/[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 535 by ph10, Thu Jun 3 19:18:24 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 457  know we are in UTF-8 mode. */ Line 475  know we are in UTF-8 mode. */
475        } \        } \
476      }      }
477    
478  /* Get the next character, testing for UTF-8 mode, and advancing the pointer */  /* Get the next character, testing for UTF-8 mode, and advancing the pointer.
479    This is called when we don't know if we are in UTF-8 mode. */
480    
481  #define GETCHARINCTEST(c, eptr) \  #define GETCHARINCTEST(c, eptr) \
482    c = *eptr++; \    c = *eptr++; \
# Line 494  if there are extra bytes. This is called Line 513  if there are extra bytes. This is called
513    
514  /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the  /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the
515  pointer, incrementing length if there are extra bytes. This is called when we  pointer, incrementing length if there are extra bytes. This is called when we
516  know we are in UTF-8 mode. */  do not know if we are in UTF-8 mode. */
517    
518  #define GETCHARLENTEST(c, eptr, len) \  #define GETCHARLENTEST(c, eptr, len) \
519    c = *eptr; \    c = *eptr; \
# Line 535  Standard C system should have one. */ Line 554  Standard C system should have one. */
554    
555  /* Private flags containing information about the compiled regex. They used to  /* Private flags containing information about the compiled regex. They used to
556  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
557  are in a 16-bit flags word. */  are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as
558    the restrictions on partial matching have been lifted. It remains for backwards
559    compatibility. */
560    
561  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */
562  #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 568  are in a 16-bit flags word. */
568  /* Options for the "extra" block produced by pcre_study(). */  /* Options for the "extra" block produced by pcre_study(). */
569    
570  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */
571    #define PCRE_STUDY_MINLEN   0x02     /* a minimum length field exists */
572    
573  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
574  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
# Line 559  time, run time, or study time, respectiv Line 581  time, run time, or study time, respectiv
581     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
582     PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \     PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \
583     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
584     PCRE_JAVASCRIPT_COMPAT)     PCRE_JAVASCRIPT_COMPAT|PCRE_UCP)
585    
586  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
587    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
588     PCRE_PARTIAL|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_NEWLINE_BITS| \
589     PCRE_NO_START_OPTIMIZE)     PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
590    
591  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
592    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
593     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS| \     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_DFA_SHORTEST| \
594     PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)     PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
595       PCRE_NO_START_OPTIMIZE)
596    
597  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
598    
# Line 853  so that PCRE works on both ASCII and EBC Line 876  so that PCRE works on both ASCII and EBC
876  #define STRING_COMMIT0              "COMMIT\0"  #define STRING_COMMIT0              "COMMIT\0"
877  #define STRING_F0                   "F\0"  #define STRING_F0                   "F\0"
878  #define STRING_FAIL0                "FAIL\0"  #define STRING_FAIL0                "FAIL\0"
879    #define STRING_MARK0                "MARK\0"
880  #define STRING_PRUNE0               "PRUNE\0"  #define STRING_PRUNE0               "PRUNE\0"
881  #define STRING_SKIP0                "SKIP\0"  #define STRING_SKIP0                "SKIP\0"
882  #define STRING_THEN                 "THEN"  #define STRING_THEN                 "THEN"
# Line 882  so that PCRE works on both ASCII and EBC Line 906  so that PCRE works on both ASCII and EBC
906  #define STRING_BSR_ANYCRLF_RIGHTPAR "BSR_ANYCRLF)"  #define STRING_BSR_ANYCRLF_RIGHTPAR "BSR_ANYCRLF)"
907  #define STRING_BSR_UNICODE_RIGHTPAR "BSR_UNICODE)"  #define STRING_BSR_UNICODE_RIGHTPAR "BSR_UNICODE)"
908  #define STRING_UTF8_RIGHTPAR        "UTF8)"  #define STRING_UTF8_RIGHTPAR        "UTF8)"
909    #define STRING_UCP_RIGHTPAR         "UCP)"
910    
911  #else  /* SUPPORT_UTF8 */  #else  /* SUPPORT_UTF8 */
912    
# Line 1105  only. */ Line 1130  only. */
1130  #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"
1131  #define STRING_F0                   STR_F "\0"  #define STRING_F0                   STR_F "\0"
1132  #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"  #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"
1133    #define STRING_MARK0                STR_M STR_A STR_R STR_K "\0"
1134  #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"
1135  #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"  #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"
1136  #define STRING_THEN                 STR_T STR_H STR_E STR_N  #define STRING_THEN                 STR_T STR_H STR_E STR_N
# Line 1134  only. */ Line 1160  only. */
1160  #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
1161  #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
1162  #define STRING_UTF8_RIGHTPAR        STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS  #define STRING_UTF8_RIGHTPAR        STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1163    #define STRING_UCP_RIGHTPAR         STR_U STR_C STR_P STR_RIGHT_PARENTHESIS
