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

Diff of /code/trunk/pcre_internal.h

Parent Directory | Revision Log | View Patch Patch

-revision 1248 by ph10,
Wed Feb 13 17:36:38 2013 UTC
+revision 1387 by ph10,
Sat Nov  2 18:29:05 2013 UTC
 Line 7
  and semantics are as close as possible to those of the Perl 5 language.
                         Written by Philip Hazel
-            Copyright (c) 1997-2012 University of Cambridge
+            Copyright (c) 1997-2013 University of Cambridge
  -----------------------------------------------------------------------------
  Redistribution and use in source and binary forms, with or without
 Line 194 
 preprocessor time in standard C environm
  typedef unsigned char pcre_uint8;
  #if USHRT_MAX == 65535
-   typedef unsigned short pcre_uint16;
+ typedef unsigned short pcre_uint16;
-   typedef short pcre_int16;
+ typedef short pcre_int16;
+ #define PCRE_UINT16_MAX USHRT_MAX
+ #define PCRE_INT16_MAX SHRT_MAX
  #elif UINT_MAX == 65535
-   typedef unsigned int pcre_uint16;
+ typedef unsigned int pcre_uint16;
-   typedef int pcre_int16;
+ typedef int pcre_int16;
+ #define PCRE_UINT16_MAX UINT_MAX
+ #define PCRE_INT16_MAX INT_MAX
  #else
- # error Cannot determine a type for 16-bit unsigned integers
+ #error Cannot determine a type for 16-bit integers
  #endif
- #if UINT_MAX == 4294967295
+ #if UINT_MAX == 4294967295U
-   typedef unsigned int pcre_uint32;
+ typedef unsigned int pcre_uint32;
-   typedef int pcre_int32;
+ typedef int pcre_int32;
- #elif ULONG_MAX == 4294967295
+ #define PCRE_UINT32_MAX UINT_MAX
-   typedef unsigned long int pcre_uint32;
+ #define PCRE_INT32_MAX INT_MAX
-   typedef long int pcre_int32;
+ #elif ULONG_MAX == 4294967295UL
+ typedef unsigned long int pcre_uint32;
+ typedef long int pcre_int32;
+ #define PCRE_UINT32_MAX ULONG_MAX
+ #define PCRE_INT32_MAX LONG_MAX
  #else
- # error Cannot determine a type for 32-bit unsigned integers
+ #error Cannot determine a type for 32-bit integers
  #endif
  /* When checking for integer overflow in pcre_compile(), we need to handle
-Line 1121 
 other. NOTE: The values also appear in p
+Line 1129 
 other. NOTE: The values also appear in p
  /* Private flags containing information about the compiled regex. They used to
- live at the top end of the options word, but that got almost full, so now they
+ live at the top end of the options word, but that got almost full, so they were
- are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as
+ moved to a 16-bit flags word - which got almost full, so now they are in a
- the restrictions on partial matching have been lifted. It remains for backwards
+-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as the
+ restrictions on partial matching have been lifted. It remains for backwards
  compatibility. */
- #define PCRE_MODE8         0x0001  /* compiled in 8 bit mode */
+ #define PCRE_MODE8         0x00000001  /* compiled in 8 bit mode */
- #define PCRE_MODE16        0x0002  /* compiled in 16 bit mode */
+ #define PCRE_MODE16        0x00000002  /* compiled in 16 bit mode */
- #define PCRE_MODE32        0x0004  /* compiled in 32 bit mode */
+ #define PCRE_MODE32        0x00000004  /* compiled in 32 bit mode */
- #define PCRE_FIRSTSET      0x0010  /* first_char is set */
+ #define PCRE_FIRSTSET      0x00000010  /* first_char is set */
- #define PCRE_FCH_CASELESS  0x0020  /* caseless first char */
+ #define PCRE_FCH_CASELESS  0x00000020  /* caseless first char */
- #define PCRE_REQCHSET      0x0040  /* req_byte is set */
+ #define PCRE_REQCHSET      0x00000040  /* req_byte is set */
- #define PCRE_RCH_CASELESS  0x0080  /* caseless requested char */
+ #define PCRE_RCH_CASELESS  0x00000080  /* caseless requested char */
- #define PCRE_STARTLINE     0x0100  /* start after \n for multiline */
+ #define PCRE_STARTLINE     0x00000100  /* start after \n for multiline */
- #define PCRE_NOPARTIAL     0x0200  /* can't use partial with this regex */
+ #define PCRE_NOPARTIAL     0x00000200  /* can't use partial with this regex */
- #define PCRE_JCHANGED      0x0400  /* j option used in regex */
+ #define PCRE_JCHANGED      0x00000400  /* j option used in regex */
- #define PCRE_HASCRORLF     0x0800  /* explicit \r or \n in pattern */
+ #define PCRE_HASCRORLF     0x00000800  /* explicit \r or \n in pattern */
- #define PCRE_HASTHEN       0x1000  /* pattern contains (*THEN) */
