/[pcre]/code/branches/pcre16/pcre_internal.h
ViewVC logotype

Diff of /code/branches/pcre16/pcre_internal.h

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

code/trunk/pcre_internal.h revision 483 by ph10, Mon Jan 4 16:19:17 2010 UTC code/branches/pcre16/pcre_internal.h revision 756 by ph10, Mon Nov 21 10:48:42 2011 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) 1997-2010 University of Cambridge             Copyright (c) 1997-2011 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 158  set, we ensure here that it has no effec Line 158  set, we ensure here that it has no effec
158  #define PCRE_CALL_CONVENTION  #define PCRE_CALL_CONVENTION
159  #endif  #endif
160    
161  /* We need to have types that specify unsigned 16-bit and 32-bit integers. We  /* We need to have types that specify unsigned 8, 16 and 32-bit integers. We
162  cannot determine these outside the compilation (e.g. by running a program as  cannot determine these outside the compilation (e.g. by running a program as
163  part of "configure") because PCRE is often cross-compiled for use on other  part of "configure") because PCRE is often cross-compiled for use on other
164  systems. Instead we make use of the maximum sizes that are available at  systems. Instead we make use of the maximum sizes that are available at
165  preprocessor time in standard C environments. */  preprocessor time in standard C environments. */
166    
167    typedef unsigned char pcre_uint8;
168    
169  #if USHRT_MAX == 65535  #if USHRT_MAX == 65535
170    typedef unsigned short pcre_uint16;    typedef unsigned short pcre_uint16;
171    typedef short pcre_int16;    typedef short pcre_int16;
# Line 184  preprocessor time in standard C environm Line 186  preprocessor time in standard C environm
186    #error Cannot determine a type for 32-bit unsigned integers    #error Cannot determine a type for 32-bit unsigned integers
187  #endif  #endif
188    
189  /* When checking for integer overflow in pcre_compile(), we need to handle  /* When checking for integer overflow in pcre_compile(), we need to handle
190  large integers. If a 64-bit integer type is available, we can use that.  large integers. If a 64-bit integer type is available, we can use that.
191  Otherwise we have to cast to double, which of course requires floating point  Otherwise we have to cast to double, which of course requires floating point
192  arithmetic. Handle this by defining a macro for the appropriate type. If  arithmetic. Handle this by defining a macro for the appropriate type. If
193  stdint.h is available, include it; it may define INT64_MAX. The macro int64_t  stdint.h is available, include it; it may define INT64_MAX. Systems that do not
194  may be set by "configure". */  have stdint.h (e.g. Solaris) may have inttypes.h. The macro int64_t may be set
195    by "configure". */
196    
197  #if HAVE_STDINT_H  #if HAVE_STDINT_H
198  #include <stdint.h>  #include <stdint.h>
199    #elif HAVE_INTTYPES_H
200    #include <inttypes.h>
201  #endif  #endif
202    
203  #if defined INT64_MAX || defined int64_t  #if defined INT64_MAX || defined int64_t
# Line 204  may be set by "configure". */ Line 209  may be set by "configure". */
209  /* All character handling must be done as unsigned characters. Otherwise there  /* All character handling must be done as unsigned characters. Otherwise there
210  are problems with top-bit-set characters and functions such as isspace().  are problems with top-bit-set characters and functions such as isspace().
211  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
212  should make things easier for callers. We define a short type for unsigned char  should make things easier for callers. */
 to save lots of typing. I tried "uchar", but it causes problems on Digital  
 Unix, where it is defined in sys/types, so use "uschar" instead. */  
213    
214  typedef unsigned char uschar;  typedef unsigned char pcre_uchar;
215    
216  /* This is an unsigned int value that no character can ever have. UTF-8  /* This is an unsigned int value that no character can ever have. UTF-8
217  characters only go up to 0x7fffffff (though Unicode doesn't go beyond  characters only go up to 0x7fffffff (though Unicode doesn't go beyond
# Line 264  used for the external interface and appe Line 267  used for the external interface and appe
267  must begin with PCRE_. */  must begin with PCRE_. */
268    
269  #ifdef CUSTOM_SUBJECT_PTR  #ifdef CUSTOM_SUBJECT_PTR
270  #define PCRE_SPTR CUSTOM_SUBJECT_PTR  #define PCRE_PUCHAR CUSTOM_SUBJECT_PTR
 #define USPTR CUSTOM_SUBJECT_PTR  
271  #else  #else
272  #define PCRE_SPTR const char *  #define PCRE_PUCHAR const pcre_uchar *
 #define USPTR const unsigned char *  
273  #endif  #endif
274    
275    
# Line 405  capturing parenthesis numbers in back re Line 406  capturing parenthesis numbers in back re
406    
407  /* When UTF-8 encoding is being used, a character is no longer just a single  /* When UTF-8 encoding is being used, a character is no longer just a single
408  byte. The macros for character handling generate simple sequences when used in  byte. The macros for character handling generate simple sequences when used in
409  byte-mode, and more complicated ones for UTF-8 characters. BACKCHAR should  byte-mode, and more complicated ones for UTF-8 characters. GETCHARLENTEST is
410  never be called in byte mode. To make sure it can never even appear when UTF-8  not used when UTF-8 is not supported, so it is not defined, and BACKCHAR should
411  support is omitted, we don't even define it. */  never be called in byte mode. To make sure they can never even appear when
412    UTF-8 support is omitted, we don't even define them. */
413    
414  #ifndef SUPPORT_UTF8  #ifndef SUPPORT_UTF8
415  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
# Line 415  support is omitted, we don't even define Line 417  support is omitted, we don't even define
417  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
418  #define GETCHARINCTEST(c, eptr) c = *eptr++;  #define GETCHARINCTEST(c, eptr) c = *eptr++;
419  #define GETCHARLEN(c, eptr, len) c = *eptr;  #define GETCHARLEN(c, eptr, len) c = *eptr;
420    /* #define GETCHARLENTEST(c, eptr, len) */
421  /* #define BACKCHAR(eptr) */  /* #define BACKCHAR(eptr) */
422    
423  #else   /* SUPPORT_UTF8 */  #else   /* SUPPORT_UTF8 */
424    
425    /* These macros were originally written in the form of loops that used data
426    from the tables whose names start with _pcre_utf8_table. They were rewritten by
427    a user so as not to use loops, because in some environments this gives a
428    significant performance advantage, and it seems never to do any harm. */
429    
430    /* Base macro to pick up the remaining bytes of a UTF-8 character, not
431    advancing the pointer. */
432    
433    #define GETUTF8(c, eptr) \
434        { \
435        if ((c & 0x20) == 0) \
436          c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
437        else if ((c & 0x10) == 0) \
438          c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
439        else if ((c & 0x08) == 0) \
440          c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
441          ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
442        else if ((c & 0x04) == 0) \
443          c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
444              ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
445              (eptr[4] & 0x3f); \
446        else \
447          c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
448              ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
449              ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
450        }
451    
452  /* Get the next UTF-8 character, not advancing the pointer. This is called when  /* Get the next UTF-8 character, not advancing the pointer. This is called when
453  we know we are in UTF-8 mode. */  we know we are in UTF-8 mode. */
454    
455  #define GETCHAR(c, eptr) \  #define GETCHAR(c, eptr) \
456    c = *eptr; \    c = *eptr; \
457    if (c >= 0xc0) \    if (c >= 0xc0) GETUTF8(c, eptr);
     { \  
     int gcii; \  
     int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \  
     int gcss = 6*gcaa; \  
     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
     for (gcii = 1; gcii <= gcaa; gcii++) \  
       { \  
       gcss -= 6; \  
       c |= (eptr[gcii] & 0x3f) << gcss; \  
       } \  
     }  
458    
459  /* Get the next UTF-8 character, testing for UTF-8 mode, and not advancing the  /* Get the next UTF-8 character, testing for UTF-8 mode, and not advancing the
460  pointer. */  pointer. */
461    
462  #define GETCHARTEST(c, eptr) \  #define GETCHARTEST(c, eptr) \
463    c = *eptr; \    c = *eptr; \
464    if (utf8 && c >= 0xc0) \    if (utf8 && c >= 0xc0) GETUTF8(c, eptr);
465    
466    /* Base macro to pick up the remaining bytes of a UTF-8 character, advancing
467    the pointer. */
468    
469    #define GETUTF8INC(c, eptr) \
470      { \      { \
471      int gcii; \      if ((c & 0x20) == 0) \
472      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \        c = ((c & 0x1f) << 6) | (*eptr++ & 0x3f); \
473      int gcss = 6*gcaa; \      else if ((c & 0x10) == 0) \
474      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \        { \
475      for (gcii = 1; gcii <= gcaa; gcii++) \        c = ((c & 0x0f) << 12) | ((*eptr & 0x3f) << 6) | (eptr[1] & 0x3f); \
476          eptr += 2; \
477          } \
478        else if ((c & 0x08) == 0) \
479          { \
480          c = ((c & 0x07) << 18) | ((*eptr & 0x3f) << 12) | \
481              ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
482          eptr += 3; \
483          } \
484        else if ((c & 0x04) == 0) \
485        { \        { \
486        gcss -= 6; \        c = ((c & 0x03) << 24) | ((*eptr & 0x3f) << 18) | \
487        c |= (eptr[gcii] & 0x3f) << gcss; \            ((eptr[1] & 0x3f) << 12) | ((eptr[2] & 0x3f) << 6) | \
488              (eptr[3] & 0x3f); \
489          eptr += 4; \
490          } \
491        else \
492          { \
493          c = ((c & 0x01) << 30) | ((*eptr & 0x3f) << 24) | \
494              ((eptr[1] & 0x3f) << 18) | ((eptr[2] & 0x3f) << 12) | \
495              ((eptr[3] & 0x3f) << 6) | (eptr[4] & 0x3f); \
496          eptr += 5; \
497        } \        } \
498      }      }
499    
# Line 460  know we are in UTF-8 mode. */ Line 502  know we are in UTF-8 mode. */
502    
503  #define GETCHARINC(c, eptr) \  #define GETCHARINC(c, eptr) \
504    c = *eptr++; \    c = *eptr++; \
505    if (c >= 0xc0) \    if (c >= 0xc0) GETUTF8INC(c, eptr);
     { \  
     int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \  
     int gcss = 6*gcaa; \  
     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
     while (gcaa-- > 0) \  
       { \  
       gcss -= 6; \  
       c |= (*eptr++ & 0x3f) << gcss; \  
       } \  
     }  
506    
507  /* 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.
