regex.h revision 8858
1/* Definitions for data structures and routines for the regular 2 expression library, version 0.12. 3 4 Copyright (C) 1985, 1989, 1990, 1991, 1992, 1993 Free Software Foundation, Inc. 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, or (at your option) 9 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#ifndef __REGEXP_LIBRARY_H__ 21#define __REGEXP_LIBRARY_H__ 22 23/* POSIX says that <sys/types.h> must be included (by the caller) before 24 <regex.h>. */ 25 26#ifdef VMS 27/* VMS doesn't have `size_t' in <sys/types.h>, even though POSIX says it 28 should be there. */ 29#include <stddef.h> 30#endif 31 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... */ 37typedef long s_reg_t; 38typedef unsigned long 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. */ 45typedef unsigned long reg_syntax_t; 46 47/* If this bit is not set, then \ inside a bracket expression is literal. 48 If set, then such a \ quotes the following character. */ 49#define RE_BACKSLASH_ESCAPE_IN_LISTS (1L) 50 51/* If this bit is not set, then + and ? are operators, and \+ and \? are 52 literals. 53 If set, then \+ and \? are operators and + and ? are literals. */ 54#define RE_BK_PLUS_QM (RE_BACKSLASH_ESCAPE_IN_LISTS << 1) 55 56/* If this bit is set, then character classes are supported. They are: 57 [:alpha:], [:upper:], [:lower:], [:digit:], [:alnum:], [:xdigit:], 58 [:space:], [:print:], [:punct:], [:graph:], and [:cntrl:]. 59 If not set, then character classes are not supported. */ 60#define RE_CHAR_CLASSES (RE_BK_PLUS_QM << 1) 61 62/* If this bit is set, then ^ and $ are always anchors (outside bracket 63 expressions, of course). 64 If this bit is not set, then it depends: 65 ^ is an anchor if it is at the beginning of a regular 66 expression or after an open-group or an alternation operator; 67 $ is an anchor if it is at the end of a regular expression, or 68 before a close-group or an alternation operator. 69 70 This bit could be (re)combined with RE_CONTEXT_INDEP_OPS, because 71 POSIX draft 11.2 says that * etc. in leading positions is undefined. 72 We already implemented a previous draft which made those constructs 73 invalid, though, so we haven't changed the code back. */ 74#define RE_CONTEXT_INDEP_ANCHORS (RE_CHAR_CLASSES << 1) 75 76/* If this bit is set, then special characters are always special 77 regardless of where they are in the pattern. 78 If this bit is not set, then special characters are special only in 79 some contexts; otherwise they are ordinary. Specifically, 80 * + ? and intervals are only special when not after the beginning, 81 open-group, or alternation operator. */ 82#define RE_CONTEXT_INDEP_OPS (RE_CONTEXT_INDEP_ANCHORS << 1) 83 84/* If this bit is set, then *, +, ?, and { cannot be first in an re or 85 immediately after an alternation or begin-group operator. */ 86#define RE_CONTEXT_INVALID_OPS (RE_CONTEXT_INDEP_OPS << 1) 87 88/* If this bit is set, then . matches newline. 89 If not set, then it doesn't. */ 90#define RE_DOT_NEWLINE (RE_CONTEXT_INVALID_OPS << 1) 91 92/* If this bit is set, then . doesn't match NUL. 93 If not set, then it does. */ 94#define RE_DOT_NOT_NULL (RE_DOT_NEWLINE << 1) 95 96/* If this bit is set, nonmatching lists [^...] do not match newline. 97 If not set, they do. */ 98#define RE_HAT_LISTS_NOT_NEWLINE (RE_DOT_NOT_NULL << 1) 99 100/* If this bit is set, either \{...\} or {...} defines an 101 interval, depending on RE_NO_BK_BRACES. 102 If not set, \{, \}, {, and } are literals. */ 103#define RE_INTERVALS (RE_HAT_LISTS_NOT_NEWLINE << 1) 104 105/* If this bit is set, +, ? and | aren't recognized as operators. 106 If not set, they are. */ 107#define RE_LIMITED_OPS (RE_INTERVALS << 1) 108 109/* If this bit is set, newline is an alternation operator. 