/[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 298 by ph10, Thu Jan 10 17:09:12 2008 UTC revision 602 by ph10, Wed May 25 08:29:03 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) 1997-2007 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
53    
54    /* We do not support both EBCDIC and UTF-8 at the same time. The "configure"
55    script prevents both being selected, but not everybody uses "configure". */
56    
57    #if defined EBCDIC && defined SUPPORT_UTF8
58    #error The use of both EBCDIC and SUPPORT_UTF8 is not supported.
59    #endif
60    
61    /* If SUPPORT_UCP is defined, SUPPORT_UTF8 must also be defined. The
62    "configure" script ensures this, but not everybody uses "configure". */
63    
64    #if defined SUPPORT_UCP && !defined SUPPORT_UTF8
65    #define SUPPORT_UTF8 1
66  #endif  #endif
67    
68  /* Use a macro for debugging printing, 'cause that eliminates the use of #ifdef  /* Use a macro for debugging printing, 'cause that eliminates the use of #ifdef
# Line 60  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 72  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 132  PCRE_EXP_DATA_DEFN only if they are not Line 144  PCRE_EXP_DATA_DEFN only if they are not
144  #  endif  #  endif
145  #endif  #endif
146    
147    /* When compiling with the MSVC compiler, it is sometimes necessary to include
148    a "calling convention" before exported function names. (This is secondhand
149    information; I know nothing about MSVC myself). For example, something like
150    
151      void __cdecl function(....)
152    
153    might be needed. In order so make this easy, all the exported functions have
154    PCRE_CALL_CONVENTION just before their names. It is rarely needed; if not
155    set, we ensure here that it has no effect. */
156    
157    #ifndef PCRE_CALL_CONVENTION
158    #define PCRE_CALL_CONVENTION
159    #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 16-bit 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
# Line 140  preprocessor time in standard C environm Line 166  preprocessor time in standard C environm
166    
167  #if USHRT_MAX == 65535  #if USHRT_MAX == 65535
168    typedef unsigned short pcre_uint16;    typedef unsigned short pcre_uint16;
169      typedef short pcre_int16;
170  #elif UINT_MAX == 65535  #elif UINT_MAX == 65535
171    typedef unsigned int pcre_uint16;    typedef unsigned int pcre_uint16;
172      typedef int pcre_int16;
173  #else  #else
174    #error Cannot determine a type for 16-bit unsigned integers    #error Cannot determine a type for 16-bit unsigned integers
175  #endif  #endif
176    
177  #if UINT_MAX == 4294967295  #if UINT_MAX == 4294967295
178    typedef unsigned int pcre_uint32;    typedef unsigned int pcre_uint32;
179      typedef int pcre_int32;
180  #elif ULONG_MAX == 4294967295  #elif ULONG_MAX == 4294967295
181    typedef unsigned long int pcre_uint32;    typedef unsigned long int pcre_uint32;
182      typedef long int pcre_int32;
183  #else  #else
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 358  capturing parenthesis numbers in back re Line 408  capturing parenthesis numbers in back re
408    
409  /* 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
410  byte. The macros for character handling generate simple sequences when used in  byte. The macros for character handling generate simple sequences when used in
411  byte-mode, and more complicated ones for UTF-8 characters. BACKCHAR should  byte-mode, and more complicated ones for UTF-8 characters. GETCHARLENTEST is
412  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
413  support is omitted, we don't even define it. */  never be called in byte mode. To make sure they can never even appear when
414    UTF-8 support is omitted, we don't even define them. */
415    
416  #ifndef SUPPORT_UTF8  #ifndef SUPPORT_UTF8
 #define NEXTCHAR(p) p++;  
417  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
418  #define GETCHARTEST(c, eptr) c = *eptr;  #define GETCHARTEST(c, eptr) c = *eptr;
419  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
420  #define GETCHARINCTEST(c, eptr) c = *eptr++;  #define GETCHARINCTEST(c, eptr) c = *eptr++;
421  #define GETCHARLEN(c, eptr, len) c = *eptr;  #define GETCHARLEN(c, eptr, len) c = *eptr;
422    /* #define GETCHARLENTEST(c, eptr, len) */
423  /* #define BACKCHAR(eptr) */  /* #define BACKCHAR(eptr) */
424    
425  #else   /* SUPPORT_UTF8 */  #else   /* SUPPORT_UTF8 */
426    
427  /* Advance a character pointer one byte in non-UTF-8 mode and by one character  /* These macros were originally written in the form of loops that used data
428  in UTF-8 mode. */  from the tables whose names start with _pcre_utf8_table. They were rewritten by
429    a user so as not to use loops, because in some environments this gives a
430    significant performance advantage, and it seems never to do any harm. */
431    
432  #define NEXTCHAR(p) \  /* Base macro to pick up the remaining bytes of a UTF-8 character, not
433    p++; \  advancing the pointer. */
434    if (utf8) { while((*p & 0xc0) == 0x80) p++; }  
435    #define GETUTF8(c, eptr) \
436        { \
437        if ((c & 0x20) == 0) \
438          c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
439        else if ((c & 0x10) == 0) \
440          c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
441        else if ((c & 0x08) == 0) \
442          c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
443          ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
444        else if ((c & 0x04) == 0) \
445          c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
446              ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
447              (eptr[4] & 0x3f); \
448        else \
449          c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
450              ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
451              ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
452        }
453    
454  /* 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
455  we know we are in UTF-8 mode. */  we know we are in UTF-8 mode. */
456    
457  #define GETCHAR(c, eptr) \  #define GETCHAR(c, eptr) \
458    c = *eptr; \    c = *eptr; \
459    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; \  
       } \  
     }  
460    
461  /* 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
462  pointer. */  pointer. */
463    
464  #define GETCHARTEST(c, eptr) \  #define GETCHARTEST(c, eptr) \
465    c = *eptr; \    c = *eptr; \
466    if (utf8 && c >= 0xc0) \    if (utf8 && c >= 0xc0) GETUTF8(c, eptr);
467    
468    /* Base macro to pick up the remaining bytes of a UTF-8 character, advancing
469    the pointer. */
470    
471    #define GETUTF8INC(c, eptr) \
472      { \      { \
473      int gcii; \      if ((c & 0x20) == 0) \
474      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \        c = ((c & 0x1f) << 6) | (*eptr++ & 0x3f); \
475      int gcss = 6*gcaa; \      else if ((c & 0x10) == 0) \
476      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \        { \
477      for (gcii = 1; gcii <= gcaa; gcii++) \        c = ((c & 0x0f) << 12) | ((*eptr & 0x3f) << 6) | (eptr[1] & 0x3f); \
478          eptr += 2; \
479          } \
480        else if ((c & 0x08) == 0) \
481          { \
482          c = ((c & 0x07) << 18) | ((*eptr & 0x3f) << 12) | \
483              ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
484          eptr += 3; \
485          } \
486        else if ((c & 0x04) == 0) \
487          { \
488          c = ((c & 0x03) << 24) | ((*eptr & 0x3f) << 18) | \
489              ((eptr[1] & 0x3f) << 12) | ((eptr[2] & 0x3f) << 6) | \
490              (eptr[3] & 0x3f); \
491          eptr += 4; \
492          } \
493        else \
494        { \        { \
495        gcss -= 6; \        c = ((c & 0x01) << 30) | ((*eptr & 0x3f) << 24) | \
496        c |= (eptr[gcii] & 0x3f) << gcss; \            ((eptr[1] & 0x3f) << 18) | ((eptr[2] & 0x3f) << 12) | \
497              ((eptr[3] & 0x3f) << 6) | (eptr[4] & 0x3f); \
498          eptr += 5; \
499        } \        } \
500      }      }
501    
# Line 421  know we are in UTF-8 mode. */ Line 504  know we are in UTF-8 mode. */
504    
505  #define GETCHARINC(c, eptr) \  #define GETCHARINC(c, eptr) \
506    c = *eptr++; \    c = *eptr++; \
507    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; \  
       } \  
     }  
508    
509  /* 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.