+ #define PCRE_HASTHEN       0x00001000  /* pattern contains (*THEN) */
+ #define PCRE_MLSET         0x00002000  /* match limit set by regex */
+ #define PCRE_RLSET         0x00004000  /* recursion limit set by regex */
+ #define PCRE_MATCH_EMPTY   0x00008000  /* pattern can match empty string */
  #if defined COMPILE_PCRE8
  #define PCRE_MODE          PCRE_MODE8
-Line 1162 
 time, run time, or study time, respectiv
+Line 1174 
 time, run time, or study time, respectiv
  #define PUBLIC_COMPILE_OPTIONS \
    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
-    PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \
+    PCRE_NO_AUTO_CAPTURE|PCRE_NO_AUTO_POSSESSIFY| \
+    PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \
     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
-    PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE)
+    PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE|PCRE_NEVER_UTF)
  #define PUBLIC_EXEC_OPTIONS \
    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
-Line 1534 
 a positive value. */
+Line 1547 
 a positive value. */
  #define STRING_UTF_RIGHTPAR            "UTF)"
  #define STRING_UCP_RIGHTPAR            "UCP)"
  #define STRING_NO_START_OPT_RIGHTPAR   "NO_START_OPT)"
+ #define STRING_LIMIT_MATCH_EQ          "LIMIT_MATCH="
+ #define STRING_LIMIT_RECURSION_EQ      "LIMIT_RECURSION="
  #else  /* SUPPORT_UTF */
-Line 1795 
 only. */
+Line 1810 
 only. */
  #define STRING_UTF_RIGHTPAR            STR_U STR_T STR_F STR_RIGHT_PARENTHESIS
  #define STRING_UCP_RIGHTPAR            STR_U STR_C STR_P STR_RIGHT_PARENTHESIS
  #define STRING_NO_START_OPT_RIGHTPAR   STR_N STR_O STR_UNDERSCORE STR_S STR_T STR_A STR_R STR_T STR_UNDERSCORE STR_O STR_P STR_T STR_RIGHT_PARENTHESIS
+ #define STRING_LIMIT_MATCH_EQ          STR_L STR_I STR_M STR_I STR_T STR_UNDERSCORE STR_M STR_A STR_T STR_C STR_H STR_EQUALS_SIGN
+ #define STRING_LIMIT_RECURSION_EQ      STR_L STR_I STR_M STR_I STR_T STR_UNDERSCORE STR_R STR_E STR_C STR_U STR_R STR_S STR_I STR_O STR_N STR_EQUALS_SIGN
  #endif  /* SUPPORT_UTF */
-Line 1835 
 only. */
+Line 1852 
 only. */
  #define PT_PXSPACE    7    /* POSIX space - Z plus 9,10,11,12,13 */
  #define PT_WORD       8    /* Word - L plus N plus underscore */
  #define PT_CLIST      9    /* Pseudo-property: match character list */
+ #define PT_UCNC      10    /* Universal Character nameable character */
+ #define PT_TABSIZE   11    /* Size of square table for autopossessify tests */
+ /* The following special properties are used only in XCLASS items, when POSIX
+ classes are specified and PCRE_UCP is set - in other words, for Unicode
+ handling of these classes. They are not available via the \p or \P escapes like
+ those in the above list, and so they do not take part in the autopossessifying
+ table. */
+ #define PT_PXGRAPH   11    /* [:graph:] - characters that mark the paper */
+ #define PT_PXPRINT   12    /* [:print:] - [:graph:] plus non-control spaces */
+ #define PT_PXPUNCT   13    /* [:punct:] - punctuation characters */
  /* Flag bits and data types for the extended class (OP_XCLASS) for classes that
  contain characters with values greater than 255. */
-Line 1849 
 contain characters with values greater t
+Line 1878 
 contain characters with values greater t
  #define XCL_NOTPROP   4    /* Unicode inverted property (ditto) */
  /* 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
+ value such as \n. They must have non-zero values, as check_escape() returns 0
-for a data character.  Also, they must appear in the same order as in the opcode
+ for a data character.  Also, they must appear in the same order as in the
- definitions below, up to ESC_z. There's a dummy for OP_ALLANY because it
+ opcode definitions below, up to ESC_z. There's a dummy for OP_ALLANY because it
  corresponds to "." in DOTALL mode rather than an escape sequence. It is also
  used for [^] in JavaScript compatibility mode, and for \C in non-utf mode. In
  non-DOTALL mode, "." behaves like \N.
-Line 1874 
 enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, E
+Line 1903 
 enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, E
         ESC_E, ESC_Q, ESC_g, ESC_k,
         ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu };
- /* Opcode table: Starting from 1 (i.e. after OP_END), the values up to
- OP_EOD must correspond in order to the list of escapes immediately above.
