/[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 459 by ph10, Sun Oct 4 09:21:39 2009 UTC revision 514 by ph10, Mon May 3 12:54:22 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 857  so that PCRE works on both ASCII and EBC Line 875  so that PCRE works on both ASCII and EBC
875  #define STRING_COMMIT0              "COMMIT\0"  #define STRING_COMMIT0              "COMMIT\0"
876  #define STRING_F0                   "F\0"  #define STRING_F0                   "F\0"
877  #define STRING_FAIL0                "FAIL\0"  #define STRING_FAIL0                "FAIL\0"
878    #define STRING_MARK0                "MARK\0"
879  #define STRING_PRUNE0               "PRUNE\0"  #define STRING_PRUNE0               "PRUNE\0"
880  #define STRING_SKIP0                "SKIP\0"  #define STRING_SKIP0                "SKIP\0"
881  #define STRING_THEN                 "THEN"  #define STRING_THEN                 "THEN"
# Line 1109  only. */ Line 1128  only. */
1128  #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"  #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"
1129  #define STRING_F0                   STR_F "\0"  #define STRING_F0                   STR_F "\0"
1130  #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"  #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"
1131    #define STRING_MARK0                STR_M STR_A STR_R STR_K "\0"
1132  #define STRING_PRUNE0               STR_P STR_R STR_U STR_N STR_E "\0"  #define STRING_PRUNE0               STR_P STR_R STR_U STR_N STR_E "\0"
1133  #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"  #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"
1134  #define STRING_THEN                 STR_T STR_H STR_E STR_N  #define STRING_THEN                 STR_T STR_H STR_E STR_N
# Line 1189  contain UTF-8 characters with values gre Line 1209  contain UTF-8 characters with values gre
1209  /* 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
1210  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
1211  their negation. Also, they must appear in the same order as in the opcode  their negation. Also, they must appear in the same order as in the opcode
1212  definitions below, up to ESC_z. There's a dummy for OP_ANY because it  definitions below, up to ESC_z. There's a dummy for OP_ALLANY because it
1213  corresponds to "." rather than an escape sequence, and another for OP_ALLANY  corresponds to "." in DOTALL mode rather than an escape sequence. It is also
1214  (which is used for [^] in JavaScript compatibility mode).  used for [^] in JavaScript compatibility mode. In non-DOTALL mode, "." behaves
1215    like \N.
1216    
1217  The final escape must be ESC_REF as subsequent values are used for  The final escape must be ESC_REF as subsequent values are used for
1218  backreferences (\1, \2, \3, etc). There are two tests in the code for an escape  backreferences (\1, \2, \3, etc). There are two tests in the code for an escape
# Line 1201  put in between that don't consume a char Line 1222  put in between that don't consume a char
1222  */  */
1223    
1224  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,
1225         ESC_W, ESC_w, ESC_dum1, ESC_dum2, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,         ESC_W, ESC_w, ESC_N, ESC_dum, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
1226         ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_g, ESC_k,         ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_g, ESC_k,
1227         ESC_REF };         ESC_REF };
1228    
# Line 1210  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, E Line 1231  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, E
1231  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.
1232    
1233  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
1234  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
1235  "coptable" in pcre_dfa_exec.c that must be updated. */  "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1236    
1237  enum {  enum {
1238    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 1229  enum { Line 1250  enum {
1250    OP_WHITESPACE,         /*  9 \s */    OP_WHITESPACE,         /*  9 \s */
1251    OP_NOT_WORDCHAR,       /* 10 \W */    OP_NOT_WORDCHAR,       /* 10 \W */
1252    OP_WORDCHAR,           /* 11 \w */    OP_WORDCHAR,           /* 11 \w */
1253    OP_ANY,            /* 12 Match any character (subject to DOTALL) */    OP_ANY,            /* 12 Match any character except newline */
1254    OP_ALLANY,         /* 13 Match any character (not subject to DOTALL) */    OP_ALLANY,         /* 13 Match any character */
1255    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
1256    OP_NOTPROP,        /* 15 \P (not Unicode property) */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
1257    OP_PROP,           /* 16 \p (Unicode property) */    OP_PROP,           /* 16 \p (Unicode property) */
# Line 1348  enum { Line 1369  enum {
1369    OP_SCOND,          /* 99 Conditional group, check empty */    OP_SCOND,          /* 99 Conditional group, check empty */
1370    
1371    /* The next two pairs must (respectively) be kept together. */    /* The next two pairs must (respectively) be kept together. */
1372    
1373    OP_CREF,           /* 100 Used to hold a capture number as condition */    OP_CREF,           /* 100 Used to hold a capture number as condition */
1374    OP_NCREF,          /* 101 Same, but generaged by a name reference*/    OP_NCREF,          /* 101 Same, but generaged by a name reference*/
1375    OP_RREF,           /* 102 Used to hold a recursion number as condition */    OP_RREF,           /* 102 Used to hold a recursion number as condition */