510    This is called when we don't know if we are in UTF-8 mode. */
511    
512  #define GETCHARINCTEST(c, eptr) \  #define GETCHARINCTEST(c, eptr) \
513    c = *eptr++; \    c = *eptr++; \
514    if (utf8 && c >= 0xc0) \    if (utf8 && c >= 0xc0) GETUTF8INC(c, eptr);
515    
516    /* Base macro to pick up the remaining bytes of a UTF-8 character, not
517    advancing the pointer, incrementing the length. */
518    
519    #define GETUTF8LEN(c, eptr, len) \
520      { \      { \
521      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      if ((c & 0x20) == 0) \
522      int gcss = 6*gcaa; \        { \
523      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \        c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
524      while (gcaa-- > 0) \        len++; \
525          } \
526        else if ((c & 0x10)  == 0) \
527        { \        { \
528        gcss -= 6; \        c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
529        c |= (*eptr++ & 0x3f) << gcss; \        len += 2; \
530          } \
531        else if ((c & 0x08)  == 0) \
532          {\
533          c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
534              ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
535          len += 3; \
536          } \
537        else if ((c & 0x04)  == 0) \
538          { \
539          c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
540              ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
541              (eptr[4] & 0x3f); \
542          len += 4; \
543          } \
544        else \
545          {\
546          c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
547              ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
548              ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
549          len += 5; \
550        } \        } \
551      }      }
552    
# Line 454  if there are extra bytes. This is called Line 555  if there are extra bytes. This is called
555    
556  #define GETCHARLEN(c, eptr, len) \  #define GETCHARLEN(c, eptr, len) \
557    c = *eptr; \    c = *eptr; \
558    if (c >= 0xc0) \    if (c >= 0xc0) GETUTF8LEN(c, eptr, len);
559      { \  
560      int gcii; \  /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the
561      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \  pointer, incrementing length if there are extra bytes. This is called when we
562      int gcss = 6*gcaa; \  do not know if we are in UTF-8 mode. */
563      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
564      for (gcii = 1; gcii <= gcaa; gcii++) \  #define GETCHARLENTEST(c, eptr, len) \
565        { \    c = *eptr; \
566        gcss -= 6; \    if (utf8 && c >= 0xc0) GETUTF8LEN(c, eptr, len);
       c |= (eptr[gcii] & 0x3f) << gcss; \  
       } \  
     len += gcaa; \  
     }  
567    
568  /* 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
569  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 474  because almost all calls are already wit Line 571  because almost all calls are already wit
571    
572  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
573    
574  #endif  #endif  /* SUPPORT_UTF8 */
575    
576    
577  /* 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 491  Standard C system should have one. */ Line 588  Standard C system should have one. */
588    
589  /* Private flags containing information about the compiled regex. They used to  /* Private flags containing information about the compiled regex. They used to
590  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
591  are in a 16-bit flags word. */  are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as
592    the restrictions on partial matching have been lifted. It remains for backwards
593    compatibility. */
594    
595  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */
596  #define PCRE_FIRSTSET      0x0002  /* first_byte is set */  #define PCRE_FIRSTSET      0x0002  /* first_byte is set */
# Line 503  are in a 16-bit flags word. */ Line 602  are in a 16-bit flags word. */
602  /* Options for the "extra" block produced by pcre_study(). */  /* Options for the "extra" block produced by pcre_study(). */
603    
604  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */
605    #define PCRE_STUDY_MINLEN   0x02     /* a minimum length field exists */
606    
607  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
608  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
# Line 510  time, run time, or study time, respectiv Line 610  time, run time, or study time, respectiv
610  #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \  #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \
611                             PCRE_NEWLINE_ANYCRLF)                             PCRE_NEWLINE_ANYCRLF)
612    
613  #define PUBLIC_OPTIONS \  #define PUBLIC_COMPILE_OPTIONS \
614    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
615     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
616     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| \
617     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
618       PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE)
619    
620  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
621    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
622     PCRE_PARTIAL|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_NEWLINE_BITS| \
623       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
624    
625  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
626    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
627     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS| \     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_DFA_SHORTEST| \
628     PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)     PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
629       PCRE_NO_START_OPTIMIZE)
630    
631  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
632    
# Line 548  variable-length repeat, or a anything ot Line 651  variable-length repeat, or a anything ot
651  #define REQ_CASELESS 0x0100    /* indicates caselessness */  #define REQ_CASELESS 0x0100    /* indicates caselessness */
652  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */
653    
654  /* Miscellaneous definitions */  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
655    environments where these macros are defined elsewhere. Unfortunately, there
656    is no way to do the same for the typedef. */
657    
658  typedef int BOOL;  typedef int BOOL;
659    
660    #ifndef FALSE
661  #define FALSE   0  #define FALSE   0
662  #define TRUE    1  #define TRUE    1
663    #endif
664    
665    /* If PCRE is to support UTF-8 on EBCDIC platforms, we cannot use normal
666    character constants like '*' because the compiler would emit their EBCDIC code,
667    which is different from their ASCII/UTF-8 code. Instead we define macros for
668    the characters so that they always use the ASCII/UTF-8 code when UTF-8 support
669    is enabled. When UTF-8 support is not enabled, the definitions use character
670    literals. Both character and string versions of each character are needed, and
671    there are some longer strings as well.
672    
673    This means that, on EBCDIC platforms, the PCRE library can handle either
674    EBCDIC, or UTF-8, but not both. To support both in the same compiled library
675    would need different lookups depending on whether PCRE_UTF8 was set or not.
676    This would make it impossible to use characters in switch/case statements,
677    which would reduce performance. For a theoretical use (which nobody has asked
678    for) in a minority area (EBCDIC platforms), this is not sensible. Any
679    application that did need both could compile two versions of the library, using
680    macros to give the functions distinct names. */
681    
682    #ifndef SUPPORT_UTF8
683    
684    /* UTF-8 support is not enabled; use the platform-dependent character literals
685    so that PCRE works on both ASCII and EBCDIC platforms, in non-UTF-mode only. */
686    
687    #define CHAR_HT                     '\t'
688    #define CHAR_VT                     '\v'
689    #define CHAR_FF                     '\f'
690    #define CHAR_CR                     '\r'
691    #define CHAR_NL                     '\n'
692    #define CHAR_BS                     '\b'
693    #define CHAR_BEL                    '\a'
694    #ifdef EBCDIC
695    #define CHAR_ESC                    '\047'
696    #define CHAR_DEL                    '\007'
697    #else
698    #define CHAR_ESC                    '\033'
699    #define CHAR_DEL                    '\177'
700    #endif
701    
702    #define CHAR_SPACE                  ' '
703    #define CHAR_EXCLAMATION_MARK       '!'
704    #define CHAR_QUOTATION_MARK         '"'
705    #define CHAR_NUMBER_SIGN            '#'
706    #define CHAR_DOLLAR_SIGN            '$'
707    #define CHAR_PERCENT_SIGN           '%'
708    #define CHAR_AMPERSAND              '&'
709    #define CHAR_APOSTROPHE             '\''
710    #define CHAR_LEFT_PARENTHESIS       '('
711    #define CHAR_RIGHT_PARENTHESIS      ')'
712    #define CHAR_ASTERISK               '*'
713    #define CHAR_PLUS                   '+'
714    #define CHAR_COMMA                  ','
715    #define CHAR_MINUS                  '-'
716    #define CHAR_DOT                    '.'
717    #define CHAR_SLASH                  '/'
718    #define CHAR_0                      '0'
719    #define CHAR_1                      '1'
720    #define CHAR_2                      '2'
721    #define CHAR_3                      '3'
722    #define CHAR_4                      '4'
723    #define CHAR_5                      '5'
724    #define CHAR_6                      '6'
725    #define CHAR_7                      '7'
726    #define CHAR_8                      '8'
727    #define CHAR_9                      '9'
728    #define CHAR_COLON                  ':'
729    #define CHAR_SEMICOLON              ';'
730    #define CHAR_LESS_THAN_SIGN         '<'
731    #define CHAR_EQUALS_SIGN            '='
732    #define CHAR_GREATER_THAN_SIGN      '>'
733    #define CHAR_QUESTION_MARK          '?'
734    #define CHAR_COMMERCIAL_AT          '@'
735    #define CHAR_A                      'A'
736    #define CHAR_B                      'B'
737    #define CHAR_C                      'C'
738    #define CHAR_D                      'D'
739    #define CHAR_E                      'E'
740    #define CHAR_F                      'F'
741    #define CHAR_G                      'G'
742    #define CHAR_H                      'H'
743    #define CHAR_I                      'I'
744    #define CHAR_J                      'J'
745    #define CHAR_K                      'K'
746    #define CHAR_L                      'L'
747    #define CHAR_M                      'M'
748    #define CHAR_N                      'N'
749    #define CHAR_O                      'O'
750    #define CHAR_P                      'P'
751    #define CHAR_Q                      'Q'
752    #define CHAR_R                      'R'
753    #define CHAR_S                      'S'
754    #define CHAR_T                      'T'
755    #define CHAR_U                      'U'
756    #define CHAR_V                      'V'
757    #define CHAR_W                      'W'
758    #define CHAR_X                      'X'
759    #define CHAR_Y                      'Y'
760    #define CHAR_Z                      'Z'
761    #define CHAR_LEFT_SQUARE_BRACKET    '['
762    #define CHAR_BACKSLASH              '\\'
763    #define CHAR_RIGHT_SQUARE_BRACKET   ']'
764    #define CHAR_CIRCUMFLEX_ACCENT      '^'
765    #define CHAR_UNDERSCORE             '_'
766    #define CHAR_GRAVE_ACCENT           '`'
767    #define CHAR_a                      'a'
768    #define CHAR_b                      'b'
769    #define CHAR_c                      'c'
770    #define CHAR_d                      'd'
771    #define CHAR_e                      'e'
772    #define CHAR_f                      'f'
773    #define CHAR_g                      'g'
774    #define CHAR_h                      'h'
775    #define CHAR_i                      'i'
776    #define CHAR_j                      'j'
777    #define CHAR_k                      'k'
778    #define CHAR_l                      'l'
779    #define CHAR_m                      'm'
780    #define CHAR_n                      'n'
781    #define CHAR_o                      'o'
782    #define CHAR_p                      'p'
783    #define CHAR_q                      'q'
784    #define CHAR_r                      'r'
785    #define CHAR_s                      's'
786    #define CHAR_t                      't'
787    #define CHAR_u                      'u'
788    #define CHAR_v                      'v'
789    #define CHAR_w                      'w'
790    #define CHAR_x                      'x'
791    #define CHAR_y                      'y'
792    #define CHAR_z                      'z'
793    #define CHAR_LEFT_CURLY_BRACKET     '{'
794    #define CHAR_VERTICAL_LINE          '|'
795    #define CHAR_RIGHT_CURLY_BRACKET    '}'
796    #define CHAR_TILDE                  '~'
797    
798    #define STR_HT                      "\t"
799    #define STR_VT                      "\v"
800    #define STR_FF                      "\f"
801    #define STR_CR                      "\r"
802    #define STR_NL                      "\n"
803    #define STR_BS                      "\b"
804    #define STR_BEL                     "\a"
805    #ifdef EBCDIC
806    #define STR_ESC                     "\047"
807    #define STR_DEL                     "\007"
808    #else
809    #define STR_ESC                     "\033"
810    #define STR_DEL                     "\177"
811    #endif
812    
813    #define STR_SPACE                   " "
814    #define STR_EXCLAMATION_MARK        "!"