110 If not set, newline is literal. */ 111#define RE_NEWLINE_ALT (RE_LIMITED_OPS << 1) 112 113/* If this bit is set, then `{...}' defines an interval, and \{ and \} 114 are literals. 115 If not set, then `\{...\}' defines an interval. */ 116#define RE_NO_BK_BRACES (RE_NEWLINE_ALT << 1) 117 118/* If this bit is set, (...) defines a group, and \( and \) are literals. 119 If not set, \(...\) defines a group, and ( and ) are literals. */ 120#define RE_NO_BK_PARENS (RE_NO_BK_BRACES << 1) 121 122/* If this bit is set, then \<digit> matches <digit>. 123 If not set, then \<digit> is a back-reference. */ 124#define RE_NO_BK_REFS (RE_NO_BK_PARENS << 1) 125 126/* If this bit is set, then | is an alternation operator, and \| is literal. 127 If not set, then \| is an alternation operator, and | is literal. */ 128#define RE_NO_BK_VBAR (RE_NO_BK_REFS << 1) 129 130/* If this bit is set, then an ending range point collating higher 131 than the starting range point, as in [z-a], is invalid. 132 If not set, then when ending range point collates higher than the 133 starting range point, the range is ignored. */ 134#define RE_NO_EMPTY_RANGES (RE_NO_BK_VBAR << 1) 135 136/* If this bit is set, then an unmatched ) is ordinary. 137 If not set, then an unmatched ) is invalid. */ 138#define RE_UNMATCHED_RIGHT_PAREN_ORD (RE_NO_EMPTY_RANGES << 1) 139 140/* If this bit is set, do not process the GNU regex operators. 141 IF not set, then the GNU regex operators are recognized. */ 142#define RE_NO_GNU_OPS (RE_UNMATCHED_RIGHT_PAREN_ORD << 1) 143 144/* This global variable defines the particular regexp syntax to use (for 145 some interfaces). When a regexp is compiled, the syntax used is 146 stored in the pattern buffer, so changing this does not affect 147 already-compiled regexps. */ 148extern reg_syntax_t re_syntax_options; 149 150/* Define combinations of the above bits for the standard possibilities. 151 (The [[[ comments delimit what gets put into the Texinfo file, so 152 don't delete them!) */ 153/* [[[begin syntaxes]]] */ 154#define RE_SYNTAX_EMACS 0 155 156#define RE_SYNTAX_AWK \ 157 (RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DOT_NOT_NULL \ 158 | RE_NO_BK_PARENS | RE_NO_BK_REFS \ 159 | RE_NO_BK_VBAR | RE_NO_EMPTY_RANGES \ 160 | RE_UNMATCHED_RIGHT_PAREN_ORD | RE_NO_GNU_OPS) 161 162#define RE_SYNTAX_GNU_AWK \ 163 (RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS) 164 165#define RE_SYNTAX_POSIX_AWK \ 166 (RE_SYNTAX_GNU_AWK | RE_NO_GNU_OPS) 167 168#define RE_SYNTAX_GREP \ 169 (RE_BK_PLUS_QM | RE_CHAR_CLASSES \ 170 | RE_HAT_LISTS_NOT_NEWLINE | RE_INTERVALS \ 171 | RE_NEWLINE_ALT) 172 173#define RE_SYNTAX_EGREP \ 174 (RE_CHAR_CLASSES | RE_CONTEXT_INDEP_ANCHORS \ 175 | RE_CONTEXT_INDEP_OPS | RE_HAT_LISTS_NOT_NEWLINE \ 176 | RE_NEWLINE_ALT | RE_NO_BK_PARENS \ 177 | RE_NO_BK_VBAR) 178 179#define RE_SYNTAX_POSIX_EGREP \ 180 (RE_SYNTAX_EGREP | RE_INTERVALS | RE_NO_BK_BRACES) 181 182/* P1003.2/D11.2, section 4.20.7.1, lines 5078ff. */ 183#define RE_SYNTAX_ED RE_SYNTAX_POSIX_BASIC 184 185#define RE_SYNTAX_SED RE_SYNTAX_POSIX_BASIC 186 187/* Syntax bits common to both basic and extended POSIX regex syntax. */ 188#define _RE_SYNTAX_POSIX_COMMON \ 189 (RE_CHAR_CLASSES | RE_DOT_NEWLINE | RE_DOT_NOT_NULL \ 190 | RE_INTERVALS | RE_NO_EMPTY_RANGES) 191 192#define RE_SYNTAX_POSIX_BASIC \ 193 (_RE_SYNTAX_POSIX_COMMON | RE_BK_PLUS_QM) 194 195/* Differs from ..._POSIX_BASIC only in that RE_BK_PLUS_QM becomes 196 RE_LIMITED_OPS, i.e., \? \+ \| are not recognized. Actually, this 197 isn't minimal, since other operators, such as \`, aren't disabled. */ 198#define RE_SYNTAX_POSIX_MINIMAL_BASIC \ 199 (_RE_SYNTAX_POSIX_COMMON | RE_LIMITED_OPS) 200 