgrind.\n\ |
| 596 | The output uses line numbers instead of page numbers, but\n\ |
| 597 | beyond that the differences are cosmetic; try both to see\n\ |
| 598 | which you like."); |
| 599 | } |
| 600 | |
| 601 | puts ("-V, --version\n\ |
| 602 | Print the version of the program.\n\ |
| 603 | -h, --help\n\ |
| 604 | Print this help message."); |
| 605 | |
| 606 | print_language_names (); |
| 607 | |
| 608 | exit (GOOD); |
| 609 | } |
| 610 | |
| 611 | \f |
| 612 | enum argument_type |
| 613 | { |
| 614 | at_language, |
| 615 | at_regexp, |
| 616 | at_filename |
| 617 | }; |
| 618 | |
| 619 | /* This structure helps us allow mixing of --lang and filenames. */ |
| 620 | typedef struct |
| 621 | { |
| 622 | enum argument_type arg_type; |
| 623 | char *what; |
| 624 | Lang_function *function; |
| 625 | } ARGUMENT; |
| 626 | |
| 627 | #ifdef VMS /* VMS specific functions */ |
| 628 | |
| 629 | #define EOS '\0' |
| 630 | |
| 631 | /* This is a BUG! ANY arbitrary limit is a BUG! |
| 632 | Won't someone please fix this? */ |
| 633 | #define MAX_FILE_SPEC_LEN 255 |
| 634 | typedef struct { |
| 635 | short curlen; |
| 636 | char body[MAX_FILE_SPEC_LEN + 1]; |
| 637 | } vspec; |
| 638 | |
| 639 | /* |
| 640 | v1.05 nmm 26-Jun-86 fn_exp - expand specification of list of file names |
| 641 | returning in each successive call the next filename matching the input |
| 642 | spec. The function expects that each in_spec passed |
| 643 | to it will be processed to completion; in particular, up to and |
| 644 | including the call following that in which the last matching name |
| 645 | is returned, the function ignores the value of in_spec, and will |
| 646 | only start processing a new spec with the following call. |
| 647 | If an error occurs, on return out_spec contains the value |
| 648 | of in_spec when the error occurred. |
| 649 | |
| 650 | With each successive filename returned in out_spec, the |
| 651 | function's return value is one. When there are no more matching |
| 652 | names the function returns zero. If on the first call no file |
| 653 | matches in_spec, or there is any other error, -1 is returned. |
| 654 | */ |
| 655 | |
| 656 | #include <rmsdef.h> |
| 657 | #include <descrip.h> |
| 658 | #define OUTSIZE MAX_FILE_SPEC_LEN |
| 659 | short |
| 660 | fn_exp (out, in) |
| 661 | vspec *out; |
| 662 | char *in; |
| 663 | { |
| 664 | static long context = 0; |
| 665 | static struct dsc$descriptor_s o; |
| 666 | static struct dsc$descriptor_s i; |
| 667 | static logical pass1 = TRUE; |
| 668 | long status; |
| 669 | short retval; |
| 670 | |
| 671 | if (pass1) |
| 672 | { |
| 673 | pass1 = FALSE; |
| 674 | o.dsc$a_pointer = (char *) out; |
| 675 | o.dsc$w_length = (short)OUTSIZE; |
| 676 | i.dsc$a_pointer = in; |
| 677 | i.dsc$w_length = (short)strlen(in); |
| 678 | i.dsc$b_dtype = DSC$K_DTYPE_T; |
| 679 | i.dsc$b_class = DSC$K_CLASS_S; |
| 680 | o.dsc$b_dtype = DSC$K_DTYPE_VT; |
| 681 | o.dsc$b_class = DSC$K_CLASS_VS; |
| 682 | } |
| 683 | if ((status = lib$find_file(&i, &o, &context, 0, 0)) == RMS$_NORMAL) |
| 684 | { |
| 685 | out->body[out->curlen] = EOS; |
| 686 | return 1; |
| 687 | } |
| 688 | else if (status == RMS$_NMF) |
| 689 | retval = 0; |
| 690 | else |
| 691 | { |
| 692 | strcpy(out->body, in); |
| 693 | retval = -1; |
| 694 | } |
| 695 | lib$find_file_end(&context); |
| 696 | pass1 = TRUE; |
| 697 | return retval; |
| 698 | } |
| 699 | |
| 700 | /* |
| 701 | v1.01 nmm 19-Aug-85 gfnames - return in successive calls the |
| 702 | name of each file specified by the provided arg expanding wildcards. |
| 703 | */ |
| 704 | char * |
| 705 | gfnames (arg, p_error) |
| 706 | char *arg; |
| 707 | logical *p_error; |
| 708 | { |
| 709 | static vspec filename = {MAX_FILE_SPEC_LEN, "\0"}; |
| 710 | |
| 711 | switch (fn_exp (&filename, arg)) |
| 712 | { |
| 713 | case 1: |
| 714 | *p_error = FALSE; |
| 715 | return filename.body; |
| 716 | case 0: |
| 717 | *p_error = FALSE; |
| 718 | return NULL; |
| 719 | default: |
| 720 | *p_error = TRUE; |
| 721 | return filename.body; |
| 722 | } |
| 723 | } |
| 724 | |
| 725 | #ifndef OLD /* Newer versions of VMS do provide `system'. */ |
| 726 | system (cmd) |
| 727 | char *cmd; |
| 728 | { |
| 729 | fprintf (stderr, "system() function not implemented under VMS\n"); |
| 730 | } |
| 731 | #endif |
| 732 | |
| 733 | #define VERSION_DELIM ';' |
| 734 | char *massage_name (s) |
| 735 | char *s; |
| 736 | { |
| 737 | char *start = s; |
| 738 | |
| 739 | for ( ; *s; s++) |
| 740 | if (*s == VERSION_DELIM) |
| 741 | { |
| 742 | *s = EOS; |
| 743 | break; |
| 744 | } |
| 745 | else |
| 746 | *s = tolower(*s); |
| 747 | return start; |
| 748 | } |
| 749 | #endif /* VMS */ |
| 750 | |
| 751 | \f |
| 752 | void |
| 753 | main (argc, argv) |
| 754 | int argc; |
| 755 | char *argv[]; |
| 756 | { |
| 757 | char cmd[100]; |
| 758 | int i; |
| 759 | unsigned int nincluded_files = 0; |
| 760 | char **included_files = xnew (argc, char *); |
| 761 | char *this_file; |
| 762 | ARGUMENT *argbuffer; |
| 763 | int current_arg = 0, file_count = 0; |
| 764 | #ifdef VMS |
| 765 | logical got_err; |
| 766 | #endif |
| 767 | |
| 768 | #ifdef DOS_NT |
| 769 | _fmode = O_BINARY; /* all of files are treated as binary files */ |
| 770 | #endif /* DOS_NT */ |
| 771 | |
| 772 | progname = argv[0]; |
| 773 | |
| 774 | /* Allocate enough no matter what happens. Overkill, but each one |
| 775 | is small. */ |
| 776 | argbuffer = xnew (argc, ARGUMENT); |
| 777 | |
| 778 | #ifdef ETAGS_REGEXPS |
| 779 | /* Set syntax for regular expression routines. */ |
| 780 | re_set_syntax (RE_SYNTAX_EMACS); |
| 781 | #endif /* ETAGS_REGEXPS */ |
| 782 | |
| 783 | /* |
| 784 | * If etags, always find typedefs and structure tags. Why not? |
| 785 | * Also default is to find macro constants. |
| 786 | */ |
| 787 | if (!CTAGS) |
| 788 | typedefs = typedefs_and_cplusplus = constantypedefs = 1; |
| 789 | |
| 790 | while (1) |
| 791 | { |
| 792 | int opt = getopt_long (argc, argv, |
| 793 | "-aCdDf:l:o:r:RStTi:BuvxwVhH", longopts, 0); |
| 794 | |
| 795 | if (opt == EOF) |
| 796 | break; |
| 797 | |
| 798 | switch (opt) |
| 799 | { |
| 800 | case 0: |
| 801 | /* If getopt returns 0, then it has already processed a |
| 802 | long-named option. We should do nothing. */ |
| 803 | break; |
| 804 | |
| 805 | case 1: |
| 806 | /* This means that a filename has been seen. Record it. */ |
| 807 | argbuffer[current_arg].arg_type = at_filename; |
| 808 | argbuffer[current_arg].what = optarg; |
| 809 | ++current_arg; |
| 810 | ++file_count; |
| 811 | break; |
| 812 | |
| 813 | /* Common options. */ |
| 814 | case 'a': |
| 815 | append_to_tagfile++; |
| 816 | break; |
| 817 | case 'C': |
| 818 | cplusplus = 1; |
| 819 | break; |
| 820 | case 'd': |
| 821 | constantypedefs = 1; |
| 822 | break; |
| 823 | case 'D': |
| 824 | constantypedefs = 0; |
| 825 | break; |
| 826 | case 'f': /* for compatibility with old makefiles */ |
| 827 | case 'o': |
| 828 | if (tagfile) |
| 829 | { |
| 830 | fprintf(stderr, |
| 831 | "%s: -%c flag may only be given once.\n", progname, opt); |
| 832 | goto usage; |
| 833 | } |
| 834 | tagfile = optarg; |
| 835 | break; |
| 836 | case 'l': |
| 837 | if (!get_language (optarg, &argbuffer[current_arg].function)) |
| 838 | { |
| 839 | fprintf (stderr, "%s: language \"%s\" not recognized.\n", |
| 840 | progname, optarg); |
| 841 | goto usage; |
| 842 | } |
| 843 | argbuffer[current_arg].arg_type = at_language; |
| 844 | ++current_arg; |
| 845 | break; |
| 846 | #ifdef ETAGS_REGEXPS |
| 847 | case 'r': |
| 848 | argbuffer[current_arg].arg_type = at_regexp; |
| 849 | argbuffer[current_arg].what = optarg; |
| 850 | ++current_arg; |
| 851 | break; |
| 852 | case 'R': |
| 853 | argbuffer[current_arg].arg_type = at_regexp; |
| 854 | argbuffer[current_arg].what = NULL; |
| 855 | ++current_arg; |
| 856 | break; |
| 857 | #endif /* ETAGS_REGEXPS */ |
| 858 | case 'S': |
| 859 | noindentypedefs++; |
| 860 | break; |
| 861 | case 'V': |
| 862 | print_version (); |
| 863 | break; |
| 864 | case 'h': |
| 865 | case 'H': |
| 866 | print_help (); |
| 867 | break; |
| 868 | case 't': |
| 869 | typedefs++; |
| 870 | break; |
| 871 | case 'T': |
| 872 | typedefs++; |
| 873 | typedefs_and_cplusplus++; |
| 874 | break; |
| 875 | #if (!CTAGS) |
| 876 | /* Etags options */ |
| 877 | case 'i': |
| 878 | included_files[nincluded_files++] = optarg; |
| 879 | break; |
| 880 | #else /* CTAGS */ |
| 881 | /* Ctags options. */ |
| 882 | case 'B': |
| 883 | searchar = '?'; |
| 884 | break; |
| 885 | case 'u': |
| 886 | update++; |
| 887 | break; |
| 888 | case 'v': |
| 889 | vgrind_style++; |
| 890 | /*FALLTHRU*/ |
| 891 | case 'x': |
| 892 | cxref_style++; |
| 893 | break; |
| 894 | case 'w': |
| 895 | no_warnings++; |
| 896 | break; |
| 897 | #endif /* CTAGS */ |
| 898 | default: |
| 899 | fprintf (stderr, |
| 900 | "%s: -%c flag not recognised.\n", progname, opt); |
| 901 | goto usage; |
| 902 | } |
| 903 | } |
| 904 | |
| 905 | for (; optind < argc; ++optind) |
| 906 | { |
| 907 | argbuffer[current_arg].arg_type = at_filename; |
| 908 | argbuffer[current_arg].what = argv[optind]; |
| 909 | ++current_arg; |
| 910 | ++file_count; |
| 911 | } |
| 912 | |
| 913 | if (nincluded_files == 0 && file_count == 0) |
| 914 | { |
| 915 | fprintf (stderr, "%s: No input files specified.\n", progname); |
| 916 | |
| 917 | usage: |
| 918 | fprintf (stderr, "%s: Try `%s --help' for a complete list of options.\n", |
| 919 | progname, progname); |
| 920 | exit (BAD); |
| 921 | } |
| 922 | |
| 923 | if (tagfile == NULL) |
| 924 | { |
| 925 | tagfile = CTAGS ? "tags" : "TAGS"; |
| 926 | } |
| 927 | cwd = etags_getcwd (); /* the current working directory */ |
| 928 | strcat (cwd, "/"); |
| 929 | if (streq (tagfile, "-")) |
| 930 | { |
| 931 | tagfiledir = cwd; |
| 932 | } |
| 933 | else |
| 934 | { |
| 935 | tagfiledir = absolute_dirname (tagfile, cwd); |
| 936 | } |
| 937 | |
| 938 | init (); /* set up boolean "functions" */ |
| 939 | |
| 940 | initbuffer (&lb); |
| 941 | initbuffer (&lbs[0].lb); |
| 942 | initbuffer (&lbs[1].lb); |
| 943 | initbuffer (&filename_lb); |
| 944 | |
| 945 | if (!CTAGS) |
| 946 | { |
| 947 | if (streq (tagfile, "-")) |
| 948 | tagf = stdout; |
| 949 | else |
| 950 | tagf = fopen (tagfile, append_to_tagfile ? "a" : "w"); |
| 951 | if (tagf == NULL) |
| 952 | { |
| 953 | perror (tagfile); |
| 954 | exit (BAD); |
| 955 | } |
| 956 | } |
| 957 | |
| 958 | /* |
| 959 | * Loop through files finding functions. |
| 960 | */ |
| 961 | for (i = 0; i < current_arg; ++i) |
| 962 | { |
| 963 | switch (argbuffer[i].arg_type) |
| 964 | { |
| 965 | case at_language: |
| 966 | lang_func = argbuffer[i].function; |
| 967 | break; |
| 968 | #ifdef ETAGS_REGEXPS |
| 969 | case at_regexp: |
| 970 | add_regex (argbuffer[i].what); |
| 971 | break; |
| 972 | #endif |
| 973 | case at_filename: |
| 974 | #ifdef VMS |
| 975 | while ((this_file = gfnames (argbuffer[i].what, &got_err)) != NULL) |
| 976 | { |
| 977 | if (got_err) |
| 978 | { |
| 979 | error ("Can't find file %s\n", this_file); |
| 980 | argc--, argv++; |
| 981 | } |
| 982 | else |
| 983 | { |
| 984 | this_file = massage_name (this_file); |
| 985 | } |
| 986 | #else |
| 987 | this_file = argbuffer[i].what; |
| 988 | #endif |
| 989 | /* Input file named "-" means read file names from stdin |
| 990 | and use them. */ |
| 991 | if (streq (this_file, "-")) |
| 992 | { |
| 993 | while (!feof (stdin)) |
| 994 | { |
| 995 | /* Use readline_internal so that regexp matching */ |
| 996 | /* is not done on filenames. */ |
| 997 | (void) readline_internal (&filename_lb, stdin); |
| 998 | if (strlen (filename_lb.buffer) > 0) |
| 999 | process_file (filename_lb.buffer); |
| 1000 | } |
| 1001 | } |
| 1002 | else |
| 1003 | process_file (this_file); |
| 1004 | #ifdef VMS |
| 1005 | } |
| 1006 | #endif |
| 1007 | break; |
| 1008 | } |
| 1009 | } |
| 1010 | |
| 1011 | if (!CTAGS) |
| 1012 | { |
| 1013 | while (nincluded_files-- > 0) |
| 1014 | fprintf (tagf, "\f\n%s,include\n", *included_files++); |
| 1015 | |
| 1016 | (void) fclose (tagf); |
| 1017 | exit (GOOD); |
| 1018 | } |
| 1019 | |
| 1020 | if (cxref_style) |
| 1021 | { |
| 1022 | put_entries (head); |
| 1023 | exit (GOOD); |
| 1024 | } |
| 1025 | if (update) |
| 1026 | { |
| 1027 | for (i = 0; i < current_arg; ++i) |
| 1028 | { |
| 1029 | if (argbuffer[i].arg_type == at_language) |
| 1030 | continue; |
| 1031 | sprintf (cmd, |
