Modules/pcre-int.h

   1 /*************************************************
   2 *      Perl-Compatible Regular Expressions       *
   3 *************************************************/
   4
   5
   6 #define PCRE_VERSION       "1.09 28-Apr-1998"
   7
   8
   9 /* This is a library of functions to support regular expressions whose syntax
  10 and semantics are as close as possible to those of the Perl 5 language. See
  11 the file Tech.Notes for some information on the internals.
  12
  13 Written by: Philip Hazel <ph10@cam.ac.uk>
  14
  15            Copyright (c) 1998 University of Cambridge
  16
  17 -----------------------------------------------------------------------------
  18 Permission is granted to anyone to use this software for any purpose on any
  19 computer system, and to redistribute it freely, subject to the following
  20 restrictions:
  21
  22 1. This software is distributed in the hope that it will be useful,
  23    but WITHOUT ANY WARRANTY; without even the implied warranty of
  24    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
  25
  26 2. The origin of this software must not be misrepresented, either by
  27    explicit claim or by omission.
  28
  29 3. Altered versions must be plainly marked as such, and must not be
  30    misrepresented as being the original software.
  31 -----------------------------------------------------------------------------
  32 */
  33
  34 /* This header contains definitions that are shared between the different
  35 modules, but which are not relevant to the outside. */
  36
  37
  38 /* To cope with SunOS4 and other systems that lack memmove() but have bcopy(),
  39 define a macro for memmove() if USE_BCOPY is defined. */
  40
  41 #ifdef USE_BCOPY
  42 #undef  memmove        /* some systems may have a macro */
  43 #define memmove(a, b, c) bcopy(b, a, c)
  44 #endif
  45
  46 /* Standard C headers plus the external interface definition */
  47
  48 #include <ctype.h>
  49 #include <limits.h>
  50 #include <setjmp.h>
  51 #include <stddef.h>
  52 #include <stdio.h>
  53 #include <stdlib.h>
  54 #include <string.h>
  55 #include "pcre.h"
  56
  57 /* In case there is no definition of offsetof() provided - though any proper
  58 Standard C system should have one. */
  59
  60 #ifndef offsetof
  61 #define offsetof(p_type,field) ((size_t)&(((p_type *)0)->field))
  62 #endif
  63
  64 /* Private options flags start at the most significant end of the two bytes.
  65 The public options defined in pcre.h start at the least significant end. Make
  66 sure they don't overlap! */
  67
  68 #define PCRE_FIRSTSET           0x8000  /* first_char is set */
  69 #define PCRE_STARTLINE          0x4000  /* start after \n for multiline */
  70 #define PCRE_COMPILED_CASELESS  0x2000  /* like it says */
  71
  72 /* Options for the "extra" block produced by pcre_study(). */
  73
  74 #define PCRE_STUDY_CASELESS 0x01     /* study was caseless */
  75 #define PCRE_STUDY_MAPPED   0x02     /* a map of starting chars exists */
  76
  77 /* Masks for identifying the public options: all permitted at compile time,
  78 only some permitted at run or study time. */
  79
  80 #ifdef FOR_PYTHON
  81 #define PUBLIC_OPTIONS \
  82   (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
  83    PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY| \
  84    PCRE_LOCALE)
  85 #else
  86 #define PUBLIC_OPTIONS \
  87   (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
  88    PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY)
  89 #endif
  90 #define PUBLIC_EXEC_OPTIONS \
  91   (PCRE_CASELESS|PCRE_ANCHORED|PCRE_MULTILINE|PCRE_NOTBOL|PCRE_NOTEOL| \
  92    PCRE_DOTALL|PCRE_DOLLAR_ENDONLY)
  93
  94 #define PUBLIC_STUDY_OPTIONS (PCRE_CASELESS)
  95
  96 /* Magic number to provide a small check against being handed junk. */
  97
  98 #define MAGIC_NUMBER  0x50435245   /* 'PCRE' */
  99
 100 /* Miscellaneous definitions */
 101
 102 typedef int BOOL;
 103
 104 #define FALSE   0
 105 #define TRUE    1
 106
 107 /* These are escaped items that aren't just an encoding of a particular data
 108 value such as \n. They must have non-zero values, as check_escape() returns
 109 their negation. Also, they must appear in the same order as in the opcode
 110 definitions below, up to ESC_Z. The final one must be ESC_REF as subsequent
