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

Diff of /code/trunk/pcre_internal.h

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 1361 by ph10, Fri Sep 6 17:47:32 2013 UTC revision 1392 by ph10, Wed Nov 6 18:00:09 2013 UTC
# Line 1174  time, run time, or study time, respectiv Line 1174  time, run time, or study time, respectiv
1174  #define PUBLIC_COMPILE_OPTIONS \  #define PUBLIC_COMPILE_OPTIONS \
1175    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
1176     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
1177     PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \     PCRE_NO_AUTO_CAPTURE|PCRE_NO_AUTO_POSSESSIFY| \
1178       PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \
1179     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
1180     PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE|PCRE_NEVER_UTF)     PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE|PCRE_NEVER_UTF)
1181    
# Line 1852  only. */ Line 1853  only. */
1853  #define PT_WORD       8    /* Word - L plus N plus underscore */  #define PT_WORD       8    /* Word - L plus N plus underscore */
1854  #define PT_CLIST      9    /* Pseudo-property: match character list */  #define PT_CLIST      9    /* Pseudo-property: match character list */
1855  #define PT_UCNC      10    /* Universal Character nameable character */  #define PT_UCNC      10    /* Universal Character nameable character */
1856    #define PT_TABSIZE   11    /* Size of square table for autopossessify tests */
1857    
1858    /* The following special properties are used only in XCLASS items, when POSIX
1859    classes are specified and PCRE_UCP is set - in other words, for Unicode
1860    handling of these classes. They are not available via the \p or \P escapes like
1861    those in the above list, and so they do not take part in the autopossessifying
1862    table. */
1863    
1864    #define PT_PXGRAPH   11    /* [:graph:] - characters that mark the paper */
1865    #define PT_PXPRINT   12    /* [:print:] - [:graph:] plus non-control spaces */
1866    #define PT_PXPUNCT   13    /* [:punct:] - punctuation characters */
1867    
1868  /* Flag bits and data types for the extended class (OP_XCLASS) for classes that  /* Flag bits and data types for the extended class (OP_XCLASS) for classes that
1869  contain characters with values greater than 255. */  contain characters with values greater than 255. */
# Line 1866  contain characters with values greater t Line 1878  contain characters with values greater t
1878  #define XCL_NOTPROP   4    /* Unicode inverted property (ditto) */  #define XCL_NOTPROP   4    /* Unicode inverted property (ditto) */
1879    
1880  /* These are escaped items that aren't just an encoding of a particular data  /* These are escaped items that aren't just an encoding of a particular data
1881  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
1882  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
1883  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
1884  corresponds to "." in DOTALL mode rather than an escape sequence. It is also  corresponds to "." in DOTALL mode rather than an escape sequence. It is also
1885  used for [^] in JavaScript compatibility mode, and for \C in non-utf mode. In  used for [^] in JavaScript compatibility mode, and for \C in non-utf mode. In
1886  non-DOTALL mode, "." behaves like \N.  non-DOTALL mode, "." behaves like \N.
# Line 1891  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, E Line 1903  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, E
1903         ESC_E, ESC_Q, ESC_g, ESC_k,         ESC_E, ESC_Q, ESC_g, ESC_k,
1904         ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu };         ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu };
1905    
 /* 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.  
1906    
1907  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  /********************** Opcode definitions ******************/
1908  that follow must also be updated to match. There are also tables called  
1909  "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */  /****** NOTE NOTE NOTE ******
1910    
1911    Starting from 1 (i.e. after OP_END), the values up to OP_EOD must correspond in
1912    order to the list of escapes immediately above. Furthermore, values up to
1913    OP_DOLLM must not be changed without adjusting the table called autoposstab in
1914    pcre_compile.c
1915    
1916    Whenever this list is updated, the two macro definitions that follow must be
1917    updated to match. The possessification table called "opcode_possessify" in
1918    pcre_compile.c must also be updated, and also the tables called "coptable"
1919    and "poptable" in pcre_dfa_exec.c.