| 1032 | "mv %s OTAGS;fgrep -v '\t%s\t' OTAGS >%s;rm OTAGS", |
| 1033 | tagfile, argbuffer[i].what, tagfile); |
| 1034 | (void) system (cmd); |
| 1035 | } |
| 1036 | append_to_tagfile++; |
| 1037 | } |
| 1038 | tagf = fopen (tagfile, append_to_tagfile ? "a" : "w"); |
| 1039 | if (tagf == NULL) |
| 1040 | { |
| 1041 | perror (tagfile); |
| 1042 | exit (GOOD); |
| 1043 | } |
| 1044 | put_entries (head); |
| 1045 | (void) fclose (tagf); |
| 1046 | if (update) |
| 1047 | { |
| 1048 | sprintf (cmd, "sort %s -o %s", tagfile, tagfile); |
| 1049 | (void) system (cmd); |
| 1050 | } |
| 1051 | exit (GOOD); |
| 1052 | } |
| 1053 | |
| 1054 | |
| 1055 | /* |
| 1056 | * Set the language, given the name. |
| 1057 | */ |
| 1058 | logical |
| 1059 | get_language (language, func) |
| 1060 | char *language; |
| 1061 | Lang_function **func; |
| 1062 | { |
| 1063 | struct lang_entry *lang; |
| 1064 | |
| 1065 | for (lang = lang_names; lang->extension; ++lang) |
| 1066 | { |
| 1067 | if (streq (language, lang->extension)) |
| 1068 | { |
| 1069 | *func = lang->function; |
| 1070 | return TRUE; |
| 1071 | } |
| 1072 | } |
| 1073 | |
| 1074 | return FALSE; |
| 1075 | } |
| 1076 | |
| 1077 | |
| 1078 | /* |
| 1079 | * This routine is called on each file argument. |
| 1080 | */ |
| 1081 | void |
| 1082 | process_file (file) |
| 1083 | char *file; |
| 1084 | { |
| 1085 | struct stat stat_buf; |
| 1086 | |
| 1087 | if (stat (file, &stat_buf) == 0 && !S_ISREG (stat_buf.st_mode)) |
| 1088 | { |
| 1089 | fprintf (stderr, "Skipping %s: it is not a regular file.\n", file); |
| 1090 | return; |
| 1091 | } |
| 1092 | if (streq (file, tagfile) && !streq (tagfile, "-")) |
| 1093 | { |
| 1094 | fprintf (stderr, "Skipping inclusion of %s in self.\n", file); |
| 1095 | return; |
| 1096 | } |
| 1097 | if (!find_entries (file)) |
| 1098 | { |
| 1099 | return; |
| 1100 | } |
| 1101 | if (!CTAGS) |
| 1102 | { |
| 1103 | char *filename; |
| 1104 | |
| 1105 | if (file[0] == '/') |
| 1106 | { |
| 1107 | /* file is an absolute filename. Canonicalise it. */ |
| 1108 | filename = absolute_filename (file, cwd); |
| 1109 | } |
| 1110 | else |
| 1111 | { |
| 1112 | /* file is a filename relative to cwd. Make it relative |
| 1113 | to the directory of the tags file. */ |
| 1114 | filename = relative_filename (file, tagfiledir); |
| 1115 | } |
| 1116 | fprintf (tagf, "\f\n%s,%d\n", filename, total_size_of_entries (head)); |
| 1117 | put_entries (head); |
| 1118 | free_tree (head); |
| 1119 | head = NULL; |
| 1120 | } |
| 1121 | } |
| 1122 | |
| 1123 | /* |
| 1124 | * This routine sets up the boolean pseudo-functions which work |
| 1125 | * by setting boolean flags dependent upon the corresponding character |
| 1126 | * Every char which is NOT in that string is not a white char. Therefore, |
| 1127 | * all of the array "_wht" is set to FALSE, and then the elements |
| 1128 | * subscripted by the chars in "white" are set to TRUE. Thus "_wht" |
| 1129 | * of a char is TRUE if it is the string "white", else FALSE. |
| 1130 | */ |
| 1131 | void |
| 1132 | init () |
| 1133 | { |
| 1134 | register char *sp; |
| 1135 | register int i; |
| 1136 | |
| 1137 | for (i = 0; i < 0177; i++) |
| 1138 | _wht[i] = _etk[i] = _itk[i] = _btk[i] = FALSE; |
| 1139 | for (sp = white; *sp; sp++) |
| 1140 | _wht[*sp] = TRUE; |
| 1141 | for (sp = endtk; *sp; sp++) |
| 1142 | _etk[*sp] = TRUE; |
| 1143 | for (sp = intk; *sp; sp++) |
| 1144 | _itk[*sp] = TRUE; |
| 1145 | for (sp = begtk; *sp; sp++) |
| 1146 | _btk[*sp] = TRUE; |
| 1147 | _wht[0] = _wht['\n']; |
| 1148 | _etk[0] = _etk['\n']; |
| 1149 | _btk[0] = _btk['\n']; |
| 1150 | _itk[0] = _itk['\n']; |
| 1151 | } |
| 1152 | |
| 1153 | /* |
| 1154 | * This routine opens the specified file and calls the function |
| 1155 | * which finds the function and type definitions. |
| 1156 | */ |
| 1157 | logical |
| 1158 | find_entries (file) |
| 1159 | char *file; |
| 1160 | { |
| 1161 | FILE *inf; |
| 1162 | char *cp; |
| 1163 | struct lang_entry *lang; |
| 1164 | NODE *old_last_node; |
| 1165 | extern NODE *last_node; |
| 1166 | |
| 1167 | inf = fopen (file, "r"); |
| 1168 | if (inf == NULL) |
| 1169 | { |
| 1170 | perror (file); |
| 1171 | return FALSE; |
| 1172 | } |
| 1173 | curfile = savestr (file); |
| 1174 | cp = etags_strrchr (file, '.'); |
| 1175 | |
| 1176 | /* If user specified a language, use it. */ |
| 1177 | if (lang_func != NULL) |
| 1178 | { |
| 1179 | lang_func (inf); |
| 1180 | fclose (inf); |
| 1181 | return TRUE; |
| 1182 | } |
| 1183 | |
| 1184 | if (cp) |
| 1185 | { |
| 1186 | ++cp; |
| 1187 | for (lang = lang_extensions; lang->extension; ++lang) |
| 1188 | { |
| 1189 | if (streq (cp, lang->extension)) |
| 1190 | { |
| 1191 | lang->function (inf); |
| 1192 | fclose (inf); |
| 1193 | return TRUE; |
| 1194 | } |
| 1195 | } |
| 1196 | } |
| 1197 | |
| 1198 | /* Try Fortran. */ |
| 1199 | old_last_node = last_node; |
| 1200 | Fortran_functions (inf); |
| 1201 | |
| 1202 | /* No Fortran entries found. Try C. */ |
| 1203 | if (old_last_node == last_node) |
| 1204 | default_C_entries (inf); |
| 1205 | fclose (inf); |
| 1206 | return TRUE; |
| 1207 | } |
| 1208 | \f |
| 1209 | /* Record a tag. */ |
| 1210 | /* Should take a TOKEN* instead!! */ |
| 1211 | void |
| 1212 | pfnote (name, is_func, named, linestart, linelen, lno, cno) |
| 1213 | char *name; /* tag name */ |
| 1214 | logical is_func; /* function or type name? */ |
| 1215 | logical named; /* tag different from text of definition? */ |
| 1216 | char *linestart; |
| 1217 | int linelen; |
| 1218 | int lno; |
| 1219 | long cno; |
| 1220 | { |
| 1221 | register char *fp; |
| 1222 | register NODE *np; |
| 1223 | char tem[51]; |
| 1224 | char c; |
| 1225 | |
| 1226 | np = xnew (1, NODE); |
| 1227 | if (np == NULL) |
| 1228 | { |
| 1229 | if (CTAGS) |
| 1230 | { |
| 1231 | /* It's okay to output early in etags -- it only disrupts the |
| 1232 | * character count of the tag entries, which is no longer used |
| 1233 | * by tags.el anyway. |
| 1234 | */ |
| 1235 | error ("too many entries to sort", 0); |
| 1236 | } |
| 1237 | put_entries (head); |
| 1238 | free_tree (head); |
| 1239 | head = NULL; |
| 1240 | np = xnew (1, NODE); |
| 1241 | } |
| 1242 | /* If ctags mode, change name "main" to M<thisfilename>. */ |
| 1243 | if (CTAGS && !cxref_style && streq (name, "main")) |
| 1244 | { |
| 1245 | fp = etags_strrchr (curfile, '/'); |
| 1246 | name = concat ("M", fp == 0 ? curfile : fp + 1, ""); |
| 1247 | fp = etags_strrchr (name, '.'); |
| 1248 | if (fp && fp[1] != '\0' && fp[2] == '\0') |
| 1249 | *fp = 0; |
| 1250 | named = TRUE; |
| 1251 | } |
| 1252 | np->name = savestr (name); |
| 1253 | np->file = curfile; |
| 1254 | np->is_func = is_func; |
| 1255 | np->named = named; |
| 1256 | np->lno = lno; |
| 1257 | /* Our char numbers are 0-base, because of C language tradition? |
| 1258 | ctags compatibility? old versions compatibility? I don't know. |
| 1259 | Anyway, since emacs's are 1-base we espect etags.el to take care |
| 1260 | of the difference. If we wanted to have 1-based numbers, we would |
| 1261 | uncomment the +1 below. */ |
| 1262 | np->cno = cno /* + 1 */ ; |
| 1263 | np->left = np->right = 0; |
| 1264 | if (!CTAGS) |
| 1265 | { |
| 1266 | c = linestart[linelen]; |
| 1267 | linestart[linelen] = 0; |
| 1268 | } |
| 1269 | else if (cxref_style == 0) |
| 1270 | { |
| 1271 | sprintf (tem, strlen (linestart) < 50 ? "%s$" : "%.50s", linestart); |
| 1272 | linestart = tem; |
| 1273 | } |
| 1274 | np->pat = savestr (linestart); |
| 1275 | if (!CTAGS) |
| 1276 | { |
| 1277 | linestart[linelen] = c; |
| 1278 | } |
| 1279 | |
| 1280 | add_node (np, &head); |
| 1281 | } |
| 1282 | |
| 1283 | /* |
| 1284 | * free_tree () |
| 1285 | * recurse on left children, iterate on right children. |
| 1286 | */ |
| 1287 | void |
| 1288 | free_tree (node) |
| 1289 | register NODE *node; |
| 1290 | { |
| 1291 | while (node) |
| 1292 | { |
| 1293 | register NODE *node_right = node->right; |
| 1294 | free_tree (node->left); |
| 1295 | free (node->name); |
| 1296 | free (node->pat); |
| 1297 | free ((char *) node); |
| 1298 | node = node_right; |
| 1299 | } |
| 1300 | } |
| 1301 | |
| 1302 | /* |
| 1303 | * add_node () |
| 1304 | * Adds a node to the tree of nodes. In etags mode, we don't keep |
| 1305 | * it sorted; we just keep a linear list. In ctags mode, maintain |
| 1306 | * an ordered tree, with no attempt at balancing. |
| 1307 | * |
| 1308 | * add_node is the only function allowed to add nodes, so it can |
| 1309 | * maintain state. |
| 1310 | */ |
| 1311 | /* Must avoid static vars within functions since some systems |
| 1312 | #define static as nothing. */ |
| 1313 | NODE *last_node = NULL; |
| 1314 | |
| 1315 | void |
| 1316 | add_node (node, cur_node_p) |
| 1317 | NODE *node, **cur_node_p; |
| 1318 | { |
| 1319 | register int dif; |
| 1320 | register NODE *cur_node = *cur_node_p; |
| 1321 | |
| 1322 | if (cur_node == NULL) |
| 1323 | { |
| 1324 | *cur_node_p = node; |
| 1325 | last_node = node; |
| 1326 | return; |
| 1327 | } |
| 1328 | |
| 1329 | if (!CTAGS) |
| 1330 | { |
| 1331 | /* Etags Mode */ |
| 1332 | if (last_node == NULL) |
| 1333 | fatal ("internal error in add_node", 0); |
| 1334 | last_node->right = node; |
| 1335 | last_node = node; |
| 1336 | } |
| 1337 | else |
| 1338 | { |
| 1339 | /* Ctags Mode */ |
| 1340 | dif = strcmp (node->name, cur_node->name); |
| 1341 | |
| 1342 | /* |
| 1343 | * If this tag name matches an existing one, then |
| 1344 | * do not add the node, but maybe print a warning. |
| 1345 | */ |
| 1346 | if (!dif) |
| 1347 | { |
| 1348 | if (node->file == cur_node->file) |
| 1349 | { |
| 1350 | if (!no_warnings) |
| 1351 | { |
| 1352 | fprintf (stderr, "Duplicate entry in file %s, line %d: %s\n", |
| 1353 | node->file, lineno, node->name); |
| 1354 | fprintf (stderr, "Second entry ignored\n"); |
| 1355 | } |
| 1356 | return; |
| 1357 | } |
| 1358 | if (!cur_node->been_warned && !no_warnings) |
| 1359 | { |
| 1360 | fprintf (stderr, |
| 1361 | "Duplicate entry in files %s and %s: %s (Warning only)\n", |
| 1362 | node->file, cur_node->file, node->name); |
| 1363 | } |
| 1364 | cur_node->been_warned = TRUE; |
| 1365 | return; |
| 1366 | } |
| 1367 | |
| 1368 | /* Maybe refuse to add duplicate nodes. */ |
| 1369 | if (!permit_duplicates) |
| 1370 | { |
| 1371 | if (streq (node->name, cur_node->name) |
| 1372 | && streq (node->file, cur_node->file)) |
| 1373 | return; |
| 1374 | } |
| 1375 | |
| 1376 | /* Actually add the node */ |
| 1377 | add_node (node, dif < 0 ? &cur_node->left : &cur_node->right); |
| 1378 | } |
| 1379 | } |
| 1380 | \f |
| 1381 | void |
| 1382 | put_entries (node) |
| 1383 | register NODE *node; |
| 1384 | { |
| 1385 | register char *sp; |
| 1386 | |
| 1387 | if (node == NULL) |
| 1388 | return; |
| 1389 | |
| 1390 | /* Output subentries that precede this one */ |
| 1391 | put_entries (node->left); |
| 1392 | |
| 1393 | /* Output this entry */ |
| 1394 | |
| 1395 | if (!CTAGS) |
| 1396 | { |
| 1397 | if (node->named) |
| 1398 | { |
| 1399 | fprintf (tagf, "%s\177%s\001%d,%d\n", |
| 1400 | node->pat, node->name, |
| 1401 | node->lno, node->cno); |
| 1402 | } |
| 1403 | else |
| 1404 | { |
| 1405 | fprintf (tagf, "%s\177%d,%d\n", |
| 1406 | node->pat, |
| 1407 | node->lno, node->cno); |
| 1408 | } |
| 1409 | } |
| 1410 | else if (!cxref_style) |
| 1411 | { |
| 1412 | fprintf (tagf, "%s\t%s\t", |
| 1413 | node->name, node->file); |
| 1414 | |
| 1415 | if (node->is_func) |
| 1416 | { /* a function */ |
| 1417 | putc (searchar, tagf); |
| 1418 | putc ('^', tagf); |
| 1419 | |
| 1420 | for (sp = node->pat; *sp; sp++) |
| 1421 | { |
| 1422 | if (*sp == '\\' || *sp == searchar) |
| 1423 | putc ('\\', tagf); |
| 1424 | putc (*sp, tagf); |
| 1425 | } |
| 1426 | putc (searchar, tagf); |
| 1427 | } |
| 1428 | else |
| 1429 | { /* a typedef; text pattern inadequate */ |
| 1430 | fprintf (tagf, "%d", node->lno); |
| 1431 | } |
| 1432 | putc ('\n', tagf); |
| 1433 | } |
| 1434 | else if (vgrind_style) |
| 1435 | fprintf (stdout, "%s %s %d\n", |
| 1436 | node->name, node->file, (node->lno + 63) / 64); |
| 1437 | else |
| 1438 | fprintf (stdout, "%-16s %3d %-16s %s\n", |
| 1439 | node->name, node->lno, node->file, node->pat); |
| 1440 | |
| 1441 | /* Output subentries that follow this one */ |
| 1442 | put_entries (node->right); |
| 1443 | } |
| 1444 | |
| 1445 | /* Length of a number's decimal representation. */ |
