1 /* Definitions for data structures and routines for the regular 2 expression library. 3 Copyright (C) 1985,1989-93,1995-98,2000,2001,2002,2003,2005,2006,2008,2011 4 Free Software Foundation, Inc. 5 This file is part of the GNU C Library. 6 7 The GNU C Library is free software; you can redistribute it and/or 8 modify it under the terms of the GNU Lesser General Public 9 License as published by the Free Software Foundation; either 10 version 2.1 of the License, or (at your option) any later version. 11 12 The GNU C Library is distributed in the hope that it will be useful, 13 but WITHOUT ANY WARRANTY; without even the implied warranty of 14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU 15 Lesser General Public License for more details. 16 17 You should have received a copy of the GNU Lesser General Public 18 License along with the GNU C Library; if not, write to the Free 19 Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 20 02111-1307 USA. */ 21 22 #ifndef _REGEX_H 23 #define _REGEX_H 1 24 25 #include <sys/types.h> 26 #include <gnu/option-groups.h> 27 28 /* Allow the use in C++ code. */ 29 #ifdef __cplusplus 30 extern "C" { 31 #endif 32 33 /* The following two types have to be signed and unsigned integer type 34 wide enough to hold a value of a pointer. For most ANSI compilers 35 ptrdiff_t and size_t should be likely OK. Still size of these two 36 types is 2 for Microsoft C. Ugh... */ 37 typedef long int s_reg_t; 38 typedef unsigned long int active_reg_t; 39 40 /* The following bits are used to determine the regexp syntax we 41 recognize. The set/not-set meanings are chosen so that Emacs syntax 42 remains the value 0. The bits are given in alphabetical order, and 43 the definitions shifted by one from the previous bit; thus, when we 44 add or remove a bit, only one other definition need change. */ 45 typedef unsigned long int reg_syntax_t; 46 47 #ifdef __USE_GNU 48 /* If this bit is not set, then \ inside a bracket expression is literal. 49 If set, then such a \ quotes the following character. */ 50 # define RE_BACKSLASH_ESCAPE_IN_LISTS ((unsigned long int) 1) 51 52 /* If this bit is not set, then + and ? are operators, and \+ and \? are 53 literals. 54 If set, then \+ and \? are operators and + and ? are literals. */ 55 # define RE_BK_PLUS_QM (RE_BACKSLASH_ESCAPE_IN_LISTS << 1) 56 57 /* If this bit is set, then character classes are supported. They are: 58 [:alpha:], [:upper:], [:lower:], [:digit:], [:alnum:], [:xdigit:], 59 [:space:], [:print:], [:punct:], [:graph:], and [:cntrl:]. 60 If not set, then character classes are not supported. */ 61 # define RE_CHAR_CLASSES (RE_BK_PLUS_QM << 1) 62 63 /* If this bit is set, then ^ and $ are always anchors (outside bracket 64 expressions, of course). 65 If this bit is not set, then it depends: 66 ^ is an anchor if it is at the beginning of a regular 67 expression or after an open-group or an alternation operator; 68 $ is an anchor if it is at the end of a regular expression, or 69 before a close-group or an alternation operator. 70 71 This bit could be (re)combined with RE_CONTEXT_INDEP_OPS, because 72 POSIX draft 11.2 says that * etc. in leading positions is undefined. 73 We already implemented a previous draft which made those constructs 74 invalid, though, so we haven't changed the code back. */ 75 # define RE_CONTEXT_INDEP_ANCHORS (RE_CHAR_CLASSES << 1) 76 77 /* If this bit is set, then special characters are always special 78 regardless of where they are in the pattern. 