508    This is called when we don't know if we are in UTF-8 mode. */
509    
510  #define GETCHARINCTEST(c, eptr) \  #define GETCHARINCTEST(c, eptr) \
511    c = *eptr++; \    c = *eptr++; \
512    if (utf8 && c >= 0xc0) \    if (utf8 && c >= 0xc0) GETUTF8INC(c, eptr);
513    
514    /* Base macro to pick up the remaining bytes of a UTF-8 character, not
515    advancing the pointer, incrementing the length. */
516    
517    #define GETUTF8LEN(c, eptr, len) \
518      { \      { \
519      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      if ((c & 0x20) == 0) \
     int gcss = 6*gcaa; \  
     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
     while (gcaa-- > 0) \  
520        { \        { \
521        gcss -= 6; \        c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
522        c |= (*eptr++ & 0x3f) << gcss; \        len++; \
523          } \
524        else if ((c & 0x10)  == 0) \
525          { \
526          c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
527          len += 2; \
528          } \
529        else if ((c & 0x08)  == 0) \
530          {\
531          c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
532              ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
533          len += 3; \
534          } \
535        else if ((c & 0x04)  == 0) \
536          { \
537          c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
538              ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
539              (eptr[4] & 0x3f); \
540          len += 4; \
541          } \
542        else \
543          {\
544          c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
545              ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
546              ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
547          len += 5; \
548        } \        } \
549      }      }
550    
# Line 493  if there are extra bytes. This is called Line 553  if there are extra bytes. This is called
553    
554  #define GETCHARLEN(c, eptr, len) \  #define GETCHARLEN(c, eptr, len) \
555    c = *eptr; \    c = *eptr; \
556    if (c >= 0xc0) \    if (c >= 0xc0) GETUTF8LEN(c, eptr, len);
     { \  
     int gcii; \  
     int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \  
     int gcss = 6*gcaa; \  
     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
     for (gcii = 1; gcii <= gcaa; gcii++) \  
       { \  
       gcss -= 6; \  
       c |= (eptr[gcii] & 0x3f) << gcss; \  
       } \  
     len += gcaa; \  
     }  
557    
558  /* 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
559  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
560  know we are in UTF-8 mode. */  do not know if we are in UTF-8 mode. */
561    
562  #define GETCHARLENTEST(c, eptr, len) \  #define GETCHARLENTEST(c, eptr, len) \
563    c = *eptr; \    c = *eptr; \
564    if (utf8 && c >= 0xc0) \    if (utf8 && c >= 0xc0) GETUTF8LEN(c, eptr, len);
     { \  
     int gcii; \  
     int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \  
     int gcss = 6*gcaa; \  
     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
     for (gcii = 1; gcii <= gcaa; gcii++) \  
       { \  
       gcss -= 6; \  
       c |= (eptr[gcii] & 0x3f) << gcss; \  
       } \  
     len += gcaa; \  
     }  
565    
566  /* If the pointer is not at the start of a character, move it back until  /* If the pointer is not at the start of a character, move it back until
567  it is. This is called only in UTF-8 mode - we don't put a test within the macro  it is. This is called only in UTF-8 mode - we don't put a test within the macro
# Line 533  because almost all calls are already wit Line 569  because almost all calls are already wit
569    
570  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
571    
572  #endif  #endif  /* SUPPORT_UTF8 */
573    
574    
575  /* In case there is no definition of offsetof() provided - though any proper  /* In case there is no definition of offsetof() provided - though any proper
# Line 544  Standard C system should have one. */ Line 580  Standard C system should have one. */
580  #endif  #endif
581    
582    
 /* These are the public options that can change during matching. */  
   
 #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  
   
583  /* Private flags containing information about the compiled regex. They used to  /* Private flags containing information about the compiled regex. They used to
584  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
585  are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as  are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as
# Line 560  compatibility. */ Line 592  compatibility. */
592  #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */  #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */
593  #define PCRE_JCHANGED      0x0010  /* j option used in regex */  #define PCRE_JCHANGED      0x0010  /* j option used in regex */
594  #define PCRE_HASCRORLF     0x0020  /* explicit \r or \n in pattern */  #define PCRE_HASCRORLF     0x0020  /* explicit \r or \n in pattern */
595    #define PCRE_HASTHEN       0x0040  /* pattern contains (*THEN) */
596    
597  /* Options for the "extra" block produced by pcre_study(). */  /* Flags for the "extra" block produced by pcre_study(). */
598    
599  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */  #define PCRE_STUDY_MAPPED  0x0001  /* a map of starting chars exists */
600  #define PCRE_STUDY_MINLEN   0x02     /* a minimum length field exists */  #define PCRE_STUDY_MINLEN  0x0002  /* a minimum length field exists */
601    
602  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
603  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
# Line 577  time, run time, or study time, respectiv Line 610  time, run time, or study time, respectiv
610     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
611     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| \
612     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
613     PCRE_JAVASCRIPT_COMPAT)     PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE)
614    
615  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
616    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
# Line 590  time, run time, or study time, respectiv Line 623  time, run time, or study time, respectiv
623     PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \     PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
624     PCRE_NO_START_OPTIMIZE)     PCRE_NO_START_OPTIMIZE)
625    
626  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS \
627       PCRE_STUDY_JIT_COMPILE
628    
629  /* Magic number to provide a small check against being handed junk. Also used  /* Magic number to provide a small check against being handed junk. Also used
630  to detect whether a pattern was compiled on a host of different endianness. */  to detect whether a pattern was compiled on a host of different endianness. */
# Line 872  so that PCRE works on both ASCII and EBC Line 906  so that PCRE works on both ASCII and EBC
906  #define STRING_COMMIT0              "COMMIT\0"  #define STRING_COMMIT0              "COMMIT\0"
907  #define STRING_F0                   "F\0"  #define STRING_F0                   "F\0"
908  #define STRING_FAIL0                "FAIL\0"  #define STRING_FAIL0                "FAIL\0"
909    #define STRING_MARK0                "MARK\0"
910  #define STRING_PRUNE0               "PRUNE\0"  #define STRING_PRUNE0               "PRUNE\0"
911  #define STRING_SKIP0                "SKIP\0"  #define STRING_SKIP0                "SKIP\0"
912  #define STRING_THEN                 "THEN"  #define STRING_THEN                 "THEN"
# Line 893  so that PCRE works on both ASCII and EBC Line 928  so that PCRE works on both ASCII and EBC
928    
929  #define STRING_DEFINE               "DEFINE"  #define STRING_DEFINE               "DEFINE"
930    
931  #define STRING_CR_RIGHTPAR          "CR)"  #define STRING_CR_RIGHTPAR             "CR)"
932  #define STRING_LF_RIGHTPAR          "LF)"  #define STRING_LF_RIGHTPAR             "LF)"
933  #define STRING_CRLF_RIGHTPAR        "CRLF)"  #define STRING_CRLF_RIGHTPAR           "CRLF)"
934  #define STRING_ANY_RIGHTPAR         "ANY)"  #define STRING_ANY_RIGHTPAR            "ANY)"
935  #define STRING_ANYCRLF_RIGHTPAR     "ANYCRLF)"  #define STRING_ANYCRLF_RIGHTPAR        "ANYCRLF)"
936  #define STRING_BSR_ANYCRLF_RIGHTPAR "BSR_ANYCRLF)"  #define STRING_BSR_ANYCRLF_RIGHTPAR    "BSR_ANYCRLF)"
937  #define STRING_BSR_UNICODE_RIGHTPAR "BSR_UNICODE)"  #define STRING_BSR_UNICODE_RIGHTPAR    "BSR_UNICODE)"
938  #define STRING_UTF8_RIGHTPAR        "UTF8)"  #define STRING_UTF8_RIGHTPAR           "UTF8)"
939    #define STRING_UCP_RIGHTPAR            "UCP)"
940    #define STRING_NO_START_OPT_RIGHTPAR   "NO_START_OPT)"
941    
942  #else  /* SUPPORT_UTF8 */  #else  /* SUPPORT_UTF8 */
943    
# Line 1124  only. */ Line 1161  only. */
1161  #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"
1162  #define STRING_F0                   STR_F "\0"  #define STRING_F0                   STR_F "\0"
1163  #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"  #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"
1164    #define STRING_MARK0                STR_M STR_A STR_R STR_K "\0"
1165  #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"
