1 /* -*- buffer-read-only: t -*- vi: set ro: */ 2 /* DO NOT EDIT! GENERATED AUTOMATICALLY! */ 3 /* Definitions for data structures and routines for the regular 4 expression library. 5 Copyright (C) 1985, 1989, 1990, 1991, 1992, 1993, 1995, 1996, 1997, 1998, 6 2000, 2001, 2002, 2003, 2005, 2006, 2009, 2010 Free Software Foundation, 7 Inc. 8 This file is part of the GNU C Library. 9 10 This program is free software; you can redistribute it and/or modify 11 it under the terms of the GNU General Public License as published by 12 the Free Software Foundation; either version 3, or (at your option) 13 any later version. 14 15 This program is distributed in the hope that it will be useful, 16 but WITHOUT ANY WARRANTY; without even the implied warranty of 17 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the 18 GNU General Public License for more details. 19 20 You should have received a copy of the GNU General Public License along 21 with this program; if not, write to the Free Software Foundation, 22 Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. */ 23 24 #ifndef _REGEX_H 25 #define _REGEX_H 1 26 27 #include <sys/types.h> 28 29 /* Allow the use in C++ code. */ 30 #ifdef __cplusplus 31 extern "C" { 32 #endif 33 34 /* Define __USE_GNU_REGEX to declare GNU extensions that violate the 35 POSIX name space rules. */ 36 #undef __USE_GNU_REGEX 37 #if (defined _GNU_SOURCE \ 38 || (!defined _POSIX_C_SOURCE && !defined _POSIX_SOURCE \ 39 && !defined _XOPEN_SOURCE)) 40 # define __USE_GNU_REGEX 1 41 #endif 42 43 #ifdef _REGEX_LARGE_OFFSETS 44 45 /* Use types and values that are wide enough to represent signed and 46 unsigned byte offsets in memory. This currently works only when 47 the regex code is used outside of the GNU C library; it is not yet 48 supported within glibc itself, and glibc users should not define 49 _REGEX_LARGE_OFFSETS. */ 50 51 /* The type of the offset of a byte within a string. 52 For historical reasons POSIX 1003.1-2004 requires that regoff_t be 53 at least as wide as off_t. However, many common POSIX platforms set 54 regoff_t to the more-sensible ssize_t and the Open Group has 55 signalled its intention to change the requirement to be that 56 regoff_t be at least as wide as ptrdiff_t and ssize_t; see XBD ERN 57 60 (2005-08-25). We don't know of any hosts where ssize_t or 58 ptrdiff_t is wider than ssize_t, so ssize_t is safe. */ 59 typedef ssize_t regoff_t; 60 61 /* The type of nonnegative object indexes. Traditionally, GNU regex 62 uses 'int' for these. Code that uses __re_idx_t should work 63 regardless of whether the type is signed. */ 64 typedef size_t __re_idx_t; 65 66 /* The type of object sizes. */ 67 typedef size_t __re_size_t; 68 69 /* The type of object sizes, in places where the traditional code 70 uses unsigned long int. */ 71 typedef size_t __re_long_size_t; 72 73 #else 74 75 /* Use types that are binary-compatible with the traditional GNU regex 76 implementation, which mishandles strings longer than INT_MAX. */ 77 78 typedef int regoff_t; 79 typedef int __re_idx_t; 80 typedef unsigned int __re_size_t; 81 typedef unsigned long int __re_long_size_t; 82 83 #endif 84 85 /* The following two types have to be signed and unsigned integer type 86 wide enough to hold a value of a pointer. For most ANSI compilers 87 ptrdiff_t and size_t should be likely OK. Still size of these two 88 