79 If this bit is not set, then special characters are special only in 80 some contexts; otherwise they are ordinary. Specifically, 81 * + ? and intervals are only special when not after the beginning, 82 open-group, or alternation operator. */ 83 # define RE_CONTEXT_INDEP_OPS (RE_CONTEXT_INDEP_ANCHORS << 1) 84 85 /* If this bit is set, then *, +, ?, and { cannot be first in an re or 86 immediately after an alternation or begin-group operator. */ 87 # define RE_CONTEXT_INVALID_OPS (RE_CONTEXT_INDEP_OPS << 1) 88 89 /* If this bit is set, then . matches newline. 90 If not set, then it doesn't. */ 91 # define RE_DOT_NEWLINE (RE_CONTEXT_INVALID_OPS << 1) 92 93 /* If this bit is set, then . doesn't match NUL. 94 If not set, then it does. */ 95 # define RE_DOT_NOT_NULL (RE_DOT_NEWLINE << 1) 96 97 /* If this bit is set, nonmatching lists [^...] do not match newline. 98 If not set, they do. */ 99 # define RE_HAT_LISTS_NOT_NEWLINE (RE_DOT_NOT_NULL << 1) 100 101 /* If this bit is set, either \{...\} or {...} defines an 102 interval, depending on RE_NO_BK_BRACES. 103 If not set, \{, \}, {, and } are literals. */ 104 # define RE_INTERVALS (RE_HAT_LISTS_NOT_NEWLINE << 1) 105 106 /* If this bit is set, +, ? and | aren't recognized as operators. 107 If not set, they are. */ 108 # define RE_LIMITED_OPS (RE_INTERVALS << 1) 109 110 /* If this bit is set, newline is an alternation operator. 111 If not set, newline is literal. */ 112 # define RE_NEWLINE_ALT (RE_LIMITED_OPS << 1) 113 114 /* If this bit is set, then `{...}' defines an interval, and \{ and \} 115 are literals. 116 If not set, then `\{...\}' defines an interval. */ 117 # define RE_NO_BK_BRACES (RE_NEWLINE_ALT << 1) 118 119 /* If this bit is set, (...) defines a group, and \( and \) are literals. 120 If not set, \(...\) defines a group, and ( and ) are literals. */ 121 # define RE_NO_BK_PARENS (RE_NO_BK_BRACES << 1) 122 123 /* If this bit is set, then \<digit> matches <digit>. 124 If not set, then \<digit> is a back-reference. */ 125 # define RE_NO_BK_REFS (RE_NO_BK_PARENS << 1) 126 127 /* If this bit is set, then | is an alternation operator, and \| is literal. 128 If not set, then \| is an alternation operator, and | is literal. */ 129 # define RE_NO_BK_VBAR (RE_NO_BK_REFS << 1) 130 131 /* If this bit is set, then an ending range point collating higher 132 than the starting range point, as in [z-a], is invalid. 133 If not set, then when ending range point collates higher than the 134 starting range point, the range is ignored. */ 135 # define RE_NO_EMPTY_RANGES (RE_NO_BK_VBAR << 1) 136 137 /* If this bit is set, then an unmatched ) is ordinary. 138 If not set, then an unmatched ) is invalid. */ 139 # define RE_UNMATCHED_RIGHT_PAREN_ORD (RE_NO_EMPTY_RANGES << 1) 140 141 /* If this bit is set, succeed as soon as we match the whole pattern, 142 without further backtracking. */ 143 # define RE_NO_POSIX_BACKTRACKING (RE_UNMATCHED_RIGHT_PAREN_ORD << 1) 144 145 /* If this bit is set, do not process the GNU regex operators. 146 If not set, then the GNU regex operators are recognized. */ 147 # define RE_NO_GNU_OPS (RE_NO_POSIX_BACKTRACKING << 1) 148 149 /* If this bit is set, turn on internal regex debugging. 