1 /* Definitions for data structures and routines for the regular 2 expression library, version 0.12. 3 4 Copyright (C) 1985, 89, 90, 91, 92, 93, 95 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 Foundation, 18 Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, 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 #if !defined (_POSIX_C_SOURCE) && !defined (_POSIX_SOURCE) && defined (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 bits are used to determine the regexp syntax we 34 recognize. The set/not-set meanings are chosen so that Emacs syntax 35 remains the value 0. The bits are given in alphabetical order, and 36 the definitions shifted by one from the previous bit; thus, when we 37 add or remove a bit, only one other definition need change. */ 38 typedef unsigned reg_syntax_t; 39 40 /* If this bit is not set, then \ inside a bracket expression is literal. 41 If set, then such a \ quotes the following character. */ 42 #define RE_BACKSLASH_ESCAPE_IN_LISTS (1) 43 44 /* If this bit is not set, then + and ? are operators, and \+ and \? are 45 literals. 46 If set, then \+ and \? are operators and + and ? are literals. */ 47 #define RE_BK_PLUS_QM (RE_BACKSLASH_ESCAPE_IN_LISTS << 1) 48 49 /* If this bit is set, then character classes are supported. They are: 50 [:alpha:], [:upper:], [:lower:], [:digit:], [:alnum:], [:xdigit:], 51 [:space:], [:print:], [:punct:], [:graph:], and [:cntrl:]. 52 If not set, then character classes are not supported. */ 53 #define RE_CHAR_CLASSES (RE_BK_PLUS_QM << 1) 54 55 /* If this bit is set, then ^ and $ are always anchors (outside bracket 56 expressions, of course). 57 If this bit is not set, then it depends: 58 ^ is an anchor if it is at the beginning of a regular 59 expression or after an open-group or an alternation operator; 60 $ is an anchor if it is at the end of a regular expression, or 61 before a close-group or an alternation operator. 62 63 This bit could be (re)combined with RE_CONTEXT_INDEP_OPS, because 64 POSIX draft 11.2 says that * etc. in leading positions is undefined. 65 We already implemented a previous draft which made those constructs 66 invalid, though, so we haven't changed the code back. */ 67 #define RE_CONTEXT_INDEP_ANCHORS (RE_CHAR_CLASSES << 1) 68 69 /* If this bit is set, then special characters are always special 70 regardless of where they are in the pattern. 71 If this bit is not set, then special characters are special only in 72 some contexts; otherwise they are ordinary. Specifically, 73 * + ? and intervals are only special when not after the beginning, 74 open-group, or alternation operator. */ 75 #define RE_CONTEXT_INDEP_OPS (RE_CONTEXT_INDEP_ANCHORS << 1) 76 77 /* If this bit is set, then *, +, ?, and { cannot be first in an re or 78 immediately after an alternation or begin-group operator. */ 79 #define RE_CONTEXT_INVALID_OPS (RE_CONTEXT_INDEP_OPS << 1) 80 81 /* If this bit is set, then . matches newline. 82 If not set, then it doesn't. */ 83 #define RE_DOT_NEWLINE (RE_CONTEXT_INVALID_OPS << 1) 84 85 /* If this bit is set, then . doesn't match NUL. 86 If not set, then it does. */ 87 #define RE_DOT_NOT_NULL (RE_DOT_NEWLINE << 1) 88 89 /* If this bit is set, nonmatching lists [^...] do not match newline. 90 If not set, they do. */ 91 #define RE_HAT_LISTS_NOT_NEWLINE (RE_DOT_NOT_NULL << 1) 92 93 /* If this bit is set, either \{...\} or {...} defines an 94 interval, depending on RE_NO_BK_BRACES. 95 If not set, \{, \}, {, and } are literals. */ 96 #define RE_INTERVALS (RE_HAT_LISTS_NOT_NEWLINE << 1) 97 98 /* If this bit is set, +, ? and | aren't recognized as operators. 99 If not set, they are. */ 100 #define RE_LIMITED_OPS (RE_INTERVALS << 1) 101 102 /* If this bit is set, newline is an alternation operator. 