types is 2 for Microsoft C. Ugh... */ 89 typedef long int s_reg_t; 90 typedef unsigned long int active_reg_t; 91 92 /* The following bits are used to determine the regexp syntax we 93 recognize. The set/not-set meanings are chosen so that Emacs syntax 94 remains the value 0. The bits are given in alphabetical order, and 95 the definitions shifted by one from the previous bit; thus, when we 96 add or remove a bit, only one other definition need change. */ 97 typedef unsigned long int reg_syntax_t; 98 99 #ifdef __USE_GNU_REGEX 100 101 /* If this bit is not set, then \ inside a bracket expression is literal. 102 If set, then such a \ quotes the following character. */ 103 # define RE_BACKSLASH_ESCAPE_IN_LISTS ((unsigned long int) 1) 104 105 /* If this bit is not set, then + and ? are operators, and \+ and \? are 106 literals. 107 If set, then \+ and \? are operators and + and ? are literals. */ 108 # define RE_BK_PLUS_QM (RE_BACKSLASH_ESCAPE_IN_LISTS << 1) 109 110 /* If this bit is set, then character classes are supported. They are: 111 [:alpha:], [:upper:], [:lower:], [:digit:], [:alnum:], [:xdigit:], 112 [:space:], [:print:], [:punct:], [:graph:], and [:cntrl:]. 113 If not set, then character classes are not supported. */ 114 # define RE_CHAR_CLASSES (RE_BK_PLUS_QM << 1) 115 116 /* If this bit is set, then ^ and $ are always anchors (outside bracket 117 expressions, of course). 118 If this bit is not set, then it depends: 119 ^ is an anchor if it is at the beginning of a regular 120 expression or after an open-group or an alternation operator; 121 $ is an anchor if it is at the end of a regular expression, or 122 before a close-group or an alternation operator. 123 124 This bit could be (re)combined with RE_CONTEXT_INDEP_OPS, because 125 POSIX draft 11.2 says that * etc. in leading positions is undefined. 126 We already implemented a previous draft which made those constructs 127 invalid, though, so we haven't changed the code back. */ 128 # define RE_CONTEXT_INDEP_ANCHORS (RE_CHAR_CLASSES << 1) 129 130 /* If this bit is set, then special characters are always special 131 regardless of where they are in the pattern. 132 If this bit is not set, then special characters are special only in 133 some contexts; otherwise they are ordinary. Specifically, 134 * + ? and intervals are only special when not after the beginning, 135 open-group, or alternation operator. */ 136 # define RE_CONTEXT_INDEP_OPS (RE_CONTEXT_INDEP_ANCHORS << 1) 137 138 /* If this bit is set, then *, +, ?, and { cannot be first in an re or 139 immediately after an alternation or begin-group operator. */ 140 # define RE_CONTEXT_INVALID_OPS (RE_CONTEXT_INDEP_OPS << 1) 141 142 /* If this bit is set, then . matches newline. 143 If not set, then it doesn't. */ 144 # define RE_DOT_NEWLINE (RE_CONTEXT_INVALID_OPS << 1) 145 146 /* If this bit is set, then . doesn't match NUL. 147 If not set, then it does. */ 148 # define RE_DOT_NOT_NULL (RE_DOT_NEWLINE << 1) 149 150 /* If this bit is set, nonmatching lists [^...] do not match newline. 151 If not set, they do. */ 152 # define RE_HAT_LISTS_NOT_NEWLINE (RE_DOT_NOT_NULL << 1) 153 154 /* If this bit is set, either \{...\} or {...} defines an 155 interval, depending on RE_NO_BK_BRACES. 156 If not set, \{, \}, {, and } are literals. */ 157 # define RE_INTERVALS (RE_HAT_LISTS_NOT_NEWLINE << 1) 158 159 /* If this bit is set, +, ? and | aren't recognized as operators. 