1920    
1921    ****** NOTE NOTE NOTE ******/
1922    
1923    
1924    /* The values between FIRST_AUTOTAB_OP and LAST_AUTOTAB_RIGHT_OP, inclusive,
1925    are used in a table for deciding whether a repeated character type can be
1926    auto-possessified. */
1927    
1928    #define FIRST_AUTOTAB_OP       OP_NOT_DIGIT
1929    #define LAST_AUTOTAB_LEFT_OP   OP_EXTUNI
1930    #define LAST_AUTOTAB_RIGHT_OP  OP_DOLLM
1931    
1932  enum {  enum {
1933    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 1928  enum { Line 1959  enum {
1959    OP_EXTUNI,         /* 22 \X (extended Unicode sequence */    OP_EXTUNI,         /* 22 \X (extended Unicode sequence */
1960    OP_EODN,           /* 23 End of data or \n at end of data (\Z) */    OP_EODN,           /* 23 End of data or \n at end of data (\Z) */
1961    OP_EOD,            /* 24 End of data (\z) */    OP_EOD,            /* 24 End of data (\z) */
1962    
1963      /* Line end assertions */
1964    
1965    OP_CIRC,           /* 25 Start of line - not multiline */    OP_DOLL,           /* 25 End of line - not multiline */
1966    OP_CIRCM,          /* 26 Start of line - multiline */    OP_DOLLM,          /* 26 End of line - multiline */
1967    OP_DOLL,           /* 27 End of line - not multiline */    OP_CIRC,           /* 27 Start of line - not multiline */
1968    OP_DOLLM,          /* 28 End of line - multiline */    OP_CIRCM,          /* 28 Start of line - multiline */
1969    
1970      /* Single characters; caseful must precede the caseless ones */
1971    
1972    OP_CHAR,           /* 29 Match one character, casefully */    OP_CHAR,           /* 29 Match one character, casefully */
1973    OP_CHARI,          /* 30 Match one character, caselessly */    OP_CHARI,          /* 30 Match one character, caselessly */
1974    OP_NOT,            /* 31 Match one character, not the given one, casefully */    OP_NOT,            /* 31 Match one character, not the given one, casefully */
# Line 1941  enum { Line 1977  enum {
1977    /* The following sets of 13 opcodes must always be kept in step because    /* The following sets of 13 opcodes must always be kept in step because
1978    the offset from the first one is used to generate the others. */    the offset from the first one is used to generate the others. */
1979    
1980    /**** Single characters, caseful, must precede the caseless ones ****/    /* Repeated characters; caseful must precede the caseless ones */
1981    
1982    OP_STAR,           /* 33 The maximizing and minimizing versions of */    OP_STAR,           /* 33 The maximizing and minimizing versions of */
1983    OP_MINSTAR,        /* 34 these six opcodes must come in pairs, with */    OP_MINSTAR,        /* 34 these six opcodes must come in pairs, with */
# Line 1959  enum { Line 1995  enum {
1995    OP_POSQUERY,       /* 44 Posesssified query, caseful */    OP_POSQUERY,       /* 44 Posesssified query, caseful */
1996    OP_POSUPTO,        /* 45 Possessified upto, caseful */    OP_POSUPTO,        /* 45 Possessified upto, caseful */
1997    
1998    /**** Single characters, caseless, must follow the caseful ones */    /* Repeated characters; caseless must follow the caseful ones */
1999    
2000    OP_STARI,          /* 46 */    OP_STARI,          /* 46 */
2001    OP_MINSTARI,       /* 47 */    OP_MINSTARI,       /* 47 */
# Line 1977  enum { Line 2013  enum {
2013    OP_POSQUERYI,      /* 57 Posesssified query, caseless */    OP_POSQUERYI,      /* 57 Posesssified query, caseless */
2014    OP_POSUPTOI,       /* 58 Possessified upto, caseless */    OP_POSUPTOI,       /* 58 Possessified upto, caseless */
2015    
2016    /**** The negated ones must follow the non-negated ones, and match them ****/    /* The negated ones must follow the non-negated ones, and match them */
2017    /**** Negated single character, caseful; must precede the caseless ones ****/    /* Negated repeated character, caseful; must precede the caseless ones */
2018    
2019    OP_NOTSTAR,        /* 59 The maximizing and minimizing versions of */    OP_NOTSTAR,        /* 59 The maximizing and minimizing versions of */
2020    OP_NOTMINSTAR,     /* 60 these six opcodes must come in pairs, with */    OP_NOTMINSTAR,     /* 60 these six opcodes must come in pairs, with */
# Line 1996  enum { Line 2032  enum {
2032    OP_NOTPOSQUERY,    /* 70 */    OP_NOTPOSQUERY,    /* 70 */