150 If not set, and debugging was on, turn it off. 151 This only works if regex.c is compiled -DDEBUG. 152 We define this bit always, so that all that's needed to turn on 153 debugging is to recompile regex.c; the calling code can always have 154 this bit set, and it won't affect anything in the normal case. */ 155 # define RE_DEBUG (RE_NO_GNU_OPS << 1) 156 157 /* If this bit is set, a syntactically invalid interval is treated as 158 a string of ordinary characters. For example, the ERE 'a{1' is 159 treated as 'a\{1'. */ 160 # define RE_INVALID_INTERVAL_ORD (RE_DEBUG << 1) 161 162 /* EGLIBC: Old regex implementation does not support these. */ 163 # ifdef __OPTION_POSIX_REGEXP_GLIBC 164 /* If this bit is set, then ignore case when matching. 165 If not set, then case is significant. */ 166 # define RE_ICASE (RE_INVALID_INTERVAL_ORD << 1) 167 168 /* This bit is used internally like RE_CONTEXT_INDEP_ANCHORS but only 169 for ^, because it is difficult to scan the regex backwards to find 170 whether ^ should be special. */ 171 # define RE_CARET_ANCHORS_HERE (RE_ICASE << 1) 172 173 /* If this bit is set, then \{ cannot be first in an bre or 174 immediately after an alternation or begin-group operator. */ 175 # define RE_CONTEXT_INVALID_DUP (RE_CARET_ANCHORS_HERE << 1) 176 177 /* If this bit is set, then no_sub will be set to 1 during 178 re_compile_pattern. */ 179 # define RE_NO_SUB (RE_CONTEXT_INVALID_DUP << 1) 180 # endif /* __OPTION_POSIX_REGEXP_GLIBC */ 181 #endif 182 183 /* This global variable defines the particular regexp syntax to use (for 184 some interfaces). When a regexp is compiled, the syntax used is 185 stored in the pattern buffer, so changing this does not affect 186 already-compiled regexps. */ 187 extern reg_syntax_t re_syntax_options; 188 189 #ifdef __USE_GNU 191 /* Define combinations of the above bits for the standard possibilities. 192 (The [[[ comments delimit what gets put into the Texinfo file, so 193 don't delete them!) */ 194 /* [[[begin syntaxes]]] */ 195 #define RE_SYNTAX_EMACS 0 196 197 #define RE_SYNTAX_AWK \ 198 (RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DOT_NOT_NULL \ 199 | RE_NO_BK_PARENS | RE_NO_BK_REFS \ 200 | RE_NO_BK_VBAR | RE_NO_EMPTY_RANGES \ 201 | RE_DOT_NEWLINE | RE_CONTEXT_INDEP_ANCHORS \ 202 | RE_CHAR_CLASSES \ 203 | RE_UNMATCHED_RIGHT_PAREN_ORD | RE_NO_GNU_OPS) 204 205 #define RE_SYNTAX_GNU_AWK \ 206 ((RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS \ 207 | RE_INVALID_INTERVAL_ORD) \ 208 & ~(RE_DOT_NOT_NULL | RE_CONTEXT_INDEP_OPS \ 209 | RE_CONTEXT_INVALID_OPS )) 210 211 #define RE_SYNTAX_POSIX_AWK \ 212 (RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS \ 213 | RE_INTERVALS | RE_NO_GNU_OPS \ 214 | RE_INVALID_INTERVAL_ORD) 215 216 #define RE_SYNTAX_GREP \ 217 (RE_BK_PLUS_QM | RE_CHAR_CLASSES \ 218 | RE_HAT_LISTS_NOT_NEWLINE | RE_INTERVALS \ 219 | RE_NEWLINE_ALT) 220 221 #define RE_SYNTAX_EGREP \ 222 (RE_CHAR_CLASSES | RE_CONTEXT_INDEP_ANCHORS \ 223 | RE_CONTEXT_INDEP_OPS | RE_HAT_LISTS_NOT_NEWLINE \ 224 | RE_NEWLINE_ALT | RE_NO_BK_PARENS \ 225 | RE_NO_BK_VBAR) 226 227 #define RE_SYNTAX_POSIX_EGREP \ 228 (RE_SYNTAX_EGREP | RE_INTERVALS | RE_NO_BK_BRACES \ 229 | RE_INVALID_INTERVAL_ORD) 230 231 /* P1003.2/D11.2, section 4.20.7.1, lines 5078ff. */ 232 #define RE_SYNTAX_ED RE_SYNTAX_POSIX_BASIC 233 234 #define RE_SYNTAX_SED