/[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 392 by ph10, Tue Mar 17 21:30:30 2009 UTC revision 459 by ph10, Sun Oct 4 09:21:39 2009 UTC
# Line 58  script prevents both being selected, but Line 58  script prevents both being selected, but
58  #error The use of both EBCDIC and SUPPORT_UTF8 is not supported.  #error The use of both EBCDIC and SUPPORT_UTF8 is not supported.
59  #endif  #endif
60    
61    /* If SUPPORT_UCP is defined, SUPPORT_UTF8 must also be defined. The
62    "configure" script ensures this, but not everybody uses "configure". */
63    
64    #if defined SUPPORT_UCP && !defined SUPPORT_UTF8
65    #define SUPPORT_UTF8 1
66    #endif
67    
68  /* Use a macro for debugging printing, 'cause that eliminates the use of #ifdef  /* Use a macro for debugging printing, 'cause that eliminates the use of #ifdef
69  inline, and there are *still* stupid compilers about that don't like indented  inline, and there are *still* stupid compilers about that don't like indented
70  pre-processor statements, or at least there were when I first wrote this. After  pre-processor statements, or at least there were when I first wrote this. After
# Line 528  Standard C system should have one. */ Line 535  Standard C system should have one. */
535    
536  /* Private flags containing information about the compiled regex. They used to  /* Private flags containing information about the compiled regex. They used to
537  live at the top end of the options word, but that got almost full, so now they  live at the top end of the options word, but that got almost full, so now they
538  are in a 16-bit flags word. */  are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as
539    the restrictions on partial matching have been lifted. It remains for backwards
540    compatibility. */
541    
542  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */
543  #define PCRE_FIRSTSET      0x0002  /* first_byte is set */  #define PCRE_FIRSTSET      0x0002  /* first_byte is set */
# Line 540  are in a 16-bit flags word. */ Line 549  are in a 16-bit flags word. */
549  /* Options for the "extra" block produced by pcre_study(). */  /* Options for the "extra" block produced by pcre_study(). */
550    
551  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */
552    #define PCRE_STUDY_MINLEN   0x02     /* a minimum length field exists */
553    
554  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
555  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
# Line 555  time, run time, or study time, respectiv Line 565  time, run time, or study time, respectiv
565     PCRE_JAVASCRIPT_COMPAT)     PCRE_JAVASCRIPT_COMPAT)
566    
567  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
568    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
569     PCRE_PARTIAL|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_NEWLINE_BITS| \
570     PCRE_NO_START_OPTIMIZE)     PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
571    
572  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
573    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
574     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS| \     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_DFA_SHORTEST| \
575     PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)     PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
576       PCRE_NO_START_OPTIMIZE)
577    
578  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
579    
# Line 874  so that PCRE works on both ASCII and EBC Line 885  so that PCRE works on both ASCII and EBC
885  #define STRING_ANYCRLF_RIGHTPAR     "ANYCRLF)"  #define STRING_ANYCRLF_RIGHTPAR     "ANYCRLF)"
886  #define STRING_BSR_ANYCRLF_RIGHTPAR "BSR_ANYCRLF)"  #define STRING_BSR_ANYCRLF_RIGHTPAR "BSR_ANYCRLF)"
887  #define STRING_BSR_UNICODE_RIGHTPAR "BSR_UNICODE)"  #define STRING_BSR_UNICODE_RIGHTPAR "BSR_UNICODE)"
888    #define STRING_UTF8_RIGHTPAR        "UTF8)"
889    
890  #else  /* SUPPORT_UTF8 */  #else  /* SUPPORT_UTF8 */
891    
# Line 1125  only. */ Line 1137  only. */
1137  #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
1138  #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
1139  #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
1140    #define STRING_UTF8_RIGHTPAR        STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1141    
1142  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF8 */
1143    
# Line 1334  enum { Line 1347  enum {
1347    OP_SCBRA,          /* 98 Start of capturing bracket, check empty */    OP_SCBRA,          /* 98 Start of capturing bracket, check empty */
1348    OP_SCOND,          /* 99 Conditional group, check empty */    OP_SCOND,          /* 99 Conditional group, check empty */
1349    
1350      /* The next two pairs must (respectively) be kept together. */
1351    
1352    OP_CREF,           /* 100 Used to hold a capture number as condition */    OP_CREF,           /* 100 Used to hold a capture number as condition */
1353    OP_RREF,           /* 101 Used to hold a recursion number as condition */    OP_NCREF,          /* 101 Same, but generaged by a name reference*/