1164    
1165  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF8 */
1166    
# Line 1166  only. */ Line 1193  only. */
1193    
1194  #define PT_ANY        0    /* Any property - matches all chars */  #define PT_ANY        0    /* Any property - matches all chars */
1195  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */
1196  #define PT_GC         2    /* General characteristic (e.g. L) */  #define PT_GC         2    /* Specified general characteristic (e.g. L) */
1197  #define PT_PC         3    /* Particular characteristic (e.g. Lu) */  #define PT_PC         3    /* Specified particular characteristic (e.g. Lu) */
1198  #define PT_SC         4    /* Script (e.g. Han) */  #define PT_SC         4    /* Script (e.g. Han) */
1199    #define PT_ALNUM      5    /* Alphanumeric - the union of L and N */
1200    #define PT_SPACE      6    /* Perl space - Z plus 9,10,12,13 */
1201    #define PT_PXSPACE    7    /* POSIX space - Z plus 9,10,11,12,13 */
1202    #define PT_WORD       8    /* Word - L plus N plus underscore */
1203    
1204  /* Flag bits and data types for the extended class (OP_XCLASS) for classes that  /* Flag bits and data types for the extended class (OP_XCLASS) for classes that
1205  contain UTF-8 characters with values greater than 255. */  contain UTF-8 characters with values greater than 255. */
# Line 1185  contain UTF-8 characters with values gre Line 1216  contain UTF-8 characters with values gre
1216  /* 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
1217  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
1218  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
1219  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
1220  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
1221  (which is used for [^] in JavaScript compatibility mode).  used for [^] in JavaScript compatibility mode. In non-DOTALL mode, "." behaves
1222    like \N.
1223    
1224    The special values ESC_DU, ESC_du, etc. are used instead of ESC_D, ESC_d, etc.
1225    when PCRE_UCP is set, when replacement of \d etc by \p sequences is required.
1226    They must be contiguous, and remain in order so that the replacements can be
1227    looked up from a table.
1228    
1229  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
1230  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 1197  put in between that don't consume a char Line 1234  put in between that don't consume a char
1234  */  */
1235    
1236  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,
1237         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,
1238         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,
1239           ESC_E, ESC_Q, ESC_g, ESC_k,
1240           ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu,
1241         ESC_REF };         ESC_REF };
1242    
   
1243  /* Opcode table: Starting from 1 (i.e. after OP_END), the values up to  /* Opcode table: Starting from 1 (i.e. after OP_END), the values up to
1244  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.
1245    
1246  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
1247  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
1248  "coptable" in pcre_dfa_exec.c that must be updated. */  "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1249    
1250  enum {  enum {
1251    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 1225  enum { Line 1263  enum {
1263    OP_WHITESPACE,         /*  9 \s */    OP_WHITESPACE,         /*  9 \s */
1264    OP_NOT_WORDCHAR,       /* 10 \W */    OP_NOT_WORDCHAR,       /* 10 \W */
1265    OP_WORDCHAR,           /* 11 \w */    OP_WORDCHAR,           /* 11 \w */
1266    OP_ANY,            /* 12 Match any character (subject to DOTALL) */    OP_ANY,            /* 12 Match any character except newline */
1267    OP_ALLANY,         /* 13 Match any character (not subject to DOTALL) */    OP_ALLANY,         /* 13 Match any character */
1268    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 */
1269    OP_NOTPROP,        /* 15 \P (not Unicode property) */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
1270    OP_PROP,           /* 16 \p (Unicode property) */    OP_PROP,           /* 16 \p (Unicode property) */
# Line 1343  enum { Line 1381  enum {
1381    OP_SCBRA,          /* 98 Start of capturing bracket, check empty */    OP_SCBRA,          /* 98 Start of capturing bracket, check empty */
1382    OP_SCOND,          /* 99 Conditional group, check empty */    OP_SCOND,          /* 99 Conditional group, check empty */
1383    
1384      /* The next two pairs must (respectively) be kept together. */
1385    
1386    OP_CREF,           /* 100 Used to hold a capture number as condition */    OP_CREF,           /* 100 Used to hold a capture number as condition */
1387    OP_RREF,           /* 101 Used to hold a recursion number as condition */    OP_NCREF,          /* 101 Same, but generaged by a name reference*/
1388    OP_DEF,            /* 102 The DEFINE condition */    OP_RREF,           /* 102 Used to hold a recursion number as condition */
1389      OP_NRREF,          /* 103 Same, but generaged by a name reference*/
1390      OP_DEF,            /* 104 The DEFINE condition */
1391    
1392    OP_BRAZERO,        /* 103 These two must remain together and in this */    OP_BRAZERO,        /* 105 These two must remain together and in this */
1393    OP_BRAMINZERO,     /* 104 order. */    OP_BRAMINZERO,     /* 106 order. */
1394    
1395    /* These are backtracking control verbs */    /* These are backtracking control verbs */
1396    
1397    OP_PRUNE,          /* 105 */    OP_MARK,           /* 107 always has an argument */
1398    OP_SKIP,           /* 106 */    OP_PRUNE,          /* 108 */
1399    OP_THEN,           /* 107 */    OP_PRUNE_ARG,      /* 109 same, but with argument */