RE_SYNTAX_POSIX_BASIC 235 236 /* Syntax bits common to both basic and extended POSIX regex syntax. */ 237 #define _RE_SYNTAX_POSIX_COMMON \ 238 (RE_CHAR_CLASSES | RE_DOT_NEWLINE | RE_DOT_NOT_NULL \ 239 | RE_INTERVALS | RE_NO_EMPTY_RANGES) 240 241 #ifdef __OPTION_POSIX_REGEXP_GLIBC 242 #define RE_SYNTAX_POSIX_BASIC \ 243 (_RE_SYNTAX_POSIX_COMMON | RE_BK_PLUS_QM | RE_CONTEXT_INVALID_DUP) 244 #else 245 #define RE_SYNTAX_POSIX_BASIC \ 246 (_RE_SYNTAX_POSIX_COMMON | RE_BK_PLUS_QM) 247 #endif 248 249 /* Differs from ..._POSIX_BASIC only in that RE_BK_PLUS_QM becomes 250 RE_LIMITED_OPS, i.e., \? \+ \| are not recognized. Actually, this 251 isn't minimal, since other operators, such as \`, aren't disabled. */ 252 #define RE_SYNTAX_POSIX_MINIMAL_BASIC \ 253 (_RE_SYNTAX_POSIX_COMMON | RE_LIMITED_OPS) 254 255 #define RE_SYNTAX_POSIX_EXTENDED \ 256 (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \ 257 | RE_CONTEXT_INDEP_OPS | RE_NO_BK_BRACES \ 258 | RE_NO_BK_PARENS | RE_NO_BK_VBAR \ 259 | RE_CONTEXT_INVALID_OPS | RE_UNMATCHED_RIGHT_PAREN_ORD) 260 261 /* Differs from ..._POSIX_EXTENDED in that RE_CONTEXT_INDEP_OPS is 262 removed and RE_NO_BK_REFS is added. */ 263 #define RE_SYNTAX_POSIX_MINIMAL_EXTENDED \ 264 (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \ 265 | RE_CONTEXT_INVALID_OPS | RE_NO_BK_BRACES \ 266 | RE_NO_BK_PARENS | RE_NO_BK_REFS \ 267 | RE_NO_BK_VBAR | RE_UNMATCHED_RIGHT_PAREN_ORD) 268 /* [[[end syntaxes]]] */ 269 270 /* Maximum number of duplicates an interval can allow. Some systems 272 (erroneously) define this in other header files, but we want our 273 value, so remove any previous define. */ 274 # ifdef RE_DUP_MAX 275 # undef RE_DUP_MAX 276 # endif 277 /* If sizeof(int) == 2, then ((1 << 15) - 1) overflows. */ 278 # define RE_DUP_MAX (0x7fff) 279 #endif 280 281 282 /* POSIX `cflags' bits (i.e., information for `regcomp'). */ 283 284 /* If this bit is set, then use extended regular expression syntax. 285 If not set, then use basic regular expression syntax. */ 286 #define REG_EXTENDED 1 287 288 /* If this bit is set, then ignore case when matching. 289 If not set, then case is significant. */ 290 #define REG_ICASE (REG_EXTENDED << 1) 291 292 /* If this bit is set, then anchors do not match at newline 293 characters in the string. 294 If not set, then anchors do match at newlines. */ 295 #define REG_NEWLINE (REG_ICASE << 1) 296 297 /* If this bit is set, then report only success or fail in regexec. 298 If not set, then returns differ between not matching and errors. */ 299 #define REG_NOSUB (REG_NEWLINE << 1) 300 301 302 /* POSIX `eflags' bits (i.e., information for regexec). */ 303 304 /* If this bit is set, then the beginning-of-line operator doesn't match 305 the beginning of the string (presumably because it's not the 306 beginning of a line). 307 If not set, then the beginning-of-line operator does match the 308 beginning of the string. */ 309 #define REG_NOTBOL 1 310 311 /* Like REG_NOTBOL, except for the end-of-line. */ 312 #define REG_NOTEOL (1 << 1) 313 314 #ifdef __OPTION_POSIX_REGEXP_GLIBC 315 /* Use PMATCH[0] to delimit the start and end of the search in the 316 buffer. */ 317 #define REG_STARTEND (1 << 2) 318 #endif 319 320 321 /* If any error codes are removed, changed, or added, update the 322 `re_error_msg' table in regex.c. */ 323 typedef enum 324 { 325 #if defined _XOPEN_SOURCE || defined __USE_XOPEN2K 326 REG_ENOSYS = -1, /* This will never happen for this implementation. */ 327 #endif 328 329 REG_NOERROR = 0, /* Success. */ 330 REG_NOMATCH, /* Didn't find a match (for regexec). */ 331 332 /* POSIX regcomp return error codes. (In the order listed in the 333 standard.) */ 334 REG_BADPAT, /* Invalid pattern. */ 335 REG_ECOLLATE, /* Inalid collating element. */ 336 REG_ECTYPE, /* Invalid character class name. */ 337 REG_EESCAPE, /* Trailing backslash. */ 338 REG_ESUBREG, /* Invalid back reference. */ 339 REG_EBRACK, /* Unmatched left bracket. */ 340 REG_EPAREN, /* Parenthesis imbalance. */ 341 REG_EBRACE, /* Unmatched \{. */ 342 REG_BADBR, /* Invalid contents of \{\}. */ 343 REG_ERANGE, /* Invalid range end. */ 344 REG_ESPACE, /* Ran out of memory. */ 345 REG_BADRPT, /* No preceding re for repetition op. */ 346 347 /* Error codes we've added. */ 348 REG_EEND, /* Premature end. */ 349 REG_ESIZE, /* Compiled pattern bigger than 2^16 bytes. */ 350 REG_ERPAREN /* Unmatched ) or \); not returned from regcomp. */ 351 } reg_errcode_t; 352 353 /* This data structure represents a compiled pattern. Before calling 355 the pattern compiler, the fields `buffer', `allocated', `fastmap', 356 and `translate' can be set. After the pattern has been compiled, 357 the fields `re_nsub', `not_bol' and `not_eol' are available. All 358 other fields are private to the regex routines. */ 359 360 #ifndef RE_TRANSLATE_TYPE 361 # define __RE_TRANSLATE_TYPE unsigned char * 362 # ifdef __USE_GNU 363 # define RE_TRANSLATE_TYPE __RE_TRANSLATE_TYPE 364 # endif 365 #endif 366 367 #ifdef __USE_GNU 368 # define __REPB_PREFIX(name) name 369 #else 370 # define __REPB_PREFIX(name) __##name 371 #endif 372 373 struct re_pattern_buffer 374 { 375 /* Space that holds the compiled pattern. It is declared as 376 `unsigned char *' because its elements are sometimes used as 377 array indexes. */ 378 unsigned char *__REPB_PREFIX(buffer); 379 380 /* Number of bytes to which `buffer' points. */ 381 unsigned long int __REPB_PREFIX(allocated); 382 383 /* Number of bytes actually used in `buffer'. */ 384 unsigned long int __REPB_PREFIX(used); 385 386 /* Syntax setting with which the pattern was compiled. */ 387 reg_syntax_t __REPB_PREFIX(syntax); 388 389 /* Pointer to a fastmap, if any, otherwise zero. re_search uses the 390 fastmap, if there is one, to skip over impossible starting points 391 for matches. */ 392 char *__REPB_PREFIX(fastmap); 393 394 /* Either a translate table to apply to all characters before 395 comparing them, or zero for no translation. The translation is 396 applied to a pattern when it is compiled and to a string when it 397 is matched. */ 398 __RE_TRANSLATE_TYPE __REPB_PREFIX(translate); 399 400 /* Number of subexpressions found by the compiler. */ 401 size_t re_nsub; 402 403 /* Zero if this pattern cannot match the empty string, one else. 404 Well, in truth it's used only in `re_search_2', to see whether or 405 not we should use the fastmap, so we don't set this absolutely 406 perfectly; see `re_compile_fastmap' (the `duplicate' case). */ 407 unsigned __REPB_PREFIX(can_be_null) : 1; 408 409 /* If REGS_UNALLOCATED, allocate space in the `regs' structure 410 for `max (RE_NREGS, re_nsub + 1)' groups. 411 If REGS_REALLOCATE, reallocate space if necessary. 