1166  #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"  #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"
1167  #define STRING_THEN                 STR_T STR_H STR_E STR_N  #define STRING_THEN                 STR_T STR_H STR_E STR_N
# Line 1145  only. */ Line 1183  only. */
1183    
1184  #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E  #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E
1185    
1186  #define STRING_CR_RIGHTPAR          STR_C STR_R STR_RIGHT_PARENTHESIS  #define STRING_CR_RIGHTPAR             STR_C STR_R STR_RIGHT_PARENTHESIS
1187  #define STRING_LF_RIGHTPAR          STR_L STR_F STR_RIGHT_PARENTHESIS  #define STRING_LF_RIGHTPAR             STR_L STR_F STR_RIGHT_PARENTHESIS
1188  #define STRING_CRLF_RIGHTPAR        STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS  #define STRING_CRLF_RIGHTPAR           STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1189  #define STRING_ANY_RIGHTPAR         STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS  #define STRING_ANY_RIGHTPAR            STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1190  #define STRING_ANYCRLF_RIGHTPAR     STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS  #define STRING_ANYCRLF_RIGHTPAR        STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1191  #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
1192  #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
1193  #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
1194    #define STRING_UCP_RIGHTPAR            STR_U STR_C STR_P STR_RIGHT_PARENTHESIS
1195    #define STRING_NO_START_OPT_RIGHTPAR   STR_N STR_O STR_UNDERSCORE STR_S STR_T STR_A STR_R STR_T STR_UNDERSCORE STR_O STR_P STR_T STR_RIGHT_PARENTHESIS
1196    
1197  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF8 */
1198    
# Line 1185  only. */ Line 1225  only. */
1225    
1226  #define PT_ANY        0    /* Any property - matches all chars */  #define PT_ANY        0    /* Any property - matches all chars */
1227  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */
1228  #define PT_GC         2    /* General characteristic (e.g. L) */  #define PT_GC         2    /* Specified general characteristic (e.g. L) */
1229  #define PT_PC         3    /* Particular characteristic (e.g. Lu) */  #define PT_PC         3    /* Specified particular characteristic (e.g. Lu) */
1230  #define PT_SC         4    /* Script (e.g. Han) */  #define PT_SC         4    /* Script (e.g. Han) */
1231    #define PT_ALNUM      5    /* Alphanumeric - the union of L and N */
1232    #define PT_SPACE      6    /* Perl space - Z plus 9,10,12,13 */
1233    #define PT_PXSPACE    7    /* POSIX space - Z plus 9,10,11,12,13 */
1234    #define PT_WORD       8    /* Word - L plus N plus underscore */
1235    
1236  /* 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
1237  contain UTF-8 characters with values greater than 255. */  contain UTF-8 characters with values greater than 255. */
# Line 1204  contain UTF-8 characters with values gre Line 1248  contain UTF-8 characters with values gre
1248  /* 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
1249  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
1250  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
1251  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
1252  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
1253  (which is used for [^] in JavaScript compatibility mode).  used for [^] in JavaScript compatibility mode, and for \C in non-utf8 mode. In
1254    non-DOTALL mode, "." behaves like \N.
1255    
1256    The special values ESC_DU, ESC_du, etc. are used instead of ESC_D, ESC_d, etc.
1257    when PCRE_UCP is set, when replacement of \d etc by \p sequences is required.
1258    They must be contiguous, and remain in order so that the replacements can be
1259    looked up from a table.
1260    
1261  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
1262  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 1216  put in between that don't consume a char Line 1266  put in between that don't consume a char
1266  */  */
1267    
1268  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,
1269         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,
1270         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,
1271           ESC_E, ESC_Q, ESC_g, ESC_k,
1272           ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu,
1273         ESC_REF };         ESC_REF };
1274    
   
1275  /* 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
1276  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.
1277    
# Line 1244  enum { Line 1295  enum {
1295    OP_WHITESPACE,         /*  9 \s */    OP_WHITESPACE,         /*  9 \s */
1296    OP_NOT_WORDCHAR,       /* 10 \W */    OP_NOT_WORDCHAR,       /* 10 \W */
1297    OP_WORDCHAR,           /* 11 \w */    OP_WORDCHAR,           /* 11 \w */
1298    OP_ANY,            /* 12 Match any character (subject to DOTALL) */  
1299    OP_ALLANY,         /* 13 Match any character (not subject to DOTALL) */    OP_ANY,            /* 12 Match any character except newline */
1300      OP_ALLANY,         /* 13 Match any character */
1301    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 */
1302    OP_NOTPROP,        /* 15 \P (not Unicode property) */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
1303    OP_PROP,           /* 16 \p (Unicode property) */    OP_PROP,           /* 16 \p (Unicode property) */
# Line 1258  enum { Line 1310  enum {
1310    OP_EODN,           /* 23 End of data or \n at end of data: \Z. */    OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
1311    OP_EOD,            /* 24 End of data: \z */    OP_EOD,            /* 24 End of data: \z */
1312    
1313    OP_OPT,            /* 25 Set runtime options */    OP_CIRC,           /* 25 Start of line - not multiline */
1314    OP_CIRC,           /* 26 Start of line - varies with multiline switch */    OP_CIRCM,          /* 26 Start of line - multiline */
1315    OP_DOLL,           /* 27 End of line - varies with multiline switch */    OP_DOLL,           /* 27 End of line - not multiline */
1316    OP_CHAR,           /* 28 Match one character, casefully */    OP_DOLLM,          /* 28 End of line - multiline */
1317    OP_CHARNC,         /* 29 Match one character, caselessly */    OP_CHAR,           /* 29 Match one character, casefully */
1318    OP_NOT,            /* 30 Match one character, not the following one */    OP_CHARI,          /* 30 Match one character, caselessly */
1319      OP_NOT,            /* 31 Match one character, not the given one, casefully */
1320    OP_STAR,           /* 31 The maximizing and minimizing versions of */    OP_NOTI,           /* 32 Match one character, not the given one, caselessly */
1321    OP_MINSTAR,        /* 32 these six opcodes must come in pairs, with */  
1322    OP_PLUS,           /* 33 the minimizing one second. */    /* The following sets of 13 opcodes must always be kept in step because
1323    OP_MINPLUS,        /* 34 This first set applies to single characters.*/    the offset from the first one is used to generate the others. */
1324    OP_QUERY,          /* 35 */  
1325    OP_MINQUERY,       /* 36 */    /**** Single characters, caseful, must precede the caseless ones ****/
1326    
1327    OP_UPTO,           /* 37 From 0 to n matches */    OP_STAR,           /* 33 The maximizing and minimizing versions of */
1328    OP_MINUPTO,        /* 38 */    OP_MINSTAR,        /* 34 these six opcodes must come in pairs, with */
1329    OP_EXACT,          /* 39 Exactly n matches */    OP_PLUS,           /* 35 the minimizing one second. */
1330      OP_MINPLUS,        /* 36 */
1331    OP_POSSTAR,        /* 40 Possessified star */    OP_QUERY,          /* 37 */
1332    OP_POSPLUS,        /* 41 Possessified plus */    OP_MINQUERY,       /* 38 */
1333    OP_POSQUERY,       /* 42 Posesssified query */  
1334    OP_POSUPTO,        /* 43 Possessified upto */    OP_UPTO,           /* 39 From 0 to n matches of one character, caseful*/
1335      OP_MINUPTO,        /* 40 */
1336    OP_NOTSTAR,        /* 44 The maximizing and minimizing versions of */    OP_EXACT,          /* 41 Exactly n matches */
1337    OP_NOTMINSTAR,     /* 45 these six opcodes must come in pairs, with */  
1338    OP_NOTPLUS,        /* 46 the minimizing one second. They must be in */    OP_POSSTAR,        /* 42 Possessified star, caseful */
1339    OP_NOTMINPLUS,     /* 47 exactly the same order as those above. */    OP_POSPLUS,        /* 43 Possessified plus, caseful */
1340    OP_NOTQUERY,       /* 48 This set applies to "not" single characters. */    OP_POSQUERY,       /* 44 Posesssified query, caseful */
1341    OP_NOTMINQUERY,    /* 49 */    OP_POSUPTO,        /* 45 Possessified upto, caseful */
1342    
1343    OP_NOTUPTO,        /* 50 From 0 to n matches */    /**** Single characters, caseless, must follow the caseful ones */
1344    OP_NOTMINUPTO,     /* 51 */  
1345    OP_NOTEXACT,       /* 52 Exactly n matches */    OP_STARI,          /* 46 */
1346      OP_MINSTARI,       /* 47 */
1347    OP_NOTPOSSTAR,     /* 53 Possessified versions */    OP_PLUSI,          /* 48 */
1348    OP_NOTPOSPLUS,     /* 54 */    OP_MINPLUSI,       /* 49 */
1349    OP_NOTPOSQUERY,    /* 55 */    OP_QUERYI,         /* 50 */
1350    OP_NOTPOSUPTO,     /* 56 */    OP_MINQUERYI,      /* 51 */
1351    
