/[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 1359 by ph10, Tue Sep 3 10:10:59 2013 UTC revision 1404 by ph10, Tue Nov 19 15:36:57 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_POSSESS| \
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 1533  a positive value. */ Line 1534  a positive value. */
1534    
1535  #define STRING_DEFINE               "DEFINE"  #define STRING_DEFINE               "DEFINE"
1536    
1537  #define STRING_CR_RIGHTPAR             "CR)"  #define STRING_CR_RIGHTPAR              "CR)"
1538  #define STRING_LF_RIGHTPAR             "LF)"  #define STRING_LF_RIGHTPAR              "LF)"
1539  #define STRING_CRLF_RIGHTPAR           "CRLF)"  #define STRING_CRLF_RIGHTPAR            "CRLF)"
1540  #define STRING_ANY_RIGHTPAR            "ANY)"  #define STRING_ANY_RIGHTPAR             "ANY)"
1541  #define STRING_ANYCRLF_RIGHTPAR        "ANYCRLF)"  #define STRING_ANYCRLF_RIGHTPAR         "ANYCRLF)"
1542  #define STRING_BSR_ANYCRLF_RIGHTPAR    "BSR_ANYCRLF)"  #define STRING_BSR_ANYCRLF_RIGHTPAR     "BSR_ANYCRLF)"
1543  #define STRING_BSR_UNICODE_RIGHTPAR    "BSR_UNICODE)"  #define STRING_BSR_UNICODE_RIGHTPAR     "BSR_UNICODE)"
1544  #define STRING_UTF8_RIGHTPAR           "UTF8)"  #define STRING_UTF8_RIGHTPAR            "UTF8)"
1545  #define STRING_UTF16_RIGHTPAR          "UTF16)"  #define STRING_UTF16_RIGHTPAR           "UTF16)"
1546  #define STRING_UTF32_RIGHTPAR          "UTF32)"  #define STRING_UTF32_RIGHTPAR           "UTF32)"
1547  #define STRING_UTF_RIGHTPAR            "UTF)"  #define STRING_UTF_RIGHTPAR             "UTF)"
1548  #define STRING_UCP_RIGHTPAR            "UCP)"  #define STRING_UCP_RIGHTPAR             "UCP)"
1549  #define STRING_NO_START_OPT_RIGHTPAR   "NO_START_OPT)"  #define STRING_NO_AUTO_POSSESS_RIGHTPAR "NO_AUTO_POSSESS)"
1550  #define STRING_LIMIT_MATCH_EQ          "LIMIT_MATCH="  #define STRING_NO_START_OPT_RIGHTPAR    "NO_START_OPT)"
1551  #define STRING_LIMIT_RECURSION_EQ      "LIMIT_RECURSION="  #define STRING_LIMIT_MATCH_EQ           "LIMIT_MATCH="
1552    #define STRING_LIMIT_RECURSION_EQ       "LIMIT_RECURSION="
1553    
1554  #else  /* SUPPORT_UTF */  #else  /* SUPPORT_UTF */
1555    
# Line 1796  only. */ Line 1798  only. */
1798    
1799  #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E  #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E
1800    
1801  #define STRING_CR_RIGHTPAR             STR_C STR_R STR_RIGHT_PARENTHESIS  #define STRING_CR_RIGHTPAR              STR_C STR_R STR_RIGHT_PARENTHESIS
1802  #define STRING_LF_RIGHTPAR             STR_L STR_F STR_RIGHT_PARENTHESIS  #define STRING_LF_RIGHTPAR              STR_L STR_F STR_RIGHT_PARENTHESIS
1803  #define STRING_CRLF_RIGHTPAR           STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS  #define STRING_CRLF_RIGHTPAR            STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1804  #define STRING_ANY_RIGHTPAR            STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS  #define STRING_ANY_RIGHTPAR             STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1805  #define STRING_ANYCRLF_RIGHTPAR        STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS  #define STRING_ANYCRLF_RIGHTPAR         STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1806  #define STRING_BSR_ANYCRLF_RIGHTPAR    STR_B STR_S STR_R STR_UNDERSCORE STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS  #define STRING_BSR_ANYCRLF_RIGHTPAR     STR_B STR_S STR_R STR_UNDERSCORE STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1807  #define STRING_BSR_UNICODE_RIGHTPAR    STR_B STR_S STR_R STR_UNDERSCORE STR_U STR_N STR_I STR_C STR_O STR_D STR_E STR_RIGHT_PARENTHESIS  #define STRING_BSR_UNICODE_RIGHTPAR     STR_B STR_S STR_R STR_UNDERSCORE STR_U STR_N STR_I STR_C STR_O STR_D STR_E STR_RIGHT_PARENTHESIS
