/[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 1348 by ph10, Fri Jul 5 10:38:37 2013 UTC revision 1408 by ph10, Tue Dec 3 16:27:00 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 1795  only. */ Line 1797  only. */
1797  #define STRING_xdigit               STR_x STR_d STR_i STR_g STR_i STR_t  #define STRING_xdigit               STR_x STR_d STR_i STR_g STR_i STR_t
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    #define STRING_WEIRD_STARTWORD      STR_LEFT_SQUARE_BRACKET STR_COLON STR_LESS_THAN_SIGN STR_COLON STR_RIGHT_SQUARE_BRACKET STR_RIGHT_SQUARE_BRACKET
1801    #define STRING_WEIRD_ENDWORD        STR_LEFT_SQUARE_BRACKET STR_COLON STR_GREATER_THAN_SIGN STR_COLON STR_RIGHT_SQUARE_BRACKET STR_RIGHT_SQUARE_BRACKET
1802    
1803  #define STRING_CR_RIGHTPAR             STR_C STR_R STR_RIGHT_PARENTHESIS  #define STRING_CR_RIGHTPAR              STR_C STR_R STR_RIGHT_PARENTHESIS
1804  #define STRING_LF_RIGHTPAR             STR_L STR_F STR_RIGHT_PARENTHESIS  #define STRING_LF_RIGHTPAR              STR_L STR_F STR_RIGHT_PARENTHESIS
1805  #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
1806  #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
1807  #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
1808  #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
1809  #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
1810  #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
1811  #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
1812  #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
1813  #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
1814  #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
1815  #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
1816  #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
1817  #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
1818    #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
1819    
1820  #endif  /* SUPPORT_UTF */  #endif  /* SUPPORT_UTF */
1821    
# Line 1852  only. */ Line 1857  only. */
1857  #define PT_WORD       8    /* Word - L plus N plus underscore */  #define PT_WORD       8    /* Word - L plus N plus underscore */
1858  #define PT_CLIST      9    /* Pseudo-property: match character list */  #define PT_CLIST      9    /* Pseudo-property: match character list */
1859  #define PT_UCNC      10    /* Universal Character nameable character */  #define PT_UCNC      10    /* Universal Character nameable character */
1860    #define PT_TABSIZE   11    /* Size of square table for autopossessify tests */
1861    
1862    /* The following special properties are used only in XCLASS items, when POSIX
1863    classes are specified and PCRE_UCP is set - in other words, for Unicode
1864    handling of these classes. They are not available via the \p or \P escapes like
1865    those in the above list, and so they do not take part in the autopossessifying
1866    table. */
1867    
1868    #define PT_PXGRAPH   11    /* [:graph:] - characters that mark the paper */
1869    #define PT_PXPRINT   12    /* [:print:] - [:graph:] plus non-control spaces */
1870    #define PT_PXPUNCT   13    /* [:punct:] - punctuation characters */
1871    
1872  /* 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
1873  contain characters with values greater than 255. */  contain characters with values greater than 255. */
# Line 1866  contain characters with values greater t Line 1882  contain characters with values greater t
1882  #define XCL_NOTPROP   4    /* Unicode inverted property (ditto) */  #define XCL_NOTPROP   4    /* Unicode inverted property (ditto) */
1883    
1884  /* 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
1885  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
1886  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
1887  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
1888  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
1889  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
1890  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 1907  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, E
1907         ESC_E, ESC_Q, ESC_g, ESC_k,         ESC_E, ESC_Q, ESC_g, ESC_k,
1908         ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu };         ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu };
1909    
 /* 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.  
1910    
1911  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  /********************** Opcode definitions ******************/
1912  that follow must also be updated to match. There are also tables called  
1913  "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */  /****** NOTE NOTE NOTE ******
1914    
1915    Starting from 1 (i.e. after OP_END), the values up to OP_EOD must correspond in
1916    order to the list of escapes immediately above. Furthermore, values up to
1917    OP_DOLLM must not be changed without adjusting the table called autoposstab in
1918    pcre_compile.c
1919    
1920    Whenever this list is updated, the two macro definitions that follow must be
1921    updated to match. The possessification table called "opcode_possessify" in
1922    pcre_compile.c must also be updated, and also the tables called "coptable"
1923    and "poptable" in pcre_dfa_exec.c.