| 1446 | int |
| 1447 | number_len (num) |
| 1448 | long num; |
| 1449 | { |
| 1450 | int len = 0; |
| 1451 | if (!num) |
| 1452 | return 1; |
| 1453 | for (; num; num /= 10) |
| 1454 | ++len; |
| 1455 | return len; |
| 1456 | } |
| 1457 | |
| 1458 | /* |
| 1459 | * Return total number of characters that put_entries will output for |
| 1460 | * the nodes in the subtree of the specified node. Works only if |
| 1461 | * we are not ctags, but called only in that case. This count |
| 1462 | * is irrelevant with the new tags.el, but is still supplied for |
| 1463 | * backward compatibility. |
| 1464 | */ |
| 1465 | int |
| 1466 | total_size_of_entries (node) |
| 1467 | register NODE *node; |
| 1468 | { |
| 1469 | register int total; |
| 1470 | |
| 1471 | if (node == NULL) |
| 1472 | return 0; |
| 1473 | |
| 1474 | total = 0; |
| 1475 | for (; node; node = node->right) |
| 1476 | { |
| 1477 | /* Count left subentries. */ |
| 1478 | total += total_size_of_entries (node->left); |
| 1479 | |
| 1480 | /* Count this entry */ |
| 1481 | total += strlen (node->pat) + 1; |
| 1482 | total += number_len ((long) node->lno) + 1 + number_len (node->cno) + 1; |
| 1483 | if (node->named) |
| 1484 | total += 1 + strlen (node->name); /* \001name */ |
| 1485 | } |
| 1486 | |
| 1487 | return total; |
| 1488 | } |
| 1489 | \f |
| 1490 | /* |
| 1491 | * The C symbol tables. |
| 1492 | */ |
| 1493 | |
| 1494 | /* Feed stuff between (but not including) %[ and %] lines to: |
| 1495 | gperf -c -k1,3 -o -p -r -t |
| 1496 | %[ |
| 1497 | struct C_stab_entry { char *name; int c_ext; enum sym_type type; } |
| 1498 | %% |
| 1499 | class, C_PLPL, st_C_struct |
| 1500 | domain, C_STAR, st_C_struct |
| 1501 | union, 0, st_C_struct |
| 1502 | struct, 0, st_C_struct |
| 1503 | enum, 0, st_C_enum |
| 1504 | typedef, 0, st_C_typedef |
| 1505 | define, 0, st_C_define |
| 1506 | long, 0, st_C_typespec |
| 1507 | short, 0, st_C_typespec |
| 1508 | int, 0, st_C_typespec |
| 1509 | char, 0, st_C_typespec |
| 1510 | float, 0, st_C_typespec |
| 1511 | double, 0, st_C_typespec |
| 1512 | signed, 0, st_C_typespec |
| 1513 | unsigned, 0, st_C_typespec |
| 1514 | auto, 0, st_C_typespec |
| 1515 | void, 0, st_C_typespec |
| 1516 | extern, 0, st_C_typespec |
| 1517 | static, 0, st_C_typespec |
| 1518 | const, 0, st_C_typespec |
| 1519 | volatile, 0, st_C_typespec |
| 1520 | %] |
| 1521 | and replace lines between %< and %> with its output. */ |
| 1522 | /*%<*/ |
| 1523 | /* C code produced by gperf version 1.8.1 (K&R C version) */ |
| 1524 | /* Command-line: gperf -c -k1,3 -o -p -r -t */ |
| 1525 | |
| 1526 | |
| 1527 | struct C_stab_entry { char *name; int c_ext; enum sym_type type; }; |
| 1528 | |
| 1529 | #define MIN_WORD_LENGTH 3 |
| 1530 | #define MAX_WORD_LENGTH 8 |
| 1531 | #define MIN_HASH_VALUE 10 |
| 1532 | #define MAX_HASH_VALUE 62 |
| 1533 | /* |
| 1534 | 21 keywords |
| 1535 | 53 is the maximum key range |
| 1536 | */ |
| 1537 | |
| 1538 | static int |
| 1539 | hash (str, len) |
| 1540 | register char *str; |
| 1541 | register int len; |
| 1542 | { |
| 1543 | static unsigned char hash_table[] = |
| 1544 | { |
| 1545 | 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, |
| 1546 | 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, |
| 1547 | 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, |
| 1548 | 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, |
| 1549 | 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, |
| 1550 | 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, |
| 1551 | 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, |
| 1552 | 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, |
| 1553 | 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, |
| 1554 | 62, 62, 62, 62, 62, 62, 62, 2, 62, 7, |
| 1555 | 6, 9, 15, 30, 62, 24, 62, 62, 1, 24, |
| 1556 | 7, 27, 13, 62, 19, 26, 18, 27, 1, 62, |
| 1557 | 62, 62, 62, 62, 62, 62, 62, 62, |
| 1558 | }; |
| 1559 | return len + hash_table[str[2]] + hash_table[str[0]]; |
| 1560 | } |
| 1561 | |
| 1562 | struct C_stab_entry * |
| 1563 | in_word_set (str, len) |
| 1564 | register char *str; |
| 1565 | register int len; |
| 1566 | { |
| 1567 | |
| 1568 | static struct C_stab_entry wordlist[] = |
| 1569 | { |
| 1570 | {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, {"",}, |
| 1571 | {"",}, |
| 1572 | {"volatile", 0, st_C_typespec}, |
| 1573 | {"",}, |
| 1574 | {"long", 0, st_C_typespec}, |
| 1575 | {"char", 0, st_C_typespec}, |
| 1576 | {"class", C_PLPL, st_C_struct}, |
| 1577 | {"",}, {"",}, {"",}, {"",}, |
| 1578 | {"const", 0, st_C_typespec}, |
| 1579 | {"",}, {"",}, {"",}, {"",}, |
| 1580 | {"auto", 0, st_C_typespec}, |
| 1581 | {"",}, {"",}, |
| 1582 | {"define", 0, st_C_define}, |
| 1583 | {"",}, |
| 1584 | {"void", 0, st_C_typespec}, |
| 1585 | {"",}, {"",}, {"",}, |
| 1586 | {"extern", 0, st_C_typespec}, |
| 1587 | {"static", 0, st_C_typespec}, |
| 1588 | {"",}, |
| 1589 | {"domain", C_STAR, st_C_struct}, |
| 1590 | {"",}, |
| 1591 | {"typedef", 0, st_C_typedef}, |
| 1592 | {"double", 0, st_C_typespec}, |
| 1593 | {"enum", 0, st_C_enum}, |
| 1594 | {"",}, {"",}, {"",}, {"",}, |
| 1595 | {"int", 0, st_C_typespec}, |
| 1596 | {"",}, |
| 1597 | {"float", 0, st_C_typespec}, |
| 1598 | {"",}, {"",}, {"",}, |
| 1599 | {"struct", 0, st_C_struct}, |
| 1600 | {"",}, {"",}, {"",}, {"",}, |
| 1601 | {"union", 0, st_C_struct}, |
| 1602 | {"",}, |
| 1603 | {"short", 0, st_C_typespec}, |
| 1604 | {"",}, {"",}, |
| 1605 | {"unsigned", 0, st_C_typespec}, |
| 1606 | {"signed", 0, st_C_typespec}, |
| 1607 | }; |
| 1608 | |
| 1609 | if (len <= MAX_WORD_LENGTH && len >= MIN_WORD_LENGTH) |
| 1610 | { |
| 1611 | register int key = hash (str, len); |
| 1612 | |
| 1613 | if (key <= MAX_HASH_VALUE && key >= MIN_HASH_VALUE) |
| 1614 | { |
| 1615 | register char *s = wordlist[key].name; |
| 1616 | |
| 1617 | if (*s == *str && strneq (str + 1, s + 1, len - 1)) |
| 1618 | return &wordlist[key]; |
| 1619 | } |
| 1620 | } |
| 1621 | return 0; |
| 1622 | } |
| 1623 | /*%>*/ |
| 1624 | |
| 1625 | enum sym_type |
| 1626 | C_symtype(str, len, c_ext) |
| 1627 | char *str; |
| 1628 | int len; |
| 1629 | int c_ext; |
| 1630 | { |
| 1631 | register struct C_stab_entry *se = in_word_set(str, len); |
| 1632 | |
| 1633 | if (se == NULL || (se->c_ext && !(c_ext & se->c_ext))) |
| 1634 | return st_none; |
| 1635 | return se->type; |
| 1636 | } |
| 1637 | \f |
| 1638 | /* |
| 1639 | * C functions are recognized using a simple finite automaton. |
| 1640 | * funcdef is its state variable. |
| 1641 | */ |
| 1642 | typedef enum |
| 1643 | { |
| 1644 | fnone, /* nothing seen */ |
| 1645 | ftagseen, /* function-like tag seen */ |
| 1646 | fstartlist, /* just after open parenthesis */ |
| 1647 | finlist, /* in parameter list */ |
| 1648 | flistseen, /* after parameter list */ |
| 1649 | fignore /* before open brace */ |
| 1650 | } FUNCST; |
| 1651 | FUNCST funcdef; |
| 1652 | |
| 1653 | |
| 1654 | /* |
| 1655 | * typedefs are recognized using a simple finite automaton. |
| 1656 | * typeddef is its state variable. |
| 1657 | */ |
| 1658 | typedef enum |
| 1659 | { |
| 1660 | tnone, /* nothing seen */ |
| 1661 | ttypedseen, /* typedef keyword seen */ |
| 1662 | tinbody, /* inside typedef body */ |
| 1663 | tend, /* just before typedef tag */ |
| 1664 | tignore /* junk after typedef tag */ |
| 1665 | } TYPEDST; |
| 1666 | TYPEDST typdef; |
| 1667 | |
| 1668 | |
| 1669 | /* |
| 1670 | * struct-like structures (enum, struct and union) are recognized |
| 1671 | * using another simple finite automaton. `structdef' is its state |
| 1672 | * variable. |
| 1673 | */ |
| 1674 | typedef enum |
| 1675 | { |
| 1676 | snone, /* nothing seen yet */ |
| 1677 | skeyseen, /* struct-like keyword seen */ |
| 1678 | stagseen, /* struct-like tag seen */ |
| 1679 | scolonseen, /* colon seen after struct-like tag */ |
| 1680 | sinbody /* in struct body: recognize member func defs*/ |
| 1681 | } STRUCTST; |
| 1682 | STRUCTST structdef; |
| 1683 | |
| 1684 | /* |
| 1685 | * When structdef is stagseen, scolonseen, or sinbody, structtag is the |
| 1686 | * struct tag, and structtype is the type of the preceding struct-like |
| 1687 | * keyword. |
| 1688 | */ |
| 1689 | char structtag[BUFSIZ]; |
| 1690 | enum sym_type structtype; |
| 1691 | |
| 1692 | /* |
| 1693 | * Yet another little state machine to deal with preprocessor lines. |
| 1694 | */ |
| 1695 | typedef enum |
| 1696 | { |
| 1697 | dnone, /* nothing seen */ |
| 1698 | dsharpseen, /* '#' seen as first char on line */ |
| 1699 | ddefineseen, /* '#' and 'define' seen */ |
| 1700 | dignorerest /* ignore rest of line */ |
| 1701 | } DEFINEST; |
| 1702 | DEFINEST definedef; |
| 1703 | |
| 1704 | /* |
| 1705 | * Set this to TRUE, and the next token considered is called a function. |
| 1706 | * Used only for GNUmacs's function-defining macros. |
| 1707 | */ |
| 1708 | logical next_token_is_func; |
| 1709 | |
| 1710 | /* |
| 1711 | * TRUE in the rules part of a yacc file, FALSE outside (parse as C). |
| 1712 | */ |
| 1713 | logical yacc_rules; |
| 1714 | |
| 1715 | /* |
| 1716 | * consider_token () |
| 1717 | * checks to see if the current token is at the start of a |
| 1718 | * function, or corresponds to a typedef, or is a struct/union/enum |
| 1719 | * tag. |
| 1720 | * |
| 1721 | * *IS_FUNC gets TRUE iff the token is a function or macro with args. |
| 1722 | * C_EXT is which language we are looking at. |
| 1723 | * |
| 1724 | * In the future we will need some way to adjust where the end of |
| 1725 | * the token is; for instance, implementing the C++ keyword |
| 1726 | * `operator' properly will adjust the end of the token to be after |
| 1727 | * whatever follows `operator'. |
| 1728 | * |
| 1729 | * Globals |
| 1730 | * funcdef IN OUT |
| 1731 | * structdef IN OUT |
| 1732 | * definedef IN OUT |
| 1733 | * typdef IN OUT |
| 1734 | * next_token_is_func IN OUT |
| 1735 | */ |
| 1736 | |
| 1737 | logical |
| 1738 | consider_token (c, tokp, c_ext, cblev, is_func) |
| 1739 | register char c; /* IN: first char after the token */ |
| 1740 | register TOKEN *tokp; /* IN: token pointer */ |
| 1741 | int c_ext; /* IN: C extensions mask */ |
| 1742 | int cblev; /* IN: curly brace level */ |
| 1743 | logical *is_func; /* OUT: function found */ |
| 1744 | { |
| 1745 | enum sym_type toktype = C_symtype(tokp->p, tokp->len, c_ext); |
| 1746 | |
| 1747 | /* |
| 1748 | * Advance the definedef state machine. |
| 1749 | */ |
| 1750 | switch (definedef) |
| 1751 | { |
| 1752 | case dnone: |
| 1753 | /* We're not on a preprocessor line. */ |
| 1754 | break; |
| 1755 | case dsharpseen: |
| 1756 | if (toktype == st_C_define) |
| 1757 | { |
| 1758 | definedef = ddefineseen; |
| 1759 | } |
| 1760 | else |
| 1761 | { |
| 1762 | definedef = dignorerest; |
| 1763 | } |
| 1764 | return FALSE; |
| 1765 | case ddefineseen: |
| 1766 | /* |
| 1767 | * Make a tag for any macro, unless it is a constant |
| 1768 | * and constantypedefs is FALSE. |
| 1769 | */ |
| 1770 | definedef = dignorerest; |
| 1771 | *is_func = (c == '('); |
| 1772 | if (!*is_func && !constantypedefs) |
| 1773 | return FALSE; |
| 1774 | else |
| 1775 | return TRUE; |
| 1776 | case dignorerest: |
| 1777 | return FALSE; |
| 1778 | default: |
| 1779 | error ("internal error: definedef value.", 0); |
| 1780 | } |
| 1781 | |
| 1782 | /* |
| 1783 | * Now typedefs |
| 1784 | */ |
| 1785 | switch (typdef) |
| 1786 | { |
| 1787 | case tnone: |
| 1788 | if (toktype == st_C_typedef) |
| 1789 | { |
| 1790 | if (typedefs) |
| 1791 | typdef = ttypedseen; |
| 1792 | funcdef = fnone; |
| 1793 | return FALSE; |
| 1794 | } |
| 1795 | break; |
| 1796 | case ttypedseen: |
| 1797 | switch (toktype) |
| 1798 | { |
| 1799 | case st_none: |
| 1800 | case st_C_typespec: |
| 1801 | typdef = tend; |
| 1802 | break; |
| 1803 | case st_C_struct: |
| 1804 | case st_C_enum: |
| 1805 | break; |
| 1806 | } |
| 1807 | /* Do not return here, so the structdef stuff has a chance. */ |
| 1808 | break; |
| 1809 | case tend: |
| 1810 | switch (toktype) |
| 1811 | { |
| 1812 | case st_C_typespec: |
| 1813 | case st_C_struct: |
| 1814 | case st_C_enum: |
| 1815 | return FALSE; |
| 1816 | } |
| 1817 | return TRUE; |
| 1818 | } |
| 1819 | |
| 1820 | /* |
| 1821 | * This structdef business is currently only invoked when cblev==0. |
| 1822 | * It should be recursively invoked whatever the curly brace level, |