201#define RE_SYNTAX_POSIX_EXTENDED \ 202 (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \ 203 | RE_CONTEXT_INDEP_OPS | RE_NO_BK_BRACES \ 204 | RE_NO_BK_PARENS | RE_NO_BK_VBAR \ 205 | RE_UNMATCHED_RIGHT_PAREN_ORD) 206 207/* Differs from ..._POSIX_EXTENDED in that RE_CONTEXT_INVALID_OPS 208 replaces RE_CONTEXT_INDEP_OPS and RE_NO_BK_REFS is added. */ 209#define RE_SYNTAX_POSIX_MINIMAL_EXTENDED \ 210 (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \ 211 | RE_CONTEXT_INVALID_OPS | RE_NO_BK_BRACES \ 212 | RE_NO_BK_PARENS | RE_NO_BK_REFS \ 213 | RE_NO_BK_VBAR | RE_UNMATCHED_RIGHT_PAREN_ORD) 214/* [[[end syntaxes]]] */ 215 216/* Maximum number of duplicates an interval can allow. Some systems 217 (erroneously) define this in other header files, but we want our 218 value, so remove any previous define. */ 219#ifdef RE_DUP_MAX 220#undef RE_DUP_MAX 221#endif 222/* if sizeof(int) == 2, then ((1 << 15) - 1) overflows */ 223#define RE_DUP_MAX (0x7fff) 224 225 226/* POSIX `cflags' bits (i.e., information for `regcomp'). */ 227 228/* If this bit is set, then use extended regular expression syntax. 229 If not set, then use basic regular expression syntax. */ 230#define REG_EXTENDED 1 231 232/* If this bit is set, then ignore case when matching. 233 If not set, then case is significant. */ 234#define REG_ICASE (REG_EXTENDED << 1) 235 236/* If this bit is set, then anchors do not match at newline 237 characters in the string. 238 If not set, then anchors do match at newlines. */ 239#define REG_NEWLINE (REG_ICASE << 1) 240 241/* If this bit is set, then report only success or fail in regexec. 242 If not set, then returns differ between not matching and errors. */ 243#define REG_NOSUB (REG_NEWLINE << 1) 244 245 246/* POSIX `eflags' bits (i.e., information for regexec). */ 247 248/* If this bit is set, then the beginning-of-line operator doesn't match 249 the beginning of the string (presumably because it's not the 250 beginning of a line). 251 If not set, then the beginning-of-line operator does match the 252 beginning of the string. */ 253#define REG_NOTBOL 1 254 255/* Like REG_NOTBOL, except for the end-of-line. */ 256#define REG_NOTEOL (1 << 1) 257 258 259/* If any error codes are removed, changed, or added, update the 260 `re_error_msg' table in regex.c. */ 261typedef enum 262{ 263 REG_NOERROR = 0, /* Success. */ 264 REG_NOMATCH, /* Didn't find a match (for regexec). */ 265 266 /* POSIX regcomp return error codes. (In the order listed in the 267 standard.) */ 268 REG_BADPAT, /* Invalid pattern. */ 269 REG_ECOLLATE, /* Not implemented. */ 270 REG_ECTYPE, /* Invalid character class name. */ 271 REG_EESCAPE, /* Trailing backslash. */ 272 REG_ESUBREG, /* Invalid back reference. */ 273 REG_EBRACK, /* Unmatched left bracket. */ 274 REG_EPAREN, /* Parenthesis imbalance. */ 275 REG_EBRACE, /* Unmatched \{. */ 276 REG_BADBR, /* Invalid contents of \{\}. */ 277 REG_ERANGE, /* Invalid range end. */ 278 REG_ESPACE, /* Ran out of memory. */ 279 REG_BADRPT, /* No preceding re for repetition op. */ 280 281 /* Error codes we've added. */ 282 REG_EEND, /* Premature end. */ 283 REG_ESIZE, /* Compiled pattern bigger than 2^16 bytes. */ 284 REG_ERPAREN /* Unmatched ) or \); not returned from regcomp. */ 285} reg_errcode_t; 286 287/* This data structure represents a compiled pattern. Before calling 288 the pattern compiler, the fields `buffer', `allocated', `fastmap', 289 `translate', and `no_sub' can be set. After the pattern has been 290 compiled, the `re_nsub' field is available. All other fields are 291 private to the regex routines. */ 292 293struct re_pattern_buffer 294{ 295/* [[[begin pattern_buffer]]] */ 296 /* Space that holds the compiled pattern. It is declared as 297 `unsigned