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 4 Free Software Foundation, Inc. 5 This file is part of the GNU C Library. 6 7 This program is free software; you can redistribute it and/or modify 8 it under the terms of the GNU General Public License as published by 9 the Free Software Foundation; either version 2, or (at your option) 10 any later version. 11 12 This program 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 15 GNU General Public License for more details. 16 17 You should have received a copy of the GNU General Public License along 18 with this program; if not, write to the Free Software Foundation, 19 Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. */ 20 21 #ifndef _REGEX_H 22 #define _REGEX_H 1 23 24 #include <sys/types.h> 25 26 /* Allow the use in C++ code. */ 27 #ifdef __cplusplus 28 extern "C" { 29 #endif 30 31 /* Define _REGEX_SOURCE to get definitions that are incompatible with 32 POSIX. */ 33 #if (!defined _REGEX_SOURCE \ 34 && (defined _GNU_SOURCE \ 35 || (!defined _POSIX_C_SOURCE && !defined _POSIX_SOURCE \ 36 && !defined _XOPEN_SOURCE))) 37 # define _REGEX_SOURCE 1 38 #endif 39 40 #if defined _REGEX_SOURCE && defined VMS 41 /* VMS doesn't have `size_t' in <sys/types.h>, even though POSIX says it 42 should be there. */ 43 # include <stddef.h> 44 #endif 45 46 #ifdef _REGEX_LARGE_OFFSETS 47 48 /* Use types and values that are wide enough to represent signed and 49 unsigned byte offsets in memory. This currently works only when 50 the regex code is used outside of the GNU C library; it is not yet 51 supported within glibc itself, and glibc users should not define 52 _REGEX_LARGE_OFFSETS. */ 53 54 /* The type of the offset of a byte within a string. 55 For historical reasons POSIX 1003.1-2004 requires that regoff_t be 56 at least as wide as off_t. This is a bit odd (and many common 57 POSIX platforms set it to the more-sensible ssize_t) but we might 58 as well conform. We don't know of any hosts where ssize_t is wider 59 than off_t, so off_t is safe. */ 60 typedef off_t regoff_t; 61 62 /* The type of nonnegative object indexes. Traditionally, GNU regex 63 uses 'int' for these. Code that uses __re_idx_t should work 64 regardless of whether the type is signed. */ 65 typedef size_t __re_idx_t; 66 67 /* The type of object sizes. */ 68 typedef size_t __re_size_t; 69 70 /* The type of object sizes, in places where the traditional code 71 uses unsigned long int. */ 72 typedef size_t __re_long_size_t; 73 74 #else 75 76 /* Use types that are binary-compatible with the traditional GNU regex 77 implementation, which mishandles strings longer than INT_MAX. */ 78 79 typedef int regoff_t; 80 typedef int __re_idx_t; 81 typedef unsigned int __re_size_t; 82 typedef unsigned long int __re_long_size_t; 83 84 #endif 85 86 /* The following two types have to be signed and unsigned integer type 87 wide enough to hold a value of a pointer. For most ANSI compilers 88 ptrdiff_t and size_t should be likely OK. Still size of these two 89 types is 2 for Microsoft C. Ugh... */ 90 typedef long int s_reg_t; 91 typedef unsigned long int active_reg_t; 92 93 /* The following bits are used to determine the regexp syntax we 94 recognize. The set/not-set meanings are chosen so that Emacs syntax 95 remains the value 0. The bits are given in alphabetical order, and 96 the definitions shifted by one from the previous bit; thus, when we 97 add or remove a bit, only one other definition need change. */ 98 typedef unsigned long int reg_syntax_t; 99 100 /* If this bit is not set, then \ inside a bracket expression is literal. 101 If set, then such a \ quotes the following character. */ 102 #define REG_BACKSLASH_ESCAPE_IN_LISTS 1ul 103 104 /* If this bit is not set, then + and ? are operators, and \+ and \? are 105 literals. 