1924    
1925    ****** NOTE NOTE NOTE ******/
1926    
1927    
1928    /* The values between FIRST_AUTOTAB_OP and LAST_AUTOTAB_RIGHT_OP, inclusive,
1929    are used in a table for deciding whether a repeated character type can be
1930    auto-possessified. */
1931    
1932    #define FIRST_AUTOTAB_OP       OP_NOT_DIGIT
1933    #define LAST_AUTOTAB_LEFT_OP   OP_EXTUNI
1934    #define LAST_AUTOTAB_RIGHT_OP  OP_DOLLM
1935    
1936  enum {  enum {
1937    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 1929  enum { Line 1964  enum {
1964    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) */
1965    OP_EOD,            /* 24 End of data (\z) */    OP_EOD,            /* 24 End of data (\z) */
1966    
1967    OP_CIRC,           /* 25 Start of line - not multiline */    /* Line end assertions */
1968    OP_CIRCM,          /* 26 Start of line - multiline */  
1969    OP_DOLL,           /* 27 End of line - not multiline */    OP_DOLL,           /* 25 End of line - not multiline */
1970    OP_DOLLM,          /* 28 End of line - multiline */    OP_DOLLM,          /* 26 End of line - multiline */
1971      OP_CIRC,           /* 27 Start of line - not multiline */
1972      OP_CIRCM,          /* 28 Start of line - multiline */
1973    
1974      /* Single characters; caseful must precede the caseless ones */
1975    
1976    OP_CHAR,           /* 29 Match one character, casefully */    OP_CHAR,           /* 29 Match one character, casefully */
1977    OP_CHARI,          /* 30 Match one character, caselessly */    OP_CHARI,          /* 30 Match one character, caselessly */
1978    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 1981  enum {
1981    /* 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
1982    the offset from the first one is used to generate the others. */    the offset from the first one is used to generate the others. */
1983    
1984    /**** Single characters, caseful, must precede the caseless ones ****/    /* Repeated characters; caseful must precede the caseless ones */
1985    
1986    OP_STAR,           /* 33 The maximizing and minimizing versions of */    OP_STAR,           /* 33 The maximizing and minimizing versions of */
1987    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 1999  enum {
1999    OP_POSQUERY,       /* 44 Posesssified query, caseful */    OP_POSQUERY,       /* 44 Posesssified query, caseful */
2000    OP_POSUPTO,        /* 45 Possessified upto, caseful */    OP_POSUPTO,        /* 45 Possessified upto, caseful */
2001    
2002    /**** Single characters, caseless, must follow the caseful ones */    /* Repeated characters; caseless must follow the caseful ones */
2003    
2004    OP_STARI,          /* 46 */    OP_STARI,          /* 46 */
2005    OP_MINSTARI,       /* 47 */    OP_MINSTARI,       /* 47 */
# Line 1977  enum { Line 2017  enum {
2017    OP_POSQUERYI,      /* 57 Posesssified query, caseless */    OP_POSQUERYI,      /* 57 Posesssified query, caseless */
2018    OP_POSUPTOI,       /* 58 Possessified upto, caseless */    OP_POSUPTOI,       /* 58 Possessified upto, caseless */
2019    
2020    /**** The negated ones must follow the non-negated ones, and match them ****/    /* The negated ones must follow the non-negated ones, and match them */
2021    /**** Negated single character, caseful; must precede the caseless ones ****/    /* Negated repeated character, caseful; must precede the caseless ones */
2022    
2023    OP_NOTSTAR,        /* 59 The maximizing and minimizing versions of */    OP_NOTSTAR,        /* 59 The maximizing and minimizing versions of */
2024    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 2036  enum {
2036    OP_NOTPOSQUERY,    /* 70 */    OP_NOTPOSQUERY,    /* 70 */
2037    OP_NOTPOSUPTO,     /* 71 */    OP_NOTPOSUPTO,     /* 71 */
2038    
2039    /**** Negated single character, caseless; must follow the caseful ones ****/    /* Negated repeated character, caseless; must follow the caseful ones */
2040    
2041    OP_NOTSTARI,       /* 72 */    OP_NOTSTARI,       /* 72 */
2042    OP_NOTMINSTARI,    /* 73 */    OP_NOTMINSTARI,    /* 73 */
# Line 2014  enum { Line 2054  enum {
2054    OP_NOTPOSQUERYI,   /* 83 */    OP_NOTPOSQUERYI,   /* 83 */
2055    OP_NOTPOSUPTOI,    /* 84 */    OP_NOTPOSUPTOI,    /* 84 */
2056    
2057    /**** Character types ****/    /* Character types */
2058    
2059    OP_TYPESTAR,       /* 85 The maximizing and minimizing versions of */    OP_TYPESTAR,       /* 85 The maximizing and minimizing versions of */