char *' because its elements are 298 sometimes used as array indexes. */ 299 unsigned char *buffer; 300 301 /* Number of bytes to which `buffer' points. */ 302 unsigned long allocated; 303 304 /* Number of bytes actually used in `buffer'. */ 305 unsigned long used; 306 307 /* Syntax setting with which the pattern was compiled. */ 308 reg_syntax_t syntax; 309 310 /* Pointer to a fastmap, if any, otherwise zero. re_search uses 311 the fastmap, if there is one, to skip over impossible 312 starting points for matches. */ 313 char *fastmap; 314 315 /* Either a translate table to apply to all characters before 316 comparing them, or zero for no translation. The translation 317 is applied to a pattern when it is compiled and to a string 318 when it is matched. */ 319 char *translate; 320 321 /* Number of subexpressions found by the compiler. */ 322 size_t re_nsub; 323 324 /* Zero if this pattern cannot match the empty string, one else. 325 Well, in truth it's used only in `re_search_2', to see 326 whether or not we should use the fastmap, so we don't set 327 this absolutely perfectly; see `re_compile_fastmap' (the 328 `duplicate' case). */ 329 unsigned can_be_null : 1; 330 331 /* If REGS_UNALLOCATED, allocate space in the `regs' structure 332 for `max (RE_NREGS, re_nsub + 1)' groups. 333 If REGS_REALLOCATE, reallocate space if necessary. 334 If REGS_FIXED, use what's there. */ 335#define REGS_UNALLOCATED 0 336#define REGS_REALLOCATE 1 337#define REGS_FIXED 2 338 unsigned regs_allocated : 2; 339 340 /* Set to zero when `regex_compile' compiles a pattern; set to one 341 by `re_compile_fastmap' if it updates the fastmap. */ 342 unsigned fastmap_accurate : 1; 343 344 /* If set, `re_match_2' does not return information about 345 subexpressions. */ 346 unsigned no_sub : 1; 347 348 /* If set, a beginning-of-line anchor doesn't match at the 349 beginning of the string. */ 350 unsigned not_bol : 1; 351 352 /* Similarly for an end-of-line anchor. */ 353 unsigned not_eol : 1; 354 355 /* If true, an anchor at a newline matches. */ 356 unsigned newline_anchor : 1; 357 358/* [[[end pattern_buffer]]] */ 359}; 360 361typedef struct re_pattern_buffer regex_t; 362 363 364/* search.c (search_buffer) in Emacs needs this one opcode value. It is 365 defined both in `regex.c' and here. */ 366#define RE_EXACTN_VALUE 1 367 368/* Type for byte offsets within the string. POSIX mandates this. */ 369typedef int regoff_t; 370 371 372/* This is the structure we store register match data in. See 373 regex.texinfo for a full description of what registers match. */ 374struct re_registers 375{ 376 unsigned num_regs; 377 regoff_t *start; 378 regoff_t *end; 379}; 380 381 382/* If `regs_allocated' is REGS_UNALLOCATED in the pattern buffer, 383 `re_match_2' returns information about at least this many registers 384 the first time a `regs' structure is passed. */ 385#ifndef RE_NREGS 386#define RE_NREGS 30 387#endif 388 389 390/* POSIX specification for registers. Aside from the different names than 391 `re_registers', POSIX uses an array of structures, instead of a 392 structure of arrays. */ 393typedef struct 394{ 395 regoff_t rm_so; /* Byte offset from string's start to substring's start. */ 396 regoff_t rm_eo; /* Byte offset from string's start to substring's end. */ 397} regmatch_t; 398 399/* Declarations for routines. */ 400 401/* To avoid duplicating every routine declaration -- once with a 402 prototype (if we are ANSI), and once without (if we aren't) -- we 403 use the following macro to declare argument types. This 404 unfortunately clutters up the declarations a bit, but I think it's 405 worth it. */ 406 407#ifdef __STDC__ 408 409#define _RE_ARGS(args) args 410 411#else /* not __STDC__ */ 412 413#define _RE_ARGS(args) () 414 415#endif /* not __STDC__ */ 416 417/* Sets the current default syntax to SYNTAX, and return the old syntax. 