1352    OP_TYPESTAR,       /* 57 The maximizing and minimizing versions of */    OP_UPTOI,          /* 52 From 0 to n matches of one character, caseless */
1353    OP_TYPEMINSTAR,    /* 58 these six opcodes must come in pairs, with */    OP_MINUPTOI,       /* 53 */
1354    OP_TYPEPLUS,       /* 59 the minimizing one second. These codes must */    OP_EXACTI,         /* 54 */
1355    OP_TYPEMINPLUS,    /* 60 be in exactly the same order as those above. */  
1356    OP_TYPEQUERY,      /* 61 This set applies to character types such as \d */    OP_POSSTARI,       /* 55 Possessified star, caseless */
1357    OP_TYPEMINQUERY,   /* 62 */    OP_POSPLUSI,       /* 56 Possessified plus, caseless */
1358      OP_POSQUERYI,      /* 57 Posesssified query, caseless */
1359    OP_TYPEUPTO,       /* 63 From 0 to n matches */    OP_POSUPTOI,       /* 58 Possessified upto, caseless */
1360    OP_TYPEMINUPTO,    /* 64 */  
1361    OP_TYPEEXACT,      /* 65 Exactly n matches */    /**** The negated ones must follow the non-negated ones, and match them ****/
1362      /**** Negated single character, caseful; must precede the caseless ones ****/
1363    OP_TYPEPOSSTAR,    /* 66 Possessified versions */  
1364    OP_TYPEPOSPLUS,    /* 67 */    OP_NOTSTAR,        /* 59 The maximizing and minimizing versions of */
1365    OP_TYPEPOSQUERY,   /* 68 */    OP_NOTMINSTAR,     /* 60 these six opcodes must come in pairs, with */
1366    OP_TYPEPOSUPTO,    /* 69 */    OP_NOTPLUS,        /* 61 the minimizing one second. They must be in */
1367      OP_NOTMINPLUS,     /* 62 exactly the same order as those above. */
1368    OP_CRSTAR,         /* 70 The maximizing and minimizing versions of */    OP_NOTQUERY,       /* 63 */
1369    OP_CRMINSTAR,      /* 71 all these opcodes must come in pairs, with */    OP_NOTMINQUERY,    /* 64 */
1370    OP_CRPLUS,         /* 72 the minimizing one second. These codes must */  
1371    OP_CRMINPLUS,      /* 73 be in exactly the same order as those above. */    OP_NOTUPTO,        /* 65 From 0 to n matches, caseful */
1372    OP_CRQUERY,        /* 74 These are for character classes and back refs */    OP_NOTMINUPTO,     /* 66 */
1373    OP_CRMINQUERY,     /* 75 */    OP_NOTEXACT,       /* 67 Exactly n matches */
1374    OP_CRRANGE,        /* 76 These are different to the three sets above. */  
1375    OP_CRMINRANGE,     /* 77 */    OP_NOTPOSSTAR,     /* 68 Possessified versions, caseful */
1376      OP_NOTPOSPLUS,     /* 69 */
1377    OP_CLASS,          /* 78 Match a character class, chars < 256 only */    OP_NOTPOSQUERY,    /* 70 */
1378    OP_NCLASS,         /* 79 Same, but the bitmap was created from a negative    OP_NOTPOSUPTO,     /* 71 */
1379                             class - the difference is relevant only when a UTF-8  
1380                             character > 255 is encountered. */    /**** Negated single character, caseless; must follow the caseful ones ****/
1381    
1382    OP_XCLASS,         /* 80 Extended class for handling UTF-8 chars within the    OP_NOTSTARI,       /* 72 */
1383                             class. This does both positive and negative. */    OP_NOTMINSTARI,    /* 73 */
1384      OP_NOTPLUSI,       /* 74 */
1385    OP_REF,            /* 81 Match a back reference */    OP_NOTMINPLUSI,    /* 75 */
1386    OP_RECURSE,        /* 82 Match a numbered subpattern (possibly recursive) */    OP_NOTQUERYI,      /* 76 */
1387    OP_CALLOUT,        /* 83 Call out to external function if provided */    OP_NOTMINQUERYI,   /* 77 */
1388    
1389    OP_ALT,            /* 84 Start of alternation */    OP_NOTUPTOI,       /* 78 From 0 to n matches, caseless */
1390    OP_KET,            /* 85 End of group that doesn't have an unbounded repeat */    OP_NOTMINUPTOI,    /* 79 */
1391    OP_KETRMAX,        /* 86 These two must remain together and in this */    OP_NOTEXACTI,      /* 80 Exactly n matches */
1392    OP_KETRMIN,        /* 87 order. They are for groups the repeat for ever. */  
1393      OP_NOTPOSSTARI,    /* 81 Possessified versions, caseless */
1394    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/    OP_NOTPOSPLUSI,    /* 82 */
1395      OP_NOTPOSQUERYI,   /* 83 */
1396    OP_ASSERT,         /* 88 Positive lookahead */    OP_NOTPOSUPTOI,    /* 84 */
1397    OP_ASSERT_NOT,     /* 89 Negative lookahead */  
1398    OP_ASSERTBACK,     /* 90 Positive lookbehind */    /**** Character types ****/
1399    OP_ASSERTBACK_NOT, /* 91 Negative lookbehind */  
1400    OP_REVERSE,        /* 92 Move pointer back - used in lookbehind assertions */    OP_TYPESTAR,       /* 85 The maximizing and minimizing versions of */
1401      OP_TYPEMINSTAR,    /* 86 these six opcodes must come in pairs, with */
1402    /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,    OP_TYPEPLUS,       /* 87 the minimizing one second. These codes must */
1403    as there's a test for >= ONCE for a subpattern that isn't an assertion. */    OP_TYPEMINPLUS,    /* 88 be in exactly the same order as those above. */
1404      OP_TYPEQUERY,      /* 89 */
1405    OP_ONCE,           /* 93 Atomic group */    OP_TYPEMINQUERY,   /* 90 */
1406    OP_BRA,            /* 94 Start of non-capturing bracket */  
1407    OP_CBRA,           /* 95 Start of capturing bracket */    OP_TYPEUPTO,       /* 91 From 0 to n matches */
1408    OP_COND,           /* 96 Conditional group */    OP_TYPEMINUPTO,    /* 92 */
1409      OP_TYPEEXACT,      /* 93 Exactly n matches */
1410    
1411      OP_TYPEPOSSTAR,    /* 94 Possessified versions */
1412      OP_TYPEPOSPLUS,    /* 95 */
1413      OP_TYPEPOSQUERY,   /* 96 */
1414      OP_TYPEPOSUPTO,    /* 97 */
1415    
1416      /* These are used for character classes and back references; only the
1417      first six are the same as the sets above. */
1418    
1419      OP_CRSTAR,         /* 98 The maximizing and minimizing versions of */
1420      OP_CRMINSTAR,      /* 99 all these opcodes must come in pairs, with */
1421      OP_CRPLUS,         /* 100 the minimizing one second. These codes must */
1422      OP_CRMINPLUS,      /* 101 be in exactly the same order as those above. */
1423      OP_CRQUERY,        /* 102 */
1424      OP_CRMINQUERY,     /* 103 */
1425    
1426      OP_CRRANGE,        /* 104 These are different to the three sets above. */
1427      OP_CRMINRANGE,     /* 105 */
1428    
1429      /* End of quantifier opcodes */
1430    
1431      OP_CLASS,          /* 106 Match a character class, chars < 256 only */
1432      OP_NCLASS,         /* 107 Same, but the bitmap was created from a negative
1433                                  class - the difference is relevant only when a
1434                                  UTF-8 character > 255 is encountered. */
1435      OP_XCLASS,         /* 108 Extended class for handling UTF-8 chars within the
1436                                  class. This does both positive and negative. */
1437      OP_REF,            /* 109 Match a back reference, casefully */
1438      OP_REFI,           /* 110 Match a back reference, caselessly */
1439      OP_RECURSE,        /* 111 Match a numbered subpattern (possibly recursive) */
1440      OP_CALLOUT,        /* 112 Call out to external function if provided */
1441    
1442      OP_ALT,            /* 113 Start of alternation */
1443      OP_KET,            /* 114 End of group that doesn't have an unbounded repeat */
1444      OP_KETRMAX,        /* 115 These two must remain together and in this */
1445      OP_KETRMIN,        /* 116 order. They are for groups the repeat for ever. */
1446      OP_KETRPOS,        /* 117 Possessive unlimited repeat. */
1447    
1448      /* The assertions must come before BRA, CBRA, ONCE, and COND, and the four
1449      asserts must remain in order. */
1450    
1451      OP_REVERSE,        /* 118 Move pointer back - used in lookbehind assertions */
1452      OP_ASSERT,         /* 119 Positive lookahead */
1453      OP_ASSERT_NOT,     /* 120 Negative lookahead */
1454      OP_ASSERTBACK,     /* 121 Positive lookbehind */
1455      OP_ASSERTBACK_NOT, /* 122 Negative lookbehind */
1456    
1457      /* ONCE, ONCE_NC, BRA, BRAPOS, CBRA, CBRAPOS, and COND must come immediately
1458      after the assertions, with ONCE first, as there's a test for >= ONCE for a
1459      subpattern that isn't an assertion. The POS versions must immediately follow
1460      the non-POS versions in each case. */
1461    
1462      OP_ONCE,           /* 123 Atomic group, contains captures */
1463      OP_ONCE_NC,        /* 124 Atomic group containing no captures */
1464      OP_BRA,            /* 125 Start of non-capturing bracket */
1465      OP_BRAPOS,         /* 126 Ditto, with unlimited, possessive repeat */
1466      OP_CBRA,           /* 127 Start of capturing bracket */
1467      OP_CBRAPOS,        /* 128 Ditto, with unlimited, possessive repeat */
1468      OP_COND,           /* 129 Conditional group */
1469    
1470    /* These three must follow the previous three, in the same order. There's a    /* These five must follow the previous five, in the same order. There's a
1471    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
1472    
1473    OP_SBRA,           /* 97 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 130 Start of non-capturing bracket, check empty  */
1474    OP_SCBRA,          /* 98 Start of capturing bracket, check empty */    OP_SBRAPOS,        /* 131 Ditto, with unlimited, possessive repeat */