2060    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 2085  enum {
2085    OP_CRRANGE,        /* 104 These are different to the three sets above. */    OP_CRRANGE,        /* 104 These are different to the three sets above. */
2086    OP_CRMINRANGE,     /* 105 */    OP_CRMINRANGE,     /* 105 */
2087    
2088      OP_CRPOSSTAR,      /* 106 Possessified versions */
2089      OP_CRPOSPLUS,      /* 107 */
2090      OP_CRPOSQUERY,     /* 108 */
2091      OP_CRPOSRANGE,     /* 109 */
2092    
2093    /* End of quantifier opcodes */    /* End of quantifier opcodes */
2094    
2095    OP_CLASS,          /* 106 Match a character class, chars < 256 only */    OP_CLASS,          /* 110 Match a character class, chars < 256 only */
2096    OP_NCLASS,         /* 107 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 111 Same, but the bitmap was created from a negative
2097                                class - the difference is relevant only when a                                class - the difference is relevant only when a
2098                                character > 255 is encountered. */                                character > 255 is encountered. */
2099    OP_XCLASS,         /* 108 Extended class for handling > 255 chars within the    OP_XCLASS,         /* 112 Extended class for handling > 255 chars within the
2100                                class. This does both positive and negative. */                                class. This does both positive and negative. */
2101    OP_REF,            /* 109 Match a back reference, casefully */    OP_REF,            /* 113 Match a back reference, casefully */
2102    OP_REFI,           /* 110 Match a back reference, caselessly */    OP_REFI,           /* 114 Match a back reference, caselessly */
2103    OP_RECURSE,        /* 111 Match a numbered subpattern (possibly recursive) */    OP_DNREF,          /* 115 Match a duplicate name backref, casefully */
2104    OP_CALLOUT,        /* 112 Call out to external function if provided */    OP_DNREFI,         /* 116 Match a duplicate name backref, caselessly */
2105      OP_RECURSE,        /* 117 Match a numbered subpattern (possibly recursive) */
2106    OP_ALT,            /* 113 Start of alternation */    OP_CALLOUT,        /* 118 Call out to external function if provided */
2107    OP_KET,            /* 114 End of group that doesn't have an unbounded repeat */  
2108    OP_KETRMAX,        /* 115 These two must remain together and in this */    OP_ALT,            /* 119 Start of alternation */
2109    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 */
2110    OP_KETRPOS,        /* 117 Possessive unlimited repeat. */    OP_KETRMAX,        /* 121 These two must remain together and in this */
2111      OP_KETRMIN,        /* 122 order. They are for groups the repeat for ever. */
2112      OP_KETRPOS,        /* 123 Possessive unlimited repeat. */
2113    
2114    /* 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
2115    asserts must remain in order. */    asserts must remain in order. */
2116    
2117    OP_REVERSE,        /* 118 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 124 Move pointer back - used in lookbehind assertions */
2118    OP_ASSERT,         /* 119 Positive lookahead */    OP_ASSERT,         /* 125 Positive lookahead */
2119    OP_ASSERT_NOT,     /* 120 Negative lookahead */    OP_ASSERT_NOT,     /* 126 Negative lookahead */
2120    OP_ASSERTBACK,     /* 121 Positive lookbehind */    OP_ASSERTBACK,     /* 127 Positive lookbehind */
2121    OP_ASSERTBACK_NOT, /* 122 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 128 Negative lookbehind */
2122    
2123    /* ONCE, ONCE_NC, BRA, BRAPOS, CBRA, CBRAPOS, and COND must come immediately    /* ONCE, ONCE_NC, BRA, BRAPOS, CBRA, CBRAPOS, and COND must come immediately
2124    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
2125    subpattern that isn't an assertion. The POS versions must immediately follow    subpattern that isn't an assertion. The POS versions must immediately follow
2126    the non-POS versions in each case. */    the non-POS versions in each case. */
2127    
2128    OP_ONCE,           /* 123 Atomic group, contains captures */    OP_ONCE,           /* 129 Atomic group, contains captures */
2129    OP_ONCE_NC,        /* 124 Atomic group containing no captures */    OP_ONCE_NC,        /* 130 Atomic group containing no captures */
2130    OP_BRA,            /* 125 Start of non-capturing bracket */    OP_BRA,            /* 131 Start of non-capturing bracket */
