| 1 | /* Tags file maker to go with GNU Emacs |
| 2 | Copyright (C) 1984, 1987, 1988, 1989, 1993 Free Software Foundation, Inc. and Ken Arnold |
| 3 | |
| 4 | This file is not considered part of GNU Emacs. |
| 5 | |
| 6 | This program is free software; you can redistribute it and/or modify |
| 7 | it under the terms of the GNU General Public License as published by |
| 8 | the Free Software Foundation; either version 2 of the License, or |
| 9 | (at your option) any later version. |
| 10 | |
| 11 | This program is distributed in the hope that it will be useful, |
| 12 | but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 13 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 14 | GNU General Public License for more details. |
| 15 | |
| 16 | You should have received a copy of the GNU General Public License |
| 17 | along with this program; if not, write to the Free Software |
| 18 | Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */ |
| 19 | |
| 20 | /* |
| 21 | * Authors: |
| 22 | * Ctags originally by Ken Arnold. |
| 23 | * FORTRAN added by Jim Kleckner. |
| 24 | * Ed Pelegri-Llopart added C typedefs. |
| 25 | * Gnu Emacs TAGS format and modifications by RMS? |
| 26 | * Sam Kendall added C++. |
| 27 | * |
| 28 | * Francesco Potorti` (pot@cnuce.cnr.it) is the current maintainer. 9.7 |
| 29 | */ |
| 30 | |
| 31 | #ifdef HAVE_CONFIG_H |
| 32 | #include <../src/config.h> |
| 33 | #endif |
| 34 | |
| 35 | #include <stdio.h> |
| 36 | #include <ctype.h> |
| 37 | #include <sys/types.h> |
| 38 | #include <sys/stat.h> |
| 39 | |
| 40 | #if !defined (S_ISREG) && defined (S_IFREG) |
| 41 | # define S_ISREG(m) (((m) & S_IFMT) == S_IFREG) |
| 42 | #endif |
| 43 | |
| 44 | #include "getopt.h" |
| 45 | |
| 46 | extern char *malloc (), *realloc (); |
| 47 | extern char *getenv (); |
| 48 | extern char *strcpy (), *strncpy (); |
| 49 | extern int strcmp (); |
| 50 | |
| 51 | char *etags_index (), *etags_rindex (); |
| 52 | char *savenstr (); |
| 53 | |
| 54 | /* Define the symbol ETAGS to make the program "etags", |
| 55 | which makes emacs-style tag tables by default. |
| 56 | Define CTAGS to make the program "ctags" compatible with the usual one. |
| 57 | Define neither one to get behavior that depends |
| 58 | on the name with which the program is invoked |
| 59 | (but we don't normally compile it that way). */ |
| 60 | |
| 61 | #if !defined(ETAGS) && !defined(CTAGS) |
| 62 | /* If neither is defined, program can be run as either. */ |
| 63 | #define ETAGS |
| 64 | #define CTAGS |
| 65 | #endif |
| 66 | |
| 67 | /* On VMS, CTAGS is not useful, so always do ETAGS. */ |
| 68 | #ifdef VMS |
| 69 | #ifndef ETAGS |
| 70 | #define ETAGS |
| 71 | #endif |
| 72 | #endif |
| 73 | |
| 74 | /* Exit codes for success and failure. */ |
| 75 | #ifdef VMS |
| 76 | #define GOOD (1) |
| 77 | #define BAD (0) |
| 78 | #else |
| 79 | #define GOOD (0) |
| 80 | #define BAD (1) |
| 81 | #endif |
| 82 | |
| 83 | /* |
| 84 | * The FILEPOS abstract type, which represents a position in a file, |
| 85 | * plus the following accessor functions: |
| 86 | * |
| 87 | * long GET_CHARNO (pos) |
| 88 | * returns absolute char number. |
| 89 | * void SET_FILEPOS (pos, fp, charno) |
| 90 | * FILE *fp; long charno; |
| 91 | * sets `pos' from the current file |
| 92 | * position of `fp' and from `charno', |
| 93 | * which must be the absolute character |
| 94 | * number corresponding to the current |
| 95 | * position of `fp'. |
| 96 | * |
| 97 | * The `pos' parameter is an lvalue expression of type FILEPOS. |
| 98 | * Parameters to the accessor functions are evaluated 0 or more times, |
| 99 | * and so must have no side effects. |
| 100 | * |
| 101 | * FILEPOS objects can also be assigned and passed to and from |
| 102 | * functions in the normal C manner. |
| 103 | * |
| 104 | * Implementation notes: the `+ 0' is to enforce rvalue-ness. |
| 105 | */ |
| 106 | |
| 107 | #ifndef DEBUG |
| 108 | /* real implementation */ |
| 109 | typedef long FILEPOS; |
| 110 | #define GET_CHARNO(pos) ((pos) + 0) |
| 111 | #define SET_FILEPOS(pos, fp, cno) ((void) ((pos) = (cno))) |
| 112 | #else |
| 113 | /* debugging implementation */ |
| 114 | typedef struct |
| 115 | { |
| 116 | long charno; |
| 117 | } FILEPOS; |
| 118 | |
| 119 | #define GET_CHARNO(pos) ((pos).charno + 0) |
| 120 | #define SET_FILEPOS(pos, fp, cno) \ |
| 121 | ((void) ((pos).charno = (cno), \ |
| 122 | (cno) != ftell (fp) ? (error ("SET_FILEPOS inconsistency"), 0) \ |
| 123 | : 0)) |
| 124 | #endif |
| 125 | |
| 126 | #define streq(s, t) (strcmp (s, t) == 0) |
| 127 | #define strneq(s, t, n) (strncmp (s, t, n) == 0) |
| 128 | #define logical int |
| 129 | |
| 130 | #define TRUE 1 |
| 131 | #define FALSE 0 |
| 132 | |
| 133 | #define iswhite(arg) (_wht[arg]) /* T if char is white */ |
| 134 | #define begtoken(arg) (_btk[arg]) /* T if char can start token */ |
| 135 | #define intoken(arg) (_itk[arg]) /* T if char can be in token */ |
| 136 | #define endtoken(arg) (_etk[arg]) /* T if char ends tokens */ |
| 137 | |
| 138 | #define max(I1,I2) ((I1) > (I2) ? (I1) : (I2)) |
| 139 | |
| 140 | struct nd_st |
| 141 | { /* sorting structure */ |
| 142 | char *name; /* function or type name */ |
| 143 | char *file; /* file name */ |
| 144 | logical is_func; /* use pattern or line no */ |
| 145 | logical named; /* list name separately */ |
| 146 | logical been_warned; /* set if noticed dup */ |
| 147 | int lno; /* line number tag is on */ |
| 148 | long cno; /* character number line starts on */ |
| 149 | char *pat; /* search pattern */ |
| 150 | struct nd_st *left, *right; /* left and right sons */ |
| 151 | }; |
| 152 | |
| 153 | long ftell (); |
| 154 | typedef struct nd_st NODE; |
| 155 | |
| 156 | logical header_file; /* TRUE if .h file, FALSE o.w. */ |
| 157 | /* boolean "functions" (see init) */ |
| 158 | logical _wht[0177], _etk[0177], _itk[0177], _btk[0177]; |
| 159 | |
| 160 | |
| 161 | char *concat (); |
| 162 | char *savenstr (); |
| 163 | char *savestr (); |
| 164 | char *xmalloc (); |
| 165 | char *xrealloc (); |
| 166 | int L_isdef (), L_isquote (); |
| 167 | int PF_funcs (); |
| 168 | int total_size_of_entries (); |
| 169 | logical consider_token (); |
| 170 | logical tail (); |
| 171 | long readline (); |
| 172 | void Asm_funcs (); |
| 173 | void C_entries (); |
| 174 | void L_funcs (); |
| 175 | void L_getit (); |
| 176 | void PAS_funcs (); |
| 177 | void Scheme_funcs (); |
| 178 | void TEX_funcs (); |
| 179 | void add_node (); |
| 180 | void error (); |
| 181 | void fatal (); |
| 182 | void find_entries (); |
| 183 | void free_tree (); |
| 184 | void getit (); |
| 185 | void getline (); |
| 186 | void init (); |
| 187 | void initbuffer (); |
| 188 | void initbuffer (); |
| 189 | void pfnote (); |
| 190 | void process_file (); |
| 191 | void put_entries (); |
| 192 | void takeprec (); |
| 193 | |
| 194 | /* |
| 195 | * MACRO |
| 196 | * xnew -- allocate storage |
| 197 | * |
| 198 | * SYNOPSIS |
| 199 | * Type *xnew (int n, Type); |
| 200 | */ |
| 201 | #define xnew(n, Type) ((Type *) xmalloc ((n) * sizeof (Type))) |
| 202 | \f |
| 203 | |
| 204 | |
| 205 | /* |
| 206 | * Symbol table stuff. |
| 207 | * |
| 208 | * Should probably be implemented with hash table; linked list for now. |
| 209 | */ |
| 210 | |
| 211 | enum sym_type |
| 212 | { |
| 213 | st_none, st_C_struct, st_C_enum, st_C_define, st_C_typedef, st_C_typespec |
| 214 | }; |
| 215 | |
| 216 | struct stab_entry |
| 217 | { |
| 218 | char *sym; |
| 219 | int symlen; |
| 220 | enum sym_type type; |
| 221 | struct stab_entry *next; |
| 222 | }; |
| 223 | |
| 224 | typedef struct stab_entry Stab_entry; |
| 225 | typedef Stab_entry *Stab; |
| 226 | |
| 227 | /* |
| 228 | * NAME |
| 229 | * Stab, Stab_entry, stab_create, stab_search, stab_find -- symbol table |
| 230 | * |
| 231 | * SYNOPSIS |
| 232 | * Types: Stab, Stab_entry, enum sym_type |
| 233 | * |
| 234 | * Stab * stab_create () |
| 235 | * |
| 236 | * Stab_entry * stab_find (stab, sym) |
| 237 | * Stab *stab; |
| 238 | * char *sym; |
| 239 | * |
| 240 | * Stab_entry * stab_search (stab, sym) |
| 241 | * Stab *stab; |
| 242 | * char *sym; |
| 243 | * |
| 244 | * DESCRIPTION |
| 245 | * stab_create creates a Stab, a symbol table object, and returns a |
| 246 | * pointer to it. stab_find finds a symbol in a Stab; it returns a |
| 247 | * pointer to the Stab_entry if found, otherwise NULL. stab_search |
| 248 | * is like stab_find, except that it creates a new Stab_entry, |
| 249 | * initialized with type = st_none, if one did not exist already |
| 250 | * (it never returns NULL). |
| 251 | * |
| 252 | * A Stab_entry is a structure that contains at least the following |
| 253 | * members: |
| 254 | * |
| 255 | * char *name; // must not be modified |
| 256 | * enum sym_type type; // should be set |
| 257 | * |
| 258 | * The type field is initially set to st_none; it should be set to |
| 259 | * something else by the caller of stab_search. Other possible values |
| 260 | * of an enum sym_type can be added. |
| 261 | */ |
| 262 | |
| 263 | Stab * |
| 264 | stab_create () |
| 265 | { |
| 266 | Stab *sp; |
| 267 | sp = xnew (1, Stab); |
| 268 | *sp = NULL; /* a Stab starts out as a null Stab_entry* */ |
| 269 | return sp; |
| 270 | } |
| 271 | |
| 272 | Stab_entry * |
| 273 | stab_find (stab, sym, symlen) |
| 274 | Stab *stab; |
| 275 | register char *sym; |
| 276 | register int symlen; |
| 277 | { |
| 278 | register Stab_entry *se; |
| 279 | for (se = *stab; se != NULL; se = se->next) |
| 280 | { |
| 281 | if (se->symlen == symlen && strneq (se->sym, sym, symlen)) |
| 282 | return se; |
| 283 | } |
| 284 | |
| 285 | return NULL; |
| 286 | } |
| 287 | |
| 288 | Stab_entry * |
| 289 | stab_search (stab, sym, symlen) |
| 290 | register Stab *stab; |
| 291 | char *sym; |
| 292 | int symlen; |
| 293 | { |
| 294 | register Stab_entry *se; |
| 295 | se = stab_find (stab, sym, symlen); |
| 296 | |
| 297 | if (se == NULL) |
| 298 | { |
| 299 | /* make a new one */ |
| 300 | se = xnew (1, Stab_entry); |
| 301 | se->sym = savenstr (sym, symlen); |
| 302 | se->symlen = symlen; |
| 303 | se->type = st_none; |
| 304 | se->next = *stab; |
| 305 | *stab = se; |
| 306 | } |
| 307 | |
| 308 | return se; |
| 309 | } |
| 310 | |
| 311 | /* |
| 312 | * NAME |
| 313 | * stab_type -- type of a symbol table entry |
| 314 | * |
| 315 | * SYNOPSIS |
| 316 | * enum sym_type stab_type (Stab_entry *se); |
| 317 | * |
| 318 | * WARNING |
| 319 | * May evaluate its argument more than once. |
| 320 | */ |
| 321 | |
| 322 | #define stab_type(se) ((se)==NULL ? st_none : (se)->type) |
| 323 | \f |
| 324 | |
| 325 | |
| 326 | typedef int LINENO; |
| 327 | |
| 328 | typedef struct |
| 329 | { |
| 330 | char *p; |
| 331 | int len; |
| 332 | LINENO lineno; |
| 333 | logical named; |
| 334 | } TOKEN; |
| 335 | |
| 336 | /* C extensions. |
| 337 | */ |
| 338 | #define C_PLPL 0x00001 /* C++ */ |
| 339 | #define C_STAR 0x00003 /* C* */ |
| 340 | #define YACC 0x10000 /* yacc file */ |
| 341 | |
| 342 | char searchar = '/'; /* use /.../ searches */ |
| 343 | |
| 344 | LINENO lineno; /* line number of current line */ |
| 345 | long charno; /* current character number */ |
| 346 | |
| 347 | long linecharno; /* charno of start of line; not used by C, but |
| 348 | * by every other language. |
| 349 | */ |
| 350 | |
| 351 | char *curfile, /* current input file name */ |
| 352 | *outfile, /* output file */ |
