Annotation of parser3/src/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: #include <stddef.h>
! 51: #include <stdio.h>
! 52: #include <stdlib.h>
! 53: #include <string.h>
! 54: #include "pcre.h"
! 55:
! 56: /* In case there is no definition of offsetof() provided - though any proper
! 57: Standard C system should have one. */
! 58:
! 59: #ifndef offsetof
! 60: #define offsetof(p_type,field) ((size_t)&(((p_type *)0)->field))
! 61: #endif
! 62:
! 63: /* These are the public options that can change during matching. */
! 64:
! 65: #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
! 66:
! 67: /* Private options flags start at the most significant end of the four bytes,
! 68: but skip the top bit so we can use ints for convenience without getting tangled
! 69: with negative values. The public options defined in pcre.h start at the least
! 70: significant end. Make sure they don't overlap, though now that we have expanded
! 71: to four bytes there is plenty of space. */
! 72:
! 73: #define PCRE_FIRSTSET 0x40000000 /* first_char is set */
! 74: #define PCRE_REQCHSET 0x20000000 /* req_char is set */
! 75: #define PCRE_STARTLINE 0x10000000 /* start after \n for multiline */
! 76: #define PCRE_INGROUP 0x08000000 /* compiling inside a group */
! 77: #define PCRE_ICHANGED 0x04000000 /* i option changes within regex */
! 78:
! 79: /* Options for the "extra" block produced by pcre_study(). */
! 80:
! 81: #define PCRE_STUDY_MAPPED 0x01 /* a map of starting chars exists */
! 82:
! 83: /* Masks for identifying the public options which are permitted at compile
! 84: time, run time or study time, respectively. */
! 85:
! 86: #define PUBLIC_OPTIONS \
! 87: (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
! 88: PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY)
! 89:
! 90: #define PUBLIC_EXEC_OPTIONS \
! 91: (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY)
! 92:
! 93: #define PUBLIC_STUDY_OPTIONS 0 /* None defined */
! 94:
! 95: /* Magic number to provide a small check against being handed junk. */
! 96:
! 97: #define MAGIC_NUMBER 0x50435245UL /* 'PCRE' */
! 98:
! 99: /* Miscellaneous definitions */
! 100:
! 101: typedef int BOOL;
! 102:
! 103: #define FALSE 0
! 104: #define TRUE 1
! 105:
! 106: /* These are escaped items that aren't just an encoding of a particular data
! 107: value such as \n. They must have non-zero values, as check_escape() returns
! 108: their negation. Also, they must appear in the same order as in the opcode
! 109: definitions below, up to ESC_z. The final one must be ESC_REF as subsequent
! 110: values are used for \1, \2, \3, etc. There is a test in the code for an escape
! 111: greater than ESC_b and less than ESC_X to detect the types that may be
! 112: repeated. If any new escapes are put in-between that don't consume a character,
! 113: that code will have to change. */
! 114:
! 115: enum { ESC_A = 1, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W, ESC_w,
! 116: ESC_Z, ESC_z, ESC_REF };
! 117:
! 118: /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets
! 119: that extract substrings. Starting from 1 (i.e. after OP_END), the values up to
! 120: OP_EOD must correspond in order to the list of escapes immediately above. */
! 121:
! 122: enum {
! 123: OP_END, /* End of pattern */
! 124:
! 125: /* Values corresponding to backslashed metacharacters */
! 126:
! 127: OP_SOD, /* Start of data: \A */
! 128: OP_NOT_WORD_BOUNDARY, /* \B */
! 129: OP_WORD_BOUNDARY, /* \b */
! 130: OP_NOT_DIGIT, /* \D */
! 131: OP_DIGIT, /* \d */
! 132: OP_NOT_WHITESPACE, /* \S */
! 133: OP_WHITESPACE, /* \s */
! 134: OP_NOT_WORDCHAR, /* \W */
! 135: OP_WORDCHAR, /* \w */
! 136: OP_EODN, /* End of data or \n at end of data: \Z. */
! 137: OP_EOD, /* End of data: \z */
! 138:
! 139: OP_OPT, /* Set runtime options */
! 140: OP_CIRC, /* Start of line - varies with multiline switch */
! 141: OP_DOLL, /* End of line - varies with multiline switch */
! 142: OP_ANY, /* Match any character */
! 143: OP_CHARS, /* Match string of characters */
! 144: OP_NOT, /* Match anything but the following char */
! 145:
! 146: OP_STAR, /* The maximizing and minimizing versions of */