1400    OP_COMMIT,         /* 108 */    OP_SKIP,           /* 110 */
1401      OP_SKIP_ARG,       /* 111 same, but with argument */
1402      OP_THEN,           /* 112 */
1403      OP_THEN_ARG,       /* 113 same, but with argument */
1404      OP_COMMIT,         /* 114 */
1405    
1406    /* These are forced failure and success verbs */    /* These are forced failure and success verbs */
1407    
1408    OP_FAIL,           /* 109 */    OP_FAIL,           /* 115 */
1409    OP_ACCEPT,         /* 110 */    OP_ACCEPT,         /* 116 */
1410      OP_CLOSE,          /* 117 Used before OP_ACCEPT to close open captures */
1411    
1412    /* This is used to skip a subpattern with a {0} quantifier */    /* This is used to skip a subpattern with a {0} quantifier */
1413    
1414    OP_SKIPZERO        /* 111 */    OP_SKIPZERO,       /* 118 */
1415    
1416      /* This is not an opcode, but is used to check that tables indexed by opcode
1417      are the correct length, in order to catch updating errors - there have been
1418      some in the past. */
1419    
1420      OP_TABLE_LENGTH
1421  };  };
1422    
1423    /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
1424    definitions that follow must also be updated to match. There are also tables
1425    called "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1426    
1427    
1428  /* 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
1429  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 1445  for debugging. The macro is referenced o
1445    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
1446    "AssertB", "AssertB not", "Reverse",                            \    "AssertB", "AssertB not", "Reverse",                            \
1447    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
1448    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \    "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \
1449    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \    "Brazero", "Braminzero",                                        \
1450    "Skip zero"    "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \
1451      "*THEN", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",                \
1452      "Close", "Skip zero"
1453    
1454    
1455  /* 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 1407  in UTF-8 mode. The code that uses this t Line 1466  in UTF-8 mode. The code that uses this t
1466    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \
1467    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
1468    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
1469    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3,                          /* \P, \p                                 */ \
1470    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
1471      1,                             /* \X                                     */ \
1472    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
1473    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
1474    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Charnc  - the minimum length           */ \
# Line 1450  in UTF-8 mode. The code that uses this t Line 1510  in UTF-8 mode. The code that uses this t
1510    1+LINK_SIZE,                   /* SBRA                                   */ \    1+LINK_SIZE,                   /* SBRA                                   */ \
1511    3+LINK_SIZE,                   /* SCBRA                                  */ \    3+LINK_SIZE,                   /* SCBRA                                  */ \
1512    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
1513    3,                             /* CREF                                   */ \    3, 3,                          /* CREF, NCREF                            */ \
1514    3,                             /* RREF                                   */ \    3, 3,                          /* RREF, NRREF                            */ \
1515    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
1516    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
1517    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \    3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG,                */ \
1518    1, 1, 1                        /* FAIL, ACCEPT, SKIPZERO                 */    1, 3, 1, 3,                    /* SKIP, SKIP_ARG, THEN, THEN_ARG,        */ \
1519      1, 1, 1, 3, 1                  /* COMMIT, FAIL, ACCEPT, CLOSE, SKIPZERO  */
1520    
1521    
1522  /* 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"
1523    condition. */
1524    
1525  #define RREF_ANY  0xffff  #define RREF_ANY  0xffff
1526    
1527  /* 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
1528  tracked. */  easily be tracked. When a new number is added, the table called eint in
1529    pcreposix.c must be updated. */
1530    
1531  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,
1532         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
# Line 1471  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4, Line 1534  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
1534         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
1535         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
1536         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
1537         ERR60, ERR61, ERR62, ERR63, ERR64 };         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERRCOUNT };
1538    
1539  /* 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
1540  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 1550  Because people can now save and re-use c
1550  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
1551  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
1552  fields are present. Currently PCRE always sets the dummy fields to zero.  fields are present. Currently PCRE always sets the dummy fields to zero.