815    #define STR_QUOTATION_MARK          "\""
816    #define STR_NUMBER_SIGN             "#"
817    #define STR_DOLLAR_SIGN             "$"
818    #define STR_PERCENT_SIGN            "%"
819    #define STR_AMPERSAND               "&"
820    #define STR_APOSTROPHE              "'"
821    #define STR_LEFT_PARENTHESIS        "("
822    #define STR_RIGHT_PARENTHESIS       ")"
823    #define STR_ASTERISK                "*"
824    #define STR_PLUS                    "+"
825    #define STR_COMMA                   ","
826    #define STR_MINUS                   "-"
827    #define STR_DOT                     "."
828    #define STR_SLASH                   "/"
829    #define STR_0                       "0"
830    #define STR_1                       "1"
831    #define STR_2                       "2"
832    #define STR_3                       "3"
833    #define STR_4                       "4"
834    #define STR_5                       "5"
835    #define STR_6                       "6"
836    #define STR_7                       "7"
837    #define STR_8                       "8"
838    #define STR_9                       "9"
839    #define STR_COLON                   ":"
840    #define STR_SEMICOLON               ";"
841    #define STR_LESS_THAN_SIGN          "<"
842    #define STR_EQUALS_SIGN             "="
843    #define STR_GREATER_THAN_SIGN       ">"
844    #define STR_QUESTION_MARK           "?"
845    #define STR_COMMERCIAL_AT           "@"
846    #define STR_A                       "A"
847    #define STR_B                       "B"
848    #define STR_C                       "C"
849    #define STR_D                       "D"
850    #define STR_E                       "E"
851    #define STR_F                       "F"
852    #define STR_G                       "G"
853    #define STR_H                       "H"
854    #define STR_I                       "I"
855    #define STR_J                       "J"
856    #define STR_K                       "K"
857    #define STR_L                       "L"
858    #define STR_M                       "M"
859    #define STR_N                       "N"
860    #define STR_O                       "O"
861    #define STR_P                       "P"
862    #define STR_Q                       "Q"
863    #define STR_R                       "R"
864    #define STR_S                       "S"
865    #define STR_T                       "T"
866    #define STR_U                       "U"
867    #define STR_V                       "V"
868    #define STR_W                       "W"
869    #define STR_X                       "X"
870    #define STR_Y                       "Y"
871    #define STR_Z                       "Z"
872    #define STR_LEFT_SQUARE_BRACKET     "["
873    #define STR_BACKSLASH               "\\"
874    #define STR_RIGHT_SQUARE_BRACKET    "]"
875    #define STR_CIRCUMFLEX_ACCENT       "^"
876    #define STR_UNDERSCORE              "_"
877    #define STR_GRAVE_ACCENT            "`"
878    #define STR_a                       "a"
879    #define STR_b                       "b"
880    #define STR_c                       "c"
881    #define STR_d                       "d"
882    #define STR_e                       "e"
883    #define STR_f                       "f"
884    #define STR_g                       "g"
885    #define STR_h                       "h"
886    #define STR_i                       "i"
887    #define STR_j                       "j"
888    #define STR_k                       "k"
889    #define STR_l                       "l"
890    #define STR_m                       "m"
891    #define STR_n                       "n"
892    #define STR_o                       "o"
893    #define STR_p                       "p"
894    #define STR_q                       "q"
895    #define STR_r                       "r"
896    #define STR_s                       "s"
897    #define STR_t                       "t"
898    #define STR_u                       "u"
899    #define STR_v                       "v"
900    #define STR_w                       "w"
901    #define STR_x                       "x"
902    #define STR_y                       "y"
903    #define STR_z                       "z"
904    #define STR_LEFT_CURLY_BRACKET      "{"
905    #define STR_VERTICAL_LINE           "|"
906    #define STR_RIGHT_CURLY_BRACKET     "}"
907    #define STR_TILDE                   "~"
908    
909    #define STRING_ACCEPT0              "ACCEPT\0"
910    #define STRING_COMMIT0              "COMMIT\0"
911    #define STRING_F0                   "F\0"
912    #define STRING_FAIL0                "FAIL\0"
913    #define STRING_MARK0                "MARK\0"
914    #define STRING_PRUNE0               "PRUNE\0"
915    #define STRING_SKIP0                "SKIP\0"
916    #define STRING_THEN                 "THEN"
917    
918    #define STRING_alpha0               "alpha\0"
919    #define STRING_lower0               "lower\0"
920    #define STRING_upper0               "upper\0"
921    #define STRING_alnum0               "alnum\0"
922    #define STRING_ascii0               "ascii\0"
923    #define STRING_blank0               "blank\0"
924    #define STRING_cntrl0               "cntrl\0"
925    #define STRING_digit0               "digit\0"
926    #define STRING_graph0               "graph\0"
927    #define STRING_print0               "print\0"
928    #define STRING_punct0               "punct\0"
929    #define STRING_space0               "space\0"
930    #define STRING_word0                "word\0"
931    #define STRING_xdigit               "xdigit"
932    
933    #define STRING_DEFINE               "DEFINE"
934    
935    #define STRING_CR_RIGHTPAR             "CR)"
936    #define STRING_LF_RIGHTPAR             "LF)"
937    #define STRING_CRLF_RIGHTPAR           "CRLF)"
938    #define STRING_ANY_RIGHTPAR            "ANY)"
939    #define STRING_ANYCRLF_RIGHTPAR        "ANYCRLF)"
940    #define STRING_BSR_ANYCRLF_RIGHTPAR    "BSR_ANYCRLF)"
941    #define STRING_BSR_UNICODE_RIGHTPAR    "BSR_UNICODE)"
942    #define STRING_UTF8_RIGHTPAR           "UTF8)"
943    #define STRING_UCP_RIGHTPAR            "UCP)"
944    #define STRING_NO_START_OPT_RIGHTPAR   "NO_START_OPT)"
945    
946    #else  /* SUPPORT_UTF8 */
947    
948    /* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This
949    works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode
950    only. */
951    
952    #define CHAR_HT                     '\011'
953    #define CHAR_VT                     '\013'
954    #define CHAR_FF                     '\014'
955    #define CHAR_CR                     '\015'
956    #define CHAR_NL                     '\012'
957    #define CHAR_BS                     '\010'
958    #define CHAR_BEL                    '\007'
959    #define CHAR_ESC                    '\033'
960    #define CHAR_DEL                    '\177'
961    
962    #define CHAR_SPACE                  '\040'
963    #define CHAR_EXCLAMATION_MARK       '\041'
964    #define CHAR_QUOTATION_MARK         '\042'
965    #define CHAR_NUMBER_SIGN            '\043'
966    #define CHAR_DOLLAR_SIGN            '\044'
967    #define CHAR_PERCENT_SIGN           '\045'
968    #define CHAR_AMPERSAND              '\046'
969    #define CHAR_APOSTROPHE             '\047'
970    #define CHAR_LEFT_PARENTHESIS       '\050'
971    #define CHAR_RIGHT_PARENTHESIS      '\051'
972    #define CHAR_ASTERISK               '\052'
973    #define CHAR_PLUS                   '\053'
974    #define CHAR_COMMA                  '\054'
975    #define CHAR_MINUS                  '\055'
976    #define CHAR_DOT                    '\056'
977    #define CHAR_SLASH                  '\057'
978    #define CHAR_0                      '\060'
979    #define CHAR_1                      '\061'
980    #define CHAR_2                      '\062'
981    #define CHAR_3                      '\063'
982    #define CHAR_4                      '\064'
983    #define CHAR_5                      '\065'
984    #define CHAR_6                      '\066'
985    #define CHAR_7                      '\067'
986    #define CHAR_8                      '\070'
987    #define CHAR_9                      '\071'
988    #define CHAR_COLON                  '\072'
989    #define CHAR_SEMICOLON              '\073'
990    #define CHAR_LESS_THAN_SIGN         '\074'
991    #define CHAR_EQUALS_SIGN            '\075'
992    #define CHAR_GREATER_THAN_SIGN      '\076'
993    #define CHAR_QUESTION_MARK          '\077'
994    #define CHAR_COMMERCIAL_AT          '\100'
995    #define CHAR_A                      '\101'
996    #define CHAR_B                      '\102'
997    #define CHAR_C                      '\103'
998    #define CHAR_D                      '\104'
999    #define CHAR_E                      '\105'
1000    #define CHAR_F                      '\106'
1001    #define CHAR_G                      '\107'
1002    #define CHAR_H                      '\110'
1003    #define CHAR_I                      '\111'
1004    #define CHAR_J                      '\112'
1005    #define CHAR_K                      '\113'
1006    #define CHAR_L                      '\114'
1007    #define CHAR_M                      '\115'
1008    #define CHAR_N                      '\116'
1009    #define CHAR_O                      '\117'
1010    #define CHAR_P                      '\120'
1011    #define CHAR_Q                      '\121'
1012    #define CHAR_R                      '\122'
1013    #define CHAR_S                      '\123'
1014    #define CHAR_T                      '\124'
1015    #define CHAR_U                      '\125'
1016    #define CHAR_V                      '\126'
1017    #define CHAR_W                      '\127'
1018    #define CHAR_X                      '\130'
1019    #define CHAR_Y                      '\131'
1020    #define CHAR_Z                      '\132'
1021    #define CHAR_LEFT_SQUARE_BRACKET    '\133'
1022    #define CHAR_BACKSLASH              '\134'
1023    #define CHAR_RIGHT_SQUARE_BRACKET   '\135'
1024    #define CHAR_CIRCUMFLEX_ACCENT      '\136'
1025    #define CHAR_UNDERSCORE             '\137'
1026    #define CHAR_GRAVE_ACCENT           '\140'
1027    #define CHAR_a                      '\141'
1028    #define CHAR_b                      '\142'
1029    #define CHAR_c                      '\143'
1030    #define CHAR_d                      '\144'
1031    #define CHAR_e                      '\145'
1032    #define CHAR_f                      '\146'
1033    #define CHAR_g                      '\147'
1034    #define CHAR_h                      '\150'
1035    #define CHAR_i                      '\151'
1036    #define CHAR_j                      '\152'
1037    #define CHAR_k                      '\153'
1038    #define CHAR_l                      '\154'
1039    #define CHAR_m                      '\155'
1040    #define CHAR_n                      '\156'
1041    #define CHAR_o                      '\157'
1042    #define CHAR_p                      '\160'
1043    #define CHAR_q                      '\161'
1044    #define CHAR_r                      '\162'
1045    #define CHAR_s                      '\163'
1046    #define CHAR_t                      '\164'
1047    #define CHAR_u                      '\165'
1048    #define CHAR_v                      '\166'
1049    #define CHAR_w                      '\167'