2131    OP_BRAPOS,         /* 126 Ditto, with unlimited, possessive repeat */    OP_BRAPOS,         /* 132 Ditto, with unlimited, possessive repeat */
2132    OP_CBRA,           /* 127 Start of capturing bracket */    OP_CBRA,           /* 133 Start of capturing bracket */
2133    OP_CBRAPOS,        /* 128 Ditto, with unlimited, possessive repeat */    OP_CBRAPOS,        /* 134 Ditto, with unlimited, possessive repeat */
2134    OP_COND,           /* 129 Conditional group */    OP_COND,           /* 135 Conditional group */
2135    
2136    /* 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
2137    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
2138    
2139    OP_SBRA,           /* 130 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 136 Start of non-capturing bracket, check empty  */
2140    OP_SBRAPOS,        /* 131 Ditto, with unlimited, possessive repeat */    OP_SBRAPOS,        /* 137 Ditto, with unlimited, possessive repeat */
2141    OP_SCBRA,          /* 132 Start of capturing bracket, check empty */    OP_SCBRA,          /* 138 Start of capturing bracket, check empty */
2142    OP_SCBRAPOS,       /* 133 Ditto, with unlimited, possessive repeat */    OP_SCBRAPOS,       /* 139 Ditto, with unlimited, possessive repeat */
2143    OP_SCOND,          /* 134 Conditional group, check empty */    OP_SCOND,          /* 140 Conditional group, check empty */
2144    
2145    /* The next two pairs must (respectively) be kept together. */    /* The next two pairs must (respectively) be kept together. */
2146    
2147    OP_CREF,           /* 135 Used to hold a capture number as condition */    OP_CREF,           /* 141 Used to hold a capture number as condition */
2148    OP_NCREF,          /* 136 Same, but generated by a name reference*/    OP_DNCREF,         /* 142 Used to point to duplicate names as a condition */
2149    OP_RREF,           /* 137 Used to hold a recursion number as condition */    OP_RREF,           /* 143 Used to hold a recursion number as condition */
2150    OP_NRREF,          /* 138 Same, but generated by a name reference*/    OP_DNRREF,         /* 144 Used to point to duplicate names as a condition */
2151    OP_DEF,            /* 139 The DEFINE condition */    OP_DEF,            /* 145 The DEFINE condition */
2152    
2153    OP_BRAZERO,        /* 140 These two must remain together and in this */    OP_BRAZERO,        /* 146 These two must remain together and in this */
2154    OP_BRAMINZERO,     /* 141 order. */    OP_BRAMINZERO,     /* 147 order. */
2155    OP_BRAPOSZERO,     /* 142 */    OP_BRAPOSZERO,     /* 148 */
2156    
2157    /* These are backtracking control verbs */    /* These are backtracking control verbs */
2158    
2159    OP_MARK,           /* 143 always has an argument */    OP_MARK,           /* 149 always has an argument */
2160    OP_PRUNE,          /* 144 */    OP_PRUNE,          /* 150 */
2161    OP_PRUNE_ARG,      /* 145 same, but with argument */    OP_PRUNE_ARG,      /* 151 same, but with argument */
2162    OP_SKIP,           /* 146 */    OP_SKIP,           /* 152 */
2163    OP_SKIP_ARG,       /* 147 same, but with argument */    OP_SKIP_ARG,       /* 153 same, but with argument */
2164    OP_THEN,           /* 148 */    OP_THEN,           /* 154 */
2165    OP_THEN_ARG,       /* 149 same, but with argument */    OP_THEN_ARG,       /* 155 same, but with argument */
2166    OP_COMMIT,         /* 150 */    OP_COMMIT,         /* 156 */
2167    
2168    /* These are forced failure and success verbs */    /* These are forced failure and success verbs */
2169    
2170    OP_FAIL,           /* 151 */    OP_FAIL,           /* 157 */
2171    OP_ACCEPT,         /* 152 */    OP_ACCEPT,         /* 158 */
2172    OP_ASSERT_ACCEPT,  /* 153 Used inside assertions */    OP_ASSERT_ACCEPT,  /* 159 Used inside assertions */
2173    OP_CLOSE,          /* 154 Used before OP_ACCEPT to close open captures */    OP_CLOSE,          /* 160 Used before OP_ACCEPT to close open captures */
2174    
2175    /* This is used to skip a subpattern with a {0} quantifier */    /* This is used to skip a subpattern with a {0} quantifier */
2176    
2177    OP_SKIPZERO,       /* 155 */    OP_SKIPZERO,       /* 161 */
2178    
2179    /* 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
2180    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 2185  enum {
2185    
2186  /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro  /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