106 If set, then \+ and \? are operators and + and ? are literals. */ 107 #define REG_BK_PLUS_QM (1ul << 1) 108 109 /* If this bit is set, then character classes are supported. They are: 110 [:alpha:], [:upper:], [:lower:], [:digit:], [:alnum:], [:xdigit:], 111 [:space:], [:print:], [:punct:], [:graph:], and [:cntrl:]. 112 If not set, then character classes are not supported. */ 113 #define REG_CHAR_CLASSES (1ul << 2) 114 115 /* If this bit is set, then ^ and $ are always anchors (outside bracket 116 expressions, of course). 117 If this bit is not set, then it depends: 118 ^ is an anchor if it is at the beginning of a regular 119 expression or after an open-group or an alternation operator; 120 $ is an anchor if it is at the end of a regular expression, or 121 before a close-group or an alternation operator. 122 123 This bit could be (re)combined with REG_CONTEXT_INDEP_OPS, because 124 POSIX draft 11.2 says that * etc. in leading positions is undefined. 125 We already implemented a previous draft which made those constructs 126 invalid, though, so we haven't changed the code back. */ 127 #define REG_CONTEXT_INDEP_ANCHORS (1ul << 3) 128 129 /* If this bit is set, then special characters are always special 130 regardless of where they are in the pattern. 131 If this bit is not set, then special characters are special only in 132 some contexts; otherwise they are ordinary. Specifically, 133 * + ? and intervals are only special when not after the beginning, 134 open-group, or alternation operator. */ 135 #define REG_CONTEXT_INDEP_OPS (1ul << 4) 136 137 /* If this bit is set, then *, +, ?, and { cannot be first in an re or 138 immediately after an alternation or begin-group operator. */ 139 #define REG_CONTEXT_INVALID_OPS (1ul << 5) 140 141 /* If this bit is set, then . matches newline. 142 If not set, then it doesn't. */ 143 #define REG_DOT_NEWLINE (1ul << 6) 144 145 /* If this bit is set, then . doesn't match NUL. 146 If not set, then it does. */ 147 #define REG_DOT_NOT_NULL (1ul << 7) 148 149 /* If this bit is set, nonmatching lists [^...] do not match newline. 150 If not set, they do. */ 151 #define REG_HAT_LISTS_NOT_NEWLINE (1ul << 8) 152 153 /* If this bit is set, either \{...\} or {...} defines an 154 interval, depending on REG_NO_BK_BRACES. 155 If not set, \{, \}, {, and } are literals. */ 156 #define REG_INTERVALS (1ul << 9) 157 158 /* If this bit is set, +, ? and | aren't recognized as operators. 159 If not set, they are. */ 160 #define REG_LIMITED_OPS (1ul << 10) 161 162 /* If this bit is set, newline is an alternation operator. 163 If not set, newline is literal. */ 164 #define REG_NEWLINE_ALT (1ul << 11) 165 166 /* If this bit is set, then `{...}' defines an interval, and \{ and \} 167 are literals. 168 If not set, then `\{...\}' defines an interval. */ 169 #define REG_NO_BK_BRACES (1ul << 12) 170 171 /* If this bit is set, (...) defines a group, and \( and \) are literals. 172 If not set, \(...\) defines a group, and ( and ) are literals. */ 173 #define REG_NO_BK_PARENS (1ul << 13) 174 175 /* If this bit is set, then \<digit> matches <digit>. 176 If not set, then \<digit> is a back-reference. */ 177 #define REG_NO_BK_REFS (1ul << 14) 178 179 /* If this bit is set, then | is an alternation operator, and \| is literal. 180 If not set, then \| is an alternation operator, and | is literal. */ 181 #define REG_NO_BK_VBAR (1ul << 15) 182 183 /* If this bit is set, then an ending range point collating higher 184 than the starting range point, as in [z-a], is invalid. 