| 1823 | * and a stack of states kept, to allow for definitions of structs |
| 1824 | * within structs. |
| 1825 | * |
| 1826 | * This structdef business is NOT invoked when we are ctags and the |
| 1827 | * file is plain C. This is because a struct tag may have the same |
| 1828 | * name as another tag, and this loses with ctags. |
| 1829 | * |
| 1830 | * This if statement deals with the typdef state machine as |
| 1831 | * follows: if typdef==ttypedseen and token is struct/union/class/enum, |
| 1832 | * return FALSE. All the other code here is for the structdef |
| 1833 | * state machine. |
| 1834 | */ |
| 1835 | switch (toktype) |
| 1836 | { |
| 1837 | case st_C_struct: |
| 1838 | case st_C_enum: |
| 1839 | if (typdef == ttypedseen |
| 1840 | || (typedefs_and_cplusplus && cblev == 0 && structdef == snone)) |
| 1841 | { |
| 1842 | structdef = skeyseen; |
| 1843 | structtype = toktype; |
| 1844 | } |
| 1845 | return FALSE; |
| 1846 | } |
| 1847 | if (structdef == skeyseen) |
| 1848 | { |
| 1849 | if (structtype == st_C_struct) |
| 1850 | { |
| 1851 | strncpy (structtag, tokp->p, tokp->len); |
| 1852 | structtag[tokp->len] = '\0'; /* for struct/union/class */ |
| 1853 | } |
| 1854 | else |
| 1855 | { |
| 1856 | structtag[0] = '\0'; /* for enum (why is it treated differently?) */ |
| 1857 | } |
| 1858 | structdef = stagseen; |
| 1859 | return TRUE; |
| 1860 | } |
| 1861 | |
| 1862 | /* Avoid entering funcdef stuff if typdef is going on. */ |
| 1863 | if (typdef != tnone) |
| 1864 | { |
| 1865 | definedef = dnone; |
| 1866 | return FALSE; |
| 1867 | } |
| 1868 | |
| 1869 | /* Detect GNU macros. */ |
| 1870 | if (definedef == dnone) |
| 1871 | if (strneq (tokp->p, "DEFUN", 5) /* Used in emacs */ |
| 1872 | #if FALSE |
| 1873 | These are defined inside C functions, so currently they |
| 1874 | are not met anyway. |
| 1875 | || strneq (tokp->p, "EXFUN", 5) /* Used in glibc */ |
| 1876 | || strneq (tokp->p, "DEFVAR_", 7) /* Used in emacs */ |
| 1877 | #endif |
| 1878 | || strneq (tokp->p, "SYSCALL", 7) /* Used in glibc (mach) */ |
| 1879 | || strneq (tokp->p, "ENTRY", 5) /* Used in glibc */ |
| 1880 | || strneq (tokp->p, "PSEUDO", 6)) /* Used in glibc */ |
| 1881 | |
| 1882 | { |
| 1883 | next_token_is_func = TRUE; |
| 1884 | return FALSE; |
| 1885 | } |
| 1886 | if (next_token_is_func) |
| 1887 | { |
| 1888 | next_token_is_func = FALSE; |
| 1889 | funcdef = fignore; |
| 1890 | *is_func = TRUE; |
| 1891 | return TRUE; |
| 1892 | } |
| 1893 | |
| 1894 | /* A function? */ |
| 1895 | switch (toktype) |
| 1896 | { |
| 1897 | case st_C_typespec: |
| 1898 | if (funcdef != finlist && funcdef != fignore) |
| 1899 | funcdef = fnone; /* should be useless */ |
| 1900 | return FALSE; |
| 1901 | default: |
| 1902 | if (funcdef == fnone) |
| 1903 | { |
| 1904 | funcdef = ftagseen; |
| 1905 | *is_func = TRUE; |
| 1906 | return TRUE; |
| 1907 | } |
| 1908 | } |
| 1909 | |
| 1910 | return FALSE; |
| 1911 | } |
| 1912 | |
| 1913 | /* |
| 1914 | * C_entries () |
| 1915 | * This routine finds functions, typedefs, #define's and |
| 1916 | * struct/union/enum definitions in C syntax and adds them |
| 1917 | * to the list. |
| 1918 | */ |
| 1919 | |
| 1920 | #define curlb (lbs[curndx].lb) |
| 1921 | #define othlb (lbs[1-curndx].lb) |
| 1922 | #define newlb (lbs[newndx].lb) |
| 1923 | #define curlinepos (lbs[curndx].linepos) |
| 1924 | #define othlinepos (lbs[1-curndx].linepos) |
| 1925 | #define newlinepos (lbs[newndx].linepos) |
| 1926 | |
| 1927 | /* Save and restore token state. This is used when preprocessor defines |
| 1928 | are handled, to avoid disturbing active function/typedef/struct states. */ |
| 1929 | #define TOKEN_SAVED_P (savetok.lineno > 0) |
| 1930 | #define SAVE_TOKEN (savetok = tok, savetok.p = (char *) tokoff, \ |
| 1931 | savetok.len = toklen, strcpy(savenameb, nameb)) |
| 1932 | #define RESTORE_TOKEN (tok = savetok, tokoff = (int) tok.p, \ |
| 1933 | toklen = tok.len, strcpy(nameb, savenameb), \ |
| 1934 | savetok.lineno = 0) |
| 1935 | |
| 1936 | #define CNL_SAVE_DEFINEDEF \ |
| 1937 | do { \ |
| 1938 | SET_FILEPOS (curlinepos, inf, charno); \ |
| 1939 | lineno++; \ |
| 1940 | charno += readline (&curlb, inf); \ |
| 1941 | lp = curlb.buffer; \ |
| 1942 | quotednl = FALSE; \ |
| 1943 | newndx = curndx; \ |
| 1944 | } while (0) |
| 1945 | |
| 1946 | #define CNL \ |
| 1947 | do { \ |
| 1948 | CNL_SAVE_DEFINEDEF; \ |
| 1949 | if (TOKEN_SAVED_P) \ |
| 1950 | RESTORE_TOKEN; \ |
| 1951 | definedef = dnone; \ |
| 1952 | } while (0) |
| 1953 | |
| 1954 | #define MAKE_TAG_FROM_NEW_LB(isfun) pfnote (nameb, isfun, tok.named, \ |
| 1955 | newlb.buffer, tokoff + toklen + 1, tok.lineno, GET_CHARNO (newlinepos)) |
| 1956 | #define MAKE_TAG_FROM_OTH_LB(isfun) pfnote (nameb, isfun, tok.named, \ |
| 1957 | othlb.buffer, tokoff + toklen + 1, tok.lineno, GET_CHARNO (othlinepos)) |
| 1958 | |
| 1959 | void |
| 1960 | C_entries (c_ext, inf) |
| 1961 | int c_ext; /* extension of C */ |
| 1962 | FILE *inf; /* input file */ |
| 1963 | { |
| 1964 | register char c; /* latest char read; '\0' for end of line */ |
| 1965 | register char *lp; /* pointer one beyond the character `c' */ |
| 1966 | int curndx, newndx; /* indices for current and new lb */ |
| 1967 | TOKEN tok; /* latest token read for funcdef & structdef */ |
| 1968 | char nameb[BUFSIZ]; /* latest token name for funcdef & structdef */ |
| 1969 | register int tokoff; /* offset in line of start of latest token */ |
| 1970 | register int toklen; /* length of latest token */ |
| 1971 | int cblev; /* current curly brace level */ |
| 1972 | int parlev; /* current parenthesis level */ |
| 1973 | logical incomm, inquote, inchar, quotednl, midtoken; |
| 1974 | logical cplpl; |
| 1975 | TOKEN savetok; /* saved token during preprocessor handling */ |
| 1976 | char savenameb[BUFSIZ]; /* ouch! */ |
| 1977 | |
| 1978 | savetok.lineno = 0; |
| 1979 | curndx = newndx = 0; |
| 1980 | lineno = 0; |
| 1981 | charno = 0; |
| 1982 | lp = curlb.buffer; |
| 1983 | *lp = 0; |
| 1984 | |
| 1985 | definedef = dnone; funcdef = fnone; typdef = tnone; structdef = snone; |
| 1986 | next_token_is_func = yacc_rules = FALSE; |
| 1987 | midtoken = inquote = inchar = incomm = quotednl = FALSE; |
| 1988 | cblev = 0; |
| 1989 | parlev = 0; |
| 1990 | cplpl = c_ext & C_PLPL; |
| 1991 | |
| 1992 | while (!feof (inf)) |
| 1993 | { |
| 1994 | c = *lp++; |
| 1995 | if (c == '\\') |
| 1996 | { |
| 1997 | /* If we're at the end of the line, the next character is a |
| 1998 | '\0'; don't skip it, because it's the thing that tells us |
| 1999 | to read the next line. */ |
| 2000 | if (*lp == '\0') |
| 2001 | { |
| 2002 | quotednl = TRUE; |
| 2003 | continue; |
| 2004 | } |
| 2005 | lp++; |
| 2006 | c = ' '; |
| 2007 | } |
| 2008 | else if (incomm) |
| 2009 | { |
| 2010 | switch (c) |
| 2011 | { |
| 2012 | case '*': |
| 2013 | if (*lp == '/') |
| 2014 | { |
| 2015 | c = *lp++; |
| 2016 | incomm = FALSE; |
| 2017 | } |
| 2018 | break; |
| 2019 | case '\0': |
| 2020 | /* Newlines inside comments do not end macro definitions in |
| 2021 | traditional cpp. */ |
| 2022 | CNL_SAVE_DEFINEDEF; |
| 2023 | break; |
| 2024 | } |
| 2025 | continue; |
| 2026 | } |
| 2027 | else if (inquote) |
| 2028 | { |
| 2029 | switch (c) |
| 2030 | { |
| 2031 | case '"': |
| 2032 | inquote = FALSE; |
| 2033 | break; |
| 2034 | case '\0': |
| 2035 | /* Newlines inside strings do not end macro definitions |
| 2036 | in traditional cpp, even though compilers don't |
| 2037 | usually accept them. */ |
| 2038 | CNL_SAVE_DEFINEDEF; |
| 2039 | break; |
| 2040 | } |
| 2041 | continue; |
| 2042 | } |
| 2043 | else if (inchar) |
| 2044 | { |
| 2045 | switch (c) |
| 2046 | { |
| 2047 | case '\0': |
| 2048 | /* Hmmm, something went wrong. */ |
| 2049 | CNL; |
| 2050 | /* FALLTHRU */ |
| 2051 | case '\'': |
| 2052 | inchar = FALSE; |
| 2053 | break; |
| 2054 | } |
| 2055 | continue; |
| 2056 | } |
| 2057 | else |
| 2058 | switch (c) |
| 2059 | { |
| 2060 | case '"': |
| 2061 | inquote = TRUE; |
| 2062 | if (funcdef != finlist && funcdef != fignore) |
| 2063 | funcdef = fnone; |
| 2064 | continue; |
| 2065 | case '\'': |
| 2066 | inchar = TRUE; |
| 2067 | if (funcdef != finlist && funcdef != fignore) |
| 2068 | funcdef = fnone; |
| 2069 | continue; |
| 2070 | case '/': |
| 2071 | if (*lp == '*') |
| 2072 | { |
| 2073 | lp++; |
| 2074 | incomm = TRUE; |
| 2075 | continue; |
| 2076 | } |
| 2077 | else if (cplpl && *lp == '/') |
| 2078 | { |
| 2079 | c = 0; |
| 2080 | break; |
| 2081 | } |
| 2082 | else |
| 2083 | break; |
| 2084 | case '%': |
| 2085 | if ((c_ext & YACC) && *lp == '%') |
| 2086 | { |
| 2087 | /* entering or exiting rules section in yacc file */ |
| 2088 | lp++; |
| 2089 | definedef = dnone; funcdef = fnone; |
| 2090 | typdef = tnone; structdef = snone; |
| 2091 | next_token_is_func = FALSE; |
| 2092 | midtoken = inquote = inchar = incomm = quotednl = FALSE; |
| 2093 | cblev = 0; |
| 2094 | yacc_rules = !yacc_rules; |
| 2095 | continue; |
| 2096 | } |
| 2097 | else |
| 2098 | break; |
| 2099 | case '#': |
| 2100 | if (definedef == dnone) |
| 2101 | { |
| 2102 | char *cp; |
| 2103 | logical cpptoken = TRUE; |
| 2104 | |
| 2105 | /* Look back on this line. If all blanks, or nonblanks |
| 2106 | followed by an end of comment, this is a preprocessor |
| 2107 | token. */ |
| 2108 | for (cp = newlb.buffer; cp < lp-1; cp++) |
| 2109 | if (!iswhite (*cp)) |
| 2110 | { |
| 2111 | if (*cp == '*' && *(cp+1) == '/') |
| 2112 | { |
| 2113 | cp++; |
| 2114 | cpptoken = TRUE; |
| 2115 | } |
| 2116 | else |
| 2117 | cpptoken = FALSE; |
| 2118 | } |
| 2119 | if (cpptoken) |
| 2120 | definedef = dsharpseen; |
| 2121 | } /* if (definedef == dnone) */ |
| 2122 | |
| 2123 | continue; |
| 2124 | } /* switch (c) */ |
| 2125 | |
| 2126 | |
| 2127 | /* Consider token only if some complicated conditions are satisfied. */ |
| 2128 | if ((definedef != dnone |
| 2129 | || (cblev == 0 && structdef != scolonseen) |
| 2130 | || (cblev == 1 && cplpl && structdef == sinbody)) |
| 2131 | && typdef != tignore |
| 2132 | && definedef != dignorerest |
| 2133 | && funcdef != finlist) |
| 2134 | { |
| 2135 | if (midtoken) |
| 2136 | { |
| 2137 | if (endtoken (c)) |
| 2138 | { |
| 2139 | if (cplpl && c == ':' && *lp == ':' && begtoken(*(lp + 1))) |
| 2140 | { |
| 2141 | /* |
| 2142 | * This handles :: in the middle, but not at the |
| 2143 | * beginning of an identifier. |
| 2144 | */ |
| 2145 | lp += 2; |
| 2146 | toklen += 3; |
| 2147 | } |
| 2148 | else |
| 2149 | { |
| 2150 | logical is_func = FALSE; |
| 2151 | |
| 2152 | tok.lineno = lineno; |
| 2153 | tok.p = newlb.buffer + tokoff; |
| 2154 | tok.len = toklen; |
| 2155 | tok.named = FALSE; |
| 2156 | if (yacc_rules |
| 2157 | || consider_token (c, &tok, c_ext, cblev, &is_func)) |
| 2158 | { |
| 2159 | if (structdef == sinbody |
| 2160 | && definedef == dnone |
| 2161 | && is_func) |
| 2162 | /* function defined in C++ class body */ |
| 2163 | { |
| 2164 | sprintf (nameb, "%s::%.*s", |
| 2165 | ((structtag[0] == '\0') |
| 2166 | ? "_anonymous_" : structtag), |
| 2167 | tok.len, tok.p); |
| 2168 | tok.named = TRUE; |
| 2169 | } |
| 2170 | else |
| 2171 | { |
| 2172 | sprintf (nameb, "%.*s", tok.len, tok.p); |
| 2173 | } |
| 2174 | |
| 2175 | if (structdef == stagseen |
| 2176 | || typdef == tend |
| 2177 | || (is_func |
| 2178 | && definedef == dignorerest)) /* macro */ |
| 2179 | tok.named = TRUE; |
| 2180 | |
| 2181 | if (definedef == dnone |
| 2182 | && (funcdef == ftagseen |
| 2183 | || structdef == stagseen |
| 2184 | || typdef == tend)) |
| 2185 | { |
| 2186 | if (newndx == curndx) |
| 2187 | curndx = 1 - curndx; /* switch line buffers */ |
| 2188 | } |
| 2189 | else |
| 2190 | MAKE_TAG_FROM_NEW_LB (is_func); |
| 2191 | } |
| 2192 | midtoken = FALSE; |
| 2193 | } |
| 2194 | } /* if (endtoken (c)) */ |
| 2195 | else if (intoken (c)) |
| 2196 | { |
| 2197 | toklen++; |
| 2198 | continue; |
| 2199 | } |
| 2200 | } /* if (midtoken) */ |
| 2201 | else if (begtoken (c)) |
| 2202 | { |
| 2203 | switch (definedef) |
| 2204 | { |
| 2205 | case dnone: |
| 2206 | switch (funcdef) |
| 2207 | { |
| 2208 | case fstartlist: |
| 2209 | funcdef = finlist; |
| 2210 | continue; |
| 2211 | case flistseen: |
| 2212 | MAKE_TAG_FROM_OTH_LB (TRUE); |
| 2213 | funcdef = fignore; |
| 2214 | break; |
| 2215 | case ftagseen: |
| 2216 | funcdef = fnone; |
| 2217 | break; |
| 2218 | } |
| 2219 | if (structdef == stagseen) |