160 If not set, they are. */ 161 # define RE_LIMITED_OPS (RE_INTERVALS << 1) 162 163 /* If this bit is set, newline is an alternation operator. 164 If not set, newline is literal. */ 165 # define RE_NEWLINE_ALT (RE_LIMITED_OPS << 1) 166 167 /* If this bit is set, then `{...}' defines an interval, and \{ and \} 168 are literals. 169 If not set, then `\{...\}' defines an interval. */ 170 # define RE_NO_BK_BRACES (RE_NEWLINE_ALT << 1) 171 172 /* If this bit is set, (...) defines a group, and \( and \) are literals. 173 If not set, \(...\) defines a group, and ( and ) are literals. */ 174 # define RE_NO_BK_PARENS (RE_NO_BK_BRACES << 1) 175 176 /* If this bit is set, then \<digit> matches <digit>. 177 If not set, then \<digit> is a back-reference. */ 178 # define RE_NO_BK_REFS (RE_NO_BK_PARENS << 1) 179 180 /* If this bit is set, then | is an alternation operator, and \| is literal. 181 If not set, then \| is an alternation operator, and | is literal. */ 182 # define RE_NO_BK_VBAR (RE_NO_BK_REFS << 1) 183 184 /* If this bit is set, then an ending range point collating higher 185 than the starting range point, as in [z-a], is invalid. 186 If not set, then when ending range point collates higher than the 187 starting range point, the range is ignored. */ 188 # define RE_NO_EMPTY_RANGES (RE_NO_BK_VBAR << 1) 189 190 /* If this bit is set, then an unmatched ) is ordinary. 191 If not set, then an unmatched ) is invalid. */ 192 # define RE_UNMATCHED_RIGHT_PAREN_ORD (RE_NO_EMPTY_RANGES << 1) 193 194 /* If this bit is set, succeed as soon as we match the whole pattern, 195 without further backtracking. */ 196 # define RE_NO_POSIX_BACKTRACKING (RE_UNMATCHED_RIGHT_PAREN_ORD << 1) 197 198 /* If this bit is set, do not process the GNU regex operators. 199 If not set, then the GNU regex operators are recognized. */ 200 # define RE_NO_GNU_OPS (RE_NO_POSIX_BACKTRACKING << 1) 201 202 /* If this bit is set, turn on internal regex debugging. 203 If not set, and debugging was on, turn it off. 204 This only works if regex.c is compiled -DDEBUG. 205 We define this bit always, so that all that's needed to turn on 206 debugging is to recompile regex.c; the calling code can always have 207 this bit set, and it won't affect anything in the normal case. */ 208 # define RE_DEBUG (RE_NO_GNU_OPS << 1) 209 210 /* If this bit is set, a syntactically invalid interval is treated as 211 a string of ordinary characters. For example, the ERE 'a{1' is 212 treated as 'a\{1'. */ 213 # define RE_INVALID_INTERVAL_ORD (RE_DEBUG << 1) 214 215 /* If this bit is set, then ignore case when matching. 216 If not set, then case is significant. */ 217 # define RE_ICASE (RE_INVALID_INTERVAL_ORD << 1) 218 219 /* This bit is used internally like RE_CONTEXT_INDEP_ANCHORS but only 220 for ^, because it is difficult to scan the regex backwards to find 221 whether ^ should be special. */ 222 # define RE_CARET_ANCHORS_HERE (RE_ICASE << 1) 223 224 /* If this bit is set, then \{ cannot be first in a regex or 225 immediately after an alternation, open-group or \} operator. */ 226 # define RE_CONTEXT_INVALID_DUP (RE_CARET_ANCHORS_HERE << 1) 227 228 /* If this bit is set, then no_sub will be set to 1 during 229 re_compile_pattern. */ 230 # define RE_NO_SUB (RE_CONTEXT_INVALID_DUP << 1) 231 232 #endif /* defined __USE_GNU_REGEX */ 233 234 /* This global variable defines the particular regexp syntax to use (for 235 some interfaces). When a regexp is compiled, the syntax used is 236 stored in the pattern buffer, so changing this does not affect 237 already-compiled regexps. */ 238 extern reg_syntax_t re_syntax_options; 239 240 #ifdef __USE_GNU_REGEX 241 /* Define combinations of the above bits for the standard possibilities. 