/[pcre]/code/trunk/internal.h

Diff of /code/trunk/internal.h

Parent Directory | Revision Log | View Patch Patch

-revision 7 by nigel,
Sat Feb 24 21:38:09 2007 UTC
+revision 53 by nigel,
Sat Feb 24 21:39:42 2007 UTC
 Line 3
  *************************************************/
- #define PCRE_VERSION       "1.02 12-Dec-1997"
  /* This is a library of functions to support regular expressions whose syntax
  and semantics are as close as possible to those of the Perl 5 language. See
  the file Tech.Notes for some information on the internals.
  Written by: Philip Hazel <ph10@cam.ac.uk>
-            Copyright (c) 1997 University of Cambridge
+            Copyright (c) 1997-2001 University of Cambridge
  -----------------------------------------------------------------------------
  Permission is granted to anyone to use this software for any purpose on any
-Line 28 
 restrictions:
+Line 25 
 restrictions:
 . Altered versions must be plainly marked as such, and must not be
     misrepresented as being the original software.
+. If PCRE is embedded in any software that is released under the GNU
+    General Purpose Licence (GPL), then the terms of that licence shall
+    supersede any condition above with which it is incompatible.
  -----------------------------------------------------------------------------
  */
  /* This header contains definitions that are shared between the different
  modules, but which are not relevant to the outside. */
- /* To cope with SunOS4 and other systems that lack memmove() but have bcopy(),
+ /* Get the definitions provided by running "configure" */
- define a macro for memmove() if USE_BCOPY is defined. */
- #ifdef USE_BCOPY
+ #include "config.h"
+ /* To cope with SunOS4 and other systems that lack memmove() but have bcopy(),
+ define a macro for memmove() if HAVE_MEMMOVE is false, provided that HAVE_BCOPY
+ is set. Otherwise, include an emulating function for those systems that have
+ neither (there some non-Unix environments where this is the case). This assumes
+ that all calls to memmove are moving strings upwards in store, which is the
+ case in PCRE. */
+ #if ! HAVE_MEMMOVE
+ #undef  memmove        /* some systems may have a macro */
+ #if HAVE_BCOPY
  #define memmove(a, b, c) bcopy(b, a, c)
+ #else
+ void *
+ pcre_memmove(unsigned char *dest, const unsigned char *src, size_t n)
+ {
+ int i;
+ dest += n;
+ src += n;
+ for (i = 0; i < n; ++i) *(--dest) =  *(--src);
+ }
+ #define memmove(a, b, c) pcre_memmove(a, b, c)
+ #endif
  #endif
  /* Standard C headers plus the external interface definition */
  #include <ctype.h>
  #include <limits.h>
- #include <setjmp.h>
  #include <stddef.h>
  #include <stdio.h>
  #include <stdlib.h>
  #include <string.h>
  #include "pcre.h"
- /* Private options flags start at the most significant end of the two bytes.
+ /* In case there is no definition of offsetof() provided - though any proper
- The public options defined in pcre.h start at the least significant end. Make
+ Standard C system should have one. */
- sure they don't overlap! */
+ #ifndef offsetof
- #define PCRE_FIRSTSET           0x8000  /* first_char is set */
+ #define offsetof(p_type,field) ((size_t)&(((p_type *)0)->field))
- #define PCRE_STARTLINE          0x4000  /* start after \n for multiline */
+ #endif
- #define PCRE_COMPILED_CASELESS  0x2000  /* like it says */
+ /* These are the public options that can change during matching. */
+ #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
+ /* Private options flags start at the most significant end of the four bytes,
+ but skip the top bit so we can use ints for convenience without getting tangled
+ with negative values. The public options defined in pcre.h start at the least
+ significant end. Make sure they don't overlap, though now that we have expanded
+ to four bytes there is plenty of space. */
+ #define PCRE_FIRSTSET      0x40000000  /* first_char is set */
+ #define PCRE_REQCHSET      0x20000000  /* req_char is set */
+ #define PCRE_STARTLINE     0x10000000  /* start after \n for multiline */
+ #define PCRE_INGROUP       0x08000000  /* compiling inside a group */
