/[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 412 by ph10, Sat Apr 11 10:34:37 2009 UTC revision 462 by ph10, Sat Oct 17 19:55:02 2009 UTC
# Line 535  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 547  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 562  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 1206  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, E Line 1210  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, E
1210  OP_EOD must correspond in order to the list of escapes immediately above.  OP_EOD must correspond in order to the list of escapes immediately above.
1211    
1212  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
1213  that follow must also be updated to match. There is also a table called  that follow must also be updated to match. There are also tables called
1214  "coptable" in pcre_dfa_exec.c that must be updated. */  "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1215    
1216  enum {  enum {
1217    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 1343  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    /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
1380    definitions that follow must also be updated to match. There are also tables
1381    called "coptable" cna "poptable" in pcre_dfa_exec.c that must be updated. */
1382    
1383    
1384  /* 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
1385  for debugging. The macro is referenced only in pcre_printint.c. */  for debugging. The macro is referenced only in pcre_printint.c. */
# Line 1388  for debugging. The macro is referenced o Line 1401  for debugging. The macro is referenced o
1401    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
1402    "AssertB", "AssertB not", "Reverse",                            \    "AssertB", "AssertB not", "Reverse",                            \
1403    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
1404    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \    "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \
1405      "Brazero", "Braminzero",                                        \
1406    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \
1407    "Skip zero"    "Close", "Skip zero"
1408    
1409    
1410  /* 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 1450  in UTF-8 mode. The code that uses this t Line 1464  in UTF-8 mode. The code that uses this t
1464    1+LINK_SIZE,                   /* SBRA                                   */ \    1+LINK_SIZE,                   /* SBRA                                   */ \
1465    3+LINK_SIZE,                   /* SCBRA                                  */ \    3+LINK_SIZE,                   /* SCBRA                                  */ \
1466    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
1467    3,                             /* CREF                                   */ \    3, 3,                          /* CREF, NCREF                            */ \
1468    3,                             /* RREF                                   */ \    3, 3,                          /* RREF, NRREF                            */ \
1469    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
1470    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
1471    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \
1472    1, 1, 1                        /* FAIL, ACCEPT, SKIPZERO                 */    1, 1, 3, 1                     /* FAIL, ACCEPT, CLOSE, SKIPZERO          */
1473    
1474    
1475  /* 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"
1476    condition. */
1477    
1478  #define RREF_ANY  0xffff  #define RREF_ANY  0xffff
1479    
# Line 1471  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4, Line 1486  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
1486         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
1487         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
1488         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
1489         ERR60, ERR61, ERR62, ERR63, ERR64 };         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65 };
1490    
1491  /* 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
1492  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 1487  Because people can now save and re-use c Line 1502  Because people can now save and re-use c
1502  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
1503  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
1504  fields are present. Currently PCRE always sets the dummy fields to zero.  fields are present. Currently PCRE always sets the dummy fields to zero.
1505  NOTE NOTE NOTE:  NOTE NOTE NOTE
1506  */  */
1507    
1508  typedef struct real_pcre {  typedef struct real_pcre {
# Line 1514  remark (see NOTE above) about extending Line 1529  remark (see NOTE above) about extending
1529    
1530  typedef struct pcre_study_data {  typedef struct pcre_study_data {
1531    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1532    pcre_uint32 options;    pcre_uint32 flags;              /* Private flags */
1533    uschar start_bits[32];    uschar start_bits[32];          /* Starting char bits */
1534      pcre_uint32 minlength;          /* Minimum subject length */
1535  } pcre_study_data;  } pcre_study_data;
1536    
1537    /* Structure for building a chain of open capturing subpatterns during
1538    compiling, so that instructions to close them can be compiled when (*ACCEPT) is
1539    encountered. */
1540    
1541    typedef struct open_capitem {
1542      struct open_capitem *next;    /* Chain link */
1543      pcre_uint16 number;           /* Capture number */
1544    } open_capitem;
1545    
1546  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
1547  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
1548    
# Line 1530  typedef struct compile_data { Line 1555  typedef struct compile_data {
1555    const uschar *start_code;     /* The start of the compiled code */    const uschar *start_code;     /* The start of the compiled code */
1556    const uschar *start_pattern;  /* The start of the pattern */    const uschar *start_pattern;  /* The start of the pattern */
1557    const uschar *end_pattern;    /* The end of the pattern */    const uschar *end_pattern;    /* The end of the pattern */
1558      open_capitem *open_caps;      /* Chain of open capture items */
1559    uschar *hwm;                  /* High watermark of workspace */    uschar *hwm;                  /* High watermark of workspace */
1560    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
1561    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
# Line 1542  typedef struct compile_data { Line 1568  typedef struct compile_data {
1568    int  external_flags;          /* External flag bits to be set */    int  external_flags;          /* External flag bits to be set */