1808  #define STRING_UTF8_RIGHTPAR           STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS  #define STRING_UTF8_RIGHTPAR            STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1809  #define STRING_UTF16_RIGHTPAR          STR_U STR_T STR_F STR_1 STR_6 STR_RIGHT_PARENTHESIS  #define STRING_UTF16_RIGHTPAR           STR_U STR_T STR_F STR_1 STR_6 STR_RIGHT_PARENTHESIS
1810  #define STRING_UTF32_RIGHTPAR          STR_U STR_T STR_F STR_3 STR_2 STR_RIGHT_PARENTHESIS  #define STRING_UTF32_RIGHTPAR           STR_U STR_T STR_F STR_3 STR_2 STR_RIGHT_PARENTHESIS
1811  #define STRING_UTF_RIGHTPAR            STR_U STR_T STR_F STR_RIGHT_PARENTHESIS  #define STRING_UTF_RIGHTPAR             STR_U STR_T STR_F STR_RIGHT_PARENTHESIS
1812  #define STRING_UCP_RIGHTPAR            STR_U STR_C STR_P STR_RIGHT_PARENTHESIS  #define STRING_UCP_RIGHTPAR             STR_U STR_C STR_P STR_RIGHT_PARENTHESIS
1813  #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_NO_AUTO_POSSESS_RIGHTPAR STR_N STR_O STR_UNDERSCORE STR_A STR_U STR_T STR_O STR_UNDERSCORE STR_P STR_O STR_S STR_S STR_E STR_S STR_S STR_RIGHT_PARENTHESIS
1814  #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_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
1815  #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  #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
1816    #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
1817    
1818  #endif  /* SUPPORT_UTF */  #endif  /* SUPPORT_UTF */
1819    
# Line 1852  only. */ Line 1855  only. */
1855  #define PT_WORD       8    /* Word - L plus N plus underscore */  #define PT_WORD       8    /* Word - L plus N plus underscore */
1856  #define PT_CLIST      9    /* Pseudo-property: match character list */  #define PT_CLIST      9    /* Pseudo-property: match character list */
1857  #define PT_UCNC      10    /* Universal Character nameable character */  #define PT_UCNC      10    /* Universal Character nameable character */
1858    #define PT_TABSIZE   11    /* Size of square table for autopossessify tests */
1859    
1860    /* The following special properties are used only in XCLASS items, when POSIX
1861    classes are specified and PCRE_UCP is set - in other words, for Unicode
1862    handling of these classes. They are not available via the \p or \P escapes like
1863    those in the above list, and so they do not take part in the autopossessifying
1864    table. */
1865    
1866    #define PT_PXGRAPH   11    /* [:graph:] - characters that mark the paper */
1867    #define PT_PXPRINT   12    /* [:print:] - [:graph:] plus non-control spaces */
1868    #define PT_PXPUNCT   13    /* [:punct:] - punctuation characters */
1869    
1870  /* 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
1871  contain characters with values greater than 255. */  contain characters with values greater than 255. */
# Line 1866  contain characters with values greater t Line 1880  contain characters with values greater t
1880  #define XCL_NOTPROP   4    /* Unicode inverted property (ditto) */  #define XCL_NOTPROP   4    /* Unicode inverted property (ditto) */
1881    
1882  /* 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
1883  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
1884  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
1885  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
1886  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
1887  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
1888  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 1905  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, E
1905         ESC_E, ESC_Q, ESC_g, ESC_k,         ESC_E, ESC_Q, ESC_g, ESC_k,
1906         ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu };         ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu };
1907    
 /* 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.  