+ #define PCRE_ICHANGED      0x04000000  /* i option changes within regex */
  /* Options for the "extra" block produced by pcre_study(). */
- #define PCRE_STUDY_CASELESS 0x01     /* study was caseless */
+ #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */
- #define PCRE_STUDY_MAPPED   0x02     /* a map of starting chars exists */
- /* Masks for identifying the public options: all permitted at compile time,
+ /* Masks for identifying the public options which are permitted at compile
- only some permitted at run or study time. */
+ time, run time or study time, respectively. */
  #define PUBLIC_OPTIONS \
    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
-    PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA)
+    PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8)
  #define PUBLIC_EXEC_OPTIONS \
-   (PCRE_CASELESS|PCRE_ANCHORED|PCRE_MULTILINE|PCRE_NOTBOL|PCRE_NOTEOL| \
+   (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY)
-    PCRE_DOTALL|PCRE_DOLLAR_ENDONLY)
- #define PUBLIC_STUDY_OPTIONS (PCRE_CASELESS)
+ #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
  /* Magic number to provide a small check against being handed junk. */
- #define MAGIC_NUMBER  0x50435245   /* 'PCRE' */
+ #define MAGIC_NUMBER  0x50435245UL   /* 'PCRE' */
  /* Miscellaneous definitions */
-Line 89 
 typedef int BOOL;
+Line 123 
 typedef int BOOL;
  #define FALSE   0
  #define TRUE    1
+ /* Escape items that are just an encoding of a particular data value. Note that
+ ESC_N is defined as yet another macro, which is set in config.h to either \n
+ (the default) or \r (which some people want). */
+ #ifndef ESC_E
+ #define ESC_E 27
+ #endif
+ #ifndef ESC_F
+ #define ESC_F '\f'
+ #endif
+ #ifndef ESC_N
+ #define ESC_N NEWLINE
+ #endif
+ #ifndef ESC_R
+ #define ESC_R '\r'
+ #endif
+ #ifndef ESC_T
+ #define ESC_T '\t'
+ #endif
  /* These are escaped items that aren't just an encoding of a particular data
  value such as \n. They must have non-zero values, as check_escape() returns
  their negation. Also, they must appear in the same order as in the opcode
- definitions below, up to ESC_Z. The final one must be ESC_REF as subsequent
+ definitions below, up to ESC_z. The final one must be ESC_REF as subsequent
  values are used for \1, \2, \3, etc. There is a test in the code for an escape
- greater than ESC_b and less than ESC_X to detect the types that may be
+ greater than ESC_b and less than ESC_Z to detect the types that may be
  repeated. If any new escapes are put in-between that don't consume a character,
  that code will have to change. */
  enum { ESC_A = 1, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W, ESC_w,
+        ESC_Z, ESC_z, ESC_REF };
-                     /* These are not Perl escapes, so can't appear in the */
-        ESC_X,       /* simple table-lookup because they must be conditional */
-                     /* on PCRE_EXTRA. */
-        ESC_Z,
-        ESC_REF };
  /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets
  that extract substrings. Starting from 1 (i.e. after OP_END), the values up to
-Line 124 
 enum {
+Line 177 
 enum {
    OP_WHITESPACE,         /* \s */
    OP_NOT_WORDCHAR,       /* \W */
    OP_WORDCHAR,           /* \w */
-   OP_CUT,            /* The analogue of Prolog's "cut" operation (extension) */
+   OP_EODN,           /* End of data or \n at end of data: \Z. */
-   OP_EOD,            /* End of data: \Z. */
+   OP_EOD,            /* End of data: \z */
+   OP_OPT,            /* Set runtime options */
    OP_CIRC,           /* Start of line - varies with multiline switch */
    OP_DOLL,           /* End of line - varies with multiline switch */
    OP_ANY,            /* Match any character */
-Line 174 
 enum {
+Line 228 
 enum {
    OP_CLASS,          /* Match a character class */
    OP_REF,            /* Match a back reference */
+   OP_RECURSE,        /* Match this pattern recursively */
    OP_ALT,            /* Start of alternation */
    OP_KET,            /* End of group that doesn't have an unbounded repeat */
    OP_KETRMAX,        /* These two must remain together and in this */