! 147: OP_MINSTAR, /* all these opcodes must come in pairs, with */
! 148: OP_PLUS, /* the minimizing one second. */
! 149: OP_MINPLUS, /* This first set applies to single characters */
! 150: OP_QUERY,
! 151: OP_MINQUERY,
! 152: OP_UPTO, /* From 0 to n matches */
! 153: OP_MINUPTO,
! 154: OP_EXACT, /* Exactly n matches */
! 155:
! 156: OP_NOTSTAR, /* The maximizing and minimizing versions of */
! 157: OP_NOTMINSTAR, /* all these opcodes must come in pairs, with */
! 158: OP_NOTPLUS, /* the minimizing one second. */
! 159: OP_NOTMINPLUS, /* This first set applies to "not" single characters */
! 160: OP_NOTQUERY,
! 161: OP_NOTMINQUERY,
! 162: OP_NOTUPTO, /* From 0 to n matches */
! 163: OP_NOTMINUPTO,
! 164: OP_NOTEXACT, /* Exactly n matches */
! 165:
! 166: OP_TYPESTAR, /* The maximizing and minimizing versions of */
! 167: OP_TYPEMINSTAR, /* all these opcodes must come in pairs, with */
! 168: OP_TYPEPLUS, /* the minimizing one second. These codes must */
! 169: OP_TYPEMINPLUS, /* be in exactly the same order as those above. */
! 170: OP_TYPEQUERY, /* This set applies to character types such as \d */
! 171: OP_TYPEMINQUERY,
! 172: OP_TYPEUPTO, /* From 0 to n matches */
! 173: OP_TYPEMINUPTO,
! 174: OP_TYPEEXACT, /* Exactly n matches */
! 175:
! 176: OP_CRSTAR, /* The maximizing and minimizing versions of */
! 177: OP_CRMINSTAR, /* all these opcodes must come in pairs, with */
! 178: OP_CRPLUS, /* the minimizing one second. These codes must */
! 179: OP_CRMINPLUS, /* be in exactly the same order as those above. */
! 180: OP_CRQUERY, /* These are for character classes and back refs */
! 181: OP_CRMINQUERY,
! 182: OP_CRRANGE, /* These are different to the three seta above. */
! 183: OP_CRMINRANGE,
! 184:
! 185: OP_CLASS, /* Match a character class */
! 186: OP_REF, /* Match a back reference */
! 187:
! 188: OP_ALT, /* Start of alternation */
! 189: OP_KET, /* End of group that doesn't have an unbounded repeat */
! 190: OP_KETRMAX, /* These two must remain together and in this */
! 191: OP_KETRMIN, /* order. They are for groups the repeat for ever. */
! 192:
! 193: /* The assertions must come before ONCE and COND */
! 194:
! 195: OP_ASSERT, /* Positive lookahead */
! 196: OP_ASSERT_NOT, /* Negative lookahead */
! 197: OP_ASSERTBACK, /* Positive lookbehind */
! 198: OP_ASSERTBACK_NOT, /* Negative lookbehind */
! 199: OP_REVERSE, /* Move pointer back - used in lookbehind assertions */
! 200:
! 201: /* ONCE and COND must come after the assertions, with ONCE first, as there's
! 202: a test for >= ONCE for a subpattern that isn't an assertion. */
! 203:
! 204: OP_ONCE, /* Once matched, don't back up into the subpattern */
! 205: OP_COND, /* Conditional group */
! 206: OP_CREF, /* Used to hold an extraction string number */
! 207:
! 208: OP_BRAZERO, /* These two must remain together and in this */
! 209: OP_BRAMINZERO, /* order. */
! 210:
! 211: OP_BRA /* This and greater values are used for brackets that
! 212: extract substrings. */
! 213: };
! 214:
! 215: /* The highest extraction number. This is limited by the number of opcodes
! 216: left after OP_BRA, i.e. 255 - OP_BRA. We actually set it somewhat lower. */
! 217:
! 218: #define EXTRACT_MAX 99
! 219:
! 220: /* The texts of compile-time error messages are defined as macros here so that
! 221: they can be accessed by the POSIX wrapper and converted into error codes. Yes,
! 222: I could have used error codes in the first place, but didn't feel like changing
! 223: just to accommodate the POSIX wrapper. */
! 224:
! 225: #define ERR1 "\\ at end of pattern"
! 226: #define ERR2 "\\c at end of pattern"
! 227: #define ERR3 "unrecognized character follows \\"
! 228: #define ERR4 "numbers out of order in {} quantifier"
! 229: #define ERR5 "number too big in {} quantifier"
! 230: #define ERR6 "missing terminating ] for character class"
! 231: #define ERR7 "invalid escape sequence in character class"
! 232: #define ERR8 "range out of order in character class"
! 233: #define ERR9 "nothing to repeat"
! 234: #define ERR10 "operand of unlimited repeat could match the empty string"
! 235: #define ERR11 "internal error: unexpected repeat"
! 236: #define ERR12 "unrecognized character after (?"
