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 /* interface types */
 111 #define __REG_WIDE_T    Tcl_UniChar
 112 #define __REG_REGOFF_T  long    /* not really right, but good enough... */
 113 #define __REG_VOID_T    VOID
 114 #define __REG_CONST     CONST
 115 /* names and declarations */
 116 #define __REG_WIDE_COMPILE      TclReComp
 117 #define __REG_WIDE_EXEC         TclReExec
 118 #define __REG_NOFRONT           /* don't want regcomp() and regexec() */
 119 #define __REG_NOCHAR            /* or the char versions */
 120 #define regfree         TclReFree
 121 #define regerror        TclReError
 122 /* --- end --- */
 123
 124
 125 /*
 126  * interface types etc.
 127  */
 128
 129 /*
 130  * regoff_t has to be large enough to hold either off_t or ssize_t,
 131  * and must be signed; it's only a guess that long is suitable, so we
 132  * offer <sys/types.h> an override.
 133  */
 134 #ifdef __REG_REGOFF_T
 135 typedef __REG_REGOFF_T regoff_t;
 136 #else
 137 typedef long regoff_t;
 138 #endif
 139
 140 /*
 141  * For benefit of old compilers, we offer <sys/types.h> the option of
 142  * overriding the `void' type used to declare nonexistent return types.
 143  */
 144 #ifdef __REG_VOID_T
 145 typedef __REG_VOID_T re_void;
 146 #else
 147 typedef void re_void;
 148 #endif
 149
 150 /*
 151  * Also for benefit of old compilers, <sys/types.h> can supply a macro
 152  * which expands to a substitute for `const'.
 153  */
 154 #ifndef __REG_CONST
 155 #define __REG_CONST     const
 156 #endif
 157
 158
 159
 160 /*
 161  * other interface types
 162  */
 163
 164 /* the biggie, a compiled RE (or rather, a front end to same) */
 165 typedef struct {
 166         int re_magic;           /* magic number */
 167         size_t re_nsub;         /* number of subexpressions */
 168         long re_info;           /* information about RE */
 169 #               define  REG_UBACKREF            000001
 170 #               define  REG_ULOOKAHEAD          000002
 171 #               define  REG_UBOUNDS             000004
 172 #               define  REG_UBRACES             000010
 173 #               define  REG_UBSALNUM            000020
 174 #               define  REG_UPBOTCH             000040
 175 #               define  REG_UBBS                000100
 176 #               define  REG_UNONPOSIX           000200
 177 #               define  REG_UUNSPEC             000400
 178 #               define  REG_UUNPORT             001000
 179 #               define  REG_ULOCALE             002000
 180 #               define  REG_UEMPTYMATCH         004000
 181 #               define  REG_UIMPOSSIBLE         010000
 182 #               define  REG_USHORTEST           020000
 183         int re_csize;           /* sizeof(character) */
 184         char *re_endp;          /* backward compatibility kludge */
 185         /* the rest is opaque pointers to hidden innards */
 186         char *re_guts;          /* `char *' is more portable than `void *' */
 187         char *re_fns;
 188 } regex_t;
 189
 190 /* result reporting (may acquire more fields later) */
 191 typedef struct {
 192         regoff_t rm_so;         /* start of substring */
 193         regoff_t rm_eo;         /* end of substring */
 194 } regmatch_t;
 195
 196 /* supplementary control and reporting */
 197 typedef struct {
 198         regmatch_t rm_extend;   /* see REG_EXPECT */
 199 } rm_detail_t;
 200
 201
 202
 203 /*
 204  * compilation
 205  ^ #ifndef __REG_NOCHAR
 206  ^ int re_comp(regex_t *, __REG_CONST char *, size_t, int);
 207  ^ #endif
 208  ^ #ifndef __REG_NOFRONT
 209  ^ int regcomp(regex_t *, __REG_CONST char *, int);
 210  ^ #endif
 211  ^ #ifdef __REG_WIDE_T
 212  ^ int __REG_WIDE_COMPILE(regex_t *, __REG_CONST __REG_WIDE_T *, size_t, int);
 213  ^ #endif
 214  */
 215 #define REG_BASIC       000000  /* BREs (convenience) */
 216 #define REG_EXTENDED    000001  /* EREs */
 217 #define REG_ADVF        000002  /* advanced features in EREs */
 218 #define REG_ADVANCED    000003  /* AREs (which are also EREs) */
 219 #define REG_QUOTE       000004  /* no special characters, none */
 220 #define REG_NOSPEC      REG_QUOTE       /* historical synonym */
 221 #define REG_ICASE       000010  /* ignore case */
 222 #define REG_NOSUB       000020  /* don't care about subexpressions */
 223 #define REG_EXPANDED    000040  /* expanded format, white space & comments */
 224 #define REG_NLSTOP      000100  /* \n doesn't match . or [^ ] */
 225 #define REG_NLANCH      000200  /* ^ matches after \n, $ before */
 226 #define REG_NEWLINE     000300  /* newlines are line terminators */
 227 #define REG_PEND        000400  /* ugh -- backward-compatibility hack */