412 If REGS_FIXED, use what's there. */ 413 #ifdef __USE_GNU 414 # define REGS_UNALLOCATED 0 415 # define REGS_REALLOCATE 1 416 # define REGS_FIXED 2 417 #endif 418 unsigned __REPB_PREFIX(regs_allocated) : 2; 419 420 /* Set to zero when `regex_compile' compiles a pattern; set to one 421 by `re_compile_fastmap' if it updates the fastmap. */ 422 unsigned __REPB_PREFIX(fastmap_accurate) : 1; 423 424 /* If set, `re_match_2' does not return information about 425 subexpressions. */ 426 unsigned __REPB_PREFIX(no_sub) : 1; 427 428 /* If set, a beginning-of-line anchor doesn't match at the beginning 429 of the string. */ 430 unsigned __REPB_PREFIX(not_bol) : 1; 431 432 /* Similarly for an end-of-line anchor. */ 433 unsigned __REPB_PREFIX(not_eol) : 1; 434 435 /* If true, an anchor at a newline matches. */ 436 unsigned __REPB_PREFIX(newline_anchor) : 1; 437 }; 438 439 typedef struct re_pattern_buffer regex_t; 440 441 /* Type for byte offsets within the string. POSIX mandates this. */ 443 typedef int regoff_t; 444 445 446 #ifdef __USE_GNU 447 /* This is the structure we store register match data in. See 448 regex.texinfo for a full description of what registers match. */ 449 struct re_registers 450 { 451 unsigned num_regs; 452 regoff_t *start; 453 regoff_t *end; 454 }; 455 456 457 /* If `regs_allocated' is REGS_UNALLOCATED in the pattern buffer, 458 `re_match_2' returns information about at least this many registers 459 the first time a `regs' structure is passed. */ 460 # ifndef RE_NREGS 461 # define RE_NREGS 30 462 # endif 463 #endif 464 465 466 /* POSIX specification for registers. Aside from the different names than 467 `re_registers', POSIX uses an array of structures, instead of a 468 structure of arrays. */ 469 typedef struct 470 { 471 regoff_t rm_so; /* Byte offset from string's start to substring's start. */ 472 regoff_t rm_eo; /* Byte offset from string's start to substring's end. */ 473 } regmatch_t; 474 475 /* Declarations for routines. */ 477 478 #ifdef __USE_GNU 479 /* Sets the current default syntax to SYNTAX, and return the old syntax. 480 You can also simply assign to the `re_syntax_options' variable. */ 481 extern reg_syntax_t re_set_syntax (reg_syntax_t __syntax); 482 483 /* Compile the regular expression PATTERN, with length LENGTH 484 and syntax given by the global `re_syntax_options', into the buffer 485 BUFFER. Return NULL if successful, and an error string if not. 486 487 To free the allocated storage, you must call `regfree' on BUFFER. 488 Note that the translate table must either have been initialised by 489 `regcomp', with a malloc'ed value, or set to NULL before calling 490 `regfree'. */ 491 extern const char *re_compile_pattern (const char *__pattern, size_t __length, 492 struct re_pattern_buffer *__buffer); 493 494 495 /* Compile a fastmap for the compiled pattern in BUFFER; used to 496 accelerate searches. Return 0 if successful and -2 if was an 497 internal error. */ 498 extern int re_compile_fastmap (struct re_pattern_buffer *__buffer); 499 500 501 /* Search in the string STRING (with length LENGTH) for the pattern 502 compiled into BUFFER. Start searching at position START, for RANGE 503 characters. Return the starting position of the match, -1 for no 504 match, or -2 for an internal error. Also return register 505 information in REGS (if REGS and BUFFER->no_sub are nonzero). */ 