    OP_KETRMIN,        /* order. They are for groups the repeat for ever. */
-   OP_ASSERT,
+   /* The assertions must come before ONCE and COND */
-   OP_ASSERT_NOT,
+   OP_ASSERT,         /* Positive lookahead */
+   OP_ASSERT_NOT,     /* Negative lookahead */
+   OP_ASSERTBACK,     /* Positive lookbehind */
+   OP_ASSERTBACK_NOT, /* Negative lookbehind */
+   OP_REVERSE,        /* Move pointer back - used in lookbehind assertions */
+   /* ONCE and COND must come after the assertions, with ONCE first, as there's
+   a test for >= ONCE for a subpattern that isn't an assertion. */
    OP_ONCE,           /* Once matched, don't back up into the subpattern */
+   OP_COND,           /* Conditional group */
+   OP_CREF,           /* Used to hold an extraction string number (cond ref) */
    OP_BRAZERO,        /* These two must remain together and in this */
    OP_BRAMINZERO,     /* order. */
+   OP_BRANUMBER,      /* Used for extracting brackets whose number is greater
+                         than can fit into an opcode. */
    OP_BRA             /* This and greater values are used for brackets that
-                         extract substrings. */
+                         extract substrings up to a basic limit. After that,
+                         use is made of OP_BRANUMBER. */
  };
- /* The highest extraction number. This is limited by the number of opcodes
+ /* The highest extraction number before we have to start using additional
- left after OP_BRA, i.e. 255 - OP_BRA. We actually set it somewhat lower. */
+ bytes. (Originally PCRE didn't have support for extraction counts highter than
+ this number.) The value is limited by the number of opcodes left after OP_BRA,
+ i.e. 255 - OP_BRA. We actually set it a bit lower to leave room for additional
+ opcodes. */
- #define EXTRACT_MAX  99
+ #define EXTRACT_BASIC_MAX  150
  /* The texts of compile-time error messages are defined as macros here so that
  they can be accessed by the POSIX wrapper and converted into error codes.  Yes,
-Line 213 
 just to accommodate the POSIX wrapper. *
+Line 286 
 just to accommodate the POSIX wrapper. *
  #define ERR10 "operand of unlimited repeat could match the empty string"
  #define ERR11 "internal error: unexpected repeat"
  #define ERR12 "unrecognized character after (?"
- #define ERR13 "too many capturing parenthesized sub-patterns"
+ #define ERR13 "unused error"
  #define ERR14 "missing )"
  #define ERR15 "back reference to non-existent subpattern"
  #define ERR16 "erroffset passed as NULL"
  #define ERR17 "unknown option bit(s) set"
  #define ERR18 "missing ) after comment"
- #define ERR19 "too many sets of parentheses"
+ #define ERR19 "parentheses nested too deeply"
  #define ERR20 "regular expression too large"
  #define ERR21 "failed to get memory"
- #define ERR22 "unmatched brackets"
+ #define ERR22 "unmatched parentheses"
  #define ERR23 "internal error: code overflow"
+ #define ERR24 "unrecognized character after (?<"
+ #define ERR25 "lookbehind assertion is not fixed length"
+ #define ERR26 "malformed number after (?("
+ #define ERR27 "conditional group contains more than two branches"
+ #define ERR28 "assertion expected after (?("
+ #define ERR29 "(?p must be followed by )"
+ #define ERR30 "unknown POSIX class name"
+ #define ERR31 "POSIX collating elements are not supported"
+ #define ERR32 "this version of PCRE is not compiled with PCRE_UTF8 support"
+ #define ERR33 "characters with values > 255 are not yet supported in classes"
+ #define ERR34 "character value in \\x{...} sequence is too large"
+ #define ERR35 "invalid condition (?(0)"
  /* All character handling must be done as unsigned characters. Otherwise there
  are problems with top-bit-set characters and functions such as isspace().
-Line 238 
 typedef unsigned char uschar;
+Line 323 
 typedef unsigned char uschar;
  runs on as long as necessary after the end. */
  typedef struct real_pcre {
-   unsigned int  magic_number;
+   unsigned long int magic_number;
-   unsigned short int options;
+   size_t size;
-   unsigned char top_bracket;
+   const unsigned char *tables;
-   unsigned char top_backref;
+   unsigned long int options;