185 If not set, the containing range is empty and does not match any string. */ 186 #define REG_NO_EMPTY_RANGES (1ul << 16) 187 188 /* If this bit is set, then an unmatched ) is ordinary. 189 If not set, then an unmatched ) is invalid. */ 190 #define REG_UNMATCHED_RIGHT_PAREN_ORD (1ul << 17) 191 192 /* If this bit is set, succeed as soon as we match the whole pattern, 193 without further backtracking. */ 194 #define REG_NO_POSIX_BACKTRACKING (1ul << 18) 195 196 /* If this bit is set, do not process the GNU regex operators. 197 If not set, then the GNU regex operators are recognized. */ 198 #define REG_NO_GNU_OPS (1ul << 19) 199 200 /* If this bit is set, turn on internal regex debugging. 201 If not set, and debugging was on, turn it off. 202 This only works if regex.c is compiled -DDEBUG. 203 We define this bit always, so that all that's needed to turn on 204 debugging is to recompile regex.c; the calling code can always have 205 this bit set, and it won't affect anything in the normal case. */ 206 #define REG_DEBUG (1ul << 20) 207 208 /* If this bit is set, a syntactically invalid interval is treated as 209 a string of ordinary characters. For example, the ERE 'a{1' is 210 treated as 'a\{1'. */ 211 #define REG_INVALID_INTERVAL_ORD (1ul << 21) 212 213 /* If this bit is set, then ignore case when matching. 214 If not set, then case is significant. */ 215 #define REG_IGNORE_CASE (1ul << 22) 216 217 /* This bit is used internally like REG_CONTEXT_INDEP_ANCHORS but only 218 for ^, because it is difficult to scan the regex backwards to find 219 whether ^ should be special. */ 220 #define REG_CARET_ANCHORS_HERE (1ul << 23) 221 222 /* If this bit is set, then \{ cannot be first in an bre or 223 immediately after an alternation or begin-group operator. */ 224 #define REG_CONTEXT_INVALID_DUP (1ul << 24) 225 226 /* If this bit is set, then no_sub will be set to 1 during 227 re_compile_pattern. */ 228 #define REG_NO_SUB (1ul << 25) 229 230 /* This global variable defines the particular regexp syntax to use (for 231 some interfaces). When a regexp is compiled, the syntax used is 232 stored in the pattern buffer, so changing this does not affect 233 already-compiled regexps. */ 234 extern reg_syntax_t re_syntax_options; 235 236 /* Define combinations of the above bits for the standard possibilities. 237 (The [[[ comments delimit what gets put into the Texinfo file, so 238 don't delete them!) */ 239 /* [[[begin syntaxes]]] */ 240 #define REG_SYNTAX_EMACS 0 241 242 #define REG_SYNTAX_AWK \ 243 (REG_BACKSLASH_ESCAPE_IN_LISTS | REG_DOT_NOT_NULL \ 244 | REG_NO_BK_PARENS | REG_NO_BK_REFS \ 245 | REG_NO_BK_VBAR | REG_NO_EMPTY_RANGES \ 246 | REG_DOT_NEWLINE | REG_CONTEXT_INDEP_ANCHORS \ 247 | REG_UNMATCHED_RIGHT_PAREN_ORD | REG_NO_GNU_OPS) 248 249 #define REG_SYNTAX_GNU_AWK \ 250 ((REG_SYNTAX_POSIX_EXTENDED | REG_BACKSLASH_ESCAPE_IN_LISTS \ 251 | REG_DEBUG) \ 252 & ~(REG_DOT_NOT_NULL | REG_INTERVALS | REG_CONTEXT_INDEP_OPS \ 253 | REG_CONTEXT_INVALID_OPS )) 254 255 #define REG_SYNTAX_POSIX_AWK \ 256 (REG_SYNTAX_POSIX_EXTENDED | REG_BACKSLASH_ESCAPE_IN_LISTS \ 257 | REG_INTERVALS | REG_NO_GNU_OPS) 258 259 #define REG_SYNTAX_GREP \ 260 (REG_BK_PLUS_QM | REG_CHAR_CLASSES \ 261 | REG_HAT_LISTS_NOT_NEWLINE | REG_INTERVALS \ 262 | REG_NEWLINE_ALT) 263 264 #define REG_SYNTAX_EGREP \ 265 (REG_CHAR_CLASSES | REG_CONTEXT_INDEP_ANCHORS \ 266 | REG_CONTEXT_INDEP_OPS | REG_HAT_LISTS_NOT_NEWLINE \ 267 | REG_NEWLINE_ALT | REG_NO_BK_PARENS \ 268 | REG_NO_BK_VBAR) 269 270 #define REG_SYNTAX_POSIX_EGREP \ 271 (REG_SYNTAX_EGREP | REG_INTERVALS | REG_NO_BK_BRACES \ 272 | REG_INVALID_INTERVAL_ORD) 273 274 /* P1003.2/D11.2, section 4.20.7.1, lines 5078ff. */ 275 #define REG_SYNTAX_ED REG_SYNTAX_POSIX_BASIC 276 277 #define