Annotation of parser3/src/lib/pcre/internal.h, revision 1.1
1.1 ! paf 1: /*************************************************
! 2: * Perl-Compatible Regular Expressions *
! 3: *************************************************/
! 4:
! 5:
! 6: /* This is a library of functions to support regular expressions whose syntax
! 7: and semantics are as close as possible to those of the Perl 5 language. See
! 8: the file Tech.Notes for some information on the internals.
! 9:
! 10: Written by: Philip Hazel <ph10@cam.ac.uk>
! 11:
! 12: Copyright (c) 1997-1999 University of Cambridge
! 13:
! 14: -----------------------------------------------------------------------------
! 15: Permission is granted to anyone to use this software for any purpose on any
! 16: computer system, and to redistribute it freely, subject to the following
! 17: restrictions:
! 18:
! 19: 1. This software is distributed in the hope that it will be useful,
! 20: but WITHOUT ANY WARRANTY; without even the implied warranty of
! 21: MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
! 22:
! 23: 2. The origin of this software must not be misrepresented, either by
! 24: explicit claim or by omission.
! 25:
! 26: 3. Altered versions must be plainly marked as such, and must not be
! 27: misrepresented as being the original software.
! 28:
! 29: 4. If PCRE is embedded in any software that is released under the GNU
! 30: General Purpose Licence (GPL), then the terms of that licence shall
! 31: supersede any condition above with which it is incompatible.
! 32: -----------------------------------------------------------------------------
! 33: */
! 34:
! 35: /* This header contains definitions that are shared between the different
! 36: modules, but which are not relevant to the outside. */
! 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: PAF@design.ru removed this, for it were used only for stupid MAX_INT for
! 51: mimimum searches, changed that to some out-of-reasonable-text-length value
! 52: of 10M
! 53: */
! 54: #define PCRE_MAX_POS (10*0x400*0x400)
! 55:
! 56: #include <stddef.h>
! 57: #include <stdio.h>
! 58: #include <stdlib.h>
! 59: #include <string.h>
! 60: #include "pcre.h"
! 61:
! 62: /* In case there is no definition of offsetof() provided - though any proper
! 63: Standard C system should have one. */
! 64:
! 65: #ifndef offsetof
! 66: #define offsetof(p_type,field) ((size_t)&(((p_type *)0)->field))
! 67: #endif
! 68:
! 69: /* These are the public options that can change during matching. */
! 70:
! 71: #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
! 72:
! 73: /* Private options flags start at the most significant end of the four bytes,
! 74: but skip the top bit so we can use ints for convenience without getting tangled
! 75: with negative values. The public options defined in pcre.h start at the least
! 76: significant end. Make sure they don't overlap, though now that we have expanded
! 77: to four bytes there is plenty of space. */
! 78:
! 79: #define PCRE_FIRSTSET 0x40000000 /* first_char is set */
! 80: #define PCRE_REQCHSET 0x20000000 /* req_char is set */
! 81: #define PCRE_STARTLINE 0x10000000 /* start after \n for multiline */
! 82: #define PCRE_INGROUP 0x08000000 /* compiling inside a group */
! 83: #define PCRE_ICHANGED 0x04000000 /* i option changes within regex */
! 84:
! 85: /* Options for the "extra" block produced by pcre_study(). */
! 86:
! 87: #define PCRE_STUDY_MAPPED 0x01 /* a map of starting chars exists */
! 88:
! 89: /* Masks for identifying the public options which are permitted at compile
! 90: time, run time or study time, respectively. */
! 91:
! 92: #define PUBLIC_OPTIONS \
! 93: (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
! 94: PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY)
! 95:
! 96: #define PUBLIC_EXEC_OPTIONS \
! 97: (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY)
! 98:
! 99: #define PUBLIC_STUDY_OPTIONS 0 /* None defined */
! 100:
! 101: /* Magic number to provide a small check against being handed junk. */
! 102:
! 103: #define MAGIC_NUMBER 0x50435245UL /* 'PCRE' */
! 104:
! 105: /* Miscellaneous definitions */
! 106:
! 107: typedef int BOOL;
! 108:
! 109: #ifndef FALSE
! 110: #define FALSE 0
! 111: #endif
! 112: #ifndef TRUE
! 113: #define TRUE 1
! 114: #endif
! 115:
! 116: /* These are escaped items that aren't just an encoding of a particular data
! 117: value such as \n. They must have non-zero values, as check_escape() returns
! 118: their negation. Also, they must appear in the same order as in the opcode
! 119: definitions below, up to ESC_z. The final one must be ESC_REF as subsequent
! 120: values are used for \1, \2, \3, etc. There is a test in the code for an escape