242 (The [[[ comments delimit what gets put into the Texinfo file, so 243 don't delete them!) */ 244 /* [[[begin syntaxes]]] */ 245 # define RE_SYNTAX_EMACS 0 246 247 # define RE_SYNTAX_AWK \ 248 (RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DOT_NOT_NULL \ 249 | RE_NO_BK_PARENS | RE_NO_BK_REFS \ 250 | RE_NO_BK_VBAR | RE_NO_EMPTY_RANGES \ 251 | RE_DOT_NEWLINE | RE_CONTEXT_INDEP_ANCHORS \ 252 | RE_UNMATCHED_RIGHT_PAREN_ORD | RE_NO_GNU_OPS) 253 254 # define RE_SYNTAX_GNU_AWK \ 255 ((RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DEBUG) \ 256 & ~(RE_DOT_NOT_NULL | RE_INTERVALS | RE_CONTEXT_INDEP_OPS \ 257 | RE_CONTEXT_INVALID_OPS )) 258 259 # define RE_SYNTAX_POSIX_AWK \ 260 (RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS \ 261 | RE_INTERVALS | RE_NO_GNU_OPS) 262 263 # define RE_SYNTAX_GREP \ 264 (RE_BK_PLUS_QM | RE_CHAR_CLASSES \ 265 | RE_HAT_LISTS_NOT_NEWLINE | RE_INTERVALS \ 266 | RE_NEWLINE_ALT) 267 268 # define RE_SYNTAX_EGREP \ 269 (RE_CHAR_CLASSES | RE_CONTEXT_INDEP_ANCHORS \ 270 | RE_CONTEXT_INDEP_OPS | RE_HAT_LISTS_NOT_NEWLINE \ 271 | RE_NEWLINE_ALT | RE_NO_BK_PARENS \ 272 | RE_NO_BK_VBAR) 273 274 # define RE_SYNTAX_POSIX_EGREP \ 275 (RE_SYNTAX_EGREP | RE_INTERVALS | RE_NO_BK_BRACES \ 276 | RE_INVALID_INTERVAL_ORD) 277 278 /* P1003.2/D11.2, section 4.20.7.1, lines 5078ff. */ 279 # define RE_SYNTAX_ED RE_SYNTAX_POSIX_BASIC 280 281 # define RE_SYNTAX_SED RE_SYNTAX_POSIX_BASIC 282 283 /* Syntax bits common to both basic and extended POSIX regex syntax. */ 284 # define _RE_SYNTAX_POSIX_COMMON \ 285 (RE_CHAR_CLASSES | RE_DOT_NEWLINE | RE_DOT_NOT_NULL \ 286 | RE_INTERVALS | RE_NO_EMPTY_RANGES) 287 288 # define RE_SYNTAX_POSIX_BASIC \ 289 (_RE_SYNTAX_POSIX_COMMON | RE_BK_PLUS_QM | RE_CONTEXT_INVALID_DUP) 290 291 /* Differs from ..._POSIX_BASIC only in that RE_BK_PLUS_QM becomes 292 RE_LIMITED_OPS, i.e., \? \+ \| are not recognized. Actually, this 293 isn't minimal, since other operators, such as \`, aren't disabled. */ 294 # define RE_SYNTAX_POSIX_MINIMAL_BASIC \ 295 (_RE_SYNTAX_POSIX_COMMON | RE_LIMITED_OPS) 296 297 # define RE_SYNTAX_POSIX_EXTENDED \ 298 (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \ 299 | RE_CONTEXT_INDEP_OPS | RE_NO_BK_BRACES \ 300 | RE_NO_BK_PARENS | RE_NO_BK_VBAR \ 301 | RE_CONTEXT_INVALID_OPS | RE_UNMATCHED_RIGHT_PAREN_ORD) 302 303 /* Differs from ..._POSIX_EXTENDED in that RE_CONTEXT_INDEP_OPS is 304 removed and RE_NO_BK_REFS is added. */ 305 # define RE_SYNTAX_POSIX_MINIMAL_EXTENDED \ 306 (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \ 307 | RE_CONTEXT_INVALID_OPS | RE_NO_BK_BRACES \ 308 | RE_NO_BK_PARENS | RE_NO_BK_REFS \ 309 | RE_NO_BK_VBAR | RE_UNMATCHED_RIGHT_PAREN_ORD) 310 /* [[[end syntaxes]]] */ 311 312 #endif /* defined __USE_GNU_REGEX */ 313 314 #ifdef __USE_GNU_REGEX 315 316 /* Maximum number of duplicates an interval can allow. POSIX-conforming 317 systems might define this in <limits.h>, but we want our 318 value, so remove any previous define. */ 319 # ifdef RE_DUP_MAX 320 # undef RE_DUP_MAX 321 # endif 322 323 /* RE_DUP_MAX is 2**15 - 1 because an earlier implementation stored 324 the counter as a 2-byte signed integer. This is no longer true, so 325 RE_DUP_MAX could be increased to (INT_MAX / 10 - 1), or to 326 ((SIZE_MAX - 2) / 10 - 1) if _REGEX_LARGE_OFFSETS is defined. 