- *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
+ /********************** Opcode definitions ******************/
- that follow must also be updated to match. There are also tables called
- "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
+ /****** NOTE NOTE NOTE ******
+ Starting from 1 (i.e. after OP_END), the values up to OP_EOD must correspond in
+ order to the list of escapes immediately above. Furthermore, values up to
+ OP_DOLLM must not be changed without adjusting the table called autoposstab in
+ pcre_compile.c
+ Whenever this list is updated, the two macro definitions that follow must be
+ updated to match. The possessification table called "opcode_possessify" in
+ pcre_compile.c must also be updated, and also the tables called "coptable"
+ and "poptable" in pcre_dfa_exec.c.
+ ****** NOTE NOTE NOTE ******/
+ /* The values between FIRST_AUTOTAB_OP and LAST_AUTOTAB_RIGHT_OP, inclusive,
+ are used in a table for deciding whether a repeated character type can be
+ auto-possessified. */
+ #define FIRST_AUTOTAB_OP       OP_NOT_DIGIT
+ #define LAST_AUTOTAB_LEFT_OP   OP_EXTUNI
+ #define LAST_AUTOTAB_RIGHT_OP  OP_DOLLM
  enum {
    OP_END,            /* 0 End of pattern */
-Line 1911 
 enum {
+Line 1959 
 enum {
    OP_EXTUNI,         /* 22 \X (extended Unicode sequence */
    OP_EODN,           /* 23 End of data or \n at end of data (\Z) */
    OP_EOD,            /* 24 End of data (\z) */
+   /* Line end assertions */
-   OP_CIRC,           /* 25 Start of line - not multiline */
+   OP_DOLL,           /* 25 End of line - not multiline */
-   OP_CIRCM,          /* 26 Start of line - multiline */
+   OP_DOLLM,          /* 26 End of line - multiline */
-   OP_DOLL,           /* 27 End of line - not multiline */
+   OP_CIRC,           /* 27 Start of line - not multiline */
-   OP_DOLLM,          /* 28 End of line - multiline */
+   OP_CIRCM,          /* 28 Start of line - multiline */
+   /* Single characters; caseful must precede the caseless ones */
    OP_CHAR,           /* 29 Match one character, casefully */
    OP_CHARI,          /* 30 Match one character, caselessly */
    OP_NOT,            /* 31 Match one character, not the given one, casefully */
-Line 1924 
 enum {
+Line 1977 
 enum {
    /* The following sets of 13 opcodes must always be kept in step because
    the offset from the first one is used to generate the others. */
-   /**** Single characters, caseful, must precede the caseless ones ****/
+   /* Repeated characters; caseful must precede the caseless ones */
    OP_STAR,           /* 33 The maximizing and minimizing versions of */
    OP_MINSTAR,        /* 34 these six opcodes must come in pairs, with */
-Line 1942 
 enum {
+Line 1995 
 enum {
    OP_POSQUERY,       /* 44 Posesssified query, caseful */
    OP_POSUPTO,        /* 45 Possessified upto, caseful */
-   /**** Single characters, caseless, must follow the caseful ones */
+   /* Repeated characters; caseless must follow the caseful ones */
    OP_STARI,          /* 46 */
    OP_MINSTARI,       /* 47 */
-Line 1960 
 enum {
+Line 2013 
 enum {
    OP_POSQUERYI,      /* 57 Posesssified query, caseless */
    OP_POSUPTOI,       /* 58 Possessified upto, caseless */
-   /**** The negated ones must follow the non-negated ones, and match them ****/
+   /* The negated ones must follow the non-negated ones, and match them */
-   /**** Negated single character, caseful; must precede the caseless ones ****/
+   /* Negated repeated character, caseful; must precede the caseless ones */
    OP_NOTSTAR,        /* 59 The maximizing and minimizing versions of */
    OP_NOTMINSTAR,     /* 60 these six opcodes must come in pairs, with */
-Line 1979 
 enum {
+Line 2032 
 enum {
    OP_NOTPOSQUERY,    /* 70 */
    OP_NOTPOSUPTO,     /* 71 */
-   /**** Negated single character, caseless; must follow the caseful ones ****/
+   /* Negated repeated character, caseless; must follow the caseful ones */
    OP_NOTSTARI,       /* 72 */
    OP_NOTMINSTARI,    /* 73 */
-Line 1997 
 enum {
+Line 2050 
 enum {
    OP_NOTPOSQUERYI,   /* 83 */
    OP_NOTPOSUPTOI,    /* 84 */
-   /**** Character types ****/
+   /* Character types */
    OP_TYPESTAR,       /* 85 The maximizing and minimizing versions of */
    OP_TYPEMINSTAR,    /* 86 these six opcodes must come in pairs, with */
-Line 2028 
 enum {
+Line 2081 
 enum {
    OP_CRRANGE,        /* 104 These are different to the three sets above. */
    OP_CRMINRANGE,     /* 105 */
+   OP_CRPOSSTAR,      /* 106 Possessified versions */
+   OP_CRPOSPLUS,      /* 107 */
+   OP_CRPOSQUERY,     /* 108 */
+   OP_CRPOSRANGE,     /* 109 */
    /* End of quantifier opcodes */
-   OP_CLASS,          /* 106 Match a character class, chars < 256 only */
+   OP_CLASS,          /* 110 Match a character class, chars < 256 only */