! 121: greater than ESC_b and less than ESC_X to detect the types that may be
! 122: repeated. If any new escapes are put in-between that don't consume a character,
! 123: that code will have to change. */
! 124:
! 125: enum { ESC_A = 1, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W, ESC_w,
! 126: ESC_Z, ESC_z, ESC_REF };
! 127:
! 128: /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets
! 129: that extract substrings. Starting from 1 (i.e. after OP_END), the values up to
! 130: OP_EOD must correspond in order to the list of escapes immediately above. */
! 131:
! 132: enum {
! 133: OP_END, /* End of pattern */
! 134:
! 135: /* Values corresponding to backslashed metacharacters */
! 136:
! 137: OP_SOD, /* Start of data: \A */
! 138: OP_NOT_WORD_BOUNDARY, /* \B */
! 139: OP_WORD_BOUNDARY, /* \b */
! 140: OP_NOT_DIGIT, /* \D */
! 141: OP_DIGIT, /* \d */
! 142: OP_NOT_WHITESPACE, /* \S */
! 143: OP_WHITESPACE, /* \s */
! 144: OP_NOT_WORDCHAR, /* \W */
! 145: OP_WORDCHAR, /* \w */
! 146: OP_EODN, /* End of data or \n at end of data: \Z. */
! 147: OP_EOD, /* End of data: \z */
! 148:
! 149: OP_OPT, /* Set runtime options */
! 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_REF, /* Match a back reference */
! 197:
! 198: OP_ALT, /* Start of alternation */
! 199: OP_KET, /* End of group that doesn't have an unbounded repeat */
! 200: OP_KETRMAX, /* These two must remain together and in this */
! 201: OP_KETRMIN, /* order. They are for groups the repeat for ever. */
! 202:
! 203: /* The assertions must come before ONCE and COND */
! 204:
! 205: OP_ASSERT, /* Positive lookahead */
! 206: OP_ASSERT_NOT, /* Negative lookahead */
! 207: OP_ASSERTBACK, /* Positive lookbehind */
! 208: OP_ASSERTBACK_NOT, /* Negative lookbehind */
! 209: OP_REVERSE, /* Move pointer back - used in lookbehind assertions */
! 210:
! 211: /* ONCE and COND must come after the assertions, with ONCE first, as there's
! 212: a test for >= ONCE for a subpattern that isn't an assertion. */
! 213:
! 214: OP_ONCE, /* Once matched, don't back up into the subpattern */
! 215: OP_COND, /* Conditional group */
! 216: OP_CREF, /* Used to hold an extraction string number */
! 217:
! 218: OP_BRAZERO, /* These two must remain together and in this */
! 219: OP_BRAMINZERO, /* order. */
! 220:
! 221: OP_BRA /* This and greater values are used for brackets that
! 222: extract substrings. */
! 223: };
! 224:
! 225: /* The highest extraction number. This is limited by the number of opcodes
! 226: left after OP_BRA, i.e. 255 - OP_BRA. We actually set it somewhat lower. */
! 227:
! 228: #define EXTRACT_MAX 99
! 229:
! 230: /* The texts of compile-time error messages are defined as macros here so that
! 231: they can be accessed by the POSIX wrapper and converted into error codes. Yes,
! 232: I could have used error codes in the first place, but didn't feel like changing
! 233: just to accommodate the POSIX wrapper. */
! 234:
! 235: #define ERR1 "\\ at end of pattern"
! 236: #define ERR2 "\\c at end of pattern"
! 237: #define ERR3 "unrecognized character follows \\"
! 238: #define ERR4 "numbers out of order in {} quantifier"
! 239: #define ERR5 "number too big in {} quantifier"
! 240: #define ERR6 "missing terminating ] for character class"
! 241: #define ERR7 "invalid escape sequence in character class"
! 242: #define ERR8 "range out of order in character class"
! 243: #define ERR9 "nothing to repeat"
! 244: #define ERR10 "operand of unlimited repeat could match the empty string"
! 245: #define ERR11 "internal error: unexpected repeat"
! 246: #define ERR12 "unrecognized character after (?"
! 247: #define ERR13 "too many capturing parenthesized sub-patterns"
! 248: #define ERR14 "missing )"
! 249: #define ERR15 "back reference to non-existent subpattern"
! 250: #define ERR16 "erroffset passed as NULL"
! 251: #define ERR17 "unknown option bit(s) set"
! 252: #define ERR18 "missing ) after comment"
! 253: #define ERR19 "too many sets of parentheses"
! 254: #define ERR20 "regular expression too large"
! 255: #define ERR21 "failed to get memory"
! 256: #define ERR22 "unmatched parentheses"
! 257: #define ERR23 "internal error: code overflow"
! 258: #define ERR24 "unrecognized character after (?<"
! 259: #define ERR25 "lookbehind assertion is not fixed length"
! 260: #define ERR26 "malformed number after (?("
! 261: #define ERR27 "conditional group contains more than two branches"
! 262: #define ERR28 "assertion expected after (?("
! 263:
! 264: /* All character handling must be done as unsigned characters. Otherwise there
! 265: are problems with top-bit-set characters and functions such as isspace().