| 2220 | structdef = snone; |
| 2221 | break; |
| 2222 | case dsharpseen: |
| 2223 | /* Take a quick peek ahead for a define directive, |
| 2224 | so we can avoid saving the token when not absolutely |
| 2225 | necessary. [This is a speed hack.] */ |
| 2226 | if (c == 'd' && strneq (lp, "efine", 5) |
| 2227 | && iswhite (*(lp + 5))) |
| 2228 | { |
| 2229 | SAVE_TOKEN; |
| 2230 | definedef = ddefineseen; |
| 2231 | lp += 6; |
| 2232 | } |
| 2233 | else |
| 2234 | definedef = dignorerest; |
| 2235 | continue; |
| 2236 | } |
| 2237 | if (!yacc_rules || lp == newlb.buffer + 1) |
| 2238 | { |
| 2239 | tokoff = lp - 1 - newlb.buffer; |
| 2240 | toklen = 1; |
| 2241 | midtoken = TRUE; |
| 2242 | } |
| 2243 | continue; |
| 2244 | } |
| 2245 | } /* if must look at token */ |
| 2246 | |
| 2247 | |
| 2248 | /* Detect end of line, colon, comma, semicolon and various braces |
| 2249 | after having handled a token.*/ |
| 2250 | switch (c) |
| 2251 | { |
| 2252 | case ':': |
| 2253 | if (definedef != dnone) |
| 2254 | break; |
| 2255 | if (structdef == stagseen) |
| 2256 | structdef = scolonseen; |
| 2257 | else |
| 2258 | switch (funcdef) |
| 2259 | { |
| 2260 | case ftagseen: |
| 2261 | if (yacc_rules) |
| 2262 | { |
| 2263 | MAKE_TAG_FROM_OTH_LB (FALSE); |
| 2264 | funcdef = fignore; |
| 2265 | } |
| 2266 | break; |
| 2267 | case fstartlist: |
| 2268 | funcdef = fnone; |
| 2269 | break; |
| 2270 | } |
| 2271 | break; |
| 2272 | case ';': |
| 2273 | if (definedef != dnone) |
| 2274 | break; |
| 2275 | if (cblev == 0) |
| 2276 | switch (typdef) |
| 2277 | { |
| 2278 | case tend: |
| 2279 | MAKE_TAG_FROM_OTH_LB (FALSE); |
| 2280 | /* FALLTHRU */ |
| 2281 | default: |
| 2282 | typdef = tnone; |
| 2283 | } |
| 2284 | if (funcdef != fignore) |
| 2285 | funcdef = fnone; |
| 2286 | if (structdef == stagseen) |
| 2287 | structdef = snone; |
| 2288 | break; |
| 2289 | case ',': |
| 2290 | if (definedef != dnone) |
| 2291 | break; |
| 2292 | if (funcdef != finlist && funcdef != fignore) |
| 2293 | funcdef = fnone; |
| 2294 | if (structdef == stagseen) |
| 2295 | structdef = snone; |
| 2296 | break; |
| 2297 | case '[': |
| 2298 | if (definedef != dnone) |
| 2299 | break; |
| 2300 | if (cblev == 0 && typdef == tend) |
| 2301 | { |
| 2302 | typdef = tignore; |
| 2303 | MAKE_TAG_FROM_OTH_LB (FALSE); |
| 2304 | break; |
| 2305 | } |
| 2306 | if (funcdef != finlist && funcdef != fignore) |
| 2307 | funcdef = fnone; |
| 2308 | if (structdef == stagseen) |
| 2309 | structdef = snone; |
| 2310 | break; |
| 2311 | case '(': |
| 2312 | if (definedef != dnone) |
| 2313 | break; |
| 2314 | switch (funcdef) |
| 2315 | { |
| 2316 | case fnone: |
| 2317 | switch (typdef) |
| 2318 | { |
| 2319 | case ttypedseen: |
| 2320 | case tend: |
| 2321 | /* Make sure that the next char is not a '*'. |
| 2322 | This handles constructs like: |
| 2323 | typedef void OperatorFun (int fun); */ |
| 2324 | if (*lp != '*') |
| 2325 | { |
| 2326 | typdef = tignore; |
| 2327 | MAKE_TAG_FROM_OTH_LB (FALSE); |
| 2328 | } |
| 2329 | break; |
| 2330 | } /* switch (typdef) */ |
| 2331 | break; |
| 2332 | case ftagseen: |
| 2333 | funcdef = fstartlist; |
| 2334 | break; |
| 2335 | case flistseen: |
| 2336 | funcdef = finlist; |
| 2337 | break; |
| 2338 | } |
| 2339 | parlev++; |
| 2340 | break; |
| 2341 | case ')': |
| 2342 | if (definedef != dnone) |
| 2343 | break; |
| 2344 | if (--parlev == 0) |
| 2345 | { |
| 2346 | switch (funcdef) |
| 2347 | { |
| 2348 | case fstartlist: |
| 2349 | case finlist: |
| 2350 | funcdef = flistseen; |
| 2351 | break; |
| 2352 | } |
| 2353 | if (cblev == 0 && typdef == tend) |
| 2354 | { |
| 2355 | typdef = tignore; |
| 2356 | MAKE_TAG_FROM_OTH_LB (FALSE); |
| 2357 | } |
| 2358 | } |
| 2359 | else if (parlev < 0) /* can happen due to ill-conceived #if's. */ |
| 2360 | parlev = 0; |
| 2361 | break; |
| 2362 | case '{': |
| 2363 | if (definedef != dnone) |
| 2364 | break; |
| 2365 | if (typdef == ttypedseen) |
| 2366 | typdef = tinbody; |
| 2367 | switch (structdef) |
| 2368 | { |
| 2369 | case skeyseen: /* unnamed struct */ |
| 2370 | structtag[0] = '\0'; |
| 2371 | structdef = sinbody; |
| 2372 | break; |
| 2373 | case stagseen: |
| 2374 | case scolonseen: /* named struct */ |
| 2375 | structdef = sinbody; |
| 2376 | MAKE_TAG_FROM_OTH_LB (FALSE); |
| 2377 | break; |
| 2378 | } |
| 2379 | switch (funcdef) |
| 2380 | { |
| 2381 | case flistseen: |
| 2382 | MAKE_TAG_FROM_OTH_LB (TRUE); |
| 2383 | /* FALLTHRU */ |
| 2384 | case fignore: |
| 2385 | funcdef = fnone; |
| 2386 | break; |
| 2387 | case fnone: |
| 2388 | /* Neutralize `extern "C" {' grot and look inside structs. */ |
| 2389 | if (cblev == 0 && structdef == snone && typdef == tnone) |
| 2390 | cblev = -1; |
| 2391 | } |
| 2392 | cblev++; |
| 2393 | break; |
| 2394 | case '*': |
| 2395 | if (definedef != dnone) |
| 2396 | break; |
| 2397 | if (funcdef == fstartlist) |
| 2398 | funcdef = fnone; /* avoid tagging `foo' in `foo (*bar()) ()' */ |
| 2399 | break; |
| 2400 | case '}': |
| 2401 | if (definedef != dnone) |
| 2402 | break; |
| 2403 | if (!noindentypedefs && lp == newlb.buffer + 1) |
| 2404 | { |
| 2405 | cblev = 0; /* reset curly brace level if first column */ |
| 2406 | parlev = 0; /* also reset paren level, just in case... */ |
| 2407 | } |
| 2408 | else if (cblev > 0) |
| 2409 | cblev--; |
| 2410 | if (cblev == 0) |
| 2411 | { |
| 2412 | if (typdef == tinbody) |
| 2413 | typdef = tend; |
| 2414 | structdef = snone; |
| 2415 | strcpy (structtag, "<error 2>"); |
| 2416 | } |
| 2417 | break; |
| 2418 | case '=': |
| 2419 | case '#': case '+': case '-': case '~': case '&': case '%': case '/': |
| 2420 | case '|': case '^': case '!': case '<': case '>': case '.': case '?': |
| 2421 | if (definedef != dnone) |
| 2422 | break; |
| 2423 | /* These surely cannot follow a function tag. */ |
| 2424 | if (funcdef != finlist && funcdef != fignore) |
| 2425 | funcdef = fnone; |
| 2426 | break; |
| 2427 | case '\0': |
| 2428 | /* If a macro spans multiple lines don't reset its state. */ |
| 2429 | if (quotednl) |
| 2430 | CNL_SAVE_DEFINEDEF; |
| 2431 | else |
| 2432 | CNL; |
| 2433 | break; |
| 2434 | } /* switch (c) */ |
| 2435 | |
| 2436 | } /* while not eof */ |
| 2437 | } |
| 2438 | |
| 2439 | /* |
| 2440 | * Process either a C++ file or a C file depending on the setting |
| 2441 | * of a global flag. |
| 2442 | */ |
| 2443 | void |
| 2444 | default_C_entries (inf) |
| 2445 | FILE *inf; |
| 2446 | { |
| 2447 | C_entries (cplusplus ? C_PLPL : 0, inf); |
| 2448 | } |
| 2449 | |
| 2450 | /* Always do C++. */ |
| 2451 | void |
| 2452 | Cplusplus_entries (inf) |
| 2453 | FILE *inf; |
| 2454 | { |
| 2455 | C_entries (C_PLPL, inf); |
| 2456 | } |
| 2457 | |
| 2458 | /* Always do C*. */ |
| 2459 | void |
| 2460 | Cstar_entries (inf) |
| 2461 | FILE *inf; |
| 2462 | { |
| 2463 | C_entries (C_STAR, inf); |
| 2464 | } |
| 2465 | |
| 2466 | /* Always do Yacc. */ |
| 2467 | void |
| 2468 | Yacc_entries (inf) |
| 2469 | FILE *inf; |
| 2470 | { |
| 2471 | C_entries (YACC, inf); |
| 2472 | } |
| 2473 | \f |
| 2474 | /* Fortran parsing */ |
| 2475 | |
| 2476 | char *dbp; |
| 2477 | int pfcnt; |
| 2478 | |
| 2479 | logical |
| 2480 | tail (cp) |
| 2481 | char *cp; |
| 2482 | { |
| 2483 | register int len = 0; |
| 2484 | |
| 2485 | while (*cp && (*cp | ' ') == (dbp[len] | ' ')) |
| 2486 | cp++, len++; |
| 2487 | if (*cp == 0) |
| 2488 | { |
| 2489 | dbp += len; |
| 2490 | return TRUE; |
| 2491 | } |
| 2492 | return FALSE; |
| 2493 | } |
| 2494 | |
| 2495 | void |
| 2496 | takeprec () |
| 2497 | { |
| 2498 | while (isspace (*dbp)) |
| 2499 | dbp++; |
| 2500 | if (*dbp != '*') |
| 2501 | return; |
| 2502 | dbp++; |
| 2503 | while (isspace (*dbp)) |
| 2504 | dbp++; |
| 2505 | if (tail ("(*)")) |
| 2506 | return; |
| 2507 | if (!isdigit (*dbp)) |
| 2508 | { |
| 2509 | --dbp; /* force failure */ |
| 2510 | return; |
| 2511 | } |
| 2512 | do |
| 2513 | dbp++; |
| 2514 | while (isdigit (*dbp)); |
| 2515 | } |
| 2516 | |
| 2517 | void |
| 2518 | getit (inf) |
| 2519 | FILE *inf; |
| 2520 | { |
| 2521 | register char *cp; |
| 2522 | char c; |
| 2523 | char nambuf[BUFSIZ]; |
| 2524 | |
| 2525 | while (isspace (*dbp)) |
| 2526 | dbp++; |
| 2527 | if (*dbp == '\0') |
| 2528 | { |
| 2529 | lineno++; |
| 2530 | linecharno = charno; |
| 2531 | charno += readline (&lb, inf); |
| 2532 | dbp = lb.buffer; |
| 2533 | if (dbp[5] != '&') |
| 2534 | return; |
| 2535 | dbp += 6; |
| 2536 | while (isspace (*dbp)) |
| 2537 | dbp++; |
| 2538 | } |
| 2539 | if (!isalpha (*dbp) |
| 2540 | && *dbp != '_' |
| 2541 | && *dbp != '$') |
| 2542 | return; |
| 2543 | for (cp = dbp + 1; |
| 2544 | (*cp |
| 2545 | && (isalpha (*cp) || isdigit (*cp) || (*cp == '_') || (*cp == '$'))); |
| 2546 | cp++) |
| 2547 | continue; |
| 2548 | c = *cp; |
| 2549 | *cp = '\0'; |
| 2550 | strcpy (nambuf, dbp); |
| 2551 | *cp = c; |
| 2552 | pfnote (nambuf, TRUE, FALSE, lb.buffer, |
| 2553 | cp - lb.buffer + 1, lineno, linecharno); |
| 2554 | pfcnt++; |
| 2555 | } |
| 2556 | |
| 2557 | void |
| 2558 | Fortran_functions (inf) |
| 2559 | FILE *inf; |
| 2560 | { |
| 2561 | lineno = 0; |
| 2562 | charno = 0; |
| 2563 | pfcnt = 0; |
| 2564 | |
| 2565 | while (!feof (inf)) |
| 2566 | { |
| 2567 | lineno++; |
| 2568 | linecharno = charno; |
| 2569 | charno += readline (&lb, inf); |
| 2570 | dbp = lb.buffer; |
| 2571 | if (*dbp == '%') |
| 2572 | dbp++; /* Ratfor escape to fortran */ |
| 2573 | while (isspace (*dbp)) |
| 2574 | dbp++; |
| 2575 | if (*dbp == 0) |
| 2576 | continue; |
| 2577 | switch (*dbp | ' ') |
| 2578 | { |
| 2579 | case 'i': |
| 2580 | if (tail ("integer")) |
| 2581 | takeprec (); |
| 2582 | break; |
| 2583 | case 'r': |
| 2584 | if (tail ("real")) |
| 2585 | takeprec (); |
| 2586 | break; |
| 2587 | case 'l': |
| 2588 | if (tail ("logical")) |
| 2589 | takeprec (); |
| 2590 | break; |
| 2591 | case 'c': |
| 2592 | if (tail ("complex") || tail ("character")) |
| 2593 | takeprec (); |
| 2594 | break; |
| 2595 | case 'd': |
| 2596 | if (tail ("double")) |
| 2597 | { |
| 2598 | while (isspace (*dbp)) |
| 2599 | dbp++; |
| 2600 | if (*dbp == 0) |
| 2601 | continue; |
| 2602 | if (tail ("precision")) |
| 2603 | break; |
| 2604 | continue; |
| 2605 | } |
| 2606 | break; |
| 2607 | } |
| 2608 | while (isspace (*dbp)) |
| 2609 | dbp++; |
| 2610 | if (*dbp == 0) |
| 2611 | continue; |
| 2612 | switch (*dbp | ' ') |
| 2613 | { |
| 2614 | case 'f': |
| 2615 | if (tail ("function")) |
| 2616 | getit (inf); |
| 2617 | continue; |
| 2618 | case 's': |
| 2619 | if (tail ("subroutine")) |
| 2620 | getit (inf); |
| 2621 | continue; |
| 2622 | case 'e': |
| 2623 | if (tail ("entry")) |
| 2624 | getit (inf); |
| 2625 | continue; |
| 2626 | case 'p': |
| 2627 | if (tail ("program")) |
| 2628 | { |
| 2629 | getit (inf); |
| 2630 | continue; |
| 2631 | } |
| 2632 | if (tail ("procedure")) |
| 2633 | getit (inf); |
| 2634 | continue; |
| 2635 | } |
| 2636 | } |
| 2637 | } |
| 2638 | \f |
| 2639 | /* |
| 2640 | * Bob Weiner, Motorola Inc., 4/3/94 |
| 2641 | * Unix and microcontroller assembly tag handling |
| 2642 | * look for '^[a-zA-Z_.$][a-zA_Z0-9_.$]*[: ^I^J]' |
| 2643 | */ |
| 2644 | void |
| 2645 | Asm_labels (inf) |
| 2646 | FILE *inf; |
| 2647 | { |
| 2648 | char nambuf[BUFSIZ]; |
| 2649 | register char *cp; |
| 2650 | char c; |
| 2651 | |
| 2652 | lineno = 0; |
| 2653 | charno = 0; |
| 2654 | pfcnt = 0; |
| 2655 | |
| 2656 | while (!feof (inf)) |
| 2657 | { |
| 2658 | lineno++; |
| 2659 | linecharno = charno; |
| 2660 | charno += readline (&lb, inf); |
| 2661 | cp = lb.buffer; |
| 2662 | |
| 2663 | /* If first char is alphabetic or one of [_.$], test for colon |
| 2664 | following identifier. */ |
| 2665 | if (isalpha (*cp) || *cp == '_' || *cp == '.' || *cp == '$') |
| 2666 | { |
| 2667 | /* Read past label. */ |
| 2668 | cp++; |
| 2669 | while (isalnum (*cp) || *cp == '_' || *cp == '.' || *cp == '$') |
| 2670 | cp++; |
| 2671 | if (*cp == ':' || isspace (*cp)) |
| 2672 | { |
| 2673 | /* Found end of label, so copy it and add it to the table. */ |
| 2674 | c = *cp; |
| 2675 | *cp = '\0'; |
| 2676 | strcpy (nambuf, lb.buffer); |
| 2677 | *cp = c; |
| 2678 | pfnote (nambuf, TRUE, FALSE, lb.buffer, |
| 2679 | cp - lb.buffer + 1, lineno, linecharno); |
| 2680 | pfcnt++; |
| 2681 | } |
| 2682 | } |
| 2683 | } |
| 2684 | } |
| 2685 | \f |
| 2686 | /* Added by Mosur Mohan, 4/22/88 */ |
| 2687 | /* Pascal parsing */ |
| 2688 | |