506 extern int re_search (struct re_pattern_buffer *__buffer, const char *__string, 507 int __length, int __start, int __range, 508 struct re_registers *__regs); 509 510 511 /* Like `re_search', but search in the concatenation of STRING1 and 512 STRING2. Also, stop searching at index START + STOP. */ 513 extern int re_search_2 (struct re_pattern_buffer *__buffer, 514 const char *__string1, int __length1, 515 const char *__string2, int __length2, int __start, 516 int __range, struct re_registers *__regs, int __stop); 517 518 519 /* Like `re_search', but return how many characters in STRING the regexp 520 in BUFFER matched, starting at position START. */ 521 extern int re_match (struct re_pattern_buffer *__buffer, const char *__string, 522 int __length, int __start, struct re_registers *__regs); 523 524 525 /* Relates to `re_match' as `re_search_2' relates to `re_search'. */ 526 extern int re_match_2 (struct re_pattern_buffer *__buffer, 527 const char *__string1, int __length1, 528 const char *__string2, int __length2, int __start, 529 struct re_registers *__regs, int __stop); 530 531 532 /* Set REGS to hold NUM_REGS registers, storing them in STARTS and 533 ENDS. Subsequent matches using BUFFER and REGS will use this memory 534 for recording register information. STARTS and ENDS must be 535 allocated with malloc, and must each be at least `NUM_REGS * sizeof 536 (regoff_t)' bytes long. 537 538 If NUM_REGS == 0, then subsequent matches should allocate their own 539 register data. 540 541 Unless this function is called, the first search or match using 542 PATTERN_BUFFER will allocate its own register data, without 543 freeing the old data. */ 544 extern void re_set_registers (struct re_pattern_buffer *__buffer, 545 struct re_registers *__regs, 546 unsigned int __num_regs, 547 regoff_t *__starts, regoff_t *__ends); 548 #endif /* Use GNU */ 549 550 #if defined _REGEX_RE_COMP || (defined _LIBC && defined __USE_BSD) 551 # ifndef _CRAY 552 /* 4.2 bsd compatibility. */ 553 extern char *re_comp (const char *); 554 extern int re_exec (const char *); 555 # endif 556 #endif 557 558 /* GCC 2.95 and later have "__restrict"; C99 compilers have 559 "restrict", and "configure" may have defined "restrict". */ 560 #ifndef __restrict 561 # if ! (2 < __GNUC__ || (2 == __GNUC__ && 95 <= __GNUC_MINOR__)) 562 # if defined restrict || 199901L <= __STDC_VERSION__ 563 # define __restrict restrict 564 # else 565 # define __restrict 566 # endif 567 # endif 568 #endif 569 /* gcc 3.1 and up support the [restrict] syntax. */ 570 #ifndef __restrict_arr 571 # if (__GNUC__ > 3 || (__GNUC__ == 3 && __GNUC_MINOR__ >= 1)) \ 572 && !defined __GNUG__ 573 # define __restrict_arr __restrict 574 # else 575 # define __restrict_arr 576 # endif 577 #endif 578 579 /* POSIX compatibility. */ 580 extern int regcomp (regex_t *__restrict __preg, 581 const char *__restrict __pattern, 582 int __cflags); 583 584 extern int regexec (const regex_t *__restrict __preg, 585 const char *__restrict __string, size_t __nmatch, 586 regmatch_t __pmatch[__restrict_arr], 587 int __eflags); 588 589 extern size_t regerror (int __errcode, const regex_t *__restrict __preg, 590 char *__restrict __errbuf, size_t __errbuf_size); 591 592 extern void regfree (regex_t *__preg); 593 594 595 #ifdef __cplusplus 596 } 597 #endif /* C++ */ 598 599 #endif /* regex.h */ 600