| 353 | *white = " \f\t\n", /* white chars */ |
| 354 | *endtk = " \t\n\"'#()[]{}=-+%*/&|^~!<>;,.:?", /* token ending chars */ |
| 355 | /* token starting chars */ |
| 356 | *begtk = "ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz$", |
| 357 | /* valid in-token chars */ |
| 358 | *intk = "ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz$0123456789"; |
| 359 | |
| 360 | int append_to_tagfile; /* -a: append to tags */ |
| 361 | int emacs_tags_format; /* emacs style output (no -e option any more) */ |
| 362 | /* The following three default to 1 for etags, but to 0 for ctags. */ |
| 363 | int typedefs; /* -t: create tags for typedefs */ |
| 364 | int typedefs_and_cplusplus; /* -T: create tags for typedefs, level */ |
| 365 | /* 0 struct/enum/union decls, and C++ */ |
| 366 | /* member functions */ |
| 367 | int constantypedefs; /* -d: create tags for C #define and enum */ |
| 368 | /* constants. Default under etags. Enum */ |
| 369 | /* constants not implemented. */ |
| 370 | /* -D: opposite of -d. Default under ctags. */ |
| 371 | int update; /* -u: update tags */ |
| 372 | int vgrind_style; /* -v: create vgrind style index output */ |
| 373 | int no_warnings; /* -w: suppress warnings */ |
| 374 | int cxref_style; /* -x: create cxref style output */ |
| 375 | int cplusplus; /* .[hc] means C++, not C */ |
| 376 | int noindentypedefs; /* -S: ignore indentation in C */ |
| 377 | |
| 378 | /* Name this program was invoked with. */ |
| 379 | char *progname; |
| 380 | |
| 381 | struct option longopts[] = { |
| 382 | { "append", no_argument, NULL, 'a' }, |
| 383 | { "backward-search", no_argument, NULL, 'B' }, |
| 384 | { "c++", no_argument, NULL, 'C' }, |
| 385 | { "cxref", no_argument, NULL, 'x' }, |
| 386 | { "defines", no_argument, NULL, 'd' }, |
| 387 | { "forward-search", no_argument, NULL, 'F' }, |
| 388 | { "help", no_argument, NULL, 'H' }, |
| 389 | { "ignore-indentation", no_argument, NULL, 'S' }, |
| 390 | { "include", required_argument, NULL, 'i' }, |
| 391 | { "no-defines", no_argument, NULL, 'D' }, |
| 392 | { "no-warn", no_argument, NULL, 'w' }, |
| 393 | { "output", required_argument, NULL, 'o' }, |
| 394 | { "typedefs", no_argument, NULL, 't' }, |
| 395 | { "typedefs-and-c++", no_argument, NULL, 'T' }, |
| 396 | { "update", no_argument, NULL, 'u' }, |
| 397 | { "version", no_argument, NULL, 'V' }, |
| 398 | { "vgrind", no_argument, NULL, 'v' }, |
| 399 | { 0 } |
| 400 | }; |
| 401 | |
| 402 | FILE *inf, /* ioptr for current input file */ |
| 403 | *outf; /* ioptr for tags file */ |
| 404 | |
| 405 | NODE *head; /* the head of the binary tree of tags */ |
| 406 | |
| 407 | int permit_duplicates = 1; /* Nonzero means allow duplicate tags. */ |
| 408 | |
| 409 | /* A `struct linebuffer' is a structure which holds a line of text. |
| 410 | `readline' reads a line from a stream into a linebuffer |
| 411 | and works regardless of the length of the line. */ |
| 412 | |
| 413 | struct linebuffer |
| 414 | { |
| 415 | long size; |
| 416 | char *buffer; |
| 417 | }; |
| 418 | |
| 419 | struct linebuffer lb; /* the current line */ |
| 420 | struct linebuffer filename_lb; /* used to read in filenames */ |
| 421 | struct |
| 422 | { |
| 423 | FILEPOS linepos; |
| 424 | struct linebuffer lb; /* used by C_entries instead of lb */ |
| 425 | } lbs[2]; |
| 426 | \f |
| 427 | void |
| 428 | print_version () |
| 429 | { |
| 430 | #ifdef CTAGS |
| 431 | printf ("CTAGS "); |
| 432 | #ifdef ETAGS |
| 433 | printf ("and "); |
| 434 | #endif |
| 435 | #endif |
| 436 | #ifdef ETAGS |
| 437 | printf ("ETAGS "); |
| 438 | #endif |
| 439 | printf ("for Emacs version 19.\n"); |
| 440 | |
| 441 | exit (0); |
| 442 | } |
| 443 | |
| 444 | void |
| 445 | print_help () |
| 446 | { |
| 447 | printf ("These are the options accepted by %s. You may use unambiguous\n\ |
| 448 | abbreviations for the long option names.\n\n", progname); |
| 449 | |
| 450 | puts ("-a, --append\n\ |
| 451 | Append tag entries to existing tags file."); |
| 452 | puts ("-C, --c++\n\ |
| 453 | Treat files with `.c' and `.h' extensions as C++ code, not C\n\ |
| 454 | code. Files with `.C', `.H', `.cxx', `.hxx', or `.cc'\n\ |
| 455 | extensions are always assumed to be C++ code."); |
| 456 | fputs ("-d, --defines\n\ |
| 457 | Create tag entries for #defines, too.", stdout); |
| 458 | |
| 459 | #ifdef ETAGS |
| 460 | fputs (" This is the default\n\ |
| 461 | behavior.", stdout); |
| 462 | #endif |
| 463 | |
| 464 | fputs ("\n\ |
| 465 | -D, --no-defines\n\ |
| 466 | Don't create tag entries for #defines.", stdout); |
| 467 | |
| 468 | #ifdef CTAGS |
| 469 | fputs (" This is the default\n\ |
| 470 | behavior.", stdout); |
| 471 | #endif |
| 472 | |
| 473 | puts ("\n\ |
| 474 | -o FILE, --output=FILE\n\ |
| 475 | Write the tags to FILE.\n\ |
| 476 | -S, --ignore-indentation\n\ |
| 477 | Don't rely on indentation quite as much as normal. Currently,\n\ |
| 478 | this means not to assume that a closing brace in the first\n\ |
| 479 | column is the final brace of a function or structure\n\ |
| 480 | definition."); |
| 481 | puts ("-t, --typedefs\n\ |
| 482 | Generate tag entries for typedefs. This is the default\n\ |
| 483 | behavior."); |
| 484 | puts ("-T, --typedefs-and-c++\n\ |
| 485 | Generate tag entries for typedefs, struct/enum/union tags, and\n\ |
| 486 | C++ member functions."); |
| 487 | |
| 488 | #ifdef ETAGS |
| 489 | puts ("-i FILE, --include=FILE\n\ |
| 490 | Include a note in tag file indicating that, when searching for\n\ |
| 491 | a tag, one should also consult the tags file FILE after\n\ |
| 492 | checking the current file."); |
| 493 | #endif |
| 494 | |
| 495 | #ifdef CTAGS |
| 496 | puts ("-B, --backward-search\n\ |
| 497 | Write the search commands for the tag entries using '?', the\n\ |
| 498 | backward-search command."); |
| 499 | puts ("-F, --forward-search\n\ |
| 500 | Write the search commands for the tag entries using '/', the\n\ |
| 501 | forward-search command."); |
| 502 | puts ("-u, --update\n\ |
| 503 | Update the tag entries for the given files, leaving tag\n\ |
| 504 | entries for other files in place. Currently, this is\n\ |
| 505 | implemented by deleting the existing entries for the given\n\ |
| 506 | files and then rewriting the new entries at the end of the\n\ |
| 507 | tags file. It is often faster to simply rebuild the entire\n\ |
| 508 | tag file than to use this."); |
| 509 | puts ("-v, --vgrind\n\ |
| 510 | Generates an index of items intended for human consumption,\n\ |
| 511 | similar to the output of vgrind. The index is sorted, and\n\ |
| 512 | gives the page number of each item."); |
| 513 | puts ("-x, --cxref\n\ |
| 514 | Like --vgrind, but in the style of cxref, rather than vgrind.\n\ |
| 515 | The output uses line numbers instead of page numbers, but\n\ |
| 516 | beyond that the differences are cosmetic; try both to see\n\ |
| 517 | which you like."); |
| 518 | puts ("-w, --no-warn\n\ |
| 519 | Suppress warning messages about entries defined in multiple\n\ |
| 520 | files."); |
| 521 | #endif |
| 522 | |
| 523 | puts ("-V, --version\n\ |
| 524 | Print the version of the program.\n\ |
| 525 | -H, --help\n\ |
| 526 | Print this help message."); |
| 527 | |
| 528 | exit (0); |
| 529 | } |
| 530 | |
| 531 | \f |
| 532 | void |
| 533 | main (argc, argv) |
| 534 | int argc; |
| 535 | char *argv[]; |
| 536 | { |
| 537 | char cmd[100]; |
| 538 | int i; |
| 539 | unsigned int nincluded_files = 0; |
| 540 | char **included_files = xnew (argc, char *); |
| 541 | char *this_file; |
| 542 | #ifdef VMS |
| 543 | char got_err; |
| 544 | |
| 545 | extern char *gfnames (); |
| 546 | extern char *massage_name (); |
| 547 | #endif |
| 548 | |
| 549 | progname = argv[0]; |
| 550 | |
| 551 | #ifndef CTAGS |
| 552 | emacs_tags_format = 1; |
| 553 | #else |
| 554 | emacs_tags_format = 0; |
| 555 | #endif |
| 556 | |
| 557 | /* |
| 558 | * If etags, always find typedefs and structure tags. Why not? |
| 559 | * Also default is to find macro constants. |
| 560 | */ |
| 561 | if (emacs_tags_format) |
| 562 | typedefs = typedefs_and_cplusplus = constantypedefs = 1; |
| 563 | |
| 564 | for (;;) |
| 565 | { |
| 566 | int opt; |
| 567 | opt = getopt_long (argc, argv, "aCdDo:f:StTi:BFuvxwVH", longopts, 0); |
| 568 | |
| 569 | if (opt == EOF) |
| 570 | break; |
| 571 | |
| 572 | switch (opt) |
| 573 | { |
| 574 | case '\0': |
| 575 | /* If getopt returns '\0', then it has already processed a |
| 576 | long-named option. We should do nothing. */ |
| 577 | break; |
| 578 | |
| 579 | /* Common options. */ |
| 580 | case 'a': |
| 581 | append_to_tagfile++; |
| 582 | break; |
| 583 | case 'C': |
| 584 | cplusplus = 1; |
| 585 | break; |
| 586 | case 'd': |
| 587 | constantypedefs = 1; |
| 588 | break; |
| 589 | case 'D': |
| 590 | constantypedefs = 0; |
| 591 | break; |
| 592 | case 'f': |
| 593 | case 'o': |
| 594 | if (outfile) |
| 595 | { |
| 596 | fprintf (stderr, |
| 597 | "%s: -%c flag may only be given once\n", progname, opt); |
| 598 | goto usage; |
| 599 | } |
| 600 | outfile = optarg; |
| 601 | break; |
| 602 | case 'S': |
| 603 | noindentypedefs++; |
| 604 | break; |
| 605 | case 't': |
| 606 | typedefs++; |
| 607 | break; |
| 608 | case 'T': |
| 609 | typedefs++; |
| 610 | typedefs_and_cplusplus++; |
| 611 | break; |
| 612 | case 'V': |
| 613 | print_version (); |
| 614 | break; |
| 615 | case 'H': |
| 616 | print_help (); |
| 617 | break; |
| 618 | |
| 619 | /* Etags options */ |
| 620 | case 'i': |
| 621 | if (!emacs_tags_format) |
| 622 | goto usage; |
| 623 | included_files[nincluded_files++] = optarg; |
| 624 | break; |
| 625 | |
| 626 | /* Ctags options. */ |
| 627 | case 'B': |
| 628 | searchar = '?'; |
| 629 | if (emacs_tags_format) |
| 630 | goto usage; |
| 631 | break; |
| 632 | case 'F': |
| 633 | searchar = '/'; |
| 634 | if (emacs_tags_format) |
| 635 | goto usage; |
| 636 | break; |
| 637 | case 'u': |
| 638 | update++; |
| 639 | if (emacs_tags_format) |
| 640 | goto usage; |
| 641 | break; |
| 642 | case 'v': |
| 643 | vgrind_style++; |
| 644 | /*FALLTHRU*/ |
| 645 | case 'x': |
| 646 | cxref_style++; |
| 647 | if (emacs_tags_format) |
| 648 | goto usage; |
| 649 | break; |
| 650 | case 'w': |
| 651 | no_warnings++; |
| 652 | if (emacs_tags_format) |
| 653 | goto usage; |
| 654 | break; |
| 655 | |
| 656 | default: |
| 657 | goto usage; |
| 658 | } |
| 659 | } |
| 660 | |
| 661 | if (optind == argc && nincluded_files == 0) |
| 662 | { |
| 663 | fprintf (stderr, "%s: No input files specified.\n", progname); |
| 664 | |
| 665 | usage: |
| 666 | fprintf (stderr, "%s: Try `%s --help' for a complete list of options.\n", |
| 667 | progname, progname); |
| 668 | exit (BAD); |
| 669 | } |
| 670 | |
| 671 | if (outfile == 0) |
| 672 | { |
| 673 | outfile = emacs_tags_format ? "TAGS" : "tags"; |
| 674 | } |
| 675 | |
| 676 | init (); /* set up boolean "functions" */ |
| 677 | |
| 678 | initbuffer (&lb); |
| 679 | initbuffer (&lbs[0].lb); |
| 680 | initbuffer (&lbs[1].lb); |
| 681 | initbuffer (&filename_lb); |
| 682 | /* |
| 683 | * loop through files finding functions |
| 684 | */ |
| 685 | if (emacs_tags_format) |
| 686 | { |
| 687 | if (streq (outfile, "-")) |
| 688 | outf = stdout; |
| 689 | else |
| 690 | outf = fopen (outfile, append_to_tagfile ? "a" : "w"); |
| 691 | if (!outf) |
| 692 | { |
| 693 | perror (outfile); |
| 694 | exit (1); |
| 695 | } |
| 696 | } |
| 697 | |
| 698 | #ifdef VMS |
| 699 | argc -= optind; |