1553  NOTE NOTE NOTE:  NOTE NOTE NOTE
1554  */  */
1555    
1556  typedef struct real_pcre {  typedef struct real_pcre {
# Line 1514  remark (see NOTE above) about extending Line 1577  remark (see NOTE above) about extending
1577    
1578  typedef struct pcre_study_data {  typedef struct pcre_study_data {
1579    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1580    pcre_uint32 options;    pcre_uint32 flags;              /* Private flags */
1581    uschar start_bits[32];    uschar start_bits[32];          /* Starting char bits */
1582      pcre_uint32 minlength;          /* Minimum subject length */
1583  } pcre_study_data;  } pcre_study_data;
1584    
1585    /* Structure for building a chain of open capturing subpatterns during
1586    compiling, so that instructions to close them can be compiled when (*ACCEPT) is
1587    encountered. This is also used to identify subpatterns that contain recursive
1588    back references to themselves, so that they can be made atomic. */
1589    
1590    typedef struct open_capitem {
1591      struct open_capitem *next;    /* Chain link */
1592      pcre_uint16 number;           /* Capture number */
1593      pcre_uint16 flag;             /* Set TRUE if recursive back ref */
1594    } open_capitem;
1595    
1596  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
1597  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
1598    
# Line 1530  typedef struct compile_data { Line 1605  typedef struct compile_data {
1605    const uschar *start_code;     /* The start of the compiled code */    const uschar *start_code;     /* The start of the compiled code */
1606    const uschar *start_pattern;  /* The start of the pattern */    const uschar *start_pattern;  /* The start of the pattern */
1607    const uschar *end_pattern;    /* The end of the pattern */    const uschar *end_pattern;    /* The end of the pattern */
1608      open_capitem *open_caps;      /* Chain of open capture items */
1609    uschar *hwm;                  /* High watermark of workspace */    uschar *hwm;                  /* High watermark of workspace */
1610    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
1611    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
# Line 1542  typedef struct compile_data { Line 1618  typedef struct compile_data {
1618    int  external_flags;          /* External flag bits to be set */    int  external_flags;          /* External flag bits to be set */
1619    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
1620    BOOL had_accept;              /* (*ACCEPT) encountered */    BOOL had_accept;              /* (*ACCEPT) encountered */
1621      BOOL check_lookbehind;        /* Lookbehinds need later checking */
1622    int  nltype;                  /* Newline type */    int  nltype;                  /* Newline type */
1623    int  nllen;                   /* Newline string length */    int  nllen;                   /* Newline string length */
1624    uschar nl[4];                 /* Newline string when fixed length */    uschar nl[4];                 /* Newline string when fixed length */
# Line 1552  branches, for testing for left recursion Line 1629  branches, for testing for left recursion
1629    
1630  typedef struct branch_chain {  typedef struct branch_chain {
1631    struct branch_chain *outer;    struct branch_chain *outer;
1632    uschar *current;    uschar *current_branch;
1633  } branch_chain;  } branch_chain;
1634    
1635  /* 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 1562  typedef struct recursion_info { Line 1639  typedef struct recursion_info {
1639    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
1640    int group_num;                /* Number of group that was called */    int group_num;                /* Number of group that was called */
1641    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 */  
1642    int *offset_save;             /* Pointer to start of saved offsets */    int *offset_save;             /* Pointer to start of saved offsets */
1643    int saved_max;                /* Number of saved offsets */    int saved_max;                /* Number of saved offsets */
1644      int save_offset_top;          /* Current value of offset_top */
1645  } recursion_info;  } recursion_info;
1646    
1647  /* 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 1666  typedef struct match_data {
1666    int    offset_max;            /* The maximum usable for return data */    int    offset_max;            /* The maximum usable for return data */
1667    int    nltype;                /* Newline type */    int    nltype;                /* Newline type */
1668    int    nllen;                 /* Newline string length */    int    nllen;                 /* Newline string length */
1669      int    name_count;            /* Number of names in name table */
1670      int    name_entry_size;       /* Size of entry in names table */
1671      uschar *name_table;           /* Table of names */
1672    uschar nl[4];                 /* Newline string when fixed */    uschar nl[4];                 /* Newline string when fixed */
1673    const uschar *lcc;            /* Points to lower casing table */    const uschar *lcc;            /* Points to lower casing table */
1674    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