 111 values are used for \1, \2, \3, etc. There is a test in the code for an escape
 112 greater than ESC_b and less than ESC_X to detect the types that may be
 113 repeated. If any new escapes are put in-between that don't consume a character,
 114 that code will have to change. */
 115
 116 enum { ESC_A = 1, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W, ESC_w,
 117
 118                     /* These are not Perl escapes, so can't appear in the */
 119        ESC_X,       /* simple table-lookup because they must be conditional */
 120                     /* on PCRE_EXTRA. */
 121        ESC_Z,
 122        ESC_REF };
 123
 124 /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets
 125 that extract substrings. Starting from 1 (i.e. after OP_END), the values up to
 126 OP_EOD must correspond in order to the list of escapes immediately above. */
 127
 128 enum {
 129   OP_END,            /* End of pattern */
 130
 131   /* Values corresponding to backslashed metacharacters */
 132
 133   OP_SOD,            /* Start of data: \A */
 134   OP_NOT_WORD_BOUNDARY,  /* \B */
 135   OP_WORD_BOUNDARY,      /* \b */
 136   OP_NOT_DIGIT,          /* \D */
 137   OP_DIGIT,              /* \d */
 138   OP_NOT_WHITESPACE,     /* \S */
 139   OP_WHITESPACE,         /* \s */
 140   OP_NOT_WORDCHAR,       /* \W */
 141   OP_WORDCHAR,           /* \w */
 142   OP_CUT,            /* The analogue of Prolog's "cut" operation (extension) */
 143   OP_EOD,            /* End of data: \Z. */
 144
 145   OP_NOT_WORD_BOUNDARY_L,  /* localized \B */
 146   OP_WORD_BOUNDARY_L,      /* localized \b */
 147   OP_NOT_WORDCHAR_L,       /* localized \W */
 148   OP_WORDCHAR_L,           /* localized \w */
 149
 150   OP_CIRC,           /* Start of line - varies with multiline switch */
 151   OP_DOLL,           /* End of line - varies with multiline switch */
 152   OP_ANY,            /* Match any character */
 153   OP_CHARS,          /* Match string of characters */
 154   OP_NOT,            /* Match anything but the following char */
 155
 156   OP_STAR,           /* The maximizing and minimizing versions of */
 157   OP_MINSTAR,        /* all these opcodes must come in pairs, with */
 158   OP_PLUS,           /* the minimizing one second. */
 159   OP_MINPLUS,        /* This first set applies to single characters */
 160   OP_QUERY,
 161   OP_MINQUERY,
 162   OP_UPTO,           /* From 0 to n matches */
 163   OP_MINUPTO,
 164   OP_EXACT,          /* Exactly n matches */
 165
 166   OP_NOTSTAR,        /* The maximizing and minimizing versions of */
 167   OP_NOTMINSTAR,     /* all these opcodes must come in pairs, with */
 168   OP_NOTPLUS,        /* the minimizing one second. */
 169   OP_NOTMINPLUS,     /* This first set applies to "not" single characters */
 170   OP_NOTQUERY,
 171   OP_NOTMINQUERY,
 172   OP_NOTUPTO,        /* From 0 to n matches */
 173   OP_NOTMINUPTO,
 174   OP_NOTEXACT,       /* Exactly n matches */
 175
 176   OP_TYPESTAR,       /* The maximizing and minimizing versions of */
 177   OP_TYPEMINSTAR,    /* all these opcodes must come in pairs, with */
 178   OP_TYPEPLUS,       /* the minimizing one second. These codes must */
 179   OP_TYPEMINPLUS,    /* be in exactly the same order as those above. */
 180   OP_TYPEQUERY,      /* This set applies to character types such as \d */
 181   OP_TYPEMINQUERY,
 182   OP_TYPEUPTO,       /* From 0 to n matches */
 183   OP_TYPEMINUPTO,
 184   OP_TYPEEXACT,      /* Exactly n matches */
 185
 186   OP_CRSTAR,         /* The maximizing and minimizing versions of */
 187   OP_CRMINSTAR,      /* all these opcodes must come in pairs, with */
 188   OP_CRPLUS,         /* the minimizing one second. These codes must */
 189   OP_CRMINPLUS,      /* be in exactly the same order as those above. */
 190   OP_CRQUERY,        /* These are for character classes and back refs */
 191   OP_CRMINQUERY,
 192   OP_CRRANGE,        /* These are different to the three seta above. */
 193   OP_CRMINRANGE,
 194
 195   OP_CLASS,          /* Match a character class */
 196   OP_NEGCLASS,       /* Match a character class, specified negatively */
 197   OP_CLASS_L,        /* Match a character class */
 198   OP_REF,            /* Match a back reference */
 199