| 700 | argv += optind; |
| 701 | while (gfnames (&argc, &argv, &got_err) != NULL) |
| 702 | { |
| 703 | if (got_err) |
| 704 | { |
| 705 | error ("Can't find file %s\n", this_file); |
| 706 | argc--, argv++; |
| 707 | } |
| 708 | else |
| 709 | { |
| 710 | this_file = massage_name (this_file); |
| 711 | #if 0 |
| 712 | } |
| 713 | } /* solely to balance out the ifdef'd parens above */ |
| 714 | #endif |
| 715 | #else |
| 716 | for (; optind < argc; optind++) |
| 717 | { |
| 718 | this_file = argv[optind]; |
| 719 | if (1) |
| 720 | { |
| 721 | #endif |
| 722 | /* Input file named "-" means read file names from stdin |
| 723 | and use them. */ |
| 724 | if (streq (this_file, "-")) |
| 725 | { |
| 726 | while (!feof (stdin)) |
| 727 | { |
| 728 | (void) readline (&filename_lb, stdin); |
| 729 | if (strlen (filename_lb.buffer) > 0) |
| 730 | process_file (filename_lb.buffer); |
| 731 | } |
| 732 | } |
| 733 | else |
| 734 | process_file (this_file); |
| 735 | } |
| 736 | } |
| 737 | |
| 738 | if (emacs_tags_format) |
| 739 | { |
| 740 | while (nincluded_files-- > 0) |
| 741 | fprintf (outf, "\f\n%s,include\n", *included_files++); |
| 742 | |
| 743 | (void) fclose (outf); |
| 744 | exit (0); |
| 745 | } |
| 746 | |
| 747 | if (cxref_style) |
| 748 | { |
| 749 | put_entries (head); |
| 750 | exit (GOOD); |
| 751 | } |
| 752 | if (update) |
| 753 | { |
| 754 | /* update cannot be set under VMS, so we may assume that argc |
| 755 | and argv have not been munged. */ |
| 756 | for (i = optind; i < argc; i++) |
| 757 | { |
| 758 | sprintf (cmd, |
| 759 | "mv %s OTAGS;fgrep -v '\t%s\t' OTAGS >%s;rm OTAGS", |
| 760 | outfile, argv[i], outfile); |
| 761 | (void) system (cmd); |
| 762 | } |
| 763 | append_to_tagfile++; |
| 764 | } |
| 765 | outf = fopen (outfile, append_to_tagfile ? "a" : "w"); |
| 766 | if (outf == NULL) |
| 767 | { |
| 768 | perror (outfile); |
| 769 | exit (GOOD); |
| 770 | } |
| 771 | put_entries (head); |
| 772 | (void) fclose (outf); |
| 773 | if (update) |
| 774 | { |
| 775 | sprintf (cmd, "sort %s -o %s", outfile, outfile); |
| 776 | (void) system (cmd); |
| 777 | } |
| 778 | exit (GOOD); |
| 779 | } |
| 780 | |
| 781 | |
| 782 | /* |
| 783 | * This routine is called on each file argument. |
| 784 | */ |
| 785 | void |
| 786 | process_file (file) |
| 787 | char *file; |
| 788 | { |
| 789 | struct stat stat_buf; |
| 790 | |
| 791 | if (stat (file, &stat_buf) || !S_ISREG (stat_buf.st_mode)) |
| 792 | { |
| 793 | fprintf (stderr, "Skipping %s: it is not a regular file.\n", file); |
| 794 | return; |
| 795 | } |
| 796 | |
| 797 | if (streq (file, outfile) && !streq (outfile, "-")) |
| 798 | { |
| 799 | fprintf (stderr, "Skipping inclusion of %s in self.\n", file); |
| 800 | return; |
| 801 | } |
| 802 | find_entries (file); |
| 803 | if (emacs_tags_format) |
| 804 | { |
| 805 | fprintf (outf, "\f\n%s,%d\n", file, total_size_of_entries (head)); |
| 806 | put_entries (head); |
| 807 | free_tree (head); |
| 808 | head = NULL; |
| 809 | } |
| 810 | } |
| 811 | |
| 812 | /* |
| 813 | * This routine sets up the boolean pseudo-functions which work |
| 814 | * by setting boolean flags dependent upon the corresponding character |
| 815 | * Every char which is NOT in that string is not a white char. Therefore, |
| 816 | * all of the array "_wht" is set to FALSE, and then the elements |
| 817 | * subscripted by the chars in "white" are set to TRUE. Thus "_wht" |
| 818 | * of a char is TRUE if it is the string "white", else FALSE. |
| 819 | */ |
| 820 | void |
| 821 | init () |
| 822 | { |
| 823 | register char *sp; |
| 824 | register int i; |
| 825 | |
| 826 | for (i = 0; i < 0177; i++) |
| 827 | _wht[i] = _etk[i] = _itk[i] = _btk[i] = FALSE; |
| 828 | for (sp = white; *sp; sp++) |
| 829 | _wht[*sp] = TRUE; |
| 830 | for (sp = endtk; *sp; sp++) |
| 831 | _etk[*sp] = TRUE; |
| 832 | for (sp = intk; *sp; sp++) |
| 833 | _itk[*sp] = TRUE; |
| 834 | for (sp = begtk; *sp; sp++) |
| 835 | _btk[*sp] = TRUE; |
| 836 | _wht[0] = _wht['\n']; |
| 837 | _etk[0] = _etk['\n']; |
| 838 | _btk[0] = _btk['\n']; |
| 839 | _itk[0] = _itk['\n']; |
| 840 | } |
| 841 | |
| 842 | /* |
| 843 | * This routine opens the specified file and calls the function |
| 844 | * which finds the function and type definitions. |
| 845 | */ |
| 846 | void |
| 847 | find_entries (file) |
| 848 | char *file; |
| 849 | { |
| 850 | char *cp; |
| 851 | void prolog_funcs (); |
| 852 | |
| 853 | inf = fopen (file, "r"); |
| 854 | if (inf == NULL) |
| 855 | { |
| 856 | perror (file); |
| 857 | return; |
| 858 | } |
| 859 | curfile = savestr (file); |
| 860 | cp = etags_rindex (file, '.'); |
| 861 | |
| 862 | header_file = (cp && (streq (cp + 1, "h"))); |
| 863 | |
| 864 | /* .tex, .aux or .bbl implies LaTeX source code */ |
| 865 | if (cp && (streq (cp + 1, "tex") || streq (cp + 1, "aux") |
| 866 | || streq (cp + 1, "bbl"))) |
| 867 | { |
| 868 | TEX_funcs (inf); |
| 869 | goto close_and_return; |
| 870 | } |
| 871 | /* .l or .el or .lisp (or .cl or .clisp or ...) implies lisp source code */ |
| 872 | if (cp && (streq (cp + 1, "l") |
| 873 | || streq (cp + 1, "el") |
| 874 | || streq (cp + 1, "lsp") |
| 875 | || streq (cp + 1, "lisp") |
| 876 | || streq (cp + 1, "cl") |
| 877 | || streq (cp + 1, "clisp"))) |
| 878 | { |
| 879 | L_funcs (inf); |
| 880 | goto close_and_return; |
| 881 | } |
| 882 | /* .scm or .sm or .scheme or ... implies scheme source code */ |
| 883 | if (cp && (streq (cp + 1, "sm") |
| 884 | || streq (cp + 1, "scm") |
| 885 | || streq (cp + 1, "scheme") |
| 886 | || streq (cp + 1, "t") |
| 887 | || streq (cp + 1, "sch") |
| 888 | || streq (cp + 1, "SM") |
| 889 | || streq (cp + 1, "SCM") |
| 890 | /* The `SCM' or `scm' prefix with a version number */ |
| 891 | || (cp[-1] == 'm' && cp[-2] == 'c' && cp[-3] == 's' |
| 892 | && string_numeric_p (cp + 1)) |
| 893 | || (cp[-1] == 'M' && cp[-2] == 'C' && cp[-3] == 'S' |
| 894 | && string_numeric_p (cp + 1)))) |
| 895 | { |
| 896 | Scheme_funcs (inf); |
| 897 | fclose (inf); |
| 898 | return; |
| 899 | } |
| 900 | /* Assume that ".s" or ".a" is assembly code. -wolfgang. |
| 901 | Or even ".sa". */ |
| 902 | if (cp && (streq (cp + 1, "s") |
| 903 | || streq (cp + 1, "a") |
| 904 | || streq (cp + 1, "sa"))) |
| 905 | { |
| 906 | Asm_funcs (inf); |
| 907 | fclose (inf); |
| 908 | return; |
| 909 | } |
| 910 | /* .C or .H or .cxx or .hxx or .cc: a C++ file */ |
| 911 | if (cp && (streq (cp + 1, "C") |
| 912 | || streq (cp + 1, "H") |
| 913 | || streq (cp + 1, "cxx") |
| 914 | || streq (cp + 1, "hxx") |
| 915 | || streq (cp + 1, "cc"))) |
| 916 | { |
| 917 | C_entries (C_PLPL); /* C++ */ |
| 918 | goto close_and_return; |
| 919 | } |
| 920 | /* .cs or .hs: a C* file */ |
| 921 | if (cp && (streq (cp + 1, "cs") |
| 922 | || streq (cp + 1, "hs"))) |
| 923 | { |
| 924 | C_entries (C_STAR); |
| 925 | goto close_and_return; |
| 926 | } |
| 927 | /* .y: a yacc file */ |
| 928 | if (cp && (streq (cp + 1, "y"))) |
| 929 | { |
| 930 | C_entries (YACC); |
| 931 | goto close_and_return; |
| 932 | } |
| 933 | /* .pl implies prolog source code */ |
| 934 | if (cp && streq (cp + 1, "pl")) |
| 935 | { |
| 936 | prolog_funcs (inf); |
| 937 | goto close_and_return; |
| 938 | } |
| 939 | /* .p or .pas: a Pascal file */ |
| 940 | if (cp && (streq (cp + 1, "p") |
| 941 | || streq (cp + 1, "pas"))) |
| 942 | { |
| 943 | PAS_funcs (inf); |
| 944 | goto close_and_return; |
| 945 | } |
| 946 | /* If .f or .for, assume it is fortran or nothing. */ |
| 947 | if (cp && (streq (cp + 1, "f") |
| 948 | || streq (cp + 1, "for"))) |
| 949 | { |
| 950 | PF_funcs (inf); |
| 951 | goto close_and_return; |
| 952 | } |
| 953 | /* if not a .c or .h or .y file, try fortran */ |
| 954 | if (cp && ((cp[1] != 'c' |
| 955 | && cp[1] != 'h' |
| 956 | && cp[1] != 'y') |
| 957 | || (cp[1] != 0 && cp[2] != 0))) |
| 958 | { |
| 959 | if (PF_funcs (inf) != 0) |
| 960 | goto close_and_return; |
| 961 | rewind (inf); /* no fortran tags found, try C */ |
| 962 | } |
| 963 | C_entries (cplusplus ? C_PLPL : 0); |
| 964 | |
| 965 | close_and_return: |
| 966 | (void) fclose (inf); |
| 967 | } |
| 968 | |
| 969 | /* Nonzero if string STR is composed of digits. */ |
| 970 | |
| 971 | int |
| 972 | string_numeric_p (str) |
| 973 | char *str; |
| 974 | { |
| 975 | while (*str) |
| 976 | { |
| 977 | if (*str < '0' || *str > '9') |
| 978 | return 0; |
| 979 | } |
| 980 | return 1; |
| 981 | } |
| 982 | \f |
| 983 | /* Record a tag. */ |
| 984 | /* Should take a TOKEN* instead!! */ |
| 985 | void |
| 986 | pfnote (name, is_func, named, linestart, linelen, lno, cno) |
| 987 | char *name; /* tag name */ |
| 988 | logical is_func; /* function or type name? */ |
| 989 | logical named; /* tag different from text of definition? */ |
| 990 | char *linestart; |
| 991 | int linelen; |
| 992 | int lno; |
| 993 | long cno; |
| 994 | { |
| 995 | register char *fp; |
| 996 | register NODE *np; |
| 997 | char tem[51]; |
| 998 | char c; |
| 999 | |
| 1000 | np = (NODE *) malloc (sizeof (NODE)); |
| 1001 | if (np == NULL) |
| 1002 | { |
| 1003 | if (!emacs_tags_format) |
| 1004 | { |
| 1005 | /* It's okay to output early in etags -- it only disrupts the |
| 1006 | * character count of the tag entries, which is no longer used |
| 1007 | * by tags.el anyway. |
| 1008 | */ |
| 1009 | error ("too many entries to sort"); |
| 1010 | } |
| 1011 | put_entries (head); |
| 1012 | free_tree (head); |
| 1013 | head = NULL; |
| 1014 | np = xnew (1, NODE); |
| 1015 | } |
| 1016 | /* If ctags mode, change name "main" to M<thisfilename>. */ |
| 1017 | if (!emacs_tags_format && !cxref_style && streq (name, "main")) |
| 1018 | { |
| 1019 | fp = etags_rindex (curfile, '/'); |
| 1020 | name = concat ("M", fp == 0 ? curfile : fp + 1, ""); |
| 1021 | fp = etags_rindex (name, '.'); |
| 1022 | if (fp && fp[1] != '\0' && fp[2] == '\0') |
| 1023 | *fp = 0; |
| 1024 | named = TRUE; |
| 1025 | } |
| 1026 | np->name = savestr (name); |
| 1027 | np->file = curfile; |
| 1028 | np->is_func = is_func; |
| 1029 | np->named = named; |
| 1030 | np->lno = lno; |
| 1031 | /* UNCOMMENT THE +1 HERE: */ |
| 1032 | np->cno = cno /* + 1 */ ; /* our char numbers are 0-base; emacs's are 1-base */ |
| 1033 | np->left = np->right = 0; |
| 1034 | if (emacs_tags_format) |
| 1035 | { |
| 1036 | c = linestart[linelen]; |
| 1037 | linestart[linelen] = 0; |
| 1038 | } |
| 1039 | else if (cxref_style == 0) |
| 1040 | { |
| 1041 | sprintf (tem, strlen (linestart) < 50 ? "%s$" : "%.50s", linestart); |
| 1042 | linestart = tem; |
| 1043 | } |
| 1044 | np->pat = savestr (linestart); |
| 1045 | if (emacs_tags_format) |
| 1046 | { |
| 1047 | linestart[linelen] = c; |
| 1048 | } |
| 1049 | |
| 1050 | add_node (np, &head); |
| 1051 | } |
| 1052 | |
| 1053 | /* |
| 1054 | * free_tree () |
| 1055 | * recurse on left children, iterate on right children. |
| 1056 | */ |
| 1057 | void |
| 1058 | free_tree (node) |
| 1059 | register NODE *node; |
| 1060 | { |
| 1061 | while (node) |
| 1062 | { |
| 1063 | register NODE *node_right = node->right; |
| 1064 | free_tree (node->left); |
| 1065 | free (node->name); |
| 1066 | free (node->pat); |
| 1067 | free ((char *) node); |
| 1068 | node = node_right; |
| 1069 | } |
| 1070 | } |
| 1071 | |
| 1072 | /* |
| 1073 | * add_node () |
| 1074 | * Adds a node to the tree of nodes. In etags mode, we don't keep |
| 1075 | * it sorted; we just keep a linear list. In ctags mode, maintain |