2033    OP_NOTPOSUPTO,     /* 71 */    OP_NOTPOSUPTO,     /* 71 */
2034    
2035    /**** Negated single character, caseless; must follow the caseful ones ****/    /* Negated repeated character, caseless; must follow the caseful ones */
2036    
2037    OP_NOTSTARI,       /* 72 */    OP_NOTSTARI,       /* 72 */
2038    OP_NOTMINSTARI,    /* 73 */    OP_NOTMINSTARI,    /* 73 */
# Line 2014  enum { Line 2050  enum {
2050    OP_NOTPOSQUERYI,   /* 83 */    OP_NOTPOSQUERYI,   /* 83 */
2051    OP_NOTPOSUPTOI,    /* 84 */    OP_NOTPOSUPTOI,    /* 84 */
2052    
2053    /**** Character types ****/    /* Character types */
2054    
2055    OP_TYPESTAR,       /* 85 The maximizing and minimizing versions of */    OP_TYPESTAR,       /* 85 The maximizing and minimizing versions of */
2056    OP_TYPEMINSTAR,    /* 86 these six opcodes must come in pairs, with */    OP_TYPEMINSTAR,    /* 86 these six opcodes must come in pairs, with */
# Line 2045  enum { Line 2081  enum {
2081    OP_CRRANGE,        /* 104 These are different to the three sets above. */    OP_CRRANGE,        /* 104 These are different to the three sets above. */
2082    OP_CRMINRANGE,     /* 105 */    OP_CRMINRANGE,     /* 105 */
2083    
2084      OP_CRPOSSTAR,      /* 106 Possessified versions */
2085      OP_CRPOSPLUS,      /* 107 */
2086      OP_CRPOSQUERY,     /* 108 */
2087      OP_CRPOSRANGE,     /* 109 */
2088    
2089    /* End of quantifier opcodes */    /* End of quantifier opcodes */
2090    
2091    OP_CLASS,          /* 106 Match a character class, chars < 256 only */    OP_CLASS,          /* 110 Match a character class, chars < 256 only */
2092    OP_NCLASS,         /* 107 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 111 Same, but the bitmap was created from a negative
2093                                class - the difference is relevant only when a                                class - the difference is relevant only when a
2094                                character > 255 is encountered. */                                character > 255 is encountered. */
2095    OP_XCLASS,         /* 108 Extended class for handling > 255 chars within the    OP_XCLASS,         /* 112 Extended class for handling > 255 chars within the
2096                                class. This does both positive and negative. */                                class. This does both positive and negative. */
2097    OP_REF,            /* 109 Match a back reference, casefully */    OP_REF,            /* 113 Match a back reference, casefully */
2098    OP_REFI,           /* 110 Match a back reference, caselessly */    OP_REFI,           /* 114 Match a back reference, caselessly */
2099    OP_DNREF,          /* 111 Match a duplicate name backref, casefully */    OP_DNREF,          /* 115 Match a duplicate name backref, casefully */
2100    OP_DNREFI,         /* 112 Match a duplicate name backref, caselessly */    OP_DNREFI,         /* 116 Match a duplicate name backref, caselessly */
2101    OP_RECURSE,        /* 113 Match a numbered subpattern (possibly recursive) */    OP_RECURSE,        /* 117 Match a numbered subpattern (possibly recursive) */
2102    OP_CALLOUT,        /* 114 Call out to external function if provided */    OP_CALLOUT,        /* 118 Call out to external function if provided */
2103    
2104    OP_ALT,            /* 115 Start of alternation */    OP_ALT,            /* 119 Start of alternation */
2105    OP_KET,            /* 116 End of group that doesn't have an unbounded repeat */    OP_KET,            /* 120 End of group that doesn't have an unbounded repeat */
2106    OP_KETRMAX,        /* 117 These two must remain together and in this */    OP_KETRMAX,        /* 121 These two must remain together and in this */
2107    OP_KETRMIN,        /* 118 order. They are for groups the repeat for ever. */    OP_KETRMIN,        /* 122 order. They are for groups the repeat for ever. */
2108    OP_KETRPOS,        /* 119 Possessive unlimited repeat. */    OP_KETRPOS,        /* 123 Possessive unlimited repeat. */
2109    
2110    /* The assertions must come before BRA, CBRA, ONCE, and COND, and the four    /* The assertions must come before BRA, CBRA, ONCE, and COND, and the four
2111    asserts must remain in order. */    asserts must remain in order. */
2112    
2113    OP_REVERSE,        /* 129 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 124 Move pointer back - used in lookbehind assertions */