1050    #define CHAR_x                      '\170'
1051    #define CHAR_y                      '\171'
1052    #define CHAR_z                      '\172'
1053    #define CHAR_LEFT_CURLY_BRACKET     '\173'
1054    #define CHAR_VERTICAL_LINE          '\174'
1055    #define CHAR_RIGHT_CURLY_BRACKET    '\175'
1056    #define CHAR_TILDE                  '\176'
1057    
1058    #define STR_HT                      "\011"
1059    #define STR_VT                      "\013"
1060    #define STR_FF                      "\014"
1061    #define STR_CR                      "\015"
1062    #define STR_NL                      "\012"
1063    #define STR_BS                      "\010"
1064    #define STR_BEL                     "\007"
1065    #define STR_ESC                     "\033"
1066    #define STR_DEL                     "\177"
1067    
1068    #define STR_SPACE                   "\040"
1069    #define STR_EXCLAMATION_MARK        "\041"
1070    #define STR_QUOTATION_MARK          "\042"
1071    #define STR_NUMBER_SIGN             "\043"
1072    #define STR_DOLLAR_SIGN             "\044"
1073    #define STR_PERCENT_SIGN            "\045"
1074    #define STR_AMPERSAND               "\046"
1075    #define STR_APOSTROPHE              "\047"
1076    #define STR_LEFT_PARENTHESIS        "\050"
1077    #define STR_RIGHT_PARENTHESIS       "\051"
1078    #define STR_ASTERISK                "\052"
1079    #define STR_PLUS                    "\053"
1080    #define STR_COMMA                   "\054"
1081    #define STR_MINUS                   "\055"
1082    #define STR_DOT                     "\056"
1083    #define STR_SLASH                   "\057"
1084    #define STR_0                       "\060"
1085    #define STR_1                       "\061"
1086    #define STR_2                       "\062"
1087    #define STR_3                       "\063"
1088    #define STR_4                       "\064"
1089    #define STR_5                       "\065"
1090    #define STR_6                       "\066"
1091    #define STR_7                       "\067"
1092    #define STR_8                       "\070"
1093    #define STR_9                       "\071"
1094    #define STR_COLON                   "\072"
1095    #define STR_SEMICOLON               "\073"
1096    #define STR_LESS_THAN_SIGN          "\074"
1097    #define STR_EQUALS_SIGN             "\075"
1098    #define STR_GREATER_THAN_SIGN       "\076"
1099    #define STR_QUESTION_MARK           "\077"
1100    #define STR_COMMERCIAL_AT           "\100"
1101    #define STR_A                       "\101"
1102    #define STR_B                       "\102"
1103    #define STR_C                       "\103"
1104    #define STR_D                       "\104"
1105    #define STR_E                       "\105"
1106    #define STR_F                       "\106"
1107    #define STR_G                       "\107"
1108    #define STR_H                       "\110"
1109    #define STR_I                       "\111"
1110    #define STR_J                       "\112"
1111    #define STR_K                       "\113"
1112    #define STR_L                       "\114"
1113    #define STR_M                       "\115"
1114    #define STR_N                       "\116"
1115    #define STR_O                       "\117"
1116    #define STR_P                       "\120"
1117    #define STR_Q                       "\121"
1118    #define STR_R                       "\122"
1119    #define STR_S                       "\123"
1120    #define STR_T                       "\124"
1121    #define STR_U                       "\125"
1122    #define STR_V                       "\126"
1123    #define STR_W                       "\127"
1124    #define STR_X                       "\130"
1125    #define STR_Y                       "\131"
1126    #define STR_Z                       "\132"
1127    #define STR_LEFT_SQUARE_BRACKET     "\133"
1128    #define STR_BACKSLASH               "\134"
1129    #define STR_RIGHT_SQUARE_BRACKET    "\135"
1130    #define STR_CIRCUMFLEX_ACCENT       "\136"
1131    #define STR_UNDERSCORE              "\137"
1132    #define STR_GRAVE_ACCENT            "\140"
1133    #define STR_a                       "\141"
1134    #define STR_b                       "\142"
1135    #define STR_c                       "\143"
1136    #define STR_d                       "\144"
1137    #define STR_e                       "\145"
1138    #define STR_f                       "\146"
1139    #define STR_g                       "\147"
1140    #define STR_h                       "\150"
1141    #define STR_i                       "\151"
1142    #define STR_j                       "\152"
1143    #define STR_k                       "\153"
1144    #define STR_l                       "\154"
1145    #define STR_m                       "\155"
1146    #define STR_n                       "\156"
1147    #define STR_o                       "\157"
1148    #define STR_p                       "\160"
1149    #define STR_q                       "\161"
1150    #define STR_r                       "\162"
1151    #define STR_s                       "\163"
1152    #define STR_t                       "\164"
1153    #define STR_u                       "\165"
1154    #define STR_v                       "\166"
1155    #define STR_w                       "\167"
1156    #define STR_x                       "\170"
1157    #define STR_y                       "\171"
1158    #define STR_z                       "\172"
1159    #define STR_LEFT_CURLY_BRACKET      "\173"
1160    #define STR_VERTICAL_LINE           "\174"
1161    #define STR_RIGHT_CURLY_BRACKET     "\175"
1162    #define STR_TILDE                   "\176"
1163    
1164    #define STRING_ACCEPT0              STR_A STR_C STR_C STR_E STR_P STR_T "\0"
1165    #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"
1166    #define STRING_F0                   STR_F "\0"
1167    #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"
1168    #define STRING_MARK0                STR_M STR_A STR_R STR_K "\0"
1169    #define STRING_PRUNE0               STR_P STR_R STR_U STR_N STR_E "\0"
1170    #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"
1171    #define STRING_THEN                 STR_T STR_H STR_E STR_N
1172    
1173    #define STRING_alpha0               STR_a STR_l STR_p STR_h STR_a "\0"
1174    #define STRING_lower0               STR_l STR_o STR_w STR_e STR_r "\0"
1175    #define STRING_upper0               STR_u STR_p STR_p STR_e STR_r "\0"
1176    #define STRING_alnum0               STR_a STR_l STR_n STR_u STR_m "\0"
1177    #define STRING_ascii0               STR_a STR_s STR_c STR_i STR_i "\0"
1178    #define STRING_blank0               STR_b STR_l STR_a STR_n STR_k "\0"
1179    #define STRING_cntrl0               STR_c STR_n STR_t STR_r STR_l "\0"
1180    #define STRING_digit0               STR_d STR_i STR_g STR_i STR_t "\0"
1181    #define STRING_graph0               STR_g STR_r STR_a STR_p STR_h "\0"
1182    #define STRING_print0               STR_p STR_r STR_i STR_n STR_t "\0"
1183    #define STRING_punct0               STR_p STR_u STR_n STR_c STR_t "\0"
1184    #define STRING_space0               STR_s STR_p STR_a STR_c STR_e "\0"
1185    #define STRING_word0                STR_w STR_o STR_r STR_d       "\0"
1186    #define STRING_xdigit               STR_x STR_d STR_i STR_g STR_i STR_t
1187    
1188    #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E
1189    
1190    #define STRING_CR_RIGHTPAR             STR_C STR_R STR_RIGHT_PARENTHESIS
1191    #define STRING_LF_RIGHTPAR             STR_L STR_F STR_RIGHT_PARENTHESIS
1192    #define STRING_CRLF_RIGHTPAR           STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1193    #define STRING_ANY_RIGHTPAR            STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1194    #define STRING_ANYCRLF_RIGHTPAR        STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1195    #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
1196    #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
1197    #define STRING_UTF8_RIGHTPAR           STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1198    #define STRING_UCP_RIGHTPAR            STR_U STR_C STR_P STR_RIGHT_PARENTHESIS
1199    #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
1200    
1201    #endif  /* SUPPORT_UTF8 */
1202    
1203  /* Escape items that are just an encoding of a particular data value. */  /* Escape items that are just an encoding of a particular data value. */
1204    
1205  #ifndef ESC_e  #ifndef ESC_e
1206  #define ESC_e 27  #define ESC_e CHAR_ESC
1207  #endif  #endif
1208    
1209  #ifndef ESC_f  #ifndef ESC_f
1210  #define ESC_f '\f'  #define ESC_f CHAR_FF
1211  #endif  #endif
1212    
1213  #ifndef ESC_n  #ifndef ESC_n
1214  #define ESC_n '\n'  #define ESC_n CHAR_NL
1215  #endif  #endif
1216    
1217  #ifndef ESC_r  #ifndef ESC_r
1218  #define ESC_r '\r'  #define ESC_r CHAR_CR
1219  #endif  #endif
1220    
1221  /* We can't officially use ESC_t because it is a POSIX reserved identifier  /* We can't officially use ESC_t because it is a POSIX reserved identifier
1222  (presumably because of all the others like size_t). */  (presumably because of all the others like size_t). */
1223    
1224  #ifndef ESC_tee  #ifndef ESC_tee
1225  #define ESC_tee '\t'  #define ESC_tee CHAR_HT
1226  #endif  #endif
1227    
1228  /* Codes for different types of Unicode property */  /* Codes for different types of Unicode property */
1229    
1230  #define PT_ANY        0    /* Any property - matches all chars */  #define PT_ANY        0    /* Any property - matches all chars */
1231  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */
1232  #define PT_GC         2    /* General characteristic (e.g. L) */  #define PT_GC         2    /* Specified general characteristic (e.g. L) */
1233  #define PT_PC         3    /* Particular characteristic (e.g. Lu) */  #define PT_PC         3    /* Specified particular characteristic (e.g. Lu) */
1234  #define PT_SC         4    /* Script (e.g. Han) */  #define PT_SC         4    /* Script (e.g. Han) */
1235    #define PT_ALNUM      5    /* Alphanumeric - the union of L and N */
1236    #define PT_SPACE      6    /* Perl space - Z plus 9,10,12,13 */
1237    #define PT_PXSPACE    7    /* POSIX space - Z plus 9,10,11,12,13 */
1238    #define PT_WORD       8    /* Word - L plus N plus underscore */
1239    
1240  /* 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
1241  contain UTF-8 characters with values greater than 255. */  contain UTF-8 characters with values greater than 255. */
# Line 603  contain UTF-8 characters with values gre Line 1252  contain UTF-8 characters with values gre
1252  /* 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
1253  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
1254  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
1255  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
1256  corresponds to "." rather than an escape sequence. The final one must be  corresponds to "." in DOTALL mode rather than an escape sequence. It is also
1257  ESC_REF as subsequent values are used for backreferences (\1, \2, \3, etc).  used for [^] in JavaScript compatibility mode. In non-DOTALL mode, "." behaves