1569    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
1570    BOOL had_accept;              /* (*ACCEPT) encountered */    BOOL had_accept;              /* (*ACCEPT) encountered */
1571      BOOL check_lookbehind;        /* Lookbehinds need later checking */
1572    int  nltype;                  /* Newline type */    int  nltype;                  /* Newline type */
1573    int  nllen;                   /* Newline string length */    int  nllen;                   /* Newline string length */
1574    uschar nl[4];                 /* Newline string when fixed length */    uschar nl[4];                 /* Newline string when fixed length */
# Line 1565  typedef struct recursion_info { Line 1592  typedef struct recursion_info {
1592    USPTR save_start;             /* Old value of mstart */    USPTR save_start;             /* Old value of mstart */
1593    int *offset_save;             /* Pointer to start of saved offsets */    int *offset_save;             /* Pointer to start of saved offsets */
1594    int saved_max;                /* Number of saved offsets */    int saved_max;                /* Number of saved offsets */
1595      int save_offset_top;          /* Current value of offset_top */
1596  } recursion_info;  } recursion_info;
1597    
1598  /* 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 1589  typedef struct match_data { Line 1617  typedef struct match_data {
1617    int    offset_max;            /* The maximum usable for return data */    int    offset_max;            /* The maximum usable for return data */
1618    int    nltype;                /* Newline type */    int    nltype;                /* Newline type */
1619    int    nllen;                 /* Newline string length */    int    nllen;                 /* Newline string length */
1620      int    name_count;            /* Number of names in name table */
1621      int    name_entry_size;       /* Size of entry in names table */
1622      uschar *name_table;           /* Table of names */
1623    uschar nl[4];                 /* Newline string when fixed */    uschar nl[4];                 /* Newline string when fixed */
1624    const uschar *lcc;            /* Points to lower casing table */    const uschar *lcc;            /* Points to lower casing table */
1625    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
# Line 1599  typedef struct match_data { Line 1630  typedef struct match_data {
1630    BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */    BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */
1631    BOOL   endonly;               /* Dollar not before final \n */    BOOL   endonly;               /* Dollar not before final \n */
1632    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notempty;              /* Empty string match not wanted */
1633    BOOL   partial;               /* PARTIAL flag */    BOOL   notempty_atstart;      /* Empty string match at start not wanted */
1634    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
1635    BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */    BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */
1636    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
# Line 1607  typedef struct match_data { Line 1638  typedef struct match_data {
1638    USPTR  end_subject;           /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
1639    USPTR  start_match_ptr;       /* Start of matched string */    USPTR  start_match_ptr;       /* Start of matched string */
1640    USPTR  end_match_ptr;         /* Subject position at end match */    USPTR  end_match_ptr;         /* Subject position at end match */
1641      USPTR  start_used_ptr;        /* Earliest consulted character */
1642      int    partial;               /* PARTIAL options */
1643    int    end_offset_top;        /* Highwater mark at end of match */    int    end_offset_top;        /* Highwater mark at end of match */
1644    int    capture_last;          /* Most recent capture number */    int    capture_last;          /* Most recent capture number */
1645    int    start_offset;          /* The start offset value */    int    start_offset;          /* The start offset value */
# Line 1623  typedef struct dfa_match_data { Line 1656  typedef struct dfa_match_data {
1656    const uschar *start_code;     /* Start of the compiled pattern */    const uschar *start_code;     /* Start of the compiled pattern */
1657    const uschar *start_subject;  /* Start of the subject string */    const uschar *start_subject;  /* Start of the subject string */
1658    const uschar *end_subject;    /* End of subject string */    const uschar *end_subject;    /* End of subject string */
1659      const uschar *start_used_ptr; /* Earliest consulted character */
1660    const uschar *tables;         /* Character tables */    const uschar *tables;         /* Character tables */
1661      int   start_offset;           /* The start offset value */
1662    int   moptions;               /* Match options */    int   moptions;               /* Match options */
1663    int   poptions;               /* Pattern options */    int   poptions;               /* Pattern options */
1664    int    nltype;                /* Newline type */    int    nltype;                /* Newline type */
# Line 1702  extern const uschar _pcre_OP_lengths[]; Line 1737  extern const uschar _pcre_OP_lengths[];
1737  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
1738  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1739    
1740  extern BOOL         _pcre_is_newline(const uschar *, int, const uschar *,  extern const uschar *_pcre_find_bracket(const uschar *, BOOL, int);
1741                        int *, BOOL);  extern BOOL          _pcre_is_newline(const uschar *, int, const uschar *,
1742  extern int          _pcre_ord2utf8(int, uschar *);                         int *, BOOL);
1743  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  extern int           _pcre_ord2utf8(int, uschar *);
1744                        const pcre_study_data *, pcre_study_data *);  extern real_pcre    *_pcre_try_flipped(const real_pcre *, real_pcre *,
1745  extern int          _pcre_valid_utf8(const uschar *, int);                         const pcre_study_data *, pcre_study_data *);
1746  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,  extern int           _pcre_valid_utf8(const uschar *, int);
1747                        int *, BOOL);  extern BOOL          _pcre_was_newline(const uschar *, int, const uschar *,
1748  extern BOOL         _pcre_xclass(int, const uschar *);                         int *, BOOL);
1749    extern BOOL          _pcre_xclass(int, const uschar *);
1750    
1751    
1752  /* Unicode character database (UCD) */  /* Unicode character database (UCD) */

Legend:
Removed from v.412  
changed lines
  Added in v.462

  ViewVC Help
Powered by ViewVC 1.1.5