327 However, there would be a huge performance problem if someone 328 actually used a pattern like a\{214748363\}, so RE_DUP_MAX retains 329 its historical value. */ 330 # define RE_DUP_MAX (0x7fff) 331 332 #endif /* defined __USE_GNU_REGEX */ 333 334 335 /* POSIX `cflags' bits (i.e., information for `regcomp'). */ 336 337 /* If this bit is set, then use extended regular expression syntax. 338 If not set, then use basic regular expression syntax. */ 339 #define REG_EXTENDED 1 340 341 /* If this bit is set, then ignore case when matching. 342 If not set, then case is significant. */ 343 #define REG_ICASE (1 << 1) 344 345 /* If this bit is set, then anchors do not match at newline 346 characters in the string. 347 If not set, then anchors do match at newlines. */ 348 #define REG_NEWLINE (1 << 2) 349 350 /* If this bit is set, then report only success or fail in regexec. 351 If not set, then returns differ between not matching and errors. */ 352 #define REG_NOSUB (1 << 3) 353 354 355 /* POSIX `eflags' bits (i.e., information for regexec). */ 356 357 /* If this bit is set, then the beginning-of-line operator doesn't match 358 the beginning of the string (presumably because it's not the 359 beginning of a line). 360 If not set, then the beginning-of-line operator does match the 361 beginning of the string. */ 362 #define REG_NOTBOL 1 363 364 /* Like REG_NOTBOL, except for the end-of-line. */ 365 #define REG_NOTEOL (1 << 1) 366 367 /* Use PMATCH[0] to delimit the start and end of the search in the 368 buffer. */ 369 #define REG_STARTEND (1 << 2) 370 371 372 /* If any error codes are removed, changed, or added, update the 373 `__re_error_msgid' table in regcomp.c. */ 374 375 typedef enum 376 { 377 _REG_ENOSYS = -1, /* This will never happen for this implementation. */ 378 _REG_NOERROR = 0, /* Success. */ 379 _REG_NOMATCH, /* Didn't find a match (for regexec). */ 380 381 /* POSIX regcomp return error codes. (In the order listed in the 382 standard.) */ 383 _REG_BADPAT, /* Invalid pattern. */ 384 _REG_ECOLLATE, /* Invalid collating element. */ 385 _REG_ECTYPE, /* Invalid character class name. */ 386 _REG_EESCAPE, /* Trailing backslash. */ 387 _REG_ESUBREG, /* Invalid back reference. */ 388 _REG_EBRACK, /* Unmatched left bracket. */ 389 _REG_EPAREN, /* Parenthesis imbalance. */ 390 _REG_EBRACE, /* Unmatched \{. */ 391 _REG_BADBR, /* Invalid contents of \{\}. */ 392 _REG_ERANGE, /* Invalid range end. */ 393 _REG_ESPACE, /* Ran out of memory. */ 394 _REG_BADRPT, /* No preceding re for repetition op. */ 395 396 /* Error codes we've added. */ 397 _REG_EEND, /* Premature end. */ 398 _REG_ESIZE, /* Compiled pattern bigger than 2^16 bytes. */ 399 _REG_ERPAREN /* Unmatched ) or \); not returned from regcomp. */ 400 } reg_errcode_t; 401 402 #ifdef _XOPEN_SOURCE 403 # define REG_ENOSYS _REG_ENOSYS 404 #endif 405 #define REG_NOERROR _REG_NOERROR 406 #define REG_NOMATCH _REG_NOMATCH 407 #define REG_BADPAT _REG_BADPAT 408 #define REG_ECOLLATE _REG_ECOLLATE 409 #define REG_ECTYPE _REG_ECTYPE 410 #define REG_EESCAPE _REG_EESCAPE 411 #define REG_ESUBREG _REG_ESUBREG 412 #define REG_EBRACK _REG_EBRACK 413 #define REG_EPAREN _REG_EPAREN 414 #define REG_EBRACE _REG_EBRACE 415 #define REG_BADBR _REG_BADBR 416 #define REG_ERANGE _REG_ERANGE 417 #define REG_ESPACE _REG_ESPACE 418 #define REG_BADRPT _REG_BADRPT 419 #define REG_EEND _REG_EEND 420 #define REG_ESIZE _REG_ESIZE 421 #define REG_ERPAREN _REG_ERPAREN 422 423 /* struct re_pattern_buffer normally uses member names like `buffer' 424 that POSIX does not allow. In POSIX mode these members have names 425 with leading `re_' (e.g., `re_buffer'). */ 426 #ifdef __USE_GNU_REGEX 427 # define _REG_RE_NAME(id) id 428 # define _REG_RM_NAME(id) id 429 #else 430 # define _REG_RE_NAME(id) re_##id 431 # define _REG_RM_NAME(id) rm_##id 432 #endif 433 434 /* The user can specify the type of the re_translate member by 435 defining the macro RE_TRANSLATE_TYPE, which defaults to unsigned 436 char *. This pollutes the POSIX name space, so in POSIX mode just 437 use unsigned char *. */ 438 #ifdef __USE_GNU_REGEX 439 # ifndef RE_TRANSLATE_TYPE 440 # define RE_TRANSLATE_TYPE unsigned char * 441 # endif 442 # define REG_TRANSLATE_TYPE RE_TRANSLATE_TYPE 443 #else 444 # define REG_TRANSLATE_TYPE unsigned char * 445 #endif 446 447 /* This data structure represents a compiled pattern. Before calling 448 the pattern compiler, the fields `buffer', `allocated', `fastmap', 449 `translate', and `no_sub' can be set. After the pattern has been 450 compiled, the `re_nsub' field is available. All other fields are 451 private to the regex routines. */ 452 453 struct re_pattern_buffer 454 { 455 /* Space that holds the compiled pattern. It is declared as 456 `unsigned char *' because its elements are sometimes used as 457 array indexes. */ 458 unsigned char *_REG_RE_NAME (buffer); 459 460 /* Number of bytes to which `buffer' points. */ 461 __re_long_size_t _REG_RE_NAME (allocated); 462 463 /* Number of bytes actually used in `buffer'. */ 464 __re_long_size_t _REG_RE_NAME (used); 465 466 /* Syntax setting with which the pattern was compiled. */ 467 reg_syntax_t _REG_RE_NAME (syntax); 468 469 /* Pointer to a fastmap, if any, otherwise zero. re_search uses the 470 fastmap, if there is one, to skip over impossible starting points 471 for matches. */ 472 char *_REG_RE_NAME (fastmap); 473 474 /* Either a translate table to apply to all characters before 475 comparing them, or zero for no translation. The translation is 476 applied to a pattern when it is compiled and to a string when it 477 is matched. */ 478 REG_TRANSLATE_TYPE _REG_RE_NAME (translate); 479 480 /* Number of subexpressions found by the compiler. */ 481 size_t re_nsub; 482 483 /* Zero if this pattern cannot match the empty string, one else. 484 Well, in truth it's used only in `re_search_2', to see whether or 485 not we should use the fastmap, so we don't set this absolutely 486 perfectly; see `re_compile_fastmap' (the `duplicate' case). */ 487 unsigned int _REG_RE_NAME (can_be_null) : 1; 488 489 /* If REGS_UNALLOCATED, allocate space in the `regs' structure 490 for `max (RE_NREGS, re_nsub + 1)' groups. 491 If REGS_REALLOCATE, reallocate space if necessary. 