REG_SYNTAX_SED REG_SYNTAX_POSIX_BASIC 278 279 /* Syntax bits common to both basic and extended POSIX regex syntax. */ 280 #define _REG_SYNTAX_POSIX_COMMON \ 281 (REG_CHAR_CLASSES | REG_DOT_NEWLINE | REG_DOT_NOT_NULL \ 282 | REG_INTERVALS | REG_NO_EMPTY_RANGES) 283 284 #define REG_SYNTAX_POSIX_BASIC \ 285 (_REG_SYNTAX_POSIX_COMMON | REG_BK_PLUS_QM | REG_CONTEXT_INVALID_DUP) 286 287 /* Differs from ..._POSIX_BASIC only in that REG_BK_PLUS_QM becomes 288 REG_LIMITED_OPS, i.e., \? \+ \| are not recognized. Actually, this 289 isn't minimal, since other operators, such as \`, aren't disabled. */ 290 #define REG_SYNTAX_POSIX_MINIMAL_BASIC \ 291 (_REG_SYNTAX_POSIX_COMMON | REG_LIMITED_OPS) 292 293 #define REG_SYNTAX_POSIX_EXTENDED \ 294 (_REG_SYNTAX_POSIX_COMMON | REG_CONTEXT_INDEP_ANCHORS \ 295 | REG_CONTEXT_INDEP_OPS | REG_NO_BK_BRACES \ 296 | REG_NO_BK_PARENS | REG_NO_BK_VBAR \ 297 | REG_CONTEXT_INVALID_OPS | REG_UNMATCHED_RIGHT_PAREN_ORD) 298 299 /* Differs from ..._POSIX_EXTENDED in that REG_CONTEXT_INDEP_OPS is 300 removed and REG_NO_BK_REFS is added. */ 301 #define REG_SYNTAX_POSIX_MINIMAL_EXTENDED \ 302 (_REG_SYNTAX_POSIX_COMMON | REG_CONTEXT_INDEP_ANCHORS \ 303 | REG_CONTEXT_INVALID_OPS | REG_NO_BK_BRACES \ 304 | REG_NO_BK_PARENS | REG_NO_BK_REFS \ 305 | REG_NO_BK_VBAR | REG_UNMATCHED_RIGHT_PAREN_ORD) 306 /* [[[end syntaxes]]] */ 307 308 /* Maximum number of duplicates an interval can allow. This is 309 distinct from RE_DUP_MAX, to conform to POSIX name space rules and 310 to avoid collisions with <limits.h>. */ 311 #define REG_DUP_MAX 32767 312 313 314 /* POSIX `cflags' bits (i.e., information for `regcomp'). */ 315 316 /* If this bit is set, then use extended regular expression syntax. 317 If not set, then use basic regular expression syntax. */ 318 #define REG_EXTENDED 1 319 320 /* If this bit is set, then ignore case when matching. 321 If not set, then case is significant. */ 322 #define REG_ICASE (1 << 1) 323 324 /* If this bit is set, then anchors do not match at newline 325 characters in the string. 326 If not set, then anchors do match at newlines. */ 327 #define REG_NEWLINE (1 << 2) 328 329 /* If this bit is set, then report only success or fail in regexec. 330 If not set, then returns differ between not matching and errors. */ 331 #define REG_NOSUB (1 << 3) 332 333 334 /* POSIX `eflags' bits (i.e., information for regexec). */ 335 336 /* If this bit is set, then the beginning-of-line operator doesn't match 337 the beginning of the string (presumably because it's not the 338 beginning of a line). 339 If not set, then the beginning-of-line operator does match the 340 beginning of the string. */ 341 #define REG_NOTBOL 1 342 343 /* Like REG_NOTBOL, except for the end-of-line. */ 344 #define REG_NOTEOL (1 << 1) 345 346 /* Use PMATCH[0] to delimit the start and end of the search in the 347 buffer. */ 348 #define REG_STARTEND (1 << 2) 349 350 351 /* If any error codes are removed, changed, or added, update the 352 `__re_error_msgid' table in regcomp.c. */ 353 354 typedef enum 355 { 356 _REG_ENOSYS = -1, /* This will never happen for this implementation. */ 357 #define REG_ENOSYS _REG_ENOSYS 358 359 _REG_NOERROR, /* Success. */ 360 #define REG_NOERROR _REG_NOERROR 361 362 _REG_NOMATCH, /* Didn't find a match (for regexec). */ 363 #define REG_NOMATCH _REG_NOMATCH 364 365 /* POSIX regcomp return error codes. (In the order listed in the 366 standard.) */ 367 368 _REG_BADPAT, /* Invalid pattern. */ 369 #define REG_BADPAT _REG_BADPAT 370 371 _REG_ECOLLATE, /* Inalid collating element. */ 372 #define REG_ECOLLATE _REG_ECOLLATE 373 374 _REG_ECTYPE, /* Invalid character class name. */ 375 #define REG_ECTYPE _REG_ECTYPE 376 377 _REG_EESCAPE, /* Trailing backslash. */ 378 #define REG_EESCAPE _REG_EESCAPE 379 380 _REG_ESUBREG, /* Invalid back reference. */ 381 #define REG_ESUBREG _REG_ESUBREG 382 383 _REG_EBRACK, /* Unmatched left bracket. */ 384 #define REG_EBRACK _REG_EBRACK 385 386 _REG_EPAREN, /* Parenthesis imbalance. */ 387 #define REG_EPAREN _REG_EPAREN 388 389 _REG_EBRACE, /* Unmatched \{. */ 390 #define REG_EBRACE _REG_EBRACE 391 392 _REG_BADBR, /* Invalid contents of \{\}. */ 393 #define REG_BADBR _REG_BADBR 394 395 _REG_ERANGE, /* Invalid range end. */ 396 #define REG_ERANGE _REG_ERANGE 397 398 _REG_ESPACE, /* Ran out of memory. */ 399 #define REG_ESPACE _REG_ESPACE 400 401 _REG_BADRPT, /* No preceding re for repetition op. */ 402 #define REG_BADRPT _REG_BADRPT 403 404 /* Error codes we've added. */ 405 406 _REG_EEND, /* Premature end. */ 407 #define REG_EEND _REG_EEND 408 409 _REG_ESIZE, /* Compiled pattern bigger than 2^16 bytes. */ 410 #define REG_ESIZE _REG_ESIZE 411 412 _REG_ERPAREN /* Unmatched ) or \); not returned from regcomp. */ 413 #define REG_ERPAREN _REG_ERPAREN 414 415 } reg_errcode_t; 416 417 /* In the traditional GNU implementation, regex.h defined member names 418 like `buffer' that POSIX does not allow. These members now have 419 names with leading `re_' (e.g., `re_buffer'). Support the old 420 names only if _REGEX_SOURCE is defined. New programs should use 421 the new names. */ 422 #ifdef _REGEX_SOURCE 423 # define _REG_RE_NAME(id) id 424 # define _REG_RM_NAME(id) id 425 #else 426 # define _REG_RE_NAME(id) re_##id 427 # define _REG_RM_NAME(id) rm_##id 428 #endif 429 430 /* The user can specify the type of the re_translate member by 431 defining the macro REG_TRANSLATE_TYPE. In the traditional GNU 432 implementation, this macro was named RE_TRANSLATE_TYPE, but POSIX 433 does not allow this. Support the old name only if _REGEX_SOURCE 434 and if the new name is not defined. New programs should use the new 435 name. */ 436 #ifndef REG_TRANSLATE_TYPE 437 # if defined _REGEX_SOURCE && defined RE_TRANSLATE_TYPE 438 # define REG_TRANSLATE_TYPE RE_TRANSLATE_TYPE 439 # else 440 # define REG_TRANSLATE_TYPE char * 441 # endif 442 #endif 443 444 /* This data structure represents a compiled pattern. Before calling 445 the pattern compiler), the fields `re_buffer', `re_allocated', `re_fastmap', 446 `re_translate', and `re_no_sub' can be set. After the pattern has been 447 compiled, the `re_nsub' field is available. All other fields are 448 private to the regex routines. */ 449 450 struct re_pattern_buffer 451 { 452 /* [[[begin pattern_buffer]]] */ 453 /* Space that holds the compiled pattern. It is declared as 454 `unsigned char *' because its elements are 455 sometimes used as array indexes. */ 456 unsigned char *_REG_RE_NAME (buffer); 457 458 /* Number of bytes to which `re_buffer' points. */ 459 __re_long_size_t _REG_RE_NAME (allocated); 460 461 /* Number of bytes actually used in `re_buffer'. */ 462 __re_long_size_t _REG_RE_NAME (used); 463 464 /* Syntax setting with which the pattern was compiled. */ 465 reg_syntax_t _REG_RE_NAME (syntax); 466 467 /* Pointer to a fastmap, if any, otherwise zero. re_search uses 468 the fastmap, if there is one, to skip over impossible 469 starting points for matches. */ 470 char *_REG_RE_NAME (fastmap); 471 472 /* Either a translate table to apply to all characters before 473 comparing them, or zero for no translation. The translation 474 is applied to a pattern when it is compiled and to a string 475 when it is matched. */ 476 REG_TRANSLATE_TYPE _REG_RE_NAME (translate); 477 478 /* Number of subexpressions found by the compiler. */ 479 size_t re_nsub; 480 481 /* Zero if this pattern cannot match the empty string, one else. 482 Well, in truth it's used only in `re_search_2', to see 483 whether or not we should use the fastmap, so we don't set 484 this absolutely perfectly; see `re_compile_fastmap' (the 485 `duplicate' case). */ 486 unsigned int _REG_RE_NAME (can_be_null) : 1; 487 488 /* If REG_UNALLOCATED, allocate space in the `regs' structure 489 for `max (REG_NREGS, re_nsub + 1)' groups. 