# Line 1597  typedef struct match_data { Line 1677  typedef struct match_data {
1677    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                /* NOTEOL flag */
1678    BOOL   utf8;                  /* UTF8 flag */    BOOL   utf8;                  /* UTF8 flag */
1679    BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */    BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */
1680      BOOL   use_ucp;               /* PCRE_UCP flag */
1681    BOOL   endonly;               /* Dollar not before final \n */    BOOL   endonly;               /* Dollar not before final \n */
1682    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notempty;              /* Empty string match not wanted */
1683    BOOL   partial;               /* PARTIAL flag */    BOOL   notempty_atstart;      /* Empty string match at start not wanted */
1684    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
1685    BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */    BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */
1686    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
# Line 1607  typedef struct match_data { Line 1688  typedef struct match_data {
1688    USPTR  end_subject;           /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
1689    USPTR  start_match_ptr;       /* Start of matched string */    USPTR  start_match_ptr;       /* Start of matched string */
1690    USPTR  end_match_ptr;         /* Subject position at end match */    USPTR  end_match_ptr;         /* Subject position at end match */
1691      USPTR  start_used_ptr;        /* Earliest consulted character */
1692      int    partial;               /* PARTIAL options */
1693    int    end_offset_top;        /* Highwater mark at end of match */    int    end_offset_top;        /* Highwater mark at end of match */
1694    int    capture_last;          /* Most recent capture number */    int    capture_last;          /* Most recent capture number */
1695    int    start_offset;          /* The start offset value */    int    start_offset;          /* The start offset value */
# Line 1614  typedef struct match_data { Line 1697  typedef struct match_data {
1697    int    eptrn;                 /* Next free eptrblock */    int    eptrn;                 /* Next free eptrblock */
1698    recursion_info *recursive;    /* Linked list of recursion data */    recursion_info *recursive;    /* Linked list of recursion data */
1699    void  *callout_data;          /* To pass back to callouts */    void  *callout_data;          /* To pass back to callouts */
1700      const uschar *mark;           /* Mark pointer to pass back */
1701  } match_data;  } match_data;
1702    
1703  /* 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 1623  typedef struct dfa_match_data { Line 1707  typedef struct dfa_match_data {
1707    const uschar *start_code;     /* Start of the compiled pattern */    const uschar *start_code;     /* Start of the compiled pattern */
1708    const uschar *start_subject;  /* Start of the subject string */    const uschar *start_subject;  /* Start of the subject string */
1709    const uschar *end_subject;    /* End of subject string */    const uschar *end_subject;    /* End of subject string */
1710      const uschar *start_used_ptr; /* Earliest consulted character */
1711    const uschar *tables;         /* Character tables */    const uschar *tables;         /* Character tables */
1712      int   start_offset;           /* The start offset value */
1713    int   moptions;               /* Match options */    int   moptions;               /* Match options */
1714    int   poptions;               /* Pattern options */    int   poptions;               /* Pattern options */
1715    int    nltype;                /* Newline type */    int    nltype;                /* Newline type */
# Line 1702  extern const uschar _pcre_OP_lengths[]; Line 1788  extern const uschar _pcre_OP_lengths[];
1788  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
1789  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1790    
1791  extern BOOL         _pcre_is_newline(const uschar *, int, const uschar *,  extern const uschar *_pcre_find_bracket(const uschar *, BOOL, int);
1792                        int *, BOOL);  extern BOOL          _pcre_is_newline(USPTR, int, USPTR, int *, BOOL);
1793  extern int          _pcre_ord2utf8(int, uschar *);  extern int           _pcre_ord2utf8(int, uschar *);
1794  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  extern real_pcre    *_pcre_try_flipped(const real_pcre *, real_pcre *,
1795                        const pcre_study_data *, pcre_study_data *);                         const pcre_study_data *, pcre_study_data *);
1796  extern int          _pcre_valid_utf8(const uschar *, int);  extern int           _pcre_valid_utf8(USPTR, int);
1797  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,  extern BOOL          _pcre_was_newline(USPTR, int, USPTR, int *, BOOL);
1798                        int *, BOOL);  extern BOOL          _pcre_xclass(int, const uschar *);
 extern BOOL         _pcre_xclass(int, const uschar *);  
1799    
1800    
1801  /* Unicode character database (UCD) */  /* Unicode character database (UCD) */

Legend:
Removed from v.412  
changed lines
  Added in v.535

  ViewVC Help
Powered by ViewVC 1.1.5