| 2689 | #define GET_NEW_LINE \ |
| 2690 | { \ |
| 2691 | linecharno = charno; lineno++; \ |
| 2692 | charno += 1 + readline (&lb, inf); \ |
| 2693 | dbp = lb.buffer; \ |
| 2694 | } |
| 2695 | |
| 2696 | /* |
| 2697 | * Locates tags for procedures & functions. Doesn't do any type- or |
| 2698 | * var-definitions. It does look for the keyword "extern" or |
| 2699 | * "forward" immediately following the procedure statement; if found, |
| 2700 | * the tag is skipped. |
| 2701 | */ |
| 2702 | void |
| 2703 | Pascal_functions (inf) |
| 2704 | FILE *inf; |
| 2705 | { |
| 2706 | struct linebuffer tline; /* mostly copied from C_entries */ |
| 2707 | long save_lcno; |
| 2708 | int save_lineno; |
| 2709 | char c, *cp; |
| 2710 | char nambuf[BUFSIZ]; |
| 2711 | |
| 2712 | logical /* each of these flags is TRUE iff: */ |
| 2713 | incomment, /* point is inside a comment */ |
| 2714 | inquote, /* point is inside '..' string */ |
| 2715 | get_tagname, /* point is after PROCEDURE/FUNCTION */ |
| 2716 | /* keyword, so next item = potential tag */ |
| 2717 | found_tag, /* point is after a potential tag */ |
| 2718 | inparms, /* point is within parameter-list */ |
| 2719 | verify_tag; /* point has passed the parm-list, so the */ |
| 2720 | /* next token will determine whether */ |
| 2721 | /* this is a FORWARD/EXTERN to be */ |
| 2722 | /* ignored, or whether it is a real tag */ |
| 2723 | |
| 2724 | lineno = 0; |
| 2725 | charno = 0; |
| 2726 | dbp = lb.buffer; |
| 2727 | *dbp = 0; |
| 2728 | initbuffer (&tline); |
| 2729 | |
| 2730 | incomment = inquote = FALSE; |
| 2731 | found_tag = FALSE; /* have a proc name; check if extern */ |
| 2732 | get_tagname = FALSE; /* have found "procedure" keyword */ |
| 2733 | inparms = FALSE; /* found '(' after "proc" */ |
| 2734 | verify_tag = FALSE; /* check if "extern" is ahead */ |
| 2735 | |
| 2736 | /* long main loop to get next char */ |
| 2737 | while (!feof (inf)) |
| 2738 | { |
| 2739 | c = *dbp++; |
| 2740 | if (c == 0) /* if end of line */ |
| 2741 | { |
| 2742 | GET_NEW_LINE; |
| 2743 | if (*dbp == 0) |
| 2744 | continue; |
| 2745 | if (!((found_tag && verify_tag) || |
| 2746 | get_tagname)) |
| 2747 | c = *dbp++; /* only if don't need *dbp pointing */ |
| 2748 | /* to the beginning of the name of */ |
| 2749 | /* the procedure or function */ |
| 2750 | } |
| 2751 | if (incomment) |
| 2752 | { |
| 2753 | if (c == '}') /* within { - } comments */ |
| 2754 | incomment = FALSE; |
| 2755 | else if (c == '*' && dbp[1] == ')') /* within (* - *) comments */ |
| 2756 | { |
| 2757 | dbp++; |
| 2758 | incomment = FALSE; |
| 2759 | } |
| 2760 | continue; |
| 2761 | } |
| 2762 | else if (inquote) |
| 2763 | { |
| 2764 | if (c == '\'') |
| 2765 | inquote = FALSE; |
| 2766 | continue; |
| 2767 | } |
| 2768 | else |
| 2769 | switch (c) |
| 2770 | { |
| 2771 | case '\'': |
| 2772 | inquote = TRUE; /* found first quote */ |
| 2773 | continue; |
| 2774 | case '{': /* found open-{-comment */ |
| 2775 | incomment = TRUE; |
| 2776 | continue; |
| 2777 | case '(': |
| 2778 | if (*dbp == '*') /* found open-(*-comment */ |
| 2779 | { |
| 2780 | incomment = TRUE; |
| 2781 | dbp++; |
| 2782 | } |
| 2783 | else if (found_tag) /* found '(' after tag, i.e., parm-list */ |
| 2784 | inparms = TRUE; |
| 2785 | continue; |
| 2786 | case ')': /* end of parms list */ |
| 2787 | if (inparms) |
| 2788 | inparms = FALSE; |
| 2789 | continue; |
| 2790 | case ';': |
| 2791 | if ((found_tag) && (!inparms)) /* end of proc or fn stmt */ |
| 2792 | { |
| 2793 | verify_tag = TRUE; |
| 2794 | break; |
| 2795 | } |
| 2796 | continue; |
| 2797 | } |
| 2798 | if ((found_tag) && (verify_tag) && (*dbp != ' ')) |
| 2799 | { |
| 2800 | /* check if this is an "extern" declaration */ |
| 2801 | if (*dbp == 0) |
| 2802 | continue; |
| 2803 | if ((*dbp == 'e') || (*dbp == 'E')) |
| 2804 | { |
| 2805 | if (tail ("extern")) /* superfluous, really! */ |
| 2806 | { |
| 2807 | found_tag = FALSE; |
| 2808 | verify_tag = FALSE; |
| 2809 | } |
| 2810 | } |
| 2811 | else if ((*dbp == 'f') || (*dbp == 'F')) |
| 2812 | { |
| 2813 | if (tail ("forward")) /* check for forward reference */ |
| 2814 | { |
| 2815 | found_tag = FALSE; |
| 2816 | verify_tag = FALSE; |
| 2817 | } |
| 2818 | } |
| 2819 | if ((found_tag) && (verify_tag)) /* not external proc, so make tag */ |
| 2820 | { |
| 2821 | found_tag = FALSE; |
| 2822 | verify_tag = FALSE; |
| 2823 | pfnote (nambuf, TRUE, FALSE, |
| 2824 | tline.buffer, cp - tline.buffer + 1, |
| 2825 | save_lineno, save_lcno); |
| 2826 | continue; |
| 2827 | } |
| 2828 | } |
| 2829 | if (get_tagname) /* grab name of proc or fn */ |
| 2830 | { |
| 2831 | if (*dbp == 0) |
| 2832 | continue; |
| 2833 | |
| 2834 | /* save all values for later tagging */ |
| 2835 | tline.size = lb.size; |
| 2836 | strcpy (tline.buffer, lb.buffer); |
| 2837 | save_lineno = lineno; |
| 2838 | save_lcno = linecharno; |
| 2839 | |
| 2840 | /* grab block name */ |
| 2841 | for (cp = dbp + 1; *cp && (!endtoken (*cp)); cp++) |
| 2842 | continue; |
| 2843 | c = cp[0]; |
| 2844 | cp[0] = 0; |
| 2845 | strcpy (nambuf, dbp); |
| 2846 | cp[0] = c; |
| 2847 | dbp = cp; /* restore dbp to e-o-token */ |
| 2848 | get_tagname = FALSE; |
| 2849 | found_tag = TRUE; |
| 2850 | continue; |
| 2851 | |
| 2852 | /* and proceed to check for "extern" */ |
| 2853 | } |
| 2854 | if (!incomment && !inquote && !found_tag && !get_tagname) |
| 2855 | { |
| 2856 | /* check for proc/fn keywords */ |
| 2857 | switch (c | ' ') |
| 2858 | { |
| 2859 | case 'p': |
| 2860 | if (tail ("rocedure")) /* c = 'p', dbp has advanced */ |
| 2861 | get_tagname = TRUE; |
| 2862 | continue; |
| 2863 | case 'f': |
| 2864 | if (tail ("unction")) |
| 2865 | get_tagname = TRUE; |
| 2866 | continue; |
| 2867 | } |
| 2868 | } |
| 2869 | } /* while not eof */ |
| 2870 | } |
| 2871 | \f |
| 2872 | /* |
| 2873 | * lisp tag functions |
| 2874 | * just look for (def or (DEF |
| 2875 | */ |
| 2876 | |
| 2877 | int |
| 2878 | L_isdef (tokp) |
| 2879 | register char *tokp; |
| 2880 | { |
| 2881 | return ((tokp[1] == 'd' || tokp[1] == 'D') |
| 2882 | && (tokp[2] == 'e' || tokp[2] == 'E') |
| 2883 | && (tokp[3] == 'f' || tokp[3] == 'F')); |
| 2884 | } |
| 2885 | |
| 2886 | int |
| 2887 | L_isquote (tokp) |
| 2888 | register char *tokp; |
| 2889 | { |
| 2890 | return ((*(++tokp) == 'q' || *tokp == 'Q') |
| 2891 | && (*(++tokp) == 'u' || *tokp == 'U') |
| 2892 | && (*(++tokp) == 'o' || *tokp == 'O') |
| 2893 | && (*(++tokp) == 't' || *tokp == 'T') |
| 2894 | && (*(++tokp) == 'e' || *tokp == 'E') |
| 2895 | && isspace(*(++tokp))); |
| 2896 | } |
| 2897 | |
| 2898 | void |
| 2899 | L_getit () |
| 2900 | { |
| 2901 | register char *cp; |
| 2902 | char c; |
| 2903 | char nambuf[BUFSIZ]; |
| 2904 | |
| 2905 | if (*dbp == '\'') /* Skip prefix quote */ |
| 2906 | dbp++; |
| 2907 | else if (*dbp == '(' && L_isquote (dbp)) /* Skip "(quote " */ |
| 2908 | { |
| 2909 | dbp += 7; |
| 2910 | while (isspace(*dbp)) |
| 2911 | dbp++; |
| 2912 | } |
| 2913 | for (cp = dbp /*+1*/; *cp && *cp != '(' && *cp != ' ' && *cp != ')'; cp++) |
| 2914 | continue; |
| 2915 | if (cp == dbp) |
| 2916 | return; |
| 2917 | |
| 2918 | c = cp[0]; |
| 2919 | cp[0] = 0; |
| 2920 | strcpy (nambuf, dbp); |
| 2921 | cp[0] = c; |
| 2922 | pfnote (nambuf, TRUE, FALSE, lb.buffer, |
| 2923 | cp - lb.buffer + 1, lineno, linecharno); |
| 2924 | pfcnt++; |
| 2925 | } |
| 2926 | |
| 2927 | void |
| 2928 | Lisp_functions (inf) |
| 2929 | FILE *inf; |
| 2930 | { |
| 2931 | lineno = 0; |
| 2932 | charno = 0; |
| 2933 | pfcnt = 0; |
| 2934 | |
| 2935 | while (!feof (inf)) |
| 2936 | { |
| 2937 | lineno++; |
| 2938 | linecharno = charno; |
| 2939 | charno += readline (&lb, inf); |
| 2940 | dbp = lb.buffer; |
| 2941 | if (dbp[0] == '(') |
| 2942 | { |
| 2943 | if (L_isdef (dbp)) |
| 2944 | { |
| 2945 | while (!isspace (*dbp)) |
| 2946 | dbp++; |
| 2947 | while (isspace (*dbp)) |
| 2948 | dbp++; |
| 2949 | L_getit (); |
| 2950 | } |
| 2951 | else |
| 2952 | { |
| 2953 | /* Check for (foo::defmumble name-defined ... */ |
| 2954 | do |
| 2955 | dbp++; |
| 2956 | while (*dbp && !isspace (*dbp) |
| 2957 | && *dbp != ':' && *dbp != '(' && *dbp != ')'); |
| 2958 | if (*dbp == ':') |
| 2959 | { |
| 2960 | do |
| 2961 | dbp++; |
| 2962 | while (*dbp == ':'); |
| 2963 | |
| 2964 | if (L_isdef (dbp - 1)) |
| 2965 | { |
| 2966 | while (!isspace (*dbp)) |
| 2967 | dbp++; |
| 2968 | while (isspace (*dbp)) |
| 2969 | dbp++; |
| 2970 | L_getit (); |
| 2971 | } |
| 2972 | } |
| 2973 | } |
| 2974 | } |
| 2975 | } |
| 2976 | } |
| 2977 | \f |
| 2978 | /* |
| 2979 | * Scheme tag functions |
| 2980 | * look for (def... xyzzy |
| 2981 | * look for (def... (xyzzy |
| 2982 | * look for (def ... ((...(xyzzy .... |
| 2983 | * look for (set! xyzzy |
| 2984 | */ |
| 2985 | |
| 2986 | void get_scheme (); |
| 2987 | |
| 2988 | void |
| 2989 | Scheme_functions (inf) |
| 2990 | FILE *inf; |
| 2991 | { |
| 2992 | lineno = 0; |
| 2993 | charno = 0; |
| 2994 | pfcnt = 0; |
| 2995 | |
| 2996 | while (!feof (inf)) |
| 2997 | { |
| 2998 | lineno++; |
| 2999 | linecharno = charno; |
| 3000 | charno += readline (&lb, inf); |
| 3001 | dbp = lb.buffer; |
| 3002 | if (dbp[0] == '(' && |
| 3003 | (dbp[1] == 'D' || dbp[1] == 'd') && |
| 3004 | (dbp[2] == 'E' || dbp[2] == 'e') && |
| 3005 | (dbp[3] == 'F' || dbp[3] == 'f')) |
| 3006 | { |
| 3007 | while (!isspace (*dbp)) |
| 3008 | dbp++; |
| 3009 | /* Skip over open parens and white space */ |
| 3010 | while (*dbp && (isspace (*dbp) || *dbp == '(')) |
| 3011 | dbp++; |
| 3012 | get_scheme (); |
| 3013 | } |
| 3014 | if (dbp[0] == '(' && |
| 3015 | (dbp[1] == 'S' || dbp[1] == 's') && |
| 3016 | (dbp[2] == 'E' || dbp[2] == 'e') && |
| 3017 | (dbp[3] == 'T' || dbp[3] == 't') && |
| 3018 | (dbp[4] == '!' || dbp[4] == '!') && |
| 3019 | (isspace (dbp[5]))) |
| 3020 | { |
| 3021 | while (!isspace (*dbp)) |
| 3022 | dbp++; |
| 3023 | /* Skip over white space */ |
| 3024 | while (isspace (*dbp)) |
| 3025 | dbp++; |
| 3026 | get_scheme (); |
| 3027 | } |
| 3028 | } |
| 3029 | } |
| 3030 | |
| 3031 | void |
| 3032 | get_scheme () |
| 3033 | { |
| 3034 | register char *cp; |
| 3035 | char c; |
| 3036 | char nambuf[BUFSIZ]; |
| 3037 | |
| 3038 | if (*dbp == 0) |
| 3039 | return; |
| 3040 | /* Go till you get to white space or a syntactic break */ |
| 3041 | for (cp = dbp + 1; *cp && *cp != '(' && *cp != ')' && !isspace (*cp); cp++) |
| 3042 | continue; |
| 3043 | /* Null terminate the string there. */ |
| 3044 | c = cp[0]; |
| 3045 | cp[0] = 0; |
| 3046 | /* Copy the string */ |
| 3047 | strcpy (nambuf, dbp); |
| 3048 | /* Unterminate the string */ |
| 3049 | cp[0] = c; |
| 3050 | /* Announce the change */ |
| 3051 | pfnote (nambuf, TRUE, FALSE, lb.buffer, cp - lb.buffer + 1, lineno, linecharno); |
| 3052 | pfcnt++; |
| 3053 | } |
| 3054 | \f |
| 3055 | /* Find tags in TeX and LaTeX input files. */ |
| 3056 | |
| 3057 | /* TEX_toktab is a table of TeX control sequences that define tags. |
| 3058 | Each TEX_tabent records one such control sequence. |
| 3059 | CONVERT THIS TO USE THE Stab TYPE!! */ |
| 3060 | struct TEX_tabent |
| 3061 | { |
| 3062 | char *name; |
| 3063 | int len; |
| 3064 | }; |
| 3065 | |
| 3066 | struct TEX_tabent *TEX_toktab = NULL; /* Table with tag tokens */ |
| 3067 | |
| 3068 | /* Default set of control sequences to put into TEX_toktab. |
| 3069 | The value of environment var TEXTAGS is prepended to this. */ |
| 3070 | |
| 3071 | char *TEX_defenv = "\ |
| 3072 | :chapter:section:subsection:subsubsection:eqno:label:ref:cite:bibitem:typeout"; |
| 3073 | |
| 3074 | void TEX_mode (); |
| 3075 | struct TEX_tabent *TEX_decode_env (); |
| 3076 | void TEX_getit (); |
| 3077 | int TEX_Token (); |
| 3078 | |
| 3079 | char TEX_esc = '\\'; |
| 3080 | char TEX_opgrp = '{'; |
| 3081 | char TEX_clgrp = '}'; |
| 3082 | |
| 3083 | /* |
| 3084 | * TeX/LaTeX scanning loop. |
| 3085 | */ |
| 3086 | void |
| 3087 | TeX_functions (inf) |
| 3088 | FILE *inf; |
| 3089 | { |
| 3090 | char *lasthit; |
| 3091 | |
| 3092 | lineno = 0; |
| 3093 | charno = 0; |
| 3094 | pfcnt = 0; |
| 3095 | |
| 3096 | /* Select either \ or ! as escape character. */ |
| 3097 | TEX_mode (inf); |
| 3098 | |
| 3099 | /* Initialize token table once from environment. */ |
| 3100 | if (!TEX_toktab) |
| 3101 | TEX_toktab = TEX_decode_env ("TEXTAGS", TEX_defenv); |
| 3102 | |
| 3103 | while (!feof (inf)) |
| 3104 | { /* Scan each line in file */ |