490 If REG_REALLOCATE, reallocate space if necessary. 491 If REG_FIXED, use what's there. */ 492 #define REG_UNALLOCATED 0 493 #define REG_REALLOCATE 1 494 #define REG_FIXED 2 495 unsigned int _REG_RE_NAME (regs_allocated) : 2; 496 497 /* Set to zero when `regex_compile' compiles a pattern; set to one 498 by `re_compile_fastmap' if it updates the fastmap. */ 499 unsigned int _REG_RE_NAME (fastmap_accurate) : 1; 500 501 /* If set, `re_match_2' does not return information about 502 subexpressions. */ 503 unsigned int _REG_RE_NAME (no_sub) : 1; 504 505 /* If set, a beginning-of-line anchor doesn't match at the 506 beginning of the string. */ 507 unsigned int _REG_RE_NAME (not_bol) : 1; 508 509 /* Similarly for an end-of-line anchor. */ 510 unsigned int _REG_RE_NAME (not_eol) : 1; 511 512 /* If true, an anchor at a newline matches. */ 513 unsigned int _REG_RE_NAME (newline_anchor) : 1; 514 515 /* [[[end pattern_buffer]]] */ 516 }; 517 518 typedef struct re_pattern_buffer regex_t; 519 520 /* This is the structure we store register match data in. See 521 regex.texinfo for a full description of what registers match. */ 522 struct re_registers 523 { 524 __re_size_t _REG_RM_NAME (num_regs); 525 regoff_t *_REG_RM_NAME (start); 526 regoff_t *_REG_RM_NAME (end); 527 }; 528 529 530 /* If `regs_allocated' is REG_UNALLOCATED in the pattern buffer, 531 `re_match_2' returns information about at least this many registers 532 the first time a `regs' structure is passed. */ 533 #ifndef REG_NREGS 534 # define REG_NREGS 30 535 #endif 536 537 538 /* POSIX specification for registers. Aside from the different names than 539 `re_registers', POSIX uses an array of structures, instead of a 540 structure of arrays. */ 541 typedef struct 542 { 543 regoff_t rm_so; /* Byte offset from string's start to substring's start. */ 544 regoff_t rm_eo; /* Byte offset from string's start to substring's end. */ 545 } regmatch_t; 546 547 /* Declarations for routines. */ 548 549 /* Sets the current default syntax to SYNTAX, and return the old syntax. 550 You can also simply assign to the `re_syntax_options' variable. */ 551 extern reg_syntax_t re_set_syntax (reg_syntax_t __syntax); 552 553 /* Compile the regular expression PATTERN, with length LENGTH 554 and syntax given by the global `re_syntax_options', into the buffer 555 BUFFER. Return NULL if successful, and an error string if not. */ 556 extern const char *re_compile_pattern (const char *__pattern, size_t __length, 557 struct re_pattern_buffer *__buffer); 558 559 560 /* Compile a fastmap for the compiled pattern in BUFFER; used to 561 accelerate searches. Return 0 if successful and -2 if was an 562 internal error. */ 563 extern int re_compile_fastmap (struct re_pattern_buffer *__buffer); 564 565 566 /* Search in the string STRING (with length LENGTH) for the pattern 567 compiled into BUFFER. Start searching at position START, for RANGE 568 characters. Return the starting position of the match, -1 for no 569 match, or -2 for an internal error. Also return register 570 information in REGS (if REGS and BUFFER->re_no_sub are nonzero). */ 571 extern regoff_t re_search (struct re_pattern_buffer *__buffer, 572 const char *__string, __re_idx_t __length, 573 __re_idx_t __start, regoff_t __range, 574 struct re_registers *__regs); 575 576 577 /* Like `re_search', but search in the concatenation of STRING1 and 578 STRING2. Also, stop searching at index START + STOP. */ 579 extern regoff_t re_search_2 (struct re_pattern_buffer *__buffer, 580 const char *__string1, __re_idx_t __length1, 581 const char *__string2, __re_idx_t __length2, 582 __re_idx_t __start, regoff_t __range, 583 struct re_registers *__regs, 584 __re_idx_t __stop); 585 586 587 /* Like `re_search', but return how many characters in STRING the regexp 588 in BUFFER matched, starting at position START. */ 589 extern regoff_t re_match (struct re_pattern_buffer *__buffer, 590 const char *__string, __re_idx_t __length, 591 __re_idx_t __start, struct re_registers *__regs); 592 593 594 /* Relates to `re_match' as `re_search_2' relates to `re_search'. */ 595 extern regoff_t re_match_2 (struct re_pattern_buffer *__buffer, 596 const char *__string1, __re_idx_t __length1, 597 const char *__string2, __re_idx_t __length2, 598 __re_idx_t __start, struct re_registers *__regs, 599 __re_idx_t __stop); 600 601 602 /* Set REGS to hold NUM_REGS registers, storing them in STARTS and 603 ENDS. Subsequent matches using BUFFER and REGS will use this memory 604 for recording register information. STARTS and ENDS must be 605 allocated with malloc, and must each be at least `NUM_REGS * sizeof 606 (regoff_t)' bytes long. 607 608 If NUM_REGS == 0, then subsequent matches should allocate their own 609 register data. 610 611 Unless this function is called, the first search or match using 612 PATTERN_BUFFER will allocate its own register data, without 613 freeing the old data. */ 614 extern void re_set_registers (struct re_pattern_buffer *__buffer, 615 struct re_registers *__regs, 616 __re_size_t __num_regs, 617 regoff_t *__starts, regoff_t *__ends); 618 619 #if defined _REGEX_RE_COMP || defined _LIBC 620 # ifndef _CRAY 621 /* 4.2 bsd compatibility. */ 622 extern char *re_comp (const char *); 623 extern int re_exec (const char *); 624 # endif 625 #endif 626 627 /* GCC 2.95 and later have "__restrict"; C99 compilers have 628 "restrict", and "configure" may have defined "restrict". */ 629 #ifndef __restrict 630 # if ! (2 < __GNUC__ || (2 == __GNUC__ && 95 <= __GNUC_MINOR__)) 631 # if defined restrict || 199901L <= __STDC_VERSION__ 632 # define __restrict restrict 633 # else 634 # define __restrict 635 # endif 636 # endif 637 #endif 638 /* gcc 3.1 and up support the [restrict] syntax, but g++ doesn't. */ 639 #ifndef __restrict_arr 640 # if (__GNUC__ > 3 || (__GNUC__ == 3 && __GNUC_MINOR__ >= 1)) && !defined __cplusplus 641 # define __restrict_arr __restrict 642 # else 643 # define __restrict_arr 644 # endif 645 #endif 646 647 /* POSIX compatibility. */ 648 extern int regcomp (regex_t *__restrict __preg, 649 const char *__restrict __pattern, 650 int __cflags); 651 652 extern int regexec (const regex_t *__restrict __preg, 653 const char *__restrict __string, size_t __nmatch, 654 regmatch_t __pmatch[__restrict_arr], 655 int __eflags); 656 657 extern size_t regerror (int __errcode, const regex_t *__restrict __preg, 658 char *__restrict __errbuf, size_t __errbuf_size); 659 660 extern void regfree (regex_t *__preg); 661 662 663 #ifdef _REGEX_SOURCE 664 665 /* Define the POSIX-compatible member names in terms of the 666 incompatible (and deprecated) names established by _REG_RE_NAME. 667 New programs should use the re_* names. */ 668 669 # define re_allocated allocated 670 # define re_buffer buffer 671 # define re_can_be_null can_be_null 672 # define re_fastmap fastmap 673 # define re_fastmap_accurate fastmap_accurate 674 # define re_newline_anchor newline_anchor 675 # define re_no_sub no_sub 676 # define re_not_bol not_bol 677 # define re_not_eol not_eol 678 # define re_regs_allocated regs_allocated 679 # define re_syntax syntax 680 # define re_translate translate 681 # define re_used used 682 683 /* Similarly for _REG_RM_NAME. */ 684 685 # define rm_end end 686 # define rm_num_regs num_regs 687 # define rm_start start 688 689 /* Undef RE_DUP_MAX first, in case the user has already included a 690 <limits.h> with an incompatible definition. 