1475    OP_SCOND,          /* 99 Conditional group, check empty */    OP_SCBRA,          /* 132 Start of capturing bracket, check empty */
1476      OP_SCBRAPOS,       /* 133 Ditto, with unlimited, possessive repeat */
1477      OP_SCOND,          /* 134 Conditional group, check empty */
1478    
1479    /* The next two pairs must (respectively) be kept together. */    /* The next two pairs must (respectively) be kept together. */
1480    
1481    OP_CREF,           /* 100 Used to hold a capture number as condition */    OP_CREF,           /* 135 Used to hold a capture number as condition */
1482    OP_NCREF,          /* 101 Same, but generaged by a name reference*/    OP_NCREF,          /* 136 Same, but generated by a name reference*/
1483    OP_RREF,           /* 102 Used to hold a recursion number as condition */    OP_RREF,           /* 137 Used to hold a recursion number as condition */
1484    OP_NRREF,          /* 103 Same, but generaged by a name reference*/    OP_NRREF,          /* 138 Same, but generated by a name reference*/
1485    OP_DEF,            /* 104 The DEFINE condition */    OP_DEF,            /* 139 The DEFINE condition */
1486    
1487    OP_BRAZERO,        /* 105 These two must remain together and in this */    OP_BRAZERO,        /* 140 These two must remain together and in this */
1488    OP_BRAMINZERO,     /* 106 order. */    OP_BRAMINZERO,     /* 141 order. */
1489      OP_BRAPOSZERO,     /* 142 */
1490    
1491    /* These are backtracking control verbs */    /* These are backtracking control verbs */
1492    
1493    OP_PRUNE,          /* 107 */    OP_MARK,           /* 143 always has an argument */
1494    OP_SKIP,           /* 108 */    OP_PRUNE,          /* 144 */
1495    OP_THEN,           /* 109 */    OP_PRUNE_ARG,      /* 145 same, but with argument */
1496    OP_COMMIT,         /* 110 */    OP_SKIP,           /* 146 */
1497      OP_SKIP_ARG,       /* 147 same, but with argument */
1498      OP_THEN,           /* 148 */
1499      OP_THEN_ARG,       /* 149 same, but with argument */
1500      OP_COMMIT,         /* 150 */
1501    
1502    /* These are forced failure and success verbs */    /* These are forced failure and success verbs */
1503    
1504    OP_FAIL,           /* 111 */    OP_FAIL,           /* 151 */
1505    OP_ACCEPT,         /* 112 */    OP_ACCEPT,         /* 152 */
1506    OP_CLOSE,          /* 113 Used before OP_ACCEPT to close open captures */    OP_ASSERT_ACCEPT,  /* 153 Used inside assertions */
1507      OP_CLOSE,          /* 154 Used before OP_ACCEPT to close open captures */
1508    
1509    /* This is used to skip a subpattern with a {0} quantifier */    /* This is used to skip a subpattern with a {0} quantifier */
1510    
1511    OP_SKIPZERO        /* 114 */    OP_SKIPZERO,       /* 155 */
1512    
1513      /* This is not an opcode, but is used to check that tables indexed by opcode
1514      are the correct length, in order to catch updating errors - there have been
1515      some in the past. */
1516    
1517      OP_TABLE_LENGTH
1518  };  };
1519    
1520  /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro  /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
1521  definitions that follow must also be updated to match. There are also tables  definitions that follow must also be updated to match. There are also tables
1522  called "coptable" cna "poptable" in pcre_dfa_exec.c that must be updated. */  called "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1523    
1524    
1525  /* 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
1526  for debugging. The macro is referenced only in pcre_printint.c. */  for debugging, and some of them are only partial names. The macro is referenced
1527    only in pcre_printint.c, which fills out the full names in many cases (and in
1528    some cases doesn't actually use these names at all). */
1529    
1530  #define OP_NAME_LIST \  #define OP_NAME_LIST \
1531    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
1532    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
1533    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
1534    "extuni",  "\\Z", "\\z",                                        \    "extuni",  "\\Z", "\\z",                                        \
1535    "Opt", "^", "$", "char", "charnc", "not",                       \    "^", "^", "$", "$", "char", "chari", "not", "noti",             \
1536    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??",                                \
1537      "{", "{", "{",                                                  \
1538    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1539    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??",                                \
1540      "{", "{", "{",                                                  \
1541      "*+","++", "?+", "{",                                           \
1542      "*", "*?", "+", "+?", "?", "??",                                \
1543      "{", "{", "{",                                                  \
1544      "*+","++", "?+", "{",                                           \
1545      "*", "*?", "+", "+?", "?", "??",                                \
1546      "{", "{", "{",                                                  \
1547    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1548    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
1549    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1550    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
1551    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Refi",                     \
1552    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Recurse", "Callout",                                           \
1553    "AssertB", "AssertB not", "Reverse",                            \    "Alt", "Ket", "KetRmax", "KetRmin", "KetRpos",                  \
1554    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \    "Reverse", "Assert", "Assert not", "AssertB", "AssertB not",    \
1555      "Once", "Once_NC",                                              \
1556      "Bra", "BraPos", "CBra", "CBraPos",                             \
1557      "Cond",                                                         \
1558      "SBra", "SBraPos", "SCBra", "SCBraPos",                         \
1559      "SCond",                                                        \
1560    "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \    "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \
1561    "Brazero", "Braminzero",                                        \    "Brazero", "Braminzero", "Braposzero",                          \
1562    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \    "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \
1563      "*THEN", "*THEN", "*COMMIT", "*FAIL",                           \
1564      "*ACCEPT", "*ASSERT_ACCEPT",                                    \
1565    "Close", "Skip zero"    "Close", "Skip zero"
1566    
1567    
# Line 1436  in UTF-8 mode. The code that uses this t Line 1579  in UTF-8 mode. The code that uses this t
1579    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \
1580    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
1581    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
1582    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3,                          /* \P, \p                                 */ \
1583    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
1584    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1,                             /* \X                                     */ \
1585      1, 1, 1, 1, 1, 1,              /* \Z, \z, ^, ^M, $, $M                   */ \
1586    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
1587    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Chari  - the minimum length            */ \
1588    2,                             /* not                                    */ \    2,                             /* not                                    */ \
1589    /* Positive single-char repeats                            ** These are  */ \    2,                             /* noti                                   */ \
1590    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \    /* Positive single-char repeats                             ** These are */ \
1591    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??       ** minima in */ \
1592      4, 4, 4,                       /* upto, minupto, exact      ** mode      */ \
1593    2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \    2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \
1594      2, 2, 2, 2, 2, 2,              /* *I, *?I, +I, +?I, ?I, ??I ** UTF-8     */ \
1595      4, 4, 4,                       /* upto I, minupto I, exact I             */ \
1596      2, 2, 2, 4,                    /* *+I, ++I, ?+I, upto+I                  */ \
1597    /* Negative single-char repeats - only for chars < 256                   */ \    /* Negative single-char repeats - only for chars < 256                   */ \
1598    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \
1599    4, 4, 4,                       /* NOT upto, minupto, exact               */ \    4, 4, 4,                       /* NOT upto, minupto, exact               */ \