| 3105 | lineno++; |
| 3106 | linecharno = charno; |
| 3107 | charno += readline (&lb, inf); |
| 3108 | dbp = lb.buffer; |
| 3109 | lasthit = dbp; |
| 3110 | while (dbp = etags_strchr (dbp, TEX_esc)) /* Look at each esc in line */ |
| 3111 | { |
| 3112 | register int i; |
| 3113 | |
| 3114 | if (!*(++dbp)) |
| 3115 | break; |
| 3116 | linecharno += dbp - lasthit; |
| 3117 | lasthit = dbp; |
| 3118 | i = TEX_Token (lasthit); |
| 3119 | if (0 <= i) |
| 3120 | { |
| 3121 | TEX_getit (lasthit, TEX_toktab[i].len); |
| 3122 | break; /* We only save a line once */ |
| 3123 | } |
| 3124 | } |
| 3125 | } |
| 3126 | } |
| 3127 | |
| 3128 | #define TEX_LESC '\\' |
| 3129 | #define TEX_SESC '!' |
| 3130 | #define TEX_cmt '%' |
| 3131 | |
| 3132 | /* Figure out whether TeX's escapechar is '\\' or '!' and set grouping |
| 3133 | chars accordingly. */ |
| 3134 | void |
| 3135 | TEX_mode (inf) |
| 3136 | FILE *inf; |
| 3137 | { |
| 3138 | int c; |
| 3139 | |
| 3140 | while ((c = getc (inf)) != EOF) |
| 3141 | { |
| 3142 | /* Skip to next line if we hit the TeX comment char. */ |
| 3143 | if (c == TEX_cmt) |
| 3144 | while (c != '\n') |
| 3145 | c = getc (inf); |
| 3146 | else if (c == TEX_LESC || c == TEX_SESC ) |
| 3147 | break; |
| 3148 | } |
| 3149 | |
| 3150 | if (c == TEX_LESC) |
| 3151 | { |
| 3152 | TEX_esc = TEX_LESC; |
| 3153 | TEX_opgrp = '{'; |
| 3154 | TEX_clgrp = '}'; |
| 3155 | } |
| 3156 | else |
| 3157 | { |
| 3158 | TEX_esc = TEX_SESC; |
| 3159 | TEX_opgrp = '<'; |
| 3160 | TEX_clgrp = '>'; |
| 3161 | } |
| 3162 | rewind (inf); |
| 3163 | } |
| 3164 | |
| 3165 | /* Read environment and prepend it to the default string. |
| 3166 | Build token table. */ |
| 3167 | struct TEX_tabent * |
| 3168 | TEX_decode_env (evarname, defenv) |
| 3169 | char *evarname; |
| 3170 | char *defenv; |
| 3171 | { |
| 3172 | register char *env, *p; |
| 3173 | |
| 3174 | struct TEX_tabent *tab; |
| 3175 | int size, i; |
| 3176 | |
| 3177 | /* Append default string to environment. */ |
| 3178 | env = getenv (evarname); |
| 3179 | if (!env) |
| 3180 | env = defenv; |
| 3181 | else |
| 3182 | env = concat (env, defenv, ""); |
| 3183 | |
| 3184 | /* Allocate a token table */ |
| 3185 | for (size = 1, p = env; p;) |
| 3186 | if ((p = etags_strchr (p, ':')) && *(++p)) |
| 3187 | size++; |
| 3188 | /* Add 1 to leave room for null terminator. */ |
| 3189 | tab = xnew (size + 1, struct TEX_tabent); |
| 3190 | |
| 3191 | /* Unpack environment string into token table. Be careful about */ |
| 3192 | /* zero-length strings (leading ':', "::" and trailing ':') */ |
| 3193 | for (i = 0; *env;) |
| 3194 | { |
| 3195 | p = etags_strchr (env, ':'); |
| 3196 | if (!p) /* End of environment string. */ |
| 3197 | p = env + strlen (env); |
| 3198 | if (p - env > 0) |
| 3199 | { /* Only non-zero strings. */ |
| 3200 | tab[i].name = savenstr (env, p - env); |
| 3201 | tab[i].len = strlen (tab[i].name); |
| 3202 | i++; |
| 3203 | } |
| 3204 | if (*p) |
| 3205 | env = p + 1; |
| 3206 | else |
| 3207 | { |
| 3208 | tab[i].name = NULL; /* Mark end of table. */ |
| 3209 | tab[i].len = 0; |
| 3210 | break; |
| 3211 | } |
| 3212 | } |
| 3213 | return tab; |
| 3214 | } |
| 3215 | |
| 3216 | /* Record a tag defined by a TeX command of length LEN and starting at NAME. |
| 3217 | The name being defined actually starts at (NAME + LEN + 1). |
| 3218 | But we seem to include the TeX command in the tag name. */ |
| 3219 | void |
| 3220 | TEX_getit (name, len) |
| 3221 | char *name; |
| 3222 | int len; |
| 3223 | { |
| 3224 | char *p = name + len; |
| 3225 | char nambuf[BUFSIZ]; |
| 3226 | |
| 3227 | if (*name == 0) |
| 3228 | return; |
| 3229 | |
| 3230 | /* Let tag name extend to next group close (or end of line) */ |
| 3231 | while (*p && *p != TEX_clgrp) |
| 3232 | p++; |
| 3233 | strncpy (nambuf, name, p - name); |
| 3234 | nambuf[p - name] = 0; |
| 3235 | |
| 3236 | pfnote (nambuf, TRUE, FALSE, lb.buffer, strlen (lb.buffer), lineno, linecharno); |
| 3237 | pfcnt++; |
| 3238 | } |
| 3239 | |
| 3240 | /* If the text at CP matches one of the tag-defining TeX command names, |
| 3241 | return the pointer to the first occurrence of that command in TEX_toktab. |
| 3242 | Otherwise return -1. |
| 3243 | Keep the capital `T' in `Token' for dumb truncating compilers |
| 3244 | (this distinguishes it from `TEX_toktab' */ |
| 3245 | int |
| 3246 | TEX_Token (cp) |
| 3247 | char *cp; |
| 3248 | { |
| 3249 | int i; |
| 3250 | |
| 3251 | for (i = 0; TEX_toktab[i].len > 0; i++) |
| 3252 | if (strneq (TEX_toktab[i].name, cp, TEX_toktab[i].len)) |
| 3253 | return i; |
| 3254 | return -1; |
| 3255 | } |
| 3256 | \f |
| 3257 | /* Support for Prolog. */ |
| 3258 | |
| 3259 | /* Whole head (not only functor, but also arguments) |
| 3260 | is gotten in compound term. */ |
| 3261 | void |
| 3262 | prolog_getit (s) |
| 3263 | char *s; |
| 3264 | { |
| 3265 | char nambuf[BUFSIZ], *save_s, tmpc; |
| 3266 | int insquote, npar; |
| 3267 | |
| 3268 | save_s = s; |
| 3269 | insquote = FALSE; |
| 3270 | npar = 0; |
| 3271 | while (1) |
| 3272 | { |
| 3273 | if (*s == '\0') /* syntax error. */ |
| 3274 | return; |
| 3275 | else if (insquote && *s == '\'' && *(s + 1) == '\'') |
| 3276 | s += 2; |
| 3277 | else if (*s == '\'') |
| 3278 | { |
| 3279 | insquote = !insquote; |
| 3280 | s++; |
| 3281 | } |
| 3282 | else if (!insquote && *s == '(') |
| 3283 | { |
| 3284 | npar++; |
| 3285 | s++; |
| 3286 | } |
| 3287 | else if (!insquote && *s == ')') |
| 3288 | { |
| 3289 | npar--; |
| 3290 | s++; |
| 3291 | if (npar == 0) |
| 3292 | break; |
| 3293 | else if (npar < 0) /* syntax error. */ |
| 3294 | return; |
| 3295 | } |
| 3296 | else if (!insquote && *s == '.' && (isspace (*(s + 1)) || *(s + 1) == '\0')) |
| 3297 | { /* fullstop. */ |
| 3298 | if (npar != 0) /* syntax error. */ |
| 3299 | return; |
| 3300 | s++; |
| 3301 | break; |
| 3302 | } |
| 3303 | else |
| 3304 | s++; |
| 3305 | } |
| 3306 | tmpc = *s; |
| 3307 | *s = '\0'; |
| 3308 | strcpy (nambuf, save_s); |
| 3309 | *s = tmpc; |
| 3310 | pfnote (nambuf, TRUE, FALSE, save_s, strlen (nambuf), lineno, linecharno); |
| 3311 | } |
| 3312 | |
| 3313 | /* It is assumed that prolog predicate starts from column 0. */ |
| 3314 | void |
| 3315 | Prolog_functions (inf) |
| 3316 | FILE *inf; |
| 3317 | { |
| 3318 | void skip_comment (), prolog_getit (); |
| 3319 | |
| 3320 | lineno = linecharno = charno = 0; |
| 3321 | while (!feof (inf)) |
| 3322 | { |
| 3323 | lineno++; |
| 3324 | linecharno += charno; |
| 3325 | charno = readline (&lb, inf) + 1; /* 1 for newline. */ |
| 3326 | dbp = lb.buffer; |
| 3327 | if (isspace (dbp[0])) /* not predicate header. */ |
| 3328 | continue; |
| 3329 | else if (dbp[0] == '%') /* comment. */ |
| 3330 | continue; |
| 3331 | else if (dbp[0] == '/' && dbp[1] == '*') /* comment. */ |
| 3332 | skip_comment (&lb, inf, &lineno, &linecharno); |
| 3333 | else /* found. */ |
| 3334 | prolog_getit (dbp); |
| 3335 | } |
| 3336 | } |
| 3337 | |
| 3338 | void |
| 3339 | skip_comment (plb, inf, plineno, plinecharno) |
| 3340 | struct linebuffer *plb; |
| 3341 | FILE *inf; |
| 3342 | int *plineno; /* result */ |
| 3343 | long *plinecharno; /* result */ |
| 3344 | { |
| 3345 | char *cp; |
| 3346 | |
| 3347 | do |
| 3348 | { |
| 3349 | for (cp = plb->buffer; *cp != '\0'; cp++) |
| 3350 | if (cp[0] == '*' && cp[1] == '/') |
| 3351 | return; |
| 3352 | (*plineno)++; |
| 3353 | *plinecharno += readline (plb, inf) + 1; /* 1 for newline. */ |
| 3354 | } |
| 3355 | while (!feof(inf)); |
| 3356 | } |
| 3357 | \f |
| 3358 | #ifdef ETAGS_REGEXPS |
| 3359 | /* Take a string like "/blah/" and turn it into "blah", making sure |
| 3360 | that the first and last characters are the same, and handling |
| 3361 | quoted separator characters. Actually, stops on the occurence of |
| 3362 | an unquoted separator. Also turns "\t" into a Tab character. |
| 3363 | Returns pointer to terminating separator. Works in place. Null |
| 3364 | terminates name string. */ |
| 3365 | char * |
| 3366 | scan_separators (name) |
| 3367 | char *name; |
| 3368 | { |
| 3369 | char sep = name[0]; |
| 3370 | char *copyto = name; |
| 3371 | logical quoted = FALSE; |
| 3372 | |
| 3373 | for (++name; *name != '\0'; ++name) |
| 3374 | { |
| 3375 | if (quoted) |
| 3376 | { |
| 3377 | if (*name == 't') |
| 3378 | *copyto++ = '\t'; |
| 3379 | else if (*name == sep) |
| 3380 | *copyto++ = sep; |
| 3381 | else |
| 3382 | { |
| 3383 | /* Something else is quoted, so preserve the quote. */ |
| 3384 | *copyto++ = '\\'; |
| 3385 | *copyto++ = *name; |
| 3386 | } |
| 3387 | quoted = FALSE; |
| 3388 | } |
| 3389 | else if (*name == '\\') |
| 3390 | quoted = TRUE; |
| 3391 | else if (*name == sep) |
| 3392 | break; |
| 3393 | else |
| 3394 | *copyto++ = *name; |
| 3395 | } |
| 3396 | |
| 3397 | /* Terminate copied string. */ |
| 3398 | *copyto = '\0'; |
| 3399 | return name; |
| 3400 | } |
| 3401 | |
| 3402 | /* Turn a name, which is an ed-style (but Emacs syntax) regular |
| 3403 | expression, into a real regular expression by compiling it. */ |
| 3404 | void |
| 3405 | add_regex (regexp_pattern) |
| 3406 | char *regexp_pattern; |
| 3407 | { |
| 3408 | char *name; |
| 3409 | const char *err; |
| 3410 | struct re_pattern_buffer *patbuf; |
| 3411 | |
| 3412 | if (regexp_pattern == NULL) |
| 3413 | { |
| 3414 | /* Remove existing regexps. */ |
| 3415 | num_patterns = 0; |
| 3416 | patterns = NULL; |
| 3417 | return; |
| 3418 | } |
| 3419 | |
| 3420 | if (regexp_pattern[0] == '\0') |
| 3421 | { |
| 3422 | error ("missing regexp", 0); |
| 3423 | return; |
| 3424 | } |
| 3425 | if (regexp_pattern[strlen(regexp_pattern)-1] != regexp_pattern[0]) |
| 3426 | { |
| 3427 | error ("%s: unterminated regexp", regexp_pattern); |
| 3428 | return; |
| 3429 | } |
| 3430 | name = scan_separators (regexp_pattern); |
| 3431 | if (regexp_pattern[0] == '\0') |
| 3432 | { |
| 3433 | error ("null regexp", 0); |
| 3434 | return; |
| 3435 | } |
| 3436 | (void) scan_separators (name); |
| 3437 | |
| 3438 | patbuf = xnew (1, struct re_pattern_buffer); |
| 3439 | patbuf->translate = NULL; |
| 3440 | patbuf->fastmap = NULL; |
| 3441 | patbuf->buffer = NULL; |
| 3442 | patbuf->allocated = 0; |
| 3443 | |
| 3444 | err = re_compile_pattern (regexp_pattern, strlen (regexp_pattern), patbuf); |
| 3445 | if (err != NULL) |
| 3446 | { |
| 3447 | error ("%s while compiling pattern", err); |
| 3448 | return; |
| 3449 | } |
| 3450 | |
| 3451 | num_patterns += 1; |
| 3452 | if (num_patterns == 1) |
| 3453 | patterns = xnew (1, struct pattern); |
| 3454 | else |
| 3455 | patterns = ((struct pattern *) |
| 3456 | xrealloc (patterns, |
| 3457 | (num_patterns * sizeof (struct pattern)))); |
| 3458 | patterns[num_patterns - 1].pattern = patbuf; |
| 3459 | patterns[num_patterns - 1].name_pattern = savestr (name); |
| 3460 | patterns[num_patterns - 1].error_signaled = FALSE; |
| 3461 | } |
| 3462 | |
| 3463 | /* |
| 3464 | * Do the subtitutions indicated by the regular expression and |
| 3465 | * arguments. |
| 3466 | */ |
| 3467 | char * |
| 3468 | substitute (in, out, regs) |
| 3469 | char *in, *out; |
| 3470 | struct re_registers *regs; |
| 3471 | { |
| 3472 | char *result = NULL, *t; |
| 3473 | int size = 0; |
| 3474 | |
| 3475 | /* Pass 1: figure out how much size to allocate. */ |
| 3476 | for (t = out; *t; ++t) |
| 3477 | { |
| 3478 | if (*t == '\\') |
| 3479 | { |
| 3480 | ++t; |
| 3481 | if (!*t) |
| 3482 | { |
| 3483 | fprintf (stderr, "%s: pattern subtitution ends prematurely\n", |
| 3484 | progname); |
| 3485 | return NULL; |
| 3486 | } |
| 3487 | if (isdigit (*t)) |
| 3488 | { |
| 3489 | int dig = *t - '0'; |
| 3490 | size += regs->end[dig] - regs->start[dig]; |
| 3491 | } |
| 3492 | } |
| 3493 | } |
| 3494 | |
| 3495 | /* Allocate space and do the substitutions. */ |
| 3496 | result = xnew (size + 1, char); |
| 3497 | size = 0; |
| 3498 | for (; *out; ++out) |
| 3499 | { |
| 3500 | if (*out == '\\') |
| 3501 | { |
| 3502 | ++out; |
| 3503 | if (isdigit (*out)) |
| 3504 | { |
| 3505 | /* Using "dig2" satisfies my debugger. Bleah. */ |
| 3506 | int dig2 = *out - '0'; |
| 3507 | strncpy (result + size, in + regs->start[dig2], |
| 3508 | regs->end[dig2] - regs->start[dig2]); |
| 3509 | size += regs->end[dig2] - regs->start[dig2]; |
| 3510 | } |
| 3511 | else |
| 3512 | { |
| 3513 | switch (*out) |
| 3514 | { |
| 3515 | case '\t': |
| 3516 | result[size++] = '\t'; |
| 3517 | break; |