 228 #define REG_EXPECT      001000  /* report details on partial/limited matches */
 229 #define REG_BOSONLY     002000  /* temporary kludge for BOS-only matches */
 230 #define REG_DUMP        004000  /* none of your business :-) */
 231 #define REG_FAKE        010000  /* none of your business :-) */
 232 #define REG_PROGRESS    020000  /* none of your business :-) */
 233
 234
 235
 236 /*
 237  * execution
 238  ^ #ifndef __REG_NOCHAR
 239  ^ int re_exec(regex_t *, __REG_CONST char *, size_t,
 240  ^                              rm_detail_t *, size_t, regmatch_t [], int);
 241  ^ #endif
 242  ^ #ifndef __REG_NOFRONT
 243  ^ int regexec(regex_t *, __REG_CONST char *, size_t, regmatch_t [], int);
 244  ^ #endif
 245  ^ #ifdef __REG_WIDE_T
 246  ^ int __REG_WIDE_EXEC(regex_t *, __REG_CONST __REG_WIDE_T *, size_t,
 247  ^                              rm_detail_t *, size_t, regmatch_t [], int);
 248  ^ #endif
 249  */
 250 #define REG_NOTBOL      0001    /* BOS is not BOL */
 251 #define REG_NOTEOL      0002    /* EOS is not EOL */
 252 #define REG_STARTEND    0004    /* backward compatibility kludge */
 253 #define REG_FTRACE      0010    /* none of your business */
 254 #define REG_MTRACE      0020    /* none of your business */
 255 #define REG_SMALL       0040    /* none of your business */
 256
 257
 258
 259 /*
 260  * misc generics (may be more functions here eventually)
 261  ^ re_void regfree(regex_t *);
 262  */
 263
 264
 265
 266 /*
 267  * error reporting
 268  * Be careful if modifying the list of error codes -- the table used by
 269  * regerror() is generated automatically from this file!
 270  *
 271  * Note that there is no wide-char variant of regerror at this time; what
 272  * kind of character is used for error reports is independent of what kind
 273  * is used in matching.
 274  *
 275  ^ extern size_t regerror(int, __REG_CONST regex_t *, char *, size_t);
 276  */
 277 #define REG_OKAY         0      /* no errors detected */
 278 #define REG_NOMATCH      1      /* failed to match */
 279 #define REG_BADPAT       2      /* invalid regexp */
 280 #define REG_ECOLLATE     3      /* invalid collating element */
 281 #define REG_ECTYPE       4      /* invalid character class */
 282 #define REG_EESCAPE      5      /* invalid escape \ sequence */
 283 #define REG_ESUBREG      6      /* invalid backreference number */
 284 #define REG_EBRACK       7      /* brackets [] not balanced */
 285 #define REG_EPAREN       8      /* parentheses () not balanced */
 286 #define REG_EBRACE       9      /* braces {} not balanced */
 287 #define REG_BADBR       10      /* invalid repetition count(s) */
 288 #define REG_ERANGE      11      /* invalid character range */
 289 #define REG_ESPACE      12      /* out of memory */
 290 #define REG_BADRPT      13      /* quantifier operand invalid */
 291 #define REG_ASSERT      15      /* "can't happen" -- you found a bug */
 292 #define REG_INVARG      16      /* invalid argument to regex function */
 293 #define REG_MIXED       17      /* character widths of regex and string differ */
 294 #define REG_BADOPT      18      /* invalid embedded option */
 295 /* two specials for debugging and testing */
 296 #define REG_ATOI        101     /* convert error-code name to number */
 297 #define REG_ITOA        102     /* convert error-code number to name */
 298
 299
 300
 301 /*
 302  * the prototypes, as possibly munched by regfwd
 303  */
 304 /* =====^!^===== begin forwards =====^!^===== */
 305 /* automatically gathered by fwd; do not hand-edit */
 306 /* === regproto.h === */
 307 #ifndef __REG_NOCHAR
 308 int re_comp _ANSI_ARGS_((regex_t *, __REG_CONST char *, size_t, int));
 309 #endif
 310 #ifndef __REG_NOFRONT
 311 int regcomp _ANSI_ARGS_((regex_t *, __REG_CONST char *, int));
 312 #endif
 313 #ifdef __REG_WIDE_T
 314 int __REG_WIDE_COMPILE _ANSI_ARGS_((regex_t *, __REG_CONST __REG_WIDE_T *, size_t, int));
 315 #endif
 316 #ifndef __REG_NOCHAR
 317 int re_exec _ANSI_ARGS_((regex_t *, __REG_CONST char *, size_t, rm_detail_t *, size_t, regmatch_t [], int));
 318 #endif
 319 #ifndef __REG_NOFRONT
 320 int regexec _ANSI_ARGS_((regex_t *, __REG_CONST char *, size_t, regmatch_t [], int));
 321 #endif
 322 #ifdef __REG_WIDE_T
 323 int __REG_WIDE_EXEC _ANSI_ARGS_((regex_t *, __REG_CONST __REG_WIDE_T *, size_t, rm_detail_t *, size_t, regmatch_t [], int));
 324 #endif
 325 re_void regfree _ANSI_ARGS_((regex_t *));
 326 extern size_t regerror _ANSI_ARGS_((int, __REG_CONST regex_t *, char *, size_t));
 327 /* automatically gathered by fwd; do not hand-edit */
 328 /* =====^!^===== end forwards =====^!^===== */
 329
 330
 331
 332 /*
 333  * more C++ voodoo
 334  */
 335 #ifdef __cplusplus
 336 }
 337 #endif
 338
 339
 340
 341 #endif