# Line 1360  enum { Line 1381  enum {
1381    
1382    /* These are backtracking control verbs */    /* These are backtracking control verbs */
1383    
1384    OP_PRUNE,          /* 107 */    OP_MARK,           /* 107 always has an argument */
1385    OP_SKIP,           /* 108 */    OP_PRUNE,          /* 108 */
1386    OP_THEN,           /* 109 */    OP_PRUNE_ARG,      /* 109 same, but with argument */
1387    OP_COMMIT,         /* 110 */    OP_SKIP,           /* 110 */
1388      OP_SKIP_ARG,       /* 111 same, but with argument */
1389      OP_THEN,           /* 112 */
1390      OP_THEN_ARG,       /* 113 same, but with argument */
1391      OP_COMMIT,         /* 114 */
1392    
1393    /* These are forced failure and success verbs */    /* These are forced failure and success verbs */
1394    
1395    OP_FAIL,           /* 111 */    OP_FAIL,           /* 115 */
1396    OP_ACCEPT,         /* 112 */    OP_ACCEPT,         /* 116 */
1397    OP_CLOSE,          /* 113 Used before OP_ACCEPT to close open captures */    OP_CLOSE,          /* 117 Used before OP_ACCEPT to close open captures */
1398    
1399    /* This is used to skip a subpattern with a {0} quantifier */    /* This is used to skip a subpattern with a {0} quantifier */
1400    
1401    OP_SKIPZERO        /* 114 */    OP_SKIPZERO,       /* 118 */
1402    
1403      /* This is not an opcode, but is used to check that tables indexed by opcode
1404      are the correct length, in order to catch updating errors - there have been
1405      some in the past. */
1406    
1407      OP_TABLE_LENGTH
1408  };  };
1409    
1410    /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
1411    definitions that follow must also be updated to match. There are also tables
1412    called "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1413    
1414    
1415  /* 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
1416  for debugging. The macro is referenced only in pcre_printint.c. */  for debugging. The macro is referenced only in pcre_printint.c. */
# Line 1399  for debugging. The macro is referenced o Line 1434  for debugging. The macro is referenced o
1434    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
1435    "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \    "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \
1436    "Brazero", "Braminzero",                                        \    "Brazero", "Braminzero",                                        \
1437    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \    "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \
1438      "*THEN", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",                \
1439    "Close", "Skip zero"    "Close", "Skip zero"
1440    
1441    
# Line 1417  in UTF-8 mode. The code that uses this t Line 1453  in UTF-8 mode. The code that uses this t
1453    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \
1454    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
1455    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
1456    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3,                          /* \P, \p                                 */ \
1457    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
1458      1,                             /* \X                                     */ \
1459    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
1460    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
1461    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Charnc  - the minimum length           */ \
# Line 1464  in UTF-8 mode. The code that uses this t Line 1501  in UTF-8 mode. The code that uses this t
1501    3, 3,                          /* RREF, NRREF                            */ \    3, 3,                          /* RREF, NRREF                            */ \
1502    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
1503    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
1504    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \    3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG,                */ \
1505    1, 1, 3, 1                     /* FAIL, ACCEPT, CLOSE, SKIPZERO          */    1, 3, 1, 3,                    /* SKIP, SKIP_ARG, THEN, THEN_ARG,        */ \
1506      1, 1, 1, 3, 1                  /* COMMIT, FAIL, ACCEPT, CLOSE, SKIPZERO  */
1507    
1508    
1509  /* A magic value for OP_RREF and OP_NRREF to indicate the "any recursion"  /* A magic value for OP_RREF and OP_NRREF to indicate the "any recursion"
# Line 1473  condition. */ Line 1511  condition. */
1511    
1512  #define RREF_ANY  0xffff  #define RREF_ANY  0xffff
1513    
1514  /* 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
1515  tracked. */  easily be tracked. When a new number is added, the table called eint in
1516    pcreposix.c must be updated. */
1517    
1518  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,
1519         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
# Line 1482  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4, Line 1521  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
1521         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
1522         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
1523         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
1524         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65 };         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERRCOUNT };
1525    
1526  /* 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