103 If not set, newline is literal. */ 104 #define RE_NEWLINE_ALT (RE_LIMITED_OPS << 1) 105 106 /* If this bit is set, then `{...}' defines an interval, and \{ and \} 107 are literals. 108 If not set, then `\{...\}' defines an interval. */ 109 #define RE_NO_BK_BRACES (RE_NEWLINE_ALT << 1) 110 111 /* If this bit is set, (...) defines a group, and \( and \) are literals. 112 If not set, \(...\) defines a group, and ( and ) are literals. */ 113 #define RE_NO_BK_PARENS (RE_NO_BK_BRACES << 1) 114 115 /* If this bit is set, then \<digit> matches <digit>. 116 If not set, then \<digit> is a back-reference. */ 117 #define RE_NO_BK_REFS (RE_NO_BK_PARENS << 1) 118 119 /* If this bit is set, then | is an alternation operator, and \| is literal. 120 If not set, then \| is an alternation operator, and | is literal. */ 121 #define RE_NO_BK_VBAR (RE_NO_BK_REFS << 1) 122 123 /* If this bit is set, then an ending range point collating higher 124 than the starting range point, as in [z-a], is invalid. 125 If not set, then when ending range point collates higher than the 126 starting range point, the range is ignored. */ 127 #define RE_NO_EMPTY_RANGES (RE_NO_BK_VBAR << 1) 128 129 /* If this bit is set, then an unmatched ) is ordinary. 130 If not set, then an unmatched ) is invalid. */ 131 #define RE_UNMATCHED_RIGHT_PAREN_ORD (RE_NO_EMPTY_RANGES << 1) 132 133 /* If this bit is set, succeed as soon as we match the whole pattern, 134 without further backtracking. */ 135 #define RE_NO_POSIX_BACKTRACKING (RE_UNMATCHED_RIGHT_PAREN_ORD << 1) 136 137 /* This global variable defines the particular regexp syntax to use (for 138 some interfaces). When a regexp is compiled, the syntax used is 139 stored in the pattern buffer, so changing this does not affect 140 already-compiled regexps. */ 141 extern reg_syntax_t re_syntax_options; 142 143 /* Define combinations of the above bits for the standard possibilities. 144 (The [[[ comments delimit what gets put into the Texinfo file, so 145 don't delete them!) */ 146 /* [[[begin syntaxes]]] */ 147 #define RE_SYNTAX_EMACS 0 148 149 #define RE_SYNTAX_AWK \ 150 (RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DOT_NOT_NULL \ 151 | RE_NO_BK_PARENS | RE_NO_BK_REFS \ 152 | RE_NO_BK_VBAR | RE_NO_EMPTY_RANGES \ 153 | RE_UNMATCHED_RIGHT_PAREN_ORD) 154 155 #define RE_SYNTAX_POSIX_AWK \ 156 (RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS) 157 158 #define RE_SYNTAX_GREP \ 159 (RE_BK_PLUS_QM | RE_CHAR_CLASSES \ 160 | RE_HAT_LISTS_NOT_NEWLINE | RE_INTERVALS \ 161 | RE_NEWLINE_ALT) 162 163 #define RE_SYNTAX_EGREP \ 164 (RE_CHAR_CLASSES | RE_CONTEXT_INDEP_ANCHORS \ 165 | RE_CONTEXT_INDEP_OPS | RE_HAT_LISTS_NOT_NEWLINE \ 166 | RE_NEWLINE_ALT | RE_NO_BK_PARENS \ 167 | RE_NO_BK_VBAR) 168 169 #define RE_SYNTAX_POSIX_EGREP \ 170 (RE_SYNTAX_EGREP | RE_INTERVALS | RE_NO_BK_BRACES) 171 172 /* P1003.2/D11.2, section 4.20.7.1, lines 5078ff. */ 173 #define RE_SYNTAX_ED RE_SYNTAX_POSIX_BASIC 174 175 #define RE_SYNTAX_SED RE_SYNTAX_POSIX_BASIC 176 177 /* Syntax bits common to both basic and extended POSIX regex syntax. */ 178 #define _RE_SYNTAX_POSIX_COMMON \ 179 (RE_CHAR_CLASSES | RE_DOT_NEWLINE | RE_DOT_NOT_NULL \ 180 | RE_INTERVALS | RE_NO_EMPTY_RANGES) 181 182 #define RE_SYNTAX_POSIX_BASIC \ 183 (_RE_SYNTAX_POSIX_COMMON | RE_BK_PLUS_QM) 184 185 /* Differs from ..._POSIX_BASIC only in that RE_BK_PLUS_QM becomes 186 RE_LIMITED_OPS, i.e., \? \+ \| are not recognized. Actually, this 187 isn't minimal, since other operators, such as \`, aren't disabled. */ 188 #define RE_SYNTAX_POSIX_MINIMAL_BASIC \ 189 (_RE_SYNTAX_POSIX_COMMON | RE_LIMITED_OPS) 190 191 #define RE_SYNTAX_POSIX_EXTENDED \ 192 (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \ 193 | RE_CONTEXT_INDEP_OPS | RE_NO_BK_BRACES \ 194 | RE_NO_BK_PARENS | RE_NO_BK_VBAR \ 195 | RE_UNMATCHED_RIGHT_PAREN_ORD) 196 197 /* Differs from ..._POSIX_EXTENDED in that RE_CONTEXT_INVALID_OPS 198 replaces RE_CONTEXT_INDEP_OPS and RE_NO_BK_REFS is added. */ 199 #define RE_SYNTAX_POSIX_MINIMAL_EXTENDED \ 200 (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \ 201 | RE_CONTEXT_INVALID_OPS | RE_NO_BK_BRACES \ 202 | RE_NO_BK_PARENS | RE_NO_BK_REFS \ 203 | RE_NO_BK_VBAR | RE_UNMATCHED_RIGHT_PAREN_ORD) 204 /* [[[end syntaxes]]] */ 205 206 /* Maximum number of duplicates an interval can allow. Some systems 207 (erroneously) define this in other header files, but we want our 208 value, so remove any previous define. */ 209 #ifdef RE_DUP_MAX 210 #undef RE_DUP_MAX 211 #endif 212 #define RE_DUP_MAX ((1 << 15) - 1) 213 214 215 /* POSIX `cflags' bits (i.e., information for `regcomp'). */ 216 217 /* If this bit is set, then use extended regular expression syntax. 218 If not set, then use basic regular expression syntax. */ 219 #define REG_EXTENDED 1 220 221 /* If this bit is set, then ignore case when matching. 222 If not set, then case is significant. */ 223 #define REG_ICASE (REG_EXTENDED << 1) 224 225 /* If this bit is set, then anchors do not match at newline 226 characters in the string. 227 If not set, then anchors do match at newlines. */ 228 #define REG_NEWLINE (REG_ICASE << 1) 229 230 /* If this bit is set, then report only success or fail in regexec. 231 If not set, then returns differ between not matching and errors. */ 232 #define REG_NOSUB (REG_NEWLINE << 1) 233 234 235 /* POSIX `eflags' bits (i.e., information for regexec). */ 236 237 /* If this bit is set, then the beginning-of-line operator doesn't match 238 the beginning of the string (presumably because it's not the 239 beginning of a line). 240 If not set, then the beginning-of-line operator does match the 241 beginning of the string. */ 242 #define REG_NOTBOL 1 243 244 /* Like REG_NOTBOL, except for the end-of-line. */ 245 #define REG_NOTEOL (1 << 1) 246 247 248 /* If any error codes are removed, changed, or added, update the 249 `re_error_msg' table in regex.c. */ 250 typedef enum 251 { 252 REG_NOERROR = 0, /* Success. */ 253 REG_NOMATCH, /* Didn't find a match (for regexec). */ 254 255 /* POSIX regcomp return error codes. (In the order listed in the 256 standard.) */ 257 REG_BADPAT, /* Invalid pattern. */ 258 REG_ECOLLATE, /* Not implemented. */ 259 REG_ECTYPE, /* Invalid character class name. */ 260 REG_EESCAPE, /* Trailing backslash. */ 261 REG_ESUBREG, /* Invalid back reference. */ 262 REG_EBRACK, /* Unmatched left bracket. */ 263 REG_EPAREN, /* Parenthesis imbalance. */ 264 REG_EBRACE, /* Unmatched \{. */ 265 REG_BADBR, /* Invalid contents of \{\}. */ 266 REG_ERANGE, /* Invalid range end. */ 267 REG_ESPACE, /* Ran out of memory. */ 268 REG_BADRPT, /* No preceding re for repetition op. */ 269 270 /* Error codes we've added. */ 271 REG_EEND, /* Premature end. */ 272 REG_ESIZE, /* Compiled pattern bigger than 2^16 bytes. */ 273 REG_ERPAREN /* Unmatched ) or \); not returned from regcomp. */ 274 } reg_errcode_t; 275 276 /* This data structure represents a compiled pattern. Before calling 277 the pattern compiler, the fields `buffer', `allocated', `fastmap', 278 `translate', and `no_sub' can be set. After the pattern has been 279 compiled, the `re_nsub' field is available. All other fields are 280 private to the regex routines. */ 281 282 #ifndef RE_TRANSLATE_TYPE 283 #define RE_TRANSLATE_TYPE char * 284 #endif 285 286 struct re_pattern_buffer 287 { 288 /* [[[begin pattern_buffer]]] */ 289 /* Space that holds the compiled pattern. It is