418 You can also simply assign to the `re_syntax_options' variable. */ 419extern reg_syntax_t re_set_syntax _RE_ARGS ((reg_syntax_t syntax)); 420 421/* Compile the regular expression PATTERN, with length LENGTH 422 and syntax given by the global `re_syntax_options', into the buffer 423 BUFFER. Return NULL if successful, and an error string if not. */ 424extern const char *re_compile_pattern 425 _RE_ARGS ((const char *pattern, size_t length, 426 struct re_pattern_buffer *buffer)); 427 428 429/* Compile a fastmap for the compiled pattern in BUFFER; used to 430 accelerate searches. Return 0 if successful and -2 if was an 431 internal error. */ 432extern int re_compile_fastmap _RE_ARGS ((struct re_pattern_buffer *buffer)); 433 434 435/* Search in the string STRING (with length LENGTH) for the pattern 436 compiled into BUFFER. Start searching at position START, for RANGE 437 characters. Return the starting position of the match, -1 for no 438 match, or -2 for an internal error. Also return register 439 information in REGS (if REGS and BUFFER->no_sub are nonzero). */ 440extern int re_search 441 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string, 442 int length, int start, int range, struct re_registers *regs)); 443 444 445/* Like `re_search', but search in the concatenation of STRING1 and 446 STRING2. Also, stop searching at index START + STOP. */ 447extern int re_search_2 448 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1, 449 int length1, const char *string2, int length2, 450 int start, int range, struct re_registers *regs, int stop)); 451 452 453/* Like `re_search', but return how many characters in STRING the regexp 454 in BUFFER matched, starting at position START. */ 455extern int re_match 456 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string, 457 int length, int start, struct re_registers *regs)); 458 459 460/* Relates to `re_match' as `re_search_2' relates to `re_search'. */ 461extern int re_match_2 462 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1, 463 int length1, const char *string2, int length2, 464 int start, struct re_registers *regs, int stop)); 465 466 467/* Set REGS to hold NUM_REGS registers, storing them in STARTS and 468 ENDS. Subsequent matches using BUFFER and REGS will use this memory 469 for recording register information. STARTS and ENDS must be 470 allocated with malloc, and must each be at least `NUM_REGS * sizeof 471 (regoff_t)' bytes long. 472 473 If NUM_REGS == 0, then subsequent matches should allocate their own 474 register data. 475 476 Unless this function is called, the first search or match using 477 PATTERN_BUFFER will allocate its own register data, without 478 freeing the old data. */ 479extern void re_set_registers 480 _RE_ARGS ((struct re_pattern_buffer *buffer, struct re_registers *regs, 481 unsigned num_regs, regoff_t *starts, regoff_t *ends)); 482 483/* 4.2 bsd compatibility. */ 484extern char *re_comp _RE_ARGS ((const char *)); 485extern int re_exec _RE_ARGS ((const char *)); 486 487/* POSIX compatibility. */ 488extern int regcomp _RE_ARGS ((regex_t *preg, const char *pattern, int cflags)); 489extern int regexec 490 _RE_ARGS ((const regex_t *preg, const char *string, size_t nmatch, 491 regmatch_t pmatch[], int eflags)); 492extern size_t regerror 493 _RE_ARGS ((int errcode, const regex_t *preg, char *errbuf, 494 size_t errbuf_size)); 495extern void regfree _RE_ARGS ((regex_t *preg)); 496 497#endif /* not __REGEXP_LIBRARY_H__ */ 498 499/* 500Local variables: 501make-backup-files: t 502version-control: t 503trim-versions-without-asking: nil 504End: 505*/ 506