492 If REGS_FIXED, use what's there. */ 493 #ifdef __USE_GNU_REGEX 494 # define REGS_UNALLOCATED 0 495 # define REGS_REALLOCATE 1 496 # define REGS_FIXED 2 497 #endif 498 unsigned int _REG_RE_NAME (regs_allocated) : 2; 499 500 /* Set to zero when `re_compile_pattern' compiles a pattern; set to 501 one by `re_compile_fastmap' if it updates the fastmap. */ 502 unsigned int _REG_RE_NAME (fastmap_accurate) : 1; 503 504 /* If set, `re_match_2' does not return information about 505 subexpressions. */ 506 unsigned int _REG_RE_NAME (no_sub) : 1; 507 508 /* If set, a beginning-of-line anchor doesn't match at the beginning 509 of the string. */ 510 unsigned int _REG_RE_NAME (not_bol) : 1; 511 512 /* Similarly for an end-of-line anchor. */ 513 unsigned int _REG_RE_NAME (not_eol) : 1; 514 515 /* If true, an anchor at a newline matches. */ 516 unsigned int _REG_RE_NAME (newline_anchor) : 1; 517 518 /* [[[end pattern_buffer]]] */ 519 }; 520 521 typedef struct re_pattern_buffer regex_t; 522 523 /* This is the structure we store register match data in. See 524 regex.texinfo for a full description of what registers match. */ 525 struct re_registers 526 { 527 __re_size_t _REG_RM_NAME (num_regs); 528 regoff_t *_REG_RM_NAME (start); 529 regoff_t *_REG_RM_NAME (end); 530 }; 531 532 533 /* If `regs_allocated' is REGS_UNALLOCATED in the pattern buffer, 534 `re_match_2' returns information about at least this many registers 535 the first time a `regs' structure is passed. */ 536 #if !defined RE_NREGS && defined __USE_GNU_REGEX 537 # define RE_NREGS 30 538 #endif 539 540 541 /* POSIX specification for registers. Aside from the different names than 542 `re_registers', POSIX uses an array of structures, instead of a 543 structure of arrays. */ 544 typedef struct 545 { 546 regoff_t rm_so; /* Byte offset from string's start to substring's start. */ 547 regoff_t rm_eo; /* Byte offset from string's start to substring's end. */ 548 } regmatch_t; 549 550 /* Declarations for routines. */ 551 552 /* Sets the current default syntax to SYNTAX, and return the old syntax. 553 You can also simply assign to the `re_syntax_options' variable. */ 554 extern reg_syntax_t re_set_syntax (reg_syntax_t __syntax); 555 556 /* Compile the regular expression PATTERN, with length LENGTH 557 and syntax given by the global `re_syntax_options', into the buffer 558 BUFFER. Return NULL if successful, and an error string if not. */ 559 extern const char *re_compile_pattern (const char *__pattern, size_t __length, 560 struct re_pattern_buffer *__buffer); 561 562 563 /* Compile a fastmap for the compiled pattern in BUFFER; used to 564 accelerate searches. Return 0 if successful and -2 if was an 565 internal error. */ 566 extern int re_compile_fastmap (struct re_pattern_buffer *__buffer); 567 568 569 /* Search in the string STRING (with length LENGTH) for the pattern 570 compiled into BUFFER. Start searching at position START, for RANGE 571 characters. Return the starting position of the match, -1 for no 572 match, or -2 for an internal error. Also return register 573 information in REGS (if REGS and BUFFER->no_sub are nonzero). */ 574 extern regoff_t re_search (struct re_pattern_buffer *__buffer, 575 const char *__string, __re_idx_t __length, 576 __re_idx_t __start, regoff_t __range, 577 struct re_registers *__regs); 578 579 580 /* Like `re_search', but search in the concatenation of STRING1 and 581 STRING2. Also, stop searching at index START + STOP. */ 582 extern regoff_t re_search_2 (struct re_pattern_buffer *__buffer, 