1908    
1909  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  /********************** Opcode definitions ******************/
1910  that follow must also be updated to match. There are also tables called  
1911  "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */  /****** NOTE NOTE NOTE ******
1912    
1913    Starting from 1 (i.e. after OP_END), the values up to OP_EOD must correspond in
1914    order to the list of escapes immediately above. Furthermore, values up to
1915    OP_DOLLM must not be changed without adjusting the table called autoposstab in
1916    pcre_compile.c
1917    
1918    Whenever this list is updated, the two macro definitions that follow must be
1919    updated to match. The possessification table called "opcode_possessify" in
1920    pcre_compile.c must also be updated, and also the tables called "coptable"
1921    and "poptable" in pcre_dfa_exec.c.
1922    
1923    ****** NOTE NOTE NOTE ******/
1924    
1925    
1926    /* The values between FIRST_AUTOTAB_OP and LAST_AUTOTAB_RIGHT_OP, inclusive,
1927    are used in a table for deciding whether a repeated character type can be
1928    auto-possessified. */
1929    
1930    #define FIRST_AUTOTAB_OP       OP_NOT_DIGIT
1931    #define LAST_AUTOTAB_LEFT_OP   OP_EXTUNI
1932    #define LAST_AUTOTAB_RIGHT_OP  OP_DOLLM
1933    
1934  enum {  enum {
1935    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 1929  enum { Line 1962  enum {
1962    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) */
1963    OP_EOD,            /* 24 End of data (\z) */    OP_EOD,            /* 24 End of data (\z) */
1964    
1965    OP_CIRC,           /* 25 Start of line - not multiline */    /* Line end assertions */
1966    OP_CIRCM,          /* 26 Start of line - multiline */  
1967    OP_DOLL,           /* 27 End of line - not multiline */    OP_DOLL,           /* 25 End of line - not multiline */
1968    OP_DOLLM,          /* 28 End of line - multiline */    OP_DOLLM,          /* 26 End of line - multiline */
1969      OP_CIRC,           /* 27 Start of line - not multiline */
1970      OP_CIRCM,          /* 28 Start of line - multiline */
1971    
1972      /* Single characters; caseful must precede the caseless ones */
1973    
1974    OP_CHAR,           /* 29 Match one character, casefully */    OP_CHAR,           /* 29 Match one character, casefully */
1975    OP_CHARI,          /* 30 Match one character, caselessly */    OP_CHARI,          /* 30 Match one character, caselessly */
1976    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 1979  enum {
1979    /* 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
1980    the offset from the first one is used to generate the others. */    the offset from the first one is used to generate the others. */
1981    
1982    /**** Single characters, caseful, must precede the caseless ones ****/    /* Repeated characters; caseful must precede the caseless ones */
1983    
1984    OP_STAR,           /* 33 The maximizing and minimizing versions of */    OP_STAR,           /* 33 The maximizing and minimizing versions of */
1985    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 1997  enum {
1997    OP_POSQUERY,       /* 44 Posesssified query, caseful */    OP_POSQUERY,       /* 44 Posesssified query, caseful */
1998    OP_POSUPTO,        /* 45 Possessified upto, caseful */    OP_POSUPTO,        /* 45 Possessified upto, caseful */
1999    
2000    /**** Single characters, caseless, must follow the caseful ones */    /* Repeated characters; caseless must follow the caseful ones */
2001    
2002    OP_STARI,          /* 46 */    OP_STARI,          /* 46 */
2003    OP_MINSTARI,       /* 47 */    OP_MINSTARI,       /* 47 */
# Line 1977  enum { Line 2015  enum {
2015    OP_POSQUERYI,      /* 57 Posesssified query, caseless */    OP_POSQUERYI,      /* 57 Posesssified query, caseless */
2016    OP_POSUPTOI,       /* 58 Possessified upto, caseless */    OP_POSUPTOI,       /* 58 Possessified upto, caseless */
2017    
2018    /**** The negated ones must follow the non-negated ones, and match them ****/    /* The negated ones must follow the non-negated ones, and match them */
2019    /**** Negated single character, caseful; must precede the caseless ones ****/    /* Negated repeated character, caseful; must precede the caseless ones */
2020    
2021    OP_NOTSTAR,        /* 59 The maximizing and minimizing versions of */    OP_NOTSTAR,        /* 59 The maximizing and minimizing versions of */
2022    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 2034  enum {
2034    OP_NOTPOSQUERY,    /* 70 */    OP_NOTPOSQUERY,    /* 70 */
