src/regex/regex.h

   1 #ifndef _REGEX_H_
   2 #define _REGEX_H_       /* never again */
   3 /*
   4  * regular expressions
   5  *
   6  * Copyright (c) 1998, 1999 Henry Spencer.  All rights reserved.
   7  *
   8  * Development of this software was funded, in part, by Cray Research Inc.,
   9  * UUNET Communications Services Inc., Sun Microsystems Inc., and Scriptics
  10  * Corporation, none of whom are responsible for the results.  The author
  11  * thanks all of them.
  12  *
  13  * Redistribution and use in source and binary forms -- with or without
  14  * modification -- are permitted for any purpose, provided that
  15  * redistributions in source form retain this entire copyright notice and
  16  * indicate the origin and nature of any modifications.
  17  *
  18  * I'd appreciate being given credit for this package in the documentation
  19  * of software which uses it, but that is not a requirement.
  20  *
  21  * THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES,
  22  * INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY
  23  * AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL
  24  * HENRY SPENCER BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
  25  * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
  26  * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
  27  * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
  28  * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
  29  * OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
  30  * ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
  31  *
  32  *
  33  *
  34  * Prototypes etc. marked with "^" within comments get gathered up (and
  35  * possibly edited) by the regfwd program and inserted near the bottom of
  36  * this file.
  37  *
  38  * We offer the option of declaring one wide-character version of the
  39  * RE functions as well as the char versions.  To do that, define
  40  * __REG_WIDE_T to the type of wide characters (unfortunately, there
  41  * is no consensus that wchar_t is suitable) and __REG_WIDE_COMPILE and
  42  * __REG_WIDE_EXEC to the names to be used for the compile and execute
  43  * functions (suggestion:  re_Xcomp and re_Xexec, where X is a letter
  44  * suggestive of the wide type, e.g. re_ucomp and re_uexec for Unicode).
  45  * For cranky old compilers, it may be necessary to do something like:
  46  * #define      __REG_WIDE_COMPILE(a,b,c,d)     re_Xcomp(a,b,c,d)
  47  * #define      __REG_WIDE_EXEC(a,b,c,d,e,f,g)  re_Xexec(a,b,c,d,e,f,g)
  48  * rather than just #defining the names as parameterless macros.
  49  *
  50  * For some specialized purposes, it may be desirable to suppress the
  51  * declarations of the "front end" functions, regcomp() and regexec(),
  52  * or of the char versions of the compile and execute functions.  To
  53  * suppress the front-end functions, define __REG_NOFRONT.  To suppress
  54  * the char versions, define __REG_NOCHAR.
  55  *
  56  * The right place to do those defines (and some others you may want, see
  57  * below) would be <sys/types.h>.  If you don't have control of that file,
  58  * the right place to add your own defines to this file is marked below.
  59  * This is normally done automatically, by the makefile and regmkhdr, based
  60  * on the contents of regcustom.h.
  61  */
  62
  63
  64
  65 /*
  66  * voodoo for C++
  67  */
  68 #ifdef __cplusplus
  69 extern "C" {
  70 #endif
  71
  72
  73
  74 /*
  75  * Add your own defines, if needed, here.
  76  */
  77
  78
  79
  80 /*
  81  * Location where a chunk of regcustom.h is automatically spliced into
  82  * this file (working from its prototype, regproto.h).
  83  */
  84 /* --- begin --- */
  85 /* ensure certain things don't sneak in from system headers */
  86 #ifdef __REG_WIDE_T
  87 #undef __REG_WIDE_T
  88 #endif
  89 #ifdef __REG_WIDE_COMPILE
  90 #undef __REG_WIDE_COMPILE
  91 #endif
  92 #ifdef __REG_WIDE_EXEC
  93 #undef __REG_WIDE_EXEC
  94 #endif
  95 #ifdef __REG_REGOFF_T
  96 #undef __REG_REGOFF_T
  97 #endif
  98 #ifdef __REG_VOID_T
  99 #undef __REG_VOID_T
 100 #endif
 101 #ifdef __REG_CONST
 102 #undef __REG_CONST
 103 #endif
 104 #ifdef __REG_NOFRONT
 105 #undef __REG_NOFRONT
 106 #endif
 107 #ifdef __REG_NOCHAR
 108 #undef __REG_NOCHAR
 109 #endif
 110 #if wxUSE_UNICODE
 111 #   define  __REG_WIDE_T        wxChar
 112 #   define  __REG_WIDE_COMPILE  wx_re_comp
 113 #   define  __REG_WIDE_EXEC     wx_re_exec
 114 #   define  __REG_NOCHAR        /* don't want the char versions */
 115 #endif
 116 #define __REG_NOFRONT           /* don't want regcomp() and regexec() */
 117 #define _ANSI_ARGS_(x)          x
 118 /* --- end --- */
 119
 120
 121 /*
 122  * interface types etc.