2187  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
2188  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
2189    pcre_dfa_exec.c that must be updated. */
2190    
2191    
2192  /* 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 2199  some cases doesn't actually use these na
2199    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
2200    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
2201    "extuni",  "\\Z", "\\z",                                        \    "extuni",  "\\Z", "\\z",                                        \
2202    "^", "^", "$", "$", "char", "chari", "not", "noti",             \    "$", "$", "^", "^", "char", "chari", "not", "noti",             \
2203    "*", "*?", "+", "+?", "?", "??",                                \    "*", "*?", "+", "+?", "?", "??",                                \
2204    "{", "{", "{",                                                  \    "{", "{", "{",                                                  \
2205    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
# Line 2167  some cases doesn't actually use these na Line 2215  some cases doesn't actually use these na
2215    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
2216    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
2217    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
2218    "class", "nclass", "xclass", "Ref", "Refi",                     \    "*+","++", "?+", "{",                                           \
2219      "class", "nclass", "xclass", "Ref", "Refi", "DnRef", "DnRefi",  \
2220    "Recurse", "Callout",                                           \    "Recurse", "Callout",                                           \
2221    "Alt", "Ket", "KetRmax", "KetRmin", "KetRpos",                  \    "Alt", "Ket", "KetRmax", "KetRmin", "KetRpos",                  \
2222    "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 2225  some cases doesn't actually use these na
2225    "Cond",                                                         \    "Cond",                                                         \
2226    "SBra", "SBraPos", "SCBra", "SCBraPos",                         \    "SBra", "SBraPos", "SCBra", "SCBraPos",                         \
2227    "SCond",                                                        \    "SCond",                                                        \
2228    "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \    "Cond ref", "Cond dnref", "Cond rec", "Cond dnrec", "Cond def", \
2229    "Brazero", "Braminzero", "Braposzero",                          \    "Brazero", "Braminzero", "Braposzero",                          \
2230    "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \    "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \
2231    "*THEN", "*THEN", "*COMMIT", "*FAIL",                           \    "*THEN", "*THEN", "*COMMIT", "*FAIL",                           \
# Line 2201  in UTF-8 mode. The code that uses this t Line 2250  in UTF-8 mode. The code that uses this t
2250    3, 3,                          /* \P, \p                                 */ \    3, 3,                          /* \P, \p                                 */ \
2251    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
2252    1,                             /* \X                                     */ \    1,                             /* \X                                     */ \
2253    1, 1, 1, 1, 1, 1,              /* \Z, \z, ^, ^M, $, $M                   */ \    1, 1, 1, 1, 1, 1,              /* \Z, \z, $, $M ^, ^M                    */ \
2254    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
2255    2,                             /* Chari  - the minimum length            */ \    2,                             /* Chari  - the minimum length            */ \
2256    2,                             /* not                                    */ \    2,                             /* not                                    */ \
# Line 2232  in UTF-8 mode. The code that uses this t Line 2281  in UTF-8 mode. The code that uses this t
2281    /* Character class & ref repeats                                         */ \    /* Character class & ref repeats                                         */ \
2282    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \
2283    1+2*IMM2_SIZE, 1+2*IMM2_SIZE,  /* CRRANGE, CRMINRANGE                    */ \    1+2*IMM2_SIZE, 1+2*IMM2_SIZE,  /* CRRANGE, CRMINRANGE                    */ \