2035    OP_NOTPOSUPTO,     /* 71 */    OP_NOTPOSUPTO,     /* 71 */
2036    
2037    /**** Negated single character, caseless; must follow the caseful ones ****/    /* Negated repeated character, caseless; must follow the caseful ones */
2038    
2039    OP_NOTSTARI,       /* 72 */    OP_NOTSTARI,       /* 72 */
2040    OP_NOTMINSTARI,    /* 73 */    OP_NOTMINSTARI,    /* 73 */
# Line 2014  enum { Line 2052  enum {
2052    OP_NOTPOSQUERYI,   /* 83 */    OP_NOTPOSQUERYI,   /* 83 */
2053    OP_NOTPOSUPTOI,    /* 84 */    OP_NOTPOSUPTOI,    /* 84 */
2054    
2055    /**** Character types ****/    /* Character types */
2056    
2057    OP_TYPESTAR,       /* 85 The maximizing and minimizing versions of */    OP_TYPESTAR,       /* 85 The maximizing and minimizing versions of */
2058    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 2083  enum {
2083    OP_CRRANGE,        /* 104 These are different to the three sets above. */    OP_CRRANGE,        /* 104 These are different to the three sets above. */
2084    OP_CRMINRANGE,     /* 105 */    OP_CRMINRANGE,     /* 105 */
2085    
2086      OP_CRPOSSTAR,      /* 106 Possessified versions */
2087      OP_CRPOSPLUS,      /* 107 */
2088      OP_CRPOSQUERY,     /* 108 */
2089      OP_CRPOSRANGE,     /* 109 */
2090    
2091    /* End of quantifier opcodes */    /* End of quantifier opcodes */
2092    
2093    OP_CLASS,          /* 106 Match a character class, chars < 256 only */    OP_CLASS,          /* 110 Match a character class, chars < 256 only */
2094    OP_NCLASS,         /* 107 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 111 Same, but the bitmap was created from a negative
2095                                class - the difference is relevant only when a                                class - the difference is relevant only when a
2096                                character > 255 is encountered. */                                character > 255 is encountered. */
2097    OP_XCLASS,         /* 108 Extended class for handling > 255 chars within the    OP_XCLASS,         /* 112 Extended class for handling > 255 chars within the
2098                                class. This does both positive and negative. */                                class. This does both positive and negative. */
2099    OP_REF,            /* 109 Match a back reference, casefully */    OP_REF,            /* 113 Match a back reference, casefully */
2100    OP_REFI,           /* 110 Match a back reference, caselessly */    OP_REFI,           /* 114 Match a back reference, caselessly */
2101    OP_RECURSE,        /* 111 Match a numbered subpattern (possibly recursive) */    OP_DNREF,          /* 115 Match a duplicate name backref, casefully */
2102    OP_CALLOUT,        /* 112 Call out to external function if provided */    OP_DNREFI,         /* 116 Match a duplicate name backref, caselessly */
2103      OP_RECURSE,        /* 117 Match a numbered subpattern (possibly recursive) */
2104    OP_ALT,            /* 113 Start of alternation */    OP_CALLOUT,        /* 118 Call out to external function if provided */
2105    OP_KET,            /* 114 End of group that doesn't have an unbounded repeat */  
2106    OP_KETRMAX,        /* 115 These two must remain together and in this */    OP_ALT,            /* 119 Start of alternation */
2107    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 */
2108    OP_KETRPOS,        /* 117 Possessive unlimited repeat. */    OP_KETRMAX,        /* 121 These two must remain together and in this */
2109      OP_KETRMIN,        /* 122 order. They are for groups the repeat for ever. */
2110      OP_KETRPOS,        /* 123 Possessive unlimited repeat. */
2111    
2112    /* 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
2113    asserts must remain in order. */    asserts must remain in order. */
2114    
2115    OP_REVERSE,        /* 118 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 124 Move pointer back - used in lookbehind assertions */
2116    OP_ASSERT,         /* 119 Positive lookahead */    OP_ASSERT,         /* 125 Positive lookahead */
2117    OP_ASSERT_NOT,     /* 120 Negative lookahead */    OP_ASSERT_NOT,     /* 126 Negative lookahead */
2118    OP_ASSERTBACK,     /* 121 Positive lookbehind */    OP_ASSERTBACK,     /* 127 Positive lookbehind */
2119    OP_ASSERTBACK_NOT, /* 122 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 128 Negative lookbehind */
2120    
2121    /* ONCE, ONCE_NC, BRA, BRAPOS, CBRA, CBRAPOS, and COND must come immediately    /* ONCE, ONCE_NC, BRA, BRAPOS, CBRA, CBRAPOS, and COND must come immediately