 123  */
 124
 125 /*
 126  * regoff_t has to be large enough to hold either off_t or ssize_t,
 127  * and must be signed; it's only a guess that long is suitable, so we
 128  * offer <sys/types.h> an override.
 129  */
 130 #ifdef __REG_REGOFF_T
 131 typedef __REG_REGOFF_T regoff_t;
 132 #else
 133 typedef long regoff_t;
 134 #endif
 135
 136 /*
 137  * For benefit of old compilers, we offer <sys/types.h> the option of
 138  * overriding the `void' type used to declare nonexistent return types.
 139  */
 140 #ifdef __REG_VOID_T
 141 typedef __REG_VOID_T re_void;
 142 #else
 143 typedef void re_void;
 144 #endif
 145
 146 /*
 147  * Also for benefit of old compilers, <sys/types.h> can supply a macro
 148  * which expands to a substitute for `const'.
 149  */
 150 #ifndef __REG_CONST
 151 #define __REG_CONST     const
 152 #endif
 153
 154
 155
 156 /*
 157  * other interface types
 158  */
 159
 160 /* the biggie, a compiled RE (or rather, a front end to same) */
 161 typedef struct {
 162         int re_magic;           /* magic number */
 163         size_t re_nsub;         /* number of subexpressions */
 164         long re_info;           /* information about RE */
 165 #               define  REG_UBACKREF            000001
 166 #               define  REG_ULOOKAHEAD          000002
 167 #               define  REG_UBOUNDS             000004
 168 #               define  REG_UBRACES             000010
 169 #               define  REG_UBSALNUM            000020
 170 #               define  REG_UPBOTCH             000040
 171 #               define  REG_UBBS                000100
 172 #               define  REG_UNONPOSIX           000200
 173 #               define  REG_UUNSPEC             000400
 174 #               define  REG_UUNPORT             001000
 175 #               define  REG_ULOCALE             002000
 176 #               define  REG_UEMPTYMATCH         004000
 177 #               define  REG_UIMPOSSIBLE         010000
 178 #               define  REG_USHORTEST           020000
 179         int re_csize;           /* sizeof(character) */
 180         char *re_endp;          /* backward compatibility kludge */
 181         /* the rest is opaque pointers to hidden innards */
 182         char *re_guts;          /* `char *' is more portable than `void *' */
 183         char *re_fns;
 184 } regex_t;
 185
 186 /* result reporting (may acquire more fields later) */
 187 typedef struct {
 188         regoff_t rm_so;         /* start of substring */
 189         regoff_t rm_eo;         /* end of substring */
 190 } regmatch_t;
 191
 192 /* supplementary control and reporting */
 193 typedef struct {
 194         regmatch_t rm_extend;   /* see REG_EXPECT */
 195 } rm_detail_t;
 196
 197
 198
 199 /*
 200  * compilation
 201  ^ #ifndef __REG_NOCHAR
 202  ^ int re_comp(regex_t *, __REG_CONST char *, size_t, int);
 203  ^ #endif
 204  ^ #ifndef __REG_NOFRONT
 205  ^ int regcomp(regex_t *, __REG_CONST char *, int);
 206  ^ #endif
 207  ^ #ifdef __REG_WIDE_T
 208  ^ int __REG_WIDE_COMPILE(regex_t *, __REG_CONST __REG_WIDE_T *, size_t, int);
 209  ^ #endif
 210  */
 211 #define REG_BASIC       000000  /* BREs (convenience) */
 212 #define REG_EXTENDED    000001  /* EREs */
 213 #define REG_ADVF        000002  /* advanced features in EREs */
 214 #define REG_ADVANCED    000003  /* AREs (which are also EREs) */
 215 #define REG_QUOTE       000004  /* no special characters, none */
 216 #define REG_NOSPEC      REG_QUOTE       /* historical synonym */
 217 #define REG_ICASE       000010  /* ignore case */
 218 #define REG_NOSUB       000020  /* don't care about subexpressions */
 219 #define REG_EXPANDED    000040  /* expanded format, white space & comments */
 220 #define REG_NLSTOP      000100  /* \n doesn't match . or [^ ] */
 221 #define REG_NLANCH      000200  /* ^ matches after \n, $ before */
 222 #define REG_NEWLINE     000300  /* newlines are line terminators */
 223 #define REG_PEND        000400  /* ugh -- backward-compatibility hack */
 224 #define REG_EXPECT      001000  /* report details on partial/limited matches */
 225 #define REG_BOSONLY     002000  /* temporary kludge for BOS-only matches */
 226 #define REG_DUMP        004000  /* none of your business :-) */
 227 #define REG_FAKE        010000  /* none of your business :-) */
 228 #define REG_PROGRESS    020000  /* none of your business :-) */
 229
 230
 231
 232 /*
 233  * execution
 234  ^ #ifndef __REG_NOCHAR
 235  ^ int re_exec(regex_t *, __REG_CONST char *, size_t,
 236  ^                              rm_detail_t *, size_t, regmatch_t [], int);
 237  ^ #endif
 238  ^ #ifndef __REG_NOFRONT
 239  ^ int regexec(regex_t *, __REG_CONST char *, size_t, regmatch_t [], int);
 240  ^ #endif
 241  ^ #ifdef __REG_WIDE_T
 242  ^ int __REG_WIDE_EXEC(regex_t *, __REG_CONST __REG_WIDE_T *, size_t,
 243  ^                              rm_detail_t *, size_t, regmatch_t [], int);
 244  ^ #endif
 245  */
 246 #define REG_NOTBOL      0001    /* BOS is not BOL */
 247 #define REG_NOTEOL      0002    /* EOS is not EOL */
 248 #define REG_STARTEND    0004    /* backward compatibility kludge */
 249 #define REG_FTRACE      0010    /* none of your business */
 250 #define REG_MTRACE      0020    /* none of your business */
 251 #define REG_SMALL       0040    /* none of your business */
 252
 253
 254
 255 /*
 256  * misc generics (may be more functions here eventually)
 257  ^ re_void regfree(regex_t *);
 258  */
 259
 260
 261
 262 /*
 263  * error reporting
 264  * Be careful if modifying the list of error codes -- the table used by
 265  * regerror() is generated automatically from this file!