2284      1, 1, 1, 1+2*IMM2_SIZE,        /* Possessive *+, ++, ?+, CRPOSRANGE      */ \
2285    1+(32/sizeof(pcre_uchar)),     /* CLASS                                  */ \    1+(32/sizeof(pcre_uchar)),     /* CLASS                                  */ \
2286    1+(32/sizeof(pcre_uchar)),     /* NCLASS                                 */ \    1+(32/sizeof(pcre_uchar)),     /* NCLASS                                 */ \
2287    0,                             /* XCLASS - variable length               */ \    0,                             /* XCLASS - variable length               */ \
2288    1+IMM2_SIZE,                   /* REF                                    */ \    1+IMM2_SIZE,                   /* REF                                    */ \
2289    1+IMM2_SIZE,                   /* REFI                                   */ \    1+IMM2_SIZE,                   /* REFI                                   */ \
2290      1+2*IMM2_SIZE,                 /* DNREF                                  */ \
2291      1+2*IMM2_SIZE,                 /* DNREFI                                 */ \
2292    1+LINK_SIZE,                   /* RECURSE                                */ \    1+LINK_SIZE,                   /* RECURSE                                */ \
2293    2+2*LINK_SIZE,                 /* CALLOUT                                */ \    2+2*LINK_SIZE,                 /* CALLOUT                                */ \
2294    1+LINK_SIZE,                   /* Alt                                    */ \    1+LINK_SIZE,                   /* Alt                                    */ \
# Line 2261  in UTF-8 mode. The code that uses this t Line 2313  in UTF-8 mode. The code that uses this t
2313    1+LINK_SIZE+IMM2_SIZE,         /* SCBRA                                  */ \    1+LINK_SIZE+IMM2_SIZE,         /* SCBRA                                  */ \
2314    1+LINK_SIZE+IMM2_SIZE,         /* SCBRAPOS                               */ \    1+LINK_SIZE+IMM2_SIZE,         /* SCBRAPOS                               */ \
2315    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
2316    1+IMM2_SIZE, 1+IMM2_SIZE,      /* CREF, NCREF                            */ \    1+IMM2_SIZE, 1+2*IMM2_SIZE,    /* CREF, DNCREF                           */ \
2317    1+IMM2_SIZE, 1+IMM2_SIZE,      /* RREF, NRREF                            */ \    1+IMM2_SIZE, 1+2*IMM2_SIZE,    /* RREF, DNRREF                           */ \
2318    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
2319    1, 1, 1,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */ \    1, 1, 1,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */ \
2320    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 2323  in UTF-8 mode. The code that uses this t
2323    1, 1, 1, 1,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */ \    1, 1, 1, 1,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */ \
2324    1+IMM2_SIZE, 1                 /* CLOSE, SKIPZERO                        */    1+IMM2_SIZE, 1                 /* CLOSE, SKIPZERO                        */
2325    
2326  /* 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. */  
2327    
2328  #define RREF_ANY  0xffff  #define RREF_ANY  0xffff
2329    
# Line 2287  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4, Line 2338  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
2338         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
2339         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
2340         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,
2341         ERR70, ERR71, ERR72, ERR73, ERR74, ERR75, ERR76, ERR77, ERR78, ERRCOUNT };         ERR70, ERR71, ERR72, ERR73, ERR74, ERR75, ERR76, ERR77, ERR78, ERR79,
2342           ERR80, ERR81, ERR82, ERR83, ERR84, ERRCOUNT };
2343    
2344  /* JIT compiling modes. The function list is indexed by them. */  /* JIT compiling modes. The function list is indexed by them. */
2345    
2346  enum { JIT_COMPILE, JIT_PARTIAL_SOFT_COMPILE, JIT_PARTIAL_HARD_COMPILE,  enum { JIT_COMPILE, JIT_PARTIAL_SOFT_COMPILE, JIT_PARTIAL_HARD_COMPILE,
2347         JIT_NUMBER_OF_COMPILE_MODES };         JIT_NUMBER_OF_COMPILE_MODES };
2348    
# Line 2407  typedef struct open_capitem { Line 2460  typedef struct open_capitem {
2460    pcre_uint16 flag;             /* Set TRUE if recursive back ref */    pcre_uint16 flag;             /* Set TRUE if recursive back ref */
2461  } open_capitem;  } open_capitem;