| 1076 | * an ordered tree, with no attempt at balancing. |
| 1077 | * |
| 1078 | * add_node is the only function allowed to add nodes, so it can |
| 1079 | * maintain state. |
| 1080 | */ |
| 1081 | /* Must avoid static vars within functions since some systems |
| 1082 | #define static as nothing. */ |
| 1083 | static NODE *last_node = NULL; |
| 1084 | |
| 1085 | void |
| 1086 | add_node (node, cur_node_p) |
| 1087 | NODE *node, **cur_node_p; |
| 1088 | { |
| 1089 | register int dif; |
| 1090 | register NODE *cur_node = *cur_node_p; |
| 1091 | |
| 1092 | if (cur_node == NULL) |
| 1093 | { |
| 1094 | *cur_node_p = node; |
| 1095 | last_node = node; |
| 1096 | return; |
| 1097 | } |
| 1098 | |
| 1099 | if (emacs_tags_format) |
| 1100 | { |
| 1101 | /* Etags Mode */ |
| 1102 | if (!last_node) |
| 1103 | fatal ("internal error in add_node"); |
| 1104 | last_node->right = node; |
| 1105 | last_node = node; |
| 1106 | } |
| 1107 | else |
| 1108 | { |
| 1109 | /* Ctags Mode */ |
| 1110 | dif = strcmp (node->name, cur_node->name); |
| 1111 | |
| 1112 | /* |
| 1113 | * If this tag name matches an existing one, then |
| 1114 | * do not add the node, but maybe print a warning. |
| 1115 | */ |
| 1116 | if (!dif) |
| 1117 | { |
| 1118 | if (node->file == cur_node->file) |
| 1119 | { |
| 1120 | if (!no_warnings) |
| 1121 | { |
| 1122 | fprintf (stderr, "Duplicate entry in file %s, line %d: %s\n", |
| 1123 | node->file, lineno, node->name); |
| 1124 | fprintf (stderr, "Second entry ignored\n"); |
| 1125 | } |
| 1126 | return; |
| 1127 | } |
| 1128 | if (!cur_node->been_warned && !no_warnings) |
| 1129 | { |
| 1130 | fprintf (stderr, |
| 1131 | "Duplicate entry in files %s and %s: %s (Warning only)\n", |
| 1132 | node->file, cur_node->file, node->name); |
| 1133 | } |
| 1134 | cur_node->been_warned = TRUE; |
| 1135 | return; |
| 1136 | } |
| 1137 | |
| 1138 | /* Maybe refuse to add duplicate nodes. */ |
| 1139 | if (!permit_duplicates) |
| 1140 | { |
| 1141 | if (!strcmp (node->name, cur_node->name) |
| 1142 | && !strcmp (node->file, cur_node->file)) |
| 1143 | return; |
| 1144 | } |
| 1145 | |
| 1146 | /* Actually add the node */ |
| 1147 | add_node (node, dif < 0 ? &cur_node->left : &cur_node->right); |
| 1148 | } |
| 1149 | } |
| 1150 | \f |
| 1151 | void |
| 1152 | put_entries (node) |
| 1153 | register NODE *node; |
| 1154 | { |
| 1155 | register char *sp; |
| 1156 | |
| 1157 | if (node == NULL) |
| 1158 | return; |
| 1159 | |
| 1160 | /* Output subentries that precede this one */ |
| 1161 | put_entries (node->left); |
| 1162 | |
| 1163 | /* Output this entry */ |
| 1164 | |
| 1165 | if (emacs_tags_format) |
| 1166 | { |
| 1167 | if (node->named) |
| 1168 | { |
| 1169 | fprintf (outf, "%s\177%s\001%d,%d\n", |
| 1170 | node->pat, node->name, |
| 1171 | node->lno, node->cno); |
| 1172 | } |
| 1173 | else |
| 1174 | { |
| 1175 | fprintf (outf, "%s\177%d,%d\n", |
| 1176 | node->pat, |
| 1177 | node->lno, node->cno); |
| 1178 | } |
| 1179 | } |
| 1180 | else if (!cxref_style) |
| 1181 | { |
| 1182 | fprintf (outf, "%s\t%s\t", |
| 1183 | node->name, node->file); |
| 1184 | |
| 1185 | if (node->is_func) |
| 1186 | { /* a function */ |
| 1187 | putc (searchar, outf); |
| 1188 | putc ('^', outf); |
| 1189 | |
| 1190 | for (sp = node->pat; *sp; sp++) |
| 1191 | { |
| 1192 | if (*sp == '\\' || *sp == searchar) |
| 1193 | putc ('\\', outf); |
| 1194 | putc (*sp, outf); |
| 1195 | } |
| 1196 | putc (searchar, outf); |
| 1197 | } |
| 1198 | else |
| 1199 | { /* a typedef; text pattern inadequate */ |
| 1200 | fprintf (outf, "%d", node->lno); |
| 1201 | } |
| 1202 | putc ('\n', outf); |
| 1203 | } |
| 1204 | else if (vgrind_style) |
| 1205 | fprintf (stdout, "%s %s %d\n", |
| 1206 | node->name, node->file, (node->lno + 63) / 64); |
| 1207 | else |
| 1208 | fprintf (stdout, "%-16s %3d %-16s %s\n", |
| 1209 | node->name, node->lno, node->file, node->pat); |
| 1210 | |
| 1211 | /* Output subentries that follow this one */ |
| 1212 | put_entries (node->right); |
| 1213 | } |
| 1214 | |
| 1215 | /* Length of a number's decimal representation. */ |
| 1216 | int |
| 1217 | number_len (num) |
| 1218 | long num; |
| 1219 | { |
| 1220 | int len = 0; |
| 1221 | if (!num) |
| 1222 | return 1; |
| 1223 | for (; num; num /= 10) |
| 1224 | ++len; |
| 1225 | return len; |
| 1226 | } |
| 1227 | |
| 1228 | /* |
| 1229 | * Return total number of characters that put_entries will output for |
| 1230 | * the nodes in the subtree of the specified node. Works only if emacs_tags_format |
| 1231 | * is set, but called only in that case. This count is irrelevant with |
| 1232 | * the new tags.el, but is still supplied for backward compatibility. |
| 1233 | */ |
| 1234 | int |
| 1235 | total_size_of_entries (node) |
| 1236 | register NODE *node; |
| 1237 | { |
| 1238 | register int total; |
| 1239 | |
| 1240 | if (node == NULL) |
| 1241 | return 0; |
| 1242 | |
| 1243 | total = 0; |
| 1244 | for (; node; node = node->right) |
| 1245 | { |
| 1246 | /* Count left subentries. */ |
| 1247 | total += total_size_of_entries (node->left); |
| 1248 | |
| 1249 | /* Count this entry */ |
| 1250 | total += strlen (node->pat) + 1; |
| 1251 | total += number_len ((long) node->lno) + 1 + number_len (node->cno) + 1; |
| 1252 | if (node->named) |
| 1253 | total += 1 + strlen (node->name); /* \001name */ |
| 1254 | } |
| 1255 | |
| 1256 | return total; |
| 1257 | } |
| 1258 | \f |
| 1259 | /* |
| 1260 | * The C symbol tables. |
| 1261 | */ |
| 1262 | |
| 1263 | Stab *C_stab, *C_PLPL_stab, *C_STAR_stab; |
| 1264 | |
| 1265 | /* |
| 1266 | * SYNOPSIS |
| 1267 | * Stab *get_C_stab (int c_ext); |
| 1268 | */ |
| 1269 | #define get_C_stab(c_ext) ((c_ext & C_STAR) ? C_STAR_stab : \ |
| 1270 | (c_ext & C_PLPL) ? C_PLPL_stab : \ |
| 1271 | C_stab) |
| 1272 | |
| 1273 | void |
| 1274 | add_keyword (stab, sym, type) |
| 1275 | Stab *stab; |
| 1276 | char *sym; |
| 1277 | enum sym_type type; |
| 1278 | { |
| 1279 | stab_search (stab, sym, strlen (sym))->type = type; |
| 1280 | } |
| 1281 | |
| 1282 | Stab * |
| 1283 | C_create_stab (c_ext) |
| 1284 | int c_ext; |
| 1285 | { |
| 1286 | Stab *stab; |
| 1287 | |
| 1288 | stab = stab_create (); |
| 1289 | |
| 1290 | /* C, C++ and C* */ |
| 1291 | if (c_ext & C_PLPL) |
| 1292 | add_keyword (stab, "class", st_C_struct); |
| 1293 | if (c_ext & C_STAR) |
| 1294 | add_keyword (stab, "domain", st_C_struct); |
| 1295 | add_keyword (stab, "union", st_C_struct); |
| 1296 | add_keyword (stab, "struct", st_C_struct); |
| 1297 | add_keyword (stab, "enum", st_C_enum); |
| 1298 | add_keyword (stab, "typedef", st_C_typedef); |
| 1299 | add_keyword (stab, "define", st_C_define); |
| 1300 | add_keyword (stab, "long", st_C_typespec); |
| 1301 | add_keyword (stab, "short", st_C_typespec); |
| 1302 | add_keyword (stab, "int", st_C_typespec); |
| 1303 | add_keyword (stab, "char", st_C_typespec); |
| 1304 | add_keyword (stab, "float", st_C_typespec); |
| 1305 | add_keyword (stab, "double", st_C_typespec); |
| 1306 | add_keyword (stab, "signed", st_C_typespec); |
| 1307 | add_keyword (stab, "unsigned", st_C_typespec); |
| 1308 | add_keyword (stab, "auto", st_C_typespec); |
| 1309 | add_keyword (stab, "void", st_C_typespec); |
| 1310 | add_keyword (stab, "extern", st_C_typespec); |
| 1311 | add_keyword (stab, "static", st_C_typespec); |
| 1312 | add_keyword (stab, "const", st_C_typespec); |
| 1313 | add_keyword (stab, "volatile", st_C_typespec); |
| 1314 | |
| 1315 | return stab; |
| 1316 | } |
| 1317 | |
| 1318 | void |
| 1319 | C_create_stabs () |
| 1320 | { |
| 1321 | C_stab = C_create_stab (0); |
| 1322 | C_PLPL_stab = C_create_stab (C_PLPL); |
| 1323 | C_STAR_stab = C_create_stab (C_STAR | C_PLPL); |
| 1324 | } |
| 1325 | \f |
| 1326 | /* |
| 1327 | * C functions are recognized using a simple finite automaton. |
| 1328 | * funcdef is its state variable. |
| 1329 | */ |
| 1330 | typedef enum |
| 1331 | { |
| 1332 | fnone, /* nothing seen */ |
| 1333 | ftagseen, /* function-like tag seen */ |
| 1334 | finlist, /* in parameter list */ |
| 1335 | flistseen, /* after parameter list */ |
| 1336 | fignore, /* before open brace */ |
| 1337 | } FUNCST; |
| 1338 | FUNCST funcdef; |
| 1339 | |
| 1340 | |
| 1341 | /* typedefs are recognized using a simple finite automaton. |
| 1342 | * typeddef is its state variable. |
| 1343 | */ |
| 1344 | typedef enum |
| 1345 | { |
| 1346 | tnone, /* nothing seen */ |
| 1347 | ttypedseen, /* typedef keyword seen */ |
| 1348 | tinbody, /* inside typedef body */ |
| 1349 | tend, /* just before typedef tag */ |
| 1350 | } TYPEDST; |
| 1351 | TYPEDST typdef; |
| 1352 | |
| 1353 | |
| 1354 | /* struct tags for C++ are recognized using another simple |
| 1355 | * finite automaton. `structdef' is its state variable. |
| 1356 | * This machinery is only invoked for C++; otherwise structdef |
| 1357 | * should remain snone. However, this machinery can easily be |
| 1358 | * adapted to find structure tags in normal C code. |
| 1359 | */ |
| 1360 | typedef enum |
| 1361 | { |
| 1362 | snone, /* nothing seen yet */ |
| 1363 | skeyseen, /* struct-like keyword seen */ |
| 1364 | stagseen, /* struct-like tag seen */ |
| 1365 | scolonseen, /* colon seen after struct-like tag */ |
| 1366 | sinbody, /* in struct body: recognize member func defs*/ |
| 1367 | } STRUCTST; |
| 1368 | STRUCTST structdef; |
| 1369 | /* |
| 1370 | * When structdef is stagseen, scolonseen, or sinbody, structtag is the |
| 1371 | * struct tag, and structkey is the preceding struct-like keyword. |
| 1372 | */ |
| 1373 | char structtag[BUFSIZ]; |
| 1374 | Stab_entry *structkey; |
| 1375 | |
| 1376 | /* |
| 1377 | * Yet another little state machine to deal with preprocessor lines. |
| 1378 | */ |
| 1379 | typedef enum |
| 1380 | { |
| 1381 | dnone, /* nothing seen */ |
| 1382 | dsharpseen, /* '#' seen as first char on line */ |
| 1383 | ddefineseen, /* '#' and 'define' seen */ |
| 1384 | dignorerest, /* ignore rest of line */ |
| 1385 | } DEFINEST; |
| 1386 | DEFINEST definedef; |
| 1387 | |
| 1388 | /* |
| 1389 | * Set this to TRUE, and the next token considered is called a function. |
| 1390 | * Used only for GNUmacs's function-defining macros. |
| 1391 | */ |
| 1392 | logical next_token_is_func; |
| 1393 | |
| 1394 | /* |
| 1395 | * TRUE in the rules part of a yacc file, FALSE outside (parse as C). |
| 1396 | */ |
| 1397 | logical yacc_rules; |
| 1398 | |
| 1399 | /* |
| 1400 | * C_entries () |
| 1401 | * This routine finds functions, typedefs, #define's and |
| 1402 | * struct/union/enum definitions in C syntax and adds them |
| 1403 | * to the list. |
| 1404 | */ |
| 1405 | |
| 1406 | #define curlb (lbs[curndx].lb) |
| 1407 | #define othlb (lbs[1-curndx].lb) |
| 1408 | #define newlb (lbs[newndx].lb) |
| 1409 | #define curlinepos (lbs[curndx].linepos) |
| 1410 | #define othlinepos (lbs[1-curndx].linepos) |
| 1411 | #define newlinepos (lbs[newndx].linepos) |
| 1412 | |
| 1413 | #define CNL_SAVE_DEFINEDEF \ |
| 1414 | do { \ |
| 1415 | SET_FILEPOS (curlinepos, inf, charno); \ |
| 1416 | lineno++; \ |
| 1417 | charno += readline (&curlb, inf); \ |
| 1418 | lp = curlb.buffer; \ |
| 1419 | quotednl = FALSE; \ |
| 1420 | newndx = curndx; \ |
| 1421 | } while (FALSE) |
| 1422 | |
| 1423 | #define CNL \ |
| 1424 | do { \ |
| 1425 | CNL_SAVE_DEFINEDEF; \ |
| 1426 | definedef = dnone; \ |
| 1427 | } while (FALSE) |
| 1428 | |