1258  There are two tests in the code for an escape greater than ESC_b and less than  like \N.
1259  ESC_Z to detect the types that may be repeated. These are the types that  
1260  consume characters. If any new escapes are put in between that don't consume a  The special values ESC_DU, ESC_du, etc. are used instead of ESC_D, ESC_d, etc.
1261  character, that code will have to change. */  when PCRE_UCP is set, when replacement of \d etc by \p sequences is required.
1262    They must be contiguous, and remain in order so that the replacements can be
1263    looked up from a table.
1264    
1265    The final escape must be ESC_REF as subsequent values are used for
1266    backreferences (\1, \2, \3, etc). There are two tests in the code for an escape
1267    greater than ESC_b and less than ESC_Z to detect the types that may be
1268    repeated. These are the types that consume characters. If any new escapes are
1269    put in between that don't consume a character, that code will have to change.
1270    */
1271    
1272  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,
1273         ESC_W, ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H, ESC_h,         ESC_W, ESC_w, ESC_N, ESC_dum, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
1274         ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_k, ESC_REF };         ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z,
1275           ESC_E, ESC_Q, ESC_g, ESC_k,
1276           ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu,
1277           ESC_REF };
1278    
1279  /* 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
1280  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.
1281    
1282  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
1283  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
1284  "coptable" in pcre_dfa_exec.c that must be updated. */  "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1285    
1286  enum {  enum {
1287    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 639  enum { Line 1299  enum {
1299    OP_WHITESPACE,         /*  9 \s */    OP_WHITESPACE,         /*  9 \s */
1300    OP_NOT_WORDCHAR,       /* 10 \W */    OP_NOT_WORDCHAR,       /* 10 \W */
1301    OP_WORDCHAR,           /* 11 \w */    OP_WORDCHAR,           /* 11 \w */
1302    OP_ANY,            /* 12 Match any character */  
1303    OP_ANYBYTE,        /* 13 Match any byte (\C); different to OP_ANY for UTF-8 */    OP_ANY,            /* 12 Match any character except newline */
1304    OP_NOTPROP,        /* 14 \P (not Unicode property) */    OP_ALLANY,         /* 13 Match any character */
1305    OP_PROP,           /* 15 \p (Unicode property) */    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
1306    OP_ANYNL,          /* 16 \R (any newline sequence) */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
1307    OP_NOT_HSPACE,     /* 17 \H (not horizontal whitespace) */    OP_PROP,           /* 16 \p (Unicode property) */
1308    OP_HSPACE,         /* 18 \h (horizontal whitespace) */    OP_ANYNL,          /* 17 \R (any newline sequence) */
1309    OP_NOT_VSPACE,     /* 19 \V (not vertical whitespace) */    OP_NOT_HSPACE,     /* 18 \H (not horizontal whitespace) */
1310    OP_VSPACE,         /* 20 \v (vertical whitespace) */    OP_HSPACE,         /* 19 \h (horizontal whitespace) */
1311    OP_EXTUNI,         /* 21 \X (extended Unicode sequence */    OP_NOT_VSPACE,     /* 20 \V (not vertical whitespace) */
1312    OP_EODN,           /* 22 End of data or \n at end of data: \Z. */    OP_VSPACE,         /* 21 \v (vertical whitespace) */
1313    OP_EOD,            /* 23 End of data: \z */    OP_EXTUNI,         /* 22 \X (extended Unicode sequence */
1314      OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
1315    OP_OPT,            /* 24 Set runtime options */    OP_EOD,            /* 24 End of data: \z */
1316    OP_CIRC,           /* 25 Start of line - varies with multiline switch */  
1317    OP_DOLL,           /* 26 End of line - varies with multiline switch */    OP_CIRC,           /* 25 Start of line - not multiline */
1318    OP_CHAR,           /* 27 Match one character, casefully */    OP_CIRCM,          /* 26 Start of line - multiline */
1319    OP_CHARNC,         /* 28 Match one character, caselessly */    OP_DOLL,           /* 27 End of line - not multiline */
1320    OP_NOT,            /* 29 Match one character, not the following one */    OP_DOLLM,          /* 28 End of line - multiline */
1321      OP_CHAR,           /* 29 Match one character, casefully */
1322    OP_STAR,           /* 30 The maximizing and minimizing versions of */    OP_CHARI,          /* 30 Match one character, caselessly */
1323    OP_MINSTAR,        /* 31 these six opcodes must come in pairs, with */    OP_NOT,            /* 31 Match one character, not the given one, casefully */
1324    OP_PLUS,           /* 32 the minimizing one second. */    OP_NOTI,           /* 32 Match one character, not the given one, caselessly */
1325    OP_MINPLUS,        /* 33 This first set applies to single characters.*/  
1326    OP_QUERY,          /* 34 */    /* The following sets of 13 opcodes must always be kept in step because
1327    OP_MINQUERY,       /* 35 */    the offset from the first one is used to generate the others. */
1328    
1329    OP_UPTO,           /* 36 From 0 to n matches */    /**** Single characters, caseful, must precede the caseless ones ****/
1330    OP_MINUPTO,        /* 37 */  
1331    OP_EXACT,          /* 38 Exactly n matches */    OP_STAR,           /* 33 The maximizing and minimizing versions of */
1332      OP_MINSTAR,        /* 34 these six opcodes must come in pairs, with */
1333    OP_POSSTAR,        /* 39 Possessified star */    OP_PLUS,           /* 35 the minimizing one second. */
1334    OP_POSPLUS,        /* 40 Possessified plus */    OP_MINPLUS,        /* 36 */
1335    OP_POSQUERY,       /* 41 Posesssified query */    OP_QUERY,          /* 37 */
1336    OP_POSUPTO,        /* 42 Possessified upto */    OP_MINQUERY,       /* 38 */
1337    
1338    OP_NOTSTAR,        /* 43 The maximizing and minimizing versions of */    OP_UPTO,           /* 39 From 0 to n matches of one character, caseful*/
1339    OP_NOTMINSTAR,     /* 44 these six opcodes must come in pairs, with */    OP_MINUPTO,        /* 40 */
1340    OP_NOTPLUS,        /* 45 the minimizing one second. They must be in */    OP_EXACT,          /* 41 Exactly n matches */
1341    OP_NOTMINPLUS,     /* 46 exactly the same order as those above. */  
1342    OP_NOTQUERY,       /* 47 This set applies to "not" single characters. */    OP_POSSTAR,        /* 42 Possessified star, caseful */
1343    OP_NOTMINQUERY,    /* 48 */    OP_POSPLUS,        /* 43 Possessified plus, caseful */
1344      OP_POSQUERY,       /* 44 Posesssified query, caseful */
1345    OP_NOTUPTO,        /* 49 From 0 to n matches */    OP_POSUPTO,        /* 45 Possessified upto, caseful */
1346    OP_NOTMINUPTO,     /* 50 */  
1347    OP_NOTEXACT,       /* 51 Exactly n matches */    /**** Single characters, caseless, must follow the caseful ones */
1348    
1349    OP_NOTPOSSTAR,     /* 52 Possessified versions */    OP_STARI,          /* 46 */
1350    OP_NOTPOSPLUS,     /* 53 */    OP_MINSTARI,       /* 47 */
1351    OP_NOTPOSQUERY,    /* 54 */    OP_PLUSI,          /* 48 */
1352    OP_NOTPOSUPTO,     /* 55 */    OP_MINPLUSI,       /* 49 */
1353      OP_QUERYI,         /* 50 */
1354    OP_TYPESTAR,       /* 56 The maximizing and minimizing versions of */    OP_MINQUERYI,      /* 51 */
1355    OP_TYPEMINSTAR,    /* 57 these six opcodes must come in pairs, with */  
1356    OP_TYPEPLUS,       /* 58 the minimizing one second. These codes must */    OP_UPTOI,          /* 52 From 0 to n matches of one character, caseless */
1357    OP_TYPEMINPLUS,    /* 59 be in exactly the same order as those above. */    OP_MINUPTOI,       /* 53 */
1358    OP_TYPEQUERY,      /* 60 This set applies to character types such as \d */    OP_EXACTI,         /* 54 */
1359    OP_TYPEMINQUERY,   /* 61 */  
1360      OP_POSSTARI,       /* 55 Possessified star, caseless */
1361    OP_TYPEUPTO,       /* 62 From 0 to n matches */    OP_POSPLUSI,       /* 56 Possessified plus, caseless */
1362    OP_TYPEMINUPTO,    /* 63 */    OP_POSQUERYI,      /* 57 Posesssified query, caseless */
1363    OP_TYPEEXACT,      /* 64 Exactly n matches */    OP_POSUPTOI,       /* 58 Possessified upto, caseless */
1364    
1365    OP_TYPEPOSSTAR,    /* 65 Possessified versions */    /**** The negated ones must follow the non-negated ones, and match them ****/
1366    OP_TYPEPOSPLUS,    /* 66 */    /**** Negated single character, caseful; must precede the caseless ones ****/
1367    OP_TYPEPOSQUERY,   /* 67 */  
1368    OP_TYPEPOSUPTO,    /* 68 */    OP_NOTSTAR,        /* 59 The maximizing and minimizing versions of */
1369      OP_NOTMINSTAR,     /* 60 these six opcodes must come in pairs, with */
1370    OP_CRSTAR,         /* 69 The maximizing and minimizing versions of */    OP_NOTPLUS,        /* 61 the minimizing one second. They must be in */
1371    OP_CRMINSTAR,      /* 70 all these opcodes must come in pairs, with */    OP_NOTMINPLUS,     /* 62 exactly the same order as those above. */
1372    OP_CRPLUS,         /* 71 the minimizing one second. These codes must */    OP_NOTQUERY,       /* 63 */
1373    OP_CRMINPLUS,      /* 72 be in exactly the same order as those above. */    OP_NOTMINQUERY,    /* 64 */
1374    OP_CRQUERY,        /* 73 These are for character classes and back refs */  
1375    OP_CRMINQUERY,     /* 74 */    OP_NOTUPTO,        /* 65 From 0 to n matches, caseful */
1376    OP_CRRANGE,        /* 75 These are different to the three sets above. */    OP_NOTMINUPTO,     /* 66 */