2114    OP_ASSERT,         /* 121 Positive lookahead */    OP_ASSERT,         /* 125 Positive lookahead */
2115    OP_ASSERT_NOT,     /* 122 Negative lookahead */    OP_ASSERT_NOT,     /* 126 Negative lookahead */
2116    OP_ASSERTBACK,     /* 123 Positive lookbehind */    OP_ASSERTBACK,     /* 127 Positive lookbehind */
2117    OP_ASSERTBACK_NOT, /* 124 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 128 Negative lookbehind */
2118    
2119    /* ONCE, ONCE_NC, BRA, BRAPOS, CBRA, CBRAPOS, and COND must come immediately    /* ONCE, ONCE_NC, BRA, BRAPOS, CBRA, CBRAPOS, and COND must come immediately
2120    after the assertions, with ONCE first, as there's a test for >= ONCE for a    after the assertions, with ONCE first, as there's a test for >= ONCE for a
2121    subpattern that isn't an assertion. The POS versions must immediately follow    subpattern that isn't an assertion. The POS versions must immediately follow
2122    the non-POS versions in each case. */    the non-POS versions in each case. */
2123    
2124    OP_ONCE,           /* 125 Atomic group, contains captures */    OP_ONCE,           /* 129 Atomic group, contains captures */
2125    OP_ONCE_NC,        /* 126 Atomic group containing no captures */    OP_ONCE_NC,        /* 130 Atomic group containing no captures */
2126    OP_BRA,            /* 127 Start of non-capturing bracket */    OP_BRA,            /* 131 Start of non-capturing bracket */
2127    OP_BRAPOS,         /* 128 Ditto, with unlimited, possessive repeat */    OP_BRAPOS,         /* 132 Ditto, with unlimited, possessive repeat */
2128    OP_CBRA,           /* 129 Start of capturing bracket */    OP_CBRA,           /* 133 Start of capturing bracket */
2129    OP_CBRAPOS,        /* 130 Ditto, with unlimited, possessive repeat */    OP_CBRAPOS,        /* 134 Ditto, with unlimited, possessive repeat */
2130    OP_COND,           /* 131 Conditional group */    OP_COND,           /* 135 Conditional group */
2131    
2132    /* These five must follow the previous five, in the same order. There's a    /* These five must follow the previous five, in the same order. There's a
2133    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
2134    
2135    OP_SBRA,           /* 132 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 136 Start of non-capturing bracket, check empty  */
2136    OP_SBRAPOS,        /* 133 Ditto, with unlimited, possessive repeat */    OP_SBRAPOS,        /* 137 Ditto, with unlimited, possessive repeat */
2137    OP_SCBRA,          /* 134 Start of capturing bracket, check empty */    OP_SCBRA,          /* 138 Start of capturing bracket, check empty */
2138    OP_SCBRAPOS,       /* 135 Ditto, with unlimited, possessive repeat */    OP_SCBRAPOS,       /* 139 Ditto, with unlimited, possessive repeat */
2139    OP_SCOND,          /* 136 Conditional group, check empty */    OP_SCOND,          /* 140 Conditional group, check empty */
2140    
2141    /* The next two pairs must (respectively) be kept together. */    /* The next two pairs must (respectively) be kept together. */
2142    
2143    OP_CREF,           /* 137 Used to hold a capture number as condition */    OP_CREF,           /* 141 Used to hold a capture number as condition */
2144    OP_NCREF,          /* 138 Same, but generated by a name reference*/    OP_DNCREF,         /* 142 Used to point to duplicate names as a condition */
2145    OP_RREF,           /* 139 Used to hold a recursion number as condition */    OP_RREF,           /* 143 Used to hold a recursion number as condition */
2146    OP_NRREF,          /* 140 Same, but generated by a name reference*/    OP_DNRREF,         /* 144 Used to point to duplicate names as a condition */
2147    OP_DEF,            /* 141 The DEFINE condition */    OP_DEF,            /* 145 The DEFINE condition */
2148    
2149    OP_BRAZERO,        /* 142 These two must remain together and in this */    OP_BRAZERO,        /* 146 These two must remain together and in this */
2150    OP_BRAMINZERO,     /* 143 order. */    OP_BRAMINZERO,     /* 147 order. */
2151    OP_BRAPOSZERO,     /* 144 */    OP_BRAPOSZERO,     /* 148 */
2152    
2153    /* These are backtracking control verbs */    /* These are backtracking control verbs */
2154    
2155    OP_MARK,           /* 145 always has an argument */    OP_MARK,           /* 149 always has an argument */
2156    OP_PRUNE,          /* 146 */    OP_PRUNE,          /* 150 */
