/[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 454 by ph10, Tue Sep 22 09:42:11 2009 UTC revision 487 by ph10, Wed Jan 6 10:26:55 2010 UTC
# Line 7  Line 7 
7  and semantics are as close as possible to those of the Perl 5 language.  and semantics are as close as possible to those of the Perl 5 language.
8    
9                         Written by Philip Hazel                         Written by Philip Hazel
10             Copyright (c) 1997-2009 University of Cambridge             Copyright (c) 1997-2010 University of Cambridge
11    
12  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
13  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 45  functions whose names all begin with "_p Line 45  functions whose names all begin with "_p
45  #ifndef PCRE_INTERNAL_H  #ifndef PCRE_INTERNAL_H
46  #define PCRE_INTERNAL_H  #define PCRE_INTERNAL_H
47    
48  /* Define DEBUG to get debugging output on stdout. */  /* Define PCRE_DEBUG to get debugging output on stdout. */
49    
50  #if 0  #if 0
51  #define DEBUG  #define PCRE_DEBUG
52  #endif  #endif
53    
54  /* We do not support both EBCDIC and UTF-8 at the same time. The "configure"  /* We do not support both EBCDIC and UTF-8 at the same time. The "configure"
# Line 74  It turns out that the Mac Debugging.h he Line 74  It turns out that the Mac Debugging.h he
74  be absolutely sure we get our version. */  be absolutely sure we get our version. */
75    
76  #undef DPRINTF  #undef DPRINTF
77  #ifdef DEBUG  #ifdef PCRE_DEBUG
78  #define DPRINTF(p) printf p  #define DPRINTF(p) printf p
79  #else  #else
80  #define DPRINTF(p) /* Nothing */  #define DPRINTF(p) /* Nothing */
# Line 86  setjmp and stdarg are used is when NO_RE Line 86  setjmp and stdarg are used is when NO_RE
86    
87  #include <ctype.h>  #include <ctype.h>
88  #include <limits.h>  #include <limits.h>
 #include <setjmp.h>  
 #include <stdarg.h>  
89  #include <stddef.h>  #include <stddef.h>
90  #include <stdio.h>  #include <stdio.h>
91  #include <stdlib.h>  #include <stdlib.h>
# Line 186  preprocessor time in standard C environm Line 184  preprocessor time in standard C environm
184    #error Cannot determine a type for 32-bit unsigned integers    #error Cannot determine a type for 32-bit unsigned integers
185  #endif  #endif
186    
187    /* When checking for integer overflow in pcre_compile(), we need to handle
188    large integers. If a 64-bit integer type is available, we can use that.
189    Otherwise we have to cast to double, which of course requires floating point
190    arithmetic. Handle this by defining a macro for the appropriate type. If
191    stdint.h is available, include it; it may define INT64_MAX. The macro int64_t
192    may be set by "configure". */
193    
194    #if HAVE_STDINT_H
195    #include <stdint.h>
196    #endif
197    
198    #if defined INT64_MAX || defined int64_t
199    #define INT64_OR_DOUBLE int64_t
200    #else
201    #define INT64_OR_DOUBLE double
202    #endif
203    
204  /* All character handling must be done as unsigned characters. Otherwise there  /* All character handling must be done as unsigned characters. Otherwise there
205  are problems with top-bit-set characters and functions such as isspace().  are problems with top-bit-set characters and functions such as isspace().
206  However, we leave the interface to the outside world as char *, because that  However, we leave the interface to the outside world as char *, because that
# Line 549  compatibility. */ Line 564  compatibility. */
564  /* Options for the "extra" block produced by pcre_study(). */  /* Options for the "extra" block produced by pcre_study(). */
565    
566  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */
567    #define PCRE_STUDY_MINLEN   0x02     /* a minimum length field exists */
568    
569  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
570  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
# Line 1209  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, E Line 1225  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, E
1225  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.