| 3518 | case '\\': |
| 3519 | *out = '\\'; |
| 3520 | break; |
| 3521 | default: |
| 3522 | result[size++] = *out; |
| 3523 | break; |
| 3524 | } |
| 3525 | } |
| 3526 | } |
| 3527 | else |
| 3528 | result[size++] = *out; |
| 3529 | } |
| 3530 | result[size] = '\0'; |
| 3531 | |
| 3532 | return result; |
| 3533 | } |
| 3534 | \f |
| 3535 | #endif /* ETAGS_REGEXPS */ |
| 3536 | /* Initialize a linebuffer for use */ |
| 3537 | void |
| 3538 | initbuffer (linebuffer) |
| 3539 | struct linebuffer *linebuffer; |
| 3540 | { |
| 3541 | linebuffer->size = 200; |
| 3542 | linebuffer->buffer = xnew (200, char); |
| 3543 | } |
| 3544 | |
| 3545 | /* |
| 3546 | * Read a line of text from `stream' into `linebuffer'. |
| 3547 | * Return the number of characters read from `stream', |
| 3548 | * which is the length of the line including the newline, if any. |
| 3549 | */ |
| 3550 | long |
| 3551 | readline_internal (linebuffer, stream) |
| 3552 | struct linebuffer *linebuffer; |
| 3553 | register FILE *stream; |
| 3554 | { |
| 3555 | char *buffer = linebuffer->buffer; |
| 3556 | register char *p = linebuffer->buffer; |
| 3557 | register char *pend; |
| 3558 | int chars_deleted; |
| 3559 | |
| 3560 | pend = p + linebuffer->size; /* Separate to avoid 386/IX compiler bug. */ |
| 3561 | |
| 3562 | while (1) |
| 3563 | { |
| 3564 | register int c = getc (stream); |
| 3565 | if (p == pend) |
| 3566 | { |
| 3567 | linebuffer->size *= 2; |
| 3568 | buffer = (char *) xrealloc (buffer, linebuffer->size); |
| 3569 | p += buffer - linebuffer->buffer; |
| 3570 | pend = buffer + linebuffer->size; |
| 3571 | linebuffer->buffer = buffer; |
| 3572 | } |
| 3573 | if (c == EOF) |
| 3574 | { |
| 3575 | chars_deleted = 0; |
| 3576 | break; |
| 3577 | } |
| 3578 | if (c == '\n') |
| 3579 | { |
| 3580 | if (p[-1] == '\r' && p > buffer) |
| 3581 | { |
| 3582 | *--p = '\0'; |
| 3583 | chars_deleted = 2; |
| 3584 | } |
| 3585 | else |
| 3586 | { |
| 3587 | *p = '\0'; |
| 3588 | chars_deleted = 1; |
| 3589 | } |
| 3590 | break; |
| 3591 | } |
| 3592 | *p++ = c; |
| 3593 | } |
| 3594 | |
| 3595 | return p - buffer + chars_deleted; |
| 3596 | } |
| 3597 | |
| 3598 | /* |
| 3599 | * Like readline_internal, above, but try to match the input |
| 3600 | * line against any existing regular expressions. |
| 3601 | */ |
| 3602 | long |
| 3603 | readline (linebuffer, stream) |
| 3604 | struct linebuffer *linebuffer; |
| 3605 | FILE *stream; |
| 3606 | { |
| 3607 | /* Read new line. */ |
| 3608 | int i; |
| 3609 | long result = readline_internal (linebuffer, stream); |
| 3610 | |
| 3611 | #ifdef ETAGS_REGEXPS |
| 3612 | /* Match against all listed patterns. */ |
| 3613 | for (i = 0; i < num_patterns; ++i) |
| 3614 | { |
| 3615 | int match = re_match (patterns[i].pattern, linebuffer->buffer, |
| 3616 | (int)result, 0, &patterns[i].regs); |
| 3617 | switch (match) |
| 3618 | { |
| 3619 | case -2: |
| 3620 | /* Some error. */ |
| 3621 | if (!patterns[i].error_signaled) |
| 3622 | { |
| 3623 | error ("error while matching pattern %d", i); |
| 3624 | patterns[i].error_signaled = TRUE; |
| 3625 | } |
| 3626 | break; |
| 3627 | case -1: |
| 3628 | /* No match. */ |
| 3629 | break; |
| 3630 | default: |
| 3631 | /* Match occurred. Construct a tag. */ |
| 3632 | if (patterns[i].name_pattern[0] != '\0') |
| 3633 | { |
| 3634 | /* Make a named tag. */ |
| 3635 | char *name = substitute (linebuffer->buffer, |
| 3636 | patterns[i].name_pattern, |
| 3637 | &patterns[i].regs); |
| 3638 | if (name != NULL) |
| 3639 | pfnote (name, TRUE, TRUE, linebuffer->buffer, |
| 3640 | match, lineno, linecharno); |
| 3641 | } |
| 3642 | else |
| 3643 | { |
| 3644 | /* Make an unnamed tag. */ |
| 3645 | pfnote (NULL, TRUE, FALSE, linebuffer->buffer, |
| 3646 | match, lineno, linecharno); |
| 3647 | } |
| 3648 | break; |
| 3649 | } |
| 3650 | } |
| 3651 | #endif /* ETAGS_REGEXPS */ |
| 3652 | |
| 3653 | return result; |
| 3654 | } |
| 3655 | |
| 3656 | /* |
| 3657 | * Read a file, but do no processing. This is used to do regexp |
| 3658 | * matching on files that have no language defined. |
| 3659 | */ |
| 3660 | void |
| 3661 | just_read_file (inf) |
| 3662 | FILE *inf; |
| 3663 | { |
| 3664 | while (!feof (inf)) |
| 3665 | { |
| 3666 | ++lineno; |
| 3667 | linecharno = charno; |
| 3668 | charno += readline (&lb, inf) + 1; |
| 3669 | } |
| 3670 | } |
| 3671 | |
| 3672 | \f |
| 3673 | char * |
| 3674 | savestr (cp) |
| 3675 | char *cp; |
| 3676 | { |
| 3677 | return savenstr (cp, strlen (cp)); |
| 3678 | } |
| 3679 | |
| 3680 | char * |
| 3681 | savenstr (cp, len) |
| 3682 | char *cp; |
| 3683 | int len; |
| 3684 | { |
| 3685 | register char *dp; |
| 3686 | |
| 3687 | dp = xnew (len + 1, char); |
| 3688 | strncpy (dp, cp, len); |
| 3689 | dp[len] = '\0'; |
| 3690 | return dp; |
| 3691 | } |
| 3692 | |
| 3693 | /* |
| 3694 | * Return the ptr in sp at which the character c last |
| 3695 | * appears; NULL if not found |
| 3696 | * |
| 3697 | * Identical to System V strrchr, included for portability. |
| 3698 | */ |
| 3699 | char * |
| 3700 | etags_strrchr (sp, c) |
| 3701 | register char *sp, c; |
| 3702 | { |
| 3703 | register char *r; |
| 3704 | |
| 3705 | r = NULL; |
| 3706 | do |
| 3707 | { |
| 3708 | if (*sp == c) |
| 3709 | r = sp; |
| 3710 | } while (*sp++); |
| 3711 | return r; |
| 3712 | } |
| 3713 | |
| 3714 | |
| 3715 | /* |
| 3716 | * Return the ptr in sp at which the character c first |
| 3717 | * appears; NULL if not found |
| 3718 | * |
| 3719 | * Identical to System V strchr, included for portability. |
| 3720 | */ |
| 3721 | char * |
| 3722 | etags_strchr (sp, c) |
| 3723 | register char *sp, c; |
| 3724 | { |
| 3725 | do |
| 3726 | { |
| 3727 | if (*sp == c) |
| 3728 | return sp; |
| 3729 | } while (*sp++); |
| 3730 | return NULL; |
| 3731 | } |
| 3732 | |
| 3733 | /* Print error message and exit. */ |
| 3734 | void |
| 3735 | fatal (s1, s2) |
| 3736 | char *s1, *s2; |
| 3737 | { |
| 3738 | error (s1, s2); |
| 3739 | exit (BAD); |
| 3740 | } |
| 3741 | |
| 3742 | /* Print error message. `s1' is printf control string, `s2' is arg for it. */ |
| 3743 | void |
| 3744 | error (s1, s2) |
| 3745 | char *s1, *s2; |
| 3746 | { |
| 3747 | fprintf (stderr, "%s: ", progname); |
| 3748 | fprintf (stderr, s1, s2); |
| 3749 | fprintf (stderr, "\n"); |
| 3750 | } |
| 3751 | |
| 3752 | /* Return a newly-allocated string whose contents |
| 3753 | concatenate those of s1, s2, s3. */ |
| 3754 | char * |
| 3755 | concat (s1, s2, s3) |
| 3756 | char *s1, *s2, *s3; |
| 3757 | { |
| 3758 | int len1 = strlen (s1), len2 = strlen (s2), len3 = strlen (s3); |
| 3759 | char *result = xnew (len1 + len2 + len3 + 1, char); |
| 3760 | |
| 3761 | strcpy (result, s1); |
| 3762 | strcpy (result + len1, s2); |
| 3763 | strcpy (result + len1 + len2, s3); |
| 3764 | result[len1 + len2 + len3] = '\0'; |
| 3765 | |
| 3766 | return result; |
| 3767 | } |
| 3768 | \f |
| 3769 | #ifdef DOS_NT |
| 3770 | char * |
| 3771 | etags_getcwd () |
| 3772 | { |
| 3773 | char *p, cwd[MAXPATHLEN + 1]; /* Fixed size is safe on MSDOS. */ |
| 3774 | getwd (cwd); |
| 3775 | p = cwd; |
| 3776 | while (*p) |
| 3777 | if (*p == '\\') |
| 3778 | *p++ = '/'; |
| 3779 | else |
| 3780 | *p++ = tolower (*p); |
| 3781 | return strdup (cwd); |
| 3782 | } |
| 3783 | #else /* not DOS_NT */ |
| 3784 | /* Does the same work as the system V getcwd, but does not need to |
| 3785 | guess buffer size in advance. */ |
| 3786 | char * |
| 3787 | etags_getcwd () |
| 3788 | { |
| 3789 | char *buf; |
| 3790 | int bufsize = 256; |
| 3791 | |
| 3792 | #ifdef HAVE_GETCWD |
| 3793 | buf = xnew (bufsize, char); |
| 3794 | while (getcwd (buf, bufsize / 2) == NULL) |
| 3795 | { |
| 3796 | if (errno != ERANGE) |
| 3797 | { |
| 3798 | perror ("pwd"); |
| 3799 | exit (BAD); |
| 3800 | } |
| 3801 | bufsize *= 2; |
| 3802 | buf = xnew (bufsize, char); |
| 3803 | } |
| 3804 | #else |
| 3805 | do |
| 3806 | { |
| 3807 | FILE *pipe; |
| 3808 | |
| 3809 | buf = xnew (bufsize, char); |
| 3810 | |
| 3811 | pipe = (FILE *) popen ("pwd 2>/dev/null", "r"); |
| 3812 | if (pipe == NULL) |
| 3813 | { |
| 3814 | perror ("pwd"); |
| 3815 | exit (BAD); |
| 3816 | } |
| 3817 | if (fgets (buf, bufsize, pipe) == NULL) |
| 3818 | { |
| 3819 | perror ("pwd"); |
| 3820 | exit (BAD); |
| 3821 | } |
| 3822 | pclose (pipe); |
| 3823 | |
| 3824 | bufsize *= 2; |
| 3825 | |
| 3826 | } while (buf[strlen (buf) - 1] != '\n'); |
| 3827 | #endif |
| 3828 | |
| 3829 | buf[strlen (buf) - 1] = '\0'; |
| 3830 | return buf; |
| 3831 | } |
| 3832 | #endif /* not DOS_NT */ |
| 3833 | |
| 3834 | /* Return a newly allocated string containing the filename |
| 3835 | of FILE relative to the absolute directory DIR (which |
| 3836 | should end with a slash). */ |
| 3837 | char * |
| 3838 | relative_filename (file, dir) |
| 3839 | char *file, *dir; |
| 3840 | { |
| 3841 | char *fp, *dp, *res; |
| 3842 | |
| 3843 | /* Find the common root of file and dir. */ |
| 3844 | fp = absolute_filename (file, cwd); |
| 3845 | dp = dir; |
| 3846 | while (*fp++ == *dp++) |
| 3847 | continue; |
| 3848 | do |
| 3849 | { |
| 3850 | fp--; |
| 3851 | dp--; |
| 3852 | } |
| 3853 | while (*fp != '/'); |
| 3854 | |
| 3855 | /* Build a sequence of "../" strings for the resulting relative filename. */ |
| 3856 | for (dp = etags_strchr (dp + 1, '/'), res = ""; |
| 3857 | dp != NULL; |
| 3858 | dp = etags_strchr (dp + 1, '/')) |
| 3859 | { |
| 3860 | res = concat (res, "../", ""); |
| 3861 | } |
| 3862 | |
| 3863 | /* Add the filename relative to the common root of file and dir. */ |
| 3864 | res = concat (res, fp + 1, ""); |
| 3865 | |
| 3866 | return res; /* temporary stub */ |
| 3867 | } |
| 3868 | |
| 3869 | /* Return a newly allocated string containing the |
| 3870 | absolute filename of FILE given CWD (which should |
| 3871 | end with a slash). */ |
| 3872 | char * |
| 3873 | absolute_filename (file, cwd) |
| 3874 | char *file, *cwd; |
| 3875 | { |
| 3876 | char *slashp, *cp, *res; |
| 3877 | |
| 3878 | if (file[0] == '/') |
| 3879 | res = concat (file, "", ""); |
| 3880 | else |
| 3881 | res = concat (cwd, file, ""); |
| 3882 | |
| 3883 | /* Delete the "/dirname/.." and "/." substrings. */ |
| 3884 | slashp = etags_strchr (res, '/'); |
| 3885 | while (slashp != NULL && slashp[0] != '\0') |
| 3886 | { |
| 3887 | if (slashp[1] == '.') |
| 3888 | { |
| 3889 | if (slashp[2] == '.' |
| 3890 | && (slashp[3] == '/' || slashp[3] == '\0')) |
| 3891 | { |
| 3892 | cp = slashp; |
| 3893 | do |
| 3894 | cp--; |
| 3895 | while (cp >= res && *cp != '/'); |
| 3896 | if (*cp == '/') |
| 3897 | { |
| 3898 | strcpy (cp, slashp + 3); |
| 3899 | } |
| 3900 | else /* else (cp == res) */ |
| 3901 | { |
| 3902 | if (slashp[3] != '\0') |
| 3903 | strcpy (cp, slashp + 4); |
| 3904 | else |
| 3905 | return "."; |
| 3906 | } |
| 3907 | slashp = cp; |
| 3908 | continue; |
| 3909 | } |
| 3910 | else if (slashp[2] == '/' || slashp[2] == '\0') |
| 3911 | { |
| 3912 | strcpy (slashp, slashp + 2); |
| 3913 | continue; |
| 3914 | } |
| 3915 | } |
| 3916 | |
| 3917 | slashp = etags_strchr (slashp + 1, '/'); |
| 3918 | } |
| 3919 | |
| 3920 | return res; |
| 3921 | } |
| 3922 | |
| 3923 | /* Return a newly allocated string containing the absolute |
| 3924 | filename of dir where FILE resides given CWD (which should |
| 3925 | end with a slash). */ |
| 3926 | char * |
| 3927 | absolute_dirname (file, cwd) |
| 3928 | char *file, *cwd; |
| 3929 | { |
| 3930 | char *slashp, *res; |
| 3931 | char save; |
| 3932 | |
| 3933 | slashp = etags_strrchr (file, '/'); |
| 3934 | if (slashp == NULL) |
| 3935 | return cwd; |
| 3936 | save = slashp[1]; |
| 3937 | slashp[1] = '\0'; |
| 3938 | res = absolute_filename (file, cwd); |
| 3939 | slashp[1] = save; |
| 3940 | |
| 3941 | return res; |
| 3942 | } |
| 3943 | |
| 3944 | /* Like malloc but get fatal error if memory is exhausted. */ |
| 3945 | char * |
| 3946 | xmalloc (size) |
| 3947 | unsigned int size; |
| 3948 | { |
| 3949 | char *result = (char *) malloc (size); |
| 3950 | if (result == NULL) |
| 3951 | fatal ("virtual memory exhausted", 0); |
| 3952 | return result; |
| 3953 | } |
| 3954 | |
| 3955 | char * |
| 3956 | xrealloc (ptr, size) |
| 3957 | char *ptr; |
| 3958 | unsigned int size; |
| 3959 | { |
| 3960 | char *result = (char *) realloc (ptr, size); |
| 3961 | if (result == NULL) |
| 3962 | fatal ("virtual memory exhausted"); |
| 3963 | return result; |
| 3964 | } |