! 237: #define ERR13 "too many capturing parenthesized sub-patterns"
! 238: #define ERR14 "missing )"
! 239: #define ERR15 "back reference to non-existent subpattern"
! 240: #define ERR16 "erroffset passed as NULL"
! 241: #define ERR17 "unknown option bit(s) set"
! 242: #define ERR18 "missing ) after comment"
! 243: #define ERR19 "too many sets of parentheses"
! 244: #define ERR20 "regular expression too large"
! 245: #define ERR21 "failed to get memory"
! 246: #define ERR22 "unmatched parentheses"
! 247: #define ERR23 "internal error: code overflow"
! 248: #define ERR24 "unrecognized character after (?<"
! 249: #define ERR25 "lookbehind assertion is not fixed length"
! 250: #define ERR26 "malformed number after (?("
! 251: #define ERR27 "conditional group contains more than two branches"
! 252: #define ERR28 "assertion expected after (?("
! 253:
! 254: /* All character handling must be done as unsigned characters. Otherwise there
! 255: are problems with top-bit-set characters and functions such as isspace().
! 256: However, we leave the interface to the outside world as char *, because that
! 257: should make things easier for callers. We define a short type for unsigned char
! 258: to save lots of typing. I tried "uchar", but it causes problems on Digital
! 259: Unix, where it is defined in sys/types, so use "uschar" instead. */
! 260:
! 261: typedef unsigned char uschar;
! 262:
! 263: /* The real format of the start of the pcre block; the actual code vector
! 264: runs on as long as necessary after the end. */
! 265:
! 266: typedef struct real_pcre {
! 267: unsigned long int magic_number;
! 268: const unsigned char *tables;
! 269: unsigned long int options;
! 270: uschar top_bracket;
! 271: uschar top_backref;
! 272: uschar first_char;
! 273: uschar req_char;
! 274: uschar code[1];
! 275: } real_pcre;
! 276:
! 277: /* The real format of the extra block returned by pcre_study(). */
! 278:
! 279: typedef struct real_pcre_extra {
! 280: uschar options;
! 281: uschar start_bits[32];
! 282: } real_pcre_extra;
! 283:
! 284:
! 285: /* Structure for passing "static" information around between the functions
! 286: doing the compiling, so that they are thread-safe. */
! 287:
! 288: typedef struct compile_data {
! 289: const uschar *lcc; /* Points to lower casing table */
! 290: const uschar *fcc; /* Points to case-flipping table */
! 291: const uschar *cbits; /* Points to character type table */
! 292: const uschar *ctypes; /* Points to table of type maps */
! 293: } compile_data;
! 294:
! 295: /* Structure for passing "static" information around between the functions
! 296: doing the matching, so that they are thread-safe. */
! 297:
! 298: typedef struct match_data {
! 299: int errorcode; /* As it says */
! 300: int *offset_vector; /* Offset vector */
! 301: int offset_end; /* One past the end */
! 302: int offset_max; /* The maximum usable for return data */
! 303: const uschar *lcc; /* Points to lower casing table */
! 304: const uschar *ctypes; /* Points to table of type maps */
! 305: BOOL offset_overflow; /* Set if too many extractions */
! 306: BOOL notbol; /* NOTBOL flag */
! 307: BOOL noteol; /* NOTEOL flag */
! 308: BOOL endonly; /* Dollar not before final \n */
! 309: BOOL notempty; /* Empty string match not wanted */
! 310: const uschar *start_subject; /* Start of the subject string */
! 311: const uschar *end_subject; /* End of the subject string */
! 312: const uschar *start_match; /* Start of this match attempt */
! 313: const uschar *end_match_ptr; /* Subject position at end match */
! 314: int end_offset_top; /* Highwater mark at end of match */
! 315: } match_data;
! 316:
! 317: /* Bit definitions for entries in the pcre_ctypes table. */
! 318:
! 319: #define ctype_space 0x01
! 320: #define ctype_letter 0x02
! 321: #define ctype_digit 0x04
! 322: #define ctype_xdigit 0x08
! 323: #define ctype_word 0x10 /* alphameric or '_' */
! 324: #define ctype_meta 0x80 /* regexp meta char or zero (end pattern) */
! 325:
! 326: /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set
! 327: of bits for a class map. */
! 328:
! 329: #define cbit_digit 0 /* for \d */
! 330: #define cbit_word 32 /* for \w */
! 331: #define cbit_space 64 /* for \s */
! 332: #define cbit_length 96 /* Length of the cbits table */
! 333:
! 334: /* Offsets of the various tables from the base tables pointer, and
! 335: total length. */
! 336:
! 337: #define lcc_offset 0
! 338: #define fcc_offset 256
! 339: #define cbits_offset 512
! 340: #define ctypes_offset (cbits_offset + cbit_length)
! 341: #define tables_length (ctypes_offset + 256)
! 342:
! 343: /* 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 / pcre