-   OP_NCLASS,         /* 107 Same, but the bitmap was created from a negative
+   OP_NCLASS,         /* 111 Same, but the bitmap was created from a negative
                                class - the difference is relevant only when a
                                character > 255 is encountered. */
-   OP_XCLASS,         /* 108 Extended class for handling > 255 chars within the
+   OP_XCLASS,         /* 112 Extended class for handling > 255 chars within the
                                class. This does both positive and negative. */
-   OP_REF,            /* 109 Match a back reference, casefully */
+   OP_REF,            /* 113 Match a back reference, casefully */
-   OP_REFI,           /* 110 Match a back reference, caselessly */
+   OP_REFI,           /* 114 Match a back reference, caselessly */
-   OP_RECURSE,        /* 111 Match a numbered subpattern (possibly recursive) */
+   OP_DNREF,          /* 115 Match a duplicate name backref, casefully */
-   OP_CALLOUT,        /* 112 Call out to external function if provided */
+   OP_DNREFI,         /* 116 Match a duplicate name backref, caselessly */
+   OP_RECURSE,        /* 117 Match a numbered subpattern (possibly recursive) */
-   OP_ALT,            /* 113 Start of alternation */
+   OP_CALLOUT,        /* 118 Call out to external function if provided */
-   OP_KET,            /* 114 End of group that doesn't have an unbounded repeat */
-   OP_KETRMAX,        /* 115 These two must remain together and in this */
+   OP_ALT,            /* 119 Start of alternation */
-   OP_KETRMIN,        /* 116 order. They are for groups the repeat for ever. */
+   OP_KET,            /* 120 End of group that doesn't have an unbounded repeat */
-   OP_KETRPOS,        /* 117 Possessive unlimited repeat. */
+   OP_KETRMAX,        /* 121 These two must remain together and in this */
+   OP_KETRMIN,        /* 122 order. They are for groups the repeat for ever. */
+   OP_KETRPOS,        /* 123 Possessive unlimited repeat. */
    /* The assertions must come before BRA, CBRA, ONCE, and COND, and the four
    asserts must remain in order. */
-   OP_REVERSE,        /* 118 Move pointer back - used in lookbehind assertions */
+   OP_REVERSE,        /* 124 Move pointer back - used in lookbehind assertions */
-   OP_ASSERT,         /* 119 Positive lookahead */
+   OP_ASSERT,         /* 125 Positive lookahead */
-   OP_ASSERT_NOT,     /* 120 Negative lookahead */
+   OP_ASSERT_NOT,     /* 126 Negative lookahead */
-   OP_ASSERTBACK,     /* 121 Positive lookbehind */
+   OP_ASSERTBACK,     /* 127 Positive lookbehind */
-   OP_ASSERTBACK_NOT, /* 122 Negative lookbehind */
+   OP_ASSERTBACK_NOT, /* 128 Negative lookbehind */
    /* ONCE, ONCE_NC, BRA, BRAPOS, CBRA, CBRAPOS, and COND must come immediately
    after the assertions, with ONCE first, as there's a test for >= ONCE for a
    subpattern that isn't an assertion. The POS versions must immediately follow
    the non-POS versions in each case. */
-   OP_ONCE,           /* 123 Atomic group, contains captures */
+   OP_ONCE,           /* 129 Atomic group, contains captures */
-   OP_ONCE_NC,        /* 124 Atomic group containing no captures */
+   OP_ONCE_NC,        /* 130 Atomic group containing no captures */
-   OP_BRA,            /* 125 Start of non-capturing bracket */
+   OP_BRA,            /* 131 Start of non-capturing bracket */
-   OP_BRAPOS,         /* 126 Ditto, with unlimited, possessive repeat */
+   OP_BRAPOS,         /* 132 Ditto, with unlimited, possessive repeat */
-   OP_CBRA,           /* 127 Start of capturing bracket */
+   OP_CBRA,           /* 133 Start of capturing bracket */
-   OP_CBRAPOS,        /* 128 Ditto, with unlimited, possessive repeat */
+   OP_CBRAPOS,        /* 134 Ditto, with unlimited, possessive repeat */
-   OP_COND,           /* 129 Conditional group */
+   OP_COND,           /* 135 Conditional group */
    /* These five must follow the previous five, in the same order. There's a
    check for >= SBRA to distinguish the two sets. */
-   OP_SBRA,           /* 130 Start of non-capturing bracket, check empty  */
+   OP_SBRA,           /* 136 Start of non-capturing bracket, check empty  */
-   OP_SBRAPOS,        /* 131 Ditto, with unlimited, possessive repeat */
+   OP_SBRAPOS,        /* 137 Ditto, with unlimited, possessive repeat */
-   OP_SCBRA,          /* 132 Start of capturing bracket, check empty */
+   OP_SCBRA,          /* 138 Start of capturing bracket, check empty */
-   OP_SCBRAPOS,       /* 133 Ditto, with unlimited, possessive repeat */
+   OP_SCBRAPOS,       /* 139 Ditto, with unlimited, possessive repeat */
-   OP_SCOND,          /* 134 Conditional group, check empty */
+   OP_SCOND,          /* 140 Conditional group, check empty */