2122    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
2123    subpattern that isn't an assertion. The POS versions must immediately follow    subpattern that isn't an assertion. The POS versions must immediately follow
2124    the non-POS versions in each case. */    the non-POS versions in each case. */
2125    
2126    OP_ONCE,           /* 123 Atomic group, contains captures */    OP_ONCE,           /* 129 Atomic group, contains captures */
2127    OP_ONCE_NC,        /* 124 Atomic group containing no captures */    OP_ONCE_NC,        /* 130 Atomic group containing no captures */
2128    OP_BRA,            /* 125 Start of non-capturing bracket */    OP_BRA,            /* 131 Start of non-capturing bracket */
2129    OP_BRAPOS,         /* 126 Ditto, with unlimited, possessive repeat */    OP_BRAPOS,         /* 132 Ditto, with unlimited, possessive repeat */
2130    OP_CBRA,           /* 127 Start of capturing bracket */    OP_CBRA,           /* 133 Start of capturing bracket */
2131    OP_CBRAPOS,        /* 128 Ditto, with unlimited, possessive repeat */    OP_CBRAPOS,        /* 134 Ditto, with unlimited, possessive repeat */
2132    OP_COND,           /* 129 Conditional group */    OP_COND,           /* 135 Conditional group */
2133    
2134    /* 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
2135    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
2136    
2137    OP_SBRA,           /* 130 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 136 Start of non-capturing bracket, check empty  */
2138    OP_SBRAPOS,        /* 131 Ditto, with unlimited, possessive repeat */    OP_SBRAPOS,        /* 137 Ditto, with unlimited, possessive repeat */
2139    OP_SCBRA,          /* 132 Start of capturing bracket, check empty */    OP_SCBRA,          /* 138 Start of capturing bracket, check empty */
2140    OP_SCBRAPOS,       /* 133 Ditto, with unlimited, possessive repeat */    OP_SCBRAPOS,       /* 139 Ditto, with unlimited, possessive repeat */
2141    OP_SCOND,          /* 134 Conditional group, check empty */    OP_SCOND,          /* 140 Conditional group, check empty */
2142    
2143    /* The next two pairs must (respectively) be kept together. */    /* The next two pairs must (respectively) be kept together. */
2144    
2145    OP_CREF,           /* 135 Used to hold a capture number as condition */    OP_CREF,           /* 141 Used to hold a capture number as condition */
2146    OP_NCREF,          /* 136 Same, but generated by a name reference*/    OP_DNCREF,         /* 142 Used to point to duplicate names as a condition */
2147    OP_RREF,           /* 137 Used to hold a recursion number as condition */    OP_RREF,           /* 143 Used to hold a recursion number as condition */
2148    OP_NRREF,          /* 138 Same, but generated by a name reference*/    OP_DNRREF,         /* 144 Used to point to duplicate names as a condition */
2149    OP_DEF,            /* 139 The DEFINE condition */    OP_DEF,            /* 145 The DEFINE condition */
2150    
2151    OP_BRAZERO,        /* 140 These two must remain together and in this */    OP_BRAZERO,        /* 146 These two must remain together and in this */
2152    OP_BRAMINZERO,     /* 141 order. */    OP_BRAMINZERO,     /* 147 order. */
2153    OP_BRAPOSZERO,     /* 142 */    OP_BRAPOSZERO,     /* 148 */
2154    
2155    /* These are backtracking control verbs */    /* These are backtracking control verbs */
2156    
2157    OP_MARK,           /* 143 always has an argument */    OP_MARK,           /* 149 always has an argument */
2158    OP_PRUNE,          /* 144 */    OP_PRUNE,          /* 150 */
2159    OP_PRUNE_ARG,      /* 145 same, but with argument */    OP_PRUNE_ARG,      /* 151 same, but with argument */
2160    OP_SKIP,           /* 146 */    OP_SKIP,           /* 152 */
2161    OP_SKIP_ARG,       /* 147 same, but with argument */    OP_SKIP_ARG,       /* 153 same, but with argument */
2162    OP_THEN,           /* 148 */    OP_THEN,           /* 154 */
2163    OP_THEN_ARG,       /* 149 same, but with argument */    OP_THEN_ARG,       /* 155 same, but with argument */
2164    OP_COMMIT,         /* 150 */    OP_COMMIT,         /* 156 */
2165    
2166    /* These are forced failure and success verbs */    /* These are forced failure and success verbs */
2167    
2168    OP_FAIL,           /* 151 */    OP_FAIL,           /* 157 */
2169    OP_ACCEPT,         /* 152 */    OP_ACCEPT,         /* 158 */
2170    OP_ASSERT_ACCEPT,  /* 153 Used inside assertions */    OP_ASSERT_ACCEPT,  /* 159 Used inside assertions */