2157    OP_PRUNE_ARG,      /* 147 same, but with argument */    OP_PRUNE_ARG,      /* 151 same, but with argument */
2158    OP_SKIP,           /* 148 */    OP_SKIP,           /* 152 */
2159    OP_SKIP_ARG,       /* 149 same, but with argument */    OP_SKIP_ARG,       /* 153 same, but with argument */
2160    OP_THEN,           /* 150 */    OP_THEN,           /* 154 */
2161    OP_THEN_ARG,       /* 151 same, but with argument */    OP_THEN_ARG,       /* 155 same, but with argument */
2162    OP_COMMIT,         /* 152 */    OP_COMMIT,         /* 156 */
2163    
2164    /* These are forced failure and success verbs */    /* These are forced failure and success verbs */
2165    
2166    OP_FAIL,           /* 153 */    OP_FAIL,           /* 157 */
2167    OP_ACCEPT,         /* 154 */    OP_ACCEPT,         /* 158 */
2168    OP_ASSERT_ACCEPT,  /* 155 Used inside assertions */    OP_ASSERT_ACCEPT,  /* 159 Used inside assertions */
2169    OP_CLOSE,          /* 156 Used before OP_ACCEPT to close open captures */    OP_CLOSE,          /* 160 Used before OP_ACCEPT to close open captures */
2170    
2171    /* This is used to skip a subpattern with a {0} quantifier */    /* This is used to skip a subpattern with a {0} quantifier */
2172    
2173    OP_SKIPZERO,       /* 157 */    OP_SKIPZERO,       /* 161 */
2174    
2175    /* This is not an opcode, but is used to check that tables indexed by opcode    /* This is not an opcode, but is used to check that tables indexed by opcode
2176    are the correct length, in order to catch updating errors - there have been    are the correct length, in order to catch updating errors - there have been
# Line 2140  enum { Line 2181  enum {
2181    
2182  /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro  /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
2183  definitions that follow must also be updated to match. There are also tables  definitions that follow must also be updated to match. There are also tables
2184  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
2185    pcre_dfa_exec.c that must be updated. */
2186    
2187    
2188  /* This macro defines textual names for all the opcodes. These are used only  /* This macro defines textual names for all the opcodes. These are used only
# Line 2153  some cases doesn't actually use these na Line 2195  some cases doesn't actually use these na
2195    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
2196    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
2197    "extuni",  "\\Z", "\\z",                                        \    "extuni",  "\\Z", "\\z",                                        \
2198    "^", "^", "$", "$", "char", "chari", "not", "noti",             \    "$", "$", "^", "^", "char", "chari", "not", "noti",             \
2199    "*", "*?", "+", "+?", "?", "??",                                \    "*", "*?", "+", "+?", "?", "??",                                \
2200    "{", "{", "{",                                                  \    "{", "{", "{",                                                  \
2201    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
# Line 2169  some cases doesn't actually use these na Line 2211  some cases doesn't actually use these na
2211    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
2212    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
2213    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
2214      "*+","++", "?+", "{",                                           \
2215    "class", "nclass", "xclass", "Ref", "Refi", "DnRef", "DnRefi",  \    "class", "nclass", "xclass", "Ref", "Refi", "DnRef", "DnRefi",  \
2216    "Recurse", "Callout",                                           \    "Recurse", "Callout",                                           \
2217    "Alt", "Ket", "KetRmax", "KetRmin", "KetRpos",                  \    "Alt", "Ket", "KetRmax", "KetRmin", "KetRpos",                  \
# Line 2178  some cases doesn't actually use these na Line 2221  some cases doesn't actually use these na
2221    "Cond",                                                         \    "Cond",                                                         \
2222    "SBra", "SBraPos", "SCBra", "SCBraPos",                         \    "SBra", "SBraPos", "SCBra", "SCBraPos",                         \
2223    "SCond",                                                        \    "SCond",                                                        \