1527  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 1532  typedef struct pcre_study_data { Line 1571  typedef struct pcre_study_data {
1571    
1572  /* Structure for building a chain of open capturing subpatterns during  /* Structure for building a chain of open capturing subpatterns during
1573  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
1574  encountered. */  encountered. This is also used to identify subpatterns that contain recursive
1575    back references to themselves, so that they can be made atomic. */
1576    
1577  typedef struct open_capitem {  typedef struct open_capitem {
1578    struct open_capitem *next;    /* Chain link */    struct open_capitem *next;    /* Chain link */
1579    pcre_uint16 number;           /* Capture number */    pcre_uint16 number;           /* Capture number */
1580      pcre_uint16 flag;             /* Set TRUE if recursive back ref */
1581  } open_capitem;  } open_capitem;
1582    
1583  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
# Line 1575  branches, for testing for left recursion Line 1616  branches, for testing for left recursion
1616    
1617  typedef struct branch_chain {  typedef struct branch_chain {
1618    struct branch_chain *outer;    struct branch_chain *outer;
1619    uschar *current;    uschar *current_branch;
1620  } branch_chain;  } branch_chain;
1621    
1622  /* 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 1585  typedef struct recursion_info { Line 1626  typedef struct recursion_info {
1626    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
1627    int group_num;                /* Number of group that was called */    int group_num;                /* Number of group that was called */
1628    const uschar *after_call;     /* "Return value": points after the call in the expr */    const uschar *after_call;     /* "Return value": points after the call in the expr */
   USPTR save_start;             /* Old value of mstart */  
1629    int *offset_save;             /* Pointer to start of saved offsets */    int *offset_save;             /* Pointer to start of saved offsets */
1630    int saved_max;                /* Number of saved offsets */    int saved_max;                /* Number of saved offsets */
1631    int offset_top;               /* Current value of offset_top */    int save_offset_top;          /* Current value of offset_top */
1632  } recursion_info;  } recursion_info;
1633    
1634  /* 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 1615  typedef struct match_data { Line 1655  typedef struct match_data {
1655    int    nllen;                 /* Newline string length */    int    nllen;                 /* Newline string length */
1656    int    name_count;            /* Number of names in name table */    int    name_count;            /* Number of names in name table */
1657    int    name_entry_size;       /* Size of entry in names table */    int    name_entry_size;       /* Size of entry in names table */
1658    uschar *name_table;           /* Table of names */    uschar *name_table;           /* Table of names */
1659    uschar nl[4];                 /* Newline string when fixed */    uschar nl[4];                 /* Newline string when fixed */
1660    const uschar *lcc;            /* Points to lower casing table */    const uschar *lcc;            /* Points to lower casing table */
1661    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
# Line 1643  typedef struct match_data { Line 1683  typedef struct match_data {
1683    int    eptrn;                 /* Next free eptrblock */    int    eptrn;                 /* Next free eptrblock */
1684    recursion_info *recursive;    /* Linked list of recursion data */    recursion_info *recursive;    /* Linked list of recursion data */
1685    void  *callout_data;          /* To pass back to callouts */    void  *callout_data;          /* To pass back to callouts */
1686      const uschar *mark;           /* Mark pointer to pass back */
1687  } match_data;  } match_data;
1688    
1689  /* A similar structure is used for the same purpose by the DFA matching  /* A similar structure is used for the same purpose by the DFA matching
# Line 1734  one of the exported public functions. Th Line 1775  one of the exported public functions. Th
1775  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1776    
1777  extern const uschar *_pcre_find_bracket(const uschar *, BOOL, int);  extern const uschar *_pcre_find_bracket(const uschar *, BOOL, int);
1778  extern BOOL          _pcre_is_newline(const uschar *, int, const uschar *,  extern BOOL          _pcre_is_newline(USPTR, int, USPTR, int *, BOOL);
                        int *, BOOL);  
1779  extern int           _pcre_ord2utf8(int, uschar *);  extern int           _pcre_ord2utf8(int, uschar *);
1780  extern real_pcre    *_pcre_try_flipped(const real_pcre *, real_pcre *,  extern real_pcre    *_pcre_try_flipped(const real_pcre *, real_pcre *,
1781                         const pcre_study_data *, pcre_study_data *);                         const pcre_study_data *, pcre_study_data *);
1782  extern int           _pcre_valid_utf8(const uschar *, int);  extern int           _pcre_valid_utf8(USPTR, int);
1783  extern BOOL          _pcre_was_newline(const uschar *, int, const uschar *,  extern BOOL          _pcre_was_newline(USPTR, int, USPTR, int *, BOOL);
                        int *, BOOL);  
1784  extern BOOL          _pcre_xclass(int, const uschar *);  extern BOOL          _pcre_xclass(int, const uschar *);
1785    
1786    

Legend:
Removed from v.459  
changed lines
  Added in v.514

  ViewVC Help
Powered by ViewVC 1.1.5