-   unsigned char first_char;
+   unsigned short int top_bracket;
-   unsigned char code[1];
+   unsigned short int top_backref;
+   uschar first_char;
+   uschar req_char;
+   uschar code[1];
  } real_pcre;
  /* The real format of the extra block returned by pcre_study(). */
  typedef struct real_pcre_extra {
-   unsigned char options;
+   uschar options;
-   unsigned char start_bits[32];
+   uschar start_bits[32];
  } real_pcre_extra;
- /* Global tables from chartables.c */
- extern uschar pcre_lcc[];
+ /* Structure for passing "static" information around between the functions
- extern uschar pcre_fcc[];
+ doing the compiling, so that they are thread-safe. */
- extern uschar pcre_cbits[];
- extern uschar pcre_ctypes[];
+ typedef struct compile_data {
+   const uschar *lcc;            /* Points to lower casing table */
+   const uschar *fcc;            /* Points to case-flipping table */
+   const uschar *cbits;          /* Points to character type table */
+   const uschar *ctypes;         /* Points to table of type maps */
+ } compile_data;
+ /* Structure for passing "static" information around between the functions
+ doing the matching, so that they are thread-safe. */
+ typedef struct match_data {
+   int    errorcode;             /* As it says */
+   int   *offset_vector;         /* Offset vector */
+   int    offset_end;            /* One past the end */
+   int    offset_max;            /* The maximum usable for return data */
+   const uschar *lcc;            /* Points to lower casing table */
+   const uschar *ctypes;         /* Points to table of type maps */
+   BOOL   offset_overflow;       /* Set if too many extractions */
+   BOOL   notbol;                /* NOTBOL flag */
+   BOOL   noteol;                /* NOTEOL flag */
+   BOOL   utf8;                  /* UTF8 flag */
+   BOOL   endonly;               /* Dollar not before final \n */
+   BOOL   notempty;              /* Empty string match not wanted */
+   const uschar *start_pattern;  /* For use when recursing */
+   const uschar *start_subject;  /* Start of the subject string */
+   const uschar *end_subject;    /* End of the subject string */
+   const uschar *start_match;    /* Start of this match attempt */
+   const uschar *end_match_ptr;  /* Subject position at end match */
+   int    end_offset_top;        /* Highwater mark at end of match */
+ } match_data;
- /* Bit definitions for entries in pcre_ctypes[]. */
+ /* Bit definitions for entries in the pcre_ctypes table. */
  #define ctype_space   0x01
  #define ctype_letter  0x02
-Line 269 
 extern uschar pcre_ctypes[];
+Line 385 
 extern uschar pcre_ctypes[];
  #define ctype_word    0x10   /* alphameric or '_' */
  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */
- /* Offsets for the bitmap tables */
+ /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set
+ of bits for a class map. Some classes are built by combining these tables. */
- #define cbit_digit    0
+ #define cbit_space     0      /* [:space:] or \s */
- #define cbit_letter  32
+ #define cbit_xdigit   32      /* [:xdigit:] */
- #define cbit_word    64
+ #define cbit_digit    64      /* [:digit:] or \d */
- #define cbit_space   96
+ #define cbit_upper    96      /* [:upper:] */
- #define cbit_length 128      /* Length of the cbits table */
+ #define cbit_lower   128      /* [:lower:] */
+ #define cbit_word    160      /* [:word:] or \w */
+ #define cbit_graph   192      /* [:graph:] */
+ #define cbit_print   224      /* [:print:] */
+ #define cbit_punct   256      /* [:punct:] */
+ #define cbit_cntrl   288      /* [:cntrl:] */
+ #define cbit_length  320      /* Length of the cbits table */
+ /* Offsets of the various tables from the base tables pointer, and
+ total length. */
+ #define lcc_offset      0
+ #define fcc_offset    256
+ #define cbits_offset  512
+ #define ctypes_offset (cbits_offset + cbit_length)
+ #define tables_length (ctypes_offset + 256)
  /* End of internal.h */

 Legend:



Removed from v.7
 


changed lines


 
Added in v.53
 Legend:



Removed from v.7
 


changed lines


 
Added in v.53
-Removed from v.7
+Added in v.53

	ViewVC Help
Powered by ViewVC 1.1.5