1226    
1227  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
1228  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
1229  "coptable" in pcre_dfa_exec.c that must be updated. */  "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1230    
1231  enum {  enum {
1232    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 1346  enum { Line 1362  enum {
1362    OP_SCBRA,          /* 98 Start of capturing bracket, check empty */    OP_SCBRA,          /* 98 Start of capturing bracket, check empty */
1363    OP_SCOND,          /* 99 Conditional group, check empty */    OP_SCOND,          /* 99 Conditional group, check empty */
1364    
1365      /* The next two pairs must (respectively) be kept together. */
1366    
1367    OP_CREF,           /* 100 Used to hold a capture number as condition */    OP_CREF,           /* 100 Used to hold a capture number as condition */
1368    OP_RREF,           /* 101 Used to hold a recursion number as condition */    OP_NCREF,          /* 101 Same, but generaged by a name reference*/
1369    OP_DEF,            /* 102 The DEFINE condition */    OP_RREF,           /* 102 Used to hold a recursion number as condition */
1370      OP_NRREF,          /* 103 Same, but generaged by a name reference*/
1371      OP_DEF,            /* 104 The DEFINE condition */
1372    
1373    OP_BRAZERO,        /* 103 These two must remain together and in this */    OP_BRAZERO,        /* 105 These two must remain together and in this */
1374    OP_BRAMINZERO,     /* 104 order. */    OP_BRAMINZERO,     /* 106 order. */
1375    
1376    /* These are backtracking control verbs */    /* These are backtracking control verbs */
1377    
1378    OP_PRUNE,          /* 105 */    OP_PRUNE,          /* 107 */
1379    OP_SKIP,           /* 106 */    OP_SKIP,           /* 108 */
1380    OP_THEN,           /* 107 */    OP_THEN,           /* 109 */
1381    OP_COMMIT,         /* 108 */    OP_COMMIT,         /* 110 */
1382    
1383    /* These are forced failure and success verbs */    /* These are forced failure and success verbs */
1384    
1385    OP_FAIL,           /* 109 */    OP_FAIL,           /* 111 */
1386    OP_ACCEPT,         /* 110 */    OP_ACCEPT,         /* 112 */
1387    OP_CLOSE,          /* 111 Used before OP_ACCEPT to close open captures */    OP_CLOSE,          /* 113 Used before OP_ACCEPT to close open captures */
1388    
1389    /* This is used to skip a subpattern with a {0} quantifier */    /* This is used to skip a subpattern with a {0} quantifier */
1390    
1391    OP_SKIPZERO        /* 112 */    OP_SKIPZERO        /* 114 */
1392  };  };
1393    
1394    /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
1395    definitions that follow must also be updated to match. There are also tables
1396    called "coptable" cna "poptable" in pcre_dfa_exec.c that must be updated. */
1397    
1398    
1399  /* 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
1400  for debugging. The macro is referenced only in pcre_printint.c. */  for debugging. The macro is referenced only in pcre_printint.c. */
# Line 1392  for debugging. The macro is referenced o Line 1416  for debugging. The macro is referenced o
1416    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
1417    "AssertB", "AssertB not", "Reverse",                            \    "AssertB", "AssertB not", "Reverse",                            \
1418    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
1419    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \    "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \
1420      "Brazero", "Braminzero",                                        \
1421    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \
1422    "Close", "Skip zero"    "Close", "Skip zero"
1423    
# Line 1454  in UTF-8 mode. The code that uses this t Line 1479  in UTF-8 mode. The code that uses this t
1479    1+LINK_SIZE,                   /* SBRA                                   */ \    1+LINK_SIZE,                   /* SBRA                                   */ \
1480    3+LINK_SIZE,                   /* SCBRA                                  */ \    3+LINK_SIZE,                   /* SCBRA                                  */ \
1481    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
1482    3,                             /* CREF                                   */ \    3, 3,                          /* CREF, NCREF                            */ \
1483    3,                             /* RREF                                   */ \    3, 3,                          /* RREF, NRREF                            */ \
1484    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
1485    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
1486    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \
1487    1, 1, 3, 1                     /* FAIL, ACCEPT, CLOSE, SKIPZERO          */    1, 1, 3, 1                     /* FAIL, ACCEPT, CLOSE, SKIPZERO          */
1488    
1489    
1490  /* 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"
1491    condition. */
1492    
1493  #define RREF_ANY  0xffff  #define RREF_ANY  0xffff
1494    
# Line 1475  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4, Line 1501  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
1501         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
1502         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
1503         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
1504         ERR60, ERR61, ERR62, ERR63, ERR64 };         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65 };
1505    
1506  /* 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
1507  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 1491  Because people can now save and re-use c Line 1517  Because people can now save and re-use c
1517  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
1518  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
1519  fields are present. Currently PCRE always sets the dummy fields to zero.  fields are present. Currently PCRE always sets the dummy fields to zero.