| 1429 | #define MAKE_TAG_FROM_NEW_LB(isfun) pfnote (nameb, isfun, tok.named, \ |
| 1430 | newlb.buffer, tokoff + toklen + 1, tok.lineno, GET_CHARNO (newlinepos)) |
| 1431 | #define MAKE_TAG_FROM_OTH_LB(isfun) pfnote (nameb, isfun, tok.named, \ |
| 1432 | othlb.buffer, tokoff + toklen + 1, tok.lineno, GET_CHARNO (othlinepos)) |
| 1433 | |
| 1434 | void |
| 1435 | C_entries (c_ext) |
| 1436 | int c_ext; /* extension of C? */ |
| 1437 | { |
| 1438 | register char c; /* latest char read; '\0' for end of line */ |
| 1439 | register char *lp; /* pointer one beyond the character `c' */ |
| 1440 | int curndx, newndx; /* indices for current and new lb */ |
| 1441 | TOKEN tok; /* latest token read for funcdef & structdef */ |
| 1442 | char nameb[BUFSIZ]; /* latest token name for funcdef & structdef */ |
| 1443 | register int tokoff; /* offset in line of start of latest token */ |
| 1444 | register int toklen; /* length of latest token */ |
| 1445 | int cblev; /* current curly brace level */ |
| 1446 | logical incomm, inquote, inchar, quotednl, midtoken; |
| 1447 | logical cplpl; |
| 1448 | |
| 1449 | curndx = newndx = 0; |
| 1450 | lineno = 0; |
| 1451 | charno = 0; |
| 1452 | lp = curlb.buffer; |
| 1453 | *lp = 0; |
| 1454 | |
| 1455 | definedef = dnone; funcdef = fnone; typdef= tnone; structdef= snone; |
| 1456 | next_token_is_func = yacc_rules = FALSE; |
| 1457 | midtoken = inquote = inchar = incomm = quotednl = FALSE; |
| 1458 | cblev = 0; |
| 1459 | cplpl = c_ext & C_PLPL; |
| 1460 | |
| 1461 | C_create_stabs (); |
| 1462 | |
| 1463 | while (!feof (inf)) |
| 1464 | { |
| 1465 | c = *lp++; |
| 1466 | if (c == '\\') |
| 1467 | { |
| 1468 | /* If we're at the end of the line, the next character is a |
| 1469 | '\0'; don't skip it, because it's the thing that tells us |
| 1470 | to read the next line. */ |
| 1471 | if (*lp == '\0') |
| 1472 | { |
| 1473 | quotednl = TRUE; |
| 1474 | continue; |
| 1475 | } |
| 1476 | lp++; |
| 1477 | c = ' '; |
| 1478 | } |
| 1479 | else if (incomm) |
| 1480 | { |
| 1481 | switch (c) |
| 1482 | { |
| 1483 | case '*': |
| 1484 | if (*lp == '/') |
| 1485 | { |
| 1486 | c = *lp++; |
| 1487 | incomm = FALSE; |
| 1488 | } |
| 1489 | break; |
| 1490 | case '\0': |
| 1491 | /* Newlines inside comments do not end macro definitions in |
| 1492 | traditional cpp. */ |
| 1493 | CNL_SAVE_DEFINEDEF; |
| 1494 | break; |
| 1495 | } |
| 1496 | continue; |
| 1497 | } |
| 1498 | else if (inquote) |
| 1499 | { |
| 1500 | switch (c) |
| 1501 | { |
| 1502 | case '"': |
| 1503 | inquote = FALSE; |
| 1504 | break; |
| 1505 | case '\0': |
| 1506 | /* Newlines inside strings, do not end macro definitions |
| 1507 | in traditional cpp, even though compilers don't |
| 1508 | usually accept them. */ |
| 1509 | CNL_SAVE_DEFINEDEF; |
| 1510 | break; |
| 1511 | } |
| 1512 | continue; |
| 1513 | } |
| 1514 | else if (inchar) |
| 1515 | { |
| 1516 | if (c == '\'') |
| 1517 | inchar = FALSE; |
| 1518 | continue; |
| 1519 | } |
| 1520 | else |
| 1521 | switch (c) |
| 1522 | { |
| 1523 | case '"': |
| 1524 | inquote = TRUE; |
| 1525 | continue; |
| 1526 | case '\'': |
| 1527 | inchar = TRUE; |
| 1528 | continue; |
| 1529 | case '/': |
| 1530 | if (*lp == '*') |
| 1531 | { |
| 1532 | lp++; |
| 1533 | incomm = TRUE; |
| 1534 | continue; |
| 1535 | } |
| 1536 | else if (cplpl && *lp == '/') |
| 1537 | { |
| 1538 | c = 0; |
| 1539 | break; |
| 1540 | } |
| 1541 | continue; |
| 1542 | case '%': |
| 1543 | if ((c_ext & YACC) && *lp == '%') |
| 1544 | { |
| 1545 | /* entering or exiting rules section in yacc file */ |
| 1546 | lp++; |
| 1547 | definedef = dnone; funcdef = fnone; |
| 1548 | typdef= tnone; structdef= snone; |
| 1549 | next_token_is_func = FALSE; |
| 1550 | midtoken = inquote = inchar = incomm = quotednl = FALSE; |
| 1551 | cblev = 0; |
| 1552 | yacc_rules = !yacc_rules; |
| 1553 | continue; |
| 1554 | } |
| 1555 | case '#': |
| 1556 | if (lp == newlb.buffer + 1 && definedef == dnone) |
| 1557 | definedef = dsharpseen; |
| 1558 | continue; |
| 1559 | } /* switch (c) */ |
| 1560 | |
| 1561 | |
| 1562 | /* Consider token only if some complicated conditions are satisfied. */ |
| 1563 | if (((cblev == 0 && structdef != scolonseen) |
| 1564 | || (cblev == 1 && cplpl && structdef == sinbody)) |
| 1565 | && definedef != dignorerest |
| 1566 | && funcdef != finlist) |
| 1567 | { |
| 1568 | if (midtoken) |
| 1569 | { |
| 1570 | if (endtoken (c)) |
| 1571 | { |
| 1572 | if (cplpl && c == ':' && *lp == ':' && intoken (*(lp + 1))) |
| 1573 | { |
| 1574 | /* |
| 1575 | * This handles :: in the middle, but not at beginning |
| 1576 | * of an identifier. |
| 1577 | */ |
| 1578 | lp += 2; |
| 1579 | toklen += 3; |
| 1580 | } |
| 1581 | else |
| 1582 | { |
| 1583 | logical is_func = FALSE; |
| 1584 | |
| 1585 | tok.lineno = lineno; |
| 1586 | tok.p = newlb.buffer + tokoff; |
| 1587 | tok.len = toklen; |
| 1588 | tok.named = FALSE; |
| 1589 | if (yacc_rules |
| 1590 | || consider_token (c, lp, &tok, |
| 1591 | c_ext, cblev, &is_func)) |
| 1592 | { |
| 1593 | if (structdef == sinbody |
| 1594 | && definedef == dnone |
| 1595 | && is_func) |
| 1596 | /* function defined in C++ class body */ |
| 1597 | { |
| 1598 | tok.named = TRUE; |
| 1599 | sprintf (nameb, "%s::%.*s", |
| 1600 | ((structtag[0] == '\0') |
| 1601 | ? "_anonymous_" : structtag), |
| 1602 | tok.len, tok.p); |
| 1603 | } |
| 1604 | else |
| 1605 | { |
| 1606 | sprintf (nameb, "%.*s", tok.len, tok.p); |
| 1607 | } |
| 1608 | |
| 1609 | if (structdef == stagseen |
| 1610 | || typdef == tend) |
| 1611 | tok.named = TRUE; |
| 1612 | |
| 1613 | if (funcdef == ftagseen |
| 1614 | || structdef == stagseen |
| 1615 | || typdef == tend) |
| 1616 | { |
| 1617 | if (newndx == curndx) |
| 1618 | curndx = 1 - curndx; /* switch line buffers */ |
| 1619 | } |
| 1620 | else |
| 1621 | MAKE_TAG_FROM_NEW_LB (is_func); |
| 1622 | } |
| 1623 | midtoken = FALSE; |
| 1624 | } |
| 1625 | } /* if (endtoken (c)) */ |
| 1626 | else if (intoken (c)) |
| 1627 | { |
| 1628 | toklen++; |
| 1629 | continue; |
| 1630 | } |
| 1631 | } /* if (midtoken) */ |
| 1632 | else if (begtoken (c)) |
| 1633 | { |
| 1634 | switch (funcdef) |
| 1635 | { |
| 1636 | case flistseen: |
| 1637 | MAKE_TAG_FROM_OTH_LB (TRUE); |
| 1638 | funcdef = fignore; |
| 1639 | break; |
| 1640 | case ftagseen: |
| 1641 | funcdef = fnone; |
| 1642 | break; |
| 1643 | } |
| 1644 | if (structdef == stagseen) |
| 1645 | structdef = snone; |
| 1646 | if (!yacc_rules || lp == newlb.buffer + 1) |
| 1647 | { |
| 1648 | tokoff = lp - 1 - newlb.buffer; |
| 1649 | toklen = 1; |
| 1650 | midtoken = TRUE; |
| 1651 | } |
| 1652 | continue; |
| 1653 | } |
| 1654 | } /* if must look at token */ |
| 1655 | |
| 1656 | |
| 1657 | /* Detect end of line, colon, comma, semicolon and various braces |
| 1658 | after having handled the last token on the line.*/ |
| 1659 | switch (c) |
| 1660 | { |
| 1661 | case ':': |
| 1662 | if (structdef == stagseen) |
| 1663 | structdef = scolonseen; |
| 1664 | else if (yacc_rules && funcdef == ftagseen) |
| 1665 | { |
| 1666 | MAKE_TAG_FROM_OTH_LB (FALSE); |
| 1667 | funcdef = fignore; |
| 1668 | } |
| 1669 | break; |
| 1670 | case ';': |
| 1671 | if (cblev == 0 && typdef == tend) |
| 1672 | { |
| 1673 | typdef = tnone; |
| 1674 | MAKE_TAG_FROM_OTH_LB (FALSE); |
| 1675 | } |
| 1676 | if (funcdef != fignore) |
| 1677 | funcdef = fnone; |
| 1678 | /* FALLTHRU */ |
| 1679 | case ',': |
| 1680 | /* FALLTHRU */ |
| 1681 | case '[': |
| 1682 | if (funcdef != finlist && funcdef != fignore) |
| 1683 | funcdef = fnone; |
| 1684 | if (structdef == stagseen) |
| 1685 | structdef = snone; |
| 1686 | break; |
| 1687 | case '(': |
| 1688 | switch (funcdef) |
| 1689 | { |
| 1690 | case ftagseen: |
| 1691 | funcdef = finlist; |
| 1692 | break; |
| 1693 | case finlist: |
| 1694 | case flistseen: |
| 1695 | funcdef = fnone; |
| 1696 | break; |
| 1697 | } |
| 1698 | break; |
| 1699 | case ')': |
| 1700 | if (funcdef == finlist) |
| 1701 | funcdef = flistseen; |
| 1702 | break; |
| 1703 | case '{': |
| 1704 | if (typdef == ttypedseen) |
| 1705 | typdef = tinbody; |
| 1706 | switch (structdef) |
| 1707 | { |
| 1708 | case skeyseen: /* unnamed struct */ |
| 1709 | structtag[0] = '\0'; |
| 1710 | structdef = sinbody; |
| 1711 | break; |
| 1712 | case stagseen: |
| 1713 | case scolonseen: /* named struct */ |
| 1714 | structdef = sinbody; |
| 1715 | MAKE_TAG_FROM_OTH_LB (FALSE); |
| 1716 | break; |
| 1717 | } |
| 1718 | switch (funcdef) |
| 1719 | { |
| 1720 | case flistseen: |
| 1721 | MAKE_TAG_FROM_OTH_LB (TRUE); |
| 1722 | /* FALLTHRU */ |
| 1723 | case fignore: |
| 1724 | funcdef = fnone; |
| 1725 | } |
| 1726 | cblev++; |
| 1727 | break; |
| 1728 | case '*': |
| 1729 | if (funcdef == flistseen) |
| 1730 | { |
| 1731 | MAKE_TAG_FROM_OTH_LB (TRUE); |
| 1732 | funcdef = fignore; |
| 1733 | } |
| 1734 | break; |
| 1735 | case '}': |
| 1736 | if (!noindentypedefs && lp == newlb.buffer + 1) |
| 1737 | cblev = 0; /* reset curly brace level if first column */ |
| 1738 | else if (cblev > 0) |
| 1739 | cblev--; |
| 1740 | if (cblev == 0) |
| 1741 | { |
| 1742 | if (typdef == tinbody) |
| 1743 | typdef = tend; |
| 1744 | structdef = snone; |
| 1745 | (void) strcpy (structtag, "<error 2>"); |
| 1746 | } |
| 1747 | break; |
| 1748 | case '\0': |
| 1749 | /* If a macro spans multiple lines don't reset its state. */ |
| 1750 | if (quotednl) |
| 1751 | CNL_SAVE_DEFINEDEF; |
| 1752 | else |
| 1753 | CNL; |
| 1754 | break; |
| 1755 | } /* switch (c) */ |
| 1756 | |
| 1757 | } /* while not eof */ |
| 1758 | } |
| 1759 | |
| 1760 | /* |
| 1761 | * consider_token () |
| 1762 | * checks to see if the current token is at the start of a |
| 1763 | * function, or corresponds to a typedef, or is a struct/union/enum |
| 1764 | * tag. |
| 1765 | * |
| 1766 | * *IS_FUNC gets TRUE iff the token is a function or macro with args. |
| 1767 | * C_EXT is which language we are looking at. |
| 1768 | * |
| 1769 | * In the future we will need some way to adjust where the end of |
| 1770 | * the token is; for instance, implementing the C++ keyword |
| 1771 | * `operator' properly will adjust the end of the token to be after |
| 1772 | * whatever follows `operator'. |
| 1773 | * |
| 1774 | * Globals |
| 1775 | * funcdef IN OUT |
| 1776 | * structdef IN OUT |
| 1777 | * definedef IN OUT |
| 1778 | * typdef IN OUT |
| 1779 | * next_token_is_func IN OUT |
| 1780 | */ |
| 1781 | |
| 1782 | logical |
| 1783 | consider_token (c, lp, tokp, c_ext, cblev, is_func) |
| 1784 | register char c; /* IN: first char after the token */ |
| 1785 | register char *lp; /* IN: lp points to 2nd char after the token */ |
| 1786 | register TOKEN *tokp; /* IN: token pointer */ |
| 1787 | int c_ext; /* IN: C extensions mask */ |
| 1788 | int cblev; /* IN: curly brace level */ |
| 1789 | logical *is_func; /* OUT */ |
| 1790 | { |
| 1791 | Stab_entry *tokse = stab_find (get_C_stab (c_ext), tokp->p, tokp->len); |
| 1792 | enum sym_type toktype = stab_type (tokse); |
| 1793 | |
| 1794 | /* |
| 1795 | * Advance the definedef state machine. |
| 1796 | */ |
| 1797 | switch (definedef) |
| 1798 | { |
| 1799 | case dnone: |
| 1800 | /* We're not on a preprocessor line. */ |
| 1801 | break; |
| 1802 | case dsharpseen: |
| 1803 | if (toktype == st_C_define) |
| 1804 | { |
| 1805 | definedef = ddefineseen; |
| 1806 | } |
| 1807 | else |
| 1808 | { |
| 1809 | definedef = dignorerest; |
| 1810 | } |
| 1811 | return (FALSE); |