2462    
2463    /* Structure for building a list of named groups during the first pass of
2464    compiling. */
2465    
2466    typedef struct named_group {
2467      const pcre_uchar  *name;          /* Points to the name in the pattern */
2468      int                length;        /* Length of the name */
2469      pcre_uint32        number;        /* Group number */
2470    } named_group;
2471    
2472  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
2473  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
2474    
# Line 2419  typedef struct compile_data { Line 2481  typedef struct compile_data {
2481    const pcre_uchar *start_code;     /* The start of the compiled code */    const pcre_uchar *start_code;     /* The start of the compiled code */
2482    const pcre_uchar *start_pattern;  /* The start of the pattern */    const pcre_uchar *start_pattern;  /* The start of the pattern */
2483    const pcre_uchar *end_pattern;    /* The end of the pattern */    const pcre_uchar *end_pattern;    /* The end of the pattern */
   open_capitem *open_caps;          /* Chain of open capture items */  
2484    pcre_uchar *hwm;                  /* High watermark of workspace */    pcre_uchar *hwm;                  /* High watermark of workspace */
2485      open_capitem *open_caps;          /* Chain of open capture items */
2486      named_group *named_groups;        /* Points to vector in pre-compile */
2487    pcre_uchar *name_table;           /* The name/number table */    pcre_uchar *name_table;           /* The name/number table */
2488    int  names_found;                 /* Number of entries so far */    int  names_found;                 /* Number of entries so far */
2489    int  name_entry_size;             /* Size of each entry */    int  name_entry_size;             /* Size of each entry */
2490      int  named_group_list_size;       /* Number of entries in the list */
2491    int  workspace_size;              /* Size of workspace */    int  workspace_size;              /* Size of workspace */
2492    unsigned int bracount;            /* Count of capturing parens as we compile */    unsigned int bracount;            /* Count of capturing parens as we compile */
2493    int  final_bracount;              /* Saved value after first pass */    int  final_bracount;              /* Saved value after first pass */
2494    int  max_lookbehind;              /* Maximum lookbehind (characters) */    int  max_lookbehind;              /* Maximum lookbehind (characters) */
2495    int  top_backref;                 /* Maximum back reference */    int  top_backref;                 /* Maximum back reference */
2496    unsigned int backref_map;         /* Bitmap of low back refs */    unsigned int backref_map;         /* Bitmap of low back refs */
2497      unsigned int namedrefcount;       /* Number of backreferences by name */
2498      int  parens_depth;                /* Depth of nested parentheses */
2499    int  assert_depth;                /* Depth of nested assertions */    int  assert_depth;                /* Depth of nested assertions */
2500    pcre_uint32 external_options;     /* External (initial) options */    pcre_uint32 external_options;     /* External (initial) options */
2501    pcre_uint32 external_flags;       /* External flag bits to be set */    pcre_uint32 external_flags;       /* External flag bits to be set */
# Line 2437  typedef struct compile_data { Line 2503  typedef struct compile_data {
2503    BOOL had_accept;                  /* (*ACCEPT) encountered */    BOOL had_accept;                  /* (*ACCEPT) encountered */
2504    BOOL had_pruneorskip;             /* (*PRUNE) or (*SKIP) encountered */    BOOL had_pruneorskip;             /* (*PRUNE) or (*SKIP) encountered */
2505    BOOL check_lookbehind;            /* Lookbehinds need later checking */    BOOL check_lookbehind;            /* Lookbehinds need later checking */
2506      BOOL dupnames;                    /* Duplicate names exist */
2507    int  nltype;                      /* Newline type */    int  nltype;                      /* Newline type */
2508    int  nllen;                       /* Newline string length */    int  nllen;                       /* Newline string length */
2509    pcre_uchar nl[4];                 /* Newline string when fixed length */    pcre_uchar nl[4];                 /* Newline string when fixed length */

Legend:
Removed from v.1348  
changed lines
  Added in v.1408

  ViewVC Help
Powered by ViewVC 1.1.5