1600    2, 2, 2, 4,                    /* Possessive *, +, ?, upto               */ \    2, 2, 2, 4,                    /* Possessive NOT *, +, ?, upto           */ \
1601      2, 2, 2, 2, 2, 2,              /* NOT *I, *?I, +I, +?I, ?I, ??I          */ \
1602      4, 4, 4,                       /* NOT upto I, minupto I, exact I         */ \
1603      2, 2, 2, 4,                    /* Possessive NOT *I, +I, ?I, upto I      */ \
1604    /* Positive type repeats                                                 */ \    /* Positive type repeats                                                 */ \
1605    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \
1606    4, 4, 4,                       /* Type upto, minupto, exact              */ \    4, 4, 4,                       /* Type upto, minupto, exact              */ \
# Line 1461  in UTF-8 mode. The code that uses this t Line 1612  in UTF-8 mode. The code that uses this t
1612   33,                             /* NCLASS                                 */ \   33,                             /* NCLASS                                 */ \
1613    0,                             /* XCLASS - variable length               */ \    0,                             /* XCLASS - variable length               */ \
1614    3,                             /* REF                                    */ \    3,                             /* REF                                    */ \
1615      3,                             /* REFI                                   */ \
1616    1+LINK_SIZE,                   /* RECURSE                                */ \    1+LINK_SIZE,                   /* RECURSE                                */ \
1617    2+2*LINK_SIZE,                 /* CALLOUT                                */ \    2+2*LINK_SIZE,                 /* CALLOUT                                */ \
1618    1+LINK_SIZE,                   /* Alt                                    */ \    1+LINK_SIZE,                   /* Alt                                    */ \
1619    1+LINK_SIZE,                   /* Ket                                    */ \    1+LINK_SIZE,                   /* Ket                                    */ \
1620    1+LINK_SIZE,                   /* KetRmax                                */ \    1+LINK_SIZE,                   /* KetRmax                                */ \
1621    1+LINK_SIZE,                   /* KetRmin                                */ \    1+LINK_SIZE,                   /* KetRmin                                */ \
1622      1+LINK_SIZE,                   /* KetRpos                                */ \
1623      1+LINK_SIZE,                   /* Reverse                                */ \
1624    1+LINK_SIZE,                   /* Assert                                 */ \    1+LINK_SIZE,                   /* Assert                                 */ \
1625    1+LINK_SIZE,                   /* Assert not                             */ \    1+LINK_SIZE,                   /* Assert not                             */ \
1626    1+LINK_SIZE,                   /* Assert behind                          */ \    1+LINK_SIZE,                   /* Assert behind                          */ \
1627    1+LINK_SIZE,                   /* Assert behind not                      */ \    1+LINK_SIZE,                   /* Assert behind not                      */ \
   1+LINK_SIZE,                   /* Reverse                                */ \  
1628    1+LINK_SIZE,                   /* ONCE                                   */ \    1+LINK_SIZE,                   /* ONCE                                   */ \
1629      1+LINK_SIZE,                   /* ONCE_NC                                */ \
1630    1+LINK_SIZE,                   /* BRA                                    */ \    1+LINK_SIZE,                   /* BRA                                    */ \
1631      1+LINK_SIZE,                   /* BRAPOS                                 */ \
1632    3+LINK_SIZE,                   /* CBRA                                   */ \    3+LINK_SIZE,                   /* CBRA                                   */ \
1633      3+LINK_SIZE,                   /* CBRAPOS                                */ \
1634    1+LINK_SIZE,                   /* COND                                   */ \    1+LINK_SIZE,                   /* COND                                   */ \
1635    1+LINK_SIZE,                   /* SBRA                                   */ \    1+LINK_SIZE,                   /* SBRA                                   */ \
1636      1+LINK_SIZE,                   /* SBRAPOS                                */ \
1637    3+LINK_SIZE,                   /* SCBRA                                  */ \    3+LINK_SIZE,                   /* SCBRA                                  */ \
1638      3+LINK_SIZE,                   /* SCBRAPOS                               */ \
1639    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
1640    3, 3,                          /* CREF, NCREF                            */ \    3, 3,                          /* CREF, NCREF                            */ \
1641    3, 3,                          /* RREF, NRREF                            */ \    3, 3,                          /* RREF, NRREF                            */ \
1642    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
1643    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1, 1,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */ \
1644    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \    3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \
1645    1, 1, 3, 1                     /* FAIL, ACCEPT, CLOSE, SKIPZERO          */    1, 3,                          /* SKIP, SKIP_ARG                         */ \
1646      1, 3,                          /* THEN, THEN_ARG                         */ \
1647      1, 1, 1, 1,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */ \
1648      3, 1                           /* CLOSE, SKIPZERO  */
1649    
1650  /* 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"
1651  condition. */  condition. */
1652    
1653  #define RREF_ANY  0xffff  #define RREF_ANY  0xffff
1654    
1655  /* 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
1656  tracked. */  easily be tracked. When a new number is added, the table called eint in
1657    pcreposix.c must be updated. */
1658    
1659  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,
1660         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
# Line 1501  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4, Line 1662  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
1662         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
1663         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
1664         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
1665         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65 };         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,
1666           ERR70, ERRCOUNT };
1667    
1668  /* 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
1669  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 1535  typedef struct real_pcre { Line 1697  typedef struct real_pcre {
1697    pcre_uint16 name_count;         /* Number of name items */    pcre_uint16 name_count;         /* Number of name items */
1698    pcre_uint16 ref_count;          /* Reference count */    pcre_uint16 ref_count;          /* Reference count */
1699    
1700    const unsigned char *tables;    /* Pointer to tables or NULL for std */    const pcre_uint8 *tables;       /* Pointer to tables or NULL for std */
1701    const unsigned char *nullpad;   /* NULL padding */    const pcre_uint8 *nullpad;      /* NULL padding */
1702  } real_pcre;  } real_pcre;
1703    
1704  /* The format of the block used to store data from pcre_study(). The same  /* The format of the block used to store data from pcre_study(). The same
# Line 1545  remark (see NOTE above) about extending Line 1707  remark (see NOTE above) about extending
1707  typedef struct pcre_study_data {  typedef struct pcre_study_data {
1708    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1709    pcre_uint32 flags;              /* Private flags */    pcre_uint32 flags;              /* Private flags */
1710    uschar start_bits[32];          /* Starting char bits */    pcre_uint8 start_bits[32];      /* Starting char bits */
1711    pcre_uint32 minlength;          /* Minimum subject length */    pcre_uint32 minlength;          /* Minimum subject length */
1712  } pcre_study_data;  } pcre_study_data;
1713    
1714  /* Structure for building a chain of open capturing subpatterns during  /* Structure for building a chain of open capturing subpatterns during
1715  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
1716  encountered. */  encountered. This is also used to identify subpatterns that contain recursive
1717    back references to themselves, so that they can be made atomic. */
1718    
1719  typedef struct open_capitem {  typedef struct open_capitem {
1720    struct open_capitem *next;    /* Chain link */    struct open_capitem *next;    /* Chain link */
1721    pcre_uint16 number;           /* Capture number */    pcre_uint16 number;           /* Capture number */
1722      pcre_uint16 flag;             /* Set TRUE if recursive back ref */
1723  } open_capitem;  } open_capitem;
1724    