1354    OP_DEF,            /* 102 The DEFINE condition */    OP_RREF,           /* 102 Used to hold a recursion number as condition */
1355      OP_NRREF,          /* 103 Same, but generaged by a name reference*/
1356      OP_DEF,            /* 104 The DEFINE condition */
1357    
1358    OP_BRAZERO,        /* 103 These two must remain together and in this */    OP_BRAZERO,        /* 105 These two must remain together and in this */
1359    OP_BRAMINZERO,     /* 104 order. */    OP_BRAMINZERO,     /* 106 order. */
1360    
1361    /* These are backtracking control verbs */    /* These are backtracking control verbs */
1362    
1363    OP_PRUNE,          /* 105 */    OP_PRUNE,          /* 107 */
1364    OP_SKIP,           /* 106 */    OP_SKIP,           /* 108 */
1365    OP_THEN,           /* 107 */    OP_THEN,           /* 109 */
1366    OP_COMMIT,         /* 108 */    OP_COMMIT,         /* 110 */
1367    
1368    /* These are forced failure and success verbs */    /* These are forced failure and success verbs */
1369    
1370    OP_FAIL,           /* 109 */    OP_FAIL,           /* 111 */
1371    OP_ACCEPT,         /* 110 */    OP_ACCEPT,         /* 112 */
1372      OP_CLOSE,          /* 113 Used before OP_ACCEPT to close open captures */
1373    
1374    /* This is used to skip a subpattern with a {0} quantifier */    /* This is used to skip a subpattern with a {0} quantifier */
1375    
1376    OP_SKIPZERO        /* 111 */    OP_SKIPZERO        /* 114 */
1377  };  };
1378    
1379    
# Line 1379  for debugging. The macro is referenced o Line 1397  for debugging. The macro is referenced o
1397    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
1398    "AssertB", "AssertB not", "Reverse",                            \    "AssertB", "AssertB not", "Reverse",                            \
1399    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
1400    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \    "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \
1401      "Brazero", "Braminzero",                                        \
1402    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \
1403    "Skip zero"    "Close", "Skip zero"
1404    
1405    
1406  /* This macro defines the length of fixed length operations in the compiled  /* This macro defines the length of fixed length operations in the compiled
# Line 1441  in UTF-8 mode. The code that uses this t Line 1460  in UTF-8 mode. The code that uses this t
1460    1+LINK_SIZE,                   /* SBRA                                   */ \    1+LINK_SIZE,                   /* SBRA                                   */ \
1461    3+LINK_SIZE,                   /* SCBRA                                  */ \    3+LINK_SIZE,                   /* SCBRA                                  */ \
1462    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
1463    3,                             /* CREF                                   */ \    3, 3,                          /* CREF, NCREF                            */ \
1464    3,                             /* RREF                                   */ \    3, 3,                          /* RREF, NRREF                            */ \
1465    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
1466    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
1467    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \
1468    1, 1, 1                        /* FAIL, ACCEPT, SKIPZERO                 */    1, 1, 3, 1                     /* FAIL, ACCEPT, CLOSE, SKIPZERO          */
1469    
1470    
1471  /* A magic value for OP_RREF to indicate the "any recursion" condition. */  /* A magic value for OP_RREF and OP_NRREF to indicate the "any recursion"
1472    condition. */
1473    
1474  #define RREF_ANY  0xffff  #define RREF_ANY  0xffff
1475    
# Line 1462  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4, Line 1482  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
1482         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
1483         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
1484         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
1485         ERR60, ERR61, ERR62, ERR63, ERR64 };         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65 };
1486    
1487  /* The real format of the start of the pcre block; the index of names and the  /* The real format of the start of the pcre block; the index of names and the
1488  code vector run on as long as necessary after the end. We store an explicit  code vector run on as long as necessary after the end. We store an explicit
# Line 1478  Because people can now save and re-use c Line 1498  Because people can now save and re-use c
1498  structure should be made at the end, and something earlier (e.g. a new  structure should be made at the end, and something earlier (e.g. a new
1499  flag in the options or one of the dummy fields) should indicate that the new  flag in the options or one of the dummy fields) should indicate that the new
1500  fields are present. Currently PCRE always sets the dummy fields to zero.  fields are present. Currently PCRE always sets the dummy fields to zero.
