/[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 499 by ph10, Wed Mar 3 20:09:39 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. Systems that do not
192    have stdint.h (e.g. Solaris) may have inttypes.h. The macro int64_t may be set
193    by "configure". */
194    
195    #if HAVE_STDINT_H
196    #include <stdint.h>
197    #elif HAVE_INTTYPES_H
198    #include <inttypes.h>
199    #endif
200    
201    #if defined INT64_MAX || defined int64_t
202    #define INT64_OR_DOUBLE int64_t
203    #else
204    #define INT64_OR_DOUBLE double
205    #endif
206    
207  /* All character handling must be done as unsigned characters. Otherwise there  /* All character handling must be done as unsigned characters. Otherwise there
208  are problems with top-bit-set characters and functions such as isspace().  are problems with top-bit-set characters and functions such as isspace().
209  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 567  compatibility. */
567  /* Options for the "extra" block produced by pcre_study(). */  /* Options for the "extra" block produced by pcre_study(). */
568    
569  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */
570    #define PCRE_STUDY_MINLEN   0x02     /* a minimum length field exists */
571    
572  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
573  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 1228  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, E
1228  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.
1229    
1230  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
1231  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
1232  "coptable" in pcre_dfa_exec.c that must be updated. */  "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1233    
1234  enum {  enum {
1235    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 1346  enum { Line 1365  enum {
1365    OP_SCBRA,          /* 98 Start of capturing bracket, check empty */    OP_SCBRA,          /* 98 Start of capturing bracket, check empty */
1366    OP_SCOND,          /* 99 Conditional group, check empty */    OP_SCOND,          /* 99 Conditional group, check empty */
1367    
1368      /* The next two pairs must (respectively) be kept together. */
1369    
1370    OP_CREF,           /* 100 Used to hold a capture number as condition */    OP_CREF,           /* 100 Used to hold a capture number as condition */
1371    OP_RREF,           /* 101 Used to hold a recursion number as condition */    OP_NCREF,          /* 101 Same, but generaged by a name reference*/
1372    OP_DEF,            /* 102 The DEFINE condition */    OP_RREF,           /* 102 Used to hold a recursion number as condition */
1373      OP_NRREF,          /* 103 Same, but generaged by a name reference*/
1374      OP_DEF,            /* 104 The DEFINE condition */
1375    
1376    OP_BRAZERO,        /* 103 These two must remain together and in this */    OP_BRAZERO,        /* 105 These two must remain together and in this */
1377    OP_BRAMINZERO,     /* 104 order. */    OP_BRAMINZERO,     /* 106 order. */
1378    
1379    /* These are backtracking control verbs */    /* These are backtracking control verbs */
1380    
1381    OP_PRUNE,          /* 105 */    OP_PRUNE,          /* 107 */
1382    OP_SKIP,           /* 106 */    OP_SKIP,           /* 108 */
1383    OP_THEN,           /* 107 */    OP_THEN,           /* 109 */
1384    OP_COMMIT,         /* 108 */    OP_COMMIT,         /* 110 */
1385    
1386    /* These are forced failure and success verbs */    /* These are forced failure and success verbs */
1387    
1388    OP_FAIL,           /* 109 */    OP_FAIL,           /* 111 */
1389    OP_ACCEPT,         /* 110 */    OP_ACCEPT,         /* 112 */
1390    OP_CLOSE,          /* 111 Used before OP_ACCEPT to close open captures */    OP_CLOSE,          /* 113 Used before OP_ACCEPT to close open captures */
1391    
1392    /* This is used to skip a subpattern with a {0} quantifier */    /* This is used to skip a subpattern with a {0} quantifier */
1393    
1394    OP_SKIPZERO        /* 112 */    OP_SKIPZERO,       /* 114 */
1395    
1396      /* This is not an opcode, but is used to check that tables indexed by opcode
1397      are the correct length, in order to catch updating errors - there have been
1398      some in the past. */
1399    
1400      OP_TABLE_LENGTH
1401  };  };
1402    
1403    /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
1404    definitions that follow must also be updated to match. There are also tables
1405    called "coptable" cna "poptable" in pcre_dfa_exec.c that must be updated. */
1406    
1407    
1408  /* 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
1409  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 1425  for debugging. The macro is referenced o
1425    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
1426    "AssertB", "AssertB not", "Reverse",                            \    "AssertB", "AssertB not", "Reverse",                            \
1427    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
1428    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \    "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \
1429      "Brazero", "Braminzero",                                        \
1430    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \
1431    "Close", "Skip zero"    "Close", "Skip zero"
1432    
# Line 1411  in UTF-8 mode. The code that uses this t Line 1445  in UTF-8 mode. The code that uses this t
1445    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \
1446    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
1447    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
1448    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3,                          /* \P, \p                                 */ \
1449    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
1450      1,                             /* \X                                     */ \
1451    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
1452    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
1453    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Charnc  - the minimum length           */ \
# Line 1454  in UTF-8 mode. The code that uses this t Line 1489  in UTF-8 mode. The code that uses this t
1489    1+LINK_SIZE,                   /* SBRA                                   */ \    1+LINK_SIZE,                   /* SBRA                                   */ \
1490    3+LINK_SIZE,                   /* SCBRA                                  */ \    3+LINK_SIZE,                   /* SCBRA                                  */ \
1491    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
1492    3,                             /* CREF                                   */ \    3, 3,                          /* CREF, NCREF                            */ \
1493    3,                             /* RREF                                   */ \    3, 3,                          /* RREF, NRREF                            */ \
1494    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
1495    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
1496    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \
1497    1, 1, 3, 1                     /* FAIL, ACCEPT, CLOSE, SKIPZERO          */    1, 1, 3, 1                     /* FAIL, ACCEPT, CLOSE, SKIPZERO          */
1498    
1499    
1500  /* 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"
1501    condition. */
1502    
1503  #define RREF_ANY  0xffff  #define RREF_ANY  0xffff
1504    
1505  /* Error code numbers. They are given names so that they can more easily be  /* Compile time error code numbers. They are given names so that they can more
1506  tracked. */  easily be tracked. When a new number is added, the table called eint in
1507    pcreposix.c must be updated. */
1508    
1509  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,
1510         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
# Line 1475  enum { ERR0, ERR1, ERR2, ERR3, ERR4, Line 1512  enum { ERR0, ERR1, ERR2, ERR3, ERR4,
1512         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
1513         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
1514         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
1515         ERR60, ERR61, ERR62, ERR63, ERR64 };         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERRCOUNT };
1516    
1517  /* 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
1518  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 1528  Because people can now save and re-use c
1528  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
1529  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
1530  fields are present. Currently PCRE always sets the dummy fields to zero.  fields are present. Currently PCRE always sets the dummy fields to zero.
1531  NOTE NOTE NOTE:  NOTE NOTE NOTE
1532  */  */
1533    
1534  typedef struct real_pcre {  typedef struct real_pcre {
# Line 1518  remark (see NOTE above) about extending Line 1555  remark (see NOTE above) about extending
1555    
1556  typedef struct pcre_study_data {  typedef struct pcre_study_data {
1557    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1558    pcre_uint32 options;    pcre_uint32 flags;              /* Private flags */
1559    uschar start_bits[32];    uschar start_bits[32];          /* Starting char bits */
1560      pcre_uint32 minlength;          /* Minimum subject length */
1561  } pcre_study_data;  } pcre_study_data;
1562    
1563  /* Structure for building a chain of open capturing subpatterns during  /* Structure for building a chain of open capturing subpatterns during
1564  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
1565  encountered. */  encountered. This is also used to identify subpatterns that contain recursive
1566    back references to themselves, so that they can be made atomic. */
1567    
1568  typedef struct open_capitem {  typedef struct open_capitem {
1569    struct open_capitem *next;    /* Chain link */    struct open_capitem *next;    /* Chain link */
1570    pcre_uint16 number;           /* Capture number */    pcre_uint16 number;           /* Capture number */
1571  } open_capitem;    pcre_uint16 flag;             /* Set TRUE if recursive back ref */
1572    } open_capitem;
1573    
1574  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