declared as 290 `unsigned char *' because its elements are 291 sometimes used as array indexes. */ 292 unsigned char *buffer; 293 294 /* Number of bytes to which `buffer' points. */ 295 unsigned long allocated; 296 297 /* Number of bytes actually used in `buffer'. */ 298 unsigned long used; 299 300 /* Syntax setting with which the pattern was compiled. */ 301 reg_syntax_t syntax; 302 303 /* Pointer to a fastmap, if any, otherwise zero. re_search uses 304 the fastmap, if there is one, to skip over impossible 305 starting points for matches. */ 306 char *fastmap; 307 308 /* Either a translate table to apply to all characters before 309 comparing them, or zero for no translation. The translation 310 is applied to a pattern when it is compiled and to a string 311 when it is matched. */ 312 RE_TRANSLATE_TYPE translate; 313 314 /* Number of subexpressions found by the compiler. */ 315 size_t re_nsub; 316 317 /* Zero if this pattern cannot match the empty string, one else. 318 Well, in truth it's used only in `re_search_2', to see 319 whether or not we should use the fastmap, so we don't set 320 this absolutely perfectly; see `re_compile_fastmap' (the 321 `duplicate' case). */ 322 unsigned can_be_null : 1; 323 324 /* If REGS_UNALLOCATED, allocate space in the `regs' structure 325 for `max (RE_NREGS, re_nsub + 1)' groups. 326 If REGS_REALLOCATE, reallocate space if necessary. 327 If REGS_FIXED, use what's there. */ 328 #define REGS_UNALLOCATED 0 329 #define REGS_REALLOCATE 1 330 #define REGS_FIXED 2 331 unsigned regs_allocated : 2; 332 333 /* Set to zero when `regex_compile' compiles a pattern; set to one 334 by `re_compile_fastmap' if it updates the fastmap. */ 335 unsigned fastmap_accurate : 1; 336 337 /* If set, `re_match_2' does not return information about 338 subexpressions. */ 339 unsigned no_sub : 1; 340 341 /* If set, a beginning-of-line anchor doesn't match at the 342 beginning of the string. */ 343 unsigned not_bol : 1; 344 345 /* Similarly for an end-of-line anchor. */ 346 unsigned not_eol : 1; 347 348 /* If true, an anchor at a newline matches. */ 349 unsigned newline_anchor : 1; 350 351 /* [[[end pattern_buffer]]] */ 352 }; 353 354 typedef struct re_pattern_buffer regex_t; 355 356 /* Type for byte offsets within the string. POSIX mandates this. */ 357 typedef int regoff_t; 358 359 360 /* This is the structure we store register match data in. See 361 regex.texinfo for a full description of what registers match. */ 362 struct re_registers 363 { 364 unsigned num_regs; 365 regoff_t *start; 366 regoff_t *end; 367 }; 368 369 370 /* If `regs_allocated' is REGS_UNALLOCATED in the pattern buffer, 371 `re_match_2' returns information about at least this many registers 372 the first time a `regs' structure is passed. */ 373 #ifndef RE_NREGS 374 #define RE_NREGS 30 375 #endif 376 377 378 /* POSIX specification for registers. Aside from the different names than 379 `re_registers', POSIX uses an array of structures, instead of a 380 structure of arrays. */ 381 typedef struct 382 { 383 regoff_t rm_so; /* Byte offset from string's start to substring's start. */ 384 regoff_t rm_eo; /* Byte offset from string's start to substring's end. */ 385 } regmatch_t; 386 387 /* Declarations for routines. */ 388 389 /* To avoid duplicating every routine declaration -- once with a 390 prototype (if we are ANSI), and once without (if we aren't) -- we 391 use the following macro to declare argument types. This 392 unfortunately clutters up the declarations a bit, but I think it's 393 worth it. */ 394 395 #if __STDC__ 396 397 #define _RE_ARGS(args) args 398 399 #else /* not __STDC__ */ 400 401 #define _RE_ARGS(args) () 402 403 #endif /* not __STDC__ */ 404 405 /* Sets the current default syntax to SYNTAX, and return the old syntax. 