1725  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
1726  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
1727    
1728  typedef struct compile_data {  typedef struct compile_data {
1729    const uschar *lcc;            /* Points to lower casing table */    const pcre_uint8 *lcc;            /* Points to lower casing table */
1730    const uschar *fcc;            /* Points to case-flipping table */    const pcre_uint8 *fcc;            /* Points to case-flipping table */
1731    const uschar *cbits;          /* Points to character type table */    const pcre_uint8 *cbits;          /* Points to character type table */
1732    const uschar *ctypes;         /* Points to table of type maps */    const pcre_uint8 *ctypes;         /* Points to table of type maps */
1733    const uschar *start_workspace;/* The start of working space */    const pcre_uchar *start_workspace;/* The start of working space */
1734    const uschar *start_code;     /* The start of the compiled code */    const pcre_uchar *start_code;     /* The start of the compiled code */
1735    const uschar *start_pattern;  /* The start of the pattern */    const pcre_uchar *start_pattern;  /* The start of the pattern */
1736    const uschar *end_pattern;    /* The end of the pattern */    const pcre_uchar *end_pattern;    /* The end of the pattern */
1737    open_capitem *open_caps;      /* Chain of open capture items */    open_capitem *open_caps;          /* Chain of open capture items */
1738    uschar *hwm;                  /* High watermark of workspace */    pcre_uchar *hwm;                  /* High watermark of workspace */
1739    uschar *name_table;           /* The name/number table */    pcre_uchar *name_table;           /* The name/number table */
1740    int  names_found;             /* Number of entries so far */    int  names_found;                 /* Number of entries so far */
1741    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;             /* Size of each entry */
1742    int  bracount;                /* Count of capturing parens as we compile */    int  bracount;                    /* Count of capturing parens as we compile */
1743    int  final_bracount;          /* Saved value after first pass */    int  final_bracount;              /* Saved value after first pass */
1744    int  top_backref;             /* Maximum back reference */    int  top_backref;                 /* Maximum back reference */
1745    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;         /* Bitmap of low back refs */
1746    int  external_options;        /* External (initial) options */    int  assert_depth;                /* Depth of nested assertions */
1747    int  external_flags;          /* External flag bits to be set */    int  external_options;            /* External (initial) options */
1748    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  external_flags;              /* External flag bits to be set */
1749    BOOL had_accept;              /* (*ACCEPT) encountered */    int  req_varyopt;                 /* "After variable item" flag for reqbyte */
1750    BOOL check_lookbehind;        /* Lookbehinds need later checking */    BOOL had_accept;                  /* (*ACCEPT) encountered */
1751    int  nltype;                  /* Newline type */    BOOL check_lookbehind;            /* Lookbehinds need later checking */
1752    int  nllen;                   /* Newline string length */    int  nltype;                      /* Newline type */
1753    uschar nl[4];                 /* Newline string when fixed length */    int  nllen;                       /* Newline string length */
1754      pcre_uchar nl[4];                 /* Newline string when fixed length */
1755  } compile_data;  } compile_data;
1756    
1757  /* Structure for maintaining a chain of pointers to the currently incomplete  /* Structure for maintaining a chain of pointers to the currently incomplete
1758  branches, for testing for left recursion. */  branches, for testing for left recursion while compiling. */
1759    
1760  typedef struct branch_chain {  typedef struct branch_chain {
1761    struct branch_chain *outer;    struct branch_chain *outer;
1762    uschar *current_branch;    pcre_uchar *current_branch;
1763  } branch_chain;  } branch_chain;
1764    
1765  /* Structure for items in a linked list that represents an explicit recursive  /* Structure for items in a linked list that represents an explicit recursive
1766  call within the pattern. */  call within the pattern; used by pcre_exec(). */
1767    
1768  typedef struct recursion_info {  typedef struct recursion_info {
1769    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
1770    int group_num;                /* Number of group that was called */    int group_num;                  /* Number of group that was called */
1771    const uschar *after_call;     /* "Return value": points after the call in the expr */    int *offset_save;               /* Pointer to start of saved offsets */
1772    USPTR save_start;             /* Old value of mstart */    int saved_max;                  /* Number of saved offsets */
1773    int *offset_save;             /* Pointer to start of saved offsets */    PCRE_PUCHAR subject_position;   /* Position at start of recursion */
   int saved_max;                /* Number of saved offsets */  
   int save_offset_top;          /* Current value of offset_top */  
1774  } recursion_info;  } recursion_info;
1775    
1776    /* A similar structure for pcre_dfa_exec(). */
1777    
1778    typedef struct dfa_recursion_info {
1779      struct dfa_recursion_info *prevrec;
1780      int group_num;
1781      PCRE_PUCHAR subject_position;
1782    } dfa_recursion_info;
1783    
1784  /* 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
1785  pointer at the start of each subpattern, so as to detect when an empty string  pointer at the start of each subpattern, so as to detect when an empty string
1786  has been matched by a subpattern - to break infinite loops. */  has been matched by a subpattern - to break infinite loops; used by
1787    pcre_exec(). */
1788    
1789  typedef struct eptrblock {  typedef struct eptrblock {
1790    struct eptrblock *epb_prev;    struct eptrblock *epb_prev;
1791    USPTR epb_saved_eptr;    PCRE_PUCHAR epb_saved_eptr;
1792  } eptrblock;  } eptrblock;
1793    
1794    
# Line 1627  typedef struct match_data { Line 1799  typedef struct match_data {
1799    unsigned long int match_call_count;      /* As it says */    unsigned long int match_call_count;      /* As it says */
1800    unsigned long int match_limit;           /* As it says */    unsigned long int match_limit;           /* As it says */
1801    unsigned long int match_limit_recursion; /* As it says */    unsigned long int match_limit_recursion; /* As it says */
1802    int   *offset_vector;         /* Offset vector */    int   *offset_vector;           /* Offset vector */
1803    int    offset_end;            /* One past the end */    int    offset_end;              /* One past the end */
1804    int    offset_max;            /* The maximum usable for return data */    int    offset_max;              /* The maximum usable for return data */
1805    int    nltype;                /* Newline type */    int    nltype;                  /* Newline type */
1806    int    nllen;                 /* Newline string length */    int    nllen;                   /* Newline string length */
1807    int    name_count;            /* Number of names in name table */    int    name_count;              /* Number of names in name table */
1808    int    name_entry_size;       /* Size of entry in names table */    int    name_entry_size;         /* Size of entry in names table */
1809    uschar *name_table;           /* Table of names */    pcre_uchar *name_table;         /* Table of names */
1810    uschar nl[4];                 /* Newline string when fixed */    pcre_uchar nl[4];               /* Newline string when fixed */
1811    const uschar *lcc;            /* Points to lower casing table */    const  pcre_uint8 *lcc;         /* Points to lower casing table */
1812    const uschar *ctypes;         /* Points to table of type maps */    const  pcre_uint8 *ctypes;      /* Points to table of type maps */
1813    BOOL   offset_overflow;       /* Set if too many extractions */    BOOL   offset_overflow;         /* Set if too many extractions */
1814    BOOL   notbol;                /* NOTBOL flag */    BOOL   notbol;                  /* NOTBOL flag */
1815    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                  /* NOTEOL flag */
1816    BOOL   utf8;                  /* UTF8 flag */    BOOL   utf8;                    /* UTF8 flag */
1817    BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */    BOOL   jscript_compat;          /* JAVASCRIPT_COMPAT flag */
1818    BOOL   endonly;               /* Dollar not before final \n */    BOOL   use_ucp;                 /* PCRE_UCP flag */
1819    BOOL   notempty;              /* Empty string match not wanted */    BOOL   endonly;                 /* Dollar not before final \n */
1820    BOOL   notempty_atstart;      /* Empty string match at start not wanted */    BOOL   notempty;                /* Empty string match not wanted */
1821    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   notempty_atstart;        /* Empty string match at start not wanted */