 266  *
 267  * Note that there is no wide-char variant of regerror at this time; what
 268  * kind of character is used for error reports is independent of what kind
 269  * is used in matching.
 270  *
 271  ^ extern size_t regerror(int, __REG_CONST regex_t *, char *, size_t);
 272  */
 273 #define REG_OKAY         0      /* no errors detected */
 274 #define REG_NOMATCH      1      /* failed to match */
 275 #define REG_BADPAT       2      /* invalid regexp */
 276 #define REG_ECOLLATE     3      /* invalid collating element */
 277 #define REG_ECTYPE       4      /* invalid character class */
 278 #define REG_EESCAPE      5      /* invalid escape \ sequence */
 279 #define REG_ESUBREG      6      /* invalid backreference number */
 280 #define REG_EBRACK       7      /* brackets [] not balanced */
 281 #define REG_EPAREN       8      /* parentheses () not balanced */
 282 #define REG_EBRACE       9      /* braces {} not balanced */
 283 #define REG_BADBR       10      /* invalid repetition count(s) */
 284 #define REG_ERANGE      11      /* invalid character range */
 285 #define REG_ESPACE      12      /* out of memory */
 286 #define REG_BADRPT      13      /* quantifier operand invalid */
 287 #define REG_ASSERT      15      /* "can't happen" -- you found a bug */
 288 #define REG_INVARG      16      /* invalid argument to regex function */
 289 #define REG_MIXED       17      /* character widths of regex and string differ */
 290 #define REG_BADOPT      18      /* invalid embedded option */
 291 /* two specials for debugging and testing */
 292 #define REG_ATOI        101     /* convert error-code name to number */
 293 #define REG_ITOA        102     /* convert error-code number to name */
 294
 295
 296
 297 /*
 298  * the prototypes, as possibly munched by regfwd
 299  */
 300 /* =====^!^===== begin forwards =====^!^===== */
 301 /* automatically gathered by fwd; do not hand-edit */
 302 /* === regproto.h === */
 303 #ifndef __REG_NOCHAR
 304 int wx_re_comp _ANSI_ARGS_((regex_t *, __REG_CONST char *, size_t, int));
 305 #endif
 306 #ifndef __REG_NOFRONT
 307 int regcomp _ANSI_ARGS_((regex_t *, __REG_CONST char *, int));
 308 #endif
 309 #ifdef __REG_WIDE_T
 310 int __REG_WIDE_COMPILE _ANSI_ARGS_((regex_t *, __REG_CONST __REG_WIDE_T *, size_t, int));
 311 #endif
 312 #ifndef __REG_NOCHAR
 313 int wx_re_exec _ANSI_ARGS_((regex_t *, __REG_CONST char *, size_t, rm_detail_t *, size_t, regmatch_t [], int));
 314 #endif
 315 #ifndef __REG_NOFRONT
 316 int regexec _ANSI_ARGS_((regex_t *, __REG_CONST char *, size_t, regmatch_t [], int));
 317 #endif
 318 #ifdef __REG_WIDE_T
 319 int __REG_WIDE_EXEC _ANSI_ARGS_((regex_t *, __REG_CONST __REG_WIDE_T *, size_t, rm_detail_t *, size_t, regmatch_t [], int));
 320 #endif
 321 re_void regfree _ANSI_ARGS_((regex_t *));
 322 extern size_t regerror _ANSI_ARGS_((int, __REG_CONST regex_t *, char *, size_t));
 323 /* automatically gathered by fwd; do not hand-edit */
 324 /* =====^!^===== end forwards =====^!^===== */
 325
 326
 327
 328 /*
 329  * more C++ voodoo
 330  */
 331 #ifdef __cplusplus
 332 }
 333 #endif
 334
 335
 336
 337 #endif