    /* The next two pairs must (respectively) be kept together. */
-   OP_CREF,           /* 135 Used to hold a capture number as condition */
+   OP_CREF,           /* 141 Used to hold a capture number as condition */
-   OP_NCREF,          /* 136 Same, but generated by a name reference*/
+   OP_DNCREF,         /* 142 Used to point to duplicate names as a condition */
-   OP_RREF,           /* 137 Used to hold a recursion number as condition */
+   OP_RREF,           /* 143 Used to hold a recursion number as condition */
-   OP_NRREF,          /* 138 Same, but generated by a name reference*/
+   OP_DNRREF,         /* 144 Used to point to duplicate names as a condition */
-   OP_DEF,            /* 139 The DEFINE condition */
+   OP_DEF,            /* 145 The DEFINE condition */
-   OP_BRAZERO,        /* 140 These two must remain together and in this */
+   OP_BRAZERO,        /* 146 These two must remain together and in this */
-   OP_BRAMINZERO,     /* 141 order. */
+   OP_BRAMINZERO,     /* 147 order. */
-   OP_BRAPOSZERO,     /* 142 */
+   OP_BRAPOSZERO,     /* 148 */
    /* These are backtracking control verbs */
-   OP_MARK,           /* 143 always has an argument */
+   OP_MARK,           /* 149 always has an argument */
-   OP_PRUNE,          /* 144 */
+   OP_PRUNE,          /* 150 */
-   OP_PRUNE_ARG,      /* 145 same, but with argument */
+   OP_PRUNE_ARG,      /* 151 same, but with argument */
-   OP_SKIP,           /* 146 */
+   OP_SKIP,           /* 152 */
-   OP_SKIP_ARG,       /* 147 same, but with argument */
+   OP_SKIP_ARG,       /* 153 same, but with argument */
-   OP_THEN,           /* 148 */
+   OP_THEN,           /* 154 */
-   OP_THEN_ARG,       /* 149 same, but with argument */
+   OP_THEN_ARG,       /* 155 same, but with argument */
-   OP_COMMIT,         /* 150 */
+   OP_COMMIT,         /* 156 */
    /* These are forced failure and success verbs */
-   OP_FAIL,           /* 151 */
+   OP_FAIL,           /* 157 */
-   OP_ACCEPT,         /* 152 */
+   OP_ACCEPT,         /* 158 */
-   OP_ASSERT_ACCEPT,  /* 153 Used inside assertions */
+   OP_ASSERT_ACCEPT,  /* 159 Used inside assertions */
-   OP_CLOSE,          /* 154 Used before OP_ACCEPT to close open captures */
+   OP_CLOSE,          /* 160 Used before OP_ACCEPT to close open captures */
    /* This is used to skip a subpattern with a {0} quantifier */
-   OP_SKIPZERO,       /* 155 */
+   OP_SKIPZERO,       /* 161 */
    /* This is not an opcode, but is used to check that tables indexed by opcode
    are the correct length, in order to catch updating errors - there have been
-Line 2121 
 enum {
+Line 2181 
 enum {
  /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
  definitions that follow must also be updated to match. There are also tables
- called "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
+ called "opcode_possessify" in pcre_compile.c and "coptable" and "poptable" in
+ pcre_dfa_exec.c that must be updated. */
  /* This macro defines textual names for all the opcodes. These are used only
-Line 2134 
 some cases doesn't actually use these na
+Line 2195 
 some cases doesn't actually use these na
    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
    "extuni",  "\\Z", "\\z",                                        \
-   "^", "^", "$", "$", "char", "chari", "not", "noti",             \
+   "$", "$", "^", "^", "char", "chari", "not", "noti",             \
    "*", "*?", "+", "+?", "?", "??",                                \
    "{", "{", "{",                                                  \
    "*+","++", "?+", "{",                                           \
-Line 2150 
 some cases doesn't actually use these na
+Line 2211 
 some cases doesn't actually use these na
    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
    "*+","++", "?+", "{",                                           \
    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
-   "class", "nclass", "xclass", "Ref", "Refi",                     \
+   "*+","++", "?+", "{",                                           \
+   "class", "nclass", "xclass", "Ref", "Refi", "DnRef", "DnRefi",  \
    "Recurse", "Callout",                                           \
    "Alt", "Ket", "KetRmax", "KetRmin", "KetRpos",                  \
    "Reverse", "Assert", "Assert not", "AssertB", "AssertB not",    \
-Line 2159 
 some cases doesn't actually use these na
+Line 2221 
 some cases doesn't actually use these na
    "Cond",                                                         \
    "SBra", "SBraPos", "SCBra", "SCBraPos",                         \
    "SCond",                                                        \
-   "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \
+   "Cond ref", "Cond dnref", "Cond rec", "Cond dnrec", "Cond def", \