1377    OP_CRMINRANGE,     /* 76 */    OP_NOTEXACT,       /* 67 Exactly n matches */
1378    
1379    OP_CLASS,          /* 77 Match a character class, chars < 256 only */    OP_NOTPOSSTAR,     /* 68 Possessified versions, caseful */
1380    OP_NCLASS,         /* 78 Same, but the bitmap was created from a negative    OP_NOTPOSPLUS,     /* 69 */
1381                             class - the difference is relevant only when a UTF-8    OP_NOTPOSQUERY,    /* 70 */
1382                             character > 255 is encountered. */    OP_NOTPOSUPTO,     /* 71 */
1383    
1384    OP_XCLASS,         /* 79 Extended class for handling UTF-8 chars within the    /**** Negated single character, caseless; must follow the caseful ones ****/
1385                             class. This does both positive and negative. */  
1386      OP_NOTSTARI,       /* 72 */
1387    OP_REF,            /* 80 Match a back reference */    OP_NOTMINSTARI,    /* 73 */
1388    OP_RECURSE,        /* 81 Match a numbered subpattern (possibly recursive) */    OP_NOTPLUSI,       /* 74 */
1389    OP_CALLOUT,        /* 82 Call out to external function if provided */    OP_NOTMINPLUSI,    /* 75 */
1390      OP_NOTQUERYI,      /* 76 */
1391    OP_ALT,            /* 83 Start of alternation */    OP_NOTMINQUERYI,   /* 77 */
1392    OP_KET,            /* 84 End of group that doesn't have an unbounded repeat */  
1393    OP_KETRMAX,        /* 85 These two must remain together and in this */    OP_NOTUPTOI,       /* 78 From 0 to n matches, caseless */
1394    OP_KETRMIN,        /* 86 order. They are for groups the repeat for ever. */    OP_NOTMINUPTOI,    /* 79 */
1395      OP_NOTEXACTI,      /* 80 Exactly n matches */
1396    
1397      OP_NOTPOSSTARI,    /* 81 Possessified versions, caseless */
1398      OP_NOTPOSPLUSI,    /* 82 */
1399      OP_NOTPOSQUERYI,   /* 83 */
1400      OP_NOTPOSUPTOI,    /* 84 */
1401    
1402      /**** Character types ****/
1403    
1404      OP_TYPESTAR,       /* 85 The maximizing and minimizing versions of */
1405      OP_TYPEMINSTAR,    /* 86 these six opcodes must come in pairs, with */
1406      OP_TYPEPLUS,       /* 87 the minimizing one second. These codes must */
1407      OP_TYPEMINPLUS,    /* 88 be in exactly the same order as those above. */
1408      OP_TYPEQUERY,      /* 89 */
1409      OP_TYPEMINQUERY,   /* 90 */
1410    
1411      OP_TYPEUPTO,       /* 91 From 0 to n matches */
1412      OP_TYPEMINUPTO,    /* 92 */
1413      OP_TYPEEXACT,      /* 93 Exactly n matches */
1414    
1415      OP_TYPEPOSSTAR,    /* 94 Possessified versions */
1416      OP_TYPEPOSPLUS,    /* 95 */
1417      OP_TYPEPOSQUERY,   /* 96 */
1418      OP_TYPEPOSUPTO,    /* 97 */
1419    
1420      /* These are used for character classes and back references; only the
1421      first six are the same as the sets above. */
1422    
1423      OP_CRSTAR,         /* 98 The maximizing and minimizing versions of */
1424      OP_CRMINSTAR,      /* 99 all these opcodes must come in pairs, with */
1425      OP_CRPLUS,         /* 100 the minimizing one second. These codes must */
1426      OP_CRMINPLUS,      /* 101 be in exactly the same order as those above. */
1427      OP_CRQUERY,        /* 102 */
1428      OP_CRMINQUERY,     /* 103 */
1429    
1430      OP_CRRANGE,        /* 104 These are different to the three sets above. */
1431      OP_CRMINRANGE,     /* 105 */
1432    
1433      /* End of quantifier opcodes */
1434    
1435      OP_CLASS,          /* 106 Match a character class, chars < 256 only */
1436      OP_NCLASS,         /* 107 Same, but the bitmap was created from a negative
1437                                  class - the difference is relevant only when a
1438                                  UTF-8 character > 255 is encountered. */
1439      OP_XCLASS,         /* 108 Extended class for handling UTF-8 chars within the
1440                                  class. This does both positive and negative. */
1441      OP_REF,            /* 109 Match a back reference, casefully */
1442      OP_REFI,           /* 110 Match a back reference, caselessly */
1443      OP_RECURSE,        /* 111 Match a numbered subpattern (possibly recursive) */
1444      OP_CALLOUT,        /* 112 Call out to external function if provided */
1445    
1446      OP_ALT,            /* 113 Start of alternation */
1447      OP_KET,            /* 114 End of group that doesn't have an unbounded repeat */
1448      OP_KETRMAX,        /* 115 These two must remain together and in this */
1449      OP_KETRMIN,        /* 116 order. They are for groups the repeat for ever. */
1450    
1451    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
1452    
1453    OP_ASSERT,         /* 87 Positive lookahead */    OP_ASSERT,         /* 117 Positive lookahead */
1454    OP_ASSERT_NOT,     /* 88 Negative lookahead */    OP_ASSERT_NOT,     /* 118 Negative lookahead */
1455    OP_ASSERTBACK,     /* 89 Positive lookbehind */    OP_ASSERTBACK,     /* 119 Positive lookbehind */
1456    OP_ASSERTBACK_NOT, /* 90 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 120 Negative lookbehind */
1457    OP_REVERSE,        /* 91 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 121 Move pointer back - used in lookbehind assertions */
1458    
1459    /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,    /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,
1460    as there's a test for >= ONCE for a subpattern that isn't an assertion. */    as there's a test for >= ONCE for a subpattern that isn't an assertion. */
1461    
1462    OP_ONCE,           /* 92 Atomic group */    OP_ONCE,           /* 122 Atomic group */
1463    OP_BRA,            /* 93 Start of non-capturing bracket */    OP_BRA,            /* 123 Start of non-capturing bracket */
1464    OP_CBRA,           /* 94 Start of capturing bracket */    OP_CBRA,           /* 124 Start of capturing bracket */
1465    OP_COND,           /* 95 Conditional group */    OP_COND,           /* 125 Conditional group */
1466    
1467    /* These three must follow the previous three, in the same order. There's a    /* These three must follow the previous three, in the same order. There's a
1468    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
1469    
1470    OP_SBRA,           /* 96 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 126 Start of non-capturing bracket, check empty  */
1471    OP_SCBRA,          /* 97 Start of capturing bracket, check empty */    OP_SCBRA,          /* 127 Start of capturing bracket, check empty */
1472    OP_SCOND,          /* 98 Conditional group, check empty */    OP_SCOND,          /* 128 Conditional group, check empty */
1473    
1474    OP_CREF,           /* 99 Used to hold a capture number as condition */    /* The next two pairs must (respectively) be kept together. */
1475    OP_RREF,           /* 100 Used to hold a recursion number as condition */  
1476    OP_DEF,            /* 101 The DEFINE condition */    OP_CREF,           /* 129 Used to hold a capture number as condition */
1477      OP_NCREF,          /* 130 Same, but generated by a name reference*/
1478      OP_RREF,           /* 131 Used to hold a recursion number as condition */
1479      OP_NRREF,          /* 132 Same, but generated by a name reference*/
1480      OP_DEF,            /* 133 The DEFINE condition */
1481    
1482    OP_BRAZERO,        /* 102 These two must remain together and in this */    OP_BRAZERO,        /* 134 These two must remain together and in this */
1483    OP_BRAMINZERO,     /* 103 order. */    OP_BRAMINZERO,     /* 135 order. */
1484    
1485    /* These are backtracking control verbs */    /* These are backtracking control verbs */
1486    
1487    OP_PRUNE,          /* 104 */    OP_MARK,           /* 136 always has an argument */
1488    OP_SKIP,           /* 105 */    OP_PRUNE,          /* 137 */
1489    OP_THEN,           /* 106 */    OP_PRUNE_ARG,      /* 138 same, but with argument */
1490    OP_COMMIT,         /* 107 */    OP_SKIP,           /* 139 */
1491      OP_SKIP_ARG,       /* 140 same, but with argument */
1492      OP_THEN,           /* 141 */
1493      OP_THEN_ARG,       /* 142 same, but with argument */
1494      OP_COMMIT,         /* 143 */
1495    
1496    /* These are forced failure and success verbs */    /* These are forced failure and success verbs */
1497    
1498    OP_FAIL,           /* 108 */    OP_FAIL,           /* 144 */
1499    OP_ACCEPT          /* 109 */    OP_ACCEPT,         /* 145 */
1500      OP_CLOSE,          /* 146 Used before OP_ACCEPT to close open captures */
1501    
1502      /* This is used to skip a subpattern with a {0} quantifier */
1503    
1504      OP_SKIPZERO,       /* 147 */
1505    
1506      /* This is not an opcode, but is used to check that tables indexed by opcode
1507      are the correct length, in order to catch updating errors - there have been
1508      some in the past. */
1509    
1510      OP_TABLE_LENGTH
1511  };  };
1512    
1513    /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
1514    definitions that follow must also be updated to match. There are also tables
1515    called "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1516    
1517    
1518  /* 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
1519  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
1520    only in pcre_printint.c, which fills out the full names in many cases (and in
1521    some cases doesn't actually use these names at all). */
1522    
1523  #define OP_NAME_LIST \  #define OP_NAME_LIST \
1524    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
1525    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
1526    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
1527    "extuni",  "\\Z", "\\z",                                        \    "extuni",  "\\Z", "\\z",                                        \