 200   OP_ALT,            /* Start of alternation */
 201   OP_KET,            /* End of group that doesn't have an unbounded repeat */
 202   OP_KETRMAX,        /* These two must remain together and in this */
 203   OP_KETRMIN,        /* order. They are for groups the repeat for ever. */
 204
 205   OP_ASSERT,
 206   OP_ASSERT_NOT,
 207   OP_ONCE,           /* Once matched, don't back up into the subpattern */
 208
 209   OP_BRAZERO,        /* These two must remain together and in this */
 210   OP_BRAMINZERO,     /* order. */
 211
 212   OP_BRA             /* This and greater values are used for brackets that
 213                         extract substrings. */
 214 };
 215
 216 /* The highest extraction number. This is limited by the number of opcodes
 217 left after OP_BRA, i.e. 255 - OP_BRA. We actually set it somewhat lower. */
 218
 219 #define EXTRACT_MAX  99
 220
 221 /* The texts of compile-time error messages are defined as macros here so that
 222 they can be accessed by the POSIX wrapper and converted into error codes.  Yes,
 223 I could have used error codes in the first place, but didn't feel like changing
 224 just to accommodate the POSIX wrapper. */
 225
 226 #define ERR1  "\\ at end of pattern"
 227 #define ERR2  "\\c at end of pattern"
 228 #define ERR3  "unrecognized character follows \\"
 229 #define ERR4  "numbers out of order in {} quantifier"
 230 #define ERR5  "number too big in {} quantifier"
 231 #define ERR6  "missing terminating ] for character class"
 232 #define ERR7  "invalid escape sequence in character class"
 233 #define ERR8  "range out of order in character class"
 234 #define ERR9  "nothing to repeat"
 235 #define ERR10 "operand of unlimited repeat could match the empty string"
 236 #define ERR11 "internal error: unexpected repeat"
 237 #define ERR12 "unrecognized character after (?"
 238 #define ERR13 "too many capturing parenthesized sub-patterns"
 239 #define ERR14 "missing )"
 240 #define ERR15 "back reference to non-existent subpattern"
 241 #define ERR16 "erroffset passed as NULL"
 242 #define ERR17 "unknown option bit(s) set"
 243 #define ERR18 "missing ) after comment"
 244 #define ERR19 "too many sets of parentheses"
 245 #define ERR20 "regular expression too large"
 246 #define ERR21 "failed to get memory"
 247 #define ERR22 "unmatched brackets"
 248 #define ERR23 "internal error: code overflow"
 249
 250 /* All character handling must be done as unsigned characters. Otherwise there
 251 are problems with top-bit-set characters and functions such as isspace().
 252 However, we leave the interface to the outside world as char *, because that
 253 should make things easier for callers. We define a short type for unsigned char
 254 to save lots of typing. I tried "uchar", but it causes problems on Digital
 255 Unix, where it is defined in sys/types, so use "uschar" instead. */
 256
 257 typedef unsigned char uschar;
 258
 259 /* The real format of the start of the pcre block; the actual code vector
 260 runs on as long as necessary after the end. */
 261
 262 typedef struct real_pcre {
 263   unsigned int  magic_number;
 264   unsigned short int options;
 265   unsigned char top_bracket;
 266   unsigned char top_backref;
 267   unsigned char first_char;
 268   unsigned char code[1];
 269 } real_pcre;
 270
 271 /* The real format of the extra block returned by pcre_study(). */
 272
 273 typedef struct real_pcre_extra {
 274   unsigned char options;
 275   unsigned char start_bits[32];
 276 } real_pcre_extra;
 277
 278 /* Global tables from chartables.c */
 279
 280 extern uschar pcre_lcc[];
 281 extern uschar pcre_fcc[];
 282 extern uschar pcre_cbits[];
 283 extern uschar pcre_ctypes[];
 284
 285 /* Bit definitions for entries in pcre_ctypes[]. */
 286
 287 #define ctype_space   0x01
 288 #define ctype_letter  0x02
 289 #define ctype_digit   0x04
 290 #define ctype_xdigit  0x08
 291 #define ctype_word    0x10   /* alphameric or '_' */
 292 #define ctype_odigit  0x20   /* octal digit */
 293 #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */
 294
 295 /* Offsets for the bitmap tables */
 296
 297 #define cbit_digit    0
 298 #define cbit_letter  32
 299 #define cbit_word    64
 300 #define cbit_space   96
 301 #define cbit_length 128      /* Length of the cbits table */
 302
 303 /* End of internal.h */