    "Brazero", "Braminzero", "Braposzero",                          \
    "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \
    "*THEN", "*THEN", "*COMMIT", "*FAIL",                           \
-Line 2184 
 in UTF-8 mode. The code that uses this t
+Line 2246 
 in UTF-8 mode. The code that uses this t
 , 3,                          /* \P, \p                                 */ \
 , 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
 ,                             /* \X                                     */ \
-, 1, 1, 1, 1, 1,              /* \Z, \z, ^, ^M, $, $M                   */ \
+, 1, 1, 1, 1, 1,              /* \Z, \z, $, $M ^, ^M                    */ \
 ,                             /* Char  - the minimum length             */ \
 ,                             /* Chari  - the minimum length            */ \
 ,                             /* not                                    */ \
-Line 2215 
 in UTF-8 mode. The code that uses this t
+Line 2277 
 in UTF-8 mode. The code that uses this t
    /* Character class & ref repeats                                         */ \
 , 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \
 +2*IMM2_SIZE, 1+2*IMM2_SIZE,  /* CRRANGE, CRMINRANGE                    */ \
+, 1, 1, 1+2*IMM2_SIZE,        /* Possessive *+, ++, ?+, CRPOSRANGE      */ \
 +(32/sizeof(pcre_uchar)),     /* CLASS                                  */ \
 +(32/sizeof(pcre_uchar)),     /* NCLASS                                 */ \
 ,                             /* XCLASS - variable length               */ \
 +IMM2_SIZE,                   /* REF                                    */ \
 +IMM2_SIZE,                   /* REFI                                   */ \
++2*IMM2_SIZE,                 /* DNREF                                  */ \
++2*IMM2_SIZE,                 /* DNREFI                                 */ \
 +LINK_SIZE,                   /* RECURSE                                */ \
 +2*LINK_SIZE,                 /* CALLOUT                                */ \
 +LINK_SIZE,                   /* Alt                                    */ \
-Line 2244 
 in UTF-8 mode. The code that uses this t
+Line 2309 
 in UTF-8 mode. The code that uses this t
 +LINK_SIZE+IMM2_SIZE,         /* SCBRA                                  */ \
 +LINK_SIZE+IMM2_SIZE,         /* SCBRAPOS                               */ \
 +LINK_SIZE,                   /* SCOND                                  */ \
-+IMM2_SIZE, 1+IMM2_SIZE,      /* CREF, NCREF                            */ \
++IMM2_SIZE, 1+2*IMM2_SIZE,    /* CREF, DNCREF                           */ \
-+IMM2_SIZE, 1+IMM2_SIZE,      /* RREF, NRREF                            */ \
++IMM2_SIZE, 1+2*IMM2_SIZE,    /* RREF, DNRREF                           */ \
 ,                             /* DEF                                    */ \
 , 1, 1,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */ \
 , 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \
-Line 2254 
 in UTF-8 mode. The code that uses this t
+Line 2319 
 in UTF-8 mode. The code that uses this t
 , 1, 1, 1,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */ \
 +IMM2_SIZE, 1                 /* CLOSE, SKIPZERO                        */
- /* A magic value for OP_RREF and OP_NRREF to indicate the "any recursion"
+ /* A magic value for OP_RREF to indicate the "any recursion" condition. */
- condition. */
  #define RREF_ANY  0xffff
-Line 2270 
 enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
+Line 2334 
 enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,
-        ERR70, ERR71, ERR72, ERR73, ERR74, ERR75, ERR76, ERR77, ERRCOUNT };
+        ERR70, ERR71, ERR72, ERR73, ERR74, ERR75, ERR76, ERR77, ERR78, ERR79,
+        ERR80, ERR81, ERRCOUNT };
  /* JIT compiling modes. The function list is indexed by them. */
  enum { JIT_COMPILE, JIT_PARTIAL_SOFT_COMPILE, JIT_PARTIAL_HARD_COMPILE,
-Line 2280 
 enum { JIT_COMPILE, JIT_PARTIAL_SOFT_COM
+Line 2345 
 enum { JIT_COMPILE, JIT_PARTIAL_SOFT_COM
  code vector run on as long as necessary after the end. We store an explicit
  offset to the name table so that if a regex is compiled on one host, saved, and
  then run on another where the size of pointers is different, all might still
- be well. For the case of compiled-on-4 and run-on-8, we include an extra
+ be well.
- pointer that is always NULL. For future-proofing, a few dummy fields were
- originally included - even though you can never get this planning right - but
- there is only one left now.
- NOTE NOTE NOTE:
- Because people can now save and re-use compiled patterns, any additions to this
- structure should be made at the end, and something earlier (e.g. a new
- flag in the options or one of the dummy fields) should indicate that the new
- fields are present. Currently PCRE always sets the dummy fields to zero.