2171    OP_CLOSE,          /* 154 Used before OP_ACCEPT to close open captures */    OP_CLOSE,          /* 160 Used before OP_ACCEPT to close open captures */
2172    
2173    /* This is used to skip a subpattern with a {0} quantifier */    /* This is used to skip a subpattern with a {0} quantifier */
2174    
2175    OP_SKIPZERO,       /* 155 */    OP_SKIPZERO,       /* 161 */
2176    
2177    /* 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
2178    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 2138  enum { Line 2183  enum {
2183    
2184  /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro  /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
2185  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
2186  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
2187    pcre_dfa_exec.c that must be updated. */
2188    
2189    
2190  /* 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 2151  some cases doesn't actually use these na Line 2197  some cases doesn't actually use these na
2197    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
2198    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
2199    "extuni",  "\\Z", "\\z",                                        \    "extuni",  "\\Z", "\\z",                                        \
2200    "^", "^", "$", "$", "char", "chari", "not", "noti",             \    "$", "$", "^", "^", "char", "chari", "not", "noti",             \
2201    "*", "*?", "+", "+?", "?", "??",                                \    "*", "*?", "+", "+?", "?", "??",                                \
2202    "{", "{", "{",                                                  \    "{", "{", "{",                                                  \
2203    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
# Line 2167  some cases doesn't actually use these na Line 2213  some cases doesn't actually use these na
2213    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
2214    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
2215    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
2216    "class", "nclass", "xclass", "Ref", "Refi",                     \    "*+","++", "?+", "{",                                           \
2217      "class", "nclass", "xclass", "Ref", "Refi", "DnRef", "DnRefi",  \
2218    "Recurse", "Callout",                                           \    "Recurse", "Callout",                                           \
2219    "Alt", "Ket", "KetRmax", "KetRmin", "KetRpos",                  \    "Alt", "Ket", "KetRmax", "KetRmin", "KetRpos",                  \
2220    "Reverse", "Assert", "Assert not", "AssertB", "AssertB not",    \    "Reverse", "Assert", "Assert not", "AssertB", "AssertB not",    \
# Line 2176  some cases doesn't actually use these na Line 2223  some cases doesn't actually use these na
2223    "Cond",                                                         \    "Cond",                                                         \
2224    "SBra", "SBraPos", "SCBra", "SCBraPos",                         \    "SBra", "SBraPos", "SCBra", "SCBraPos",                         \
2225    "SCond",                                                        \    "SCond",                                                        \
2226    "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \    "Cond ref", "Cond dnref", "Cond rec", "Cond dnrec", "Cond def", \
2227    "Brazero", "Braminzero", "Braposzero",                          \    "Brazero", "Braminzero", "Braposzero",                          \
2228    "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \    "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \
2229    "*THEN", "*THEN", "*COMMIT", "*FAIL",                           \    "*THEN", "*THEN", "*COMMIT", "*FAIL",                           \
# Line 2201  in UTF-8 mode. The code that uses this t Line 2248  in UTF-8 mode. The code that uses this t
2248    3, 3,                          /* \P, \p                                 */ \    3, 3,                          /* \P, \p                                 */ \
2249    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
2250    1,                             /* \X                                     */ \    1,                             /* \X                                     */ \
2251    1, 1, 1, 1, 1, 1,              /* \Z, \z, ^, ^M, $, $M                   */ \    1, 1, 1, 1, 1, 1,              /* \Z, \z, $, $M ^, ^M                    */ \
2252    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
2253    2,                             /* Chari  - the minimum length            */ \    2,                             /* Chari  - the minimum length            */ \