1822    BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */    BOOL   hitend;                  /* Hit the end of the subject at some point */
1823    const uschar *start_code;     /* For use when recursing */    BOOL   bsr_anycrlf;             /* \R is just any CRLF, not full Unicode */
1824    USPTR  start_subject;         /* Start of the subject string */    BOOL   hasthen;                 /* Pattern contains (*THEN) */
1825    USPTR  end_subject;           /* End of the subject string */    const  pcre_uchar *start_code;  /* For use when recursing */
1826    USPTR  start_match_ptr;       /* Start of matched string */    PCRE_PUCHAR start_subject;      /* Start of the subject string */
1827    USPTR  end_match_ptr;         /* Subject position at end match */    PCRE_PUCHAR end_subject;        /* End of the subject string */
1828    USPTR  start_used_ptr;        /* Earliest consulted character */    PCRE_PUCHAR start_match_ptr;    /* Start of matched string */
1829    int    partial;               /* PARTIAL options */    PCRE_PUCHAR end_match_ptr;      /* Subject position at end match */
1830    int    end_offset_top;        /* Highwater mark at end of match */    PCRE_PUCHAR start_used_ptr;     /* Earliest consulted character */
1831    int    capture_last;          /* Most recent capture number */    int    partial;                 /* PARTIAL options */
1832    int    start_offset;          /* The start offset value */    int    end_offset_top;          /* Highwater mark at end of match */
1833    eptrblock *eptrchain;         /* Chain of eptrblocks for tail recursions */    int    capture_last;            /* Most recent capture number */
1834    int    eptrn;                 /* Next free eptrblock */    int    start_offset;            /* The start offset value */
1835    recursion_info *recursive;    /* Linked list of recursion data */    int    match_function_type;     /* Set for certain special calls of MATCH() */
1836    void  *callout_data;          /* To pass back to callouts */    eptrblock *eptrchain;           /* Chain of eptrblocks for tail recursions */
1837      int    eptrn;                   /* Next free eptrblock */
1838      recursion_info *recursive;      /* Linked list of recursion data */
1839      void  *callout_data;            /* To pass back to callouts */
1840      const  pcre_uchar *mark;        /* Mark pointer to pass back */
1841      const  pcre_uchar *once_target; /* Where to back up to for atomic groups */
1842  } match_data;  } match_data;
1843    
1844  /* 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
1845  functions. */  functions. */
1846    
1847  typedef struct dfa_match_data {  typedef struct dfa_match_data {
1848    const uschar *start_code;     /* Start of the compiled pattern */    const pcre_uchar *start_code;     /* Start of the compiled pattern */
1849    const uschar *start_subject;  /* Start of the subject string */    const pcre_uchar *start_subject ; /* Start of the subject string */
1850    const uschar *end_subject;    /* End of subject string */    const pcre_uchar *end_subject;    /* End of subject string */
1851    const uschar *start_used_ptr; /* Earliest consulted character */    const pcre_uchar *start_used_ptr; /* Earliest consulted character */
1852    const uschar *tables;         /* Character tables */    const pcre_uint8 *tables;         /* Character tables */
1853    int   start_offset;           /* The start offset value */    int   start_offset;               /* The start offset value */
1854    int   moptions;               /* Match options */    int   moptions;                   /* Match options */
1855    int   poptions;               /* Pattern options */    int   poptions;                   /* Pattern options */
1856    int    nltype;                /* Newline type */    int   nltype;                     /* Newline type */
1857    int    nllen;                 /* Newline string length */    int   nllen;                      /* Newline string length */
1858    uschar nl[4];                 /* Newline string when fixed */    pcre_uchar nl[4];                 /* Newline string when fixed */
1859    void  *callout_data;          /* To pass back to callouts */    void *callout_data;               /* To pass back to callouts */
1860      dfa_recursion_info *recursive;    /* Linked list of recursion data */
1861  } dfa_match_data;  } dfa_match_data;
1862    
1863  /* Bit definitions for entries in the pcre_ctypes table. */  /* Bit definitions for entries in the pcre_ctypes table. */
# Line 1732  of the exported public functions. They h Line 1910  of the exported public functions. They h
1910  but are not part of the PCRE public API. The data for these tables is in the  but are not part of the PCRE public API. The data for these tables is in the
1911  pcre_tables.c module. */  pcre_tables.c module. */
1912    
1913  extern const int    _pcre_utf8_table1[];  extern const int            _pcre_utf8_table1[];
1914  extern const int    _pcre_utf8_table2[];  extern const int            _pcre_utf8_table2[];
1915  extern const int    _pcre_utf8_table3[];  extern const int            _pcre_utf8_table3[];
1916  extern const uschar _pcre_utf8_table4[];  extern const pcre_uint8     _pcre_utf8_table4[];
1917    
1918  extern const int    _pcre_utf8_table1_size;  extern const int            _pcre_utf8_table1_size;
1919    
1920  extern const char   _pcre_utt_names[];  extern const char           _pcre_utt_names[];
1921  extern const ucp_type_table _pcre_utt[];  extern const ucp_type_table _pcre_utt[];
1922  extern const int _pcre_utt_size;  extern const int            _pcre_utt_size;
1923    
1924  extern const uschar _pcre_default_tables[];  extern const pcre_uint8     _pcre_default_tables[];
1925    
1926  extern const uschar _pcre_OP_lengths[];  extern const pcre_uint8     _pcre_OP_lengths[];
1927    
1928    
1929  /* Internal shared functions. These are functions that are used by more than  /* Internal shared functions. These are functions that are used by more than
1930  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
1931  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1932    
1933  extern const uschar *_pcre_find_bracket(const uschar *, BOOL, int);  extern const pcre_uchar *_pcre_find_bracket(const pcre_uchar *, BOOL, int);
1934  extern BOOL          _pcre_is_newline(USPTR, int, USPTR, int *, BOOL);  extern BOOL              _pcre_is_newline(PCRE_PUCHAR, int, PCRE_PUCHAR,
1935  extern int           _pcre_ord2utf8(int, uschar *);                             int *, BOOL);
1936  extern real_pcre    *_pcre_try_flipped(const real_pcre *, real_pcre *,  extern int               _pcre_ord2utf8(int, pcre_uint8 *);
1937                         const pcre_study_data *, pcre_study_data *);  extern real_pcre        *_pcre_try_flipped(const real_pcre *, real_pcre *,
1938  extern int           _pcre_valid_utf8(USPTR, int);                             const pcre_study_data *, pcre_study_data *);
1939  extern BOOL          _pcre_was_newline(USPTR, int, USPTR, int *, BOOL);  extern int               _pcre_valid_utf8(PCRE_PUCHAR, int, int *);
1940  extern BOOL          _pcre_xclass(int, const uschar *);  extern BOOL              _pcre_was_newline(PCRE_PUCHAR, int, PCRE_PUCHAR,
1941                               int *, BOOL);
1942    extern BOOL              _pcre_xclass(int, const pcre_uchar *);
1943    
1944    #ifdef SUPPORT_JIT
1945    extern void              _pcre_jit_compile(const real_pcre *, pcre_extra *);
1946    extern int               _pcre_jit_exec(const real_pcre *, void *, PCRE_SPTR,
1947                               int, int, int, int, int *, int);
1948    extern void              _pcre_jit_free(void *);
1949    #endif
1950    
1951  /* Unicode character database (UCD) */  /* Unicode character database (UCD) */
1952    
1953  typedef struct {  typedef struct {
1954    uschar script;    pcre_uint8 script;
1955    uschar chartype;    pcre_uint8 chartype;
1956    pcre_int32 other_case;    pcre_int32 other_case;
1957  } ucd_record;  } ucd_record;
1958    
1959  extern const ucd_record  _pcre_ucd_records[];  extern const ucd_record  _pcre_ucd_records[];
1960  extern const uschar      _pcre_ucd_stage1[];  extern const pcre_uint8  _pcre_ucd_stage1[];
1961  extern const pcre_uint16 _pcre_ucd_stage2[];  extern const pcre_uint16 _pcre_ucd_stage2[];
1962  extern const int         _pcre_ucp_gentype[];  extern const int         _pcre_ucp_gentype[];
1963    #ifdef SUPPORT_JIT
1964    extern const int         _pcre_ucp_typerange[];
1965    #endif
1966    
1967  /* UCD access macros */  /* UCD access macros */
1968    
1969  #define UCD_BLOCK_SIZE 128  #define UCD_BLOCK_SIZE 128
1970  #define GET_UCD(ch) (_pcre_ucd_records + \  #define GET_UCD(ch) (_pcre_ucd_records + \
1971          _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \          _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \
1972          UCD_BLOCK_SIZE + ch % UCD_BLOCK_SIZE])          UCD_BLOCK_SIZE + (ch) % UCD_BLOCK_SIZE])
1973    
1974  #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype  #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
1975  #define UCD_SCRIPT(ch)    GET_UCD(ch)->script  #define UCD_SCRIPT(ch)    GET_UCD(ch)->script

Legend:
Removed from v.483  
changed lines
  Added in v.756

  ViewVC Help
Powered by ViewVC 1.1.5