- NOTE NOTE NOTE
- */
- #if defined COMPILE_PCRE8
+ The size of the structure must be a multiple of 8 bytes. For the case of
- #define REAL_PCRE real_pcre
+ compiled-on-4 and run-on-8, we include an extra pointer that is always NULL so
- #elif defined COMPILE_PCRE16
+ that there are an even number of pointers which therefore are a multiple of 8
- #define REAL_PCRE real_pcre16
+ bytes.
- #elif defined COMPILE_PCRE32
- #define REAL_PCRE real_pcre32
+ It is necessary to fork the struct for the 32 bit library, since it needs to
- #endif
+ use pcre_uint32 for first_char and req_char. We can't put an ifdef inside the
+ typedef because pcretest needs access to the struct of the 8-, 16- and 32-bit
- /* It is necessary to fork the struct for 32 bit, since it needs to use
+ variants.
-  * pcre_uchar for first_char and req_char. Can't put an ifdef inside the
-  * typedef since pcretest needs access to  the struct of the 8-, 16-
+ *** WARNING ***
-  * and 32-bit variants. */
+ When new fields are added to these structures, remember to adjust the code in
+ pcre_byte_order.c that is concerned with swapping the byte order of the fields
+ when a compiled regex is reloaded on a host with different endianness.
+ *** WARNING ***
+ There is also similar byte-flipping code in pcretest.c, which is used for
+ testing the byte-flipping features. It must also be kept in step.
+ *** WARNING ***
+ */
  typedef struct real_pcre8_or_16 {
    pcre_uint32 magic_number;
    pcre_uint32 size;               /* Total that was malloced */
    pcre_uint32 options;            /* Public options */
-   pcre_uint16 flags;              /* Private flags */
+   pcre_uint32 flags;              /* Private flags */
+   pcre_uint32 limit_match;        /* Limit set from regex */
+   pcre_uint32 limit_recursion;    /* Limit set from regex */
+   pcre_uint16 first_char;         /* Starting character */
+   pcre_uint16 req_char;           /* This character must be seen */
    pcre_uint16 max_lookbehind;     /* Longest lookbehind (characters) */
    pcre_uint16 top_bracket;        /* Highest numbered group */
    pcre_uint16 top_backref;        /* Highest numbered back reference */
-   pcre_uint16 first_char;         /* Starting character */
-   pcre_uint16 req_char;           /* This character must be seen */
    pcre_uint16 name_table_offset;  /* Offset to name table that follows */
    pcre_uint16 name_entry_size;    /* Size of any name items */
    pcre_uint16 name_count;         /* Number of name items */
    pcre_uint16 ref_count;          /* Reference count */
+   pcre_uint16 dummy1;             /* To ensure size is a multiple of 8 */
+   pcre_uint16 dummy2;             /* To ensure size is a multiple of 8 */
+   pcre_uint16 dummy3;             /* To ensure size is a multiple of 8 */
    const pcre_uint8 *tables;       /* Pointer to tables or NULL for std */
-   const pcre_uint8 *nullpad;      /* NULL padding */
+   void             *nullpad;      /* NULL padding */
  } real_pcre8_or_16;
  typedef struct real_pcre8_or_16 real_pcre;
-Line 2331 
 typedef struct real_pcre32 {
+Line 2397 
 typedef struct real_pcre32 {
    pcre_uint32 magic_number;
    pcre_uint32 size;               /* Total that was malloced */
    pcre_uint32 options;            /* Public options */
-   pcre_uint16 flags;              /* Private flags */
+   pcre_uint32 flags;              /* Private flags */
+   pcre_uint32 limit_match;        /* Limit set from regex */
+   pcre_uint32 limit_recursion;    /* Limit set from regex */
+   pcre_uint32 first_char;         /* Starting character */
+   pcre_uint32 req_char;           /* This character must be seen */
    pcre_uint16 max_lookbehind;     /* Longest lookbehind (characters) */
    pcre_uint16 top_bracket;        /* Highest numbered group */
    pcre_uint16 top_backref;        /* Highest numbered back reference */
-   pcre_uint32 first_char;         /* Starting character */
-   pcre_uint32 req_char;           /* This character must be seen */
    pcre_uint16 name_table_offset;  /* Offset to name table that follows */
    pcre_uint16 name_entry_size;    /* Size of any name items */
    pcre_uint16 name_count;         /* Number of name items */
    pcre_uint16 ref_count;          /* Reference count */
-   pcre_uint16 dummy1;             /* for later expansion */
+   pcre_uint16 dummy;              /* To ensure size is a multiple of 8 */
-   pcre_uint16 dummy2;             /* for later expansion */
    const pcre_uint8 *tables;       /* Pointer to tables or NULL for std */
-   void *nullpad;                  /* for later expansion */
+   void             *nullpad;      /* NULL padding */
  } real_pcre32;
+ #if defined COMPILE_PCRE8
+ #define REAL_PCRE real_pcre
+ #elif defined COMPILE_PCRE16
+ #define REAL_PCRE real_pcre16
+ #elif defined COMPILE_PCRE32
+ #define REAL_PCRE real_pcre32
+ #endif
  /* Assert that the size of REAL_PCRE is divisible by 8 */
  typedef int __assert_real_pcre_size_divisible_8[(sizeof(REAL_PCRE) % 8) == 0 ? 1 : -1];