2254    2,                             /* not                                    */ \    2,                             /* not                                    */ \
# Line 2232  in UTF-8 mode. The code that uses this t Line 2279  in UTF-8 mode. The code that uses this t
2279    /* Character class & ref repeats                                         */ \    /* Character class & ref repeats                                         */ \
2280    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \
2281    1+2*IMM2_SIZE, 1+2*IMM2_SIZE,  /* CRRANGE, CRMINRANGE                    */ \    1+2*IMM2_SIZE, 1+2*IMM2_SIZE,  /* CRRANGE, CRMINRANGE                    */ \
2282      1, 1, 1, 1+2*IMM2_SIZE,        /* Possessive *+, ++, ?+, CRPOSRANGE      */ \
2283    1+(32/sizeof(pcre_uchar)),     /* CLASS                                  */ \    1+(32/sizeof(pcre_uchar)),     /* CLASS                                  */ \
2284    1+(32/sizeof(pcre_uchar)),     /* NCLASS                                 */ \    1+(32/sizeof(pcre_uchar)),     /* NCLASS                                 */ \
2285    0,                             /* XCLASS - variable length               */ \    0,                             /* XCLASS - variable length               */ \
2286    1+IMM2_SIZE,                   /* REF                                    */ \    1+IMM2_SIZE,                   /* REF                                    */ \
2287    1+IMM2_SIZE,                   /* REFI                                   */ \    1+IMM2_SIZE,                   /* REFI                                   */ \
2288      1+2*IMM2_SIZE,                 /* DNREF                                  */ \
2289      1+2*IMM2_SIZE,                 /* DNREFI                                 */ \
2290    1+LINK_SIZE,                   /* RECURSE                                */ \    1+LINK_SIZE,                   /* RECURSE                                */ \
2291    2+2*LINK_SIZE,                 /* CALLOUT                                */ \    2+2*LINK_SIZE,                 /* CALLOUT                                */ \
2292    1+LINK_SIZE,                   /* Alt                                    */ \    1+LINK_SIZE,                   /* Alt                                    */ \
# Line 2261  in UTF-8 mode. The code that uses this t Line 2311  in UTF-8 mode. The code that uses this t
2311    1+LINK_SIZE+IMM2_SIZE,         /* SCBRA                                  */ \    1+LINK_SIZE+IMM2_SIZE,         /* SCBRA                                  */ \
2312    1+LINK_SIZE+IMM2_SIZE,         /* SCBRAPOS                               */ \    1+LINK_SIZE+IMM2_SIZE,         /* SCBRAPOS                               */ \
2313    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
2314    1+IMM2_SIZE, 1+IMM2_SIZE,      /* CREF, NCREF                            */ \    1+IMM2_SIZE, 1+2*IMM2_SIZE,    /* CREF, DNCREF                           */ \
2315    1+IMM2_SIZE, 1+IMM2_SIZE,      /* RREF, NRREF                            */ \    1+IMM2_SIZE, 1+2*IMM2_SIZE,    /* RREF, DNRREF                           */ \
2316    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
2317    1, 1, 1,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */ \    1, 1, 1,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */ \
2318    3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \    3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \
# Line 2271  in UTF-8 mode. The code that uses this t Line 2321  in UTF-8 mode. The code that uses this t
2321    1, 1, 1, 1,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */ \    1, 1, 1, 1,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */ \
2322    1+IMM2_SIZE, 1                 /* CLOSE, SKIPZERO                        */    1+IMM2_SIZE, 1                 /* CLOSE, SKIPZERO                        */
2323    
2324  /* 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. */  
2325    
2326  #define RREF_ANY  0xffff  #define RREF_ANY  0xffff
2327    
# Line 2287  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4, Line 2336  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
2336         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
2337         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
2338         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,
2339         ERR70, ERR71, ERR72, ERR73, ERR74, ERR75, ERR76, ERR77, ERR78, ERRCOUNT };         ERR70, ERR71, ERR72, ERR73, ERR74, ERR75, ERR76, ERR77, ERR78, ERR79,
2340           ERR80, ERR81, ERR82, ERR83, ERR84, ERRCOUNT };
2341    