2224    "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \    "Cond ref", "Cond dnref", "Cond rec", "Cond dnrec", "Cond def", \
2225    "Brazero", "Braminzero", "Braposzero",                          \    "Brazero", "Braminzero", "Braposzero",                          \
2226    "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \    "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \
2227    "*THEN", "*THEN", "*COMMIT", "*FAIL",                           \    "*THEN", "*THEN", "*COMMIT", "*FAIL",                           \
# Line 2203  in UTF-8 mode. The code that uses this t Line 2246  in UTF-8 mode. The code that uses this t
2246    3, 3,                          /* \P, \p                                 */ \    3, 3,                          /* \P, \p                                 */ \
2247    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
2248    1,                             /* \X                                     */ \    1,                             /* \X                                     */ \
2249    1, 1, 1, 1, 1, 1,              /* \Z, \z, ^, ^M, $, $M                   */ \    1, 1, 1, 1, 1, 1,              /* \Z, \z, $, $M ^, ^M                    */ \
2250    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
2251    2,                             /* Chari  - the minimum length            */ \    2,                             /* Chari  - the minimum length            */ \
2252    2,                             /* not                                    */ \    2,                             /* not                                    */ \
# Line 2234  in UTF-8 mode. The code that uses this t Line 2277  in UTF-8 mode. The code that uses this t
2277    /* Character class & ref repeats                                         */ \    /* Character class & ref repeats                                         */ \
2278    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \
2279    1+2*IMM2_SIZE, 1+2*IMM2_SIZE,  /* CRRANGE, CRMINRANGE                    */ \    1+2*IMM2_SIZE, 1+2*IMM2_SIZE,  /* CRRANGE, CRMINRANGE                    */ \
2280      1, 1, 1, 1+2*IMM2_SIZE,        /* Possessive *+, ++, ?+, CRPOSRANGE      */ \
2281    1+(32/sizeof(pcre_uchar)),     /* CLASS                                  */ \    1+(32/sizeof(pcre_uchar)),     /* CLASS                                  */ \
2282    1+(32/sizeof(pcre_uchar)),     /* NCLASS                                 */ \    1+(32/sizeof(pcre_uchar)),     /* NCLASS                                 */ \
2283    0,                             /* XCLASS - variable length               */ \    0,                             /* XCLASS - variable length               */ \
# Line 2265  in UTF-8 mode. The code that uses this t Line 2309  in UTF-8 mode. The code that uses this t
2309    1+LINK_SIZE+IMM2_SIZE,         /* SCBRA                                  */ \    1+LINK_SIZE+IMM2_SIZE,         /* SCBRA                                  */ \
2310    1+LINK_SIZE+IMM2_SIZE,         /* SCBRAPOS                               */ \    1+LINK_SIZE+IMM2_SIZE,         /* SCBRAPOS                               */ \
2311    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
2312    1+IMM2_SIZE, 1+IMM2_SIZE,      /* CREF, NCREF                            */ \    1+IMM2_SIZE, 1+2*IMM2_SIZE,    /* CREF, DNCREF                           */ \
2313    1+IMM2_SIZE, 1+IMM2_SIZE,      /* RREF, NRREF                            */ \    1+IMM2_SIZE, 1+2*IMM2_SIZE,    /* RREF, DNRREF                           */ \
2314    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
2315    1, 1, 1,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */ \    1, 1, 1,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */ \
2316    3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \    3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \
# Line 2275  in UTF-8 mode. The code that uses this t Line 2319  in UTF-8 mode. The code that uses this t
2319    1, 1, 1, 1,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */ \    1, 1, 1, 1,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */ \
2320    1+IMM2_SIZE, 1                 /* CLOSE, SKIPZERO                        */    1+IMM2_SIZE, 1                 /* CLOSE, SKIPZERO                        */
2321    
2322  /* 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. */  
2323    