-Line 2380 
 typedef struct open_capitem {
+Line 2455 
 typedef struct open_capitem {
    pcre_uint16 flag;             /* Set TRUE if recursive back ref */
  } open_capitem;
+ /* Structure for building a list of named groups during the first pass of
+ compiling. */
+ typedef struct named_group {
+   const pcre_uchar  *name;          /* Points to the name in the pattern */
+   int                length;        /* Length of the name */
+   pcre_uint32        number;        /* Group number */
+ } named_group;
  /* Structure for passing "static" information around between the functions
  doing the compiling, so that they are thread-safe. */
-Line 2392 
 typedef struct compile_data {
+Line 2476 
 typedef struct compile_data {
    const pcre_uchar *start_code;     /* The start of the compiled code */
    const pcre_uchar *start_pattern;  /* The start of the pattern */
    const pcre_uchar *end_pattern;    /* The end of the pattern */
-   open_capitem *open_caps;          /* Chain of open capture items */
    pcre_uchar *hwm;                  /* High watermark of workspace */
+   open_capitem *open_caps;          /* Chain of open capture items */
+   named_group *named_groups;        /* Points to vector in pre-compile */
    pcre_uchar *name_table;           /* The name/number table */
    int  names_found;                 /* Number of entries so far */
    int  name_entry_size;             /* Size of each entry */
+   int  named_group_list_size;       /* Number of entries in the list */
    int  workspace_size;              /* Size of workspace */
-   unsigned int  bracount;           /* Count of capturing parens as we compile */
+   unsigned int bracount;            /* Count of capturing parens as we compile */
    int  final_bracount;              /* Saved value after first pass */
    int  max_lookbehind;              /* Maximum lookbehind (characters) */
    int  top_backref;                 /* Maximum back reference */
    unsigned int backref_map;         /* Bitmap of low back refs */
+   unsigned int namedrefcount;       /* Number of backreferences by name */
    int  assert_depth;                /* Depth of nested assertions */
-   int  external_options;            /* External (initial) options */
+   pcre_uint32 external_options;     /* External (initial) options */
-   int  external_flags;              /* External flag bits to be set */
+   pcre_uint32 external_flags;       /* External flag bits to be set */
    int  req_varyopt;                 /* "After variable item" flag for reqbyte */
    BOOL had_accept;                  /* (*ACCEPT) encountered */
    BOOL had_pruneorskip;             /* (*PRUNE) or (*SKIP) encountered */
    BOOL check_lookbehind;            /* Lookbehinds need later checking */
+   BOOL dupnames;                    /* Duplicate names exist */
    int  nltype;                      /* Newline type */
    int  nllen;                       /* Newline string length */
    pcre_uchar nl[4];                 /* Newline string when fixed length */
-Line 2431 
 typedef struct recursion_info {
+Line 2519 
 typedef struct recursion_info {
    unsigned int group_num;         /* Number of group that was called */
    int *offset_save;               /* Pointer to start of saved offsets */
    int saved_max;                  /* Number of saved offsets */
    int saved_capture_last;         /* Last capture number */
    PCRE_PUCHAR subject_position;   /* Position at start of recursion */
  } recursion_info;
-Line 2468 
 typedef struct match_data {
+Line 2556 
 typedef struct match_data {
    int    nllen;                   /* Newline string length */
    int    name_count;              /* Number of names in name table */
    int    name_entry_size;         /* Size of entry in names table */
+   unsigned int skip_arg_count;    /* For counting SKIP_ARGs */
+   unsigned int ignore_skip_arg;   /* For re-run when SKIP arg name not found */
    pcre_uchar *name_table;         /* Table of names */
    pcre_uchar nl[4];               /* Newline string when fixed */
    const  pcre_uint8 *lcc;         /* Points to lower casing table */
-Line 2484 
 typedef struct match_data {
+Line 2574 
 typedef struct match_data {
    BOOL   hitend;                  /* Hit the end of the subject at some point */
    BOOL   bsr_anycrlf;             /* \R is just any CRLF, not full Unicode */
    BOOL   hasthen;                 /* Pattern contains (*THEN) */
-   BOOL   ignore_skip_arg;         /* For re-run when SKIP name not found */
    const  pcre_uchar *start_code;  /* For use when recursing */
    PCRE_PUCHAR start_subject;      /* Start of the subject string */
    PCRE_PUCHAR end_subject;        /* End of the subject string */

 Legend:



Removed from v.1248
 


changed lines


 
Added in v.1387
 Legend:



Removed from v.1248
 


changed lines


 
Added in v.1387
-Removed from v.1248
+Added in v.1387

	ViewVC Help
Powered by ViewVC 1.1.5