| 1812 | case ddefineseen: |
| 1813 | /* |
| 1814 | * Make a tag for any macro. |
| 1815 | */ |
| 1816 | definedef = dignorerest; |
| 1817 | *is_func = (c == '('); |
| 1818 | if (!*is_func && !constantypedefs) |
| 1819 | return (FALSE); |
| 1820 | else |
| 1821 | return (TRUE); |
| 1822 | case dignorerest: |
| 1823 | return (FALSE); |
| 1824 | default: |
| 1825 | error ("internal error: definedef value"); |
| 1826 | } |
| 1827 | |
| 1828 | /* |
| 1829 | * Now typedefs |
| 1830 | */ |
| 1831 | switch (typdef) |
| 1832 | { |
| 1833 | case tnone: |
| 1834 | if (toktype == st_C_typedef) |
| 1835 | { |
| 1836 | if (typedefs) |
| 1837 | typdef = ttypedseen; |
| 1838 | return (FALSE); |
| 1839 | } |
| 1840 | break; |
| 1841 | case ttypedseen: |
| 1842 | switch (toktype) |
| 1843 | { |
| 1844 | case st_none: |
| 1845 | case st_C_typespec: |
| 1846 | typdef = tend; |
| 1847 | break; |
| 1848 | case st_C_struct: |
| 1849 | case st_C_enum: |
| 1850 | break; |
| 1851 | } |
| 1852 | /* Do not return here, so the structdef stuff has a chance. */ |
| 1853 | break; |
| 1854 | case tend: |
| 1855 | switch (toktype) |
| 1856 | { |
| 1857 | case st_C_typespec: |
| 1858 | case st_C_struct: |
| 1859 | case st_C_enum: |
| 1860 | return (FALSE); |
| 1861 | } |
| 1862 | return (TRUE); |
| 1863 | } |
| 1864 | |
| 1865 | /* |
| 1866 | * This structdef business is currently only invoked when cblev==0. |
| 1867 | * It should be recursively invoked whatever the curly brace level, |
| 1868 | * and a stack of states kept, to allow for definitions of structs |
| 1869 | * within structs. |
| 1870 | * |
| 1871 | * This structdef business is NOT invoked when we are ctags and the |
| 1872 | * file is plain C. This is because a struct tag may have the same |
| 1873 | * name as another tag, and this loses with ctags. |
| 1874 | * |
| 1875 | * This if statement deals with the typdef state machine as |
| 1876 | * follows: if typdef==ttypedseen and token is struct/union/class/enum, |
| 1877 | * return (FALSE). All the other code here is for the structdef |
| 1878 | * state machine. |
| 1879 | */ |
| 1880 | switch (toktype) |
| 1881 | { |
| 1882 | case st_C_struct: |
| 1883 | case st_C_enum: |
| 1884 | if (typdef == ttypedseen |
| 1885 | || (typedefs_and_cplusplus && cblev == 0 && structdef == snone)) |
| 1886 | { |
| 1887 | structdef = skeyseen; |
| 1888 | structkey = tokse; |
| 1889 | } |
| 1890 | return (FALSE); |
| 1891 | } |
| 1892 | if (structdef == skeyseen) |
| 1893 | { |
| 1894 | if (stab_type (structkey) == st_C_struct) |
| 1895 | { |
| 1896 | (void) strncpy (structtag, tokp->p, tokp->len); |
| 1897 | structtag[tokp->len] = '\0'; /* for struct/union/class */ |
| 1898 | } |
| 1899 | else |
| 1900 | { |
| 1901 | structtag[0] = '\0'; /* for enum (why is it treated differently?) */ |
| 1902 | } |
| 1903 | structdef = stagseen; |
| 1904 | return (TRUE); |
| 1905 | } |
| 1906 | |
| 1907 | /* Avoid entering funcdef stuff if typdef is going on. */ |
| 1908 | if (typdef != tnone) |
| 1909 | { |
| 1910 | definedef = dnone; |
| 1911 | return (FALSE); |
| 1912 | } |
| 1913 | |
| 1914 | /* Detect GNUmacs's function-defining macros. */ |
| 1915 | if (definedef == dnone) |
| 1916 | { |
| 1917 | if (strneq (tokp->p, "DEF", 3) |
| 1918 | || strneq (tokp->p, "ENTRY", 5) |
| 1919 | || strneq (tokp->p, "SYSCALL", 7) |
| 1920 | || strneq (tokp->p, "PSEUDO", 6)) |
| 1921 | { |
| 1922 | next_token_is_func = TRUE; |
| 1923 | return (FALSE); |
| 1924 | } |
| 1925 | if (strneq (tokp->p, "EXFUN", 5)) |
| 1926 | { |
| 1927 | next_token_is_func = FALSE; |
| 1928 | return (FALSE); |
| 1929 | } |
| 1930 | } |
| 1931 | if (next_token_is_func) |
| 1932 | { |
| 1933 | next_token_is_func = FALSE; |
| 1934 | funcdef = fnone; |
| 1935 | *is_func = TRUE; /* to force search string in ctags */ |
| 1936 | return (TRUE); |
| 1937 | } |
| 1938 | |
| 1939 | /* A function? */ |
| 1940 | switch (toktype) |
| 1941 | { |
| 1942 | case st_C_typespec: |
| 1943 | funcdef = fnone; /* should be useless */ |
| 1944 | return (FALSE); |
| 1945 | default: |
| 1946 | if (funcdef == fnone) |
| 1947 | { |
| 1948 | funcdef = ftagseen; |
| 1949 | *is_func = TRUE; |
| 1950 | return (TRUE); |
| 1951 | } |
| 1952 | } |
| 1953 | |
| 1954 | return (FALSE); |
| 1955 | } |
| 1956 | \f |
| 1957 | /* Fortran parsing */ |
| 1958 | |
| 1959 | char *dbp; |
| 1960 | int pfcnt; |
| 1961 | |
| 1962 | int |
| 1963 | PF_funcs (fi) |
| 1964 | FILE *fi; |
| 1965 | { |
| 1966 | lineno = 0; |
| 1967 | charno = 0; |
| 1968 | pfcnt = 0; |
| 1969 | |
| 1970 | while (!feof (fi)) |
| 1971 | { |
| 1972 | lineno++; |
| 1973 | linecharno = charno; |
| 1974 | charno += readline (&lb, fi); |
| 1975 | dbp = lb.buffer; |
| 1976 | if (*dbp == '%') |
| 1977 | dbp++; /* Ratfor escape to fortran */ |
| 1978 | while (isspace (*dbp)) |
| 1979 | dbp++; |
| 1980 | if (*dbp == 0) |
| 1981 | continue; |
| 1982 | switch (*dbp | ' ') |
| 1983 | { |
| 1984 | case 'i': |
| 1985 | if (tail ("integer")) |
| 1986 | takeprec (); |
| 1987 | break; |
| 1988 | case 'r': |
| 1989 | if (tail ("real")) |
| 1990 | takeprec (); |
| 1991 | break; |
| 1992 | case 'l': |
| 1993 | if (tail ("logical")) |
| 1994 | takeprec (); |
| 1995 | break; |
| 1996 | case 'c': |
| 1997 | if (tail ("complex") || tail ("character")) |
| 1998 | takeprec (); |
| 1999 | break; |
| 2000 | case 'd': |
| 2001 | if (tail ("double")) |
| 2002 | { |
| 2003 | while (isspace (*dbp)) |
| 2004 | dbp++; |
| 2005 | if (*dbp == 0) |
| 2006 | continue; |
| 2007 | if (tail ("precision")) |
| 2008 | break; |
| 2009 | continue; |
| 2010 | } |
| 2011 | break; |
| 2012 | } |
| 2013 | while (isspace (*dbp)) |
| 2014 | dbp++; |
| 2015 | if (*dbp == 0) |
| 2016 | continue; |
| 2017 | switch (*dbp | ' ') |
| 2018 | { |
| 2019 | case 'f': |
| 2020 | if (tail ("function")) |
| 2021 | getit (); |
| 2022 | continue; |
| 2023 | case 's': |
| 2024 | if (tail ("subroutine")) |
| 2025 | getit (); |
| 2026 | continue; |
| 2027 | case 'e': |
| 2028 | if (tail ("entry")) |
| 2029 | getit (); |
| 2030 | continue; |
| 2031 | case 'p': |
| 2032 | if (tail ("program")) |
| 2033 | { |
| 2034 | getit (); |
| 2035 | continue; |
| 2036 | } |
| 2037 | if (tail ("procedure")) |
| 2038 | getit (); |
| 2039 | continue; |
| 2040 | } |
| 2041 | } |
| 2042 | return (pfcnt); |
| 2043 | } |
| 2044 | |
| 2045 | logical |
| 2046 | tail (cp) |
| 2047 | char *cp; |
| 2048 | { |
| 2049 | register int len = 0; |
| 2050 | |
| 2051 | while (*cp && (*cp & ~' ') == ((*(dbp + len)) & ~' ')) |
| 2052 | cp++, len++; |
| 2053 | if (*cp == 0) |
| 2054 | { |
| 2055 | dbp += len; |
| 2056 | return (1); |
| 2057 | } |
| 2058 | return (0); |
| 2059 | } |
| 2060 | |
| 2061 | void |
| 2062 | takeprec () |
| 2063 | { |
| 2064 | while (isspace (*dbp)) |
| 2065 | dbp++; |
| 2066 | if (*dbp != '*') |
| 2067 | return; |
| 2068 | dbp++; |
| 2069 | while (isspace (*dbp)) |
| 2070 | dbp++; |
| 2071 | if (!isdigit (*dbp)) |
| 2072 | { |
| 2073 | --dbp; /* force failure */ |
| 2074 | return; |
| 2075 | } |
| 2076 | do |
| 2077 | dbp++; |
| 2078 | while (isdigit (*dbp)); |
| 2079 | } |
| 2080 | |
| 2081 | void |
| 2082 | getit () |
| 2083 | { |
| 2084 | register char *cp; |
| 2085 | char c; |
| 2086 | char nambuf[BUFSIZ]; |
| 2087 | |
| 2088 | while (isspace (*dbp)) |
| 2089 | dbp++; |
| 2090 | if (*dbp == 0 |
| 2091 | || (!isalpha (*dbp) |
| 2092 | && *dbp != '_' |
| 2093 | && *dbp != '$')) |
| 2094 | return; |
| 2095 | for (cp = dbp + 1; *cp && (isalpha (*cp) || isdigit (*cp) |
| 2096 | || (*cp == '_') || (*cp == '$')); cp++) |
| 2097 | continue; |
| 2098 | c = cp[0]; |
| 2099 | cp[0] = 0; |
| 2100 | (void) strcpy (nambuf, dbp); |
| 2101 | cp[0] = c; |
| 2102 | pfnote (nambuf, TRUE, FALSE, lb.buffer, cp - lb.buffer + 1, lineno, linecharno); |
| 2103 | pfcnt++; |
| 2104 | } |
| 2105 | |
| 2106 | /* Handle a file of assembler code. */ |
| 2107 | |
| 2108 | void |
| 2109 | Asm_funcs (fi) |
| 2110 | FILE *fi; |
| 2111 | { |
| 2112 | int i; |
| 2113 | register char c; |
| 2114 | |
| 2115 | lineno = 0; |
| 2116 | charno = 0; |
| 2117 | pfcnt = 0; |
| 2118 | |
| 2119 | while (!feof (fi)) |
| 2120 | { |
| 2121 | lineno++; |
| 2122 | linecharno = charno; |
| 2123 | charno += readline (&lb, fi); |
| 2124 | dbp = lb.buffer; |
| 2125 | |
| 2126 | for (i = 0; ((c = dbp[i]) && !isspace (c)) && (c != ':'); i++) |
| 2127 | ; |
| 2128 | |
| 2129 | if ((i > 0) && (c == ':')) |
| 2130 | getit (); |
| 2131 | } |
| 2132 | } |
| 2133 | \f |
| 2134 | /* Added by Mosur Mohan, 4/22/88 */ |
| 2135 | /* Pascal parsing */ |
| 2136 | |
| 2137 | #define GET_NEW_LINE \ |
| 2138 | { \ |
| 2139 | linecharno = charno; lineno++; \ |
| 2140 | charno += 1 + readline (&lb, inf); \ |
| 2141 | dbp = lb.buffer; \ |
| 2142 | } |
| 2143 | |
| 2144 | /* Locates tags for procedures & functions. |
| 2145 | * Doesn't do any type- or var-definitions. |
| 2146 | * It does look for the keyword "extern" or "forward" |
| 2147 | * immediately following the procedure statement; |
| 2148 | * if found, the tag is skipped. |
| 2149 | */ |
| 2150 | |
| 2151 | void |
| 2152 | PAS_funcs (fi) |
| 2153 | FILE *fi; |
| 2154 | { |
| 2155 | struct linebuffer tline; /* mostly copied from C_entries */ |
| 2156 | long save_lcno; |
| 2157 | int save_lineno; |
| 2158 | char c, *cp; |
| 2159 | char nambuf[BUFSIZ]; |
| 2160 | |
| 2161 | logical /* each of these flags is TRUE iff: */ |
| 2162 | incomm1, /* point is inside {..} comment */ |
| 2163 | incomm2, /* point is inside (*..*) comment */ |
| 2164 | inquote, /* point is inside '..' string */ |
| 2165 | get_tagname, /* point is after PROCEDURE/FUNCTION */ |
| 2166 | /* keyword, so next item = potential tag */ |
| 2167 | found_tag, /* point is after a potential tag */ |
| 2168 | inparms, /* point is within parameter-list */ |
| 2169 | verify_tag; /* point has passed the parm-list, so the */ |
| 2170 | /* next token will determine whether */ |
| 2171 | /* this is a FORWARD/EXTERN to be */ |
| 2172 | /* ignored, or whether it is a real tag */ |
| 2173 | |
| 2174 | lineno = 0; |
| 2175 | charno = 0; |
| 2176 | dbp = lb.buffer; |
| 2177 | *dbp = 0; |
| 2178 | initbuffer (&tline); |
| 2179 | |
| 2180 | incomm1 = incomm2 = inquote = FALSE; |
| 2181 | found_tag = FALSE; /* have a proc name; check if extern */ |
| 2182 | get_tagname = FALSE; /* have found "procedure" keyword */ |
| 2183 | inparms = FALSE; /* found '(' after "proc" */ |
| 2184 | verify_tag = FALSE; /* check if "extern" is ahead */ |
| 2185 | |
| 2186 | /* long main loop to get next char */ |
| 2187 | while (!feof (fi)) |
| 2188 | { |
| 2189 | c = *dbp++; |
| 2190 | if (c == 0) /* if end of line */ |
| 2191 | { |
| 2192 | GET_NEW_LINE; |
| 2193 | if (*dbp == 0) |
| 2194 | continue; |
| 2195 | if (!((found_tag && verify_tag) || |
| 2196 | get_tagname)) |
| 2197 | c = *dbp++; /* only if don't need *dbp pointing */ |
| 2198 | /* to the beginning of the name of */ |
| 2199 | /* the procedure or function */ |
| 2200 | } |
| 2201 | if (incomm1) /* within { - } comments */ |
| 2202 | { |
| 2203 | if (c == '}') |
| 2204 | incomm1 = FALSE; |
| 2205 | continue; |
| 2206 | } |
| 2207 | else if (incomm2) /* within (* - *) comments */ |
| 2208 | { |
| 2209 | if (c == '*') |
| 2210 | { |
| 2211 | while ((c = *dbp++) == '*') |
| 2212 | continue; |
| 2213 | if (c == 0) |
| 2214 | GET_NEW_LINE; |
| 2215 | if (c == ')') |
| 2216 | incomm2 = FALSE; |
| 2217 | } |
| 2218 | continue; |
| 2219 | } |
| 2220 | else if (inquote) |
| 2221 | { |