1528    "Opt", "^", "$", "char", "charnc", "not",                       \    "^", "^", "$", "$", "char", "chari", "not", "noti",             \
1529    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??",                                \
1530      "{", "{", "{",                                                  \
1531    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1532    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??",                                \
1533      "{", "{", "{",                                                  \
1534      "*+","++", "?+", "{",                                           \
1535      "*", "*?", "+", "+?", "?", "??",                                \
1536      "{", "{", "{",                                                  \
1537      "*+","++", "?+", "{",                                           \
1538      "*", "*?", "+", "+?", "?", "??",                                \
1539      "{", "{", "{",                                                  \
1540    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1541    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
1542    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1543    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
1544    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Refi",                     \
1545      "Recurse", "Callout",                                           \
1546    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
1547    "AssertB", "AssertB not", "Reverse",                            \    "AssertB", "AssertB not", "Reverse",                            \
1548    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
1549    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \    "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \
1550    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT"    "Brazero", "Braminzero",                                        \
1551      "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \
1552      "*THEN", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",                \
1553      "Close", "Skip zero"
1554    
1555    
1556  /* 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 814  in UTF-8 mode. The code that uses this t Line 1566  in UTF-8 mode. The code that uses this t
1566    1,                             /* End                                    */ \    1,                             /* End                                    */ \
1567    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \
1568    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
1569    1, 1,                          /* Any, Anybyte                           */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
1570    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3,                          /* \P, \p                                 */ \
1571    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
1572    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1,                             /* \X                                     */ \
1573      1, 1, 1, 1, 1, 1,              /* \Z, \z, ^, ^M, $, $M                   */ \
1574    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
1575    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Chari  - the minimum length            */ \
1576    2,                             /* not                                    */ \    2,                             /* not                                    */ \
1577    /* Positive single-char repeats                            ** These are  */ \    2,                             /* noti                                   */ \
1578    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \    /* Positive single-char repeats                             ** These are */ \
1579    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??       ** minima in */ \
1580      4, 4, 4,                       /* upto, minupto, exact      ** mode      */ \
1581    2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \    2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \
1582      2, 2, 2, 2, 2, 2,              /* *I, *?I, +I, +?I, ?I, ??I ** UTF-8     */ \
1583      4, 4, 4,                       /* upto I, minupto I, exact I             */ \
1584      2, 2, 2, 4,                    /* *+I, ++I, ?+I, upto+I                  */ \
1585    /* Negative single-char repeats - only for chars < 256                   */ \    /* Negative single-char repeats - only for chars < 256                   */ \
1586    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \
1587    4, 4, 4,                       /* NOT upto, minupto, exact               */ \    4, 4, 4,                       /* NOT upto, minupto, exact               */ \
1588    2, 2, 2, 4,                    /* Possessive *, +, ?, upto               */ \    2, 2, 2, 4,                    /* Possessive NOT *, +, ?, upto           */ \
1589      2, 2, 2, 2, 2, 2,              /* NOT *I, *?I, +I, +?I, ?I, ??I          */ \
1590      4, 4, 4,                       /* NOT upto I, minupto I, exact I         */ \
1591      2, 2, 2, 4,                    /* Possessive NOT *I, +I, ?I, upto I      */ \
1592    /* Positive type repeats                                                 */ \    /* Positive type repeats                                                 */ \
1593    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \
1594    4, 4, 4,                       /* Type upto, minupto, exact              */ \    4, 4, 4,                       /* Type upto, minupto, exact              */ \
# Line 840  in UTF-8 mode. The code that uses this t Line 1600  in UTF-8 mode. The code that uses this t
1600   33,                             /* NCLASS                                 */ \   33,                             /* NCLASS                                 */ \
1601    0,                             /* XCLASS - variable length               */ \    0,                             /* XCLASS - variable length               */ \
1602    3,                             /* REF                                    */ \    3,                             /* REF                                    */ \
1603      3,                             /* REFI                                   */ \
1604    1+LINK_SIZE,                   /* RECURSE                                */ \    1+LINK_SIZE,                   /* RECURSE                                */ \
1605    2+2*LINK_SIZE,                 /* CALLOUT                                */ \    2+2*LINK_SIZE,                 /* CALLOUT                                */ \
1606    1+LINK_SIZE,                   /* Alt                                    */ \    1+LINK_SIZE,                   /* Alt                                    */ \
# Line 858  in UTF-8 mode. The code that uses this t Line 1619  in UTF-8 mode. The code that uses this t
1619    1+LINK_SIZE,                   /* SBRA                                   */ \    1+LINK_SIZE,                   /* SBRA                                   */ \
1620    3+LINK_SIZE,                   /* SCBRA                                  */ \    3+LINK_SIZE,                   /* SCBRA                                  */ \
1621    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
1622    3,                             /* CREF                                   */ \    3, 3,                          /* CREF, NCREF                            */ \
1623    3,                             /* RREF                                   */ \    3, 3,                          /* RREF, NRREF                            */ \
1624    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
1625    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
1626    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \    3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \
1627    1, 1                           /* FAIL, ACCEPT                           */    1, 3,                          /* SKIP, SKIP_ARG                         */ \
1628      1+LINK_SIZE, 3+LINK_SIZE,      /* THEN, THEN_ARG                         */ \
1629      1, 1, 1, 3, 1                  /* COMMIT, FAIL, ACCEPT, CLOSE, SKIPZERO  */ \
1630    
1631    /* A magic value for OP_RREF and OP_NRREF to indicate the "any recursion"
1632  /* A magic value for OP_RREF to indicate the "any recursion" condition. */  condition. */
1633    
1634  #define RREF_ANY  0xffff  #define RREF_ANY  0xffff
1635    
1636  /* 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
1637  tracked. */  easily be tracked. When a new number is added, the table called eint in
1638    pcreposix.c must be updated. */
1639    
1640  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,
1641         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
# Line 879  enum { ERR0, ERR1, ERR2, ERR3, ERR4, Line 1643  enum { ERR0, ERR1, ERR2, ERR3, ERR4,
1643         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
1644         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
1645         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
1646         ERR60, ERR61, ERR62, ERR63 };         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68,
1647           ERRCOUNT };
1648    
1649  /* 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
1650  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 895  Because people can now save and re-use c Line 1660  Because people can now save and re-use c
1660  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
1661  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
1662  fields are present. Currently PCRE always sets the dummy fields to zero.  fields are present. Currently PCRE always sets the dummy fields to zero.