2342  /* JIT compiling modes. The function list is indexed by them. */  /* JIT compiling modes. The function list is indexed by them. */
2343    
2344  enum { JIT_COMPILE, JIT_PARTIAL_SOFT_COMPILE, JIT_PARTIAL_HARD_COMPILE,  enum { JIT_COMPILE, JIT_PARTIAL_SOFT_COMPILE, JIT_PARTIAL_HARD_COMPILE,
2345         JIT_NUMBER_OF_COMPILE_MODES };         JIT_NUMBER_OF_COMPILE_MODES };
2346    
# Line 2407  typedef struct open_capitem { Line 2458  typedef struct open_capitem {
2458    pcre_uint16 flag;             /* Set TRUE if recursive back ref */    pcre_uint16 flag;             /* Set TRUE if recursive back ref */
2459  } open_capitem;  } open_capitem;
2460    
2461  /* 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
2462  compiling. */  compiling. */
2463    
2464  typedef struct named_group {  typedef struct named_group {
2465    const pcre_uchar  *name;          /* Points to the name in the pattern */    const pcre_uchar  *name;          /* Points to the name in the pattern */
2466    int                length;        /* Length of the name */    int                length;        /* Length of the name */
2467    pcre_uint32        number;        /* Group number */    pcre_uint32        number;        /* Group number */
2468  } named_group;  } named_group;
2469    
2470  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
2471  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
# Line 2434  typedef struct compile_data { Line 2485  typedef struct compile_data {
2485    pcre_uchar *name_table;           /* The name/number table */    pcre_uchar *name_table;           /* The name/number table */
2486    int  names_found;                 /* Number of entries so far */    int  names_found;                 /* Number of entries so far */
2487    int  name_entry_size;             /* Size of each entry */    int  name_entry_size;             /* Size of each entry */
2488    int  named_group_list_size;       /* Number of entries in the list */    int  named_group_list_size;       /* Number of entries in the list */
2489    int  workspace_size;              /* Size of workspace */    int  workspace_size;              /* Size of workspace */
2490    unsigned int bracount;            /* Count of capturing parens as we compile */    unsigned int bracount;            /* Count of capturing parens as we compile */
2491    int  final_bracount;              /* Saved value after first pass */    int  final_bracount;              /* Saved value after first pass */
2492    int  max_lookbehind;              /* Maximum lookbehind (characters) */    int  max_lookbehind;              /* Maximum lookbehind (characters) */
2493    int  top_backref;                 /* Maximum back reference */    int  top_backref;                 /* Maximum back reference */
2494    unsigned int backref_map;         /* Bitmap of low back refs */    unsigned int backref_map;         /* Bitmap of low back refs */
2495      unsigned int namedrefcount;       /* Number of backreferences by name */
2496      int  parens_depth;                /* Depth of nested parentheses */
2497    int  assert_depth;                /* Depth of nested assertions */    int  assert_depth;                /* Depth of nested assertions */
2498    pcre_uint32 external_options;     /* External (initial) options */    pcre_uint32 external_options;     /* External (initial) options */
2499    pcre_uint32 external_flags;       /* External flag bits to be set */    pcre_uint32 external_flags;       /* External flag bits to be set */
# Line 2448  typedef struct compile_data { Line 2501  typedef struct compile_data {
2501    BOOL had_accept;                  /* (*ACCEPT) encountered */    BOOL had_accept;                  /* (*ACCEPT) encountered */
2502    BOOL had_pruneorskip;             /* (*PRUNE) or (*SKIP) encountered */    BOOL had_pruneorskip;             /* (*PRUNE) or (*SKIP) encountered */
2503    BOOL check_lookbehind;            /* Lookbehinds need later checking */    BOOL check_lookbehind;            /* Lookbehinds need later checking */
2504      BOOL dupnames;                    /* Duplicate names exist */
2505    int  nltype;                      /* Newline type */    int  nltype;                      /* Newline type */
2506    int  nllen;                       /* Newline string length */    int  nllen;                       /* Newline string length */
2507    pcre_uchar nl[4];                 /* Newline string when fixed length */    pcre_uchar nl[4];                 /* Newline string when fixed length */

Legend:
Removed from v.1359  
changed lines
  Added in v.1404

  ViewVC Help
Powered by ViewVC 1.1.5