406 You can also simply assign to the `re_syntax_options' variable. */ 407 extern reg_syntax_t re_set_syntax _RE_ARGS ((reg_syntax_t syntax)); 408 409 /* Compile the regular expression PATTERN, with length LENGTH 410 and syntax given by the global `re_syntax_options', into the buffer 411 BUFFER. Return NULL if successful, and an error string if not. */ 412 extern const char *re_compile_pattern 413 _RE_ARGS ((const char *pattern, int length, 414 struct re_pattern_buffer *buffer)); 415 416 417 /* Compile a fastmap for the compiled pattern in BUFFER; used to 418 accelerate searches. Return 0 if successful and -2 if was an 419 internal error. */ 420 extern int re_compile_fastmap _RE_ARGS ((struct re_pattern_buffer *buffer)); 421 422 423 /* Search in the string STRING (with length LENGTH) for the pattern 424 compiled into BUFFER. Start searching at position START, for RANGE 425 characters. Return the starting position of the match, -1 for no 426 match, or -2 for an internal error. Also return register 427 information in REGS (if REGS and BUFFER->no_sub are nonzero). */ 428 extern int re_search 429 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string, 430 int length, int start, int range, struct re_registers *regs)); 431 432 433 /* Like `re_search', but search in the concatenation of STRING1 and 434 STRING2. Also, stop searching at index START + STOP. */ 435 extern int re_search_2 436 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1, 437 int length1, const char *string2, int length2, 438 int start, int range, struct re_registers *regs, int stop)); 439 440 441 /* Like `re_search', but return how many characters in STRING the regexp 442 in BUFFER matched, starting at position START. */ 443 extern int re_match 444 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string, 445 int length, int start, struct re_registers *regs)); 446 447 448 /* Relates to `re_match' as `re_search_2' relates to `re_search'. */ 449 extern int re_match_2 450 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1, 451 int length1, const char *string2, int length2, 452 int start, struct re_registers *regs, int stop)); 453 454 455 /* Set REGS to hold NUM_REGS registers, storing them in STARTS and 456 ENDS. Subsequent matches using BUFFER and REGS will use this memory 457 for recording register information. STARTS and ENDS must be 458 allocated with malloc, and must each be at least `NUM_REGS * sizeof 459 (regoff_t)' bytes long. 460 461 If NUM_REGS == 0, then subsequent matches should allocate their own 462 register data. 463 464 Unless this function is called, the first search or match using 465 PATTERN_BUFFER will allocate its own register data, without 466 freeing the old data. */ 467 extern void re_set_registers 468 _RE_ARGS ((struct re_pattern_buffer *buffer, struct re_registers *regs, 469 unsigned num_regs, regoff_t *starts, regoff_t *ends)); 470 471 #ifdef _REGEX_RE_COMP 472 /* 4.2 bsd compatibility. */ 473 extern char *re_comp _RE_ARGS ((const char *)); 474 extern int re_exec _RE_ARGS ((const char *)); 475 #endif 476 477 /* POSIX compatibility. */ 478 extern int regcomp _RE_ARGS ((regex_t *preg, const char *pattern, int cflags)); 479 extern int regexec 480 _RE_ARGS ((const regex_t *preg, const char *string, size_t nmatch, 481 regmatch_t pmatch[], int eflags)); 482 extern size_t regerror 483 _RE_ARGS ((int errcode, const regex_t *preg, char *errbuf, 484 size_t errbuf_size)); 485 extern void regfree _RE_ARGS ((regex_t *preg)); 486 487 #endif /* not __REGEXP_LIBRARY_H__ */ 488 489 /* 490 Local variables: 491 make-backup-files: t 492 version-control: t 493 trim-versions-without-asking: nil 494 End: 495 */ 496