| 2222 | if (c == '\'') |
| 2223 | inquote = FALSE; |
| 2224 | continue; |
| 2225 | } |
| 2226 | else |
| 2227 | switch (c) |
| 2228 | { |
| 2229 | case '\'': |
| 2230 | inquote = TRUE; /* found first quote */ |
| 2231 | continue; |
| 2232 | case '{': /* found open-{-comment */ |
| 2233 | incomm1 = TRUE; |
| 2234 | continue; |
| 2235 | case '(': |
| 2236 | if (*dbp == '*') /* found open-(*-comment */ |
| 2237 | { |
| 2238 | incomm2 = TRUE; |
| 2239 | dbp++; |
| 2240 | } |
| 2241 | else if (found_tag) /* found '(' after tag, i.e., parm-list */ |
| 2242 | inparms = TRUE; |
| 2243 | continue; |
| 2244 | case ')': /* end of parms list */ |
| 2245 | if (inparms) |
| 2246 | inparms = FALSE; |
| 2247 | continue; |
| 2248 | case ';': |
| 2249 | if ((found_tag) && (!inparms)) /* end of proc or fn stmt */ |
| 2250 | { |
| 2251 | verify_tag = TRUE; |
| 2252 | break; |
| 2253 | } |
| 2254 | continue; |
| 2255 | } |
| 2256 | if ((found_tag) && (verify_tag) && (*dbp != ' ')) |
| 2257 | { |
| 2258 | /* check if this is an "extern" declaration */ |
| 2259 | if (*dbp == 0) |
| 2260 | continue; |
| 2261 | if ((*dbp == 'e') || (*dbp == 'E')) |
| 2262 | { |
| 2263 | if (tail ("extern")) /* superfluous, really! */ |
| 2264 | { |
| 2265 | found_tag = FALSE; |
| 2266 | verify_tag = FALSE; |
| 2267 | } |
| 2268 | } |
| 2269 | else if ((*dbp == 'f') || (*dbp == 'F')) |
| 2270 | { |
| 2271 | if (tail ("forward")) /* check for forward reference */ |
| 2272 | { |
| 2273 | found_tag = FALSE; |
| 2274 | verify_tag = FALSE; |
| 2275 | } |
| 2276 | } |
| 2277 | if ((found_tag) && (verify_tag)) /* not external proc, so make tag */ |
| 2278 | { |
| 2279 | found_tag = FALSE; |
| 2280 | verify_tag = FALSE; |
| 2281 | pfnote (nambuf, TRUE, FALSE, |
| 2282 | tline.buffer, cp - tline.buffer + 1, |
| 2283 | save_lineno, save_lcno); |
| 2284 | continue; |
| 2285 | } |
| 2286 | } |
| 2287 | if (get_tagname) /* grab name of proc or fn */ |
| 2288 | { |
| 2289 | if (*dbp == 0) |
| 2290 | continue; |
| 2291 | |
| 2292 | /* save all values for later tagging */ |
| 2293 | tline.size = lb.size; |
| 2294 | strcpy (tline.buffer, lb.buffer); |
| 2295 | save_lineno = lineno; |
| 2296 | save_lcno = linecharno; |
| 2297 | |
| 2298 | /* grab block name */ |
| 2299 | for (cp = dbp + 1; *cp && (!endtoken (*cp)); cp++) |
| 2300 | continue; |
| 2301 | c = cp[0]; |
| 2302 | cp[0] = 0; |
| 2303 | strcpy (nambuf, dbp); |
| 2304 | cp[0] = c; |
| 2305 | dbp = cp; /* restore dbp to e-o-token */ |
| 2306 | get_tagname = FALSE; |
| 2307 | found_tag = TRUE; |
| 2308 | continue; |
| 2309 | |
| 2310 | /* and proceed to check for "extern" */ |
| 2311 | } |
| 2312 | if ((!incomm1) && (!incomm2) && (!inquote) && |
| 2313 | (!found_tag) && (!get_tagname)) |
| 2314 | { |
| 2315 | /* check for proc/fn keywords */ |
| 2316 | switch (c | ' ') |
| 2317 | { |
| 2318 | case 'p': |
| 2319 | if (tail ("rocedure")) /* c = 'p', dbp has advanced */ |
| 2320 | get_tagname = TRUE; |
| 2321 | continue; |
| 2322 | case 'f': |
| 2323 | if (tail ("unction")) |
| 2324 | get_tagname = TRUE; |
| 2325 | continue; |
| 2326 | } |
| 2327 | } |
| 2328 | } /* while not e-o-f */ |
| 2329 | } |
| 2330 | \f |
| 2331 | /* |
| 2332 | * lisp tag functions |
| 2333 | * just look for (def or (DEF |
| 2334 | */ |
| 2335 | |
| 2336 | void |
| 2337 | L_funcs (fi) |
| 2338 | FILE *fi; |
| 2339 | { |
| 2340 | lineno = 0; |
| 2341 | charno = 0; |
| 2342 | pfcnt = 0; |
| 2343 | |
| 2344 | while (!feof (fi)) |
| 2345 | { |
| 2346 | lineno++; |
| 2347 | linecharno = charno; |
| 2348 | charno += readline (&lb, fi); |
| 2349 | dbp = lb.buffer; |
| 2350 | if (dbp[0] == '(') |
| 2351 | { |
| 2352 | if (L_isdef (dbp)) |
| 2353 | { |
| 2354 | while (!isspace (*dbp)) |
| 2355 | dbp++; |
| 2356 | while (isspace (*dbp)) |
| 2357 | dbp++; |
| 2358 | L_getit (); |
| 2359 | } |
| 2360 | else |
| 2361 | { |
| 2362 | /* Check for (foo::defmumble name-defined ... */ |
| 2363 | do |
| 2364 | dbp++; |
| 2365 | while (*dbp && !isspace (*dbp) |
| 2366 | && *dbp != ':' && *dbp != '(' && *dbp != ')'); |
| 2367 | if (*dbp == ':') |
| 2368 | { |
| 2369 | do |
| 2370 | dbp++; |
| 2371 | while (*dbp == ':'); |
| 2372 | |
| 2373 | if (L_isdef (dbp - 1)) |
| 2374 | { |
| 2375 | while (!isspace (*dbp)) |
| 2376 | dbp++; |
| 2377 | while (isspace (*dbp)) |
| 2378 | dbp++; |
| 2379 | L_getit (); |
| 2380 | } |
| 2381 | } |
| 2382 | } |
| 2383 | } |
| 2384 | } |
| 2385 | } |
| 2386 | |
| 2387 | int |
| 2388 | L_isdef (dbp) |
| 2389 | register char *dbp; |
| 2390 | { |
| 2391 | return ((dbp[1] == 'd' || dbp[1] == 'D') |
| 2392 | && (dbp[2] == 'e' || dbp[2] == 'E') |
| 2393 | && (dbp[3] == 'f' || dbp[3] == 'F')); |
| 2394 | } |
| 2395 | |
| 2396 | int |
| 2397 | L_isquote (dbp) |
| 2398 | register char *dbp; |
| 2399 | { |
| 2400 | return ((*(++dbp) == 'q' || *dbp == 'Q') |
| 2401 | && (*(++dbp) == 'u' || *dbp == 'U') |
| 2402 | && (*(++dbp) == 'o' || *dbp == 'O') |
| 2403 | && (*(++dbp) == 't' || *dbp == 'T') |
| 2404 | && (*(++dbp) == 'e' || *dbp == 'E') |
| 2405 | && isspace(*(++dbp))); |
| 2406 | } |
| 2407 | |
| 2408 | void |
| 2409 | L_getit () |
| 2410 | { |
| 2411 | register char *cp; |
| 2412 | char c; |
| 2413 | char nambuf[BUFSIZ]; |
| 2414 | |
| 2415 | if (*dbp == '\'') /* Skip prefix quote */ |
| 2416 | dbp++; |
| 2417 | else if (*dbp == '(' && L_isquote (dbp)) /* Skip "(quote " */ |
| 2418 | { |
| 2419 | dbp += 7; |
| 2420 | while (isspace(*dbp)) |
| 2421 | dbp++; |
| 2422 | } |
| 2423 | for (cp = dbp /*+1*/; *cp && *cp != '(' && *cp != ' ' && *cp != ')'; cp++) |
| 2424 | continue; |
| 2425 | if (cp == dbp) |
| 2426 | return; |
| 2427 | |
| 2428 | c = cp[0]; |
| 2429 | cp[0] = 0; |
| 2430 | (void) strcpy (nambuf, dbp); |
| 2431 | cp[0] = c; |
| 2432 | pfnote (nambuf, TRUE, FALSE, lb.buffer, |
| 2433 | cp - lb.buffer + 1, lineno, linecharno); |
| 2434 | pfcnt++; |
| 2435 | } |
| 2436 | \f |
| 2437 | /* |
| 2438 | * Scheme tag functions |
| 2439 | * look for (def... xyzzy |
| 2440 | * look for (def... (xyzzy |
| 2441 | * look for (def ... ((...(xyzzy .... |
| 2442 | * look for (set! xyzzy |
| 2443 | */ |
| 2444 | |
| 2445 | static void get_scheme (); |
| 2446 | |
| 2447 | void |
| 2448 | Scheme_funcs (fi) |
| 2449 | FILE *fi; |
| 2450 | { |
| 2451 | lineno = 0; |
| 2452 | charno = 0; |
| 2453 | pfcnt = 0; |
| 2454 | |
| 2455 | while (!feof (fi)) |
| 2456 | { |
| 2457 | lineno++; |
| 2458 | linecharno = charno; |
| 2459 | charno += readline (&lb, fi); |
| 2460 | dbp = lb.buffer; |
| 2461 | if (dbp[0] == '(' && |
| 2462 | (dbp[1] == 'D' || dbp[1] == 'd') && |
| 2463 | (dbp[2] == 'E' || dbp[2] == 'e') && |
| 2464 | (dbp[3] == 'F' || dbp[3] == 'f')) |
| 2465 | { |
| 2466 | while (!isspace (*dbp)) |
| 2467 | dbp++; |
| 2468 | /* Skip over open parens and white space */ |
| 2469 | while (*dbp && (isspace (*dbp) || *dbp == '(')) |
| 2470 | dbp++; |
| 2471 | get_scheme (); |
| 2472 | } |
| 2473 | if (dbp[0] == '(' && |
| 2474 | (dbp[1] == 'S' || dbp[1] == 's') && |
| 2475 | (dbp[2] == 'E' || dbp[2] == 'e') && |
| 2476 | (dbp[3] == 'T' || dbp[3] == 't') && |
| 2477 | (dbp[4] == '!' || dbp[4] == '!') && |
| 2478 | (isspace (dbp[5]))) |
| 2479 | { |
| 2480 | while (!isspace (*dbp)) |
| 2481 | dbp++; |
| 2482 | /* Skip over white space */ |
| 2483 | while (isspace (*dbp)) |
| 2484 | dbp++; |
| 2485 | get_scheme (); |
| 2486 | } |
| 2487 | } |
| 2488 | } |
| 2489 | |
| 2490 | static void |
| 2491 | get_scheme () |
| 2492 | { |
| 2493 | register char *cp; |
| 2494 | char c; |
| 2495 | char nambuf[BUFSIZ]; |
| 2496 | |
| 2497 | if (*dbp == 0) |
| 2498 | return; |
| 2499 | /* Go till you get to white space or a syntactic break */ |
| 2500 | for (cp = dbp + 1; *cp && *cp != '(' && *cp != ')' && !isspace (*cp); cp++) |
| 2501 | continue; |
| 2502 | /* Null terminate the string there. */ |
| 2503 | c = cp[0]; |
| 2504 | cp[0] = 0; |
| 2505 | /* Copy the string */ |
| 2506 | strcpy (nambuf, dbp); |
| 2507 | /* Unterminate the string */ |
| 2508 | cp[0] = c; |
| 2509 | /* Announce the change */ |
| 2510 | pfnote (nambuf, TRUE, FALSE, lb.buffer, cp - lb.buffer + 1, lineno, linecharno); |
| 2511 | pfcnt++; |
| 2512 | } |
| 2513 | \f |
| 2514 | /* Find tags in TeX and LaTeX input files. */ |
| 2515 | |
| 2516 | /* TEX_toktab is a table of TeX control sequences that define tags. |
| 2517 | Each TEX_tabent records one such control sequence. |
| 2518 | CONVERT THIS TO USE THE Stab TYPE!! */ |
| 2519 | |
| 2520 | struct TEX_tabent |
| 2521 | { |
| 2522 | char *name; |
| 2523 | int len; |
| 2524 | }; |
| 2525 | |
| 2526 | struct TEX_tabent *TEX_toktab = NULL; /* Table with tag tokens */ |
| 2527 | |
| 2528 | /* Default set of control sequences to put into TEX_toktab. |
| 2529 | The value of environment var TEXTAGS is prepended to this. */ |
| 2530 | |
| 2531 | static char *TEX_defenv = |
| 2532 | ":chapter:section:subsection:subsubsection:eqno:label:ref:cite:bibitem:typeout"; |
| 2533 | |
| 2534 | void TEX_mode (); |
| 2535 | struct TEX_tabent *TEX_decode_env (); |
| 2536 | void TEX_getit (); |
| 2537 | int TEX_Token (); |
| 2538 | |
| 2539 | static char TEX_esc = '\\'; |
| 2540 | static char TEX_opgrp = '{'; |
| 2541 | static char TEX_clgrp = '}'; |
| 2542 | |
| 2543 | /* |
| 2544 | * TeX/LaTeX scanning loop. |
| 2545 | */ |
| 2546 | |
| 2547 | void |
| 2548 | TEX_funcs (fi) |
| 2549 | FILE *fi; |
| 2550 | { |
| 2551 | char *lasthit; |
| 2552 | |
| 2553 | lineno = 0; |
| 2554 | charno = 0; |
| 2555 | pfcnt = 0; |
| 2556 | |
| 2557 | /* Select either \ or ! as escape character. */ |
| 2558 | TEX_mode (fi); |
| 2559 | |
| 2560 | /* Initialize token table once from environment. */ |
| 2561 | if (!TEX_toktab) |
| 2562 | TEX_toktab = TEX_decode_env ("TEXTAGS", TEX_defenv); |
| 2563 | |
| 2564 | while (!feof (fi)) |
| 2565 | { /* Scan each line in file */ |
| 2566 | lineno++; |
| 2567 | linecharno = charno; |
| 2568 | charno += readline (&lb, fi); |
| 2569 | dbp = lb.buffer; |
| 2570 | lasthit = dbp; |
| 2571 | while (dbp = etags_index (dbp, TEX_esc)) /* Look at each escape in line */ |
| 2572 | { |
| 2573 | register int i; |
| 2574 | |
| 2575 | if (!*(++dbp)) |
| 2576 | break; |
| 2577 | linecharno += dbp - lasthit; |
| 2578 | lasthit = dbp; |
| 2579 | i = TEX_Token (lasthit); |
| 2580 | if (0 <= i) |
| 2581 | { |
| 2582 | TEX_getit (lasthit, TEX_toktab[i].len); |
| 2583 | break; /* We only save a line once */ |
| 2584 | } |
| 2585 | } |
| 2586 | } |
| 2587 | } |
| 2588 | |
| 2589 | #define TEX_LESC '\\' |
| 2590 | #define TEX_SESC '!' |
| 2591 | #define TEX_cmt '%' |
| 2592 | |
| 2593 | /* Figure out whether TeX's escapechar is '\\' or '!' and set grouping */ |
| 2594 | /* chars accordingly. */ |
| 2595 | |
| 2596 | void |
| 2597 | TEX_mode (f) |
| 2598 | FILE *f; |
| 2599 | { |
| 2600 | int c; |
| 2601 | |
| 2602 | while ((c = getc (f)) != EOF) |
| 2603 | { |
| 2604 | /* Skip to next line if we hit the TeX comment char. */ |
| 2605 | if (c == TEX_cmt) |
| 2606 | while (c != '\n') |
| 2607 | c = getc (f); |
| 2608 | else if (c == TEX_LESC || c == TEX_SESC ) |
| 2609 | break; |
| 2610 | } |
| 2611 | |
| 2612 | if (c == TEX_LESC) |
| 2613 | { |
| 2614 | TEX_esc = TEX_LESC; |
| 2615 | TEX_opgrp = '{'; |
| 2616 | TEX_clgrp = '}'; |
| 2617 | } |
| 2618 | else |
| 2619 | { |
| 2620 | TEX_esc = TEX_SESC; |
| 2621 | TEX_opgrp = '<'; |
| 2622 | TEX_clgrp = '>'; |
| 2623 | } |
| 2624 | rewind (f); |
| 2625 | } |
| 2626 | |
| 2627 | /* Read environment and prepend it to the default string. */ |