1663  NOTE NOTE NOTE:  NOTE NOTE NOTE
1664  */  */
1665    
1666  typedef struct real_pcre {  typedef struct real_pcre {
# Line 922  remark (see NOTE above) about extending Line 1687  remark (see NOTE above) about extending
1687    
1688  typedef struct pcre_study_data {  typedef struct pcre_study_data {
1689    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1690    pcre_uint32 options;    pcre_uint32 flags;              /* Private flags */
1691    uschar start_bits[32];    uschar start_bits[32];          /* Starting char bits */
1692      pcre_uint32 minlength;          /* Minimum subject length */
1693  } pcre_study_data;  } pcre_study_data;
1694    
1695    /* Structure for building a chain of open capturing subpatterns during
1696    compiling, so that instructions to close them can be compiled when (*ACCEPT) is
1697    encountered. This is also used to identify subpatterns that contain recursive
1698    back references to themselves, so that they can be made atomic. */
1699    
1700    typedef struct open_capitem {
1701      struct open_capitem *next;    /* Chain link */
1702      pcre_uint16 number;           /* Capture number */
1703      pcre_uint16 flag;             /* Set TRUE if recursive back ref */
1704    } open_capitem;
1705    
1706  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
1707  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
1708    
# Line 938  typedef struct compile_data { Line 1715  typedef struct compile_data {
1715    const uschar *start_code;     /* The start of the compiled code */    const uschar *start_code;     /* The start of the compiled code */
1716    const uschar *start_pattern;  /* The start of the pattern */    const uschar *start_pattern;  /* The start of the pattern */
1717    const uschar *end_pattern;    /* The end of the pattern */    const uschar *end_pattern;    /* The end of the pattern */
1718      open_capitem *open_caps;      /* Chain of open capture items */
1719    uschar *hwm;                  /* High watermark of workspace */    uschar *hwm;                  /* High watermark of workspace */
1720    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
1721    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
# Line 950  typedef struct compile_data { Line 1728  typedef struct compile_data {
1728    int  external_flags;          /* External flag bits to be set */    int  external_flags;          /* External flag bits to be set */
1729    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
1730    BOOL had_accept;              /* (*ACCEPT) encountered */    BOOL had_accept;              /* (*ACCEPT) encountered */
1731      BOOL check_lookbehind;        /* Lookbehinds need later checking */
1732    int  nltype;                  /* Newline type */    int  nltype;                  /* Newline type */
1733    int  nllen;                   /* Newline string length */    int  nllen;                   /* Newline string length */
1734    uschar nl[4];                 /* Newline string when fixed length */    uschar nl[4];                 /* Newline string when fixed length */
# Line 960  branches, for testing for left recursion Line 1739  branches, for testing for left recursion
1739    
1740  typedef struct branch_chain {  typedef struct branch_chain {
1741    struct branch_chain *outer;    struct branch_chain *outer;
1742    uschar *current;    uschar *current_branch;
1743  } branch_chain;  } branch_chain;
1744    
1745  /* 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 970  typedef struct recursion_info { Line 1749  typedef struct recursion_info {
1749    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
1750    int group_num;                /* Number of group that was called */    int group_num;                /* Number of group that was called */
1751    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 */  
1752    int *offset_save;             /* Pointer to start of saved offsets */    int *offset_save;             /* Pointer to start of saved offsets */
1753    int saved_max;                /* Number of saved offsets */    int saved_max;                /* Number of saved offsets */
1754      int save_offset_top;          /* Current value of offset_top */
1755  } recursion_info;  } recursion_info;
1756    
1757  /* 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 997  typedef struct match_data { Line 1776  typedef struct match_data {
1776    int    offset_max;            /* The maximum usable for return data */    int    offset_max;            /* The maximum usable for return data */
1777    int    nltype;                /* Newline type */    int    nltype;                /* Newline type */
1778    int    nllen;                 /* Newline string length */    int    nllen;                 /* Newline string length */
1779      int    name_count;            /* Number of names in name table */
1780      int    name_entry_size;       /* Size of entry in names table */
1781      uschar *name_table;           /* Table of names */
1782    uschar nl[4];                 /* Newline string when fixed */    uschar nl[4];                 /* Newline string when fixed */
1783    const uschar *lcc;            /* Points to lower casing table */    const uschar *lcc;            /* Points to lower casing table */
1784    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
# Line 1004  typedef struct match_data { Line 1786  typedef struct match_data {
1786    BOOL   notbol;                /* NOTBOL flag */    BOOL   notbol;                /* NOTBOL flag */
1787    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                /* NOTEOL flag */
1788    BOOL   utf8;                  /* UTF8 flag */    BOOL   utf8;                  /* UTF8 flag */
1789      BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */
1790      BOOL   use_ucp;               /* PCRE_UCP flag */
1791    BOOL   endonly;               /* Dollar not before final \n */    BOOL   endonly;               /* Dollar not before final \n */
1792    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notempty;              /* Empty string match not wanted */
1793    BOOL   partial;               /* PARTIAL flag */    BOOL   notempty_atstart;      /* Empty string match at start not wanted */
1794    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
1795    BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */    BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */
1796    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
# Line 1014  typedef struct match_data { Line 1798  typedef struct match_data {
1798    USPTR  end_subject;           /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
1799    USPTR  start_match_ptr;       /* Start of matched string */    USPTR  start_match_ptr;       /* Start of matched string */
1800    USPTR  end_match_ptr;         /* Subject position at end match */    USPTR  end_match_ptr;         /* Subject position at end match */
1801      USPTR  start_used_ptr;        /* Earliest consulted character */
1802      int    partial;               /* PARTIAL options */
1803    int    end_offset_top;        /* Highwater mark at end of match */    int    end_offset_top;        /* Highwater mark at end of match */
1804    int    capture_last;          /* Most recent capture number */    int    capture_last;          /* Most recent capture number */
1805    int    start_offset;          /* The start offset value */    int    start_offset;          /* The start offset value */
# Line 1021  typedef struct match_data { Line 1807  typedef struct match_data {
1807    int    eptrn;                 /* Next free eptrblock */    int    eptrn;                 /* Next free eptrblock */
1808    recursion_info *recursive;    /* Linked list of recursion data */    recursion_info *recursive;    /* Linked list of recursion data */
1809    void  *callout_data;          /* To pass back to callouts */    void  *callout_data;          /* To pass back to callouts */
1810      const uschar *mark;           /* Mark pointer to pass back */
1811  } match_data;  } match_data;
1812    
1813  /* 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 1030  typedef struct dfa_match_data { Line 1817  typedef struct dfa_match_data {
1817    const uschar *start_code;     /* Start of the compiled pattern */    const uschar *start_code;     /* Start of the compiled pattern */
1818    const uschar *start_subject;  /* Start of the subject string */    const uschar *start_subject;  /* Start of the subject string */
1819    const uschar *end_subject;    /* End of subject string */    const uschar *end_subject;    /* End of subject string */
1820      const uschar *start_used_ptr; /* Earliest consulted character */
1821    const uschar *tables;         /* Character tables */    const uschar *tables;         /* Character tables */
1822      int   start_offset;           /* The start offset value */
1823    int   moptions;               /* Match options */    int   moptions;               /* Match options */
1824    int   poptions;               /* Pattern options */    int   poptions;               /* Pattern options */
1825    int    nltype;                /* Newline type */    int    nltype;                /* Newline type */
# Line 1109  extern const uschar _pcre_OP_lengths[]; Line 1898  extern const uschar _pcre_OP_lengths[];
1898  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
1899  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1900    
1901  extern BOOL         _pcre_is_newline(const uschar *, int, const uschar *,  extern const uschar *_pcre_find_bracket(const uschar *, BOOL, int);
1902                        int *, BOOL);  extern BOOL          _pcre_is_newline(USPTR, int, USPTR, int *, BOOL);
1903  extern int          _pcre_ord2utf8(int, uschar *);  extern int           _pcre_ord2utf8(int, uschar *);
1904  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  extern real_pcre    *_pcre_try_flipped(const real_pcre *, real_pcre *,
1905                        const pcre_study_data *, pcre_study_data *);                         const pcre_study_data *, pcre_study_data *);
1906  extern int          _pcre_ucp_findprop(const unsigned int, int *, int *);  extern int           _pcre_valid_utf8(USPTR, int, int *);
1907  extern unsigned int _pcre_ucp_othercase(const unsigned int);  extern BOOL          _pcre_was_newline(USPTR, int, USPTR, int *, BOOL);
1908  extern int          _pcre_valid_utf8(const uschar *, int);  extern BOOL          _pcre_xclass(int, const uschar *);
1909  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,  
1910                        int *, BOOL);  
1911  extern BOOL         _pcre_xclass(int, const uschar *);  /* Unicode character database (UCD) */
1912    
1913    typedef struct {
1914      uschar script;
1915      uschar chartype;
1916      pcre_int32 other_case;
1917    } ucd_record;
1918    
1919    extern const ucd_record  _pcre_ucd_records[];
1920    extern const uschar      _pcre_ucd_stage1[];
1921    extern const pcre_uint16 _pcre_ucd_stage2[];
1922    extern const int         _pcre_ucp_gentype[];
1923    
1924    
1925    /* UCD access macros */
1926    
1927    #define UCD_BLOCK_SIZE 128
1928    #define GET_UCD(ch) (_pcre_ucd_records + \
1929            _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \
1930            UCD_BLOCK_SIZE + (ch) % UCD_BLOCK_SIZE])
1931    
1932    #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
1933    #define UCD_SCRIPT(ch)    GET_UCD(ch)->script
1934    #define UCD_CATEGORY(ch)  _pcre_ucp_gentype[UCD_CHARTYPE(ch)]
1935    #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
1936    
1937  #endif  #endif
1938    

Legend:
Removed from v.298  
changed lines
  Added in v.602

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12