1501  NOTE NOTE NOTE:  NOTE NOTE NOTE
1502  */  */
1503    
1504  typedef struct real_pcre {  typedef struct real_pcre {
# Line 1505  remark (see NOTE above) about extending Line 1525  remark (see NOTE above) about extending
1525    
1526  typedef struct pcre_study_data {  typedef struct pcre_study_data {
1527    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1528    pcre_uint32 options;    pcre_uint32 flags;              /* Private flags */
1529    uschar start_bits[32];    uschar start_bits[32];          /* Starting char bits */
1530      pcre_uint32 minlength;          /* Minimum subject length */
1531  } pcre_study_data;  } pcre_study_data;
1532    
1533    /* Structure for building a chain of open capturing subpatterns during
1534    compiling, so that instructions to close them can be compiled when (*ACCEPT) is
1535    encountered. */
1536    
1537    typedef struct open_capitem {
1538      struct open_capitem *next;    /* Chain link */
1539      pcre_uint16 number;           /* Capture number */
1540    } open_capitem;
1541    
1542  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
1543  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
1544    
# Line 1521  typedef struct compile_data { Line 1551  typedef struct compile_data {
1551    const uschar *start_code;     /* The start of the compiled code */    const uschar *start_code;     /* The start of the compiled code */
1552    const uschar *start_pattern;  /* The start of the pattern */    const uschar *start_pattern;  /* The start of the pattern */
1553    const uschar *end_pattern;    /* The end of the pattern */    const uschar *end_pattern;    /* The end of the pattern */
1554      open_capitem *open_caps;      /* Chain of open capture items */
1555    uschar *hwm;                  /* High watermark of workspace */    uschar *hwm;                  /* High watermark of workspace */
1556    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
1557    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
# Line 1533  typedef struct compile_data { Line 1564  typedef struct compile_data {
1564    int  external_flags;          /* External flag bits to be set */    int  external_flags;          /* External flag bits to be set */
1565    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
1566    BOOL had_accept;              /* (*ACCEPT) encountered */    BOOL had_accept;              /* (*ACCEPT) encountered */
1567      BOOL check_lookbehind;        /* Lookbehinds need later checking */
1568    int  nltype;                  /* Newline type */    int  nltype;                  /* Newline type */
1569    int  nllen;                   /* Newline string length */    int  nllen;                   /* Newline string length */
1570    uschar nl[4];                 /* Newline string when fixed length */    uschar nl[4];                 /* Newline string when fixed length */
# Line 1556  typedef struct recursion_info { Line 1588  typedef struct recursion_info {
1588    USPTR save_start;             /* Old value of mstart */    USPTR save_start;             /* Old value of mstart */
1589    int *offset_save;             /* Pointer to start of saved offsets */    int *offset_save;             /* Pointer to start of saved offsets */
1590    int saved_max;                /* Number of saved offsets */    int saved_max;                /* Number of saved offsets */
1591      int offset_top;               /* Current value of offset_top */
1592  } recursion_info;  } recursion_info;
1593    
1594  /* Structure for building a chain of data for holding the values of the subject  /* Structure for building a chain of data for holding the values of the subject
# Line 1580  typedef struct match_data { Line 1613  typedef struct match_data {
1613    int    offset_max;            /* The maximum usable for return data */    int    offset_max;            /* The maximum usable for return data */
1614    int    nltype;                /* Newline type */    int    nltype;                /* Newline type */
1615    int    nllen;                 /* Newline string length */    int    nllen;                 /* Newline string length */
1616      int    name_count;            /* Number of names in name table */
1617      int    name_entry_size;       /* Size of entry in names table */
1618      uschar *name_table;           /* Table of names */
1619    uschar nl[4];                 /* Newline string when fixed */    uschar nl[4];                 /* Newline string when fixed */
1620    const uschar *lcc;            /* Points to lower casing table */    const uschar *lcc;            /* Points to lower casing table */