1575  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 1583  typedef struct compile_data {
1583    const uschar *start_code;     /* The start of the compiled code */    const uschar *start_code;     /* The start of the compiled code */
1584    const uschar *start_pattern;  /* The start of the pattern */    const uschar *start_pattern;  /* The start of the pattern */
1585    const uschar *end_pattern;    /* The end of the pattern */    const uschar *end_pattern;    /* The end of the pattern */
1586    open_capitem *open_caps;      /* Chain of open capture items */    open_capitem *open_caps;      /* Chain of open capture items */
1587    uschar *hwm;                  /* High watermark of workspace */    uschar *hwm;                  /* High watermark of workspace */
1588    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
1589    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
# Line 1556  typedef struct compile_data { Line 1596  typedef struct compile_data {
1596    int  external_flags;          /* External flag bits to be set */    int  external_flags;          /* External flag bits to be set */
1597    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
1598    BOOL had_accept;              /* (*ACCEPT) encountered */    BOOL had_accept;              /* (*ACCEPT) encountered */
1599    BOOL check_lookbehind;        /* Lookbehinds need later checking */    BOOL check_lookbehind;        /* Lookbehinds need later checking */
1600    int  nltype;                  /* Newline type */    int  nltype;                  /* Newline type */
1601    int  nllen;                   /* Newline string length */    int  nllen;                   /* Newline string length */
1602    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 1607  branches, for testing for left recursion
1607    
1608  typedef struct branch_chain {  typedef struct branch_chain {
1609    struct branch_chain *outer;    struct branch_chain *outer;
1610    uschar *current;    uschar *current_branch;
1611  } branch_chain;  } branch_chain;
1612    
1613  /* 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 1620  typedef struct recursion_info {
1620    USPTR save_start;             /* Old value of mstart */    USPTR save_start;             /* Old value of mstart */
1621    int *offset_save;             /* Pointer to start of saved offsets */    int *offset_save;             /* Pointer to start of saved offsets */
1622    int saved_max;                /* Number of saved offsets */    int saved_max;                /* Number of saved offsets */
1623    int offset_top;               /* Current value of offset_top */    int save_offset_top;          /* Current value of offset_top */
1624  } recursion_info;  } recursion_info;
1625    
1626  /* 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 1645  typedef struct match_data {
1645    int    offset_max;            /* The maximum usable for return data */    int    offset_max;            /* The maximum usable for return data */
1646    int    nltype;                /* Newline type */    int    nltype;                /* Newline type */
1647    int    nllen;                 /* Newline string length */    int    nllen;                 /* Newline string length */
1648      int    name_count;            /* Number of names in name table */
1649      int    name_entry_size;       /* Size of entry in names table */
1650      uschar *name_table;           /* Table of names */
1651    uschar nl[4];                 /* Newline string when fixed */    uschar nl[4];                 /* Newline string when fixed */
1652    const uschar *lcc;            /* Points to lower casing table */    const uschar *lcc;            /* Points to lower casing table */
1653    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 1765  extern const uschar _pcre_OP_lengths[];
1765  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
1766  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1767    
1768  extern BOOL         _pcre_is_newline(const uschar *, int, const uschar *,  extern const uschar *_pcre_find_bracket(const uschar *, BOOL, int);
1769                        int *, BOOL);  extern BOOL          _pcre_is_newline(USPTR, int, USPTR, int *, BOOL);
1770  extern int          _pcre_ord2utf8(int, uschar *);  extern int           _pcre_ord2utf8(int, uschar *);
1771  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  extern real_pcre    *_pcre_try_flipped(const real_pcre *, real_pcre *,
1772                        const pcre_study_data *, pcre_study_data *);                         const pcre_study_data *, pcre_study_data *);
1773  extern int          _pcre_valid_utf8(const uschar *, int);  extern int           _pcre_valid_utf8(USPTR, int);
1774  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,  extern BOOL          _pcre_was_newline(USPTR, int, USPTR, int *, BOOL);
1775                        int *, BOOL);  extern BOOL          _pcre_xclass(int, const uschar *);
 extern BOOL         _pcre_xclass(int, const uschar *);  
1776    
1777    
1778  /* Unicode character database (UCD) */  /* Unicode character database (UCD) */
Legend:
Removed from v.454  
changed lines
  Added in v.499
webmaster@exim.org
 ViewVC Help
Powered by ViewVC 1.1.12