583 const char *__string1, __re_idx_t __length1, 584 const char *__string2, __re_idx_t __length2, 585 __re_idx_t __start, regoff_t __range, 586 struct re_registers *__regs, 587 __re_idx_t __stop); 588 589 590 /* Like `re_search', but return how many characters in STRING the regexp 591 in BUFFER matched, starting at position START. */ 592 extern regoff_t re_match (struct re_pattern_buffer *__buffer, 593 const char *__string, __re_idx_t __length, 594 __re_idx_t __start, struct re_registers *__regs); 595 596 597 /* Relates to `re_match' as `re_search_2' relates to `re_search'. */ 598 extern regoff_t re_match_2 (struct re_pattern_buffer *__buffer, 599 const char *__string1, __re_idx_t __length1, 600 const char *__string2, __re_idx_t __length2, 601 __re_idx_t __start, struct re_registers *__regs, 602 __re_idx_t __stop); 603 604 605 /* Set REGS to hold NUM_REGS registers, storing them in STARTS and 606 ENDS. Subsequent matches using BUFFER and REGS will use this memory 607 for recording register information. STARTS and ENDS must be 608 allocated with malloc, and must each be at least `NUM_REGS * sizeof 609 (regoff_t)' bytes long. 610 611 If NUM_REGS == 0, then subsequent matches should allocate their own 612 register data. 613 614 Unless this function is called, the first search or match using 615 BUFFER will allocate its own register data, without freeing the old 616 data. */ 617 extern void re_set_registers (struct re_pattern_buffer *__buffer, 618 struct re_registers *__regs, 619 __re_size_t __num_regs, 620 regoff_t *__starts, regoff_t *__ends); 621 622 #if defined _REGEX_RE_COMP || defined _LIBC 623 # ifndef _CRAY 624 /* 4.2 bsd compatibility. */ 625 extern char *re_comp (const char *); 626 extern int re_exec (const char *); 627 # endif 628 #endif 629 630 /* GCC 2.95 and later have "__restrict"; C99 compilers have 631 "restrict", and "configure" may have defined "restrict". 632 Other compilers use __restrict, __restrict__, and _Restrict, and 633 'configure' might #define 'restrict' to those words, so pick a 634 different name. */ 635 #ifndef _Restrict_ 636 # if 199901L <= __STDC_VERSION__ 637 # define _Restrict_ restrict 638 # elif 2 < __GNUC__ || (2 == __GNUC__ && 95 <= __GNUC_MINOR__) 639 # define _Restrict_ __restrict 640 # else 641 # define _Restrict_ 642 # endif 643 #endif 644 /* gcc 3.1 and up support the [restrict] syntax. Don't trust 645 sys/cdefs.h's definition of __restrict_arr, though, as it 646 mishandles gcc -ansi -pedantic. */ 647 #ifndef _Restrict_arr_ 648 # if ((199901L <= __STDC_VERSION__ \ 649 || ((3 < __GNUC__ || (3 == __GNUC__ && 1 <= __GNUC_MINOR__)) \ 650 && !__STRICT_ANSI__)) \ 651 && !defined __GNUG__) 652 # define _Restrict_arr_ _Restrict_ 653 # else 654 # define _Restrict_arr_ 655 # endif 656 #endif 657 658 /* POSIX compatibility. */ 659 extern int regcomp (regex_t *_Restrict_ __preg, 660 const char *_Restrict_ __pattern, 661 int __cflags); 662 663 extern int regexec (const regex_t *_Restrict_ __preg, 664 const char *_Restrict_ __string, size_t __nmatch, 665 regmatch_t __pmatch[_Restrict_arr_], 666 int __eflags); 667 668 extern size_t regerror (int __errcode, const regex_t *_Restrict_ __preg, 669 char *_Restrict_ __errbuf, size_t __errbuf_size); 670 671 extern void regfree (regex_t *__preg); 672 673 674 #ifdef __cplusplus 675 } 676 #endif /* C++ */ 677 678 #endif /* regex.h */ 679