! 266: However, we leave the interface to the outside world as char *, because that
! 267: should make things easier for callers. We define a short type for unsigned char
! 268: to save lots of typing. I tried "uchar", but it causes problems on Digital
! 269: Unix, where it is defined in sys/types, so use "uschar" instead. */
! 270:
! 271: typedef unsigned char uschar;
! 272:
! 273: /* The real format of the start of the pcre block; the actual code vector
! 274: runs on as long as necessary after the end. */
! 275:
! 276: typedef struct real_pcre {
! 277: unsigned long int magic_number;
! 278: const unsigned char *tables;
! 279: unsigned long int options;
! 280: uschar top_bracket;
! 281: uschar top_backref;
! 282: uschar first_char;
! 283: uschar req_char;
! 284: uschar code[1];
! 285: } real_pcre;
! 286:
! 287: /* The real format of the extra block returned by pcre_study(). */
! 288:
! 289: typedef struct real_pcre_extra {
! 290: uschar options;
! 291: uschar start_bits[32];
! 292: } real_pcre_extra;
! 293:
! 294:
! 295: /* Structure for passing "static" information around between the functions
! 296: doing the compiling, so that they are thread-safe. */
! 297:
! 298: typedef struct compile_data {
! 299: const uschar *lcc; /* Points to lower casing table */
! 300: const uschar *fcc; /* Points to case-flipping table */
! 301: const uschar *cbits; /* Points to character type table */
! 302: const uschar *ctypes; /* Points to table of type maps */
! 303: } compile_data;
! 304:
! 305: /* Structure for passing "static" information around between the functions
! 306: doing the matching, so that they are thread-safe. */
! 307:
! 308: typedef struct match_data {
! 309: int errorcode; /* As it says */
! 310: int *offset_vector; /* Offset vector */
! 311: int offset_end; /* One past the end */
! 312: int offset_max; /* The maximum usable for return data */
! 313: const uschar *lcc; /* Points to lower casing table */
! 314: const uschar *ctypes; /* Points to table of type maps */
! 315: BOOL offset_overflow; /* Set if too many extractions */
! 316: BOOL notbol; /* NOTBOL flag */
! 317: BOOL noteol; /* NOTEOL flag */
! 318: BOOL endonly; /* Dollar not before final \n */
! 319: BOOL notempty; /* Empty string match not wanted */
! 320: const uschar *start_subject; /* Start of the subject string */
! 321: const uschar *end_subject; /* End of the subject string */
! 322: const uschar *start_match; /* Start of this match attempt */
! 323: const uschar *end_match_ptr; /* Subject position at end match */
! 324: int end_offset_top; /* Highwater mark at end of match */
! 325: } match_data;
! 326:
! 327: /* Bit definitions for entries in the pcre_ctypes table. */
! 328:
! 329: #define ctype_space 0x01
! 330: #define ctype_letter 0x02
! 331: #define ctype_digit 0x04
! 332: #define ctype_xdigit 0x08
! 333: #define ctype_word 0x10 /* alphameric or '_' */
! 334: #define ctype_meta 0x80 /* regexp meta char or zero (end pattern) */
! 335:
! 336: /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set
! 337: of bits for a class map. */
! 338:
! 339: #define cbit_digit 0 /* for \d */
! 340: #define cbit_word 32 /* for \w */
! 341: #define cbit_space 64 /* for \s */
! 342: #define cbit_length 96 /* Length of the cbits table */
! 343:
! 344: /* Offsets of the various tables from the base tables pointer, and
! 345: total length. */
! 346:
! 347: #define lcc_offset 0
! 348: #define fcc_offset 256
! 349: #define cbits_offset 512
! 350: #define ctypes_offset (cbits_offset + cbit_length)
! 351: #define tables_length (ctypes_offset + 256)
! 352:
! 353: /* End of internal.h */
E-mail:
![[BACK]](/cvsweb/icons/back.gif){kind=icon}
Return to internal.h CVS log ![[TXT]](/cvsweb/icons/text.gif){kind=icon}
![[DIR]](/cvsweb/icons/dir.gif){kind=icon}
Up to [parser3project] / parser3 / src / lib / pcre