691 692 On GNU systems, the most common spelling for RE_DUP_MAX's value in 693 <limits.h> is (0x7ffff), so define RE_DUP_MAX to that, not to 694 REG_DUP_MAX. This avoid some duplicate-macro-definition warnings 695 with programs that include <limits.h> after this file. 696 697 New programs should not assume that regex.h defines RE_DUP_MAX; to 698 get the value of RE_DUP_MAX, they should instead include <limits.h> 699 and possibly invoke the sysconf function. */ 700 701 # undef RE_DUP_MAX 702 # define RE_DUP_MAX (0x7fff) 703 704 /* Define the following symbols for backward source compatibility. 705 These symbols violate the POSIX name space rules, and new programs 706 should avoid them. */ 707 708 # define REGS_FIXED REG_FIXED 709 # define REGS_REALLOCATE REG_REALLOCATE 710 # define REGS_UNALLOCATED REG_UNALLOCATED 711 # define RE_BACKSLASH_ESCAPE_IN_LISTS REG_BACKSLASH_ESCAPE_IN_LISTS 712 # define RE_BK_PLUS_QM REG_BK_PLUS_QM 713 # define RE_CARET_ANCHORS_HERE REG_CARET_ANCHORS_HERE 714 # define RE_CHAR_CLASSES REG_CHAR_CLASSES 715 # define RE_CONTEXT_INDEP_ANCHORS REG_CONTEXT_INDEP_ANCHORS 716 # define RE_CONTEXT_INDEP_OPS REG_CONTEXT_INDEP_OPS 717 # define RE_CONTEXT_INVALID_DUP REG_CONTEXT_INVALID_DUP 718 # define RE_CONTEXT_INVALID_OPS REG_CONTEXT_INVALID_OPS 719 # define RE_DEBUG REG_DEBUG 720 # define RE_DOT_NEWLINE REG_DOT_NEWLINE 721 # define RE_DOT_NOT_NULL REG_DOT_NOT_NULL 722 # define RE_HAT_LISTS_NOT_NEWLINE REG_HAT_LISTS_NOT_NEWLINE 723 # define RE_ICASE REG_IGNORE_CASE /* avoid collision with REG_ICASE */ 724 # define RE_INTERVALS REG_INTERVALS 725 # define RE_INVALID_INTERVAL_ORD REG_INVALID_INTERVAL_ORD 726 # define RE_LIMITED_OPS REG_LIMITED_OPS 727 # define RE_NEWLINE_ALT REG_NEWLINE_ALT 728 # define RE_NO_BK_BRACES REG_NO_BK_BRACES 729 # define RE_NO_BK_PARENS REG_NO_BK_PARENS 730 # define RE_NO_BK_REFS REG_NO_BK_REFS 731 # define RE_NO_BK_VBAR REG_NO_BK_VBAR 732 # define RE_NO_EMPTY_RANGES REG_NO_EMPTY_RANGES 733 # define RE_NO_GNU_OPS REG_NO_GNU_OPS 734 # define RE_NO_POSIX_BACKTRACKING REG_NO_POSIX_BACKTRACKING 735 # define RE_NO_SUB REG_NO_SUB 736 # define RE_NREGS REG_NREGS 737 # define RE_SYNTAX_AWK REG_SYNTAX_AWK 738 # define RE_SYNTAX_ED REG_SYNTAX_ED 739 # define RE_SYNTAX_EGREP REG_SYNTAX_EGREP 740 # define RE_SYNTAX_EMACS REG_SYNTAX_EMACS 741 # define RE_SYNTAX_GNU_AWK REG_SYNTAX_GNU_AWK 742 # define RE_SYNTAX_GREP REG_SYNTAX_GREP 743 # define RE_SYNTAX_POSIX_AWK REG_SYNTAX_POSIX_AWK 744 # define RE_SYNTAX_POSIX_BASIC REG_SYNTAX_POSIX_BASIC 745 # define RE_SYNTAX_POSIX_EGREP REG_SYNTAX_POSIX_EGREP 746 # define RE_SYNTAX_POSIX_EXTENDED REG_SYNTAX_POSIX_EXTENDED 747 # define RE_SYNTAX_POSIX_MINIMAL_BASIC REG_SYNTAX_POSIX_MINIMAL_BASIC 748 # define RE_SYNTAX_POSIX_MINIMAL_EXTENDED REG_SYNTAX_POSIX_MINIMAL_EXTENDED 749 # define RE_SYNTAX_SED REG_SYNTAX_SED 750 # define RE_UNMATCHED_RIGHT_PAREN_ORD REG_UNMATCHED_RIGHT_PAREN_ORD 751 # ifndef RE_TRANSLATE_TYPE 752 # define RE_TRANSLATE_TYPE REG_TRANSLATE_TYPE 753 # endif 754 755 #endif /* defined _REGEX_SOURCE */ 756 757 #ifdef __cplusplus 758 } 759 #endif /* C++ */ 760 761 #endif /* regex.h */ 762 763 /* 764 Local variables: 765 make-backup-files: t 766 version-control: t 767 trim-versions-without-asking: nil 768 End: 769 */ 770