2324  #define RREF_ANY  0xffff  #define RREF_ANY  0xffff
2325    
# Line 2291  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4, Line 2334  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
2334         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
2335         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
2336         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,
2337         ERR70, ERR71, ERR72, ERR73, ERR74, ERR75, ERR76, ERR77, ERR78, ERRCOUNT };         ERR70, ERR71, ERR72, ERR73, ERR74, ERR75, ERR76, ERR77, ERR78, ERR79,
2338           ERR80, ERR81, ERR82, ERR83, ERRCOUNT };
2339    
2340  /* JIT compiling modes. The function list is indexed by them. */  /* JIT compiling modes. The function list is indexed by them. */
2341    
2342  enum { JIT_COMPILE, JIT_PARTIAL_SOFT_COMPILE, JIT_PARTIAL_HARD_COMPILE,  enum { JIT_COMPILE, JIT_PARTIAL_SOFT_COMPILE, JIT_PARTIAL_HARD_COMPILE,
2343         JIT_NUMBER_OF_COMPILE_MODES };         JIT_NUMBER_OF_COMPILE_MODES };
2344    
# Line 2411  typedef struct open_capitem { Line 2456  typedef struct open_capitem {
2456    pcre_uint16 flag;             /* Set TRUE if recursive back ref */    pcre_uint16 flag;             /* Set TRUE if recursive back ref */
2457  } open_capitem;  } open_capitem;
2458    
2459  /* Structure for building a list of named groups during the first pass of  /* Structure for building a list of named groups during the first pass of
2460  compiling. */  compiling. */
2461    
2462  typedef struct named_group {  typedef struct named_group {
2463    const pcre_uchar  *name;          /* Points to the name in the pattern */    const pcre_uchar  *name;          /* Points to the name in the pattern */
2464    int                length;        /* Length of the name */    int                length;        /* Length of the name */
2465    pcre_uint32        number;        /* Group number */    pcre_uint32        number;        /* Group number */
2466  } named_group;  } named_group;
2467    
2468  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
2469  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
# Line 2438  typedef struct compile_data { Line 2483  typedef struct compile_data {
2483    pcre_uchar *name_table;           /* The name/number table */    pcre_uchar *name_table;           /* The name/number table */
2484    int  names_found;                 /* Number of entries so far */    int  names_found;                 /* Number of entries so far */
2485    int  name_entry_size;             /* Size of each entry */    int  name_entry_size;             /* Size of each entry */
2486    int  named_group_list_size;       /* Number of entries in the list */    int  named_group_list_size;       /* Number of entries in the list */
2487    int  workspace_size;              /* Size of workspace */    int  workspace_size;              /* Size of workspace */
2488    unsigned int bracount;            /* Count of capturing parens as we compile */    unsigned int bracount;            /* Count of capturing parens as we compile */
2489    int  final_bracount;              /* Saved value after first pass */    int  final_bracount;              /* Saved value after first pass */
2490    int  max_lookbehind;              /* Maximum lookbehind (characters) */    int  max_lookbehind;              /* Maximum lookbehind (characters) */
2491    int  top_backref;                 /* Maximum back reference */    int  top_backref;                 /* Maximum back reference */
2492    unsigned int backref_map;         /* Bitmap of low back refs */    unsigned int backref_map;         /* Bitmap of low back refs */
2493    unsigned int namedrefcount;       /* Number of backreferences by name */    unsigned int namedrefcount;       /* Number of backreferences by name */
2494      int  parens_depth;                /* Depth of nested parentheses */
2495    int  assert_depth;                /* Depth of nested assertions */    int  assert_depth;                /* Depth of nested assertions */
2496    pcre_uint32 external_options;     /* External (initial) options */    pcre_uint32 external_options;     /* External (initial) options */
2497    pcre_uint32 external_flags;       /* External flag bits to be set */    pcre_uint32 external_flags;       /* External flag bits to be set */

Legend:
Removed from v.1361  
changed lines
  Added in v.1392

  ViewVC Help
Powered by ViewVC 1.1.5