1520  NOTE NOTE NOTE:  NOTE NOTE NOTE
1521  */  */
1522    
1523  typedef struct real_pcre {  typedef struct real_pcre {
# Line 1518  remark (see NOTE above) about extending Line 1544  remark (see NOTE above) about extending
1544    
1545  typedef struct pcre_study_data {  typedef struct pcre_study_data {
1546    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1547    pcre_uint32 options;    pcre_uint32 flags;              /* Private flags */
1548    uschar start_bits[32];    uschar start_bits[32];          /* Starting char bits */
1549      pcre_uint32 minlength;          /* Minimum subject length */
1550  } pcre_study_data;  } pcre_study_data;
1551    
1552  /* Structure for building a chain of open capturing subpatterns during  /* Structure for building a chain of open capturing subpatterns during
1553  compiling, so that instructions to close them can be compiled when (*ACCEPT) is  compiling, so that instructions to close them can be compiled when (*ACCEPT) is
1554  encountered. */  encountered. */
1555    
1556  typedef struct open_capitem {  typedef struct open_capitem {
1557    struct open_capitem *next;    /* Chain link */    struct open_capitem *next;    /* Chain link */
1558    pcre_uint16 number;           /* Capture number */    pcre_uint16 number;           /* Capture number */
1559  } open_capitem;  } open_capitem;
1560    
1561  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
1562  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
# Line 1543  typedef struct compile_data { Line 1570  typedef struct compile_data {
1570    const uschar *start_code;     /* The start of the compiled code */    const uschar *start_code;     /* The start of the compiled code */
1571    const uschar *start_pattern;  /* The start of the pattern */    const uschar *start_pattern;  /* The start of the pattern */
1572    const uschar *end_pattern;    /* The end of the pattern */    const uschar *end_pattern;    /* The end of the pattern */
1573    open_capitem *open_caps;      /* Chain of open capture items */    open_capitem *open_caps;      /* Chain of open capture items */
1574    uschar *hwm;                  /* High watermark of workspace */    uschar *hwm;                  /* High watermark of workspace */
1575    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
1576    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
# Line 1556  typedef struct compile_data { Line 1583  typedef struct compile_data {
1583    int  external_flags;          /* External flag bits to be set */    int  external_flags;          /* External flag bits to be set */
1584    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
1585    BOOL had_accept;              /* (*ACCEPT) encountered */    BOOL had_accept;              /* (*ACCEPT) encountered */
1586    BOOL check_lookbehind;        /* Lookbehinds need later checking */    BOOL check_lookbehind;        /* Lookbehinds need later checking */
1587    int  nltype;                  /* Newline type */    int  nltype;                  /* Newline type */
1588    int  nllen;                   /* Newline string length */    int  nllen;                   /* Newline string length */
1589    uschar nl[4];                 /* Newline string when fixed length */    uschar nl[4];                 /* Newline string when fixed length */
# Line 1567  branches, for testing for left recursion Line 1594  branches, for testing for left recursion
1594    
1595  typedef struct branch_chain {  typedef struct branch_chain {
1596    struct branch_chain *outer;    struct branch_chain *outer;
1597    uschar *current;    uschar *current_branch;
1598  } branch_chain;  } branch_chain;
1599    
1600  /* Structure for items in a linked list that represents an explicit recursive  /* Structure for items in a linked list that represents an explicit recursive
# Line 1580  typedef struct recursion_info { Line 1607  typedef struct recursion_info {
1607    USPTR save_start;             /* Old value of mstart */    USPTR save_start;             /* Old value of mstart */
1608    int *offset_save;             /* Pointer to start of saved offsets */    int *offset_save;             /* Pointer to start of saved offsets */
1609    int saved_max;                /* Number of saved offsets */    int saved_max;                /* Number of saved offsets */
1610    int offset_top;               /* Current value of offset_top */    int save_offset_top;          /* Current value of offset_top */
1611  } recursion_info;  } recursion_info;
1612    
1613  /* 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 1605  typedef struct match_data { Line 1632  typedef struct match_data {
1632    int    offset_max;            /* The maximum usable for return data */    int    offset_max;            /* The maximum usable for return data */
1633    int    nltype;                /* Newline type */    int    nltype;                /* Newline type */
1634    int    nllen;                 /* Newline string length */    int    nllen;                 /* Newline string length */
1635      int    name_count;            /* Number of names in name table */
1636      int    name_entry_size;       /* Size of entry in names table */
1637      uschar *name_table;           /* Table of names */
1638    uschar nl[4];                 /* Newline string when fixed */    uschar nl[4];                 /* Newline string when fixed */
1639    const uschar *lcc;            /* Points to lower casing table */    const uschar *lcc;            /* Points to lower casing table */
1640    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
# Line 1722  extern const uschar _pcre_OP_lengths[]; Line 1752  extern const uschar _pcre_OP_lengths[];
1752  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
1753  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1754    
1755  extern BOOL         _pcre_is_newline(const uschar *, int, const uschar *,  extern const uschar *_pcre_find_bracket(const uschar *, BOOL, int);
1756                        int *, BOOL);  extern BOOL          _pcre_is_newline(USPTR, int, USPTR, int *, BOOL);
1757  extern int          _pcre_ord2utf8(int, uschar *);  extern int           _pcre_ord2utf8(int, uschar *);
1758  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  extern real_pcre    *_pcre_try_flipped(const real_pcre *, real_pcre *,
1759                        const pcre_study_data *, pcre_study_data *);                         const pcre_study_data *, pcre_study_data *);
1760  extern int          _pcre_valid_utf8(const uschar *, int);  extern int           _pcre_valid_utf8(USPTR, int);
1761  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,  extern BOOL          _pcre_was_newline(USPTR, int, USPTR, int *, BOOL);
1762                        int *, BOOL);  extern BOOL          _pcre_xclass(int, const uschar *);
 extern BOOL         _pcre_xclass(int, const uschar *);  
1763    
1764    
1765  /* Unicode character database (UCD) */  /* Unicode character database (UCD) */

Legend:
Removed from v.454  
changed lines
  Added in v.487

  ViewVC Help
Powered by ViewVC 1.1.5