1621    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
# Line 1590  typedef struct match_data { Line 1626  typedef struct match_data {
1626    BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */    BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */
1627    BOOL   endonly;               /* Dollar not before final \n */    BOOL   endonly;               /* Dollar not before final \n */
1628    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notempty;              /* Empty string match not wanted */
1629    BOOL   partial;               /* PARTIAL flag */    BOOL   notempty_atstart;      /* Empty string match at start not wanted */
1630    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
1631    BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */    BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */
1632    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
# Line 1598  typedef struct match_data { Line 1634  typedef struct match_data {
1634    USPTR  end_subject;           /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
1635    USPTR  start_match_ptr;       /* Start of matched string */    USPTR  start_match_ptr;       /* Start of matched string */
1636    USPTR  end_match_ptr;         /* Subject position at end match */    USPTR  end_match_ptr;         /* Subject position at end match */
1637      USPTR  start_used_ptr;        /* Earliest consulted character */
1638      int    partial;               /* PARTIAL options */
1639    int    end_offset_top;        /* Highwater mark at end of match */    int    end_offset_top;        /* Highwater mark at end of match */
1640    int    capture_last;          /* Most recent capture number */    int    capture_last;          /* Most recent capture number */
1641    int    start_offset;          /* The start offset value */    int    start_offset;          /* The start offset value */
# Line 1614  typedef struct dfa_match_data { Line 1652  typedef struct dfa_match_data {
1652    const uschar *start_code;     /* Start of the compiled pattern */    const uschar *start_code;     /* Start of the compiled pattern */
1653    const uschar *start_subject;  /* Start of the subject string */    const uschar *start_subject;  /* Start of the subject string */
1654    const uschar *end_subject;    /* End of subject string */    const uschar *end_subject;    /* End of subject string */
1655      const uschar *start_used_ptr; /* Earliest consulted character */
1656    const uschar *tables;         /* Character tables */    const uschar *tables;         /* Character tables */
1657      int   start_offset;           /* The start offset value */
1658    int   moptions;               /* Match options */    int   moptions;               /* Match options */
1659    int   poptions;               /* Pattern options */    int   poptions;               /* Pattern options */
1660    int    nltype;                /* Newline type */    int    nltype;                /* Newline type */
# Line 1693  extern const uschar _pcre_OP_lengths[]; Line 1733  extern const uschar _pcre_OP_lengths[];
1733  one of the exported public functions. They have to be "external" in the C  one of the exported public functions. They have to be "external" in the C
1734  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1735    
1736  extern BOOL         _pcre_is_newline(const uschar *, int, const uschar *,  extern const uschar *_pcre_find_bracket(const uschar *, BOOL, int);
1737                        int *, BOOL);  extern BOOL          _pcre_is_newline(const uschar *, int, const uschar *,
1738  extern int          _pcre_ord2utf8(int, uschar *);                         int *, BOOL);
1739  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  extern int           _pcre_ord2utf8(int, uschar *);
1740                        const pcre_study_data *, pcre_study_data *);  extern real_pcre    *_pcre_try_flipped(const real_pcre *, real_pcre *,
1741  extern int          _pcre_valid_utf8(const uschar *, int);                         const pcre_study_data *, pcre_study_data *);
1742  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,  extern int           _pcre_valid_utf8(const uschar *, int);
1743                        int *, BOOL);  extern BOOL          _pcre_was_newline(const uschar *, int, const uschar *,
1744  extern BOOL         _pcre_xclass(int, const uschar *);                         int *, BOOL);
1745    extern BOOL          _pcre_xclass(int, const uschar *);
1746    
1747    
1748  /* Unicode character database (UCD) */  /* Unicode character database (UCD) */

Legend:
Removed from v.392  
changed lines
  Added in v.459

  ViewVC Help
Powered by ViewVC 1.1.5