| 2628 | /* Build token table. */ |
| 2629 | |
| 2630 | struct TEX_tabent * |
| 2631 | TEX_decode_env (evarname, defenv) |
| 2632 | char *evarname; |
| 2633 | char *defenv; |
| 2634 | { |
| 2635 | register char *env, *p; |
| 2636 | |
| 2637 | struct TEX_tabent *tab; |
| 2638 | int size, i; |
| 2639 | |
| 2640 | /* Append default string to environment. */ |
| 2641 | env = getenv (evarname); |
| 2642 | if (!env) |
| 2643 | env = defenv; |
| 2644 | else |
| 2645 | env = concat (env, defenv, ""); |
| 2646 | |
| 2647 | /* Allocate a token table */ |
| 2648 | for (size = 1, p = env; p;) |
| 2649 | if ((p = etags_index (p, ':')) && *(++p)) |
| 2650 | size++; |
| 2651 | /* Add 1 to leave room for null terminator. */ |
| 2652 | tab = xnew (size + 1, struct TEX_tabent); |
| 2653 | |
| 2654 | /* Unpack environment string into token table. Be careful about */ |
| 2655 | /* zero-length strings (leading ':', "::" and trailing ':') */ |
| 2656 | for (i = 0; *env;) |
| 2657 | { |
| 2658 | p = etags_index (env, ':'); |
| 2659 | if (!p) /* End of environment string. */ |
| 2660 | p = env + strlen (env); |
| 2661 | if (p - env > 0) |
| 2662 | { /* Only non-zero strings. */ |
| 2663 | tab[i].name = savenstr (env, p - env); |
| 2664 | tab[i].len = strlen (tab[i].name); |
| 2665 | i++; |
| 2666 | } |
| 2667 | if (*p) |
| 2668 | env = p + 1; |
| 2669 | else |
| 2670 | { |
| 2671 | tab[i].name = NULL; /* Mark end of table. */ |
| 2672 | tab[i].len = 0; |
| 2673 | break; |
| 2674 | } |
| 2675 | } |
| 2676 | return tab; |
| 2677 | } |
| 2678 | |
| 2679 | /* Record a tag defined by a TeX command of length LEN and starting at NAME. |
| 2680 | The name being defined actually starts at (NAME + LEN + 1). |
| 2681 | But we seem to include the TeX command in the tag name. */ |
| 2682 | |
| 2683 | void |
| 2684 | TEX_getit (name, len) |
| 2685 | char *name; |
| 2686 | int len; |
| 2687 | { |
| 2688 | char *p = name + len; |
| 2689 | char nambuf[BUFSIZ]; |
| 2690 | |
| 2691 | if (*name == 0) |
| 2692 | return; |
| 2693 | |
| 2694 | /* Let tag name extend to next group close (or end of line) */ |
| 2695 | while (*p && *p != TEX_clgrp) |
| 2696 | p++; |
| 2697 | (void) strncpy (nambuf, name, p - name); |
| 2698 | nambuf[p - name] = 0; |
| 2699 | |
| 2700 | pfnote (nambuf, TRUE, FALSE, lb.buffer, strlen (lb.buffer), lineno, linecharno); |
| 2701 | pfcnt++; |
| 2702 | } |
| 2703 | |
| 2704 | /* If the text at CP matches one of the tag-defining TeX command names, |
| 2705 | return the etags_index of that command in TEX_toktab. |
| 2706 | Otherwise return -1. */ |
| 2707 | |
| 2708 | /* Keep the capital `T' in `Token' for dumb truncating compilers |
| 2709 | (this distinguishes it from `TEX_toktab' */ |
| 2710 | int |
| 2711 | TEX_Token (cp) |
| 2712 | char *cp; |
| 2713 | { |
| 2714 | int i; |
| 2715 | |
| 2716 | for (i = 0; TEX_toktab[i].len > 0; i++) |
| 2717 | if (strncmp (TEX_toktab[i].name, cp, TEX_toktab[i].len) == 0) |
| 2718 | return i; |
| 2719 | return -1; |
| 2720 | } |
| 2721 | \f |
| 2722 | /* Support for Prolog. */ |
| 2723 | |
| 2724 | /* whole head (not only functor, but also arguments) |
| 2725 | is gotten in compound term. */ |
| 2726 | |
| 2727 | void |
| 2728 | prolog_getit (s, lineno, linecharno) |
| 2729 | char *s; |
| 2730 | int lineno; |
| 2731 | long linecharno; |
| 2732 | { |
| 2733 | char nambuf[BUFSIZ], *save_s, tmpc; |
| 2734 | int insquote, npar; |
| 2735 | |
| 2736 | save_s = s; |
| 2737 | insquote = FALSE; |
| 2738 | npar = 0; |
| 2739 | while (1) |
| 2740 | { |
| 2741 | if (*s == '\0') /* syntax error. */ |
| 2742 | return; |
| 2743 | else if (insquote && *s == '\'' && *(s + 1) == '\'') |
| 2744 | s += 2; |
| 2745 | else if (*s == '\'') |
| 2746 | { |
| 2747 | insquote = !insquote; |
| 2748 | s++; |
| 2749 | } |
| 2750 | else if (!insquote && *s == '(') |
| 2751 | { |
| 2752 | npar++; |
| 2753 | s++; |
| 2754 | } |
| 2755 | else if (!insquote && *s == ')') |
| 2756 | { |
| 2757 | npar--; |
| 2758 | s++; |
| 2759 | if (npar == 0) |
| 2760 | break; |
| 2761 | else if (npar < 0) /* syntax error. */ |
| 2762 | return; |
| 2763 | } |
| 2764 | else if (!insquote && *s == '.' && (isspace (*(s + 1)) || *(s + 1) == '\0')) |
| 2765 | { /* fullstop. */ |
| 2766 | if (npar != 0) /* syntax error. */ |
| 2767 | return; |
| 2768 | s++; |
| 2769 | break; |
| 2770 | } |
| 2771 | else |
| 2772 | s++; |
| 2773 | } |
| 2774 | tmpc = *s; |
| 2775 | *s = '\0'; |
| 2776 | strcpy (nambuf, save_s); |
| 2777 | *s = tmpc; |
| 2778 | pfnote (nambuf, TRUE, FALSE, save_s, strlen (nambuf), lineno, linecharno); |
| 2779 | } |
| 2780 | |
| 2781 | /* It is assumed that prolog predicate starts from column 0. */ |
| 2782 | |
| 2783 | void |
| 2784 | prolog_funcs (fi) |
| 2785 | FILE *fi; |
| 2786 | { |
| 2787 | void skip_comment (), prolog_getit (); |
| 2788 | |
| 2789 | lineno = linecharno = charno = 0; |
| 2790 | while (!feof (fi)) |
| 2791 | { |
| 2792 | lineno++; |
| 2793 | linecharno += charno; |
| 2794 | charno = readline (&lb, fi) + 1; /* 1 for newline. */ |
| 2795 | dbp = lb.buffer; |
| 2796 | if (isspace (dbp[0])) /* not predicate header. */ |
| 2797 | continue; |
| 2798 | else if (dbp[0] == '%') /* comment. */ |
| 2799 | continue; |
| 2800 | else if (dbp[0] == '/' && dbp[1] == '*') /* comment. */ |
| 2801 | skip_comment (&lb, fi, &lineno, &linecharno); |
| 2802 | else /* found. */ |
| 2803 | prolog_getit (dbp, lineno, linecharno); |
| 2804 | } |
| 2805 | } |
| 2806 | |
| 2807 | void |
| 2808 | skip_comment (plb, fi, plineno, plinecharno) |
| 2809 | struct linebuffer *plb; |
| 2810 | FILE *fi; |
| 2811 | int *plineno; /* result */ |
| 2812 | long *plinecharno; /* result */ |
| 2813 | { |
| 2814 | while (!substr ("*/", plb->buffer)) |
| 2815 | { |
| 2816 | (*plineno)++; |
| 2817 | *plinecharno += readline (plb, fi) + 1; |
| 2818 | } /* 1 for newline. */ |
| 2819 | } |
| 2820 | |
| 2821 | /* Return TRUE if 'sub' exists somewhere in 's'. */ |
| 2822 | |
| 2823 | int |
| 2824 | substr (sub, s) |
| 2825 | char *sub; |
| 2826 | char *s; |
| 2827 | { |
| 2828 | while (*s && (s = etags_index (s, *sub))) |
| 2829 | if (prestr (sub, s)) |
| 2830 | return (TRUE); |
| 2831 | else |
| 2832 | s++; |
| 2833 | return (FALSE); |
| 2834 | } |
| 2835 | |
| 2836 | /* Return TRUE if 'pre' is prefix of string 's'. */ |
| 2837 | |
| 2838 | int |
| 2839 | prestr (pre, s) |
| 2840 | char *pre; |
| 2841 | char *s; |
| 2842 | { |
| 2843 | if (*pre == '\0') |
| 2844 | return (TRUE); |
| 2845 | else if (*pre == *s) |
| 2846 | return (prestr (pre + 1, s + 1)); |
| 2847 | else |
| 2848 | return (FALSE); |
| 2849 | } |
| 2850 | \f |
| 2851 | /* Initialize a linebuffer for use */ |
| 2852 | |
| 2853 | void |
| 2854 | initbuffer (linebuffer) |
| 2855 | struct linebuffer *linebuffer; |
| 2856 | { |
| 2857 | linebuffer->size = 200; |
| 2858 | linebuffer->buffer = xnew (200, char); |
| 2859 | } |
| 2860 | |
| 2861 | /* |
| 2862 | * Read a line of text from `stream' into `linebuffer'. |
| 2863 | * Return the number of characters read from `stream', |
| 2864 | * which is the length of the line including the newline, if any. |
| 2865 | */ |
| 2866 | long |
| 2867 | readline (linebuffer, stream) |
| 2868 | struct linebuffer *linebuffer; |
| 2869 | register FILE *stream; |
| 2870 | { |
| 2871 | char *buffer = linebuffer->buffer; |
| 2872 | register char *p = linebuffer->buffer; |
| 2873 | register char *pend; |
| 2874 | int newline; /* 1 if ended with newline, 0 if ended with EOF */ |
| 2875 | |
| 2876 | pend = p + linebuffer->size; /* Separate to avoid 386/IX compiler bug. */ |
| 2877 | |
| 2878 | while (1) |
| 2879 | { |
| 2880 | register int c = getc (stream); |
| 2881 | if (p == pend) |
| 2882 | { |
| 2883 | linebuffer->size *= 2; |
| 2884 | buffer = (char *) xrealloc (buffer, linebuffer->size); |
| 2885 | p += buffer - linebuffer->buffer; |
| 2886 | pend = buffer + linebuffer->size; |
| 2887 | linebuffer->buffer = buffer; |
| 2888 | } |
| 2889 | if (c < 0 || c == '\n') |
| 2890 | { |
| 2891 | *p = 0; |
| 2892 | newline = (c == '\n' ? 1 : 0); |
| 2893 | break; |
| 2894 | } |
| 2895 | *p++ = c; |
| 2896 | } |
| 2897 | |
| 2898 | return p - buffer + newline; |
| 2899 | } |
| 2900 | \f |
| 2901 | char * |
| 2902 | savestr (cp) |
| 2903 | char *cp; |
| 2904 | { |
| 2905 | return savenstr (cp, strlen (cp)); |
| 2906 | } |
| 2907 | |
| 2908 | char * |
| 2909 | savenstr (cp, len) |
| 2910 | char *cp; |
| 2911 | int len; |
| 2912 | { |
| 2913 | register char *dp; |
| 2914 | |
| 2915 | dp = xnew (len + 1, char); |
| 2916 | (void) strncpy (dp, cp, len); |
| 2917 | dp[len] = '\0'; |
| 2918 | return dp; |
| 2919 | } |
| 2920 | |
| 2921 | /* |
| 2922 | * Return the ptr in sp at which the character c last |
| 2923 | * appears; NULL if not found |
| 2924 | * |
| 2925 | * Identical to v7 rindex, included for portability. |
| 2926 | */ |
| 2927 | |
| 2928 | char * |
| 2929 | etags_rindex (sp, c) |
| 2930 | register char *sp, c; |
| 2931 | { |
| 2932 | register char *r; |
| 2933 | |
| 2934 | r = NULL; |
| 2935 | do |
| 2936 | { |
| 2937 | if (*sp == c) |
| 2938 | r = sp; |
| 2939 | } while (*sp++); |
| 2940 | return (r); |
| 2941 | } |
| 2942 | |
| 2943 | |
| 2944 | /* |
| 2945 | * Return the ptr in sp at which the character c first |
| 2946 | * appears; NULL if not found |
| 2947 | * |
| 2948 | * Identical to v7 index, included for portability. |
| 2949 | */ |
| 2950 | |
| 2951 | char * |
| 2952 | etags_index (sp, c) |
| 2953 | register char *sp, c; |
| 2954 | { |
| 2955 | do |
| 2956 | { |
| 2957 | if (*sp == c) |
| 2958 | return (sp); |
| 2959 | } while (*sp++); |
| 2960 | return (NULL); |
| 2961 | } |
| 2962 | |
| 2963 | /* Print error message and exit. */ |
| 2964 | |
| 2965 | /* VARARGS1 */ |
| 2966 | void |
| 2967 | fatal (s1, s2) |
| 2968 | char *s1, *s2; |
| 2969 | { |
| 2970 | error (s1, s2); |
| 2971 | exit (1); |
| 2972 | } |
| 2973 | |
| 2974 | /* Print error message. `s1' is printf control string, `s2' is arg for it. */ |
| 2975 | |
| 2976 | /* VARARGS1 */ |
| 2977 | void |
| 2978 | error (s1, s2) |
| 2979 | char *s1, *s2; |
| 2980 | { |
| 2981 | fprintf (stderr, "%s: ", progname); |
| 2982 | fprintf (stderr, s1, s2); |
| 2983 | fprintf (stderr, "\n"); |
| 2984 | } |
| 2985 | |
| 2986 | /* Return a newly-allocated string whose contents concatenate those of s1, s2, s3. */ |
| 2987 | |
| 2988 | char * |
| 2989 | concat (s1, s2, s3) |
| 2990 | char *s1, *s2, *s3; |
| 2991 | { |
| 2992 | int len1 = strlen (s1), len2 = strlen (s2), len3 = strlen (s3); |
| 2993 | char *result = xnew (len1 + len2 + len3 + 1, char); |
| 2994 | |
| 2995 | (void) strcpy (result, s1); |
| 2996 | (void) strcpy (result + len1, s2); |
| 2997 | (void) strcpy (result + len1 + len2, s3); |
| 2998 | *(result + len1 + len2 + len3) = 0; |
| 2999 | |
| 3000 | return result; |
| 3001 | } |
| 3002 | |
| 3003 | /* Like malloc but get fatal error if memory is exhausted. */ |
| 3004 | |
| 3005 | char * |
| 3006 | xmalloc (size) |
| 3007 | int size; |
| 3008 | { |
| 3009 | char *result = malloc (size); |
| 3010 | if (!result) |
| 3011 | fatal ("virtual memory exhausted", 0); |
| 3012 | return result; |
| 3013 | } |
| 3014 | |
| 3015 | char * |
| 3016 | xrealloc (ptr, size) |
| 3017 | char *ptr; |
| 3018 | int size; |
| 3019 | { |
| 3020 | char *result = realloc (ptr, size); |
| 3021 | if (!result) |
| 3022 | fatal ("virtual memory exhausted"); |
| 3023 | return result; |
| 3024 | } |